[2023-11-24 22:50:59,105][root][INFO] - args.local_rank 1 [2023-11-24 22:50:59,105][root][INFO] - WORLD_SIZE 4 [2023-11-24 22:50:59,126][root][INFO] - args.local_rank 3 [2023-11-24 22:50:59,126][root][INFO] - WORLD_SIZE 4 [2023-11-24 22:50:59,132][root][INFO] - args.local_rank 2 [2023-11-24 22:50:59,132][root][INFO] - WORLD_SIZE 4 [2023-11-24 22:50:59,159][root][INFO] - args.local_rank 0 [2023-11-24 22:50:59,159][root][INFO] - WORLD_SIZE 4 [2023-11-24 22:51:00,049][torch.distributed.distributed_c10d][INFO] - Added key: store_based_barrier_key:1 to store for rank: 3 [2023-11-24 22:51:00,052][torch.distributed.distributed_c10d][INFO] - Added key: store_based_barrier_key:1 to store for rank: 2 [2023-11-24 22:51:00,055][torch.distributed.distributed_c10d][INFO] - Added key: store_based_barrier_key:1 to store for rank: 1 [2023-11-24 22:51:00,057][torch.distributed.distributed_c10d][INFO] - Added key: store_based_barrier_key:1 to store for rank: 0 [2023-11-24 22:51:00,057][torch.distributed.distributed_c10d][INFO] - Rank 0: Completed store-based barrier for key:store_based_barrier_key:1 with 4 nodes. [2023-11-24 22:51:00,057][root][INFO] - Initialized host lccpu28.cse.ust.hk as d.rank 0 on device=cuda:0, n_gpu=1, world size=4 [2023-11-24 22:51:00,057][root][INFO] - 16-bits training: True [2023-11-24 22:51:00,059][root][INFO] - CFG (after gpu configuration): [2023-11-24 22:51:00,059][torch.distributed.distributed_c10d][INFO] - Rank 3: Completed store-based barrier for key:store_based_barrier_key:1 with 4 nodes. [2023-11-24 22:51:00,060][root][INFO] - Initialized host lccpu28.cse.ust.hk as d.rank 3 on device=cuda:3, n_gpu=1, world size=4 [2023-11-24 22:51:00,060][root][INFO] - 16-bits training: True [2023-11-24 22:51:00,061][root][INFO] - train_datasets: ['dl', 'cm'] [2023-11-24 22:51:00,062][torch.distributed.distributed_c10d][INFO] - Rank 2: Completed store-based barrier for key:store_based_barrier_key:1 with 4 nodes. [2023-11-24 22:51:00,062][root][INFO] - Initialized host lccpu28.cse.ust.hk as d.rank 2 on device=cuda:2, n_gpu=1, world size=4 [2023-11-24 22:51:00,063][root][INFO] - 16-bits training: True [2023-11-24 22:51:00,064][root][INFO] - train_datasets: ['dl', 'cm'] [2023-11-24 22:51:00,064][root][INFO] - biencoder: alias: vdr sequence_length: 512 dropout: 0.1 shared_encoder: false semiparametric: true device: null encoder_q: base_model_id: bert-base-uncased max_seq_len: 128 pretrained: true norm: false shift_vocab_num: 1000 prefix: encoder_q. encoder_p: base_model_id: bert-base-uncased max_seq_len: 256 pretrained: true norm: false shift_vocab_num: 1000 prefix: encoder_p. datasets: nq_train: _target_: src.data.biencoder_dataset.BiEncoderDataset file: ${HOME}/data/train/nq-train.jsonl trivia_train: _target_: src.data.biencoder_dataset.BiEncoderDataset file: ${HOME}/data/train/trivia-train.jsonl webq_train: _target_: src.data.biencoder_dataset.BiEncoderDataset file: ${HOME}/data/train/webq-train.jsonl dl: _target_: src.data.biencoder_dataset.BiEncoderDataset file: ${HOME}/data/train/dl_10m.jsonl cm: _target_: src.data.biencoder_dataset.BiEncoderDataset file: ${HOME}/data/train/cm_10m.jsonl marco_dev: _target_: src.data.biencoder_dataset.BiEncoderDataset file: ${HOME}/data/train/msmarco-dev.jsonl marco_dev_30neg: _target_: src.data.biencoder_dataset.BiEncoderDataset file: ${HOME}/data/train/msmarco-dev-30neg.jsonl marco_train_bm25: _target_: src.data.biencoder_dataset.BiEncoderDataset file: ${HOME}/data/train/marco_bm25_20230119.jsonl marco_train_1000: _target_: src.data.biencoder_dataset.BiEncoderDataset file: ${HOME}/data/train/marco_train_1000.jsonl HOME: /export/data/jzhoubu/workspace/VDR-dense train_datasets: - dl - cm dev_datasets: - marco_dev_30neg datastore: null output_dir: null train_sampling_rates: null batch_size: 256 num_train_epochs: 5 num_warmup_epochs: 1 num_save_epochs: 1 num_eval_epochs: 1 hard_negatives: 0 other_negatives: 0 ret_negatives: 0 train_insert_title: true valid_insert_title: false adam_eps: 1.0e-08 adam_betas: (0.9, 0.999) learning_rate: 2.0e-05 max_grad_norm: 2.0 log_batch_step: 100 train_rolling_loss_step: 100 weight_decay: 0.0 sym_loss: true do_lower_case: true seed: 12345 checkpoint_file_name: vdr model_file: null local_rank: 0 local-rank: 0 device: cuda:0 distributed_world_size: 4 distributed_port: null no_cuda: false n_gpu: 1 fp16: true [2023-11-24 22:51:00,064][root][INFO] - dev_datasets: ['marco_dev_30neg'] [2023-11-24 22:51:00,064][root][INFO] - train_datasets: ['dl', 'cm'] [2023-11-24 22:51:00,065][root][INFO] - ***** Initializing components for training ***** [2023-11-24 22:51:00,065][torch.distributed.distributed_c10d][INFO] - Rank 1: Completed store-based barrier for key:store_based_barrier_key:1 with 4 nodes. [2023-11-24 22:51:00,065][root][INFO] - Initialized host lccpu28.cse.ust.hk as d.rank 1 on device=cuda:1, n_gpu=1, world size=4 [2023-11-24 22:51:00,066][root][INFO] - 16-bits training: True [2023-11-24 22:51:00,066][root][INFO] - dev_datasets: ['marco_dev_30neg'] [2023-11-24 22:51:00,067][root][INFO] - train_datasets: ['dl', 'cm'] [2023-11-24 22:51:00,067][root][INFO] - dev_datasets: ['marco_dev_30neg'] [2023-11-24 22:51:00,067][root][INFO] - ***** Initializing components for training ***** [2023-11-24 22:51:00,068][root][INFO] - ***** Initializing components for training ***** [2023-11-24 22:51:00,069][root][INFO] - dev_datasets: ['marco_dev_30neg'] [2023-11-24 22:51:00,070][root][INFO] - ***** Initializing components for training ***** [2023-11-24 22:51:03,042][root][INFO] - model.embedding_q.sum: -656919.625 [2023-11-24 22:51:03,045][root][INFO] - model.embedding_p.sum: -656919.625 [2023-11-24 22:51:05,020][root][INFO] - Initializing task/set data ['dl', 'cm'] [2023-11-24 22:51:05,020][root][INFO] - Initializing task/set data ['dl', 'cm'] [2023-11-24 22:51:05,026][root][INFO] - Initializing task/set data ['dl', 'cm'] [2023-11-24 22:51:05,030][root][INFO] - Initializing task/set data ['dl', 'cm'] [2023-11-24 22:54:40,327][root][INFO] - Load data size: 10000000 [2023-11-24 22:54:40,444][root][INFO] - Load data size: 10000000 [2023-11-24 22:54:45,855][root][INFO] - Load data size: 10000000 [2023-11-24 22:54:52,998][root][INFO] - Load data size: 10000000 [2023-11-24 22:57:51,942][root][INFO] - Load data size: 10000000 [2023-11-24 22:57:58,127][root][INFO] - Load data size: 10000000 [2023-11-24 22:57:58,861][root][INFO] - samples_per_shard=2500000, shard_start_idx=5000000, shard_end_idx=7500000, max_iterations=9765 [2023-11-24 22:57:58,862][root][INFO] - samples_per_shard=2500000, shard_start_idx=5000000, shard_end_idx=7500000, max_iterations=9765 [2023-11-24 22:57:58,862][root][INFO] - rank=2; Multi set data sizes [10000000, 10000000] [2023-11-24 22:57:58,862][root][INFO] - rank=2; Multi set total data 20000000 [2023-11-24 22:57:58,862][root][INFO] - rank=2; Multi set sampling_rates None [2023-11-24 22:57:58,862][root][INFO] - rank=2; Multi set max_iterations per dataset [9765, 9765] [2023-11-24 22:57:58,862][root][INFO] - rank=2; Multi set max_iterations 19530 [2023-11-24 22:57:58,862][root][INFO] - Total iterations per epoch=19530 [2023-11-24 22:57:58,862][root][INFO] - Total updates=97650 [2023-11-24 22:57:58,862][root][INFO] - Warmup updates=19530 [2023-11-24 22:57:58,862][root][INFO] - Eval step = 19530 [2023-11-24 22:57:58,862][root][INFO] - ***** Training ***** [2023-11-24 22:58:06,112][root][INFO] - samples_per_shard=2500000, shard_start_idx=0, shard_end_idx=2500000, max_iterations=9765 [2023-11-24 22:58:06,112][root][INFO] - samples_per_shard=2500000, shard_start_idx=0, shard_end_idx=2500000, max_iterations=9765 [2023-11-24 22:58:06,112][root][INFO] - rank=0; Multi set data sizes [10000000, 10000000] [2023-11-24 22:58:06,112][root][INFO] - rank=0; Multi set total data 20000000 [2023-11-24 22:58:06,112][root][INFO] - rank=0; Multi set sampling_rates None [2023-11-24 22:58:06,112][root][INFO] - rank=0; Multi set max_iterations per dataset [9765, 9765] [2023-11-24 22:58:06,113][root][INFO] - rank=0; Multi set max_iterations 19530 [2023-11-24 22:58:06,113][root][INFO] - Total iterations per epoch=19530 [2023-11-24 22:58:06,113][root][INFO] - Total updates=97650 [2023-11-24 22:58:06,113][root][INFO] - Warmup updates=19530 [2023-11-24 22:58:06,113][root][INFO] - Eval step = 19530 [2023-11-24 22:58:06,113][root][INFO] - ***** Training ***** [2023-11-24 22:58:07,598][root][INFO] - Saved checkpoint at ./vdr_0 [2023-11-24 22:58:07,599][root][INFO] - ***** Epoch 1 ***** [2023-11-24 22:58:07,600][root][INFO] - rank=2; Iteration start [2023-11-24 22:58:07,600][root][INFO] - rank=2; Multi set iteration: iteration ptr per set: [0, 0] [2023-11-24 22:58:07,600][root][INFO] - rank=2; Multi set iteration: source 0, batches to be taken: 9765 [2023-11-24 22:58:07,600][root][INFO] - rank=2; Multi set iteration: source 1, batches to be taken: 9765 [2023-11-24 22:58:07,606][root][INFO] - rank=2; data_src_indices len=19530 [2023-11-24 22:58:10,117][root][INFO] - Load data size: 10000000 [2023-11-24 22:58:16,627][root][INFO] - Saved checkpoint at ./vdr_0 [2023-11-24 22:58:16,627][root][INFO] - ***** Epoch 1 ***** [2023-11-24 22:58:16,629][root][INFO] - rank=0; Iteration start [2023-11-24 22:58:16,629][root][INFO] - rank=0; Multi set iteration: iteration ptr per set: [0, 0] [2023-11-24 22:58:16,629][root][INFO] - rank=0; Multi set iteration: source 0, batches to be taken: 9765 [2023-11-24 22:58:16,629][root][INFO] - rank=0; Multi set iteration: source 1, batches to be taken: 9765 [2023-11-24 22:58:16,634][root][INFO] - rank=0; data_src_indices len=19530 [2023-11-24 22:58:17,326][root][INFO] - samples_per_shard=2500000, shard_start_idx=7500000, shard_end_idx=10000000, max_iterations=9765 [2023-11-24 22:58:17,326][root][INFO] - samples_per_shard=2500000, shard_start_idx=7500000, shard_end_idx=10000000, max_iterations=9765 [2023-11-24 22:58:17,326][root][INFO] - rank=3; Multi set data sizes [10000000, 10000000] [2023-11-24 22:58:17,326][root][INFO] - rank=3; Multi set total data 20000000 [2023-11-24 22:58:17,326][root][INFO] - rank=3; Multi set sampling_rates None [2023-11-24 22:58:17,326][root][INFO] - rank=3; Multi set max_iterations per dataset [9765, 9765] [2023-11-24 22:58:17,326][root][INFO] - rank=3; Multi set max_iterations 19530 [2023-11-24 22:58:17,327][root][INFO] - Total iterations per epoch=19530 [2023-11-24 22:58:17,327][root][INFO] - Total updates=97650 [2023-11-24 22:58:17,327][root][INFO] - Warmup updates=19530 [2023-11-24 22:58:17,327][root][INFO] - Eval step = 19530 [2023-11-24 22:58:17,327][root][INFO] - ***** Training ***** [2023-11-24 22:58:18,039][root][INFO] - Load data size: 10000000 [2023-11-24 22:58:25,598][root][INFO] - samples_per_shard=2500000, shard_start_idx=2500000, shard_end_idx=5000000, max_iterations=9765 [2023-11-24 22:58:25,598][root][INFO] - samples_per_shard=2500000, shard_start_idx=2500000, shard_end_idx=5000000, max_iterations=9765 [2023-11-24 22:58:25,598][root][INFO] - rank=1; Multi set data sizes [10000000, 10000000] [2023-11-24 22:58:25,598][root][INFO] - rank=1; Multi set total data 20000000 [2023-11-24 22:58:25,598][root][INFO] - rank=1; Multi set sampling_rates None [2023-11-24 22:58:25,598][root][INFO] - rank=1; Multi set max_iterations per dataset [9765, 9765] [2023-11-24 22:58:25,598][root][INFO] - rank=1; Multi set max_iterations 19530 [2023-11-24 22:58:25,598][root][INFO] - Total iterations per epoch=19530 [2023-11-24 22:58:25,598][root][INFO] - Total updates=97650 [2023-11-24 22:58:25,598][root][INFO] - Warmup updates=19530 [2023-11-24 22:58:25,598][root][INFO] - Eval step = 19530 [2023-11-24 22:58:25,598][root][INFO] - ***** Training ***** [2023-11-24 22:58:26,301][root][INFO] - Saved checkpoint at ./vdr_0 [2023-11-24 22:58:26,301][root][INFO] - ***** Epoch 1 ***** [2023-11-24 22:58:26,303][root][INFO] - rank=3; Iteration start [2023-11-24 22:58:26,303][root][INFO] - rank=3; Multi set iteration: iteration ptr per set: [0, 0] [2023-11-24 22:58:26,303][root][INFO] - rank=3; Multi set iteration: source 0, batches to be taken: 9765 [2023-11-24 22:58:26,303][root][INFO] - rank=3; Multi set iteration: source 1, batches to be taken: 9765 [2023-11-24 22:58:26,308][root][INFO] - rank=3; data_src_indices len=19530 [2023-11-24 22:58:34,736][root][INFO] - Saved checkpoint at ./vdr_0 [2023-11-24 22:58:34,737][root][INFO] - ***** Epoch 1 ***** [2023-11-24 22:58:34,738][root][INFO] - rank=1; Iteration start [2023-11-24 22:58:34,738][root][INFO] - rank=1; Multi set iteration: iteration ptr per set: [0, 0] [2023-11-24 22:58:34,738][root][INFO] - rank=1; Multi set iteration: source 0, batches to be taken: 9765 [2023-11-24 22:58:34,738][root][INFO] - rank=1; Multi set iteration: source 1, batches to be taken: 9765 [2023-11-24 22:58:34,744][root][INFO] - rank=1; data_src_indices len=19530 [2023-11-24 22:58:41,008][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 2.80 | max: 9.68 | min: 0.43 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 3.24 | max: 11.01 | min: 0.79 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.5/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: they also worked on other projects, such as " z - force " at six flags great america. ### ### [P_TEXT]: six flags america at six flags america, thomas town was renamed " whistlestop park " and ### ### tony hawk's halfpipe slide was renamed " halfpipe ". for 2012, the park added apocalypse, a stand - ### ### up roller coaster from bolliger & mabillard, which featured two inversions and a ten - story drop. ### ### the roller coaster had previously operated as iron wolf at six flags great america. in 2013, the ### ### park added the six - slide complex " bonzai pipelines " to hurricane harbor. in 2014, six flags re ### ### - themed a section of the park to feature mardi gras. the area would feature a new roller ### ### =============================================== v_q ================================================ ### ### (',', 0, 14) ('and', 1, 12) ('the', 2, 19) ('.', 3, 21) ('project', 4, 5658) ('projects', 5, 8382) ### ### ('of', 6, 30) ('flags', 7, 11) ('z', 8, 10648) ('a', 9, 36) ('in', 10, 48) ('to', 11, 34) ### ### ('-', 12, 44) ('worked', 13, 536) ('was', 14, 86) ('with', 15, 60) ('##lusion', 16, 433) ### ### ('##tooth', 17, 5170) ('that', 18, 77) ('##media', 19, 6740) ### ### =============================================== v_p ================================================ ### ### ('coaster', 14182, 0) ('whistle', 2051, 1) ('inversion', 11932, 2) ('apocalypse', 9641, 3) ### ### ('hawk', 1435, 4) ('roller', 10456, 5) ('bon', 18935, 6) ('hurricane', 27341, 7) ('##sto', 1465, 8) ### ### ('##pipe', 1350, 9) ('slide', 29276, 10) ('flags', 7, 11) ('and', 1, 12) ('america', 44, 13) ### ### (',', 0, 14) ('##lli', 7589, 15) ('park', 8725, 16) ('##di', 700, 17) ('##zai', 5714, 18) ### ### ('the', 2, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('and', 1, 12) (',', 0, 14) ('the', 2, 19) ('flags', 7, 11) ('.', 3, 21) ('of', 6, 30) ('a', 9, 36) ### ### ('to', 11, 34) ('america', 44, 13) ('in', 10, 48) ('-', 12, 44) ('whistle', 2051, 1) ### ### ('with', 15, 60) ('is', 20, 55) ('hawk', 1435, 4) ('or', 23, 65) ('was', 14, 86) ('that', 18, 77) ### ### ('his', 22, 78) ('an', 33, 64) ### ############################################################################################################ [2023-11-24 22:58:41,583][root][INFO] - Epoch: 1: Step: 1/19530, loss[v]=695.431030, lr=0.000000, acc@1[1]=385.0/1024=0.3759765625, acc@1[2]=200.5/1024=0.19580078125 [2023-11-24 23:00:35,725][root][INFO] - Train batch 100 [2023-11-24 23:00:35,726][root][INFO] - Avg. loss per last 100 batches: 627.430166 [2023-11-24 23:00:36,451][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 2.79 | max: 9.70 | min: 0.43 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 3.21 | max: 11.03 | min: 0.77 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: marmots are relatively large ground squirrel s in the genus marmota, with 15 species ### ### living in asia, europe and north america. ### ### [P_TEXT]: afghan snowfinch burrows and hollows made by rodents, including ground squirrel s, marmot ### ### s, and in one recorded case williams'jerboa ( " allactaga williamsi " ). in particular, it is ### ### associated with the yellow ground squirrel ( " spermophilus fulvus " ). several of the other ### ### southern snowfinch species have similar nesting preferences, as does at least one species that ### ### occurs in the afghan snowfinch's range, the isabelline wheatear. the materials used to line the ### ### burrows it nests in are the hairs of squirrels, sheep, and dromedaries ; and feather s. ### ### =============================================== v_q ================================================ ### ### ('##mot', 0, 12) ('asia', 1, 10719) ('europe', 2, 20884) ('s', 3, 16) ('squirrel', 4, 13) ### ### (',', 5, 20) ('genus', 6, 4441) ('.', 7, 23) ('asian', 8, 593) ('of', 9, 41) ('and', 10, 22) ### ### ('species', 11, 31) ('to', 12, 37) ('squirrels', 13, 2) ('the', 14, 27) ('large', 15, 11685) ### ### ('america', 16, 23565) ('americans', 17, 15503) ('a', 18, 33) ('european', 19, 2164) ### ### =============================================== v_p ================================================ ### ### ('##cta', 190, 0) ('nesting', 2831, 1) ('squirrels', 13, 2) ('sperm', 3528, 3) ('afghan', 23587, 4) ### ### ('##dar', 11011, 5) ('hairs', 17106, 6) ('##rb', 3099, 7) ('feather', 4889, 8) ('##hil', 3133, 9) ### ### ('wheat', 2080, 10) ('isabel', 22133, 11) ('##mot', 0, 12) ('squirrel', 4, 13) ('range', 2563, 14) ### ### ('snow', 19471, 15) ('s', 3, 16) ('hollow', 26025, 17) ('alla', 18859, 18) ('sheep', 8793, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##mot', 0, 12) ('squirrels', 13, 2) ('s', 3, 16) ('squirrel', 4, 13) (',', 5, 20) ('.', 7, 23) ### ### ('and', 10, 22) ('##cta', 190, 0) ('the', 14, 27) ('species', 11, 31) ('a', 18, 33) ('to', 12, 37) ### ### ('of', 9, 41) ('in', 22, 47) ('-', 24, 55) ('mar', 20, 85) (';', 68, 36) ('was', 30, 56) ### ### ('rodents', 252, 21) ('nesting', 2831, 1) ### ############################################################################################################ [2023-11-24 23:00:36,957][root][INFO] - Epoch: 1: Step: 101/19530, loss[v]=645.869690, lr=0.000000, acc@1[1]=399.5/1024=0.39013671875, acc@1[2]=181.0/1024=0.1767578125 [2023-11-24 23:02:27,456][root][INFO] - Train batch 200 [2023-11-24 23:02:27,457][root][INFO] - Avg. loss per last 100 batches: 534.616008 [2023-11-24 23:02:28,165][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 2.71 | max: 9.67 | min: 0.40 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 3.07 | max: 10.94 | min: 0.69 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: a medley of songs originally performed by bobby brown — " don't be cruel ", " every ### ### little step ", " rock wit'cha " and " roni " — is also featured on the album, ### ### [P_TEXT]: usher ( musician ) cnn, " the first record that i ever bought was bobby brown's " don't ### ### be cruel " ", and that he admired him " just for being an edgy artist, being a true representative ### ### of r & b and rhythmic music ". usher's " live " album would feature a medley performance of brown's ### ### original singles - " don't be cruel ", " every little step ", " rock wit'cha ", and " roni ". usher ### ### was inspired by frank sinatra's " my way " album ### ### =============================================== v_q ================================================ ### ### ('ron', 0, 11) (',', 1, 8) ('the', 2, 10) ('cruel', 3, 6) ('and', 4, 12) ('.', 5, 9) ### ### ('brown', 6, 3) ('album', 7, 13) ('every', 8, 21) ('to', 9, 15) ('of', 10, 18) ('a', 11, 23) ### ### ('was', 12, 49) ('medley', 13, 2) ('in', 14, 28) ('rock', 15, 36) ('or', 16, 57) ('is', 17, 78) ### ### ('-', 18, 24) ('step', 19, 5) ### ### =============================================== v_p ================================================ ### ### ('usher', 18156, 0) ('sinatra', 3113, 1) ('medley', 13, 2) ('brown', 6, 3) ('way', 9254, 4) ### ### ('step', 19, 5) ('cruel', 3, 6) ('live', 3493, 7) (',', 1, 8) ('.', 5, 9) ('the', 2, 10) ### ### ('ron', 0, 11) ('and', 4, 12) ('album', 7, 13) ('ed', 2295, 14) ('to', 9, 15) ('cha', 78, 16) ### ### ('rhythmic', 5016, 17) ('of', 10, 18) ('ever', 1117, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ron', 0, 11) ('cruel', 3, 6) (',', 1, 8) ('the', 2, 10) ('brown', 6, 3) ('.', 5, 9) ### ### ('and', 4, 12) ('album', 7, 13) ('medley', 13, 2) ('to', 9, 15) ('every', 8, 21) ('step', 19, 5) ### ### ('of', 10, 18) ('a', 11, 23) ('-', 18, 24) ('in', 14, 28) ('rock', 15, 36) ('was', 12, 49) ### ### ('cha', 78, 16) ('bobby', 28, 27) ### ############################################################################################################ [2023-11-24 23:02:28,673][root][INFO] - Epoch: 1: Step: 201/19530, loss[v]=444.419128, lr=0.000000, acc@1[1]=400.5/1024=0.39111328125, acc@1[2]=298.5/1024=0.29150390625 [2023-11-24 23:04:19,245][root][INFO] - Train batch 300 [2023-11-24 23:04:19,246][root][INFO] - Avg. loss per last 100 batches: 378.172511 [2023-11-24 23:04:19,985][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 2.62 | max: 9.74 | min: 0.37 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 2.85 | max: 10.75 | min: 0.56 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: northwest side is the location of the main campus of the university of texas at san ### ### antonio, the university of texas health science center at san antonio, and the northwest campus of ### ### the university of the incarnate word, which includes the rosenberg ### ### [P_TEXT]: gayle benson in san antonio. on october 22, 2014, gayle benson and her husband ### ### commemorated their tenth anniversary by renewing their wedding vows at st. louis cathedral in new ### ### orleans. while married, gayle and tom benson funded the construction of the gayle and tom benson ### ### stadium at the university of the incarnate word in san antonio, which opened on campus on september ### ### 1, 2008. in january 2012, benson and her husband were awarded the " pro ecclesia et pontifice " for ### ### their service to catholic church by pope benedict xvi. in november 2012, gayle benson and her ### ### husband, tom, donated $ 7. 5 million towards the construction of ### ### =============================================== v_q ================================================ ### ### ('northwest', 0, 11436) ('texas', 1, 495) (',', 2, 5) ('and', 3, 9) ('inca', 4, 51) ('.', 5, 4) ### ### ('of', 6, 20) ('the', 7, 6) ('health', 8, 27135) ('university', 9, 27) ('antonio', 10, 22) ### ### ('side', 11, 18840) ('word', 12, 29) ('to', 13, 14) ('a', 14, 15) ('campus', 15, 64) ('-', 16, 62) ### ### ('southwest', 17, 14327) ('in', 18, 32) ('san', 19, 30) ### ### =============================================== v_p ================================================ ### ### ('anniversary', 373, 0) ('orleans', 5739, 1) ('benson', 16867, 2) ('gay', 17460, 3) ('.', 5, 4) ### ### (',', 2, 5) ('the', 7, 6) ('benedict', 4394, 7) ('cathedral', 13146, 8) ('and', 3, 9) ### ### ('##le', 349, 10) ('tom', 18199, 11) ('louis', 26696, 12) ('vows', 24053, 13) ('to', 13, 14) ### ### ('a', 14, 15) ('stadium', 12001, 16) ('##cles', 9962, 17) ('awarded', 20765, 18) ### ### ('construction', 6535, 19) ### ### ============================================ v_q * v_p ============================================= ### ### (',', 2, 5) ('.', 5, 4) ('and', 3, 9) ('the', 7, 6) ('of', 6, 20) ('to', 13, 14) ('a', 14, 15) ### ### ('antonio', 10, 22) ('university', 9, 27) ('word', 12, 29) ('inca', 4, 51) ('anniversary', 373, 0) ### ### ('san', 19, 30) ('in', 18, 32) ('-', 16, 62) ('campus', 15, 64) ('"', 25, 41) ('that', 21, 44) ### ### ('an', 26, 47) ('texas', 1, 495) ### ############################################################################################################ [2023-11-24 23:04:20,494][root][INFO] - Epoch: 1: Step: 301/19530, loss[v]=324.638947, lr=0.000000, acc@1[1]=394.5/1024=0.38525390625, acc@1[2]=367.0/1024=0.3583984375 [2023-11-24 23:06:11,371][root][INFO] - Train batch 400 [2023-11-24 23:06:11,372][root][INFO] - Avg. loss per last 100 batches: 246.364684 [2023-11-24 23:06:12,110][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 2.49 | max: 9.59 | min: 0.33 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 2.59 | max: 10.59 | min: 0.42 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: elkins, and ohio valley ) accepted invitations to join the great midwest athletic ### ### conference. ### ### [P_TEXT]: salem university been a charter member of the west virginia intercollegiate athletic ### ### conference ( wviac ) in 1924, but left that conference in 2010. since that time, the salem tigers ### ### have been members of the eastern college athletic conference in 17 men's and women's sports ### ### including basketball, baseball, softball, soccer, tennis, water polo, swimming, cross country, ### ### volleyball, and track. salem is one of four west virginia schools that joined the great midwest ### ### athletic conference ( g - mac ) in july 2013. the other three schools were former wviac colleagues ### ### of salem that had been cast adrift in the breakup of the ### ### =============================================== v_q ================================================ ### ### ('midwest', 0, 19) ('ohio', 1, 1070) (',', 2, 10) ('elk', 3, 12191) ('.', 4, 6) ('the', 5, 13) ### ### ('and', 6, 15) ('to', 7, 25) ('athletic', 8, 11) ('conference', 9, 21) ('a', 10, 24) ('(', 11, 38) ### ### ('in', 12, 27) (')', 13, 106) ('conferences', 14, 49) ('of', 15, 33) ('invitations', 16, 27876) ### ### ('##ins', 17, 3321) ('or', 18, 104) ('that', 19, 44) ### ### =============================================== v_p ================================================ ### ### ('softball', 20067, 0) ('soccer', 13799, 1) ('swimming', 16182, 2) ('salem', 23102, 3) ### ### ('baseball', 14878, 4) ('volleyball', 6526, 5) ('.', 4, 6) ('three', 6406, 7) ('tennis', 25449, 8) ### ### ('##via', 2265, 9) (',', 2, 10) ('athletic', 8, 11) ('basketball', 17406, 12) ('the', 5, 13) ### ### ('eastern', 4206, 14) ('and', 6, 15) ('mac', 8473, 16) ('intercollegiate', 1780, 17) ### ### ('2013', 15552, 18) ('midwest', 0, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('midwest', 0, 19) ('.', 4, 6) (',', 2, 10) ('the', 5, 13) ('and', 6, 15) ('athletic', 8, 11) ### ### ('conference', 9, 21) ('to', 7, 25) ('a', 10, 24) ('in', 12, 27) ('(', 11, 38) ('of', 15, 33) ### ### ('conferences', 14, 49) ('that', 19, 44) ('-', 25, 32) ('is', 20, 54) ('was', 34, 48) ### ### ('had', 23, 58) ('an', 33, 52) (')', 13, 106) ### ############################################################################################################ [2023-11-24 23:06:12,616][root][INFO] - Epoch: 1: Step: 401/19530, loss[v]=206.831085, lr=0.000000, acc@1[1]=418.5/1024=0.40869140625, acc@1[2]=418.5/1024=0.40869140625 [2023-11-24 23:08:03,300][root][INFO] - Train batch 500 [2023-11-24 23:08:03,300][root][INFO] - Avg. loss per last 100 batches: 163.614058 [2023-11-24 23:08:04,017][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 2.37 | max: 9.67 | min: 0.29 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 2.42 | max: 10.60 | min: 0.35 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.8/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: a new squad was constructed, most notably including the arrivals of midfielders maurizio ### ### ciaramitaro and simone bentivoglio, defender cesar and forward antimo iunco. ### ### [P_TEXT]: antimo iunco antimo iunco ( born 6 june 1984 ) is an italian footballer. iunco started ### ### his career at hometown club brindisi. he followed the team to play at serie c2 after winning the ### ### serie d group h champion in 2002. in january 2004, he joined verona of serie b in a co - ownership ### ### deal. he was made permanent in june 2004. after verona relegated to serie c1, he moved to cross ### ### town rival chievo ( in a co - ownership deal ), which also relegated but from serie a, where he won ### ### serie a ### ### =============================================== v_q ================================================ ### ### (',', 0, 6) ('.', 1, 4) ('##ivo', 2, 1825) ('##un', 3, 8) ('##glio', 4, 378) ('the', 5, 12) ### ### ('cesar', 6, 9553) ('anti', 7, 3) ('and', 8, 7) ('##zio', 9, 6337) ('to', 10, 19) ('in', 11, 18) ### ### ('a', 12, 11) ('of', 13, 22) ('##ram', 14, 2361) ('-', 15, 42) ('his', 16, 54) ('with', 17, 32) ### ### ('an', 18, 45) ('most', 19, 8436) ### ### =============================================== v_p ================================================ ### ### ('verona', 4220, 0) ('ownership', 17218, 1) ('##ev', 14643, 2) ('anti', 7, 3) ('.', 1, 4) ### ### ('c2', 17951, 5) (',', 0, 6) ('and', 8, 7) ('##un', 3, 8) ('##ind', 5874, 9) ### ### ('permanent', 6971, 10) ('a', 12, 11) ('the', 5, 12) ('2004', 27975, 13) ('from', 56, 14) ### ### ('serie', 13319, 15) ('but', 71, 16) ('(', 39, 17) ('in', 11, 18) ('to', 10, 19) ### ### ============================================ v_q * v_p ============================================= ### ### (',', 0, 6) ('.', 1, 4) ('##un', 3, 8) ('anti', 7, 3) ('and', 8, 7) ('the', 5, 12) ('a', 12, 11) ### ### ('in', 11, 18) ('to', 10, 19) ('of', 13, 22) ('-', 15, 42) ('(', 39, 17) ('with', 17, 32) ### ### ('from', 56, 14) ('is', 32, 24) ('but', 71, 16) ('an', 18, 45) ('his', 16, 54) ('at', 42, 27) ### ### ('for', 29, 43) ### ############################################################################################################ [2023-11-24 23:08:04,523][root][INFO] - Epoch: 1: Step: 501/19530, loss[v]=125.714134, lr=0.000001, acc@1[1]=452.0/1024=0.44140625, acc@1[2]=622.0/1024=0.607421875 [2023-11-24 23:09:55,768][root][INFO] - Train batch 600 [2023-11-24 23:09:55,768][root][INFO] - Avg. loss per last 100 batches: 119.000208 [2023-11-24 23:09:56,484][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 2.26 | max: 9.78 | min: 0.26 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 2.28 | max: 10.55 | min: 0.29 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the minifigures planned for inclusion in the set were katherine johnson, computer ### ### scientist margaret hamilton ; astronaut, physicist and educator sally ride ; astronomer nancy grace ### ### roman ; and astronaut and physician mae jemison ( who is also african american ). ### ### [P_TEXT]: taylor richardson taylor denise richardson ( born july 15, 2003 ; also known as astronaut ### ### starbright ) is an american advocate, activist, speaker, student and philanthropist. she has ### ### crowdfunded over $ 40, 000 to send girls to see the films " a wrinkle in time " and " hidden ### ### figures ". she attended space camp and has expressed interest in becoming an astronaut and doctor. ### ### richardson was born in columbia, south carolina and attends the bolles school. her mother is ### ### latonja richardson. richardson is an aspiring astronaut, and admires mae jemison. she cites ### ### jemison's book " find where the ### ### =============================================== v_q ================================================ ### ### ('set', 0, 1636) ('astronaut', 1, 1) ('ride', 2, 17422) (',', 3, 6) ('physicist', 4, 598) ### ### ('and', 5, 10) ('the', 6, 15) ('physician', 7, 255) ('.', 8, 8) ('computer', 9, 3147) ### ### ('##fi', 10, 19011) ('##gur', 11, 1336) ('astronomer', 12, 271) ('mini', 13, 9793) ### ### ('sets', 14, 11812) ('of', 15, 19) ('a', 16, 13) ('inclusion', 17, 2876) ('in', 18, 30) ### ### ('mae', 19, 5) ### ### =============================================== v_p ================================================ ### ### ('##bright', 6238, 0) ('astronaut', 1, 1) ('richardson', 20223, 2) ('columbia', 10140, 3) ### ### ('##lles', 15373, 4) ('mae', 19, 5) (',', 3, 6) ('carolina', 16451, 7) ('.', 8, 8) ### ### ('speaker', 3139, 9) ('and', 5, 10) ('doctor', 172, 11) ('student', 2297, 12) ('a', 16, 13) ### ### ('star', 1953, 14) ('the', 6, 15) ('##ink', 12552, 16) ('where', 461, 17) ('school', 1375, 18) ### ### ('of', 15, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('astronaut', 1, 1) (',', 3, 6) ('and', 5, 10) ('.', 8, 8) ('the', 6, 15) ('mae', 19, 5) ### ### ('a', 16, 13) ('of', 15, 19) ('in', 18, 30) ('to', 20, 28) ('astronauts', 26, 20) ### ### ('##ison', 39, 22) ('jem', 30, 34) ('was', 28, 50) ('is', 35, 39) ('-', 36, 42) ('his', 33, 53) ### ### ('spaceship', 38, 51) ('that', 42, 56) ('doctor', 172, 11) ### ############################################################################################################ [2023-11-24 23:09:56,992][root][INFO] - Epoch: 1: Step: 601/19530, loss[v]=117.859261, lr=0.000001, acc@1[1]=436.0/1024=0.42578125, acc@1[2]=578.5/1024=0.56494140625 [2023-11-24 23:11:48,173][root][INFO] - Train batch 700 [2023-11-24 23:11:48,174][root][INFO] - Avg. loss per last 100 batches: 95.231666 [2023-11-24 23:11:48,897][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 2.16 | max: 9.90 | min: 0.23 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 2.14 | max: 10.51 | min: 0.24 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the hangover is a 2009 american comedy film directed by todd phillips, co - produced with ### ### daniel goldberg, and written by jon lucas and scott moore. ### ### [P_TEXT]: the hangover part ii april 2009, warner bros. hired todd phillips, who directed " the ### ### hangover ", to write a sequel with scot armstrong. the deal, reached two months before the release ### ### of " the hangover " in june 2009, came as result of " the hangover "'s positive screen tests, and a ### ### trailer which drew a strong reaction from audiences at showest. the writers from the first film, ### ### scott moore and jon lucas, decided not to come back to write the sequel because, according to ### ### lucas, " they were done with that story... and didn't want to just write " hangover " sequels their ### ### whole careers ". ### ### =============================================== v_q ================================================ ### ### ('goldberg', 0, 27199) ('lucas', 1, 1) ('.', 2, 7) (',', 3, 6) ('phillips', 4, 3) ('and', 5, 8) ### ### ('the', 6, 10) ('written', 7, 266) ('film', 8, 12) ('moore', 9, 2) ('hang', 10, 9) ('hung', 11, 19) ### ### ('a', 12, 17) ('todd', 13, 22) ('of', 14, 23) ('to', 15, 16) ('in', 16, 25) ('jon', 17, 4) ### ### ('an', 18, 56) ('was', 19, 41) ### ### =============================================== v_p ================================================ ### ### ('armstrong', 25866, 0) ('lucas', 1, 1) ('moore', 9, 2) ('phillips', 4, 3) ('jon', 17, 4) ### ### ('trailer', 23863, 5) (',', 3, 6) ('.', 2, 7) ('and', 5, 8) ('hang', 10, 9) ('the', 6, 10) ### ### ('sequel', 18985, 11) ('film', 8, 12) ('show', 4323, 13) ('story', 2765, 14) ('##ot', 2244, 15) ### ### ('to', 15, 16) ('a', 12, 17) ('sc', 7333, 18) ('hung', 11, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('lucas', 1, 1) ('phillips', 4, 3) ('.', 2, 7) (',', 3, 6) ('moore', 9, 2) ('and', 5, 8) ### ### ('the', 6, 10) ('hang', 10, 9) ('film', 8, 12) ('jon', 17, 4) ('hung', 11, 19) ('a', 12, 17) ### ### ('todd', 13, 22) ('to', 15, 16) ('of', 14, 23) ('in', 16, 25) ('scott', 22, 30) ('was', 19, 41) ### ### ('2009', 34, 21) ('-', 23, 36) ### ############################################################################################################ [2023-11-24 23:11:49,403][root][INFO] - Epoch: 1: Step: 701/19530, loss[v]=87.887337, lr=0.000001, acc@1[1]=445.5/1024=0.43505859375, acc@1[2]=652.0/1024=0.63671875 [2023-11-24 23:13:40,649][root][INFO] - Train batch 800 [2023-11-24 23:13:40,650][root][INFO] - Avg. loss per last 100 batches: 76.008819 [2023-11-24 23:13:41,387][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 2.07 | max: 10.12 | min: 0.20 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 2.01 | max: 10.59 | min: 0.20 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the chernobyl nuclear power plant, officially the vladimir ilyich lenin nuclear power ### ### plant, is a closed nuclear power plant located near the abandoned city of pripyat in northern ### ### ukraine, northwest of the city of chernobyl, from the belarus – ukraine border, and about north of ### ### kyiv. ### ### [P_TEXT]: chernobyl over the course of history. jews were introduced to the city in the 16th ### ### century, and a now - defunct monastery was established near the city in 1626. by the end of the ### ### 18th century, chernobyl was a major centre of hasidic judaism under the twersky dynasty, who left ### ### chernobyl after the city was subject to pogroms in the early 20th century. the jewish community was ### ### later murdered during the holocaust. chernobyl was chosen as the site of ukraine's first nuclear ### ### power plant in 1972, located north of the city, which opened in 1977. chernobyl was evacuated on ### ### =============================================== v_q ================================================ ### ### ('##nob', 0, 0) ('ukraine', 1, 4) ('nuclear', 2, 1) ('.', 3, 5) (',', 4, 7) ('kyiv', 5, 159) ### ### ('and', 6, 13) ('lenin', 7, 11466) ('the', 8, 8) ('ukrainian', 9, 9) ('cher', 10, 3) ('a', 11, 10) ### ### ('##yl', 12, 15) ('city', 13, 6) ('border', 14, 10478) ('##yi', 15, 10488) ('in', 16, 19) ### ### ('of', 17, 17) ('to', 18, 20) ('belarus', 19, 3427) ### ### =============================================== v_p ================================================ ### ### ('##nob', 0, 0) ('nuclear', 2, 1) ('##wer', 3057, 2) ('cher', 10, 3) ('ukraine', 1, 4) ('.', 3, 5) ### ### ('city', 13, 6) (',', 4, 7) ('the', 8, 8) ('ukrainian', 9, 9) ('a', 11, 10) ('##gr', 9233, 11) ### ### ('history', 6843, 12) ('and', 6, 13) ('judaism', 9158, 14) ('##yl', 12, 15) ('chosen', 5617, 16) ### ### ('of', 17, 17) ('near', 40, 18) ('in', 16, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##nob', 0, 0) ('nuclear', 2, 1) ('ukraine', 1, 4) ('.', 3, 5) (',', 4, 7) ('cher', 10, 3) ### ### ('the', 8, 8) ('and', 6, 13) ('city', 13, 6) ('ukrainian', 9, 9) ('a', 11, 10) ('##yl', 12, 15) ### ### ('of', 17, 17) ('in', 16, 19) ('to', 18, 20) ('that', 26, 25) ('power', 25, 26) ('-', 23, 29) ### ### ('near', 40, 18) ('kyiv', 5, 159) ### ############################################################################################################ [2023-11-24 23:13:41,899][root][INFO] - Epoch: 1: Step: 801/19530, loss[v]=64.231216, lr=0.000001, acc@1[1]=480.0/1024=0.46875, acc@1[2]=790.0/1024=0.771484375 [2023-11-24 23:15:33,233][root][INFO] - Train batch 900 [2023-11-24 23:15:33,233][root][INFO] - Avg. loss per last 100 batches: 64.324941 [2023-11-24 23:15:33,961][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 1.98 | max: 10.29 | min: 0.18 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 1.89 | max: 10.55 | min: 0.16 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: kutztown university of pennsylvania ( kutztown university or ku ) is a public university ### ### in kutztown, pennsylvania. ### ### [P_TEXT]: kutztown, pennsylvania kutztown is a borough in berks county, pennsylvania, united ### ### states, southwest of allentown and northeast of reading. as of the 2010 census, the borough had a ### ### population of 5, 012. kutztown university of pennsylvania is located just outside the borough ### ### limits to the southwest. george ( coots ) kutz purchased of land that became kutztown on june 16, ### ### 1755, from peter wentz who owned much of what is now maxatawny township. kutz first laid out his ### ### plans for the town in 1779. ### ### =============================================== v_q ================================================ ### ### ('pennsylvania', 0, 5) ('##tz', 1, 6) (',', 2, 8) ('ku', 3, 13) ('.', 4, 7) ('university', 5, 29) ### ### ('and', 6, 12) ('the', 7, 11) ('a', 8, 17) ('##town', 9, 3) ('to', 10, 23) ('of', 11, 30) ### ### ('or', 12, 77) ('in', 13, 26) ('that', 14, 18) ('-', 15, 36) ('an', 16, 41) ('was', 17, 25) ### ### ('his', 18, 31) ('(', 19, 37) ### ### =============================================== v_p ================================================ ### ### ('##wny', 675, 0) ('reading', 451, 1) ('##rks', 7574, 2) ('##town', 9, 3) ('borough', 5087, 4) ### ### ('pennsylvania', 0, 5) ('##tz', 1, 6) ('.', 4, 7) (',', 2, 8) ('be', 187, 9) ('went', 326, 10) ### ### ('the', 7, 11) ('and', 6, 12) ('ku', 3, 13) ('township', 14481, 14) ('go', 493, 15) ### ### ('allen', 22690, 16) ('a', 8, 17) ('that', 14, 18) ('##z', 1591, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('pennsylvania', 0, 5) ('##tz', 1, 6) (',', 2, 8) ('.', 4, 7) ('##town', 9, 3) ('ku', 3, 13) ### ### ('the', 7, 11) ('and', 6, 12) ('a', 8, 17) ('university', 5, 29) ('to', 10, 23) ('that', 14, 18) ### ### ('of', 11, 30) ('in', 13, 26) ('was', 17, 25) ('his', 18, 31) ('-', 15, 36) ('is', 22, 32) ### ### ('an', 16, 41) ('(', 19, 37) ### ############################################################################################################ [2023-11-24 23:15:34,466][root][INFO] - Epoch: 1: Step: 901/19530, loss[v]=49.448261, lr=0.000001, acc@1[1]=495.0/1024=0.4833984375, acc@1[2]=823.5/1024=0.80419921875 [2023-11-24 23:17:25,609][root][INFO] - Train batch 1000 [2023-11-24 23:17:25,609][root][INFO] - Avg. loss per last 100 batches: 54.445927 [2023-11-24 23:17:26,341][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 1.87 | max: 10.47 | min: 0.15 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 1.74 | max: 10.58 | min: 0.13 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: many examples of this carbocation have been characterized, but under normal operating ### ### conditions these ### ### [P_TEXT]: paraben in neutral water, raman spectroscopy has confirmed that chlorine is predominantly ### ### present as hypochlorous acid ( hclo ). parabens can react with hclo to form mono - and di - ### ### chlorinated products through electrophilic aromatic substitution. the electrophilic attack of the ### ### chlorine forms a carbocation that is stabilized by donated electron density from the hydroxyl group ### ### of the paraben. this step is endergonic due to the loss of aromaticity, though the hydroxyl group ### ### acts as an activating group that increases the rate. a base can then abstract a proton from the ### ### carbon containing the chlorine, ### ### =============================================== v_q ================================================ ### ### ('.', 0, 7) (',', 1, 10) ('examples', 2, 8109) ('and', 3, 8) ('the', 4, 12) ('##cation', 5, 160) ### ### ('conditions', 6, 27071) ('but', 7, 25) ('of', 8, 17) ('a', 9, 21) ('##bo', 10, 48) ### ### ('this', 11, 51) ('operating', 12, 2631) ('to', 13, 28) ('in', 14, 24) ('normal', 15, 2826) ### ### ('an', 16, 46) ('many', 17, 869) ('that', 18, 31) ('characterized', 19, 24164) ### ### =============================================== v_p ================================================ ### ### ('neutral', 243, 0) ('##ben', 3046, 1) ('substitution', 9586, 2) ('hydro', 9046, 3) ### ### ('electron', 17469, 4) ('rama', 24425, 5) ('##phi', 13291, 6) ('.', 0, 7) ('and', 3, 8) ### ### ('para', 6085, 9) (',', 1, 10) ('electro', 1408, 11) ('the', 4, 12) ('from', 35, 13) ### ### ('rate', 3751, 14) ('carbon', 18063, 15) ('##yp', 14177, 16) ('of', 8, 17) ('##ous', 15155, 18) ### ### ('water', 4021, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('.', 0, 7) (',', 1, 10) ('and', 3, 8) ('the', 4, 12) ('of', 8, 17) ('a', 9, 21) ('but', 7, 25) ### ### ('in', 14, 24) ('to', 13, 28) ('neutral', 243, 0) ('from', 35, 13) ('that', 18, 31) ### ### ('##bo', 10, 48) ('is', 22, 35) ('this', 11, 51) ('an', 16, 46) ('was', 21, 41) ('-', 26, 40) ### ### ('with', 29, 42) ('his', 39, 37) ### ############################################################################################################ [2023-11-24 23:17:26,850][root][INFO] - Epoch: 1: Step: 1001/19530, loss[v]=54.619186, lr=0.000001, acc@1[1]=494.0/1024=0.482421875, acc@1[2]=742.5/1024=0.72509765625 [2023-11-24 23:19:17,629][root][INFO] - Train batch 1100 [2023-11-24 23:19:17,629][root][INFO] - Avg. loss per last 100 batches: 44.293861 [2023-11-24 23:19:18,344][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 1.67 | max: 10.56 | min: 0.11 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 1.47 | max: 10.41 | min: 0.08 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: jenny " was chosen as the fourth single off the album, released only in finland on 3 may. ### ### [P_TEXT]: jenny ( studio killers song ) " jenny ", also titled " jenny ( i wanna ruin our ### ### friendship ) " in late 2020 following its resurgence on tiktok, is a song written, produced and ### ### performed by the virtual band studio killers, released in may 2013 as the fourth single from their ### ### self - titled debut studio album. played in the key of b major, the song, inspired by singer chubby ### ### cherry's crush on a person named jenny, discusses a girl falling in love with her best friend. the ### ### single was a number - two hit on the finnish singles chart, and was received positively ### ### =============================================== v_q ================================================ ### ### ('finland', 0, 30) ('jenny', 1, 1) ('chosen', 2, 1005) ('.', 3, 4) (',', 4, 5) ('the', 5, 10) ### ### ('finnish', 6, 3) ('and', 7, 12) ('may', 8, 59) ('album', 9, 24) ('a', 10, 17) ('to', 11, 32) ### ### ('off', 12, 1407) ('of', 13, 33) ('fourth', 14, 80) ('that', 15, 34) ('was', 16, 31) ('is', 17, 44) ### ### ('in', 18, 37) ('an', 19, 45) ### ### =============================================== v_p ================================================ ### ### ('cherry', 23397, 0) ('jenny', 1, 1) ('killers', 18191, 2) ('finnish', 6, 3) ('.', 3, 4) ### ### (',', 4, 5) ('2020', 17796, 6) ('wanna', 4620, 7) ('killer', 14298, 8) ('friendship', 1209, 9) ### ### ('the', 5, 10) ('virtual', 17594, 11) ('and', 7, 12) ('chu', 392, 13) ('##bby', 28285, 14) ### ### ('##kt', 29294, 15) ('##ok', 9268, 16) ('a', 10, 17) ('ruin', 17863, 18) ('best', 2233, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('jenny', 1, 1) ('.', 3, 4) ('finnish', 6, 3) (',', 4, 5) ('the', 5, 10) ('finland', 0, 30) ### ### ('and', 7, 12) ('a', 10, 17) ('album', 9, 24) ('to', 11, 32) ('of', 13, 33) ('may', 8, 59) ### ### ('was', 16, 31) ('his', 20, 28) ('that', 15, 34) ('in', 18, 37) (';', 21, 36) ('is', 17, 44) ### ### ('an', 19, 45) ('her', 35, 29) ### ############################################################################################################ [2023-11-24 23:19:18,853][root][INFO] - Epoch: 1: Step: 1101/19530, loss[v]=32.571823, lr=0.000001, acc@1[1]=513.0/1024=0.5009765625, acc@1[2]=888.5/1024=0.86767578125 [2023-11-24 23:21:09,882][root][INFO] - Train batch 1200 [2023-11-24 23:21:09,882][root][INFO] - Avg. loss per last 100 batches: 34.476421 [2023-11-24 23:21:10,624][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 1.42 | max: 10.47 | min: 0.07 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 1.08 | max: 10.05 | min: 0.04 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: she composed the song " koi humdum na raha " originally sung by ashok kumar in the film " ### ### jeevan naiya " ( 1936 ) and later sung by kishore kumar in " jhumroo " ( 1961 ). ### ### [P_TEXT]: jhumroo soundtrack was composed by kishore kumar. the music was arranged by s. d. ### ### burman's musical band. the song " koi humdum na raha " was a cover of the song of the same name, ### ### composed by saraswati devi for " jeevan naiya " ( 1936 ). and was sung by the actor – and ### ### occasional singer – ashok kumar. kishore kumar was five years old when he heard his elder brother ### ### ashok kumar sing " koi humdum na raha ". twenty - five years later, when kishore kumar was ### ### composing the music for jhumroo, he went to his brother asking for permission to sing " koi ### ### =============================================== v_q ================================================ ### ### ('##shore', 0, 1) ('hum', 1, 4) ('##du', 2, 10) ('kumar', 3, 2) (',', 4, 7) ('.', 5, 8) ### ### ('na', 6, 16) ('ashok', 7, 18) ('##iya', 8, 30) ('the', 9, 12) ('and', 10, 13) ('ra', 11, 29) ### ### ('##hum', 12, 3) ('##m', 13, 25) ('film', 14, 198) ('##ro', 15, 26) ('je', 16, 37) ('to', 17, 32) ### ### ('a', 18, 23) ('j', 19, 19) ### ### =============================================== v_p ================================================ ### ### ('burma', 8060, 0) ('##shore', 0, 1) ('kumar', 3, 2) ('##hum', 12, 3) ('hum', 1, 4) ### ### ('permission', 236, 5) ('##sw', 588, 6) (',', 4, 7) ('.', 5, 8) ('devi', 4971, 9) ('##du', 2, 10) ### ### ('sara', 15111, 11) ('the', 9, 12) ('and', 10, 13) ('brother', 3433, 14) ('ki', 23, 15) ### ### ('na', 6, 16) ('singer', 507, 17) ('ashok', 7, 18) ('j', 19, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##shore', 0, 1) ('hum', 1, 4) ('kumar', 3, 2) ('##du', 2, 10) ('##hum', 12, 3) (',', 4, 7) ### ### ('.', 5, 8) ('na', 6, 16) ('ashok', 7, 18) ('the', 9, 12) ('and', 10, 13) ('j', 19, 19) ### ### ('##iya', 8, 30) ('ki', 23, 15) ('##m', 13, 25) ('ra', 11, 29) ('##ro', 15, 26) ('a', 18, 23) ### ### ('##ha', 22, 21) ('to', 17, 32) ### ############################################################################################################ [2023-11-24 23:21:11,128][root][INFO] - Epoch: 1: Step: 1201/19530, loss[v]=24.700787, lr=0.000001, acc@1[1]=523.0/1024=0.5107421875, acc@1[2]=902.0/1024=0.880859375 [2023-11-24 23:23:01,449][root][INFO] - Train batch 1300 [2023-11-24 23:23:01,450][root][INFO] - Avg. loss per last 100 batches: 25.122790 [2023-11-24 23:23:02,163][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 1.10 | max: 10.42 | min: 0.04 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.66 | max: 9.26 | min: 0.02 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: regular contributors to the magazine, both past and present, include its former features ### ### editor, brian kellow, william ashbrook, scott barnes, jochen breiholz, erika davidson, justin ### ### davidson, peter g. davis, matthew gurewitsch, joel honig, tim page, judith malafronte, mark thomas ### ### ketterson, ### ### [P_TEXT]: peter g. davis contributor to " opera news " and " musical america ". davis married scott ### ### parris in 2009. they remained married until the former's death. davis suffered a stroke in 2018. he ### ### died on february 13, 2021, at mount sinai west in new york city. he was 84, and suffered a brief ### ### illness prior to his death. ### ### =============================================== v_q ================================================ ### ### ('##front', 0, 5064) ('##holz', 1, 29129) ('page', 2, 1226) ('##tters', 3, 24357) ### ### ('barnes', 4, 24869) ('mark', 5, 2219) ('judith', 6, 21719) ('ash', 7, 381) ('davidson', 8, 18670) ### ### ('magazine', 9, 21311) ('hon', 10, 10440) (',', 11, 12) ('.', 12, 8) ('##chen', 13, 7894) ### ### ('g', 14, 4) ('##ig', 15, 8000) ('erika', 16, 22941) ('davis', 17, 5) ('present', 18, 389) ### ### ('##brook', 19, 29294) ### ### =============================================== v_p ================================================ ### ### ('par', 4270, 0) ('opera', 7353, 1) ('sinai', 27843, 2) ('stroke', 365, 3) ('g', 14, 4) ### ### ('davis', 17, 5) ('peter', 41, 6) ('##ris', 26018, 7) ('.', 12, 8) ('2021', 13819, 9) ### ### ('mount', 5756, 10) ('scott', 53, 11) (',', 11, 12) ('a', 39, 13) ('the', 23, 14) ### ### ('york', 3649, 15) ('and', 25, 16) ('musical', 14281, 17) ('death', 8557, 18) ('died', 1415, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('g', 14, 4) ('davis', 17, 5) ('.', 12, 8) (',', 11, 12) ('peter', 41, 6) ('the', 23, 14) ### ### ('and', 25, 16) ('a', 39, 13) ('scott', 53, 11) ('his', 52, 22) (';', 49, 25) ('that', 54, 23) ### ### ('to', 44, 32) ('of', 48, 27) ('was', 46, 30) ('in', 50, 37) ('stroke', 365, 3) ('is', 56, 33) ### ### ('he', 67, 28) ('at', 69, 31) ### ############################################################################################################ [2023-11-24 23:23:02,672][root][INFO] - Epoch: 1: Step: 1301/19530, loss[v]=13.695017, lr=0.000001, acc@1[1]=625.5/1024=0.61083984375, acc@1[2]=921.0/1024=0.8994140625 [2023-11-24 23:24:53,259][root][INFO] - Train batch 1400 [2023-11-24 23:24:53,260][root][INFO] - Avg. loss per last 100 batches: 15.475730 [2023-11-24 23:24:54,012][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.76 | max: 10.04 | min: 0.02 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.34 | max: 7.98 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the screenplay by richard maibaum and michael g. wilson takes its characters and combines ### ### elements from the plots from two short stories from ### ### [P_TEXT]: richard maibaum draft of " the man with the golden gun " ( 1974 ). he was one of the many ### ### writers who worked on " the spy who loved me " ( 1977 ), sharing credit with christopher wood. ### ### maibaum was not used on " moonraker " ( 1979 ), the producers preferring wood. instead, maibaum ### ### worked on a bond spoof, " " ( 1980 ). maibaum was brought back to work on the bonds in association ### ### with michael g. wilson, broccoli's step - son. their first movie together was " for your eyes only ### ### " ( 1981 ). it ### ### =============================================== v_q ================================================ ### ### ('##baum', 0, 0) ('mai', 1, 1) ('wilson', 2, 8) ('short', 3, 3417) ('michael', 4, 19) ### ### ('screenplay', 5, 23866) ('g', 6, 18) ('richard', 7, 4) ('.', 8, 9) (',', 9, 12) ### ### ('plot', 10, 11233) ('and', 11, 21) ('plots', 12, 14198) ('by', 13, 119) ('the', 14, 17) ### ### ('takes', 15, 4492) ('two', 16, 664) ('took', 17, 2221) ('a', 18, 22) ('long', 19, 2058) ### ### =============================================== v_p ================================================ ### ### ('##baum', 0, 0) ('mai', 1, 1) ('bonds', 14998, 2) ('##cco', 28742, 3) ('richard', 7, 4) ### ### ('christopher', 4568, 5) ('moon', 7750, 6) ('bond', 20836, 7) ('wilson', 2, 8) ('.', 8, 9) ### ### ('wood', 1209, 10) ('spy', 1533, 11) (',', 9, 12) ('gun', 8923, 13) ('golden', 16436, 14) ### ### ('eyes', 7343, 15) ('bro', 14789, 16) ('the', 14, 17) ('g', 6, 18) ('michael', 4, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##baum', 0, 0) ('mai', 1, 1) ('wilson', 2, 8) ('richard', 7, 4) ('.', 8, 9) (',', 9, 12) ### ### ('michael', 4, 19) ('g', 6, 18) ('the', 14, 17) ('and', 11, 21) ('a', 18, 22) (';', 25, 25) ### ### ('your', 54, 20) ('to', 27, 46) ('his', 40, 32) ('he', 34, 42) ('that', 37, 37) ('of', 28, 50) ### ### ('my', 49, 30) ('was', 36, 41) ### ############################################################################################################ [2023-11-24 23:24:54,516][root][INFO] - Epoch: 1: Step: 1401/19530, loss[v]=9.726568, lr=0.000001, acc@1[1]=646.5/1024=0.63134765625, acc@1[2]=921.5/1024=0.89990234375 [2023-11-24 23:26:44,981][root][INFO] - Train batch 1500 [2023-11-24 23:26:44,982][root][INFO] - Avg. loss per last 100 batches: 8.262563 [2023-11-24 23:26:45,700][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.47 | max: 9.39 | min: 0.01 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.17 | max: 6.28 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: his descendants went on to found the rana dynasty of nepal. ### ### [P_TEXT]: rana dynasty rana, who belonged to kunwar family which was then considered a minor noble ### ### family of kshatriya status. jang bahadur was a son of gorkhali governor bal narsingh kunwar and ### ### nephew of mathabarsingh thapa, the reigning prime minister of nepal ( 1843 – 1845 ) from the thapa ### ### dynasty. bal narsingh kunwar was the son of kaji ranajit kunwar and grandson of " sardar " ram ### ### krishna kunwar, who was prominent military general of king prithvi narayan shah. ram krishna kunwar ### ### =============================================== v_q ================================================ ### ### ('nepal', 0, 2) ('rana', 1, 1) ('descendants', 2, 686) ('dynasty', 3, 22) ('.', 4, 11) (',', 5, 18) ### ### ('family', 6, 73) ('nepali', 7, 40) ('the', 8, 26) ('descendant', 9, 547) ('his', 10, 50) ### ### ('a', 11, 35) ('and', 12, 37) ('kathmandu', 13, 53) (';', 14, 41) ('to', 15, 90) ### ### ('alumni', 16, 531) ('burkina', 17, 184) ('of', 18, 220) ('in', 19, 185) ### ### =============================================== v_p ================================================ ### ### ('bahadur', 77, 0) ('rana', 1, 1) ('nepal', 0, 2) ('##war', 19055, 3) ('shah', 210, 4) ### ### ('math', 6403, 5) ('ram', 467, 6) ('jang', 12343, 7) ('##ji', 2406, 8) ('##pa', 6094, 9) ### ### ('kun', 2005, 10) ('.', 4, 11) ('pri', 8285, 12) ('##hal', 26421, 13) ('narayan', 9291, 14) ### ### ('##ing', 929, 15) ('##rk', 20603, 16) ('##jit', 21585, 17) (',', 5, 18) ('krishna', 1086, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('nepal', 0, 2) ('rana', 1, 1) ('.', 4, 11) (',', 5, 18) ('bahadur', 77, 0) ('dynasty', 3, 22) ### ### ('the', 8, 26) ('shah', 210, 4) ('a', 11, 35) ('nepali', 7, 40) ('and', 12, 37) ('?', 24, 25) ### ### ('ram', 467, 6) ('his', 10, 50) (';', 14, 41) ('family', 6, 73) ('kathmandu', 13, 53) ### ### ('go', 262, 20) ('he', 21, 48) ('domesday', 58, 34) ### ############################################################################################################ [2023-11-24 23:26:46,206][root][INFO] - Epoch: 1: Step: 1501/19530, loss[v]=4.229795, lr=0.000002, acc@1[1]=747.5/1024=0.72998046875, acc@1[2]=928.0/1024=0.90625 [2023-11-24 23:28:36,705][root][INFO] - Train batch 1600 [2023-11-24 23:28:36,706][root][INFO] - Avg. loss per last 100 batches: 3.779751 [2023-11-24 23:28:37,450][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.28 | max: 8.08 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.08 | max: 4.78 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: this was the mets'fifth double of the inning, which set a new league championship series ### ### record. ### ### [P_TEXT]: dale scott, 2001, and 2011, calling balls and strikes. he has also worked in six league ### ### championship series ( 1996, 1999, 2000, 2002, 2009, 2013 ) and in twelve division series ( 1995, ### ### 1997, 1998, 2001, 2003, 2004, 2005, 2007, 2008, 2011, 2014, 2015 ). when the al introduced red ### ### shirts in 1996, scott frequently was the only umpire ### ### =============================================== v_q ================================================ ### ### ('inning', 0, 4236) ('mets', 1, 16978) ('.', 2, 6) ('double', 3, 9923) (',', 4, 14) ### ### ('new', 5, 3816) ('league', 6, 3) ('set', 7, 503) ('a', 8, 106) ('and', 9, 67) ('the', 10, 20) ### ### ('championship', 11, 17) ('twins', 12, 1030) (';', 13, 100) ('doubles', 14, 16167) ### ### ('sets', 15, 6016) ('series', 16, 95) ('this', 17, 98) ('third', 18, 2287) ('baseball', 19, 1165) ### ### =============================================== v_p ================================================ ### ### ('dale', 4089, 0) ('strikes', 1528, 1) ('umpire', 9225, 2) ('league', 6, 3) ('strike', 46, 4) ### ### ('balls', 8818, 5) ('.', 2, 6) ('scott', 7881, 7) ('red', 49, 8) ('division', 5098, 9) ### ### ('striking', 549, 10) ('ball', 86, 11) ('struck', 598, 12) ('al', 4725, 13) (',', 4, 14) ### ### ('shirt', 4455, 15) ('?', 27, 16) ('championship', 11, 17) ('serra', 1428, 18) ('anger', 272, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('league', 6, 3) ('.', 2, 6) (',', 4, 14) ('strike', 46, 4) ('red', 49, 8) ('championship', 11, 17) ### ### ('the', 10, 20) ('?', 27, 16) ('ball', 86, 11) ('strikes', 1528, 1) ('dale', 4089, 0) ### ### ('nato', 29, 24) ('striking', 549, 10) ('and', 9, 67) ('struck', 598, 12) ('domesday', 53, 21) ### ### ('a', 8, 106) (';', 13, 100) ('ring', 91, 27) ('kent', 180, 22) ### ############################################################################################################ [2023-11-24 23:28:37,953][root][INFO] - Epoch: 1: Step: 1601/19530, loss[v]=2.975074, lr=0.000002, acc@1[1]=778.0/1024=0.759765625, acc@1[2]=864.5/1024=0.84423828125 [2023-11-24 23:30:28,337][root][INFO] - Train batch 1700 [2023-11-24 23:30:28,338][root][INFO] - Avg. loss per last 100 batches: 1.714653 [2023-11-24 23:30:29,065][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.18 | max: 6.53 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.06 | max: 4.57 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it features the tenth doctor without an official companion. ### ### [P_TEXT]: tenth doctor michelle ryan and two by catherine tate. a number of tenth doctor novels ### ### were also abridged to become audiobooks, again featuring david tennant's voice alongside other cast ### ### members such as freema agyeman and television series guest stars such as georgia moffett, reggie ### ### yates and anthony head ; the last of these scheduled is " judgement of the judoon ", for december ### ### 2010. on 26 october 2015, it was announced tennant would reprise his role for, a series of full - ### ### cast audio stories produced by big finish productions. the first volume of ### ### =============================================== v_q ================================================ ### ### ('companion', 0, 4208) ('tenth', 1, 6) ('.', 2, 14) ('official', 3, 6294) (',', 4, 26) ### ### ('doctor', 5, 21) ('the', 6, 164) ('fourth', 7, 45) ('third', 8, 53) ('without', 9, 4633) ### ### ('part', 10, 59) ('and', 11, 786) ('second', 12, 29) ('prehistoric', 13, 139) ### ### ('unofficial', 14, 5497) (';', 15, 319) ('guide', 16, 897) ('fifth', 17, 43) ('formal', 18, 296) ### ### ('ninth', 19, 144) ### ### =============================================== v_p ================================================ ### ### ('tate', 4306, 0) ('ryan', 6647, 1) ('free', 798, 2) ('michelle', 11302, 3) ('judo', 949, 4) ### ### ('catherine', 3998, 5) ('tenth', 1, 6) ('anthony', 3250, 7) ('reggie', 20307, 8) ### ### ('yates', 23297, 9) ('head', 8830, 10) ('david', 2450, 11) ('ten', 216, 12) ('ag', 12330, 13) ### ### ('.', 2, 14) ('georgia', 5080, 15) ('big', 5552, 16) ('##nant', 29157, 17) ('audio', 5410, 18) ### ### ('finish', 5134, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('tenth', 1, 6) ('.', 2, 14) ('doctor', 5, 21) (',', 4, 26) ('second', 12, 29) ('?', 25, 32) ### ### ('free', 798, 2) ('ten', 216, 12) ('fourth', 7, 45) ('judo', 949, 4) ('fifth', 17, 43) ### ### ('third', 8, 53) ('full', 287, 24) ('part', 10, 59) ('tate', 4306, 0) ('kent', 51, 49) ### ### ('sixth', 22, 79) ('the', 6, 164) ('nato', 21, 98) ('david', 2450, 11) ### ############################################################################################################ [2023-11-24 23:30:29,568][root][INFO] - Epoch: 1: Step: 1701/19530, loss[v]=0.760071, lr=0.000002, acc@1[1]=900.5/1024=0.87939453125, acc@1[2]=969.0/1024=0.9462890625 [2023-11-24 23:32:20,214][root][INFO] - Train batch 1800 [2023-11-24 23:32:20,215][root][INFO] - Avg. loss per last 100 batches: 1.104677 [2023-11-24 23:32:20,937][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.11 | max: 5.33 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.70 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he is best known as the ringleader of the players involved in the 1919 black sox scandal. ### ### [P_TEXT]: bill burns ( baseball ) go on to break burns's record on august 4, 1989, this time losing ### ### a perfect game with one out to go. prior to the start of the 1919 world series, a group of players ### ### from the chicago white sox agreed to intentionally lose the world series in exchange for money from ### ### gamblers. burns met with eddie cicotte and chick gandil at the ansonia, a hotel in new york city ### ### during the formative stages of the event. it is likely that burns operated on behalf ### ### =============================================== v_q ================================================ ### ### ('black', 0, 31) ('sox', 1, 12) ('ring', 2, 42) ('.', 3, 18) (',', 4, 40) ('scandal', 5, 8693) ### ### ('players', 6, 1566) ('best', 7, 122) ('blue', 8, 228) ('##leader', 9, 19003) ('the', 10, 127) ### ### ('white', 11, 7) ('red', 12, 149) ('player', 13, 1189) ('?', 14, 29) ('crap', 15, 1073) ### ### ('green', 16, 125) ('and', 17, 360) ('dioxide', 18, 311) ('1919', 19, 28) ### ### =============================================== v_p ================================================ ### ### ('bill', 5043, 0) ('behalf', 10317, 1) ('burns', 18750, 2) ('ci', 9110, 3) ('chick', 21361, 4) ### ### ('##cott', 27621, 5) ('anson', 26250, 6) ('white', 11, 7) ('gamble', 18070, 8) ('baseball', 38, 9) ### ### ('perfect', 3186, 10) ('##di', 26784, 11) ('sox', 1, 12) ('##ia', 9844, 13) ('august', 1654, 14) ### ### ('world', 10871, 15) ('hotel', 1825, 16) ('burned', 2404, 17) ('.', 3, 18) ('york', 1043, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('sox', 1, 12) ('.', 3, 18) ('black', 0, 31) ('white', 11, 7) ('ring', 2, 42) ('baseball', 38, 9) ### ### (',', 4, 40) ('?', 14, 29) ('1919', 19, 28) ('out', 67, 27) ('best', 7, 122) ('go', 691, 21) ### ### ('bill', 5043, 0) ('york', 1043, 19) ('the', 10, 127) ('august', 1654, 14) ('red', 12, 149) ### ### ('blue', 8, 228) ('perfect', 3186, 10) ('green', 16, 125) ### ############################################################################################################ [2023-11-24 23:32:21,441][root][INFO] - Epoch: 1: Step: 1801/19530, loss[v]=1.161919, lr=0.000002, acc@1[1]=864.0/1024=0.84375, acc@1[2]=896.0/1024=0.875 [2023-11-24 23:34:11,917][root][INFO] - Train batch 1900 [2023-11-24 23:34:11,918][root][INFO] - Avg. loss per last 100 batches: 0.863821 [2023-11-24 23:34:12,644][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.09 | max: 4.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.05 | max: 4.61 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the city's skyline is dominated by several college buildings, along with the spire of the ### ### our lady and the english martyrs church, and the chimney of addenbrooke's hospital. ### ### [P_TEXT]: james coward ( raf officer ) his thigh and stem the bleeding. upon landing in a field, he ### ### met a youth who quickly found a doctor. within half an hour, he was taken to addenbrooke's hospital ### ### in cambridge to have his left leg amputated. he was promoted to flight lieutenant on 3 september ### ### 1940. having taken five months to recover from the surgery, coward joined the personal staff of ### ### prime minister winston churchill. his main duty at chequers, the prime minister's country ### ### residence, was to judge if reports were urgent enough to warrant waking ### ### =============================================== v_q ================================================ ### ### ('chimney', 0, 12747) ('skyline', 1, 11912) ('hospital', 2, 7) ('.', 3, 28) (',', 4, 100) ### ### ('martyrs', 5, 19415) ('spire', 6, 15522) ('college', 7, 1504) ('##brook', 8, 8) ### ### ('english', 9, 8710) ('city', 10, 5927) ('add', 11, 12) ('exempt', 12, 759) ('our', 13, 670) ### ### ('the', 14, 819) ('?', 15, 29) ('lady', 16, 3047) ('##en', 17, 85) ('england', 18, 3932) ### ### ('and', 19, 2723) ### ### =============================================== v_p ================================================ ### ### ('coward', 21919, 0) ('raf', 1689, 1) ('churchill', 14909, 2) ('##que', 21142, 3) ### ### ('cambridge', 463, 4) ('james', 4012, 5) ('thigh', 17586, 6) ('hospital', 2, 7) ('##brook', 8, 8) ### ### ('bleeding', 6943, 9) ('che', 19013, 10) ('surgery', 10262, 11) ('add', 11, 12) ### ### ('winston', 26704, 13) ('flight', 1021, 14) ('##rs', 5955, 15) ('staff', 2252, 16) ### ### ('urgent', 9778, 17) ('amp', 15759, 18) ('minister', 7184, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('hospital', 2, 7) ('##brook', 8, 8) ('add', 11, 12) ('.', 3, 28) ('?', 15, 29) ('added', 20, 34) ### ### ('cambridge', 463, 4) (',', 4, 100) ('raf', 1689, 1) ('adds', 26, 43) ('##en', 17, 85) ### ### ('dioxide', 22, 63) ('flight', 1021, 14) ('anger', 30, 77) ('##frame', 654, 31) ('james', 4012, 5) ### ### ('addition', 50, 61) ('staff', 2252, 16) ('adding', 56, 60) ('lieutenant', 2115, 21) ### ############################################################################################################ [2023-11-24 23:34:13,148][root][INFO] - Epoch: 1: Step: 1901/19530, loss[v]=1.163045, lr=0.000002, acc@1[1]=859.0/1024=0.8388671875, acc@1[2]=887.0/1024=0.8662109375 [2023-11-24 23:36:03,728][root][INFO] - Train batch 2000 [2023-11-24 23:36:03,728][root][INFO] - Avg. loss per last 100 batches: 0.767048 [2023-11-24 23:36:04,470][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.07 | max: 4.62 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.06 | max: 4.52 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: this project became the construction of " 60163 tornado ", carried out by locomotive ### ### construction co ltd, a wholly owned subsidiary of the a1 trust. ### ### [P_TEXT]: winton train as the train travelled through the netherlands from emmerich am rhein to ### ### hook of holland, it was hauled by locomotive no. 01 1075. no. 01 1075 is a preserved steam ### ### locomotive built in 1940 and based at the stoom stichting nederland ( ssn ) railway museum in ### ### rotterdam. travelling through england ( from harwich to london ), the train was hauled by no. 60163 ### ### " tornado ", a british mainline steam locomotive built in 2008 by the a1 steam locomotive trust, ### ### the construction of which began in 1994 ### ### =============================================== v_q ================================================ ### ### ('locomotive', 0, 12) ('tornado', 1, 3) ('a1', 2, 10) ('trust', 3, 7) ('.', 4, 22) ### ### ('construction', 5, 41) (',', 6, 919) ('ltd', 7, 6045) ('project', 8, 1670) ('locomotives', 9, 50) ### ### ('trusts', 10, 81) ('wholly', 11, 20635) ('the', 12, 1105) ('##3', 13, 43) ('?', 14, 31) ### ### ('60', 15, 35) ('tornadoes', 16, 55) ('a2', 17, 66) ('##orestation', 18, 318) ('snow', 19, 1010) ### ### =============================================== v_p ================================================ ### ### ('win', 2329, 0) ('hook', 20075, 1) ('##erland', 23505, 2) ('tornado', 1, 3) ('holland', 11755, 4) ### ### ('##ton', 5978, 5) ('rotterdam', 24558, 6) ('trust', 3, 7) ('em', 7185, 8) ('##oom', 27567, 9) ### ### ('a1', 2, 10) ('train', 807, 11) ('locomotive', 0, 12) ('##wich', 27537, 13) ('london', 824, 14) ### ### ('steam', 356, 15) ('railway', 46, 16) ('winning', 1020, 17) ('ha', 1221, 18) ('am', 2957, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('tornado', 1, 3) ('trust', 3, 7) ('locomotive', 0, 12) ('a1', 2, 10) ('.', 4, 22) ### ### ('construction', 5, 41) ('?', 14, 31) ('locomotives', 9, 50) ('railway', 46, 16) ('60', 15, 35) ### ### ('##3', 13, 43) ('trusts', 10, 81) ('tornadoes', 16, 55) ('steam', 356, 15) ('win', 2329, 0) ### ### ('dioxide', 20, 59) ('train', 807, 11) ('a2', 17, 66) ('london', 824, 14) ('m1', 24, 61) ### ############################################################################################################ [2023-11-24 23:36:04,977][root][INFO] - Epoch: 1: Step: 2001/19530, loss[v]=1.208893, lr=0.000002, acc@1[1]=878.5/1024=0.85791015625, acc@1[2]=912.5/1024=0.89111328125 [2023-11-24 23:37:55,784][root][INFO] - Train batch 2100 [2023-11-24 23:37:55,784][root][INFO] - Avg. loss per last 100 batches: 0.711753 [2023-11-24 23:37:56,522][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.07 | max: 4.50 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.06 | max: 4.66 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he replaced former envoy alexander beglov, who was removed from the position and made ### ### acting governor of saint petersburg. ### ### [P_TEXT]: alexander beglov in the regional group " petersburg " ). on 7 december 2003, he was ### ### elected to the state duma and refused the mandate. from 2003 to 2004 he served as first deputy ### ### plenipotentiary representative in the northwestern federal district. in may 2004, he became a ### ### presidential aide and head of the presidential control department. in november 2004, at the v ### ### congress of united russia, he was elected to the party's supreme council. in october 2005, beglov ### ### became a member of the presidential council for the implementation of priority national projects ### ### and demographic policy, and in july 2006 he became a ### ### =============================================== v_q ================================================ ### ### ('envoy', 0, 1915) ('alexander', 1, 1) ('##lov', 2, 15) ('petersburg', 3, 29) ('.', 4, 32) ### ### (',', 5, 1495) ('governor', 6, 752) ('beg', 7, 4) ('saint', 8, 4599) ('russian', 9, 37) ### ### ('russia', 10, 14) ('replace', 11, 1279) ('the', 12, 991) ('replaced', 13, 8935) ### ### ('replacing', 14, 4089) ('and', 15, 2076) ('former', 16, 6606) ('representatives', 17, 49) ### ### ('st', 18, 3705) ('##ов', 19, 46) ### ### =============================================== v_p ================================================ ### ### ('##iary', 13073, 0) ('alexander', 1, 1) ('du', 4735, 2) ('control', 1232, 3) ('beg', 7, 4) ### ### ('federal', 6281, 5) ('council', 2365, 6) ('pl', 2899, 7) ('district', 1982, 8) ### ### ('supreme', 3324, 9) ('##ent', 16032, 10) ('v', 3653, 11) ('aide', 5747, 12) ('deputy', 722, 13) ### ### ('russia', 10, 14) ('##lov', 2, 15) ('party', 6921, 16) ('policy', 5694, 17) ### ### ('presidential', 14327, 18) ('department', 12376, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('alexander', 1, 1) ('##lov', 2, 15) ('beg', 7, 4) ('petersburg', 3, 29) ('.', 4, 32) ### ### ('russia', 10, 14) ('russian', 9, 37) ('?', 20, 42) ('representatives', 17, 49) ('##ов', 19, 46) ### ### ('control', 1232, 3) ('deputy', 722, 13) ('envoy', 0, 1915) ('governor', 6, 752) ('2004', 903, 21) ### ### ('district', 1982, 8) ('council', 2365, 6) ('dioxide', 21, 71) ('united', 1348, 20) ('gap', 82, 44) ### ############################################################################################################ [2023-11-24 23:37:57,023][root][INFO] - Epoch: 1: Step: 2101/19530, loss[v]=0.417921, lr=0.000002, acc@1[1]=940.0/1024=0.91796875, acc@1[2]=977.0/1024=0.9541015625 [2023-11-24 23:39:47,478][root][INFO] - Train batch 2200 [2023-11-24 23:39:47,479][root][INFO] - Avg. loss per last 100 batches: 0.645170 [2023-11-24 23:39:48,195][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.07 | max: 4.58 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.06 | max: 4.73 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the team formed as a result of neath rfc and swansea rfc combining to create a new merged ### ### entity, as part of the new regional structure of welsh rugby, that began in 2003. ### ### [P_TEXT]: swansea rfc stage of the european cup. this period also included disputes with the welsh ### ### rugby union over the way the league structure was being run in wales following rugby union's move ### ### to being a professional game, which culminated in the club's boycott of the 1998 / 99 league ### ### season, in what was referred to as the rebel season. the 2003 / 04 season has seen a significant ### ### change with the introduction of regional rugby in wales. swansea rugby football club ltd, alongside ### ### neath rfc are co - owners of the ospreys. as a result, swansea rfc returned to being an amateur ### ### team. ### ### =============================================== v_q ================================================ ### ### ('rugby', 0, 9) ('rfc', 1, 0) ('neat', 2, 4) ('swansea', 3, 1) ('.', 4, 20) ('welsh', 5, 14) ### ### (',', 6, 1839) ('##h', 7, 21) ('regional', 8, 19) ('team', 9, 43) ('wales', 10, 10) ### ### ('entity', 11, 6796) ('the', 12, 802) ('football', 13, 28) ('2003', 14, 30) ('merged', 15, 11731) ### ### ('maori', 16, 372) ('combined', 17, 8054) ('?', 18, 36) ('catalan', 19, 84) ### ### =============================================== v_p ================================================ ### ### ('rfc', 1, 0) ('swansea', 3, 1) ('ltd', 5456, 2) ('amateur', 1214, 3) ('neat', 2, 4) ### ### ('cup', 317, 5) ('boycott', 15774, 6) ('rebel', 11291, 7) ('season', 212, 8) ('rugby', 0, 9) ### ### ('wales', 10, 10) ('european', 942, 11) ('stage', 8020, 12) ('1998', 291, 13) ('welsh', 5, 14) ### ### ('club', 60, 15) ('04', 14446, 16) ('##ys', 17490, 17) ('game', 702, 18) ('regional', 8, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('rfc', 1, 0) ('swansea', 3, 1) ('neat', 2, 4) ('rugby', 0, 9) ('welsh', 5, 14) ('.', 4, 20) ### ### ('wales', 10, 10) ('regional', 8, 19) ('##h', 7, 21) ('football', 13, 28) ('2003', 14, 30) ### ### ('team', 9, 43) ('?', 18, 36) ('club', 60, 15) ('cup', 317, 5) ('season', 212, 8) ### ### ('cardiff', 43, 31) ('salt', 77, 23) ('amateur', 1214, 3) ('1998', 291, 13) ### ############################################################################################################ [2023-11-24 23:39:48,699][root][INFO] - Epoch: 1: Step: 2201/19530, loss[v]=0.922256, lr=0.000002, acc@1[1]=894.5/1024=0.87353515625, acc@1[2]=895.5/1024=0.87451171875 [2023-11-24 23:41:39,561][root][INFO] - Train batch 2300 [2023-11-24 23:41:39,562][root][INFO] - Avg. loss per last 100 batches: 0.621532 [2023-11-24 23:41:40,294][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.06 | max: 4.39 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.06 | max: 4.49 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: brian snoddy is an artist whose work has appeared in role - playing game s. brian snoddy ### ### formed privateer press with friend matt wilson and writer matt staroscik to publish their own d20 ### ### books. ### ### [P_TEXT]: matt wilson ( artist ) matthew d. " matt " wilson is an artist whose work has appeared in ### ### role - playing game s. he is one of the founders, the owner, and the ceo of privateer press. matt ### ### wilson got his start in the role - playing game industry with alderac entertainment group around ### ### 1995, and worked there as an artist and art director before doing art direction for wizards of the ### ### coast and fasa. wilson formed privateer press with friend brian snoddy and writer matt staroscik to ### ### publish their own d20 ### ### =============================================== v_q ================================================ ### ### ('brian', 0, 18) ('wilson', 1, 0) ('##no', 2, 8) ('.', 3, 28) (',', 4, 2110) ('##ddy', 5, 4) ### ### ('matt', 6, 1) ('##os', 7, 20) ('d', 8, 7) ('press', 9, 14) ('books', 10, 4929) ('star', 11, 30) ### ### ('writer', 12, 37) ('role', 13, 21) ('##ci', 14, 3) ('privateer', 15, 10) ('game', 16, 25) ### ### ('artist', 17, 19) ('the', 18, 1116) ('nick', 19, 4393) ### ### =============================================== v_p ================================================ ### ### ('wilson', 1, 0) ('matt', 6, 1) ('##sa', 6659, 2) ('##ci', 14, 3) ('##ddy', 5, 4) ### ### ('wizards', 19920, 5) ('##ac', 26165, 6) ('d', 8, 7) ('##no', 2, 8) ('entertainment', 1976, 9) ### ### ('privateer', 15, 10) ('al', 7014, 11) ('art', 29, 12) ('matthew', 3711, 13) ('press', 9, 14) ### ### ('fa', 9426, 15) ('coast', 2245, 16) ('founders', 7326, 17) ('brian', 0, 18) ('artist', 17, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('wilson', 1, 0) ('##no', 2, 8) ('matt', 6, 1) ('##ddy', 5, 4) ('brian', 0, 18) ('d', 8, 7) ### ### ('##ci', 14, 3) ('press', 9, 14) ('##os', 7, 20) ('.', 3, 28) ('privateer', 15, 10) ### ### ('role', 13, 21) ('artist', 17, 19) ('game', 16, 25) ('art', 29, 12) ('star', 11, 30) ### ### ('playing', 26, 27) ('writer', 12, 37) ('s', 20, 34) ('?', 37, 36) ### ############################################################################################################ [2023-11-24 23:41:40,798][root][INFO] - Epoch: 1: Step: 2301/19530, loss[v]=0.755674, lr=0.000002, acc@1[1]=902.5/1024=0.88134765625, acc@1[2]=931.5/1024=0.90966796875 [2023-11-24 23:43:31,676][root][INFO] - Train batch 2400 [2023-11-24 23:43:31,676][root][INFO] - Avg. loss per last 100 batches: 0.623181 [2023-11-24 23:43:32,417][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.06 | max: 4.32 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.06 | max: 4.58 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in fact, formula _ 35 can be interpreted as measures of moneyness ( in standard ### ### deviations ) and formula _ 34 as probabilities of expiring itm ( " percent moneyness " ), in the ### ### respective numeraire, as discussed below. ### ### [P_TEXT]: moneyness all inputs ), and the function can depend on the other parameters of the black ### ### – scholes model, notably time to expiry, interest rates, and implied volatility ( concretely the ### ### atm implied volatility ), yielding a function : where " s " is the spot price of the underlying, " ### ### k " is the strike price, " τ " is the time to expiry, " r " is the risk - free rate, and " σ " is ### ### the implied volatility. the forward price " f " can be computed from the spot price " s " and the ### ### risk - free rate " r. " all of these are observables except for the implied volatility, which can ### ### computed from the ### ### =============================================== v_q ================================================ ### ### ('##bilities', 0, 1741) ('##aire', 1, 22156) ('money', 2, 2) ('deviation', 3, 7846) ### ### ('formula', 4, 11527) ('35', 5, 979) ('.', 6, 36) (',', 7, 896) ('_', 8, 16435) ### ### ('percent', 9, 9220) ('nu', 10, 4031) ('##m', 11, 13373) ('##ness', 12, 9) ('pro', 13, 10053) ### ### ('the', 14, 354) ('##mer', 15, 9482) ('ex', 16, 24) ('##edance', 17, 624) ('##ba', 18, 17579) ### ### ('assets', 19, 90) ### ### =============================================== v_p ================================================ ### ### ('##lity', 980, 0) ('underlying', 7631, 1) ('money', 2, 2) ('rate', 1459, 3) ('atm', 25462, 4) ### ### ('risk', 705, 5) ('price', 483, 6) ('rates', 7137, 7) ('r', 8903, 8) ('##ness', 12, 9) ### ### ('forward', 6082, 10) ('strike', 3215, 11) ('##holes', 20491, 12) ('σ', 10728, 13) ### ### ('implied', 18954, 14) ('spot', 9942, 15) ('k', 6364, 16) ('vol', 5006, 17) ('##ati', 12086, 18) ### ### ('f', 5924, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('money', 2, 2) ('##ness', 12, 9) ('.', 6, 36) ('ex', 16, 24) ('##pi', 22, 23) ('model', 98, 22) ### ### ('##lity', 980, 0) ('price', 483, 6) ('risk', 705, 5) ('?', 20, 55) ('##bilities', 0, 1741) ### ### ('cash', 28, 46) ('rate', 1459, 3) ('35', 5, 979) (',', 7, 896) ('assets', 19, 90) ### ### ('wealth', 63, 45) ('the', 14, 354) ('bother', 32, 68) ('strike', 3215, 11) ### ############################################################################################################ [2023-11-24 23:43:32,918][root][INFO] - Epoch: 1: Step: 2401/19530, loss[v]=0.393481, lr=0.000002, acc@1[1]=949.0/1024=0.9267578125, acc@1[2]=980.5/1024=0.95751953125 [2023-11-24 23:45:23,481][root][INFO] - Train batch 2500 [2023-11-24 23:45:23,482][root][INFO] - Avg. loss per last 100 batches: 0.560124 [2023-11-24 23:45:24,206][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.06 | max: 4.34 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.05 | max: 4.64 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.3/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: yeshwanthpur is a sub locality in the north western part of bengaluru city in the india n ### ### state of karnataka. ### ### [P_TEXT]: bangalore 2014, hence bangalore was renamed to " bengaluru " on 1 november 2014. a ### ### discovery of stone age artefacts during the 2001 census of india at jalahalli, sidhapura and ### ### jadigenahalli, all of which are located on bangalore's outskirts today, suggest probable human ### ### settlement around 4000 bce. around 1, 000 bce ( iron age ), burial grounds were established at ### ### koramangala and chikkajala on the outskirts of bangalore. coins of the roman emperor s augustus, ### ### tiberius, and claudius found at yeswanthpur ### ### =============================================== v_q ================================================ ### ### ('bengal', 0, 6) ('##nt', 1, 5686) ('##ur', 2, 14789) ('.', 3, 33) ('##hwa', 4, 23950) ### ### ('karnataka', 5, 84) (',', 6, 1521) ('##uru', 7, 17) ('yes', 8, 5) ('india', 9, 32) ### ### ('##hp', 10, 15162) ('the', 11, 928) ('city', 12, 9964) ('north', 13, 16325) ('##nts', 14, 11789) ### ### ('locality', 15, 19284) ('sub', 16, 16496) ('n', 17, 13708) ('indian', 18, 51) ('pinyin', 19, 2172) ### ### =============================================== v_p ================================================ ### ### ('bangalore', 4255, 0) ('claudius', 24835, 1) ('##all', 27254, 2) ('tiberius', 25988, 3) ### ### ('outskirts', 3547, 4) ('yes', 8, 5) ('bengal', 0, 6) ('sid', 14624, 7) ('##lli', 16402, 8) ### ### ('##rama', 19396, 9) ('2014', 3712, 10) ('##pur', 1078, 11) ('chi', 9332, 12) ('age', 2802, 13) ### ### ('##jal', 28735, 14) ('bce', 7590, 15) ('##wan', 19145, 16) ('##uru', 7, 17) ('iron', 9844, 18) ### ### ('##la', 6731, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('bengal', 0, 6) ('yes', 8, 5) ('##uru', 7, 17) ('.', 3, 33) ('india', 9, 32) ('karnataka', 5, 84) ### ### ('indian', 18, 51) ('?', 20, 53) ('##pur', 1078, 11) ('hamlet', 126, 40) (',', 6, 1521) ### ### ('bangalore', 4255, 0) ('outskirts', 3547, 4) ('dioxide', 25, 66) ('mysore', 56, 57) ### ### ('age', 2802, 13) ('##nt', 1, 5686) ('the', 11, 928) ('2014', 3712, 10) ('stone', 2804, 25) ### ############################################################################################################ [2023-11-24 23:45:24,710][root][INFO] - Epoch: 1: Step: 2501/19530, loss[v]=0.518248, lr=0.000003, acc@1[1]=943.0/1024=0.9208984375, acc@1[2]=979.5/1024=0.95654296875 [2023-11-24 23:47:15,693][root][INFO] - Train batch 2600 [2023-11-24 23:47:15,694][root][INFO] - Avg. loss per last 100 batches: 0.584746 [2023-11-24 23:47:16,410][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.08 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.05 | max: 4.36 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the band released two albums in 2002 and in 2010, plus a live album in 2003, and features ### ### four different singers : russell allen ( symphony x ), damian wilson ( threshold, headspace ), dan ### ### swano ( edge of sanity, nightingale ), and floor jansen ( after forever, revamp, nightwish ). ### ### [P_TEXT]: damian wilson damian wilson is an english musician, songwriter and vocalist who has ### ### appeared on over 70 separate album releases. damian is mostly known in the progressive metal genre, ### ### as the lead singer of headspace and ( formerly ) threshold. he also sang and toured with bands and ### ### projects such as for star one, ayreon, landmarq, praying mantis and with rick wakeman's english ### ### rock ensemble. as a solo artist he has released 5 solo albums, a dvd and a retrospective ### ### compilation album. damian began recording ### ### =============================================== v_q ================================================ ### ### ('swan', 0, 2002) ('threshold', 1, 0) ('##pace', 2, 7) ('##sen', 3, 463) ('forever', 4, 1833) ### ### ('floor', 5, 669) ('symphony', 6, 7874) ('x', 7, 6719) ('jan', 8, 2580) ('wilson', 9, 4) ### ### ('allen', 10, 9578) ('nightingale', 11, 19834) ('.', 12, 28) ('night', 13, 1575) ### ### ('russell', 14, 1216) ('band', 15, 57) (',', 16, 919) ('sanity', 17, 19084) ('edge', 18, 178) ### ### ('dan', 19, 4315) ### ### =============================================== v_p ================================================ ### ### ('threshold', 1, 0) ('wake', 1559, 1) ('damian', 20, 2) ('##yre', 17063, 3) ('wilson', 9, 4) ### ### ('ensemble', 5713, 5) ('##tis', 18638, 6) ('##pace', 2, 7) ('rick', 3346, 8) ('praying', 23118, 9) ### ### ('##mar', 14260, 10) ('land', 2107, 11) ('retrospective', 27786, 12) ('##man', 4000, 13) ### ### ('star', 10904, 14) ('one', 1205, 15) ('genre', 9658, 16) ('rock', 664, 17) ('##q', 12351, 18) ### ### ('separate', 469, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('threshold', 1, 0) ('##pace', 2, 7) ('wilson', 9, 4) ('damian', 20, 2) ('.', 12, 28) ### ### ('heads', 25, 27) ('band', 15, 57) ('singer', 34, 38) ('album', 29, 42) ('head', 36, 36) ### ### ('##sen', 3, 463) ('?', 35, 46) ('edge', 18, 178) ('swan', 0, 2002) ('singers', 22, 95) ### ### ('floor', 5, 669) ('space', 30, 59) ('wake', 1559, 1) (',', 16, 919) ('separate', 469, 19) ### ############################################################################################################ [2023-11-24 23:47:16,911][root][INFO] - Epoch: 1: Step: 2601/19530, loss[v]=0.808300, lr=0.000003, acc@1[1]=886.0/1024=0.865234375, acc@1[2]=912.0/1024=0.890625 [2023-11-24 23:49:07,573][root][INFO] - Train batch 2700 [2023-11-24 23:49:07,573][root][INFO] - Avg. loss per last 100 batches: 0.521159 [2023-11-24 23:49:08,300][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.14 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.05 | max: 4.59 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the winner of the champions league qualifies for the following year's champions league, ### ### the uefa super cup and the fifa club world cup. ### ### [P_TEXT]: uefa super cup height ( the uefa champions league trophy weighs and the uefa europa ### ### league trophy ). the new model, which is a larger version of the previous trophy was introduced in ### ### 2006 and weighs and measures in height. until 2008, a team which won three times in a row or five ### ### in total received an original copy of the trophy and a special mark of recognition. since then, the ### ### original trophy has been kept exclusively by uefa. milan and barcelona have achieved this honour, ### ### winning a total of five times each but ### ### =============================================== v_q ================================================ ### ### ('uefa', 0, 0) ('.', 1, 35) ('champions', 2, 18) (',', 3, 3816) ('fifa', 4, 924) ('league', 5, 12) ### ### ('cup', 6, 5) ('qu', 7, 20007) ('club', 8, 687) ('super', 9, 3) ('the', 10, 1062) ### ### ('world', 11, 5741) ('winner', 12, 399) ('clubs', 13, 4634) ('{', 14, 1412) ('string', 15, 669) ### ### ('?', 16, 59) ('##fies', 17, 15440) ('case', 18, 194) ('bother', 19, 191) ### ### =============================================== v_p ================================================ ### ### ('uefa', 0, 0) ('height', 2378, 1) ('weighs', 8222, 2) ('super', 9, 3) ('milan', 3019, 4) ### ### ('cup', 6, 5) ('barcelona', 632, 6) ('weighed', 4087, 7) ('trophy', 1437, 8) ### ### ('recognition', 6629, 9) ('honour', 10557, 10) ('special', 803, 11) ('league', 5, 12) ### ### ('model', 1331, 13) ('mark', 5816, 14) ('weighing', 16265, 15) ('weight', 9021, 16) ### ### ('2006', 2932, 17) ('champions', 2, 18) ('copy', 12283, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('uefa', 0, 0) ('cup', 6, 5) ('super', 9, 3) ('league', 5, 12) ('champions', 2, 18) ('.', 1, 35) ### ### ('barcelona', 632, 6) ('height', 2378, 1) ('fifa', 4, 924) ('special', 803, 11) ('?', 16, 59) ### ### ('milan', 3019, 4) ('trophy', 1437, 8) ('club', 8, 687) ('dioxide', 27, 44) ('model', 1331, 13) ### ### ('size', 201, 33) ('nato', 34, 49) ('weighed', 4087, 7) ('team', 954, 22) ### ############################################################################################################ [2023-11-24 23:49:08,801][root][INFO] - Epoch: 1: Step: 2701/19530, loss[v]=0.295906, lr=0.000003, acc@1[1]=963.0/1024=0.9404296875, acc@1[2]=993.5/1024=0.97021484375 [2023-11-24 23:50:59,533][root][INFO] - Train batch 2800 [2023-11-24 23:50:59,534][root][INFO] - Avg. loss per last 100 batches: 0.523413 [2023-11-24 23:51:00,279][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.16 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.05 | max: 4.59 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the " classic 39 " syndicated episodes of " the honeymooners " were filmed using ### ### electronicam ( as well as the daily five - minute syndicated series " les paul & mary ford at home ### ### " in 1954 – 55 ), but with the introduction of a practical videotape recorder only one year away, ### ### the electronicam system never saw widespread use. ### ### [P_TEXT]: the honeymooners remaining on his contract gleason said, " the excellence of the material ### ### could not be maintained, and i had too much fondness for the show to cheapen it. ” gleason ### ### subsequently sold the films of the " classic 39 " episodes of the show to cbs for $ 1. 5 million. ### ### in 1955, many television shows ( including " the jackie gleason show " ) were performed live and ### ### recorded using kinescope technology, though sitcoms already largely were recorded on film, e. g., " ### ### amos'n'andy ", " the adventures of ozzie and harriet ", " my little margie ", and " ### ### =============================================== v_q ================================================ ### ### ('les', 0, 9520) ('honeymoon', 1, 1) ('ford', 2, 6536) ('electronic', 3, 2954) ('##am', 4, 15951) ### ### ('recorder', 5, 11504) ('mary', 6, 3436) ('.', 7, 59) ('syndicated', 8, 18374) ('paul', 9, 10306) ### ### ('classic', 10, 14) (',', 11, 2706) ('##ta', 12, 12064) ('home', 13, 2506) ('##ers', 14, 8) ### ### ('system', 15, 2168) ('systems', 16, 1327) ('practical', 17, 3689) ('episode', 18, 40) ### ### ('the', 19, 715) ### ### =============================================== v_p ================================================ ### ### ('amos', 17497, 0) ('honeymoon', 1, 1) ('cheap', 3042, 2) ('harriet', 11528, 3) ### ### ('##cope', 25572, 4) ('andy', 11024, 5) ('oz', 9125, 6) ('kin', 8136, 7) ('##ers', 14, 8) ### ### ('sitcom', 2337, 9) ('mar', 1326, 10) ('excellence', 8123, 11) ('adventures', 5536, 12) ### ### ('jackie', 9206, 13) ('classic', 10, 14) ('##gie', 3136, 15) ('g', 5001, 16) ('##son', 9712, 17) ### ### ('cbs', 3897, 18) ('39', 25, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('honeymoon', 1, 1) ('##ers', 14, 8) ('classic', 10, 14) ('episodes', 22, 21) ('39', 25, 19) ### ### ('episode', 18, 40) ('film', 26, 26) ('.', 7, 59) ('1955', 43, 34) ('37', 41, 47) ### ### ('cheap', 3042, 2) ('?', 33, 61) ('mar', 1326, 10) ('1954', 23, 120) ('1956', 48, 50) ### ### ('television', 338, 32) ('material', 887, 24) ('sitcom', 2337, 9) ('35', 50, 55) ('sold', 1204, 22) ### ############################################################################################################ [2023-11-24 23:51:00,781][root][INFO] - Epoch: 1: Step: 2801/19530, loss[v]=0.333504, lr=0.000003, acc@1[1]=958.5/1024=0.93603515625, acc@1[2]=985.5/1024=0.96240234375 [2023-11-24 23:52:51,685][root][INFO] - Train batch 2900 [2023-11-24 23:52:51,686][root][INFO] - Avg. loss per last 100 batches: 0.519804 [2023-11-24 23:52:52,415][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.08 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.44 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: roughly scythe - shaped, it extends immediately from the delta of the buna in the north, ### ### across the port city of shengjin, to the cape of rodon in the south. ### ### [P_TEXT]: albanian adriatic sea coast the albanian adriatic sea coast ( — ) stretches in the ### ### southeastern adriatic sea beginning at the gulf of drin in the north, across the port cities of ### ### shengjin, durres and vlore, to the bay of vlore in the south, where the albanian riviera as well as ### ### the albanian ionian sea coast begins. albania is geographically located in south and southeast ### ### europe within the balkan peninsula. it borders on montenegro to the northwest, kosovo ### ### =============================================== v_q ================================================ ### ### ('##ythe', 0, 23923) ('cape', 1, 4740) ('##jin', 2, 2) ('sheng', 3, 19) ('bun', 4, 18088) ### ### ('.', 5, 35) ('delta', 6, 6913) (',', 7, 1683) ('rod', 8, 11961) ('port', 9, 17) ### ### ('##on', 10, 11161) ('sc', 11, 16771) ('##a', 12, 185) ('the', 13, 1092) ('city', 14, 323) ### ### ('size', 15, 154) ('north', 16, 42) ('across', 17, 24) ('shaped', 18, 22324) ('?', 19, 37) ### ### =============================================== v_p ================================================ ### ### ('##vier', 13076, 0) ('##lore', 25266, 1) ('##jin', 2, 2) ('balkan', 13885, 3) ('gulf', 1232, 4) ### ### ('albanian', 7079, 5) ('coast', 1812, 6) ('montenegro', 10555, 7) ('kosovo', 8287, 8) ### ### ('##rre', 8622, 9) ('southeast', 1864, 10) ('albania', 2223, 11) ('dr', 3816, 12) ('du', 4509, 13) ### ### ('##in', 9171, 14) ('peninsula', 2319, 15) ('adriatic', 28384, 16) ('port', 9, 17) ('v', 3464, 18) ### ### ('sheng', 3, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##jin', 2, 2) ('sheng', 3, 19) ('port', 9, 17) ('.', 5, 35) ('across', 17, 24) ('north', 16, 42) ### ### ('gulf', 1232, 4) ('##a', 12, 185) ('south', 56, 30) ('coast', 1812, 6) ('stretched', 98, 25) ### ### ('?', 19, 37) ('cape', 1, 4740) (',', 7, 1683) ('southeast', 1864, 10) ('northwest', 1062, 20) ### ### ('size', 15, 154) ('albania', 2223, 11) ('bay', 1008, 21) ('peninsula', 2319, 15) ### ############################################################################################################ [2023-11-24 23:52:52,917][root][INFO] - Epoch: 1: Step: 2901/19530, loss[v]=0.564588, lr=0.000003, acc@1[1]=908.5/1024=0.88720703125, acc@1[2]=944.5/1024=0.92236328125 [2023-11-24 23:54:43,549][root][INFO] - Train batch 3000 [2023-11-24 23:54:43,550][root][INFO] - Avg. loss per last 100 batches: 0.482366 [2023-11-24 23:54:44,292][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.26 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.43 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 2018, the museum of modern art in new york city mounted a retrospective exhibition " ### ### judson dance theater : the work ### ### [P_TEXT]: judson dance theater music theory with john cage. " a concert of dance ", the first ### ### judson concert, took place on july 6, 1962, and included the work of 14 choreographers performed by ### ### 17 people, some of whom were students in the dunn composition class. other performers in the ### ### concert were members of the merce cunningham dance company, as well as visual artists, filmmakers, ### ### and composers. the concert included works by yvonne rainer, steve paxton, david gordon, alex and ### ### deborah hay, fred herko, elaine summers ### ### =============================================== v_q ================================================ ### ### ('##dson', 0, 0) ('.', 1, 36) (',', 2, 4118) ('dance', 3, 7) ('museum', 4, 3026) ('ju', 5, 9) ### ### ('modern', 6, 4242) ('art', 7, 500) ('york', 8, 689) ('2018', 9, 7362) ('theater', 10, 18) ### ### ('the', 11, 510) ('museums', 12, 4349) ('utah', 13, 100) ('theatre', 14, 31) ### ### ('exhibition', 15, 10697) ('contemporary', 16, 4339) ('retrospective', 17, 27227) ### ### ('knelt', 18, 674) ('archaeology', 19, 142) ### ### =============================================== v_p ================================================ ### ### ('##dson', 0, 0) ('paxton', 10456, 1) ('dunn', 15938, 2) ('gordon', 4324, 3) ### ### ('cunningham', 5125, 4) ('hay', 5708, 5) ('alex', 4770, 6) ('dance', 3, 7) ('rainer', 24633, 8) ### ### ('ju', 5, 9) ('yvonne', 14868, 10) ('cage', 11529, 11) ('elaine', 11209, 12) ('theory', 420, 13) ### ### ('mer', 10339, 14) ('her', 1925, 15) ('deborah', 16683, 16) ('choreographer', 4151, 17) ### ### ('theater', 10, 18) ('composition', 11148, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##dson', 0, 0) ('dance', 3, 7) ('ju', 5, 9) ('theater', 10, 18) ('.', 1, 36) ('theatre', 14, 31) ### ### ('theory', 420, 13) ('work', 21, 35) ('art', 7, 500) ('york', 8, 689) ('music', 491, 25) ### ### ('cabin', 56, 43) ('utah', 13, 100) ('david', 1325, 20) ('her', 1925, 15) ('gordon', 4324, 3) ### ### ('the', 11, 510) ('class', 1445, 21) ('cunningham', 5125, 4) ('ballet', 43, 48) ### ############################################################################################################ [2023-11-24 23:54:44,793][root][INFO] - Epoch: 1: Step: 3001/19530, loss[v]=0.275538, lr=0.000003, acc@1[1]=957.5/1024=0.93505859375, acc@1[2]=991.5/1024=0.96826171875 [2023-11-24 23:56:35,640][root][INFO] - Train batch 3100 [2023-11-24 23:56:35,641][root][INFO] - Avg. loss per last 100 batches: 0.471210 [2023-11-24 23:56:36,364][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.05 | max: 4.21 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.05 | max: 4.43 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 1280 charles of anjou ordered the giustiziere of basilicata the supply of 400 salme ( ### ### equal to 185 liters ) of " vino rubeo melfi e " ( red wine of melfi ) when planning a summer stay ### ### at castel lagopesole ### ### [P_TEXT]: vulture ( region ) area takes its name from the extinct volcano monte vulture ( 1326 m ). ### ### sights include the two lakes of monticchio within the crater of the volcano, and the castles of ### ### frederick ii of hohenstaufen at castel lagopesole and melfi. the fertile volcanic soil of the ### ### vulture is suitable for the cultivation of grapes and olives. the doc wine aglianico del vulture is ### ### produced in the region, as are aglianico dolce, a dessert wine, aglianico di filiano, malvasia del ### ### vulture and moscato del vulture. the area of monte vulture has numerous ### ### =============================================== v_q ================================================ ### ### ('wine', 0, 21) ('basilica', 1, 715) ('mel', 2, 30) ('rub', 3, 15845) ('.', 4, 70) ### ### ('##ust', 5, 20895) ('##pes', 6, 9) ('sal', 7, 12380) ('charles', 8, 2041) ('vin', 9, 8001) ### ### ('##iz', 10, 18212) ('##fi', 11, 18) (',', 12, 1906) ('##jou', 13, 29342) ('caste', 14, 6) ### ### ('la', 15, 45) ('##me', 16, 9619) ('##eo', 17, 16883) ('e', 18, 3566) ('##go', 19, 35) ### ### =============================================== v_p ================================================ ### ### ('vulture', 23777, 0) ('crater', 2209, 1) ('frederick', 88, 2) ('mont', 1332, 3) ### ### ('extinct', 1019, 4) ('##chio', 24757, 5) ('caste', 14, 6) ('grapes', 1084, 7) ('mal', 10400, 8) ### ### ('##pes', 6, 9) ('volcano', 10488, 10) ('##ico', 12855, 11) ('##ole', 33, 12) ('monte', 16385, 13) ### ### ('soil', 1458, 14) ('olive', 25689, 15) ('volcanic', 128, 16) ('##vas', 12015, 17) ('##fi', 11, 18) ### ### ('##lce', 29055, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('wine', 0, 21) ('##pes', 6, 9) ('##fi', 11, 18) ('caste', 14, 6) ('mel', 2, 30) ('##go', 19, 35) ### ### ('la', 15, 45) ('##ole', 33, 12) ('.', 4, 70) ('frederick', 88, 2) ('fi', 78, 27) ### ### ('volcanic', 128, 16) ('basilica', 1, 715) ('castle', 29, 57) ('?', 28, 61) ('extinct', 1019, 4) ### ### ('mont', 1332, 3) ('crater', 2209, 1) ('grapes', 1084, 7) ('soil', 1458, 14) ### ############################################################################################################ [2023-11-24 23:56:36,868][root][INFO] - Epoch: 1: Step: 3101/19530, loss[v]=0.657111, lr=0.000003, acc@1[1]=903.5/1024=0.88232421875, acc@1[2]=936.0/1024=0.9140625 [2023-11-24 23:58:27,281][root][INFO] - Train batch 3200 [2023-11-24 23:58:27,282][root][INFO] - Avg. loss per last 100 batches: 0.476037 [2023-11-24 23:58:28,015][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.09 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.40 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the town of eston lies at the foot of eston hills, a ridge approximately above sea level, ### ### and a part of the cleveland hills. ### ### [P_TEXT]: bolckow, vaughan park colliery to make coke, and ironstone from whitby on the coast. the ### ### pig iron produced at witton was transported to middlesbrough for further forging or casting. in ### ### 1850, vaughan and his mining geologist john marley discovered iron ore, conveniently situated near ### ### eston in the cleveland hills of yorkshire. unknown to anyone at the time, this vein was part of the ### ### cleveland ironstone formation, which was already being mined in grosmont by losh, wilson and bell. ### ### to ### ### =============================================== v_q ================================================ ### ### ('est', 0, 3) ('cleveland', 1, 9) ('##on', 2, 34) ('.', 3, 45) (',', 4, 1723) ('hills', 5, 23) ### ### ('the', 6, 1752) ('town', 7, 2239) ('ridge', 8, 2042) ('lies', 9, 198) ('sea', 10, 3706) ### ### ('level', 11, 794) ('kent', 12, 639) ('hill', 13, 65) ('##ons', 14, 517) ('?', 15, 53) ### ### ('anger', 16, 95) ('ohio', 17, 930) ('polymer', 18, 487) ('harsh', 19, 76) ### ### =============================================== v_p ================================================ ### ### ('marley', 12536, 0) ('bell', 9991, 1) ('##stone', 5597, 2) ('est', 0, 3) ('vaughan', 7073, 4) ### ### ('wit', 13060, 5) ('wilson', 1637, 6) ('##hit', 14667, 7) ('yorkshire', 707, 8) ('cleveland', 1, 9) ### ### ('los', 7865, 10) ('##kow', 22535, 11) ('coke', 21868, 12) ('pig', 18415, 13) ('##mont', 8722, 14) ### ### ('bo', 7825, 15) ('park', 1478, 16) ('##by', 14929, 17) ('##ging', 6678, 18) ('casting', 11444, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('est', 0, 3) ('cleveland', 1, 9) ('hills', 5, 23) ('##on', 2, 34) ('.', 3, 45) ('##ton', 60, 21) ### ### ('yorkshire', 707, 8) ('wilson', 1637, 6) ('mining', 201, 32) ('park', 1478, 16) ### ### ('##stone', 5597, 2) (',', 4, 1723) ('?', 15, 53) ('hill', 13, 65) ('anyone', 1509, 27) ### ### ('##os', 1755, 29) ('vaughan', 7073, 4) ('formation', 2686, 22) ('lies', 9, 198) ('bell', 9991, 1) ### ############################################################################################################ [2023-11-24 23:58:28,515][root][INFO] - Epoch: 1: Step: 3201/19530, loss[v]=0.604083, lr=0.000003, acc@1[1]=911.5/1024=0.89013671875, acc@1[2]=939.5/1024=0.91748046875 [2023-11-25 00:00:18,908][root][INFO] - Train batch 3300 [2023-11-25 00:00:18,909][root][INFO] - Avg. loss per last 100 batches: 0.458698 [2023-11-25 00:00:19,648][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.05 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.34 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: on 15 may 1847 the line was extended to hamm and duisburg station became a through ### ### station on the line from dusseldorf to oberhausen. ### ### [P_TEXT]: dusseldorf - oberbilk and meerbusch. dusseldorf - oberbilk station connects it with the ### ### central station, the international airport, other parts of dusseldorf, cologne, leverkusen, ### ### solingen, ratingen ost, duisburg, mulheim and essen. there are two tram lines and some bus lines as ### ### well. oberbilk was a part of bilk from 1384. there were only a few farms then and agriculture ### ### dominated until 1838 when dusseldorf - elberfelder eisenbahn took its railway through oberbilk ; ### ### the coln - mindener eisenbahn line followed in ### ### =============================================== v_q ================================================ ### ### ('dusseldorf', 0, 0) ('ham', 1, 7150) ('##ausen', 2, 12276) ('.', 3, 46) ('##burg', 4, 4) ### ### (',', 5, 738) ('station', 6, 26) ('du', 7, 28) ('line', 8, 38) ('##m', 9, 6631) ('##rh', 10, 27666) ### ### ('##is', 11, 57) ('the', 12, 247) ('railway', 13, 47) ('obe', 14, 23) ('1847', 15, 19862) ### ### ('berlin', 16, 166) ('##ham', 17, 6214) ('ah', 18, 1333) ('?', 19, 58) ### ### =============================================== v_p ================================================ ### ### ('dusseldorf', 0, 0) ('##ingen', 25360, 1) ('##feld', 4796, 2) ('lever', 9402, 3) ('##burg', 4, 4) ### ### ('airport', 25, 5) ('sol', 4848, 6) ('##bahn', 10204, 7) ('rating', 18053, 8) ('central', 1863, 9) ### ### ('##lk', 21956, 10) ('el', 8615, 11) ('##ber', 10612, 12) ('##rb', 18553, 13) ('col', 8725, 14) ### ### ('##kus', 17381, 15) ('##bus', 28074, 16) ('essen', 24396, 17) ('bi', 2204, 18) ('mind', 8008, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('dusseldorf', 0, 0) ('##burg', 4, 4) ('station', 6, 26) ('du', 7, 28) ('line', 8, 38) ('.', 3, 46) ### ### ('obe', 14, 23) ('airport', 25, 5) ('railway', 13, 47) ('##is', 11, 57) ('lines', 70, 34) ### ### (',', 5, 738) ('?', 19, 58) ('the', 12, 247) ('central', 1863, 9) ('##feld', 4796, 2) ### ### ('agriculture', 1293, 24) ('bi', 2204, 18) ('ham', 1, 7150) ('bus', 2299, 21) ### ############################################################################################################ [2023-11-25 00:00:20,152][root][INFO] - Epoch: 1: Step: 3301/19530, loss[v]=0.673872, lr=0.000003, acc@1[1]=920.0/1024=0.8984375, acc@1[2]=927.5/1024=0.90576171875 [2023-11-25 00:02:10,639][root][INFO] - Train batch 3400 [2023-11-25 00:02:10,640][root][INFO] - Avg. loss per last 100 batches: 0.462063 [2023-11-25 00:02:11,352][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.07 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.48 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the first vicar of this church, from 1887 to 1894, was charles farrar forster. ### ### [P_TEXT]: charles farrar forster charles farrar forster ( 29 february 1848 – 28 august 1894 ) was ### ### curate of the parish of lockwood near huddersfield, vicar of st andrew's church in huddersfield, ### ### and the first vicar of the church of st michael and all angels, beckwithshaw. in spite of suffering ### ### severe pain due to heart disease during his working years, he was known for his dedication to ### ### parochial work and for the large congregations who came to hear his sermons. for all his bodily ### ### fragility he was notable for his personal charisma, causing various newspapers to remark upon ### ### =============================================== v_q ================================================ ### ### ('forster', 0, 0) ('vicar', 1, 8) ('.', 2, 65) ('##rar', 3, 5) ('far', 4, 12) (',', 5, 6447) ### ### ('church', 6, 26) ('charles', 7, 9) ('the', 8, 3758) ('1887', 9, 10058) ('1894', 10, 37) ### ### ('much', 11, 231) ('1893', 12, 64) ('1885', 13, 717) ('prehistoric', 14, 167) ('1902', 15, 101) ### ### ('1987', 16, 9299) ('bishop', 17, 1261) ('parish', 18, 16) ('uk', 19, 718) ### ### =============================================== v_p ================================================ ### ### ('forster', 0, 0) ('lock', 14228, 1) ('huddersfield', 19843, 2) ('##wood', 19341, 3) ### ### ('##with', 27694, 4) ('##rar', 3, 5) ('char', 2608, 6) ('##shaw', 28525, 7) ('vicar', 1, 8) ### ### ('charles', 7, 9) ('heart', 12224, 10) ('andrew', 3407, 11) ('far', 4, 12) ('beck', 14379, 13) ### ### ('angels', 14808, 14) ('michael', 1878, 15) ('parish', 18, 16) ('##gil', 22022, 17) ### ### ('hear', 3538, 18) ('sermons', 18593, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('forster', 0, 0) ('vicar', 1, 8) ('##rar', 3, 5) ('far', 4, 12) ('charles', 7, 9) ### ### ('church', 6, 26) ('.', 2, 65) ('1894', 10, 37) ('parish', 18, 16) ('1893', 12, 64) ### ### ('1994', 20, 54) ('char', 2608, 6) ('pain', 1013, 25) ('michael', 1878, 15) ('1902', 15, 101) ### ### ('causing', 630, 32) ('?', 24, 69) ('leeds', 207, 39) ('andrew', 3407, 11) ('much', 11, 231) ### ############################################################################################################ [2023-11-25 00:02:11,852][root][INFO] - Epoch: 1: Step: 3401/19530, loss[v]=0.349157, lr=0.000003, acc@1[1]=958.5/1024=0.93603515625, acc@1[2]=989.5/1024=0.96630859375 [2023-11-25 00:04:01,999][root][INFO] - Train batch 3500 [2023-11-25 00:04:02,000][root][INFO] - Avg. loss per last 100 batches: 0.427837 [2023-11-25 00:04:02,719][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.09 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.60 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the team advanced through the tournament, defeating number one seeds baylor and stanford ### ### en route to the championship against notre dame. ### ### [P_TEXT]: notre dame fighting irish women's basketball 21 tries. that win was followed by an upset ### ### of the number one - ranked uconn huskies ( making notre dame the first team ever to beat both ### ### tennessee and uconn in the same tournament ) to advance the fighting irish to the 2011 championship ### ### game, where it lost to texas a & m. the irish would return to the championship game in 2012, losing ### ### to unbeaten baylor after winning the big east regular season title and beating uconn again to reach ### ### the final. in the 2012 – 2013 season, the irish, led by diggins and shooting guard kayla mcbride, ### ### =============================================== v_q ================================================ ### ### ('stanford', 0, 830) ('notre', 1, 0) ('seeds', 2, 431) ('.', 3, 55) (',', 4, 5660) ('baylor', 5, 4) ### ### ('dame', 6, 5) ('tournament', 7, 24) ('team', 8, 53) ('the', 9, 8190) ('seed', 10, 2229) ### ### ('advanced', 11, 74) ('championship', 12, 34) ('harvard', 13, 1174) ('texas', 14, 10) ### ### ('seeded', 15, 3616) ('cavalry', 16, 90) ('brand', 17, 1590) ('indiana', 18, 2209) ### ### ('tournaments', 19, 161) ### ### =============================================== v_p ================================================ ### ### ('notre', 1, 0) ('kayla', 8216, 1) ('huskies', 6634, 2) ('mcbride', 18880, 3) ('baylor', 5, 4) ### ### ('dame', 6, 5) ('##gins', 26107, 6) ('uc', 13727, 7) ('basketball', 5914, 8) ('fighting', 1037, 9) ### ### ('texas', 14, 10) ('tennessee', 106, 11) ('##on', 20697, 12) ('irish', 631, 13) ('big', 1650, 14) ### ### ('shooting', 6254, 15) ('advance', 65, 16) ('east', 5743, 17) ('dig', 12127, 18) ('21', 5208, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('notre', 1, 0) ('baylor', 5, 4) ('dame', 6, 5) ('tournament', 7, 24) ('.', 3, 55) ### ### ('texas', 14, 10) ('championship', 12, 34) ('team', 8, 53) ('tennessee', 106, 11) ### ### ('advance', 65, 16) ('stanford', 0, 830) ('advanced', 11, 74) ('seeds', 2, 431) ### ### ('fighting', 1037, 9) ('irish', 631, 13) ('2012', 395, 25) ('else', 99, 40) ('2013', 220, 35) ### ### ('upset', 831, 22) ('big', 1650, 14) ### ############################################################################################################ [2023-11-25 00:04:03,222][root][INFO] - Epoch: 1: Step: 3501/19530, loss[v]=0.298491, lr=0.000004, acc@1[1]=968.5/1024=0.94580078125, acc@1[2]=991.5/1024=0.96826171875 [2023-11-25 00:05:53,460][root][INFO] - Train batch 3600 [2023-11-25 00:05:53,461][root][INFO] - Avg. loss per last 100 batches: 0.419280 [2023-11-25 00:05:54,198][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.09 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.05 | max: 4.21 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the regional capital is potenza. ### ### [P_TEXT]: tommy paul both his father and mother were from san fele, province of potenza, ### ### basilicata. the fifth of thirteen children, paul and his brothers bartolomeo " al ", michele " ### ### mickey ", and vito antonio " tony, jr. ", followed their brother tommy into the ring. at seventeen, ### ### paul followed his brothers into jack singer's gym, hoping for a tryout. singer recognized paul's ### ### rough abilities, and after consenting to train him, entered him in a number of tournaments. paul ### ### began boxing as an amateur in 1926, and in 1927 won the niagara district amateur bantamweight ### ### championship in march, the ### ### =============================================== v_q ================================================ ### ### ('##za', 0, 23) ('pot', 1, 15) ('.', 2, 59) ('regional', 3, 1504) (',', 4, 5668) ('##en', 5, 52) ### ### ('capital', 6, 1707) ('the', 7, 3386) ('region', 8, 853) ('za', 9, 1853) ('##ம', 10, 498) ### ### ('local', 11, 1652) ('anger', 12, 183) ('prehistoric', 13, 139) ('?', 14, 64) ('tightly', 15, 332) ### ### ('，', 16, 626) ('stuffing', 17, 360) ('php', 18, 428) ('rural', 19, 601) ### ### =============================================== v_p ================================================ ### ### ('tommy', 6252, 0) ('niagara', 14158, 1) ('basilica', 1037, 2) ('consent', 6198, 3) ### ### ('paul', 6088, 4) ('mickey', 7642, 5) ('tony', 6379, 6) ('boxing', 7328, 7) ('michele', 26070, 8) ### ### ('jack', 4886, 9) ('amateur', 5784, 10) ('rough', 6743, 11) ('gym', 9351, 12) ('antonio', 5427, 13) ### ### ('##weight', 24674, 14) ('pot', 1, 15) ('##olo', 15228, 16) ('fe', 17770, 17) ('al', 9142, 18) ### ### ('ring', 468, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('pot', 1, 15) ('##za', 0, 23) ('.', 2, 59) ('##en', 5, 52) ('basilica', 1037, 2) ### ### ('district', 180, 27) ('ring', 468, 19) ('regional', 3, 1504) ('tommy', 6252, 0) ### ### ('consent', 6198, 3) ('paul', 6088, 4) ('?', 14, 64) ('jack', 4886, 9) ('recognize', 320, 45) ### ### ('region', 8, 853) ('tony', 6379, 6) ('mickey', 7642, 5) ('amateur', 5784, 10) ('capital', 6, 1707) ### ### ('province', 987, 38) ### ############################################################################################################ [2023-11-25 00:05:54,701][root][INFO] - Epoch: 1: Step: 3601/19530, loss[v]=0.535399, lr=0.000004, acc@1[1]=930.5/1024=0.90869140625, acc@1[2]=946.5/1024=0.92431640625 [2023-11-25 00:07:45,344][root][INFO] - Train batch 3700 [2023-11-25 00:07:45,347][root][INFO] - Avg. loss per last 100 batches: 0.439973 [2023-11-25 00:07:46,075][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.05 | max: 4.07 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.05 | max: 4.39 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: a summer school for approximately 300 young musicians, now known as the tanglewood music ### ### center. ### ### [P_TEXT]: tanglewood music center the tanglewood music center is an annual summer music academy in ### ### lenox, massachusetts, united states, in which emerging professional musicians participate in ### ### performances, master classes and workshops. the center operates as a part of the tanglewood music ### ### festival, an outdoor concert series and the summer home of the boston symphony orchestra ( bso ). ### ### the tanglewood music center ( tmc ) was founded in 1940 as the berkshire music center by the bso's ### ### music director, serge koussevitzky, three years after the establishment of tanglewood as the summer ### ### home ### ### =============================================== v_q ================================================ ### ### ('tangle', 0, 1) ('.', 1, 45) ('##wood', 2, 3) ('music', 3, 9) (',', 4, 5727) ('musicians', 5, 30) ### ### ('center', 6, 6) ('young', 7, 3899) ('school', 8, 1622) ('summer', 9, 14) ('tangled', 10, 24) ### ### ('the', 11, 911) ('musician', 12, 79) ('classical', 13, 345) ('musical', 14, 46) ### ### ('centers', 15, 23) ('300', 16, 5037) ('wild', 17, 171) ('ancient', 18, 544) ('youth', 19, 1641) ### ### =============================================== v_p ================================================ ### ### ('##ox', 24406, 0) ('tangle', 0, 1) ('berkshire', 13042, 2) ('##wood', 2, 3) ('##mc', 28449, 4) ### ### ('bs', 16135, 5) ('center', 6, 6) ('##ky', 19003, 7) ('boston', 5056, 8) ('music', 3, 9) ### ### ('len', 19107, 10) ('symphony', 39, 11) ('massachusetts', 2082, 12) ('##itz', 17114, 13) ### ### ('summer', 9, 14) ('##uss', 24748, 15) ('orchestra', 295, 16) ('festival', 3640, 17) ### ### ('academy', 348, 18) ('##ev', 18558, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('tangle', 0, 1) ('##wood', 2, 3) ('music', 3, 9) ('center', 6, 6) ('summer', 9, 14) ### ### ('musicians', 5, 30) ('tangled', 10, 24) ('.', 1, 45) ('symphony', 39, 11) ('centers', 15, 23) ### ### ('centre', 23, 21) ('orchestra', 295, 16) ('musician', 12, 79) ('academy', 348, 18) ### ### ('musical', 14, 46) ('massachusetts', 2082, 12) ('school', 8, 1622) ('the', 11, 911) ### ### ('master', 1012, 22) ('?', 33, 49) ### ############################################################################################################ [2023-11-25 00:07:46,575][root][INFO] - Epoch: 1: Step: 3701/19530, loss[v]=0.322618, lr=0.000004, acc@1[1]=967.0/1024=0.9443359375, acc@1[2]=991.0/1024=0.9677734375 [2023-11-25 00:09:36,975][root][INFO] - Train batch 3800 [2023-11-25 00:09:36,978][root][INFO] - Avg. loss per last 100 batches: 0.418859 [2023-11-25 00:09:37,695][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.02 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.05 | max: 4.53 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: bohemia, to the west, consists of a basin drained by the elbe ( ) and the vltava rivers, ### ### surrounded by mostly low mountains, such as the krkonose range of the sudetes. ### ### [P_TEXT]: husinec ( prague - east district ) husinec ( ) is a municipality and village in prague - ### ### east district in the central bohemian region of the czech republic. it has about 1, 500 ### ### inhabitants. it lies on the vltava river. the " nuclear physics institute of the cas " was ### ### established in rez in 1955. since 2007 it is a public research institution. ### ### =============================================== v_q ================================================ ### ### ('.', 0, 30) ('czech', 1, 6) ('bohemia', 2, 9402) ('sud', 3, 7919) ('##ono', 4, 24118) ### ### ('##rk', 5, 16801) (',', 6, 5143) ('##lta', 7, 8) ('el', 8, 13683) ('##be', 9, 19682) ('v', 10, 11) ### ### ('##va', 11, 22) ('k', 12, 9335) ('##se', 13, 5892) ('slovakia', 14, 39) ('range', 15, 3062) ### ### ('##ete', 16, 22646) ('prague', 17, 3) ('rivers', 18, 61) ('river', 19, 13) ### ### =============================================== v_p ================================================ ### ### ('##sin', 11146, 0) ('cas', 14268, 1) ('hu', 8085, 2) ('prague', 17, 3) ('##ec', 4974, 4) ### ### ('nuclear', 1255, 5) ('czech', 1, 6) ('bohemian', 24, 7) ('##lta', 7, 8) ('research', 313, 9) ### ### ('inhabitants', 1468, 10) ('v', 10, 11) ('institute', 3298, 12) ('river', 19, 13) ### ### ('municipality', 2984, 14) ('##z', 13438, 15) ('district', 1204, 16) ('re', 11945, 17) ### ### ('2007', 3810, 18) ('1955', 7445, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('czech', 1, 6) ('##lta', 7, 8) ('prague', 17, 3) ('v', 10, 11) ('bohemian', 24, 7) ### ### ('##va', 11, 22) ('river', 19, 13) ('.', 0, 30) ('slovakia', 14, 39) ('rivers', 18, 61) ### ### ('research', 313, 9) ('nuclear', 1255, 5) ('surrounded', 32, 36) ('east', 384, 20) ### ### ('village', 302, 23) ('inhabitants', 1468, 10) ('czechoslovakia', 27, 132) ('district', 1204, 16) ### ### ('lies', 65, 33) ('##ec', 4974, 4) ### ############################################################################################################ [2023-11-25 00:09:38,199][root][INFO] - Epoch: 1: Step: 3801/19530, loss[v]=0.639777, lr=0.000004, acc@1[1]=915.0/1024=0.8935546875, acc@1[2]=938.0/1024=0.916015625 [2023-11-25 00:11:29,582][root][INFO] - Train batch 3900 [2023-11-25 00:11:29,582][root][INFO] - Avg. loss per last 100 batches: 0.423704 [2023-11-25 00:11:30,316][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.17 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.05 | max: 4.47 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: its placement within the broader division of sections of philadelphia is debated, but it ### ### is often considered as a dividing line between or belonging to both the lower northeast section of ### ### philadelphia and north philadelphia. ### ### [P_TEXT]: gloria casarez gloria casarez ( december 13, 1971 – october 19, 2014 ) was an american ### ### civil rights leader and lgbt activist in philadelphia. casarez served as philadelphia's first ### ### director of lesbian, gay, bisexual and transgender ( lgbt ) affairs. during her tenure as director, ### ### philadelphia ranked as the number one city nationwide for lgbt equality. casarez served as the ### ### executive director of gay and lesbian latino aids education initiative ( galaei ) from 1999 – 2008. ### ### casarez was born in philadelphia and was raised catholic by a single mother, elisa gonzales, she ### ### grew up in the kensington neighborhood of north philadelphia and ### ### =============================================== v_q ================================================ ### ### ('philadelphia', 0, 5) ('.', 1, 44) ('northeast', 2, 321) (',', 3, 3944) ('section', 4, 5592) ### ### ('division', 5, 6369) ('pennsylvania', 6, 42) ('sections', 7, 7594) ('line', 8, 2037) ### ### ('broader', 9, 5951) ('lower', 10, 3598) ('north', 11, 27) ('debated', 12, 21485) ('the', 13, 2516) ### ### ('dividing', 14, 25459) ('separating', 15, 384) ('placement', 16, 8803) ('southeast', 17, 2490) ### ### ('northwest', 18, 3379) ('belonging', 19, 1704) ### ### =============================================== v_p ================================================ ### ### ('gonzales', 19058, 0) ('casa', 10541, 1) ('gloria', 7423, 2) ('bisexual', 14351, 3) ### ### ('kensington', 12653, 4) ('philadelphia', 0, 5) ('gala', 7840, 6) ('aids', 19150, 7) ### ### ('initiative', 250, 8) ('equality', 13391, 9) ('##z', 13803, 10) ('gay', 10826, 11) ### ### ('latino', 6921, 12) ('elisa', 26149, 13) ('nationwide', 5682, 14) ('##ei', 18122, 15) ### ### ('transgender', 20075, 16) ('lesbian', 12098, 17) ('rights', 14893, 18) ('neighborhood', 46, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('philadelphia', 0, 5) ('.', 1, 44) ('north', 11, 27) ('pennsylvania', 6, 42) ### ### ('neighborhood', 46, 19) ('initiative', 250, 8) ('northeast', 2, 321) ('pittsburgh', 20, 54) ### ### ('separating', 15, 384) ('##з', 1151, 23) ('gloria', 7423, 2) (',', 3, 3944) ('line', 8, 2037) ### ### ('?', 84, 49) ('reading', 47, 73) ('gala', 7840, 6) ('farther', 25, 229) ('americans', 209, 43) ### ### ('bother', 143, 46) ('casa', 10541, 1) ### ############################################################################################################ [2023-11-25 00:11:30,816][root][INFO] - Epoch: 1: Step: 3901/19530, loss[v]=0.596841, lr=0.000004, acc@1[1]=912.0/1024=0.890625, acc@1[2]=942.5/1024=0.92041015625 [2023-11-25 00:13:21,463][root][INFO] - Train batch 4000 [2023-11-25 00:13:21,464][root][INFO] - Avg. loss per last 100 batches: 0.415199 [2023-11-25 00:13:22,191][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.05 | max: 4.15 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.05 | max: 4.52 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it is serviced by both the gold line and the red line during the day, and only the gold ### ### line after 9 : 00 pm. ### ### [P_TEXT]: red line ( marta ) the red line is a rapid transit line in the marta rail system. it ### ### operates between north springs and airport stations, running through sandy springs, dunwoody, ### ### atlanta, east point and college park. the red line was originally called the north - south line ### ### until marta switched to a color - based naming system in october 2009. the north - south line, from ### ### its launch, was considered one line, denoted with an orange color on old system maps until 2006 ### ### when the north branch and the northeast branch were ### ### =============================================== v_q ================================================ ### ### ('line', 0, 2) ('gold', 1, 3329) ('red', 2, 3) ('.', 3, 42) (',', 4, 3103) ('pm', 5, 20462) ### ### ('silver', 6, 434) ('00', 7, 12419) ('the', 8, 2811) ('lines', 9, 18) ('copper', 10, 1828) ### ### ('railway', 11, 189) ('day', 12, 4042) ('station', 13, 37) ('junction', 14, 4577) ### ### ('route', 15, 531) ('rail', 16, 16) ('bronze', 17, 5806) ('sour', 18, 948) ('blue', 19, 74) ### ### =============================================== v_p ================================================ ### ### ('springs', 15298, 0) ('marta', 18631, 1) ('line', 0, 2) ('red', 2, 3) ('sandy', 2177, 4) ### ### ('branch', 193, 5) ('atlanta', 1778, 6) ('color', 4232, 7) ('dun', 21849, 8) ('naming', 12816, 9) ### ### ('point', 3483, 10) ('airport', 316, 11) ('launch', 3579, 12) ('maps', 549, 13) ('##y', 14434, 14) ### ### ('college', 2305, 15) ('rail', 16, 16) ('system', 282, 17) ('lines', 9, 18) ('rapid', 3329, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('line', 0, 2) ('red', 2, 3) ('lines', 9, 18) ('rail', 16, 16) ('.', 3, 42) ('branch', 193, 5) ### ### ('station', 13, 37) ('orange', 80, 20) ('airport', 316, 11) ('transit', 43, 28) ('system', 282, 17) ### ### ('maps', 549, 13) ('sandy', 2177, 4) ('atlanta', 1778, 6) ('operates', 57, 30) ('black', 20, 50) ### ### ('sand', 32, 38) ('railway', 11, 189) ('white', 31, 41) ('gold', 1, 3329) ### ############################################################################################################ [2023-11-25 00:13:22,691][root][INFO] - Epoch: 1: Step: 4001/19530, loss[v]=0.219933, lr=0.000004, acc@1[1]=976.5/1024=0.95361328125, acc@1[2]=1003.0/1024=0.9794921875 [2023-11-25 00:15:13,401][root][INFO] - Train batch 4100 [2023-11-25 00:15:13,402][root][INFO] - Avg. loss per last 100 batches: 0.399202 [2023-11-25 00:15:14,136][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.07 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.63 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 1987, the new universe line saw a revamp under new editor - in - chief tom defalco, ### ### and byrne took over writing and art - breakdowns on the line's flagship title, " star brand " ( ### ### renamed " the star brand " during byrne's term on the book ). ### ### [P_TEXT]: star brand was unique in that, unlike nearly every other superhero comic up to that time, ### ### it was not told from omniscient point - of - view, and instead showed only ken connell's firsthand ### ### experiences. after issue # 7, shooter was fired from marvel, and romita and williamson left the ### ### series. " star brand " was taken down to bi - monthly status, and consisted of fill - in issues for ### ### the next half year. john byrne, at the prompting of new universe editor howard mackie ( a close ### ### friend of byrne's ), took over both the writing and breakdowns with issue # 11. shortly after, a ### ### meeting between ### ### =============================================== v_q ================================================ ### ### ('.', 0, 74) ('brand', 1, 0) ('universe', 2, 13) ('byrne', 3, 5) ('star', 4, 1) ('def', 5, 14940) ### ### (',', 6, 5083) ('brands', 7, 35) ('line', 8, 3872) ('tom', 9, 2046) ('##co', 10, 21947) ### ### ('flagship', 11, 10781) ('editor', 12, 15) ('breakdown', 13, 4) ('1987', 14, 5842) ### ### ('book', 15, 3215) ('##al', 16, 22227) ('writing', 17, 46) ('art', 18, 2987) ('comics', 19, 52) ### ### =============================================== v_p ================================================ ### ### ('brand', 1, 0) ('star', 4, 1) ('mack', 6734, 2) ('rom', 7295, 3) ('breakdown', 13, 4) ### ### ('byrne', 3, 5) ('shooter', 23034, 6) ('connell', 11448, 7) ('williamson', 8727, 8) ### ### ('fill', 3560, 9) ('unique', 604, 10) ('ken', 8959, 11) ('marvel', 516, 12) ('universe', 2, 13) ### ### ('##ita', 14667, 14) ('editor', 12, 15) ('stars', 20, 16) ('howard', 1064, 17) ('view', 2454, 18) ### ### ('monthly', 18777, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('brand', 1, 0) ('star', 4, 1) ('byrne', 3, 5) ('universe', 2, 13) ('breakdown', 13, 4) ### ### ('editor', 12, 15) ('brands', 7, 35) ('stars', 20, 16) ('.', 0, 74) ('writing', 17, 46) ### ### ('comics', 19, 52) ('new', 27, 44) ('comic', 99, 22) ('marvel', 516, 12) ('unique', 604, 10) ### ### ('series', 219, 27) ('howard', 1064, 17) ('issue', 352, 26) ('collapse', 35, 67) ('break', 42, 66) ### ############################################################################################################ [2023-11-25 00:15:14,640][root][INFO] - Epoch: 1: Step: 4101/19530, loss[v]=0.264213, lr=0.000004, acc@1[1]=958.0/1024=0.935546875, acc@1[2]=993.5/1024=0.97021484375 [2023-11-25 00:17:05,113][root][INFO] - Train batch 4200 [2023-11-25 00:17:05,114][root][INFO] - Avg. loss per last 100 batches: 0.380953 [2023-11-25 00:17:05,828][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.03 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.50 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: by the end of the decade, the genre became a cultural phenomenon, with standalone games ### ### such as " playerunknown's battlegrounds " ( 2017 ), " fortnite battle royale " ( 2017 ), " apex ### ### legends " ( 2019 ), and " " ( 2020 ), each having received tens of millions of players within ### ### [P_TEXT]: 2010s in culture free - to - play battle royale game s such as epic games'" fortnite ### ### battle royale " and respawn entertainment's " apex legends ". this model, dubbed " freemium ", uses ### ### in - game microtransactions and loot box es to generate revenue. these microtransactions have ### ### garnered ethical concerns ; the intrusive nature of in - game microtransactions can lead to ### ### children accidentally or purposefully rack up a high bill from purchasing in - game items, while ### ### the compulsion loop caused by loot boxes has drawn comparisons to gambling addiction. ### ### =============================================== v_q ================================================ ### ### ('.', 0, 120) ('royale', 1, 3) ('apex', 2, 9) ('##grounds', 3, 17251) ('legends', 4, 12) ### ### (',', 5, 10194) ('battle', 6, 24) ('fort', 7, 6) ('##unk', 8, 19260) ('2017', 9, 365) ### ### ('cultural', 10, 51) ('2020', 11, 11545) ('##te', 12, 31) ('game', 13, 20) ('legend', 14, 48) ### ### ('genre', 15, 4014) ('player', 16, 551) ('games', 17, 10) ('the', 18, 3892) ('2019', 19, 11954) ### ### =============================================== v_p ================================================ ### ### ('compulsion', 15081, 0) ('##mium', 25785, 1) ('addiction', 590, 2) ('royale', 1, 3) ### ### ('culture', 46, 4) ('loop', 5038, 5) ('fort', 7, 6) ('##wn', 14215, 7) ('rack', 20520, 8) ### ### ('apex', 2, 9) ('games', 17, 10) ('box', 1704, 11) ('legends', 4, 12) ('boxes', 11061, 13) ### ### ('free', 5901, 14) ('epic', 6044, 15) ('res', 13312, 16) ('gambling', 4310, 17) ### ### ('accidentally', 2763, 18) ('entertainment', 245, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('royale', 1, 3) ('apex', 2, 9) ('fort', 7, 6) ('legends', 4, 12) ('battle', 6, 24) ### ### ('games', 17, 10) ('game', 13, 20) ('##te', 12, 31) ('legend', 14, 48) ('cultural', 10, 51) ### ### ('culture', 46, 4) ('##ni', 22, 38) ('.', 0, 120) ('addiction', 590, 2) ('entertainment', 245, 19) ### ### ('2017', 9, 365) ('box', 1704, 11) ('micro', 1563, 23) ('play', 1081, 29) ('player', 16, 551) ### ############################################################################################################ [2023-11-25 00:17:06,329][root][INFO] - Epoch: 1: Step: 4201/19530, loss[v]=0.421096, lr=0.000004, acc@1[1]=935.0/1024=0.9130859375, acc@1[2]=965.0/1024=0.9423828125 [2023-11-25 00:18:56,798][root][INFO] - Train batch 4300 [2023-11-25 00:18:56,799][root][INFO] - Avg. loss per last 100 batches: 0.394182 [2023-11-25 00:18:57,525][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.02 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.04 | max: 4.36 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: sprint world championships ( sprint canoe ) in 1990 and 2001, and again in 2010. ### ### [P_TEXT]: denys tourtchenkov denys tourtchenkov ( ; born october 7, 1978 in volgograd, soviet union ### ### ) is a russia n sprint canoer who competed in the early 2000s. he won a complete set of medals at ### ### the 2001 icf canoe sprint world championships in poznan with a gold ( k - 4 500 m ), a silver ( k - ### ### 4 200 ), and a bronze ( k - 4 1000 m ). ### ### =============================================== v_q ================================================ ### ### ('canoe', 0, 1) ('.', 1, 79) ('sprint', 2, 7) (',', 3, 7344) ('1990', 4, 1207) ('2001', 5, 13) ### ### ('2010', 6, 1203) ('championships', 7, 10) ('world', 8, 35) ('2000', 9, 38) ('the', 10, 4563) ### ### ('championship', 11, 36) ('1999', 12, 64) ('1991', 13, 413) ('2002', 14, 42) ('race', 15, 4091) ### ### ('1989', 16, 2771) ('pace', 17, 111) ('fish', 18, 754) ('，', 19, 260) ### ### =============================================== v_p ================================================ ### ### ('deny', 2606, 0) ('canoe', 0, 1) ('tour', 7095, 2) ('ic', 2611, 3) ('poznan', 11362, 4) ### ### ('soviet', 384, 5) ('##tch', 15012, 6) ('sprint', 2, 7) ('denying', 2778, 8) ('##enko', 23192, 9) ### ### ('championships', 7, 10) ('##go', 17834, 11) ('vol', 8848, 12) ('2001', 5, 13) ('russia', 782, 14) ### ### ('medals', 8589, 15) ('silver', 1440, 16) ('##f', 13330, 17) ('500', 6498, 18) ('1000', 3133, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('canoe', 0, 1) ('sprint', 2, 7) ('championships', 7, 10) ('2001', 5, 13) ('world', 8, 35) ### ### ('.', 1, 79) ('2000', 9, 38) ('championship', 11, 36) ('deny', 2606, 0) ('soviet', 384, 5) ### ### ('competed', 96, 23) ('2002', 14, 42) ('ic', 2611, 3) ('russia', 782, 14) ('boat', 22, 49) ### ### ('1990', 4, 1207) ('1999', 12, 64) ('denied', 774, 21) ('2010', 6, 1203) ('denying', 2778, 8) ### ############################################################################################################ [2023-11-25 00:18:58,028][root][INFO] - Epoch: 1: Step: 4301/19530, loss[v]=0.544464, lr=0.000004, acc@1[1]=918.0/1024=0.896484375, acc@1[2]=946.0/1024=0.923828125 [2023-11-25 00:20:48,873][root][INFO] - Train batch 4400 [2023-11-25 00:20:48,873][root][INFO] - Avg. loss per last 100 batches: 0.393480 [2023-11-25 00:20:49,594][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.04 | max: 4.49 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.7/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: a lifeboat, " eliza fernley ", was launched from southport in response to distress signal ### ### s from " mexico ". ### ### [P_TEXT]: southport the first support pile. the opening of the pier was celebrated on 2 august ### ### 1860. on the night of 9 december 1886, the worst lifeboat disaster in the history of the uk ### ### occurred off the shores of southport. a cargo ship called the " mexico " was on its way to south ### ### america when it found itself in difficulty. lifeboats from lytham, st. annes and southport set off ### ### to try to rescue those aboard the vessel. the crews battled against storm - force winds as they ### ### rowed towards the casualty. the entire crew from the st. anne's boat was lost ### ### =============================================== v_q ================================================ ### ### ('lifeboat', 0, 0) ('eliza', 1, 11261) ('##ley', 2, 13700) ('fern', 3, 21953) ('.', 4, 194) ### ### ('mexico', 5, 7) ('southport', 6, 1) (',', 7, 2491) ('mexican', 8, 34) ('signal', 9, 1638) ### ### ('sailing', 10, 874) ('liverpool', 11, 39) ('ship', 12, 22) ('sailed', 13, 2580) ### ### ('naval', 14, 1249) ('maritime', 15, 4348) ('the', 16, 1783) ('malta', 17, 71) ('launch', 18, 1353) ### ### ('rescued', 19, 76) ### ### =============================================== v_p ================================================ ### ### ('lifeboat', 0, 0) ('southport', 6, 1) ('anne', 2865, 2) ('pile', 2399, 3) ('##yt', 24329, 4) ### ### ('##ham', 6865, 5) ('worst', 2624, 6) ('mexico', 5, 7) ('pier', 3226, 8) ('rowe', 7803, 9) ### ### ('boat', 31, 10) ('support', 6299, 11) ('1860', 3768, 12) ('storm', 2300, 13) ('cargo', 2729, 14) ### ### ('disaster', 2455, 15) ('1886', 15946, 16) ('opening', 10930, 17) ('casualty', 24753, 18) ### ### ('aboard', 900, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('lifeboat', 0, 0) ('southport', 6, 1) ('mexico', 5, 7) ('mexican', 8, 34) ('ship', 12, 22) ### ### ('boat', 31, 10) ('liverpool', 11, 39) ('rescue', 28, 21) ('.', 4, 194) ('##herton', 24, 28) ### ### ('vessel', 26, 30) ('malta', 17, 71) ('rescued', 19, 76) ('bury', 38, 49) ('blackpool', 43, 45) ### ### ('lancashire', 21, 78) ('pile', 2399, 3) ('anne', 2865, 2) ('hamburg', 42, 57) ('uk', 270, 24) ### ############################################################################################################ [2023-11-25 00:20:50,097][root][INFO] - Epoch: 1: Step: 4401/19530, loss[v]=0.216435, lr=0.000005, acc@1[1]=973.0/1024=0.9501953125, acc@1[2]=997.0/1024=0.9736328125 [2023-11-25 00:22:40,518][root][INFO] - Train batch 4500 [2023-11-25 00:22:40,518][root][INFO] - Avg. loss per last 100 batches: 0.366283 [2023-11-25 00:22:41,264][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.00 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.04 | max: 4.48 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.3/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: she is one half of the musical duo wendy & lisa, formed with wendy melvoin in 1986. ### ### [P_TEXT]: wendy & amp ; lisa wendy & lisa ( briefly known as girl bros. ) are a music duo ### ### consisting of wendy melvoin and lisa coleman. they began working with prince in the early 1980s and ### ### were part of his band the revolution, before branching out on their own and releasing their debut ### ### album in 1987. in recent years they have turned their attention to writing music for film and ### ### television and have won an emmy award. they have released five full - length albums, the most ### ### recent being " white flags of winter chimneys ", released in december 2008. in 1980, lisa coleman ### ### replaced gayle chapman in prince's touring band on keyboards and ### ### =============================================== v_q ================================================ ### ### ('wendy', 0, 4) ('lisa', 1, 7) ('.', 2, 138) ('mel', 3, 12) ('##vo', 4, 11) ('musical', 5, 119) ### ### (',', 6, 3696) ('##in', 7, 17) ('&', 8, 16) ('duo', 9, 33) ('1986', 10, 131) ('music', 11, 24) ### ### ('the', 12, 3104) ('1985', 13, 188) ('1987', 14, 34) ('singer', 15, 5192) ('half', 16, 1377) ### ### ('kylie', 17, 166) ('madeira', 18, 1055) ('song', 19, 951) ### ### =============================================== v_p ================================================ ### ### ('chimneys', 26144, 0) ('flags', 13795, 1) ('amp', 18996, 2) ('revolution', 7406, 3) ### ### ('wendy', 0, 4) ('chapman', 11362, 5) ('prince', 7400, 6) ('lisa', 1, 7) ('coleman', 17370, 8) ### ### ('winter', 7565, 9) ('white', 2746, 10) ('##vo', 4, 11) ('mel', 3, 12) ('gay', 7621, 13) ### ### ('bros', 20510, 14) ('emmy', 24547, 15) ('&', 8, 16) ('##in', 7, 17) ('film', 32, 18) ### ### ('writing', 8621, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('wendy', 0, 4) ('lisa', 1, 7) ('mel', 3, 12) ('##vo', 4, 11) ('##in', 7, 17) ('&', 8, 16) ### ### ('music', 11, 24) ('duo', 9, 33) ('1987', 14, 34) ('film', 32, 18) ('.', 2, 138) ('band', 55, 22) ### ### ('musical', 5, 119) ('1980s', 134, 23) ('album', 94, 31) ('1986', 10, 131) ('white', 2746, 10) ### ### ('revolution', 7406, 3) ('girl', 920, 27) ('summer', 696, 32) ### ############################################################################################################ [2023-11-25 00:22:41,766][root][INFO] - Epoch: 1: Step: 4501/19530, loss[v]=0.194957, lr=0.000005, acc@1[1]=970.0/1024=0.947265625, acc@1[2]=996.5/1024=0.97314453125 [2023-11-25 00:24:32,629][root][INFO] - Train batch 4600 [2023-11-25 00:24:32,630][root][INFO] - Avg. loss per last 100 batches: 0.365419 [2023-11-25 00:24:33,371][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.07 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.04 | max: 4.62 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: a world map based on the " geography " of ptolemy was famous for its size and was ### ### displayed in the portico of one of the schools. ### ### [P_TEXT]: geography ( ptolemy ) to be 26 : 10 european maps, 4 african maps, and 12 asian maps. as ### ### early as the 1420s, these canonical maps were complemented by extra - ptolemaic regional maps ### ### depicting, e. g., scandinavia. the original treatise by marinus of tyre that formed the basis of ### ### ptolemy's " geography " has been completely lost. a world map based on ptolemy was displayed in ### ### augustodunum ( autun, france ) in late roman times. pappus, writing at alexandria in the 4th ### ### century, produced a commentary on ptolemy's " geography " and used it ### ### =============================================== v_q ================================================ ### ### ('ptolemy', 0, 3) ('geography', 1, 0) ('.', 2, 50) ('size', 3, 180) ('portico', 4, 7842) ### ### ('map', 5, 31) ('world', 6, 24) ('school', 7, 4355) (',', 8, 1804) ('maps', 9, 12) ### ### ('schools', 10, 10914) ('greece', 11, 283) ('egypt', 12, 44) ('apollo', 13, 864) ('the', 14, 5906) ### ### ('sizes', 15, 6882) ('##ης', 16, 373) ('diameter', 17, 1426) ('commanded', 18, 2162) ### ### ('polo', 19, 5118) ### ### =============================================== v_p ================================================ ### ### ('geography', 1, 0) ('alexandria', 132, 1) ('tyre', 26531, 2) ('ptolemy', 0, 3) ('##tu', 14162, 4) ### ### ('marin', 11524, 5) ('##dun', 29356, 6) ('##pp', 25364, 7) ('##ole', 22261, 8) ('au', 12129, 9) ### ### ('extra', 6958, 10) ('pt', 6923, 11) ('maps', 9, 12) ('regional', 2525, 13) ### ### ('canonical', 22548, 14) ('asian', 956, 15) ('##0s', 28382, 16) ('complement', 17103, 17) ### ### ('##us', 15906, 18) ('treatise', 8854, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ptolemy', 0, 3) ('geography', 1, 0) ('maps', 9, 12) ('map', 5, 31) ('world', 6, 24) ('.', 2, 50) ### ### ('alexandria', 132, 1) ('roman', 30, 26) ('egypt', 12, 44) ('size', 3, 180) ('region', 23, 43) ### ### ('asian', 956, 15) ('africa', 199, 35) ('france', 903, 21) ('greece', 11, 283) ### ### ('regional', 2525, 13) ('completely', 85, 46) ('european', 1041, 32) ('?', 40, 54) ### ### ('african', 2472, 20) ### ############################################################################################################ [2023-11-25 00:24:33,875][root][INFO] - Epoch: 1: Step: 4601/19530, loss[v]=0.290194, lr=0.000005, acc@1[1]=970.5/1024=0.94775390625, acc@1[2]=1001.5/1024=0.97802734375 [2023-11-25 00:26:24,721][root][INFO] - Train batch 4700 [2023-11-25 00:26:24,722][root][INFO] - Avg. loss per last 100 batches: 0.380623 [2023-11-25 00:26:25,452][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.05 | max: 3.78 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.04 | max: 4.57 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he was known for action - comedy and spaghetti western roles with his long - time film ### ### partner terence hill. ### ### [P_TEXT]: watch out, we're mad! watch out, we're mad! ( ) is a 1974 italian comedy film, directed ### ### by marcello fondato and starring the comedy team of terence hill and bud spencer. it is an ### ### international co - production between italy and spain. somewhere in spain, off - road race car ### ### driver ben ( spencer ) is anticipating the next rallycross race, because of the prize : a red puma ### ### dune buggy with a yellow top. however, he does not count on " the kid " ( hill ) entering the ### ### competition, who also has a talent for the ### ### =============================================== v_q ================================================ ### ### ('spaghetti', 0, 15755) ('terence', 1, 0) ('hill', 2, 4) ('.', 3, 152) ('film', 4, 22) ### ### ('comedy', 5, 34) ('action', 6, 5217) (',', 7, 6232) ('western', 8, 2727) ('terry', 9, 62) ### ### ('partner', 10, 4891) ('pasta', 11, 11975) ('the', 12, 1580) ('movie', 13, 64) ### ### ('partners', 14, 2154) ('pizza', 15, 3527) ('long', 16, 2140) ('picture', 17, 628) ### ### ('hills', 18, 106) ('actor', 19, 9962) ### ### =============================================== v_p ================================================ ### ### ('terence', 1, 0) ('spencer', 8871, 1) ('kid', 6796, 2) ('watch', 6044, 3) ('hill', 2, 4) ### ### ('##ato', 26763, 5) ('dune', 24811, 6) ('ben', 2790, 7) ('rally', 7223, 8) ('mad', 2241, 9) ### ### ('bud', 2951, 10) ('out', 4661, 11) ('we', 1805, 12) ('somewhere', 4902, 13) ('fond', 13979, 14) ### ### ('pu', 17472, 15) ('marcel', 11123, 16) ('1974', 7515, 17) ('yellow', 2516, 18) ('race', 1597, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('terence', 1, 0) ('hill', 2, 4) ('film', 4, 22) ('comedy', 5, 34) ('.', 3, 152) ('spain', 44, 27) ### ### ('terry', 9, 62) ('italy', 54, 29) ('top', 99, 26) ('road', 116, 31) ('italian', 57, 40) ### ### ('movie', 13, 64) ('ben', 2790, 7) ('we', 1805, 12) ('production', 161, 37) ('mad', 2241, 9) ### ### ('starring', 27, 55) ('bud', 2951, 10) ('competition', 386, 35) ('race', 1597, 19) ### ############################################################################################################ [2023-11-25 00:26:25,953][root][INFO] - Epoch: 1: Step: 4701/19530, loss[v]=0.641409, lr=0.000005, acc@1[1]=913.0/1024=0.8916015625, acc@1[2]=942.5/1024=0.92041015625 [2023-11-25 00:28:16,483][root][INFO] - Train batch 4800 [2023-11-25 00:28:16,485][root][INFO] - Avg. loss per last 100 batches: 0.371798 [2023-11-25 00:28:17,209][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.05 | max: 4.54 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: hitzkirch and the surrounding 10 municipalities failed, when five of the eleven voted ### ### against the merger. ### ### [P_TEXT]: hitzkirch hitzkirch is a municipality in the district of hochdorf in the canton of ### ### lucerne in switzerland. on 1 january 2009, gelfingen, hamikon, mosen, muswangen, retschwil and sulz ### ### joined hitzkirch. on 1 january 2021, altwis merged into hitzkirch. hitzkirchis is located in the ### ### seetal valley, at the foot of the lindenberg mountain. after the 2021 merger, hitzkirch has an ### ### area, ( as of the 2004 / 09 survey ), of. of this area, ### ### =============================================== v_q ================================================ ### ### ('##kir', 0, 0) ('.', 1, 72) ('merger', 2, 33) ('##ch', 3, 7) ('hit', 4, 14) (',', 5, 8174) ### ### ('##z', 6, 25) ('municipality', 7, 38) ('municipal', 8, 68) ('hitting', 9, 37) ('merged', 10, 49) ### ### ('municipalities', 11, 59) ('##ching', 12, 50) ('mergers', 13, 211) ('eleven', 14, 9665) ### ### ('voted', 15, 11939) ('the', 16, 5279) ('voter', 17, 1728) ('consolidation', 18, 717) ### ### ('combined', 19, 852) ### ### =============================================== v_p ================================================ ### ### ('##kir', 0, 0) ('##ange', 27221, 1) ('##rne', 14505, 2) ('linden', 22825, 3) ('##tal', 14482, 4) ### ### ('##sen', 14307, 5) ('##wil', 18254, 6) ('##ch', 3, 7) ('##berg', 1124, 8) ('survey', 1085, 9) ### ### ('##orf', 14062, 10) ('ham', 6634, 11) ('luce', 15204, 12) ('canton', 9218, 13) ('hit', 4, 14) ### ### ('##hd', 16596, 15) ('##sw', 22407, 16) ('##wi', 16917, 17) ('mountain', 638, 18) ### ### ('2021', 23938, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##kir', 0, 0) ('##ch', 3, 7) ('hit', 4, 14) ('merger', 2, 33) ('##z', 6, 25) ### ### ('municipality', 7, 38) ('hitting', 9, 37) ('.', 1, 72) ('merged', 10, 49) ('district', 20, 42) ### ### ('##ching', 12, 50) ('municipalities', 11, 59) ('municipal', 8, 68) ('##з', 31, 47) ### ### ('mountain', 638, 18) ('survey', 1085, 9) ('##berg', 1124, 8) ('gel', 647, 22) ('see', 691, 24) ### ### ('switzerland', 964, 21) ### ############################################################################################################ [2023-11-25 00:28:17,711][root][INFO] - Epoch: 1: Step: 4801/19530, loss[v]=0.237030, lr=0.000005, acc@1[1]=973.5/1024=0.95068359375, acc@1[2]=998.5/1024=0.97509765625 [2023-11-25 00:30:08,086][root][INFO] - Train batch 4900 [2023-11-25 00:30:08,087][root][INFO] - Avg. loss per last 100 batches: 0.354240 [2023-11-25 00:30:08,833][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.04 | max: 4.40 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the origin of batman depicts the events that cause a young bruce wayne to become batman. ### ### [P_TEXT]: batman kane, finger, and future dc writers accompanied batman with supporting characters, ### ### including his sidekick robin, allies alfred pennyworth and james gordon, and foes such as catwoman, ### ### the scarecrow, the penguin, and his archenemy, the joker. batman's origin story features him ### ### swearing vengeance against criminals after witnessing the murder of his parents thomas and martha ; ### ### he trains himself physically and intellectually, crafts a bat - inspired ### ### =============================================== v_q ================================================ ### ### ('batman', 0, 0) ('bruce', 1, 237) ('.', 2, 110) ('wayne', 3, 81) ('young', 4, 2899) ### ### (',', 5, 10622) ('origin', 6, 34) ('superman', 7, 54) ('events', 8, 15493) ('origins', 9, 67) ### ### ('teenage', 10, 8281) ('the', 11, 2455) ('youth', 12, 3598) ('robin', 13, 6) ('comics', 14, 1160) ### ### ('cause', 15, 3168) ('gordon', 16, 17) ('event', 17, 12737) ('depicts', 18, 1369) ### ### ('evolution', 19, 1144) ### ### =============================================== v_p ================================================ ### ### ('batman', 0, 0) ('penguin', 1428, 1) ('finger', 6695, 2) ('joker', 34, 3) ('crafts', 18711, 4) ### ### ('kane', 1982, 5) ('robin', 13, 6) ('penny', 12237, 7) ('scare', 10409, 8) ('cat', 1655, 9) ### ### ('##crow', 16921, 10) ('foe', 20595, 11) ('bat', 44, 12) ('martha', 5520, 13) ### ### ('##woman', 25479, 14) ('sidekick', 4937, 15) ('alfred', 30, 16) ('gordon', 16, 17) ### ### ('##worth', 23338, 18) ('vengeance', 13459, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('batman', 0, 0) ('robin', 13, 6) ('origin', 6, 34) ('gordon', 16, 17) ('joker', 34, 3) ### ### ('alfred', 30, 16) ('bat', 44, 12) ('wayne', 3, 81) ('.', 2, 110) ('bruce', 1, 237) ('dc', 33, 26) ### ### ('superman', 7, 54) ('penguin', 1428, 1) ('cat', 1655, 9) ('kane', 1982, 5) ('origins', 9, 67) ### ### ('thomas', 1909, 20) ('finger', 6695, 2) ('story', 121, 37) ('sidekick', 4937, 15) ### ############################################################################################################ [2023-11-25 00:30:09,334][root][INFO] - Epoch: 1: Step: 4901/19530, loss[v]=0.248426, lr=0.000005, acc@1[1]=970.5/1024=0.94775390625, acc@1[2]=1001.5/1024=0.97802734375 [2023-11-25 00:31:59,711][root][INFO] - Train batch 5000 [2023-11-25 00:31:59,711][root][INFO] - Avg. loss per last 100 batches: 0.364867 [2023-11-25 00:32:00,446][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.04 | max: 4.49 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: fritzsche along with speer and shirach were eventually communed by lutheran pastor henry ### ### f. gerecke and were administered the eucharist. ### ### [P_TEXT]: henry f. gerecke back to christ while they stood trial. fritz sauckel, ( the nazi labour ### ### leader ) eventually took communion during the trial, as did albert speer ( minister of armaments ### ### and war production ), hans fritzsche ( a nazi propagandist ) and baldur von schirach ( leader of ### ### the hitler youth ). following their convictions, joachim von ribbentrop ( foreign minister ), ### ### wilhelm keitel ( the wehrmacht field marshal ) all also took communion with gerecke. a fortnight ### ### after the executions of those convicted war criminals who had been hung at the ### ### =============================================== v_q ================================================ ### ### ('##sche', 0, 19) ('fritz', 1, 1) ('sp', 2, 21) ('.', 3, 71) ('commune', 4, 5939) ('shi', 5, 16134) ### ### ('henry', 6, 9) ('##rac', 7, 26213) ('##eer', 8, 28) ('eu', 9, 9420) ('lutheran', 10, 17129) ### ### ('##cha', 11, 11330) ('##re', 12, 17) ('ge', 13, 20) ('f', 14, 29) (',', 15, 9251) ### ### ('##h', 16, 19254) ('pastor', 17, 4001) ('##cke', 18, 22) ('##rist', 19, 26097) ### ### =============================================== v_p ================================================ ### ### ('##uck', 5280, 0) ('fritz', 1, 1) ('hung', 2687, 2) ('communion', 2306, 3) ('bald', 10365, 4) ### ### ('armament', 23071, 5) ('albert', 3650, 6) ('convicted', 8675, 7) ('marshal', 15848, 8) ### ### ('henry', 6, 9) ('production', 4007, 10) ('##hir', 5799, 11) ('youth', 4289, 12) ### ### ('trial', 7994, 13) ('##tro', 27416, 14) ('prop', 5117, 15) ('foreign', 5619, 16) ('##re', 12, 17) ### ### ('christ', 26, 18) ('##sche', 0, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('fritz', 1, 1) ('##sche', 0, 19) ('sp', 2, 21) ('henry', 6, 9) ('##re', 12, 17) ('##eer', 8, 28) ### ### ('ge', 13, 20) ('f', 14, 29) ('##cke', 18, 22) ('.', 3, 71) ('christ', 26, 18) ('minister', 39, 27) ### ### ('nazi', 62, 31) ('##eers', 21, 69) ('wilhelm', 325, 24) ('communion', 2306, 3) ('hung', 2687, 2) ### ### ('##uck', 5280, 0) ('albert', 3650, 6) ('hitler', 1435, 30) ### ############################################################################################################ [2023-11-25 00:32:00,946][root][INFO] - Epoch: 1: Step: 5001/19530, loss[v]=0.172345, lr=0.000005, acc@1[1]=978.5/1024=0.95556640625, acc@1[2]=1000.0/1024=0.9765625 [2023-11-25 00:33:51,113][root][INFO] - Train batch 5100 [2023-11-25 00:33:51,113][root][INFO] - Avg. loss per last 100 batches: 0.339275 [2023-11-25 00:33:51,848][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 4.02 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.04 | max: 4.56 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the new worcestershire royal hospital is now nearer the motorway, close to warndon ### ### villages and leads off charles hastings way. ### ### [P_TEXT]: worcester, england falls within the administrative area of worcester city council. ### ### worcester guildhall houses the local council and dates from 1721 ( see history ). notable suburbs ### ### include barbourne, blackpole, cherry orchard, claines, diglis, dines green, henwick, northwick, red ### ### hill, ronkswood, st peter the great ( also known as st peter's ), tolladine, warndon and warndon ### ### villages ( once the largest housing development in the country when the area was being constructed ### ### in the late 1980s / very early 1990s ). most of worcester is ### ### =============================================== v_q ================================================ ### ### ('hastings', 0, 10281) ('##don', 1, 17) ('hospital', 2, 1759) ('warn', 3, 6) ('.', 4, 144) ### ### ('worcestershire', 5, 203) ('charles', 6, 11085) ('royal', 7, 2763) ('motorway', 8, 16365) ### ### ('way', 9, 10912) (',', 10, 4812) ('warning', 11, 44) ('warned', 12, 45) ('worcester', 13, 0) ### ### ('village', 14, 39) ('leads', 15, 2671) ('nearest', 16, 2380) ('new', 17, 8735) ('road', 18, 9542) ### ### ('cambridge', 19, 2012) ### ### =============================================== v_p ================================================ ### ### ('worcester', 13, 0) ('##pole', 7677, 1) ('orchard', 7408, 2) ('toll', 10905, 3) ### ### ('cherry', 2670, 4) ('peter', 1250, 5) ('warn', 3, 6) ('##bourne', 13008, 7) ('##lis', 19755, 8) ### ### ('##adi', 23292, 9) ('red', 4031, 10) ('##aine', 18860, 11) ('bar', 10618, 12) ('dig', 11210, 13) ### ### ('hill', 154, 14) ('##wood', 3471, 15) ('guild', 12992, 16) ('##don', 1, 17) ('ron', 7022, 18) ### ### ('england', 46, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('warn', 3, 6) ('##don', 1, 17) ('worcester', 13, 0) ('village', 14, 39) ('warning', 11, 44) ### ### ('villages', 27, 21) ('warned', 12, 45) ('england', 46, 19) ('.', 4, 144) ### ### ('worcestershire', 5, 203) ('hill', 154, 14) ('peter', 1250, 5) ('hospital', 2, 1759) ### ### ('green', 651, 23) ('cherry', 2670, 4) ('administrative', 506, 29) ('council', 1136, 27) ### ### ('britain', 62, 47) ('red', 4031, 10) ('##pole', 7677, 1) ### ############################################################################################################ [2023-11-25 00:33:52,351][root][INFO] - Epoch: 1: Step: 5101/19530, loss[v]=0.507682, lr=0.000005, acc@1[1]=934.0/1024=0.912109375, acc@1[2]=954.5/1024=0.93212890625 [2023-11-25 00:35:42,885][root][INFO] - Train batch 5200 [2023-11-25 00:35:42,886][root][INFO] - Avg. loss per last 100 batches: 0.342666 [2023-11-25 00:35:43,612][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.04 | max: 4.50 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: brown fumbled the ball as he was being tackled, and two officials ( back judge ray ### ### douglas and field judge dick dolack ) initially ruled the ball recovered for the vikings by jeff ### ### siemon ### ### [P_TEXT]: jeff siemon taken by the vikings with a pick acquired from the new england patriots in ### ### the trade for joe kapp, the vikings'starting quarterback in super bowl iv. siemon played for the ### ### vikings for 11 seasons, retiring after the strike - shortened 1982 campaign. during that time, he ### ### was the starting middle linebacker in four nfc championship games over the course of five years ( ### ### 1973 – 1977 ), winning three : 1973, 1974, and 1976, losing one : 1977, and three super bowls ( ### ### viii, ix, xi ), all losses. he was also a vital part of the vikings'1975 ### ### =============================================== v_q ================================================ ### ### ('##emon', 0, 0) ('vikings', 1, 2) ('.', 2, 127) ('douglas', 3, 12546) ('brown', 4, 1107) ### ### ('si', 5, 9) ('judge', 6, 15068) ('ray', 7, 6889) ('jeff', 8, 4) ('dick', 9, 3999) ### ### ('##la', 10, 19382) ('##ck', 11, 14865) ('football', 12, 2483) ('viking', 13, 31) ### ### ('officials', 14, 4306) ('field', 15, 9770) (',', 16, 4757) ('do', 17, 6873) ('norse', 18, 201) ### ### ('ball', 19, 3136) ### ### =============================================== v_p ================================================ ### ### ('##emon', 0, 0) ('##pp', 25102, 1) ('vikings', 1, 2) ('nfc', 7344, 3) ('jeff', 8, 4) ### ### ('iv', 2824, 5) ('linebacker', 293, 6) ('super', 9569, 7) ('patriots', 1381, 8) ('si', 5, 9) ### ### ('viii', 4809, 10) ('strike', 6825, 11) ('xi', 11335, 12) ('bowls', 12476, 13) ('ix', 10901, 14) ### ### ('bowl', 1312, 15) ('quarterback', 38, 16) ('losses', 6610, 17) ('campaign', 3772, 18) ### ### ('england', 1280, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##emon', 0, 0) ('vikings', 1, 2) ('jeff', 8, 4) ('si', 5, 9) ('viking', 13, 31) ### ### ('quarterback', 38, 16) ('.', 2, 127) ('linebacker', 293, 6) ('championship', 108, 24) ### ### ('joe', 207, 20) ('patriots', 1381, 8) ('pick', 280, 30) ('brown', 4, 1107) ('bowl', 1312, 15) ### ### ('norse', 18, 201) ('iv', 2824, 5) ('england', 1280, 19) ('starting', 188, 40) ('game', 26, 86) ### ### ('nfc', 7344, 3) ### ############################################################################################################ [2023-11-25 00:35:44,113][root][INFO] - Epoch: 1: Step: 5201/19530, loss[v]=0.189568, lr=0.000005, acc@1[1]=973.5/1024=0.95068359375, acc@1[2]=1008.5/1024=0.98486328125 [2023-11-25 00:37:34,341][root][INFO] - Train batch 5300 [2023-11-25 00:37:34,341][root][INFO] - Avg. loss per last 100 batches: 0.346778 [2023-11-25 00:37:35,086][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.05 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.04 | max: 4.73 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: alma galarza flores is a puerto rican vocalist best known for her work with danny rivera ### ### in albums " controversia ", " ofrenda " and " mi cancion es paz ". ### ### [P_TEXT]: ofrenda ( danny rivera album ) ofrenda ( " offering " ) is a puerto rican christmas music ### ### and puerto rican folk music album by danny rivera. this album was danny rivera's second christmas ### ### album. as special guests he reunited legendary local folk group trio vegabajeno ( fernandito ### ### alvarez, benito de jesus, pepito maduro and jorge hernandez ). jorge hernandez was also a member of ### ### the group alpha iv which had been recording with danny for several years. from his previous album, ### ### controversia he recruited singer alma galarza of vicente carattini y los cantores de san juan with ### ### the condition she quit " los cantores " to join him on lead vocals full - time in the studio and on ### ### =============================================== v_q ================================================ ### ### ('##rza', 0, 7) ('danny', 1, 8) ('flores', 2, 19481) ('rivera', 3, 9) ('.', 4, 84) ('gala', 5, 20) ### ### ('rican', 6, 46) ('paz', 7, 23255) ('alma', 8, 17) ('##ren', 9, 3) ('##vers', 10, 28) ### ### ('##da', 11, 1) ('##tro', 12, 26) ('puerto', 13, 21) ('album', 14, 35) ('singer', 15, 59) ### ### ('vocalist', 16, 1774) (',', 17, 4362) ('mi', 18, 22451) ('con', 19, 42) ### ### =============================================== v_p ================================================ ### ### ('offering', 4210, 0) ('##da', 11, 1) ('alpha', 8374, 2) ('##ren', 9, 3) ('cantor', 22151, 4) ### ### ('##uro', 21211, 5) ('iv', 5917, 6) ('##rza', 0, 7) ('danny', 1, 8) ('rivera', 3, 9) ### ### ('cara', 5046, 10) ('##ito', 14237, 11) ('folk', 4279, 12) ('hernandez', 4243, 13) ### ### ('alvarez', 959, 14) ('vicente', 5179, 15) ('jesus', 4630, 16) ('alma', 8, 17) ('jorge', 538, 18) ### ### ('vega', 9964, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##rza', 0, 7) ('danny', 1, 8) ('rivera', 3, 9) ('##da', 11, 1) ('##ren', 9, 3) ('gala', 5, 20) ### ### ('alma', 8, 17) ('##vers', 10, 28) ('##tro', 12, 26) ('puerto', 13, 21) ('album', 14, 35) ### ### ('rican', 6, 46) ('con', 19, 42) ('music', 22, 41) ('singer', 15, 59) ('.', 4, 84) ### ### ('vocals', 30, 44) ('##ia', 26, 48) ('jorge', 538, 18) ('offering', 4210, 0) ### ############################################################################################################ [2023-11-25 00:37:35,586][root][INFO] - Epoch: 1: Step: 5301/19530, loss[v]=0.222010, lr=0.000005, acc@1[1]=968.5/1024=0.94580078125, acc@1[2]=1003.0/1024=0.9794921875 [2023-11-25 00:39:25,822][root][INFO] - Train batch 5400 [2023-11-25 00:39:25,823][root][INFO] - Avg. loss per last 100 batches: 0.337227 [2023-11-25 00:39:26,558][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.05 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.04 | max: 4.51 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 1811, jose gervasio artigas, who became uruguay's national hero, launched a successful ### ### revolt against the spanish authorities, defeating them on 18 may at the battle of las piedras. ### ### [P_TEXT]: instructions of the year xiii the instructions of the year xiii ( ) were the mandate ### ### brought by the representatives from the oriental province to the assembly of the year xiii of the ### ### united provinces of the river plate. in 1813, a meeting called by the second triumvirate was meant ### ### to define the type of government for the new nation. the people of the oriental province ( what is ### ### now uruguay ) drafted a federalist document, opposed to the centralism of the second triumvirate. ### ### these instructions were the origin of an unending conflict between the oriental leader, jose ### ### gervasio artigas, and the authorities that be in buenos aires. some scholars see the influence ### ### =============================================== v_q ================================================ ### ### ('.', 0, 314) ('##dra', 1, 26738) ('uruguay', 2, 5) ('pie', 3, 4609) ('##sio', 4, 7) ('art', 5, 16) ### ### ('spanish', 6, 1857) ('las', 7, 23477) ('##iga', 8, 6) ('battle', 9, 2960) ('revolt', 10, 11059) ### ### (',', 11, 9085) ('ge', 12, 13) ('spain', 13, 645) ('rebellion', 14, 4150) ('jose', 15, 10) ### ### ('napoleon', 16, 123) ('authorities', 17, 27) ('uruguayan', 18, 64) ('argentina', 19, 76) ### ### =============================================== v_p ================================================ ### ### ('oriental', 2110, 0) ('plate', 1406, 1) ('xiii', 11465, 2) ('tri', 3395, 3) ('##nding', 19301, 4) ### ### ('uruguay', 2, 5) ('##iga', 8, 6) ('##sio', 4, 7) ('aires', 296, 8) ('nation', 1022, 9) ### ### ('jose', 15, 10) ('##vira', 18886, 11) ('provinces', 13894, 12) ('ge', 12, 13) ### ### ('mandate', 5980, 14) ('instructions', 11827, 15) ('art', 5, 16) ('river', 2520, 17) ### ### ('year', 18527, 18) ('representatives', 830, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('uruguay', 2, 5) ('##sio', 4, 7) ('art', 5, 16) ('##iga', 8, 6) ('jose', 15, 10) ('ge', 12, 13) ### ### ('authorities', 17, 27) ('.', 0, 314) ('aires', 296, 8) ('plate', 1406, 1) ('1812', 34, 43) ### ### ('##rva', 27, 49) ('uruguayan', 18, 64) ('oriental', 2110, 0) ('meant', 113, 29) ### ### ('nation', 1022, 9) ('government', 48, 44) ('argentina', 19, 76) ('federal', 495, 20) ### ### ('1813', 203, 30) ### ############################################################################################################ [2023-11-25 00:39:27,064][root][INFO] - Epoch: 1: Step: 5401/19530, loss[v]=0.541290, lr=0.000006, acc@1[1]=921.5/1024=0.89990234375, acc@1[2]=949.5/1024=0.92724609375 [2023-11-25 00:41:17,629][root][INFO] - Train batch 5500 [2023-11-25 00:41:17,630][root][INFO] - Avg. loss per last 100 batches: 0.314700 [2023-11-25 00:41:18,354][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.05 | max: 4.05 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.05 | max: 4.88 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: 65 squadron at raf duxford, 72 squadron at raf north weald, 85 / 145 at raf west malling ### ### with gloster meteor nf. 11s, and 257 squadron and 263 squadron at raf wattisham. ### ### [P_TEXT]: michael donaldson ( raf officer ) commander at raf wattisham in 1987 and following ### ### promotion to air commodore on 1 april 1992, he became senior air staff officer at no. 11 group. he ### ### simultaneously held the appointment of commandant royal observer corps ( roc ) at a time when the ### ### majority of the roc had been stood down, leaving a small number of nuclear reporting cell observers ### ### serving at various armed forces hqs all over the uk. promoted to air vice marshal, he served as ### ### commandant of the raf staff college, bracknell from 1993 until leaving the ### ### =============================================== v_q ================================================ ### ### ('.', 0, 68) ('mall', 1, 6737) ('##isha', 2, 10) ('##ald', 3, 8468) ('squadron', 4, 696) ### ### ('raf', 5, 9) ('##ford', 6, 12540) ('du', 7, 18893) ('##x', 8, 17280) ('85', 9, 6133) ### ### ('72', 10, 7013) ('meteor', 11, 16797) ('we', 12, 410) ('65', 13, 8977) ('watt', 14, 3) ### ### (',', 15, 6992) ('west', 16, 7519) ('##los', 17, 26925) ('##ing', 18, 8969) ('##m', 19, 32) ### ### =============================================== v_p ================================================ ### ### ('marshal', 8523, 0) ('donaldson', 20272, 1) ('commandant', 10086, 2) ('watt', 14, 3) ### ### ('roc', 26600, 4) ('michael', 1656, 5) ('bra', 12917, 6) ('nuclear', 675, 7) ('##nell', 20955, 8) ### ### ('raf', 5, 9) ('##isha', 2, 10) ('observer', 10603, 11) ('commodore', 13307, 12) ### ### ('staff', 5475, 13) ('college', 8135, 14) ('cell', 9721, 15) ('air', 350, 16) ### ### ('simultaneously', 18577, 17) ('vice', 18950, 18) ('hq', 22346, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##isha', 2, 10) ('raf', 5, 9) ('watt', 14, 3) ('11', 20, 21) ('##m', 19, 32) ('.', 0, 68) ### ### ('nuclear', 675, 7) ('air', 350, 16) ('michael', 1656, 5) ('squadron', 4, 696) ('we', 12, 410) ### ### ('##ם', 24, 108) ('forces', 351, 39) ('uk', 693, 33) ('marshal', 8523, 0) ('group', 1654, 26) ### ### ('1987', 2747, 23) ('commandant', 10086, 2) ('staff', 5475, 13) ('1993', 1894, 31) ### ############################################################################################################ [2023-11-25 00:41:18,855][root][INFO] - Epoch: 1: Step: 5501/19530, loss[v]=0.448148, lr=0.000006, acc@1[1]=938.0/1024=0.916015625, acc@1[2]=963.5/1024=0.94091796875 [2023-11-25 00:43:09,273][root][INFO] - Train batch 5600 [2023-11-25 00:43:09,273][root][INFO] - Avg. loss per last 100 batches: 0.348318 [2023-11-25 00:43:09,990][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.05 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.04 | max: 4.45 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he appointed fellow officer maximian as augustus, co - emperor, in 286. ### ### [P_TEXT]: disembowelment king of spain. the assassin was interrogated and condemned to death almost ### ### immediately. on 14 july, after suffering various tortures during each of the five days since the ### ### assassination, gerard was disemboweled and dismembered while still alive, after which his heart was ### ### torn out and then he was beheaded by his dutch executioners. christian tradition states that ### ### erasmus of formiae, also known as saint elmo, was finally executed by disembowelment in about a. d. ### ### 303, after he had suffered extreme forms of torture during the persecutions of emperor diocletian ### ### and maximian. ### ### =============================================== v_q ================================================ ### ### ('maxim', 0, 1) ('augustus', 1, 22191) ('.', 2, 86) ('##ian', 3, 25) ('officer', 4, 6475) ### ### ('emperor', 5, 29) ('roman', 6, 1285) ('officers', 7, 5271) (',', 8, 10863) ('romans', 9, 10613) ### ### ('napoleon', 10, 97) ('co', 11, 16619) ('##ians', 12, 93) ('286', 13, 27665) ('imperial', 14, 83) ### ### ('command', 15, 10725) ('romania', 16, 1703) ('hannah', 17, 5653) ('appointed', 18, 2039) ### ### ('escort', 19, 5086) ### ### =============================================== v_p ================================================ ### ### ('elm', 14681, 0) ('maxim', 0, 1) ('persecution', 64, 2) ('##mbo', 20635, 3) ('##iae', 23408, 4) ### ### ('erasmus', 27384, 5) ('##ent', 15898, 6) ('gerard', 10308, 7) ('assassin', 9792, 8) ### ### ('torn', 2274, 9) ('torture', 8863, 10) ('assassination', 7019, 11) ('di', 10159, 12) ### ### ('execution', 1138, 13) ('##let', 25247, 14) ('christian', 472, 15) ('form', 6150, 16) ### ### ('##oc', 19687, 17) ('dutch', 11271, 18) ('tradition', 1243, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('maxim', 0, 1) ('##ian', 3, 25) ('emperor', 5, 29) ('persecution', 64, 2) ('.', 2, 86) ### ### ('christian', 472, 15) ('king', 189, 26) ('execution', 1138, 13) ('torn', 2274, 9) ### ### ('tradition', 1243, 19) ('##ians', 12, 93) ('napoleon', 10, 97) ('finally', 336, 36) ### ### ('imperial', 14, 83) ('executed', 1521, 24) ('spain', 2298, 21) ('roman', 6, 1285) ### ### ('assassination', 7019, 11) ('form', 6150, 16) ('300', 23, 78) ### ############################################################################################################ [2023-11-25 00:43:10,493][root][INFO] - Epoch: 1: Step: 5601/19530, loss[v]=0.428386, lr=0.000006, acc@1[1]=935.0/1024=0.9130859375, acc@1[2]=966.0/1024=0.943359375 [2023-11-25 00:45:01,268][root][INFO] - Train batch 5700 [2023-11-25 00:45:01,268][root][INFO] - Avg. loss per last 100 batches: 0.334904 [2023-11-25 00:45:02,010][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.05 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.04 | max: 4.51 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it rises in northern wicomico county, close to the delaware state line, and flows ### ### generally southwest, through salisbury, its head of navigation is monie bay on the eastern edge of ### ### the chesapeake bay between mt. ### ### [P_TEXT]: quantico, maryland quantico is an unincorporated community and census - designated place ### ### in wicomico county, maryland, united states, along the former stage route from vienna to the port ### ### of whitehaven on the wicomico river. it consists of approximately 35 homes on the quantico creek. ### ### the community lies four miles ( 6 km ) from hebron along maryland route 347. there are two ### ### churches, a general store, and a post office in the community. it is part of the salisbury, ### ### maryland - delaware metropolitan statistical area. as of the ### ### =============================================== v_q ================================================ ### ### ('delaware', 0, 12) ('salisbury', 1, 1) ('.', 2, 78) ('mon', 3, 15617) ('chesapeake', 4, 14103) ### ### ('##com', 5, 8) ('wi', 6, 7) ('##ico', 7, 5) ('##ie', 8, 17059) ('bay', 9, 6236) ('county', 10, 18) ### ### ('mt', 11, 15124) (',', 12, 6574) ('pennsylvania', 13, 338) ('virginia', 14, 1351) ### ### ('flow', 15, 7610) ('navigation', 16, 19523) ('baltimore', 17, 442) ('maryland', 18, 6) ### ### ('flowed', 19, 1596) ### ### =============================================== v_p ================================================ ### ### ('##haven', 11764, 0) ('salisbury', 1, 1) ('quan', 27890, 2) ('##tic', 7422, 3) ('vienna', 4555, 4) ### ### ('##ico', 7, 5) ('maryland', 18, 6) ('wi', 6, 7) ('##com', 5, 8) ('store', 4218, 9) ### ### ('stage', 4488, 10) ('##bron', 29492, 11) ('delaware', 0, 12) ('river', 22, 13) ('white', 303, 14) ### ### ('churches', 18753, 15) ('statistical', 11341, 16) ('designated', 421, 17) ('county', 10, 18) ### ### ('34', 16514, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('salisbury', 1, 1) ('delaware', 0, 12) ('##ico', 7, 5) ('##com', 5, 8) ('wi', 6, 7) ### ### ('county', 10, 18) ('maryland', 18, 6) ('river', 22, 13) ('.', 2, 78) ('white', 303, 14) ### ### ('port', 242, 20) ('designated', 421, 17) ('vienna', 4555, 4) ('##tic', 7422, 3) ### ### ('pennsylvania', 13, 338) ('store', 4218, 9) ('route', 818, 25) ('stage', 4488, 10) ### ### ('unincorporated', 510, 32) ('baltimore', 17, 442) ### ############################################################################################################ [2023-11-25 00:45:02,512][root][INFO] - Epoch: 1: Step: 5701/19530, loss[v]=0.460562, lr=0.000006, acc@1[1]=941.5/1024=0.91943359375, acc@1[2]=953.5/1024=0.93115234375 [2023-11-25 00:46:53,356][root][INFO] - Train batch 5800 [2023-11-25 00:46:53,357][root][INFO] - Avg. loss per last 100 batches: 0.341781 [2023-11-25 00:46:54,087][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.05 | max: 3.82 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.04 | max: 4.70 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it is classified as an airport of entry by nav canada and is staffed by the canada border ### ### services agency ( cbsa ). ### ### [P_TEXT]: coronach / scobey border station airport customs officials require two hours'advance ### ### notice prior to landing, and landings are allowed only during the border crossing's normal hours of ### ### operation. the airport is one of six airports that straddle the canada – us border. the others are ### ### avey field state airport, whetstone international airport, coutts / ross international airport, ### ### international peace garden airport, and piney pinecreek border airport. the airport is classified ### ### as an airport of entry by nav canada and is staffed by the canada border services agency. cbsa ### ### officers ### ### =============================================== v_q ================================================ ### ### ('canada', 0, 12) ('.', 1, 142) ('border', 2, 19) ('cbs', 3, 8) ('##v', 4, 31) ('airport', 5, 14) ### ### ('agency', 6, 32) ('entry', 7, 24) ('services', 8, 34) ('na', 9, 28) ('canadian', 10, 55) ### ### ('airfield', 11, 61) (',', 12, 5627) ('agencies', 13, 48) ('staffed', 14, 33) ('service', 15, 69) ### ### ('air', 16, 753) ('plane', 17, 509) ('toronto', 18, 358) ('aircraft', 19, 91) ### ### =============================================== v_p ================================================ ### ### ('corona', 8553, 0) ('##obe', 25690, 1) ('ave', 18058, 2) ('##stone', 1594, 3) ('##cre', 28016, 4) ### ### ('##ut', 15973, 5) ('##ch', 18221, 6) ('ross', 3191, 7) ('cbs', 3, 8) ('pine', 4946, 9) ### ### ('##het', 26053, 10) ('garden', 2376, 11) ('canada', 0, 12) ('advance', 1393, 13) ### ### ('airport', 5, 14) ('field', 2143, 15) ('hours', 10009, 16) ('##dle', 19786, 17) ### ### ('##ek', 25838, 18) ('border', 2, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('canada', 0, 12) ('cbs', 3, 8) ('airport', 5, 14) ('border', 2, 19) ('entry', 7, 24) ### ### ('##v', 4, 31) ('agency', 6, 32) ('na', 9, 28) ('services', 8, 34) ('staffed', 14, 33) ### ### ('canadian', 10, 55) ('agencies', 13, 48) ('.', 1, 142) ('customs', 199, 26) ('airfield', 11, 61) ### ### ('##stone', 1594, 3) ('advance', 1393, 13) ('station', 794, 21) ('garden', 2376, 11) ### ### ('field', 2143, 15) ### ############################################################################################################ [2023-11-25 00:46:54,588][root][INFO] - Epoch: 1: Step: 5801/19530, loss[v]=0.407067, lr=0.000006, acc@1[1]=942.5/1024=0.92041015625, acc@1[2]=954.5/1024=0.93212890625 [2023-11-25 00:48:45,194][root][INFO] - Train batch 5900 [2023-11-25 00:48:45,195][root][INFO] - Avg. loss per last 100 batches: 0.325221 [2023-11-25 00:48:45,927][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.06 | max: 4.10 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.05 | max: 4.94 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the tubules are lined with a layer of cells ( germ cells ) that develop from puberty ### ### through old age into sperm cells ( also known as spermatozoa or male gametes ). ### ### [P_TEXT]: sperm, uterus and uterine tubes. during fertilization, the sperm provides three essential ### ### parts to the oocyte : ( 1 ) a signalling or activating factor, which causes the metabolically ### ### dormant oocyte to activate ; ( 2 ) the haploid paternal genome ; ( 3 ) the centriole, which is ### ### responsible for forming the centrosome and microtubule system. the spermatozoa of animal s are ### ### produced through spermatogenesis inside the male gonad s ( testicle s ) via meiotic division. the ### ### initial spermatozoon process takes ### ### =============================================== v_q ================================================ ### ### ('sperm', 0, 0) ('tub', 1, 265) ('game', 2, 4545) ('pub', 3, 12277) ('.', 4, 158) ### ### ('##tes', 5, 10281) ('##rm', 6, 14095) ('ge', 7, 18946) ('##ules', 8, 13036) ('male', 9, 28) ### ### ('age', 10, 5157) ('layer', 11, 3025) ('lined', 12, 6070) ('cells', 13, 5246) ('##zo', 14, 21) ### ### ('males', 15, 66) ('cell', 16, 8710) ('##ule', 17, 24) ('##ato', 18, 31) ('layers', 19, 3855) ### ### =============================================== v_p ================================================ ### ### ('sperm', 0, 0) ('##nad', 27498, 1) ('##yte', 27233, 2) ('genome', 258, 3) ('##rio', 23297, 4) ### ### ('metabolic', 7273, 5) ('##ivating', 5378, 6) ('animal', 1438, 7) ('mei', 5500, 8) ### ### ('##zoo', 24498, 9) ('paternal', 6803, 10) ('dormant', 5268, 11) ('factor', 8668, 12) ### ### ('##genesis', 2801, 13) ('##pl', 20019, 14) ('ut', 2101, 15) ('cent', 23518, 16) ### ### ('centro', 22094, 17) ('##some', 25079, 18) ('##erus', 4831, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('sperm', 0, 0) ('male', 9, 28) ('##zo', 14, 21) ('##ule', 17, 24) ('##ato', 18, 31) ### ### ('males', 15, 66) ('genome', 258, 3) ('.', 4, 158) ('tub', 1, 265) ('animal', 1438, 7) ### ### ('##a', 30, 63) ('ut', 2101, 15) ('##genesis', 2801, 13) ('##ivating', 5378, 6) ### ### ('division', 1078, 38) ('system', 444, 49) ('mei', 5500, 8) ('dormant', 5268, 11) ### ### ('##erus', 4831, 19) ('metabolic', 7273, 5) ### ############################################################################################################ [2023-11-25 00:48:46,432][root][INFO] - Epoch: 1: Step: 5901/19530, loss[v]=0.235818, lr=0.000006, acc@1[1]=982.0/1024=0.958984375, acc@1[2]=1010.5/1024=0.98681640625 [2023-11-25 00:50:37,585][root][INFO] - Train batch 6000 [2023-11-25 00:50:37,586][root][INFO] - Avg. loss per last 100 batches: 0.349760 [2023-11-25 00:50:38,311][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.05 | max: 3.80 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.05 | max: 4.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: besides teresa malone and carl reddix, who had been indicted by the u. s. attorney for ### ### the southern district of mississippi, the civil case defendants included previously uncharged ### ### michael reddix ; utah's management & training corporation ; florida's geo group, inc. ; cornell ### ### companies, inc. ; wexford health sources, inc. ; drug testing corporation ; global tel * link ### ### corporation and sentinel offender services, l. l. c. ### ### [P_TEXT]: sentinel offender services internal audit by los angeles county found that one in four of ### ### the sentinel - made ankle monitors used to monitor serious criminals were faulty. sentinel ### ### attributed many of these problems to errors by county deputies. in february 2017, sentinel offender ### ### services was included in a lawsuit brought by mississippi attorney general jim hood along with ### ### global tel link, wexford health sources, geo group and many others for their roles in alleged ### ### violations of mississippi ’ s public ethics, racketeering and antitrust laws. ### ### =============================================== v_q ================================================ ### ### ('cornell', 0, 3235) ('teresa', 1, 11230) ('carl', 2, 10162) ('malone', 3, 11685) ('.', 4, 217) ### ### ('utah', 5, 6176) ('sources', 6, 11) ('##di', 7, 25050) ('testing', 8, 2137) ('offender', 9, 2) ### ### ('corporation', 10, 1716) ('mississippi', 11, 7) ('attorney', 12, 23) ('training', 13, 367) ### ### ('link', 14, 6) ('sentinel', 15, 0) ('geo', 16, 1) ('source', 17, 32) ('wexford', 18, 3) ### ### ('global', 19, 10) ### ### =============================================== v_p ================================================ ### ### ('sentinel', 15, 0) ('geo', 16, 1) ('offender', 9, 2) ('wexford', 18, 3) ('services', 40, 4) ### ### ('jim', 3502, 5) ('link', 14, 6) ('mississippi', 11, 7) ('deputies', 19726, 8) ('hood', 9730, 9) ### ### ('global', 19, 10) ('sources', 6, 11) ('##trust', 17796, 12) ('ankle', 11592, 13) ### ### ('county', 2721, 14) ('tel', 27, 15) ('angeles', 6879, 16) ('ethics', 11476, 17) ### ### ('##ering', 5353, 18) ('health', 24, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('sentinel', 15, 0) ('offender', 9, 2) ('geo', 16, 1) ('sources', 6, 11) ('mississippi', 11, 7) ### ### ('wexford', 18, 3) ('link', 14, 6) ('global', 19, 10) ('attorney', 12, 23) ('health', 24, 19) ### ### ('tel', 27, 15) ('source', 17, 32) ('services', 40, 4) ('offenders', 37, 27) ('group', 31, 42) ### ### ('lawsuit', 90, 38) ('.', 4, 217) ('links', 82, 43) ('service', 217, 29) ('florida', 28, 99) ### ############################################################################################################ [2023-11-25 00:50:38,812][root][INFO] - Epoch: 1: Step: 6001/19530, loss[v]=0.492516, lr=0.000006, acc@1[1]=927.5/1024=0.90576171875, acc@1[2]=956.5/1024=0.93408203125 [2023-11-25 00:52:29,384][root][INFO] - Train batch 6100 [2023-11-25 00:52:29,385][root][INFO] - Avg. loss per last 100 batches: 0.356942 [2023-11-25 00:52:30,115][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.05 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.04 | max: 4.64 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the popularity of westerns continued to grow in the 1940s, with the release of films such ### ### as " the ox - bow incident " ( 1943 ), " my darling clementine " ( 1946 ), " fort apache " ### ### [P_TEXT]: fort apache ( film ) fort apache is a 1948 american western film directed by john ford ### ### and starring john wayne and henry fonda. the film was the first of the director's " cavalry trilogy ### ### " and was followed by " she wore a yellow ribbon " ( 1949 ) and " rio grande " ( 1950 ), both also ### ### starring wayne. the screenplay was inspired by james warner bellah's short story " massacre " ( ### ### 1947 ). the historical sources for " massacre " have been attributed both to george armstrong ### ### custer and the battle of little bighorn and to ### ### =============================================== v_q ================================================ ### ### ('apache', 0, 0) ('.', 1, 151) ('ox', 2, 19067) ('clement', 3, 8135) ('##ine', 4, 13217) ### ### ('film', 5, 20) ('incident', 6, 1402) ('bow', 7, 5043) ('western', 8, 30) ('darling', 9, 6760) ### ### ('fort', 10, 3) ('1943', 11, 569) ('1940s', 12, 1457) ('1946', 13, 40) ('popular', 14, 2766) ### ### (',', 15, 7826) ('arizona', 16, 123) ('films', 17, 47) ('popularity', 18, 1321) ('my', 19, 3240) ### ### =============================================== v_p ================================================ ### ### ('apache', 0, 0) ('cavalry', 1428, 1) ('bella', 6525, 2) ('fort', 10, 3) ('grande', 5099, 4) ### ### ('ribbon', 7927, 5) ('custer', 12081, 6) ('massacre', 3161, 7) ('fond', 18851, 8) ### ### ('wayne', 7223, 9) ('armstrong', 22201, 10) ('##horn', 4633, 11) ('rio', 13476, 12) ### ### ('ford', 2036, 13) ('battle', 6113, 14) ('wore', 10836, 15) ('1948', 49, 16) ('yellow', 1441, 17) ### ### ('warner', 4778, 18) ('little', 5503, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('apache', 0, 0) ('fort', 10, 3) ('film', 5, 20) ('western', 8, 30) ('1946', 13, 40) ### ### ('1948', 49, 16) ('.', 1, 151) ('1947', 30, 31) ('films', 17, 47) ('cavalry', 1428, 1) ### ### ('1949', 207, 21) ('movie', 20, 48) ('wearing', 246, 27) ('massacre', 3161, 7) ('ford', 2036, 13) ### ### ('yellow', 1441, 17) ('grande', 5099, 4) ('bella', 6525, 2) ('1950', 78, 38) ('arizona', 16, 123) ### ############################################################################################################ [2023-11-25 00:52:30,616][root][INFO] - Epoch: 1: Step: 6101/19530, loss[v]=0.198389, lr=0.000006, acc@1[1]=972.0/1024=0.94921875, acc@1[2]=1000.0/1024=0.9765625 [2023-11-25 00:54:21,471][root][INFO] - Train batch 6200 [2023-11-25 00:54:21,472][root][INFO] - Avg. loss per last 100 batches: 0.300543 [2023-11-25 00:54:22,210][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.05 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.04 | max: 4.87 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: suining is located in the center of the sichuan basin and on the central reaches of the ### ### fu river, bordering chongqing, guang'an and nanchong to the east, neijiang and ziyang to the south, ### ### the provincial capital of chengdu to the west, and deyang and mianyang to the north. ### ### [P_TEXT]: g42 shanghai – chengdu expressway, where a 100 km / h ( 62 mph ) speed limit is enforced. ### ### at linshui county, guang'an, the expressway encounters g65 baotou - maoming expressway and turns ### ### northwest bound. it then bypasses guang'an and continues northwest until intersecting with the ### ### unsigned northbound nanchong ring expressway and southbound g75 lanzhou - haikou expressway in ### ### nanchong. the expressway would then continue southwest bound and meet g93 chengdu - chongqing ring ### ### expressway in chuanshan district, suining, and bypasses suining to the north of its downtown. it ### ### crosses the fu river and intersects with ### ### =============================================== v_q ================================================ ### ### ('.', 0, 91) ('sui', 1, 3) ('##ning', 2, 4) ('##chon', 3, 5) ('fu', 4, 6) ('##iji', 5, 14915) ### ### ('river', 6, 24) ('##ang', 7, 25) ('sichuan', 8, 3565) ('cheng', 9, 10) ('##du', 10, 35) ### ### ('mia', 11, 10440) ('basin', 12, 5395) ('z', 13, 11621) ('##yang', 14, 15843) ('##qing', 15, 26) ### ### ('nan', 16, 1) ('china', 17, 158) ('ne', 18, 16257) ('##nya', 19, 27024) ### ### =============================================== v_p ================================================ ### ### ('ring', 827, 0) ('nan', 16, 1) ('lin', 7010, 2) ('sui', 1, 3) ('##ning', 2, 4) ('##chon', 3, 5) ### ### ('fu', 4, 6) ('mao', 73, 7) ('bao', 8986, 8) ('##ans', 9221, 9) ('cheng', 9, 10) ### ### ('##tou', 28332, 11) ('hai', 18262, 12) ('shanghai', 50, 13) ('##75', 20993, 14) ### ### ('expressway', 24641, 15) ('##ming', 72, 16) ('chu', 1031, 17) ('##42', 12536, 18) ### ### ('##zhou', 2681, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('sui', 1, 3) ('##ning', 2, 4) ('##chon', 3, 5) ('fu', 4, 6) ('cheng', 9, 10) ('nan', 16, 1) ### ### ('river', 6, 24) ('##ang', 7, 25) ('##du', 10, 35) ('##qing', 15, 26) ('gu', 22, 29) ### ### ('##g', 23, 42) ('shanghai', 50, 13) ('chong', 27, 37) ('.', 0, 91) ('mao', 73, 7) ### ### ('##ming', 72, 16) ('ring', 827, 0) ('county', 84, 32) ('chu', 1031, 17) ### ############################################################################################################ [2023-11-25 00:54:22,714][root][INFO] - Epoch: 1: Step: 6201/19530, loss[v]=0.452041, lr=0.000006, acc@1[1]=940.5/1024=0.91845703125, acc@1[2]=959.0/1024=0.9365234375 [2023-11-25 00:56:13,644][root][INFO] - Train batch 6300 [2023-11-25 00:56:13,645][root][INFO] - Avg. loss per last 100 batches: 0.321073 [2023-11-25 00:56:14,367][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.05 | max: 3.73 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.04 | max: 4.64 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: easy a ( stylized as easy a ) is a 2010 american teen comedy - drama film directed by ### ### will gluck, written by bert v. royal and starring emma stone, stanley tucci, patricia clarkson, ### ### thomas haden church, dan byrd, amanda bynes, penn badgley, cam gigandet, lisa kudrow, aly michalka, ### ### and malcolm mcdowell. ### ### [P_TEXT]: the boy least likely to the song " when life gives me lemons, i make lemonade " was used ### ### in the 2010 emma stone comedy " easy a " and in the 2018 film " peter rabbit ", which were both ### ### directed by will gluck. musicians known to have toured with the boy least likely to include : ### ### =============================================== v_q ================================================ ### ### ('easy', 0, 0) ('byrd', 1, 21788) ('.', 2, 790) ('bert', 3, 14541) ('film', 4, 21) ### ### ('patricia', 5, 6034) ('##nes', 6, 22377) ('amanda', 7, 6514) ('royal', 8, 13778) ### ### ('##and', 9, 23088) ('2010', 10, 14) ('stone', 11, 16) ('church', 12, 6813) ('malcolm', 13, 4867) ### ### ('mcdowell', 14, 14878) ('##cci', 15, 29061) ('will', 16, 8) ('penn', 17, 14727) ### ### ('stanley', 18, 6414) ('emma', 19, 9) ### ### =============================================== v_p ================================================ ### ### ('easy', 0, 0) ('lemon', 6488, 1) ('rabbit', 8174, 2) ('likely', 1453, 3) ('##ade', 17630, 4) ### ### ('boy', 6128, 5) ('##lu', 21, 6) ('least', 17095, 7) ('will', 16, 8) ('emma', 19, 9) ### ### ('peter', 1534, 10) ('##ck', 20, 11) ('life', 6320, 12) ('song', 3160, 13) ('2010', 10, 14) ### ### ('make', 2303, 15) ('stone', 11, 16) ('musicians', 8867, 17) ('2018', 95, 18) ('rabbits', 6207, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('easy', 0, 0) ('will', 16, 8) ('##lu', 21, 6) ('emma', 19, 9) ('##ck', 20, 11) ('2010', 10, 14) ### ### ('stone', 11, 16) ('film', 4, 21) ('easier', 22, 20) ('g', 35, 23) ('comedy', 44, 22) ### ### ('ease', 41, 30) ('simple', 43, 32) ('2018', 95, 18) ('2009', 42, 43) ('likely', 1453, 3) ### ### ('.', 2, 790) ('hard', 46, 47) ('peter', 1534, 10) ('movie', 45, 49) ### ############################################################################################################ [2023-11-25 00:56:14,868][root][INFO] - Epoch: 1: Step: 6301/19530, loss[v]=0.418083, lr=0.000006, acc@1[1]=933.5/1024=0.91162109375, acc@1[2]=965.5/1024=0.94287109375 [2023-11-25 00:58:05,618][root][INFO] - Train batch 6400 [2023-11-25 00:58:05,619][root][INFO] - Avg. loss per last 100 batches: 0.294282 [2023-11-25 00:58:06,343][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.05 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.04 | max: 4.50 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the biggest predator of the region is the puma followed by the maned wolf, pampas fox, ### ### geoffroy's cat, lesser grison as well as the omnivorous white - eared opossum and molinas hog - ### ### nosed skunk ### ### [P_TEXT]: geoffroy's cat of 106. 3. the geoffroy's cat occurs in chile, bolivia, brazil, argentina, ### ### paraguay and uruguay. it inhabits pampas and savanna landscapes in the gran chaco from southern ### ### bolivia to the straits of magellan. in the andes it occurs up to altitudes of. it prefers open ### ### woodland or scrubland with plenty of cover. it also occurs in grasslands and marshy areas. it seems ### ### to be common in central south american regions. in bolivia, it is the second most common cat after ### ### the ocelot. the geoffroy's cat is nocturnal and a ### ### =============================================== v_q ================================================ ### ### ('##ison', 0, 24577) ('.', 1, 107) ('##oss', 2, 22576) ('##pas', 3, 13) ('pu', 4, 18762) ### ### ('##roy', 5, 4) ('fox', 6, 1607) ('##unk', 7, 16541) ('wolf', 8, 6886) ('sk', 9, 20230) ### ### ('lesser', 10, 25266) ('##ma', 11, 19824) ('ear', 12, 10665) ('hog', 13, 9844) ### ### ('molina', 14, 19441) ('geoff', 15, 1) ('predator', 16, 7180) ('##um', 17, 25248) ('cat', 18, 6) ### ### ('mane', 19, 25820) ### ### =============================================== v_p ================================================ ### ### ('straits', 12267, 0) ('geoff', 15, 1) ('savanna', 9192, 2) ('nocturnal', 19418, 3) ('##roy', 5, 4) ### ### ('landscapes', 7505, 5) ('cat', 18, 6) ('bolivia', 1604, 7) ('##llan', 24954, 8) ('scrub', 2417, 9) ### ### ('##cel', 17395, 10) ('pam', 25, 11) ('gran', 14326, 12) ('##pas', 3, 13) ('plenty', 8298, 14) ### ### ('cha', 8699, 15) ('paraguay', 3146, 16) ('regions', 83, 17) ('chile', 864, 18) ('##co', 19291, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##roy', 5, 4) ('geoff', 15, 1) ('##pas', 3, 13) ('cat', 18, 6) ('pam', 25, 11) ('region', 24, 46) ### ### ('regions', 83, 17) ('.', 1, 107) ('argentina', 67, 28) ('bolivia', 1604, 7) ('chile', 864, 18) ### ### ('cats', 26, 61) ('grassland', 147, 42) ('scrub', 2417, 9) ('brazil', 1145, 26) ('common', 718, 31) ### ### ('paraguay', 3146, 16) ('savanna', 9192, 2) ('fox', 6, 1607) ('landscapes', 7505, 5) ### ############################################################################################################ [2023-11-25 00:58:06,847][root][INFO] - Epoch: 1: Step: 6401/19530, loss[v]=0.153605, lr=0.000007, acc@1[1]=980.5/1024=0.95751953125, acc@1[2]=1003.5/1024=0.97998046875 [2023-11-25 00:59:57,907][root][INFO] - Train batch 6500 [2023-11-25 00:59:57,908][root][INFO] - Avg. loss per last 100 batches: 0.300716 [2023-11-25 00:59:58,648][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.05 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.04 | max: 4.44 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: government organised these atolls into 21 administrative divisions. ### ### [P_TEXT]: administrative divisions of the maldives islands, 17 atolls and 4 cities in the maldives. ### ### during the regime of maumoon abdul gayyoom, the administrative divisions consisted of 20 ### ### administrative atolls, all controlled by the central government in male. in 2008, in an attempt of ### ### decentralization, the nasheed government divided the country into seven provinces. this maneuver ### ### was the first time that governing power has been moved away from the capital male'since 1117 a. d. ### ### however, the bill submitted by the government to decentralize the country according to this system ### ### was rejected by the opposition - majority parliament, who amended the ### ### =============================================== v_q ================================================ ### ### ('atoll', 0, 3) ('.', 1, 219) ('administrative', 2, 14) ('government', 3, 30) ('islands', 4, 20) ### ### ('island', 5, 54) ('division', 6, 38) ('hawaiian', 7, 201) ('divisions', 8, 21) (',', 9, 11111) ### ### ('subdistrict', 10, 15672) ('arrondissement', 11, 9319) ('fiji', 12, 185) ('archipelago', 13, 1063) ### ### ('samoa', 14, 2254) ('21', 15, 88) ('crete', 16, 1429) ('township', 17, 6930) ### ### ('administratively', 18, 166) ('realizing', 19, 266) ### ### =============================================== v_p ================================================ ### ### ('maldives', 126, 0) ('##eed', 26684, 1) ('male', 1395, 2) ('atoll', 0, 3) ('nash', 6739, 4) ### ### ('##om', 20977, 5) ('##yo', 28257, 6) ('gay', 7021, 7) ('provinces', 17093, 8) ('amended', 8485, 9) ### ### ('##ization', 4522, 10) ('111', 14086, 11) ('cities', 6208, 12) ('bill', 12666, 13) ### ### ('administrative', 2, 14) ('ma', 7631, 15) ('##oon', 16925, 16) ('regime', 2124, 17) ### ### ('##um', 21771, 18) ('seven', 1258, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('atoll', 0, 3) ('administrative', 2, 14) ('islands', 4, 20) ('government', 3, 30) ### ### ('divisions', 8, 21) ('division', 6, 38) ('maldives', 126, 0) ('island', 5, 54) ('.', 1, 219) ### ### ('divided', 56, 33) ('male', 1395, 2) ('capital', 626, 23) ('males', 348, 29) ### ### ('parliament', 529, 27) ('seven', 1258, 19) ('hawaiian', 7, 201) ('regime', 2124, 17) ### ### ('21', 15, 88) ('nash', 6739, 4) ('governing', 1035, 36) ### ############################################################################################################ [2023-11-25 00:59:59,150][root][INFO] - Epoch: 1: Step: 6501/19530, loss[v]=0.145068, lr=0.000007, acc@1[1]=983.5/1024=0.96044921875, acc@1[2]=1001.5/1024=0.97802734375 [2023-11-25 01:01:50,106][root][INFO] - Train batch 6600 [2023-11-25 01:01:50,107][root][INFO] - Avg. loss per last 100 batches: 0.300766 [2023-11-25 01:01:50,850][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.05 | max: 3.68 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.04 | max: 4.72 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: into the blues is the sixteenth studio album by british singer - songwriter joan ### ### armatrading, released on 1 may 2007. ### ### [P_TEXT]: starlight ( joan armatrading album ) starlight is the eighteenth studio album by british ### ### singer - songwriter joan armatrading, and was released in 2012 as a digital album on the ### ### hypertension label, ( hyp 12287 ) and then as a cd in 2013 on the 429 records ( ftn 17925 ) and ### ### savoy labels. the album was recorded at bumpkin studios, armatrading's own purpose - built studio. ### ### " starlight " was recorded during 2011 and is the final album in a trilogy of albums by armatrading ### ### each concerned with a specific musical genre. the first two albums, " into the blues " and " this ### ### charming life ", focused ### ### =============================================== v_q ================================================ ### ### ('blues', 0, 15) ('joan', 1, 7) ('into', 2, 26) ('.', 3, 119) ('arm', 4, 12) ('sixteenth', 5, 1334) ### ### ('##rad', 6, 14) ('album', 7, 18) ('2007', 8, 191) ('onto', 9, 74) ('##ing', 10, 39) ### ### ('singer', 11, 46) ('second', 12, 2529) ('##at', 13, 24) ('tenth', 14, 1960) ('jazz', 15, 420) ### ### ('fifth', 16, 3118) (',', 17, 12279) ('felix', 18, 1060) ('british', 19, 55) ### ### =============================================== v_p ================================================ ### ### ('##tension', 11850, 0) ('##light', 17824, 1) ('savoy', 17908, 2) ('bump', 5389, 3) ### ### ('charming', 7966, 4) ('##kin', 18905, 5) ('star', 6974, 6) ('joan', 1, 7) ('concerned', 2015, 8) ### ### ('##yp', 25766, 9) ('trilogy', 7848, 10) ('label', 198, 11) ('arm', 4, 12) ('hyper', 17991, 13) ### ### ('##rad', 6, 14) ('blues', 0, 15) ('ft', 6420, 16) ('eighteenth', 43, 17) ('album', 7, 18) ### ### ('2011', 216, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('joan', 1, 7) ('blues', 0, 15) ('arm', 4, 12) ('into', 2, 26) ('##rad', 6, 14) ('album', 7, 18) ### ### ('##at', 13, 24) ('##ing', 10, 39) ('singer', 11, 46) ('.', 3, 119) ('eighteenth', 43, 17) ### ### ('cd', 67, 27) ('label', 198, 11) ('onto', 9, 74) ('2011', 216, 19) ('musical', 101, 29) ### ### ('british', 19, 55) ('concerned', 2015, 8) ('2007', 8, 191) ('albums', 128, 42) ### ############################################################################################################ [2023-11-25 01:01:51,352][root][INFO] - Epoch: 1: Step: 6601/19530, loss[v]=0.395524, lr=0.000007, acc@1[1]=936.0/1024=0.9140625, acc@1[2]=963.0/1024=0.9404296875 [2023-11-25 01:03:42,126][root][INFO] - Train batch 6700 [2023-11-25 01:03:42,126][root][INFO] - Avg. loss per last 100 batches: 0.315360 [2023-11-25 01:03:42,869][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.05 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.04 | max: 4.68 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: administrative supervision was transferred to the viceroyalty of new granada when this ### ### was created in 1717, and in 1777 to the new captaincy general of venezuela. ### ### [P_TEXT]: ruy fernandez de fuenmayor ruy fernandez de fuenmayor ( santo domingo, 1603 - venezuela, ### ### 1651 ) was a spanish soldier who was governor of venezuela province between 1637 and 1644. ### ### fuenmayor was born in 1603, the son of fernandez de fuenmayor and juana figueroa. he served in ### ### various positions in spain, puerto rico and hispaniola. he was general of the galleys in ### ### hispaniola, and in 1635 led an attack on buccaneers who had settled on tortuga, expelling them from ### ### the island for several years. in this raid, he led 250 spanish soldiers. ### ### =============================================== v_q ================================================ ### ### ('venezuela', 0, 0) ('.', 1, 171) ('granada', 2, 17897) ('viceroy', 3, 15574) ### ### ('supervision', 4, 6899) ('captain', 5, 2996) ('new', 6, 12620) ('1777', 7, 9603) ### ### ('administrative', 8, 5382) ('castile', 9, 1486) ('generals', 10, 88) ('general', 11, 41) ### ### ('newly', 12, 375) ('venezuelan', 13, 42) ('caracas', 14, 47) ('ж', 15, 1405) ('spain', 16, 19) ### ### ('transferred', 17, 948) ('##ty', 18, 20684) ('##cy', 19, 22154) ### ### =============================================== v_p ================================================ ### ### ('venezuela', 0, 0) ('##tu', 7392, 1) ('##pani', 13694, 2) ('##may', 26316, 3) ('##ola', 5864, 4) ### ### ('raid', 2417, 5) ('galley', 9187, 6) ('domingo', 13809, 7) ('fernandez', 3014, 8) ### ### ('tor', 12945, 9) ('##uer', 27498, 10) ('ru', 12279, 11) ('santo', 13089, 12) ('fu', 4251, 13) ### ### ('buccaneers', 27335, 14) ('governor', 20, 15) ('rico', 15772, 16) ('##ga', 11369, 17) ### ### ('163', 12262, 18) ('spain', 16, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('venezuela', 0, 0) ('spain', 16, 19) ('general', 11, 41) ('governor', 20, 15) ### ### ('venezuelan', 13, 42) ('province', 54, 21) ('caracas', 14, 47) ('.', 1, 171) ('spanish', 72, 32) ### ### ('generals', 10, 88) ('raid', 2417, 5) ('settled', 719, 25) ('provinces', 221, 36) ### ### ('fernandez', 3014, 8) ('##ola', 5864, 4) ('fu', 4251, 13) ('##tu', 7392, 1) ('juan', 2482, 23) ### ### ('soldier', 2480, 24) ('newly', 12, 375) ### ############################################################################################################ [2023-11-25 01:03:43,373][root][INFO] - Epoch: 1: Step: 6701/19530, loss[v]=0.366110, lr=0.000007, acc@1[1]=942.0/1024=0.919921875, acc@1[2]=966.0/1024=0.943359375 [2023-11-25 01:05:34,598][root][INFO] - Train batch 6800 [2023-11-25 01:05:34,599][root][INFO] - Avg. loss per last 100 batches: 0.283988 [2023-11-25 01:05:35,326][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.05 | max: 3.94 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.04 | max: 4.52 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: robb, an older sister sansa, two younger brothers bran and rickon, and an older ### ### illegitimate half - brother, jon snow. ### ### [P_TEXT]: robb stark the beginning of " a game of thrones " ( 17 on the tv series ). he is the ### ### oldest legitimate son of eddard " ned " stark and his wife catelyn, and has five siblings : younger ### ### sisters sansa and arya, younger brothers bran and rickon, and a bastard half - brother jon snow. he ### ### is also a close friend to eddard's ward theon greyjoy. robb is constantly accompanied by his ### ### direwolf, grey wind. like most of his siblings, robb inherited his mother's tully features, being ### ### strong and fast with ### ### =============================================== v_q ================================================ ### ### ('robb', 0, 1) ('sans', 1, 4) ('bran', 2, 14) ('snow', 3, 5) ('jon', 4, 7) ('.', 5, 212) ### ### ('rick', 6, 10) ('##on', 7, 31) ('brother', 8, 36) ('sister', 9, 56) ('illegitimate', 10, 672) ### ### ('brothers', 11, 37) ('sisters', 12, 29) ('stark', 13, 0) ('##ン', 14, 237) ('##а', 15, 77) ### ### ('##a', 16, 83) ('##ons', 17, 91) ('teenage', 18, 2753) ('bother', 19, 130) ### ### =============================================== v_p ================================================ ### ### ('stark', 13, 0) ('robb', 0, 1) ('tully', 2896, 2) ('arya', 1334, 3) ('sans', 1, 4) ('snow', 3, 5) ### ### ('theo', 5879, 6) ('jon', 4, 7) ('grey', 3076, 8) ('throne', 3132, 9) ('rick', 6, 10) ### ### ('wind', 7879, 11) ('##joy', 28062, 12) ('##dar', 17281, 13) ('bran', 2, 14) ('ned', 811, 15) ### ### ('##lf', 20149, 16) ('ed', 6630, 17) ('dire', 19455, 18) ('##ely', 25764, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('robb', 0, 1) ('sans', 1, 4) ('snow', 3, 5) ('jon', 4, 7) ('bran', 2, 14) ('rick', 6, 10) ### ### ('stark', 13, 0) ('##on', 7, 31) ('brother', 8, 36) ('sisters', 12, 29) ('brothers', 11, 37) ### ### ('sister', 9, 56) ('legitimate', 66, 22) ('inherited', 94, 23) ('.', 5, 212) ('arya', 1334, 3) ### ### ('siblings', 32, 33) ('tully', 2896, 2) ('son', 34, 35) ('bastard', 404, 20) ### ############################################################################################################ [2023-11-25 01:05:35,828][root][INFO] - Epoch: 1: Step: 6801/19530, loss[v]=0.401066, lr=0.000007, acc@1[1]=936.0/1024=0.9140625, acc@1[2]=959.5/1024=0.93701171875 [2023-11-25 01:07:26,538][root][INFO] - Train batch 6900 [2023-11-25 01:07:26,539][root][INFO] - Avg. loss per last 100 batches: 0.291552 [2023-11-25 01:07:27,270][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.05 | max: 3.72 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.04 | max: 4.64 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: i - 10 - d in van horn, bus. ### ### [P_TEXT]: u. s. route 90 state of florida. it also includes part of the between tallahassee and ### ### lake city, florida. us 90 begins at an intersection with bl i - 10 and sh 54 in downtown van horn. ### ### it then heads south - southeast towards marfa, where the route begins to head generally east. the ### ### route is mostly two lanes west of uvalde. at this point, it becomes a four - lane surface road ### ### until it reaches western bexar county where it becomes a freeway, joining i - 10 in downtown san ### ### antonio. this concurrency with i - 10 continues intermittently into western houston ### ### =============================================== v_q ================================================ ### ### ('horn', 0, 0) ('van', 1, 20) ('.', 2, 629) ('bus', 3, 2763) ('10', 4, 12) ('d', 5, 9356) ### ### ('i', 6, 24) ('interstate', 7, 10014) ('buses', 8, 2426) ('и', 9, 408) ('horns', 10, 57) ### ### ('intercity', 11, 5859) ('tucson', 12, 849) ('11', 13, 113) (',', 14, 7323) ### ### ('romanized', 15, 12335) ('9', 16, 75) ('ii', 17, 108) ('ten', 18, 45) ('idaho', 19, 10489) ### ### =============================================== v_p ================================================ ### ### ('horn', 0, 0) ('##ssee', 22856, 1) ('##ald', 13218, 2) ('90', 2206, 3) ('##xa', 17208, 4) ### ### ('##fa', 17129, 5) ('houston', 791, 6) ('uv', 19044, 7) ('lake', 1492, 8) ('mar', 4044, 9) ### ### ('florida', 3968, 10) ('downtown', 297, 11) ('10', 4, 12) ('##aha', 25909, 13) ### ### ('concurrency', 15852, 14) ('antonio', 332, 15) ('route', 39, 16) ('tall', 9311, 17) ### ### ('54', 11425, 18) ('surface', 2348, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('horn', 0, 0) ('van', 1, 20) ('10', 4, 12) ('i', 6, 24) ('route', 39, 16) ('horns', 10, 57) ### ### ('downtown', 297, 11) ('antonio', 332, 15) ('freeway', 51, 29) ('.', 2, 629) ('houston', 791, 6) ### ### ('ten', 18, 45) ('90', 2206, 3) ('intersection', 77, 30) ('b', 215, 25) ('highway', 29, 42) ### ### ('lake', 1492, 8) ('county', 406, 26) ('u', 174, 32) ('we', 33, 49) ### ############################################################################################################ [2023-11-25 01:07:27,773][root][INFO] - Epoch: 1: Step: 6901/19530, loss[v]=0.398911, lr=0.000007, acc@1[1]=929.0/1024=0.9072265625, acc@1[2]=960.5/1024=0.93798828125 [2023-11-25 01:09:18,823][root][INFO] - Train batch 7000 [2023-11-25 01:09:18,824][root][INFO] - Avg. loss per last 100 batches: 0.334074 [2023-11-25 01:09:19,567][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.05 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.04 | max: 4.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: yorkshire towns of morley and rothwell, the villages that surround the towns, and the old ### ### pit village of middleton. ### ### [P_TEXT]: morley, west yorkshire west yorkshire, the morley and rothwell parliamentary constituency ### ### was abolished at the 2010 election and replaced by the morley and outwood constituency. the town's ### ### coat of arms featured the symbolic principal industries of the municipal borough : textile ### ### manufacturing, coal mining and quarry ing. morley town hall is sometimes used for music recordings. ### ### television programmes, " heartbeat " and " emmerdale " have used its disused magistrates court and ### ### a cobblestone d street to one side. it hosts concert s ### ### =============================================== v_q ================================================ ### ### ('middleton', 0, 4969) ('morley', 1, 0) ('pit', 2, 5796) ('yorkshire', 3, 17) ('roth', 4, 3) ### ### ('##well', 5, 15) ('.', 6, 209) ('village', 7, 553) ('leeds', 8, 219) ('towns', 9, 68) ### ### ('manchester', 10, 1494) ('bbc', 11, 1754) ('town', 12, 24) ('lancashire', 13, 1342) ### ### ('rochester', 14, 101) ('wood', 15, 71) ('cumbria', 16, 3960) ('england', 17, 3776) ### ### ('distract', 18, 124) ('bradford', 19, 192) ### ### =============================================== v_p ================================================ ### ### ('morley', 1, 0) ('heartbeat', 14290, 1) ('quarry', 5271, 2) ('roth', 4, 3) ('ing', 22828, 4) ### ### ('arms', 10033, 5) ('textile', 2490, 6) ('magistrates', 18287, 7) ('election', 2527, 8) ### ### ('concert', 10847, 9) ('coat', 958, 10) ('##wood', 7443, 11) ('##dale', 3023, 12) ('d', 13063, 13) ### ### ('principal', 2444, 14) ('##well', 5, 15) ('##tone', 11755, 16) ('yorkshire', 3, 17) ### ### ('out', 6916, 18) ('coal', 4577, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('morley', 1, 0) ('roth', 4, 3) ('yorkshire', 3, 17) ('##well', 5, 15) ('town', 12, 24) ### ### ('borough', 68, 27) ('towns', 9, 68) ('.', 6, 209) ('constituency', 254, 20) ('coat', 958, 10) ### ### ('manufacturing', 671, 22) ('textile', 2490, 6) ('leeds', 8, 219) ('election', 2527, 8) ### ### ('municipal', 933, 29) ('village', 7, 553) ('court', 1430, 21) ('quarry', 5271, 2) ### ### ('principal', 2444, 14) ('wood', 15, 71) ### ############################################################################################################ [2023-11-25 01:09:20,072][root][INFO] - Epoch: 1: Step: 7001/19530, loss[v]=0.165500, lr=0.000007, acc@1[1]=981.0/1024=0.9580078125, acc@1[2]=999.5/1024=0.97607421875 [2023-11-25 01:11:10,962][root][INFO] - Train batch 7100 [2023-11-25 01:11:10,963][root][INFO] - Avg. loss per last 100 batches: 0.302686 [2023-11-25 01:11:11,705][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.05 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.04 | max: 4.53 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.1/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: historian john guy ( 1988 ) argued that " england was economically healthier, more ### ### expansive, and more optimistic under the tudors " than at any time in a hundred years. ### ### [P_TEXT]: john guy ( historian ) by the university of cambridge. during his academic career, guy ### ### has held posts at st andrews university ( where he is honorary professor and was sometime vice - ### ### principal for research ), bristol university, uc berkeley, rochester and johns hopkins. guy ### ### currently teaches at cambridge university, as a fellow of clare college, where he teaches part - ### ### time so he can devote more time to his writing and broadcasting career. guy specializes in the ### ### history of tudor england and has written extensively on the subject. his books have been critically ### ### acclaimed, with " my heart is my own : the ### ### =============================================== v_q ================================================ ### ### ('guy', 0, 1) ('england', 1, 17) ('john', 2, 8) ('tudor', 3, 3) ('.', 4, 80) ('health', 5, 4077) ### ### ('britain', 6, 40) ('historian', 7, 15) ('optimistic', 8, 11404) ('hundred', 9, 9018) ### ### ('medieval', 10, 73) ('##ier', 11, 22462) ('london', 12, 411) ('oxford', 13, 473) ### ### ('english', 14, 692) ('scotland', 15, 1113) ('economically', 16, 8481) ('expansive', 17, 16377) ### ### ('centuries', 18, 1672) ('uk', 19, 565) ### ### =============================================== v_p ================================================ ### ### ('clare', 11399, 0) ('guy', 0, 1) ('hopkins', 17945, 2) ('tudor', 3, 3) ('andrews', 12823, 4) ### ### ('rochester', 3417, 5) ('berkeley', 7867, 6) ('bristol', 25, 7) ('john', 2, 8) ('cambridge', 39, 9) ### ### ('honorary', 3493, 10) ('broadcasting', 11731, 11) ('heart', 5541, 12) ('johns', 5968, 13) ### ### ('sometime', 2271, 14) ('historian', 7, 15) ('principal', 6800, 16) ('england', 1, 17) ### ### ('uc', 19268, 18) ('university', 3518, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('guy', 0, 1) ('tudor', 3, 3) ('john', 2, 8) ('england', 1, 17) ('historian', 7, 15) ### ### ('bristol', 25, 7) ('britain', 6, 40) ('cambridge', 39, 9) ('history', 22, 23) ('.', 4, 80) ### ### ('historians', 28, 39) ('rochester', 3417, 5) ('medieval', 10, 73) ('guys', 91, 37) ### ### ('sometime', 2271, 14) ('honorary', 3493, 10) ('clare', 11399, 0) ('historical', 24, 57) ### ### ('berkeley', 7867, 6) ('research', 1545, 22) ### ############################################################################################################ [2023-11-25 01:11:12,207][root][INFO] - Epoch: 1: Step: 7101/19530, loss[v]=0.177497, lr=0.000007, acc@1[1]=979.5/1024=0.95654296875, acc@1[2]=1003.5/1024=0.97998046875 [2023-11-25 01:13:02,992][root][INFO] - Train batch 7200 [2023-11-25 01:13:02,993][root][INFO] - Avg. loss per last 100 batches: 0.293865 [2023-11-25 01:13:03,741][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.05 | max: 3.69 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.04 | max: 4.88 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: its shallow, warm waters cause frequent, sometimes heavy lake - effect snow s from late ### ### fall through spring. ### ### [P_TEXT]: lake - effect snow the downwind shores. this uplifting can produce narrow but very ### ### intense bands of precipitation, which deposit at a rate of many inches of snow each hour, often ### ### resulting in a large amount of total snowfall. the areas affected by lake - effect snow are called ### ### snowbelt s. these include areas east of the great lakes, the west coasts of northern japan, the ### ### kamchatka peninsula in russia, and areas near the great salt lake, black sea, caspian sea, baltic ### ### sea, adriatic sea ### ### =============================================== v_q ================================================ ### ### ('.', 0, 1183) ('snow', 1, 6) ('spring', 2, 2830) ('lake', 3, 1) ('fall', 4, 6407) ('effect', 5, 2) ### ### ('warm', 6, 1304) ('heavy', 7, 1033) ('shallow', 8, 1731) ('frequent', 9, 1608) ('river', 10, 82) ### ### ('late', 11, 7856) ('waters', 12, 8648) ('effects', 13, 32) ('freshwater', 14, 2682) ### ### ('glacier', 15, 2679) ('water', 16, 883) ('is', 17, 14345) (',', 18, 18332) ('sometimes', 19, 1369) ### ### =============================================== v_p ================================================ ### ### ('##belt', 28615, 0) ('lake', 3, 1) ('effect', 5, 2) ('##lifting', 21528, 3) ('##wind', 26051, 4) ### ### ('baltic', 15650, 5) ('snow', 1, 6) ('lakes', 26, 7) ('japan', 5103, 8) ('rate', 1675, 9) ### ### ('peninsula', 485, 10) ('##cha', 25599, 11) ('salt', 4336, 12) ('kam', 13967, 13) ### ### ('precipitation', 29, 14) ('adriatic', 25706, 15) ('belt', 15618, 16) ('snowfall', 44, 17) ### ### ('great', 6783, 18) ('russia', 1924, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('snow', 1, 6) ('lake', 3, 1) ('effect', 5, 2) ('lakes', 26, 7) ('precipitation', 29, 14) ### ### ('effects', 13, 32) ('snowfall', 44, 17) ('peninsula', 485, 10) ('sea', 321, 23) ('.', 0, 1183) ### ### ('rate', 1675, 9) ('river', 10, 82) ('often', 28, 46) ('russia', 1924, 19) ('intense', 631, 30) ### ### ('japan', 5103, 8) ('salt', 4336, 12) ('warm', 6, 1304) ('inches', 3902, 21) ('ski', 24, 70) ### ############################################################################################################ [2023-11-25 01:13:04,246][root][INFO] - Epoch: 1: Step: 7201/19530, loss[v]=0.165516, lr=0.000007, acc@1[1]=976.0/1024=0.953125, acc@1[2]=1009.0/1024=0.9853515625 [2023-11-25 01:14:55,072][root][INFO] - Train batch 7300 [2023-11-25 01:14:55,073][root][INFO] - Avg. loss per last 100 batches: 0.281193 [2023-11-25 01:14:55,810][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.05 | max: 3.72 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.04 | max: 4.54 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: frans post to dutch brazil to have them record the country's landscape, inhabitants, ### ### flora and fauna. ### ### [P_TEXT]: dutch brazil the dutch captured the province of maranhao, meaning that dutch control now ### ### extended across the entire coastline between the amazon and sao francisco rivers. maurits claimed ### ### to have always loved brazil due to its beauty and its people, and under his rule, the colony ### ### thrived. his patronage of dutch golden age painters, such as albert eckhout and frans post, to ### ### depict brazil's richness resulted in works showing different races, landscapes, and still lifes. he ### ### also invited naturalists georg marcgraf and willem piso to brazil. they collected and published a ### ### =============================================== v_q ================================================ ### ### ('brazil', 0, 1) ('.', 1, 118) ('dutch', 2, 0) ('fran', 3, 32) ('brazilian', 4, 50) ('post', 5, 3) ### ### ('flora', 6, 4194) ('inhabitants', 7, 1954) ('landscape', 8, 53) ('fauna', 9, 8273) ### ### ('brasil', 10, 283) ('country', 11, 9075) ('netherlands', 12, 51) ('species', 13, 3123) ### ### ('geological', 14, 1145) ('vegetation', 15, 244) (',', 16, 12981) ('postal', 17, 340) ### ### ('wingspan', 18, 343) ('portugal', 19, 1272) ### ### =============================================== v_p ================================================ ### ### ('dutch', 2, 0) ('brazil', 0, 1) ('##out', 14521, 2) ('post', 5, 3) ('georg', 13018, 4) ### ### ('age', 5681, 5) ('##uri', 18467, 6) ('painters', 10353, 7) ('albert', 3580, 8) ### ### ('patronage', 7717, 9) ('##so', 19986, 10) ('francisco', 1018, 11) ('rivers', 5189, 12) ### ### ('mara', 9040, 13) ('marc', 12836, 14) ('naturalist', 189, 15) ('ec', 9729, 16) ('colony', 107, 17) ### ### ('landscapes', 37, 18) ('##gra', 22388, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('brazil', 0, 1) ('dutch', 2, 0) ('post', 5, 3) ('fran', 3, 32) ('brazilian', 4, 50) ### ### ('landscape', 8, 53) ('landscapes', 37, 18) ('.', 1, 118) ('netherlands', 12, 51) ### ### ('colony', 107, 17) ('captured', 59, 25) ('naturalist', 189, 15) ('people', 24, 43) ### ### ('province', 206, 28) ('francisco', 1018, 11) ('willem', 781, 22) ('amazon', 407, 33) ### ### ('beauty', 786, 29) ('rich', 737, 31) ('albert', 3580, 8) ### ############################################################################################################ [2023-11-25 01:14:56,310][root][INFO] - Epoch: 1: Step: 7301/19530, loss[v]=0.151958, lr=0.000007, acc@1[1]=984.0/1024=0.9609375, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-25 01:16:47,019][root][INFO] - Train batch 7400 [2023-11-25 01:16:47,020][root][INFO] - Avg. loss per last 100 batches: 0.298540 [2023-11-25 01:16:47,741][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.05 | max: 3.71 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.04 | max: 4.54 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 1999, the vincent scully prize was established by the national building museum to ### ### honor individuals who have exhibited exemplary practice, scholarship or criticism in architecture, ### ### historic preservation and urban design. ### ### [P_TEXT]: vincent scully prize the vincent scully prize was established in 1999 to recognize ### ### exemplary practice, scholarship or criticism in architecture, historic preservation and urban ### ### design. created by the national building museum in washington, d. c., the award first honored the ### ### distinguished yale professor and namesake of the award, author and educator, vincent scully. the ### ### museum ’ s website states that the prize is awarded annually, however no award was made in 2003, ### ### 2004, 2015 or 2016. these omissions are not explained on the website. the 2014 prize was presented ### ### to former talk show host charlie rose. the museum ### ### =============================================== v_q ================================================ ### ### ('scully', 0, 0) ('building', 1, 6) ('vincent', 2, 1) ('.', 3, 199) ('museum', 4, 11) ### ### ('prize', 5, 7) ('national', 6, 23) ('design', 7, 14) ('architecture', 8, 37) ('historic', 9, 13) ### ### ('urban', 10, 9) ('preservation', 11, 5) ('nacional', 12, 929) ('build', 13, 54) ### ### ('criticism', 14, 25) ('construction', 15, 77) ('buildings', 16, 51) ('prizes', 17, 16) ### ### ('scholar', 18, 232) ('architect', 19, 306) ### ### =============================================== v_p ================================================ ### ### ('scully', 0, 0) ('vincent', 2, 1) ('rose', 1553, 2) ('charlie', 2507, 3) ('yale', 4040, 4) ### ### ('preservation', 11, 5) ('building', 1, 6) ('prize', 5, 7) ('washington', 641, 8) ('urban', 10, 9) ### ### ('educator', 4122, 10) ('museum', 4, 11) ('namesake', 16747, 12) ('historic', 9, 13) ### ### ('design', 7, 14) ('website', 5289, 15) ('prizes', 17, 16) ('host', 4287, 17) ('2014', 5513, 18) ### ### ('distinguished', 2008, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('scully', 0, 0) ('vincent', 2, 1) ('building', 1, 6) ('prize', 5, 7) ('museum', 4, 11) ### ### ('preservation', 11, 5) ('design', 7, 14) ('national', 6, 23) ('historic', 9, 13) ('urban', 10, 9) ### ### ('architecture', 8, 37) ('prizes', 17, 16) ('criticism', 14, 25) ('scholarship', 26, 32) ### ### ('honored', 38, 22) ('.', 3, 199) ('1999', 28, 36) ('practice', 39, 24) ('build', 13, 54) ### ### ('buildings', 16, 51) ### ############################################################################################################ [2023-11-25 01:16:48,243][root][INFO] - Epoch: 1: Step: 7401/19530, loss[v]=0.311324, lr=0.000008, acc@1[1]=951.5/1024=0.92919921875, acc@1[2]=982.5/1024=0.95947265625 [2023-11-25 01:18:38,922][root][INFO] - Train batch 7500 [2023-11-25 01:18:38,922][root][INFO] - Avg. loss per last 100 batches: 0.282597 [2023-11-25 01:18:39,649][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.05 | max: 3.80 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.04 | max: 4.93 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: david wrote 62 of the episodes of " seinfeld ", including 1992's " the contest ", for ### ### which he won a primetime emmy award and which " tv guide " ranked as episode 1 on its list of " ### ### tv's top 100 episodes of all time ". ### ### [P_TEXT]: seinfeld seinfeld ( ; ) is an american sitcom television series created by larry david ### ### and jerry seinfeld. it aired on nbc from july 5, 1989, to may 14, 1998, over nine seasons and 180 ### ### episodes. the show stars seinfeld as a fictionalized version of himself who is the main protagonist ### ### and focuses on his personal life with three of his friends – george costanza ( jason alexander ), ### ### former girlfriend elaine benes ( julia louis - dreyfus ), and neighbor across ### ### =============================================== v_q ================================================ ### ### ('contest', 0, 11005) ('david', 1, 8) ('guide', 2, 1579) ('##feld', 3, 0) ('tv', 4, 49) ### ### ('.', 5, 134) ('episode', 6, 35) ('se', 7, 1) ('television', 8, 38) ('time', 9, 13818) ### ### ('1992', 10, 1640) ('competition', 11, 1977) ('##in', 12, 16) ('guides', 13, 19406) ### ### ('contests', 14, 14616) ('emmy', 15, 20692) ('episodes', 16, 39) ('list', 17, 11711) ### ### ('awards', 18, 5177) ('award', 19, 20659) ### ### =============================================== v_p ================================================ ### ### ('##feld', 3, 0) ('se', 7, 1) ('elaine', 4448, 2) ('larry', 1662, 3) ('ben', 566, 4) ### ### ('julia', 1022, 5) ('##fus', 26893, 6) ('girlfriend', 3101, 7) ('david', 1, 8) ('##nza', 19263, 9) ### ### ('seasons', 1632, 10) ('louis', 6766, 11) ('sitcom', 75, 12) ('costa', 7058, 13) ### ### ('neighbor', 10219, 14) ('alexander', 5552, 15) ('##in', 12, 16) ('jason', 750, 17) ('nbc', 60, 18) ### ### ('180', 19646, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##feld', 3, 0) ('david', 1, 8) ('se', 7, 1) ('episode', 6, 35) ('##in', 12, 16) ### ### ('television', 8, 38) ('tv', 4, 49) ('aired', 23, 24) ('episodes', 16, 39) ('sitcom', 75, 12) ### ### ('nbc', 60, 18) ('series', 43, 30) ('.', 5, 134) ('ben', 566, 4) ('julia', 1022, 5) ### ### ('george', 252, 20) ('season', 115, 29) ('larry', 1662, 3) ('jerry', 393, 22) ('show', 168, 31) ### ############################################################################################################ [2023-11-25 01:18:40,150][root][INFO] - Epoch: 1: Step: 7501/19530, loss[v]=0.151161, lr=0.000008, acc@1[1]=981.0/1024=0.9580078125, acc@1[2]=1003.0/1024=0.9794921875 [2023-11-25 01:20:31,057][root][INFO] - Train batch 7600 [2023-11-25 01:20:31,058][root][INFO] - Avg. loss per last 100 batches: 0.290622 [2023-11-25 01:20:31,790][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.05 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.04 | max: 4.62 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he made eighteen of twenty attempts with the team, and was hired to drive for the team ### ### full - time in 2010, with the team being offered richard childress racing chassis and hendrick ### ### motorsports engines. ### ### [P_TEXT]: hendrick motorsports sam roadside assistance 500 at talladega, where he was involved in a ### ### second hard crash. after testing, earnhardt was deemed medically unfit to race. regan smith, ### ### scheduled to drive for phoenix racing, instead drove the no. 88 at charlotte and kansas in what ### ### turned out to be a tryout that led to smith joining earnhardt's xfinity team for 2013. earnhardt ### ### returned at martinsville and finished out the season. earnhardt finished fifth in cup series points ### ### in 2013, after winning two poles ( at kentucky speedway and dover international speedway ), and ### ### =============================================== v_q ================================================ ### ### ('.', 0, 270) ('hen', 1, 3) ('richard', 2, 4503) ('child', 3, 5125) ('##ress', 4, 24279) ### ### ('2010', 5, 8309) ('chassis', 6, 1609) ('racing', 7, 27) ('motorsports', 8, 0) ('engine', 9, 1050) ### ### ('race', 10, 52) ('##drick', 11, 2) ('vehicles', 12, 97) ('nascar', 13, 612) ('ferrari', 14, 1763) ### ### ('team', 15, 39) ('2011', 16, 1269) ('twenty', 17, 2249) ('baby', 18, 3042) ('2009', 19, 1656) ### ### =============================================== v_p ================================================ ### ### ('motorsports', 8, 0) ('regan', 3986, 1) ('##drick', 11, 2) ('hen', 1, 3) ('martins', 6584, 4) ### ### ('charlotte', 2292, 5) ('speedway', 183, 6) ('##hardt', 13444, 7) ('smith', 8560, 8) ### ### ('tall', 9754, 9) ('phoenix', 201, 10) ('dover', 4936, 11) ('##fin', 14560, 12) ### ### ('kentucky', 4463, 13) ('x', 5702, 14) ('earn', 7402, 15) ('88', 12532, 16) ('try', 2172, 17) ### ### ('kansas', 645, 18) ('poles', 12439, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('hen', 1, 3) ('motorsports', 8, 0) ('##drick', 11, 2) ('racing', 7, 27) ('speedway', 183, 6) ### ### ('crash', 56, 24) ('team', 15, 39) ('race', 10, 52) ('motorsport', 38, 34) ('phoenix', 201, 10) ### ### ('series', 41, 36) ('.', 0, 270) ('points', 202, 29) ('kansas', 645, 18) ('cup', 476, 22) ### ### ('charlotte', 2292, 5) ('testing', 152, 35) ('season', 341, 28) ('2013', 895, 20) ### ### ('regan', 3986, 1) ### ############################################################################################################ [2023-11-25 01:20:32,290][root][INFO] - Epoch: 1: Step: 7601/19530, loss[v]=0.202031, lr=0.000008, acc@1[1]=988.0/1024=0.96484375, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-25 01:22:22,924][root][INFO] - Train batch 7700 [2023-11-25 01:22:22,925][root][INFO] - Avg. loss per last 100 batches: 0.309657 [2023-11-25 01:22:23,650][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.05 | max: 3.71 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.04 | max: 4.91 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: one available position was that of captain for the expedition's vessel, the " endurance ### ### ". ### ### [P_TEXT]: endurance ( 1912 ship ) structure groaned and wracked under the strain. carpenter harry ### ### mcnish noted that the solid oak beams supporting the upper deck were being visibly bent " like a ### ### piece of cane ". on deck the ship's masts were whipping back and forth as their stepping points on ### ### the keel were distorted. despite these disconcerting signs, captain frank worsley noted that the ### ### strength of the ship's structure was causing the ice itself to break up as it piled against the ### ### hull – "... just as it appears she can stand no more, the huge floe weighing possibly a million ### ### =============================================== v_q ================================================ ### ### ('endurance', 0, 0) ('expedition', 1, 2980) ('.', 2, 158) ('vessel', 3, 79) ('ship', 4, 13) ### ### ('captain', 5, 35) ('sailing', 6, 1145) ('yacht', 7, 1666) ('antarctic', 8, 286) ('sailed', 9, 397) ### ### ('patience', 10, 41) ('boat', 11, 110) ('explorer', 12, 7765) ('ships', 13, 43) ### ### ('vessels', 14, 157) ('available', 15, 5973) ('hms', 16, 6300) ('antarctica', 17, 19342) ### ### ('deck', 18, 18) ('position', 19, 8558) ### ### =============================================== v_p ================================================ ### ### ('endurance', 0, 0) ('cane', 3933, 1) ('piled', 8665, 2) ('##nish', 20161, 3) ('groaned', 1949, 4) ### ### ('strain', 8756, 5) ('keel', 20851, 6) ('distorted', 4539, 7) ('wo', 24704, 8) ### ### ('whipping', 28046, 9) ('bent', 6365, 10) ('carpenter', 3928, 11) ('hull', 32, 12) ('ship', 4, 13) ### ### ('harry', 4050, 14) ('weighing', 22308, 15) ('mast', 2036, 16) ('oak', 14429, 17) ('deck', 18, 18) ### ### ('frank', 1630, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('endurance', 0, 0) ('ship', 4, 13) ('captain', 5, 35) ('deck', 18, 18) ('patience', 10, 41) ### ### ('hull', 32, 12) ('ships', 13, 43) ('vessel', 3, 79) ('.', 2, 158) ('groaned', 1949, 4) ### ### ('cane', 3933, 1) ('frank', 1630, 19) ('ice', 1411, 23) ('structure', 812, 28) ('boat', 11, 110) ### ### ('mast', 2036, 16) ('strength', 889, 31) ('distorted', 4539, 7) ('carpenter', 3928, 11) ### ### ('expedition', 1, 2980) ### ############################################################################################################ [2023-11-25 01:22:24,153][root][INFO] - Epoch: 1: Step: 7701/19530, loss[v]=0.143944, lr=0.000008, acc@1[1]=981.0/1024=0.9580078125, acc@1[2]=1007.0/1024=0.9833984375 [2023-11-25 01:24:15,034][root][INFO] - Train batch 7800 [2023-11-25 01:24:15,035][root][INFO] - Avg. loss per last 100 batches: 0.272646 [2023-11-25 01:24:15,783][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.05 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.04 | max: 4.73 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: for articulated locomotives that have two wheelsets, such as garratt s, which are ### ### effectively two locomotives joined by a common boiler, each wheelset is denoted separately, with a ### ### plus sign ( + ) between them. ### ### [P_TEXT]: 0 - 4 - 4 - 0 in the whyte notation for the classification of steam locomotive wheel ### ### arrangement, a 0 - 4 - 4 - 0 is a locomotive with no leading truck, two sets of four driving wheel ### ### s, and no trailing truck. examples of this type were constructed as shay, heisler, climax, mallet, ### ### meyer, bmag and double fairlie locomotives. a similar configuration was used on some garratt s, but ### ### it is referred to as 0 - 4 - 0 + 0 - 4 - 0. the uic classification is ### ### =============================================== v_q ================================================ ### ### ('sign', 0, 5781) ('##tt', 1, 23) ('wheels', 2, 41) ('.', 3, 90) ('##rra', 4, 5) ('ga', 5, 24) ### ### ('plus', 6, 1372) ('articulated', 7, 11664) ('boiler', 8, 258) ('locomotive', 9, 19) ### ### ('locomotives', 10, 26) ('+', 11, 51) ('railway', 12, 708) ('common', 13, 1818) ('train', 14, 1140) ### ### ('station', 15, 4594) ('signs', 16, 12952) ('electric', 17, 4139) ('railroad', 18, 1204) ### ### ('australia', 19, 282) ### ### =============================================== v_p ================================================ ### ### ('shay', 6597, 0) ('climax', 3967, 1) ('meyer', 4707, 2) ('why', 3556, 3) ### ### ('classification', 4863, 4) ('##rra', 4, 5) ('mall', 12490, 6) ('ui', 7476, 7) ('##et', 24, 8) ### ### ('notation', 47, 9) ('##lie', 25519, 10) ('wheel', 44, 11) ('0', 4560, 12) ('truck', 720, 13) ### ### ('fair', 10847, 14) ('##ler', 19546, 15) ('arrangement', 17849, 16) ('trailing', 15372, 17) ### ### ('##te', 19968, 18) ('locomotive', 9, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##rra', 4, 5) ('##tt', 1, 23) ('ga', 5, 24) ('locomotive', 9, 19) ('locomotives', 10, 26) ### ### ('wheels', 2, 41) ('##et', 24, 8) ('notation', 47, 9) ('wheel', 44, 11) ('+', 11, 51) ### ### ('steam', 71, 20) ('.', 3, 90) ('truck', 720, 13) ('engine', 22, 55) ('type', 245, 30) ### ### ('shay', 6597, 0) ('climax', 3967, 1) ('boiler', 8, 258) ('why', 3556, 3) ('meyer', 4707, 2) ### ############################################################################################################ [2023-11-25 01:24:16,286][root][INFO] - Epoch: 1: Step: 7801/19530, loss[v]=0.458947, lr=0.000008, acc@1[1]=939.0/1024=0.9169921875, acc@1[2]=965.0/1024=0.9423828125 [2023-11-25 01:26:07,190][root][INFO] - Train batch 7900 [2023-11-25 01:26:07,191][root][INFO] - Avg. loss per last 100 batches: 0.299910 [2023-11-25 01:26:07,932][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.05 | max: 3.75 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.04 | max: 5.01 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: on 6 july 2011, it was announced that " stop the clocks " would finally be released on ### ### noel gallagher's debut solo album, " noel gallagher's high flying birds ". ### ### [P_TEXT]: stop the clocks ". when the release of " stop the clocks " was first announced in july ### ### 2006, speculation was rife that the unreleased song of the same name would be included on the ### ### record as a bonus track. however, noel gallagher told fans at a q & a session that the song was ### ### considered for inclusion, but they weren't happy with any of the many versions they have recorded. ### ### it would later be released on noel gallagher's high flying birds'2011 debut album. gallagher also ### ### confirmed that the title was chosen to sum up what was ### ### =============================================== v_q ================================================ ### ### ('noel', 0, 3) ('clocks', 1, 0) ('stop', 2, 1) ('flying', 3, 8) ('gallagher', 4, 4) ('clock', 5, 5) ### ### ('.', 6, 171) ('birds', 7, 2) ('high', 8, 12) ('album', 9, 23) ('2011', 10, 11) ('bird', 11, 21) ### ### ('solo', 12, 5594) ('stopping', 13, 24) ('stops', 14, 18) ('fly', 15, 99) ('debut', 16, 14) ### ### ('ireland', 17, 1240) ('frontman', 18, 1495) ('music', 19, 51) ### ### =============================================== v_p ================================================ ### ### ('clocks', 1, 0) ('stop', 2, 1) ('birds', 7, 2) ('noel', 0, 3) ('gallagher', 4, 4) ('clock', 5, 5) ### ### ('speculation', 7834, 6) ('##fe', 22555, 7) ('flying', 3, 8) ('bonus', 5703, 9) ### ### ('unreleased', 8762, 10) ('2011', 10, 11) ('high', 8, 12) ('inclusion', 5635, 13) ('debut', 16, 14) ### ### ('2006', 8579, 15) ('confirmed', 239, 16) ('ri', 15905, 17) ('stops', 14, 18) ('q', 3946, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('clocks', 1, 0) ('stop', 2, 1) ('noel', 0, 3) ('gallagher', 4, 4) ('flying', 3, 8) ('birds', 7, 2) ### ### ('clock', 5, 5) ('high', 8, 12) ('2011', 10, 11) ('album', 9, 23) ('bird', 11, 21) ### ### ('debut', 16, 14) ('stops', 14, 18) ('stopping', 13, 24) ('song', 24, 20) ('stopped', 20, 27) ### ### ('.', 6, 171) ('confirmed', 239, 16) ('ticking', 37, 32) ('recordings', 34, 34) ### ############################################################################################################ [2023-11-25 01:26:08,434][root][INFO] - Epoch: 1: Step: 7901/19530, loss[v]=0.373510, lr=0.000008, acc@1[1]=946.5/1024=0.92431640625, acc@1[2]=978.0/1024=0.955078125 [2023-11-25 01:27:59,043][root][INFO] - Train batch 8000 [2023-11-25 01:27:59,044][root][INFO] - Avg. loss per last 100 batches: 0.280232 [2023-11-25 01:27:59,788][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.05 | max: 3.68 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.04 | max: 4.69 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: myles francis goodwyn ( born miles francis goodwin, ) is a canadian record producer, ### ### guitarist, lead vocalist, main songwriter, and founding member of the veteran canadian hard rock ### ### band april wine. ### ### [P_TEXT]: fast train " fast train " is the first single by the canadian rock band april wine, from ### ### their 1971 self - titled debut album. it was a top 40 hit in canada, peaking at # 38 on the " rpm " ### ### 100 singles chart. it also peaked at # 23 on the canadian chum singles chart, on june 26, 1971. the ### ### success of this single established myles goodwyn as the band's main songwriter, and made it ### ### possible for april wine to record a second album. " fast train " appears on ### ### =============================================== v_q ================================================ ### ### ('myles', 0, 2) ('goodwin', 1, 17450) ('##wyn', 2, 5) ('wine', 3, 4) ('miles', 4, 12477) ### ### ('francis', 5, 19566) ('.', 6, 100) ('good', 7, 9) ('april', 8, 7) ('canadian', 9, 16) ### ### ('canada', 10, 13) ('hard', 11, 78) ('rock', 12, 18) ('band', 13, 19) ('veteran', 14, 14226) ### ### ('producer', 15, 12271) ('album', 16, 12) ('producers', 17, 4293) ('guitarist', 18, 980) ### ### ('frontman', 19, 1007) ### ### =============================================== v_p ================================================ ### ### ('fast', 3054, 0) ('rpm', 10500, 1) ('myles', 0, 2) ('chu', 17154, 3) ('wine', 3, 4) ### ### ('##wyn', 2, 5) ('train', 2765, 6) ('april', 8, 7) ('1971', 9148, 8) ('good', 7, 9) ### ### ('self', 15268, 10) ('38', 3659, 11) ('album', 16, 12) ('canada', 10, 13) ('debut', 3194, 14) ### ### ('100', 8377, 15) ('canadian', 9, 16) ('trains', 21730, 17) ('rock', 12, 18) ('band', 13, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('myles', 0, 2) ('##wyn', 2, 5) ('wine', 3, 4) ('april', 8, 7) ('good', 7, 9) ('canadian', 9, 16) ### ### ('canada', 10, 13) ('rock', 12, 18) ('band', 13, 19) ('album', 16, 12) ('.', 6, 100) ### ### ('bands', 20, 27) ('hard', 11, 78) ('fast', 3054, 0) ('june', 21, 38) ('train', 2765, 6) ### ### ('march', 29, 30) ('may', 22, 40) ('songwriter', 33, 36) ('##ם', 55, 35) ### ############################################################################################################ [2023-11-25 01:28:00,292][root][INFO] - Epoch: 1: Step: 8001/19530, loss[v]=0.390880, lr=0.000008, acc@1[1]=943.5/1024=0.92138671875, acc@1[2]=967.0/1024=0.9443359375 [2023-11-25 01:29:51,314][root][INFO] - Train batch 8100 [2023-11-25 01:29:51,314][root][INFO] - Avg. loss per last 100 batches: 0.269416 [2023-11-25 01:29:52,063][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.05 | max: 3.70 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.04 | max: 4.89 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 77.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: one of the greatest names in brazilian popular music, noel gave a new twist to samba, ### ### combining its afro - brazilian roots with a more urban, witty language and making it a vehicle for ### ### ironic social commentary. ### ### [P_TEXT]: samba in the partnership between ismael silva and noel rosa in " para me livrar do mal ". ### ### with the growing demand for new sambas by the singers, the practice of buying and selling ### ### compositions has also become common. this transaction usually took place in two different ways : ### ### the author negotiated only the sale of the samba recording – that is, he remained as the author of ### ### the composition, but he would not receive any part of the gains obtained from the sales of the ### ### records, which were divided between the buyer and the record label – or the entire composition ### ### =============================================== v_q ================================================ ### ### ('noel', 0, 2) ('samba', 1, 0) ('brazilian', 2, 104) ('.', 3, 190) ('brazil', 4, 63) ### ### ('ironic', 5, 2338) ('popular', 6, 799) ('afro', 7, 17615) ('twist', 8, 3726) ('music', 9, 45) ### ### ('commentary', 10, 15885) ('roots', 11, 8193) ('portuguese', 12, 341) ('social', 13, 12880) ### ### ('language', 14, 13469) ('brasil', 15, 3904) ('popularity', 16, 757) ('musical', 17, 86) ### ### ('urban', 18, 3759) ('africa', 19, 2310) ### ### =============================================== v_p ================================================ ### ### ('samba', 1, 0) ('rosa', 8488, 1) ('noel', 0, 2) ('mal', 9674, 3) ('silva', 470, 4) ### ### ('singers', 52, 5) ('para', 1694, 6) ('liv', 24376, 7) ('buying', 949, 8) ('##rar', 27411, 9) ### ### ('gains', 24786, 10) ('transaction', 16930, 11) ('composition', 925, 12) ('##el', 20213, 13) ### ### ('partnership', 10014, 14) ('compositions', 5623, 15) ('selling', 5349, 16) ('me', 2474, 17) ### ### ('recording', 1778, 18) ('sale', 1088, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('samba', 1, 0) ('noel', 0, 2) ('singers', 52, 5) ('brazil', 4, 63) ('music', 9, 45) ### ### ('brazilian', 2, 104) ('silva', 470, 4) ('.', 3, 190) ('buying', 949, 8) ('para', 1694, 6) ### ### ('composition', 925, 12) ('musical', 17, 86) ('singing', 75, 37) ('album', 38, 46) ### ### ('recordings', 143, 33) ('sale', 1088, 19) ('popular', 6, 799) ('rosa', 8488, 1) ### ### ('recording', 1778, 18) ('portuguese', 12, 341) ### ############################################################################################################ [2023-11-25 01:29:52,564][root][INFO] - Epoch: 1: Step: 8101/19530, loss[v]=0.163162, lr=0.000008, acc@1[1]=984.0/1024=0.9609375, acc@1[2]=1009.0/1024=0.9853515625 [2023-11-25 01:31:43,408][root][INFO] - Train batch 8200 [2023-11-25 01:31:43,409][root][INFO] - Avg. loss per last 100 batches: 0.293553 [2023-11-25 01:31:44,142][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.05 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.04 | max: 4.71 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: portuguese traders and jesuit missionaries under the " padroado " system were active in ### ### both vietnamese realms of " đang trong " ( cochinchina or quinan ) and " đang ngoai " ( tonkin ) ### ### [P_TEXT]: tonkin tonkin, also spelled " tongkin ", " tonquin " or " tongking ", is an exonym ### ### referring to the northern region of vietnam. during the 17th and 18th centuries, this term referred ### ### to the domain đang ngoai under trinh lords'control, including both the northern and thanh - nghe ### ### regions, north of the gianh river. from 1884 to early 1945, this term was used for the french ### ### protectorate of tonkin, composed of only the northern region. " tonkin " is an alternative spelling ### ### of 東 京 " tongking " ( compare nanking spelt as " nankin " ), ### ### =============================================== v_q ================================================ ### ### ('##ado', 0, 18366) ('.', 1, 84) ('pad', 2, 4270) ('tr', 3, 13384) ('vietnamese', 4, 18) ### ### ('##nan', 5, 19844) ('portuguese', 6, 2744) ('vietnam', 7, 3) ('ton', 8, 0) ('##chin', 9, 15209) ### ### ('##ang', 10, 39) ('##ong', 11, 9479) ('missionaries', 12, 6925) ('qui', 13, 9593) ('đ', 14, 37) ### ### ('##ro', 15, 22759) ('cochin', 16, 20087) ('system', 17, 6247) ('traders', 18, 3061) ### ### ('##ai', 19, 31) ### ### =============================================== v_p ================================================ ### ### ('ton', 8, 0) ('domain', 455, 1) ('protectorate', 3055, 2) ('vietnam', 7, 3) ('##kin', 24, 4) ### ### ('lords', 953, 5) ('tong', 628, 6) ('##he', 10108, 7) ('##king', 9870, 8) ('gia', 17063, 9) ### ### ('nan', 1848, 10) ('than', 8559, 11) ('ngo', 23, 12) ('spelling', 2036, 13) ('regions', 487, 14) ### ### ('##quin', 27940, 15) ('##nh', 16637, 16) ('northern', 1982, 17) ('vietnamese', 4, 18) ### ### ('##elt', 26750, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ton', 8, 0) ('vietnam', 7, 3) ('vietnamese', 4, 18) ('##kin', 24, 4) ('ngo', 23, 12) ### ### ('##ang', 10, 39) ('đ', 14, 37) ('##ai', 19, 31) ('.', 1, 84) ('domain', 455, 1) ### ### ('french', 160, 20) ('tong', 628, 6) ('lords', 953, 5) ('regions', 487, 14) ('river', 296, 25) ### ### ('centuries', 105, 38) ('protectorate', 3055, 2) ('region', 216, 33) ('nan', 1848, 10) ### ### ('ngos', 129, 41) ### ############################################################################################################ [2023-11-25 01:31:44,643][root][INFO] - Epoch: 1: Step: 8201/19530, loss[v]=0.216998, lr=0.000008, acc@1[1]=972.5/1024=0.94970703125, acc@1[2]=995.5/1024=0.97216796875 [2023-11-25 01:33:35,402][root][INFO] - Train batch 8300 [2023-11-25 01:33:35,403][root][INFO] - Avg. loss per last 100 batches: 0.285455 [2023-11-25 01:33:36,130][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.05 | max: 3.75 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.04 | max: 4.65 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.5/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the fairbairn – sykes fighting knife is a double - edged fighting knife resembling a ### ### dagger or poignard with a foil grip developed by william ewart fairbairn and eric anthony sykes in ### ### shanghai based on ideas which the two men had before world war ii while serving on the shanghai ### ### municipal police in china. ### ### [P_TEXT]: v - 42 stiletto the v - 42 stiletto was a stiletto and fighting knife issued during world ### ### war ii to the first special service force ( 1st ssf or fssf, also known as devil's brigade ), a ### ### joint american / canadian commando unit. based on the fairbairn – sykes commando knife designed by ### ### william e. fairbairn and eric a. sykes, the " fighting commando knife, type v - 42 " used a narrow ### ### - profile, double - edged stiletto blade made of high carbon steel. the v - 42 was primarily ### ### designed by officers of the fssf, including ### ### =============================================== v_q ================================================ ### ### ('knife', 0, 8) ('sykes', 1, 1) ('fair', 2, 12) ('anthony', 3, 15626) ('##bai', 4, 5) ### ### ('dagger', 5, 80) ('shanghai', 6, 6661) ('police', 7, 822) ('.', 8, 193) ('##wart', 9, 28203) ### ### ('eric', 10, 16) ('foil', 11, 8286) ('##ign', 12, 23116) ('municipal', 13, 2126) ('##rn', 14, 14) ### ### ('knives', 15, 43) ('china', 16, 2609) ('fighting', 17, 15) ('##ard', 18, 13994) ('sword', 19, 115) ### ### =============================================== v_p ================================================ ### ### ('commando', 914, 0) ('sykes', 1, 1) ('##etto', 24670, 2) ('stil', 16694, 3) ('carbon', 3805, 4) ### ### ('##bai', 4, 5) ('steel', 311, 6) ('ss', 5352, 7) ('knife', 0, 8) ('devil', 4773, 9) ### ### ('##ss', 17449, 10) ('42', 13849, 11) ('fair', 2, 12) ('profile', 3109, 13) ('##rn', 14, 14) ### ### ('fighting', 17, 15) ('eric', 10, 16) ('double', 20, 17) ('type', 2215, 18) ('v', 12884, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('sykes', 1, 1) ('knife', 0, 8) ('fair', 2, 12) ('##bai', 4, 5) ('eric', 10, 16) ('##rn', 14, 14) ### ### ('fighting', 17, 15) ('double', 20, 17) ('blade', 21, 28) ('william', 23, 36) ('knives', 15, 43) ### ### ('edged', 25, 32) ('war', 29, 41) ('dagger', 5, 80) ('steel', 311, 6) ('commando', 914, 0) ### ### ('e', 28, 49) ('.', 8, 193) ('fight', 26, 55) ('officers', 304, 29) ### ############################################################################################################ [2023-11-25 01:33:36,634][root][INFO] - Epoch: 1: Step: 8301/19530, loss[v]=0.258960, lr=0.000009, acc@1[1]=960.5/1024=0.93798828125, acc@1[2]=982.5/1024=0.95947265625 [2023-11-25 01:35:27,640][root][INFO] - Train batch 8400 [2023-11-25 01:35:27,641][root][INFO] - Avg. loss per last 100 batches: 0.276154 [2023-11-25 01:35:28,384][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.05 | max: 3.76 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.04 | max: 4.66 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he completed the piano concerto in d major for the left hand first. ### ### [P_TEXT]: jacqueline blancard " etudes ". in 1938, she became the first pianist to make a recording ### ### of ravel's concerto for the left hand, under charles munch. she also performed three piano ### ### concertos by mozart in new york. ### ### =============================================== v_q ================================================ ### ### ('concerto', 0, 7) ('hand', 1, 8) ('.', 2, 2187) ('piano', 3, 9) ('major', 4, 19277) ### ### ('d', 5, 18123) ('left', 6, 15) ('symphony', 7, 73) ('minor', 8, 14681) ('sonata', 9, 34) ### ### ('hands', 10, 18) ('soloist', 11, 32) ('fingers', 12, 31) ('initial', 13, 1386) ### ### ('pupils', 14, 2417) ('pianist', 15, 11) ('orchestra', 16, 404) ('mozart', 17, 3) ### ### ('ballet', 18, 1183) ('tchaikovsky', 19, 2405) ### ### =============================================== v_p ================================================ ### ### ('etudes', 2188, 0) ('blanca', 14363, 1) ('##nch', 18390, 2) ('mozart', 17, 3) ('rave', 16748, 4) ### ### ('jacqueline', 20700, 5) ('##rd', 22207, 6) ('concerto', 0, 7) ('hand', 1, 8) ('piano', 3, 9) ### ### ('charles', 4928, 10) ('pianist', 15, 11) ('mu', 3865, 12) ('1938', 15188, 13) ('##l', 23372, 14) ### ### ('left', 6, 15) ('she', 8615, 16) ('york', 14055, 17) ('hands', 10, 18) ('recording', 10526, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('concerto', 0, 7) ('hand', 1, 8) ('piano', 3, 9) ('mozart', 17, 3) ('left', 6, 15) ### ### ('pianist', 15, 11) ('hands', 10, 18) ('etudes', 2188, 0) ('sonata', 9, 34) ('soloist', 11, 32) ### ### ('fingers', 12, 31) ('symphony', 7, 73) ('leaves', 26, 25) ('beethoven', 22, 38) ### ### ('blanca', 14363, 1) ('charles', 4928, 10) ('mu', 3865, 12) ('.', 2, 2187) ('musicians', 86, 26) ### ### ('##nch', 18390, 2) ### ############################################################################################################ [2023-11-25 01:35:28,888][root][INFO] - Epoch: 1: Step: 8401/19530, loss[v]=0.402179, lr=0.000009, acc@1[1]=956.5/1024=0.93408203125, acc@1[2]=977.5/1024=0.95458984375 [2023-11-25 01:37:19,567][root][INFO] - Train batch 8500 [2023-11-25 01:37:19,567][root][INFO] - Avg. loss per last 100 batches: 0.267466 [2023-11-25 01:37:20,295][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.05 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 4.64 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: via abscisic aldehyde. ### ### [P_TEXT]: abscisic aldehyde abscisic aldehyde is an intermediate in the biosynthesis of the plant ### ### hormone abscisic acid. it is produced by the dehydrodgenation of xanthoxin by xanthtoxin ### ### dehydrogenases, which is an nad + dependent short - chain dehydrogenas, followed by selective ### ### oxidation by abscisic aldehyde oxygenase. ### ### =============================================== v_q ================================================ ### ### ('abs', 0, 2) ('##cis', 1, 0) ('.', 2, 1558) ('##deh', 3, 6) ('al', 4, 34) ('flavor', 5, 6369) ### ### ('chemical', 6, 4031) ('absently', 7, 557) ('via', 8, 13274) ('##ه', 9, 44) ('##ic', 10, 36) ### ### ('compounds', 11, 2355) ('drug', 12, 8595) ('derived', 13, 1779) ('realizing', 14, 205) ### ### ('oil', 15, 906) ('substances', 16, 12476) ('manpower', 17, 50) ('##₂', 18, 48) ### ### ('molecules', 19, 2032) ### ### =============================================== v_p ================================================ ### ### ('##cis', 1, 0) ('##nation', 2181, 1) ('abs', 0, 2) ('##xin', 479, 3) ('oxygen', 24, 4) ### ### ('hormone', 1981, 5) ('##deh', 3, 6) ('##dge', 22328, 7) ('dependent', 7672, 8) ('plant', 1117, 9) ### ### ('selective', 4640, 10) ('##genase', 531, 11) ('oxidation', 1662, 12) ('bio', 1806, 13) ### ### ('##ase', 3211, 14) ('##ant', 18538, 15) ('nad', 8222, 16) ('chain', 149, 17) ('##as', 11083, 18) ### ### ('short', 18364, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##cis', 1, 0) ('abs', 0, 2) ('##deh', 3, 6) ('oxygen', 24, 4) ('al', 4, 34) ('##xin', 479, 3) ### ### ('chain', 149, 17) ('##genase', 531, 11) ('##ic', 10, 36) ('##nation', 2181, 1) ('plant', 1117, 9) ### ### ('hormone', 1981, 5) ('oxidation', 1662, 12) ('bio', 1806, 13) ('##yde', 486, 24) ('##ه', 9, 44) ### ### ('selective', 4640, 10) ('##ase', 3211, 14) ('.', 2, 1558) ('dependent', 7672, 8) ### ############################################################################################################ [2023-11-25 01:37:20,800][root][INFO] - Epoch: 1: Step: 8501/19530, loss[v]=0.164298, lr=0.000009, acc@1[1]=981.0/1024=0.9580078125, acc@1[2]=1010.5/1024=0.98681640625 [2023-11-25 01:39:11,627][root][INFO] - Train batch 8600 [2023-11-25 01:39:11,628][root][INFO] - Avg. loss per last 100 batches: 0.260547 [2023-11-25 01:39:12,366][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.05 | max: 3.69 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.04 | max: 4.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the first major reform came in 1917 when outrage over the election - year murder of a ### ### police officer led to the shrinking of the city council from two houses to just one. ### ### [P_TEXT]: mark squilla mark f. squilla is a democratic politician and member of the city council of ### ### philadelphia, pennsylvania. squilla is active in philadelphia democratic politics, having held a ### ### variety of positions. in the 2004 democratic primary, he unsuccessfully challenged incumbent state ### ### representative william keller for a seat in the state house. in 2011, he ran for city council, ### ### seeking to succeed retiring democrat frank dicicco in the first district. he ultimately came out on ### ### - top of a four - man democratic primary field, and faced no opposition ### ### =============================================== v_q ================================================ ### ### ('.', 0, 145) ('council', 1, 6) ('police', 2, 2238) ('reform', 3, 1511) ('murder', 4, 4774) ### ### ('1917', 5, 12687) ('city', 6, 27) ('outrage', 7, 12110) ('officer', 8, 10884) ('election', 9, 62) ### ### ('house', 10, 13) ('killing', 11, 5447) ('terror', 12, 1209) ('1918', 13, 12454) ('cities', 14, 55) ### ### ('assassination', 15, 9813) ('massacre', 16, 15687) ('kill', 17, 6184) ('mayor', 18, 1095) ### ### ('anger', 19, 735) ### ### =============================================== v_p ================================================ ### ### ('keller', 17065, 0) ('##cic', 27737, 1) ('sq', 21492, 2) ('philadelphia', 1118, 3) ### ### ('##uil', 25893, 4) ('retiring', 10970, 5) ('council', 1, 6) ('mark', 8522, 7) ### ### ('pennsylvania', 6565, 8) ('##la', 22803, 9) ('field', 2072, 10) ('##co', 21903, 11) ### ### ('opposition', 1324, 12) ('house', 10, 13) ('democratic', 2058, 14) ('william', 741, 15) ### ### ('2004', 15712, 16) ('primary', 5341, 17) ('state', 3199, 18) ('challenged', 16473, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('council', 1, 6) ('city', 6, 27) ('house', 10, 13) ('.', 0, 145) ('election', 9, 62) ### ### ('philadelphia', 1118, 3) ('cities', 14, 55) ('william', 741, 15) ('politician', 66, 40) ### ### ('houses', 38, 49) ('opposition', 1324, 12) ('district', 295, 34) ('field', 2072, 10) ### ### ('democratic', 2058, 14) ('politics', 1513, 20) ('frank', 2075, 22) ('police', 2, 2238) ### ### ('state', 3199, 18) ('reform', 3, 1511) ('pennsylvania', 6565, 8) ### ############################################################################################################ [2023-11-25 01:39:12,869][root][INFO] - Epoch: 1: Step: 8601/19530, loss[v]=0.374924, lr=0.000009, acc@1[1]=940.0/1024=0.91796875, acc@1[2]=969.0/1024=0.9462890625 [2023-11-25 01:41:03,991][root][INFO] - Train batch 8700 [2023-11-25 01:41:03,992][root][INFO] - Avg. loss per last 100 batches: 0.256547 [2023-11-25 01:41:04,731][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.05 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.04 | max: 4.82 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: matten bei interlaken lies in the bernese oberland in the alps, in the bodeli ( the ### ### tongue of land between of lake thun and lake brienz ). ### ### [P_TEXT]: bodeli through the bodeli. on the bodeli are situated the villages and towns of ### ### unterseen, interlaken and matten, which form a closed settlement area, and at the southern border ### ### are the villages of wilderswil and bonigen. from the south a hill range, the ruuge, rises up. ### ### between 1870 and 1874 the bodeli railway ( bodelibahn ) was constructed to link the steamer quay at ### ### darligen on lake thun with the quay at bonigen on lake brienz. the railway company cunningly ### ### arranged for the route to cross the aare ### ### =============================================== v_q ================================================ ### ### ('matt', 0, 9) ('##del', 1, 0) ('inter', 2, 15) ('brien', 3, 1) ('th', 4, 17) ('alps', 5, 12464) ### ### ('##un', 6, 18) ('##en', 7, 24) ('bern', 8, 13810) ('bei', 9, 7963) ('.', 10, 645) ### ### ('##lake', 11, 5) ('bo', 12, 7) ('lake', 13, 13) ('obe', 14, 23700) ('##ʂ', 15, 736) ### ### ('##i', 16, 55) ('german', 17, 7838) ('##rland', 18, 27502) ('##ese', 19, 23260) ### ### =============================================== v_p ================================================ ### ### ('##del', 1, 0) ('brien', 3, 1) ('##igen', 13911, 2) ('bon', 14879, 3) ('wilder', 14062, 4) ### ### ('##lake', 11, 5) ('##uge', 20400, 6) ('bo', 12, 7) ('aa', 7458, 8) ('matt', 0, 9) ### ### ('##een', 16173, 10) ('##li', 1483, 11) ('##ters', 3706, 12) ('lake', 13, 13) ### ### ('steamer', 14680, 14) ('inter', 2, 15) ('quay', 21468, 16) ('th', 4, 17) ('##un', 6, 18) ### ### ('dar', 16456, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##del', 1, 0) ('brien', 3, 1) ('matt', 0, 9) ('inter', 2, 15) ('th', 4, 17) ('##un', 6, 18) ### ### ('##lake', 11, 5) ('bo', 12, 7) ('##en', 7, 24) ('lake', 13, 13) ('##z', 26, 35) ('range', 96, 22) ### ### ('##i', 16, 55) ('##li', 1483, 11) ('village', 47, 43) ('.', 10, 645) ('##gen', 829, 26) ### ### ('##ters', 3706, 12) ('##n', 21, 107) ('aa', 7458, 8) ### ############################################################################################################ [2023-11-25 01:41:05,234][root][INFO] - Epoch: 1: Step: 8701/19530, loss[v]=0.147968, lr=0.000009, acc@1[1]=979.0/1024=0.9560546875, acc@1[2]=1009.0/1024=0.9853515625 [2023-11-25 01:42:55,746][root][INFO] - Train batch 8800 [2023-11-25 01:42:55,746][root][INFO] - Avg. loss per last 100 batches: 0.259708 [2023-11-25 01:42:56,479][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.04 | max: 3.60 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.04 | max: 4.88 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.2/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 73.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: after a few years of civil wars during the reigns of fruela ii, alfonso froilaz and ### ### alfonso iv, ramiro ii ( 931 – 951 ) assumed the throne and brought stability to the kingdom. ### ### [P_TEXT]: county of monzon the county of monzon was a marcher county of the kingdom of leon in the ### ### tenth and eleventh centuries, during a period of renewed external threat ( the caliphate of cordoba ### ### ) and disintegration of royal authority. the county was created by ramiro ii for ansur fernandez in ### ### 943 and was ruled by his descendants, the banu ansur ( " banu anshur " ) or ansurez, for decades. ### ### the seat of the county was initially at the castle of curiel and later at " monteson " ; to its ### ### east the river pisuerga served as ### ### =============================================== v_q ================================================ ### ### ('.', 0, 213) ('##uel', 1, 26683) ('##iro', 2, 10) ('##oi', 3, 28092) ('93', 4, 168) ### ### ('##la', 5, 14973) ('fr', 6, 22025) ('ii', 7, 31) ('ram', 8, 12) ('spanish', 9, 4064) ### ### ('castile', 10, 3897) ('spain', 11, 472) ('reign', 12, 1662) ('stability', 13, 1944) ### ### ('alfonso', 14, 17066) ('iv', 15, 1890) ('95', 16, 128) ('madrid', 17, 563) ('##ʂ', 18, 639) ### ### ('kingdom', 19, 22) ### ### =============================================== v_p ================================================ ### ### ('##sure', 28097, 0) ('##zon', 11121, 1) ('##rie', 21526, 2) ('##sur', 23943, 3) ### ### ('##shu', 26638, 4) ('castle', 69, 5) ('leon', 593, 6) ('ban', 20804, 7) ('monte', 615, 8) ### ### ('##son', 18682, 9) ('##iro', 2, 10) ('cordoba', 5752, 11) ('ram', 8, 12) ('caliphate', 160, 13) ### ### ('mon', 21313, 14) ('marche', 25874, 15) ('fernandez', 1826, 16) ('##sin', 11413, 17) ### ### ('county', 2088, 18) ('threat', 6660, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##iro', 2, 10) ('ram', 8, 12) ('ii', 7, 31) ('kingdom', 19, 22) ('castle', 69, 5) ### ### ('centuries', 35, 23) ('caliphate', 160, 13) ('leon', 593, 6) ('monte', 615, 8) ('.', 0, 213) ### ### ('##z', 42, 41) ('94', 230, 27) ('93', 4, 168) ('fernandez', 1826, 16) ('river', 599, 30) ### ### ('ruled', 332, 35) ('county', 2088, 18) ('##zon', 11121, 1) ('cordoba', 5752, 11) ### ### ('royal', 822, 36) ### ############################################################################################################ [2023-11-25 01:42:56,983][root][INFO] - Epoch: 1: Step: 8801/19530, loss[v]=0.411974, lr=0.000009, acc@1[1]=932.0/1024=0.91015625, acc@1[2]=959.5/1024=0.93701171875 [2023-11-25 01:44:47,561][root][INFO] - Train batch 8900 [2023-11-25 01:44:47,562][root][INFO] - Avg. loss per last 100 batches: 0.266808 [2023-11-25 01:44:48,308][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.05 | max: 3.76 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 4.69 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 29.0/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: independiente won three league titles at the 60s decade ; 1960, 1963 and 1967. ### ### [P_TEXT]: 1967 argentine primera division the 1967 primera division season was the 76th season of ### ### top - flight football in argentina. it was the first season in which the metropolitano and nacional ### ### championships format was used. estudiantes ( lp ) ( 2nd title ) won the metropolitano while ### ### independiente won the nacional tournament ( 8th title ). at the end of the reclasificatorio, union ### ### de santa fe and deportivo espanol were relegated. the " campeonato metropolitano " was the first ### ### championship of the season. it was contested between twenty - two teams directed affiliated with ### ### the argentine football association from march 3 to august 6. the format championship was in two ### ### rounds. the first round saw the 22 ### ### =============================================== v_q ================================================ ### ### ('##iente', 0, 0) ('ind', 1, 19) ('1963', 2, 80) ('1967', 3, 7) ('.', 4, 72) ('##ep', 5, 24) ### ### ('60s', 6, 10396) ('##end', 7, 44) ('1960', 8, 2380) ('league', 9, 1188) ('1968', 10, 46) ### ### ('decade', 11, 3167) ('67', 12, 76) ('1966', 13, 53) ('70s', 14, 10592) ('1964', 15, 198) ### ### ('1962', 16, 781) ('club', 17, 323) ('atletico', 18, 1066) ('championship', 19, 36) ### ### =============================================== v_p ================================================ ### ### ('##iente', 0, 0) ('deportivo', 48, 1) ('metropolitan', 5159, 2) ('nacional', 63, 3) ### ### ('##antes', 16955, 4) ('lp', 2783, 5) ('argentine', 38, 6) ('1967', 3, 7) ('primera', 77, 8) ### ### ('campeonato', 449, 9) ('##ific', 29041, 10) ('tournament', 83, 11) ('argentina', 21, 12) ### ### ('union', 2372, 13) ('association', 2114, 14) ('##las', 18821, 15) ('championships', 114, 16) ### ### ('##pan', 22221, 17) ('format', 3578, 18) ('ind', 1, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##iente', 0, 0) ('1967', 3, 7) ('ind', 1, 19) ('##ep', 5, 24) ('##end', 7, 44) ### ### ('deportivo', 48, 1) ('argentina', 21, 12) ('argentine', 38, 6) ('1968', 10, 46) ('season', 20, 23) ### ### ('nacional', 63, 3) ('division', 23, 26) ('championship', 19, 36) ('relegated', 40, 20) ### ### ('.', 4, 72) ('primera', 77, 8) ('1963', 2, 80) ('tournament', 83, 11) ('football', 26, 43) ### ### ('1966', 13, 53) ### ############################################################################################################ [2023-11-25 01:44:48,809][root][INFO] - Epoch: 1: Step: 8901/19530, loss[v]=0.119682, lr=0.000009, acc@1[1]=983.5/1024=0.96044921875, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 01:46:39,525][root][INFO] - Train batch 9000 [2023-11-25 01:46:39,526][root][INFO] - Avg. loss per last 100 batches: 0.275937 [2023-11-25 01:46:40,250][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.04 | max: 3.55 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.04 | max: 4.88 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: fez ", " the talos principle ", and " watch dogs ", have incorporated qr codes as story ### ### and gameplay elements. ### ### [P_TEXT]: the talos principle as well as signs from previous adventurers in the world in the form ### ### of qr code s left as graffiti on various walls, and holograms that once collected play audio ### ### recordings. the player character, an unnamed android, is awoken in a serene environment. a ### ### disembodied entity named elohim instructs the android to explore the worlds he has created for it, ### ### and to solve the various puzzles to collect sigils, but warns it not to climb a tower at the centre ### ### of these worlds. as the android progresses, it becomes evident that these worlds ### ### =============================================== v_q ================================================ ### ### ('q', 0, 8) ('watch', 1, 5891) ('tal', 2, 0) ('##z', 3, 25591) ('fe', 4, 13139) ('dogs', 5, 5163) ### ### ('principle', 6, 4) ('game', 7, 986) ('##os', 8, 10) ('gameplay', 9, 520) ('.', 10, 227) ### ### ('dog', 11, 2444) ('code', 12, 13) ('##з', 13, 7510) ('##r', 14, 62) ('story', 15, 10262) ### ### ('franchise', 16, 5906) ('codes', 17, 53) ('characters', 18, 43) ('##ʂ', 19, 1700) ### ### =============================================== v_p ================================================ ### ### ('tal', 2, 0) ('android', 603, 1) ('##logram', 8511, 2) ('unnamed', 14434, 3) ('principle', 6, 4) ### ### ('graffiti', 6986, 5) ('worlds', 1700, 6) ('awoke', 3085, 7) ('q', 0, 8) ('signs', 21955, 9) ### ### ('##os', 8, 10) ('##gil', 23664, 11) ('serene', 18017, 12) ('code', 12, 13) ### ### ('recordings', 14394, 14) ('adventurer', 18537, 15) ('puzzles', 1244, 16) ('tower', 1125, 17) ### ### ('entity', 12356, 18) ('##oh', 17634, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('tal', 2, 0) ('q', 0, 8) ('principle', 6, 4) ('##os', 8, 10) ('code', 12, 13) ### ### ('character', 23, 25) ('principles', 24, 40) ('characters', 18, 43) ('player', 79, 20) ### ### ('android', 603, 1) ('codes', 17, 53) ('##r', 14, 62) ('.', 10, 227) ('worlds', 1700, 6) ### ### ('gameplay', 9, 520) ('tower', 1125, 17) ('puzzles', 1244, 16) ('awoke', 3085, 7) ('game', 7, 986) ### ### ('players', 398, 34) ### ############################################################################################################ [2023-11-25 01:46:40,755][root][INFO] - Epoch: 1: Step: 9001/19530, loss[v]=0.194799, lr=0.000009, acc@1[1]=972.5/1024=0.94970703125, acc@1[2]=1008.5/1024=0.98486328125 [2023-11-25 01:48:30,800][root][INFO] - Train batch 9100 [2023-11-25 01:48:30,801][root][INFO] - Avg. loss per last 100 batches: 0.251219 [2023-11-25 01:48:31,529][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.05 | max: 3.83 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 4.81 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: brian herbert and anderson have also published " hunters of dune " ( 2006 ) and " ### ### sandworms of dune " ( 2007 ), sequels to frank herbert's final novel " " ( 1985 ) which complete ### ### the chronological progression of his original series and wrap up storylines that began with his " ### ### heretics of dune " ( 1984 ). ### ### [P_TEXT]: brian herbert " dune " universe based on his father's notes, though, according to the ### ### younger herbert, there are no immediate plans to publish it. herbert is known for his ### ### collaborations with author kevin j. anderson, with whom he has written multiple prequels to his ### ### father's landmark 1965 science fiction novel, " dune ", all of which have made the " new york times ### ### " best seller list. the duo began with the trilogies " prelude to dune " ( 1999 – 2001 ) and " ### ### legends of dune " ( 2002 – 2004 ). brian and anderson next published " hunters of dune " ( 2006 ) ### ### and " sandworms of dune " ( 2007 ), two sequels to frank herbert's ### ### =============================================== v_q ================================================ ### ### ('##tics', 0, 4067) ('##worm', 1, 2) ('herbert', 2, 0) ('hunters', 3, 5) ('dune', 4, 3) ### ### ('anderson', 5, 7) ('brian', 6, 1) ('.', 7, 270) ('frank', 8, 15) ('hunter', 9, 26) ### ### ('sand', 10, 11) ('here', 11, 412) ('1985', 12, 2396) ('##tic', 13, 12036) ('novel', 14, 16) ### ### ('storyline', 15, 10950) ('final', 16, 7194) ('original', 17, 3894) ('series', 18, 3083) ### ### ('1984', 19, 4909) ### ### =============================================== v_p ================================================ ### ### ('herbert', 2, 0) ('brian', 6, 1) ('##worm', 1, 2) ('dune', 4, 3) ('kevin', 1570, 4) ### ### ('hunters', 3, 5) ('prelude', 9290, 6) ('anderson', 5, 7) ('##logies', 2839, 8) ### ### ('legends', 6064, 9) ('landmark', 1872, 10) ('sand', 10, 11) ('younger', 5117, 12) ### ### ('prequel', 184, 13) ('notes', 6513, 14) ('frank', 8, 15) ('novel', 14, 16) ('1965', 1857, 17) ### ### ('list', 3605, 18) ('duo', 4236, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('herbert', 2, 0) ('##worm', 1, 2) ('dune', 4, 3) ('brian', 6, 1) ('hunters', 3, 5) ### ### ('anderson', 5, 7) ('sand', 10, 11) ('frank', 8, 15) ('novel', 14, 16) ('hunter', 9, 26) ### ### ('hunting', 24, 31) ('sequel', 28, 30) ('universe', 46, 21) ('2006', 20, 52) ('published', 27, 42) ### ### ('prequel', 184, 13) ('sequels', 39, 28) ('2007', 23, 53) ('kevin', 1570, 4) ('book', 21, 78) ### ############################################################################################################ [2023-11-25 01:48:32,035][root][INFO] - Epoch: 1: Step: 9101/19530, loss[v]=0.375651, lr=0.000009, acc@1[1]=936.0/1024=0.9140625, acc@1[2]=965.0/1024=0.9423828125 [2023-11-25 01:50:22,447][root][INFO] - Train batch 9200 [2023-11-25 01:50:22,448][root][INFO] - Avg. loss per last 100 batches: 0.269618 [2023-11-25 01:50:23,198][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.04 | max: 3.60 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.04 | max: 4.81 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the remaining decade of his life was dominated by two concerns : finding employment and ### ### participating in controversies, most famously with richard and gabriel harvey. ### ### [P_TEXT]: ramism laurence chaderton became the leading ramist, and gabriel harvey lectured on the ### ### rhetoric of ramus. marshall mcluhan's dissertation on thomas nashe ( via the classical trivium ), ### ### who was involved in a high - profile literary quarrel with harvey, was shaped by his interest in ### ### aligning harvey with dialectic and the plain style ( logic in the sense of ramus ), and nashe with ### ### the full resources of elizabethan rhetoric. after chaderton, there was a succession of important ### ### theologians using ramist logic, including william perkins, and william ames ( amesius ), who made ### ### ramist dialectic integral to ### ### =============================================== v_q ================================================ ### ### ('gabriel', 0, 6) ('harvey', 1, 3) ('richard', 2, 3672) ('.', 3, 194) ('employment', 4, 5989) ### ### ('concerns', 5, 1077) ('finding', 6, 8324) ('decade', 7, 5455) ('life', 8, 8608) ('find', 9, 2439) ### ### ('unemployment', 10, 1080) ('participated', 11, 3341) ('controversy', 12, 3083) ('dick', 13, 13852) ### ### ('employed', 14, 4319) ('concern', 15, 6211) ('concerned', 16, 6210) ('raphael', 17, 2636) ### ### ('participants', 18, 8821) ('hire', 19, 6499) ### ### =============================================== v_p ================================================ ### ### ('ram', 13745, 0) ('perkins', 7004, 1) ('nash', 11181, 2) ('harvey', 1, 3) ('##erton', 26241, 4) ### ### ('ames', 13821, 5) ('gabriel', 0, 6) ('logic', 3307, 7) ('dissertation', 6529, 8) ('##ism', 133, 9) ### ### ('rhetoric', 3038, 10) ('quarrel', 3258, 11) ('dialect', 10501, 12) ('plain', 3979, 13) ### ### ('marshall', 5934, 14) ('thomas', 416, 15) ('integral', 1361, 16) ('chad', 6992, 17) ### ### ('##lu', 28050, 18) ('##vi', 24565, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('gabriel', 0, 6) ('harvey', 1, 3) ('##ism', 133, 9) ('thomas', 416, 15) ('.', 3, 194) ### ### ('william', 304, 26) ('integral', 1361, 16) ('logic', 3307, 7) ('rhetoric', 3038, 10) ### ### ('perkins', 7004, 1) ('quarrel', 3258, 11) ('plain', 3979, 13) ('classical', 1039, 29) ### ### ('dissertation', 6529, 8) ('richard', 2, 3672) ('nash', 11181, 2) ('marshall', 5934, 14) ### ### ('style', 2665, 27) ('ram', 13745, 0) ('elizabeth', 4321, 20) ### ############################################################################################################ [2023-11-25 01:50:23,701][root][INFO] - Epoch: 1: Step: 9201/19530, loss[v]=0.446091, lr=0.000009, acc@1[1]=933.0/1024=0.9111328125, acc@1[2]=959.5/1024=0.93701171875 [2023-11-25 01:52:14,318][root][INFO] - Train batch 9300 [2023-11-25 01:52:14,319][root][INFO] - Avg. loss per last 100 batches: 0.263032 [2023-11-25 01:52:15,044][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.05 | max: 3.62 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.04 | max: 4.85 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the cambodia national football team ( ), nicknamed the " angkor warriors " or " kouprey ### ### kampuchea ", represents cambodia in international football and is administered by the football ### ### federation of cambodia ( ffc ) which is part of the asian football confederation and sub - ### ### confederation aff. ### ### [P_TEXT]: football in cambodia the sport of football in the country of cambodia is run by the ### ### cambodian football federation. the association administers the national football team as well as ### ### the metfone c - league. football is the most popular sport in cambodia, although sepak takraw ( ### ### which looks like volleyball ), mixed martial arts and wrestling are also popular. the national ### ### team's best performance was at the 1972 afc asian cup where they reached the semi finals, arguably ### ### the best generation in cambodian football history. later, due to political plague, cambodia fell of ### ### favor ### ### =============================================== v_q ================================================ ### ### ('cambodia', 0, 0) ('football', 1, 7) ('.', 2, 517) ('##rey', 3, 23774) ('federation', 4, 9) ### ### ('warriors', 5, 4459) ('##kor', 6, 25188) ('ff', 7, 20769) ('khmer', 8, 27) ('asian', 9, 6) ### ### ('national', 10, 39) ('team', 11, 30) ('af', 12, 4093) ('##che', 13, 24122) ('ang', 14, 22189) ### ### ('confederation', 15, 1599) ('cambodian', 16, 1) ('footballer', 17, 618) ('##up', 18, 20568) ### ### ('represents', 19, 2075) ### ### =============================================== v_p ================================================ ### ### ('cambodia', 0, 0) ('cambodian', 16, 1) ('##fo', 13342, 2) ('volleyball', 1996, 3) ### ### ('sep', 17106, 4) ('plague', 13913, 5) ('asian', 9, 6) ('football', 1, 7) ('met', 15928, 8) ### ### ('federation', 4, 9) ('c', 4804, 10) ('wrestling', 901, 11) ('1972', 2787, 12) ### ### ('generation', 4649, 13) ('##ne', 21094, 14) ('martial', 14135, 15) ('cup', 71, 16) ### ### ('tak', 10520, 17) ('afc', 536, 18) ('##ak', 21445, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('cambodia', 0, 0) ('football', 1, 7) ('federation', 4, 9) ('asian', 9, 6) ('cambodian', 16, 1) ### ### ('khmer', 8, 27) ('team', 11, 30) ('national', 10, 39) ('league', 33, 22) ('cup', 71, 16) ### ### ('teams', 32, 42) ('wrestling', 901, 11) ('.', 2, 517) ('afc', 536, 18) ('volleyball', 1996, 3) ### ### ('administer', 103, 32) ('association', 381, 24) ('asia', 26, 54) ('sport', 824, 20) ### ### ('1972', 2787, 12) ### ############################################################################################################ [2023-11-25 01:52:15,548][root][INFO] - Epoch: 1: Step: 9301/19530, loss[v]=0.321029, lr=0.000010, acc@1[1]=944.5/1024=0.92236328125, acc@1[2]=981.5/1024=0.95849609375 [2023-11-25 01:54:05,874][root][INFO] - Train batch 9400 [2023-11-25 01:54:05,875][root][INFO] - Avg. loss per last 100 batches: 0.272109 [2023-11-25 01:54:06,610][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.04 | max: 3.65 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.04 | max: 4.84 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in january 2019, a production of " all in a row " sparked controversy over the way that ### ### an autistic person was being represented in puppet form. ### ### [P_TEXT]: all in a row ( play ) all in a row is a play by alex oates about a family with an 11 - ### ### year - old child on the autism spectrum. the play explores the experiences of the parents of a ### ### nonverbal, sometimes violent, autistic boy and the emotions that they experience on the night ### ### before he is taken to a residential school for children with disabilities. the play starred charlie ### ### brooks, simon lipkin, michael fox, and hugh purves, and was produced by paul virides productions at ### ### the southwark playhouse. an early draft of " all in a row " was ### ### =============================================== v_q ================================================ ### ### ('row', 0, 0) ('##tist', 1, 11) ('all', 2, 21) ('puppet', 3, 12172) ('au', 4, 33) ('.', 5, 1905) ### ### ('person', 6, 1167) ('2019', 7, 26504) ('controversy', 8, 5239) ('form', 9, 4200) ('##ic', 10, 42) ### ### ('theatre', 11, 528) ('finale', 12, 3447) ('production', 13, 53) ('troupe', 14, 2428) ### ### ('forms', 15, 2364) ('puppets', 16, 3240) ('represents', 17, 1350) ('"', 18, 690) ### ### ('musical', 19, 565) ### ### =============================================== v_p ================================================ ### ### ('row', 0, 0) ('##ride', 22646, 1) ('autism', 64, 2) ('fox', 3380, 3) ('brooks', 6214, 4) ### ### ('##wark', 26275, 5) ('##ves', 13665, 6) ('##ates', 19245, 7) ('playhouse', 375, 8) ### ### ('##bal', 23979, 9) ('spectrum', 5684, 10) ('##tist', 1, 11) ('draft', 6680, 12) ### ### ('residential', 12667, 13) ('lip', 18197, 14) ('alex', 9292, 15) ('charlie', 7967, 16) ### ### ('disabilities', 2959, 17) ('simon', 1805, 18) ('boy', 2776, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('row', 0, 0) ('##tist', 1, 11) ('all', 2, 21) ('au', 4, 33) ('autism', 64, 2) ('##ic', 10, 42) ### ### ('productions', 20, 32) ('play', 51, 30) ('playhouse', 375, 8) ('rows', 21, 46) ('child', 114, 36) ### ### ('production', 13, 53) ('starred', 351, 28) ('fox', 3380, 3) ('children', 819, 27) ### ### ('simon', 1805, 18) ('disabilities', 2959, 17) ('boy', 2776, 19) ('hugh', 2944, 20) ### ### ('brooks', 6214, 4) ### ############################################################################################################ [2023-11-25 01:54:07,110][root][INFO] - Epoch: 1: Step: 9401/19530, loss[v]=0.157845, lr=0.000010, acc@1[1]=980.0/1024=0.95703125, acc@1[2]=1002.5/1024=0.97900390625 [2023-11-25 01:55:57,901][root][INFO] - Train batch 9500 [2023-11-25 01:55:57,902][root][INFO] - Avg. loss per last 100 batches: 0.270867 [2023-11-25 01:55:58,644][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.05 | max: 3.82 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.04 | max: 4.98 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: science, but a weaker version known as the “ folkloric no free lunch theorem ” has been ### ### drawn upon by william a. dembski in support of intelligent design. ### ### [P_TEXT]: william a. dembski specified complexity cannot be purchased without intelligence. " ### ### dembski's work was strongly criticized within the scientific community, which argued that there ### ### were a number of major logical inconsistencies and evidential gaps in dembski's hypothesis. david ### ### wolpert, co - creator of the no free lunch theorem on which dembski based his book, characterized ### ### his arguments as " fatally informal and imprecise, " " written in jello, " reminiscent of ### ### philosophical discussion " of art, music, and literature, as well as much of ethics " rather than ### ### of scientific debate. mathematician mark perakh has stated he believes dembski overemphasizes his ### ### own self - importance in his ### ### =============================================== v_q ================================================ ### ### ('dem', 0, 10) ('theorem', 1, 11) ('design', 2, 2173) ('intelligent', 3, 447) ('lunch', 4, 3) ### ### ('william', 5, 8) ('.', 6, 433) ('##bs', 7, 6) ('##lor', 8, 25889) ('folk', 9, 6767) ### ### ('science', 10, 100) ('free', 11, 32) ('##ki', 12, 41) ('no', 13, 7) ('scientific', 14, 9) ### ### ('scientists', 15, 97) ('##あ', 16, 93) ('designer', 17, 983) ('sparhawk', 18, 782) ### ### ('mathematics', 19, 28) ### ### =============================================== v_p ================================================ ### ### ('##akh', 28088, 0) ('##llo', 22064, 1) ('##cise', 15453, 2) ('lunch', 4, 3) ('ethics', 4810, 4) ### ### ('hypothesis', 49, 5) ('##bs', 7, 6) ('no', 13, 7) ('william', 5, 8) ('scientific', 14, 9) ### ### ('dem', 0, 10) ('theorem', 1, 11) ('complexity', 743, 12) ('intelligence', 224, 13) ### ### ('##lp', 26746, 14) ('logical', 1055, 15) ('evident', 3837, 16) ('je', 26751, 17) ### ### ('mathematician', 689, 18) ('gaps', 5835, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('lunch', 4, 3) ('dem', 0, 10) ('theorem', 1, 11) ('william', 5, 8) ('##bs', 7, 6) ('no', 13, 7) ### ### ('scientific', 14, 9) ('free', 11, 32) ('##ki', 12, 41) ('hypothesis', 49, 5) ### ### ('mathematics', 19, 28) ('intelligence', 224, 13) ('science', 10, 100) ('literature', 270, 21) ### ### ('philosophical', 160, 30) ('complexity', 743, 12) ('intelligent', 3, 447) ### ### ('mathematician', 689, 18) ('logical', 1055, 15) ('theory', 24, 60) ### ############################################################################################################ [2023-11-25 01:55:59,146][root][INFO] - Epoch: 1: Step: 9501/19530, loss[v]=0.250330, lr=0.000010, acc@1[1]=971.5/1024=0.94873046875, acc@1[2]=1000.5/1024=0.97705078125 [2023-11-25 01:57:50,031][root][INFO] - Train batch 9600 [2023-11-25 01:57:50,032][root][INFO] - Avg. loss per last 100 batches: 0.244497 [2023-11-25 01:57:50,759][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.05 | max: 3.66 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.04 | max: 4.95 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: by 1789 she was the " prima donna " in the theatrical troupe run by johann friedel at the ### ### suburban theater auf der wieden. ### ### [P_TEXT]: the magic flute the night, " o zittre nicht, mein lieber sohn " and " der holle rache ### ### kocht in meinem herzen " require high f6, rare in opera. at the low end, the part of sarastro, ### ### premiered by franz xaver gerl, includes a conspicuous f2 in a few locations. the opera was ### ### premiered in vienna on 30 september 1791 at the suburban freihaus - theater auf der wieden. mozart ### ### conducted the orchestra, schikaneder himself played papageno, while the role of the queen of the ### ### night was sung by mozart's sister - in - law josepha hofer. ### ### =============================================== v_q ================================================ ### ### ('fried', 0, 15181) ('johann', 1, 12133) ('##ede', 2, 2) ('auf', 3, 34) ('##el', 4, 25102) ### ### ('suburban', 5, 6) ('.', 6, 646) ('troupe', 7, 2521) ('donna', 8, 14337) ('prima', 9, 11556) ### ### ('wi', 10, 61) ('1789', 11, 176) ('theatre', 12, 73) ('theater', 13, 43) ('ballet', 14, 145) ### ### ('opera', 15, 29) ('theatrical', 16, 2587) ('von', 17, 7525) ('der', 18, 58) ### ### ('shakespeare', 19, 3715) ### ### =============================================== v_p ================================================ ### ### ('flute', 2599, 0) ('magic', 3642, 1) ('##ede', 2, 2) ('papa', 5934, 3) ('##zen', 19767, 4) ### ### ('mozart', 125, 5) ('suburban', 5, 6) ('franz', 806, 7) ('##rl', 28414, 8) ('##hn', 12057, 9) ### ### ('conspicuous', 8050, 10) ('koch', 17937, 11) ('##aver', 28634, 12) ('##itt', 5472, 13) ### ### ('##fer', 27638, 14) ('joseph', 2232, 15) ('##gen', 17255, 16) ('night', 4691, 17) ### ### ('sara', 3191, 18) ('##haus', 8844, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ede', 2, 2) ('suburban', 5, 6) ('auf', 3, 34) ('opera', 15, 29) ('theater', 13, 43) ### ### ('wi', 10, 61) ('vienna', 25, 20) ('mozart', 125, 5) ('der', 18, 58) ('theatre', 12, 73) ### ### ('1791', 55, 37) ('queen', 266, 21) ('franz', 806, 7) ('flute', 2599, 0) ('1789', 11, 176) ### ### ('magic', 3642, 1) ('1790', 23, 76) ('.', 6, 646) ('joseph', 2232, 15) ('ballet', 14, 145) ### ############################################################################################################ [2023-11-25 01:57:51,263][root][INFO] - Epoch: 1: Step: 9601/19530, loss[v]=0.309714, lr=0.000010, acc@1[1]=954.0/1024=0.931640625, acc@1[2]=970.5/1024=0.94775390625 [2023-11-25 01:59:41,750][root][INFO] - Train batch 9700 [2023-11-25 01:59:41,750][root][INFO] - Avg. loss per last 100 batches: 0.275258 [2023-11-25 01:59:42,494][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.05 | max: 3.64 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.04 | max: 5.02 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the leaders of the confederation at the time of conquest were " zipa " tisquesusa, " ### ### zaque " quemuenchatocha, " iraca " sugamuxi and tundama in the northernmost portion of their ### ### territories. ### ### [P_TEXT]: nobsa rosa de viterbo in the west and in the south tibasosa and santa rosa de viterbo. ### ### nobsa is derived from chibcha and means " decent bath of today ". in the times before the spanish ### ### conquest, the area of nobsa was part of the muisca confederation, a loose confederation of the ### ### muisca. nobsa was ruled by the " tundama " of tundama. modern nobsa was founded on january 9, 1593 ### ### by jeronimo holguin and misael millan. main economical activities of nobsa are agriculture, ### ### livestock ### ### =============================================== v_q ================================================ ### ### ('tun', 0, 1) ('##ques', 1, 24925) ('zip', 2, 17197) ('##dam', 3, 2) ('ira', 4, 23420) ### ### ('que', 5, 21269) ('confederation', 6, 11) ('za', 7, 12764) ('.', 8, 1250) ('tis', 9, 16736) ### ### ('##gam', 10, 23952) ('##ux', 11, 25253) ('##que', 12, 21232) ('##to', 13, 21944) ### ### ('##usa', 14, 7418) ('conquest', 15, 17) ('su', 16, 12459) ('##ca', 17, 19) ('##mu', 18, 2326) ### ### ('guatemala', 19, 2464) ### ### =============================================== v_p ================================================ ### ### ('##bs', 20325, 0) ('tun', 0, 1) ('##dam', 3, 2) ('bath', 11307, 3) ('rosa', 3424, 4) ### ### ('##gui', 5003, 5) ('mill', 6978, 6) ('##bas', 21840, 7) ('no', 12137, 8) ('mis', 21082, 9) ### ### ('decent', 24474, 10) ('confederation', 6, 11) ('##osa', 18503, 12) ('##bc', 20941, 13) ### ### ('##bo', 24666, 14) ('loose', 3218, 15) ('ti', 14189, 16) ('conquest', 15, 17) ('##imo', 7199, 18) ### ### ('##ca', 17, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('tun', 0, 1) ('##dam', 3, 2) ('confederation', 6, 11) ('conquest', 15, 17) ('##ca', 17, 19) ### ### ('spanish', 22, 25) ('spain', 26, 40) ('##a', 146, 32) ('agriculture', 626, 21) ('159', 1178, 20) ### ### ('rosa', 3424, 4) ('ruled', 873, 30) ('##gui', 5003, 5) ('loose', 3218, 15) ('santa', 568, 38) ### ### ('livestock', 2671, 22) ('federation', 37, 75) ('activities', 393, 42) ('mill', 6978, 6) ### ### ('.', 8, 1250) ### ############################################################################################################ [2023-11-25 01:59:42,995][root][INFO] - Epoch: 1: Step: 9701/19530, loss[v]=0.413273, lr=0.000010, acc@1[1]=958.0/1024=0.935546875, acc@1[2]=976.5/1024=0.95361328125 [2023-11-25 02:01:34,052][root][INFO] - Train batch 9800 [2023-11-25 02:01:34,053][root][INFO] - Avg. loss per last 100 batches: 0.255598 [2023-11-25 02:01:34,780][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.04 | max: 3.66 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 4.72 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: all four of the famous four doctors ( bulgarus, martinus gosia, hugo de porta ravennate ### ### and jacobus de boragine ) were pupils of irnerius ; however, while there is currently no ### ### insuperable difficulty in substantiating this claim with regard to bulgarus, friedrich karl von ### ### savigny considered the evidence to be insufficient to support this claim. ### ### [P_TEXT]: four doctors of bologna the four doctors of bologna ( latin : quatuor doctores ) were ### ### italian jurists and glossator s of the 12th century, based in the university of bologna : bulgarus, ### ### martinus gosia, jacobus de boragine and hugo de porta ravennate. their teachings in the law school ### ### of bologna were based on gloss es and commentaries on the rediscovered " corpus juris civilis " of ### ### justinian. martinus may have studied with the founder of legal scholarship in bologna, irnerius. ### ### the revived ### ### =============================================== v_q ================================================ ### ### ('ir', 0, 7) ('##gny', 1, 28657) ('hugo', 2, 3) ('raven', 3, 17) ('##ner', 4, 31) ('von', 5, 14554) ### ### ('##sia', 6, 27) ('martin', 7, 10) ('port', 8, 21) ('##lga', 9, 4) ('##vi', 10, 24220) ### ### ('##rag', 11, 9) ('##rus', 12, 16) ('##ine', 13, 24) ('friedrich', 14, 11662) ('##ius', 15, 41) ### ### ('pupils', 16, 330) ('jacob', 17, 28) ('bo', 18, 39) ('.', 19, 769) ### ### =============================================== v_p ================================================ ### ### ('bologna', 4125, 0) ('corpus', 8109, 1) ('doctors', 46, 2) ('hugo', 2, 3) ('##lga', 9, 4) ### ### ('gloss', 9665, 5) ('##ator', 14833, 6) ('ir', 0, 7) ('rediscovered', 4079, 8) ('##rag', 11, 9) ### ### ('martin', 7, 10) ('doctor', 21, 11) ('jurist', 20227, 12) ('legal', 1282, 13) ### ### ('commentaries', 2769, 14) ('law', 889, 15) ('##rus', 12, 16) ('raven', 3, 17) ('civil', 3977, 18) ### ### ('##uo', 22649, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ir', 0, 7) ('hugo', 2, 3) ('##lga', 9, 4) ('raven', 3, 17) ('martin', 7, 10) ('##rag', 11, 9) ### ### ('##rus', 12, 16) ('port', 8, 21) ('##ner', 4, 31) ('##sia', 6, 27) ('doctor', 21, 11) ### ### ('##ine', 13, 24) ('jacob', 17, 28) ('##nate', 20, 29) ('doctors', 46, 2) ('bo', 18, 39) ### ### ('go', 22, 38) ('##ius', 15, 41) ('##us', 25, 40) ('four', 30, 30) ### ############################################################################################################ [2023-11-25 02:01:35,285][root][INFO] - Epoch: 1: Step: 9801/19530, loss[v]=0.355705, lr=0.000010, acc@1[1]=950.5/1024=0.92822265625, acc@1[2]=971.0/1024=0.9482421875 [2023-11-25 02:03:25,808][root][INFO] - Train batch 9900 [2023-11-25 02:03:25,809][root][INFO] - Avg. loss per last 100 batches: 0.255376 [2023-11-25 02:03:26,528][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.04 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.04 | max: 5.09 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the howrah - nagpur - mumbai line of the indian railways ( via bhusawal, nagpur, gondia, ### ### bilaspur, rourkela, kharagpur ) and is connected with most major cities. ### ### [P_TEXT]: south east central railway zone the south east central railway ( abbreviated secr ) is ### ### one of the eighteen railway zones in india. the zone office is headquartered at bilaspur and ### ### comprises the bilaspur, nagpur and raipur division. this zone was formerly part of the south ### ### eastern railway. it was inaugurated on 20 september 1998 and dedicated to the nation on 1 april ### ### 2003. bilaspur railway station is a regional hub for the system. it is the busiest junction in ### ### chhattisgarh and the fourth - busiest in central india. the zone's major stations ### ### =============================================== v_q ================================================ ### ### ('##las', 0, 0) ('##ndi', 1, 24682) ('##gpur', 2, 4) ('.', 3, 320) ('##hara', 4, 13060) ### ### ('##awa', 5, 25765) ('##hus', 6, 25386) ('na', 7, 25) ('rourke', 8, 18035) ('mumbai', 9, 1272) ### ### ('india', 10, 26) ('go', 11, 12030) ('##pur', 12, 15) ('line', 13, 597) ('indian', 14, 187) ### ### ('bi', 15, 7) ('railway', 16, 12) ('bengal', 17, 12308) ('maharashtra', 18, 1438) ### ### ('##rah', 19, 27551) ### ### =============================================== v_p ================================================ ### ### ('##las', 0, 0) ('sec', 20989, 1) ('zone', 2293, 2) ('busiest', 2939, 3) ('##gpur', 2, 4) ### ### ('eighteen', 11518, 5) ('rai', 5586, 6) ('bi', 15, 7) ('south', 6729, 8) ('eastern', 1809, 9) ### ### ('central', 358, 10) ('ch', 11897, 11) ('railway', 16, 12) ('##tis', 23154, 13) ### ### ('inaugurated', 4206, 14) ('##pur', 12, 15) ('##garh', 3945, 16) ('division', 301, 17) ### ### ('nation', 8880, 18) ('fourth', 9306, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##las', 0, 0) ('##gpur', 2, 4) ('bi', 15, 7) ('na', 7, 25) ('##pur', 12, 15) ('india', 10, 26) ### ### ('railway', 16, 12) ('junction', 41, 24) ('railways', 21, 33) ('station', 67, 27) ### ### ('central', 358, 10) ('system', 195, 21) ('.', 3, 320) ('division', 301, 17) ('zone', 2293, 2) ### ### ('busiest', 2939, 3) ('eastern', 1809, 9) ('east', 1113, 20) ('indian', 14, 187) ### ### ('railroad', 94, 44) ### ############################################################################################################ [2023-11-25 02:03:27,029][root][INFO] - Epoch: 1: Step: 9901/19530, loss[v]=0.375643, lr=0.000010, acc@1[1]=943.5/1024=0.92138671875, acc@1[2]=975.5/1024=0.95263671875 [2023-11-25 02:05:17,462][root][INFO] - Train batch 10000 [2023-11-25 02:05:17,463][root][INFO] - Avg. loss per last 100 batches: 0.255200 [2023-11-25 02:05:18,202][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.04 | max: 3.55 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.04 | max: 5.13 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the pampulha park area includes one of the largest soccer stadiums in the world, the ### ### mineirao stadium, and the sao francisco de assis church, widely known as igreja da pampulha, ### ### designed by brazilian modernist architect oscar niemeyer. ### ### [P_TEXT]: church of saint francis of assisi the church of saint francis of assisi ( igreja de sao ### ### francisco de assis, commonly known as the igreja da pampulha ) is a chapel in pampulha region of ### ### belo horizonte, in the state of minas gerais, southeastern brazil. it was designed by the brazilian ### ### architect oscar niemeyer in the organic modern style. it is the first listed modern architectural ### ### monument in brazil and consists of four undulating concrete parabolas with outdoor mosaics. the ### ### interior hosts a mural by candido portinari, and the exterior features a landscape designed by ### ### roberto burle marx. the curved lines of ### ### =============================================== v_q ================================================ ### ### ('##ira', 0, 20878) ('##pu', 1, 4) ('stadium', 2, 433) ('ass', 3, 5) ('pam', 4, 17) ### ### ('mine', 5, 3635) ('##yer', 6, 12) ('##gre', 7, 13) ('park', 8, 2716) ('.', 9, 740) ### ### ('brazilian', 10, 29) ('##ha', 11, 49) ('sao', 12, 38) ('brazil', 13, 28) ('##l', 14, 62) ### ### ('oscar', 15, 7) ('##eme', 16, 19) ('church', 17, 3) ('francisco', 18, 50) ('##is', 19, 51) ### ### =============================================== v_p ================================================ ### ### ('##bola', 28090, 0) ('mosaic', 6475, 1) ('horizon', 14885, 2) ('church', 17, 3) ('##pu', 1, 4) ### ### ('ass', 3, 5) ('candi', 17716, 6) ('oscar', 15, 7) ('francis', 2959, 8) ('##isi', 13229, 9) ### ### ('roberto', 972, 10) ('##ina', 8613, 11) ('##yer', 6, 12) ('##gre', 7, 13) ('organic', 6333, 14) ### ### ('southeastern', 9708, 15) ('bel', 13906, 16) ('pam', 4, 17) ('marx', 7227, 18) ('##eme', 16, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##pu', 1, 4) ('ass', 3, 5) ('pam', 4, 17) ('##yer', 6, 12) ('##gre', 7, 13) ('church', 17, 3) ### ### ('oscar', 15, 7) ('##eme', 16, 19) ('brazilian', 10, 29) ('brazil', 13, 28) ('sao', 12, 38) ### ### ('##ha', 11, 49) ('##ja', 23, 31) ('francisco', 18, 50) ('##is', 19, 51) ('##l', 14, 62) ### ### ('architect', 45, 35) ('и', 20, 67) ('##ل', 26, 61) ('architecture', 52, 47) ### ############################################################################################################ [2023-11-25 02:05:18,703][root][INFO] - Epoch: 1: Step: 10001/19530, loss[v]=0.161193, lr=0.000010, acc@1[1]=978.5/1024=0.95556640625, acc@1[2]=1008.0/1024=0.984375 [2023-11-25 02:07:09,655][root][INFO] - Train batch 10100 [2023-11-25 02:07:09,655][root][INFO] - Avg. loss per last 100 batches: 0.273255 [2023-11-25 02:07:10,385][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.69 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 4.89 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: ( paisley, like the leaders of the nationalist social democratic and labour party ( sdlp ### ### ) and sinn fein chose not to become a minister ) but those dup members serving as ministers ( peter ### ### robinson and nigel dodds ) refused to attend meetings of the executive committee ( cabinet ) in ### ### protest at sinn fein's participation. ### ### [P_TEXT]: 2010 united kingdom general election in northern ireland allowing the reformation of a ### ### devolved government at stormont. the dup's ian paisley became first minister, sharing power with ### ### sinn fein's martin mcguinness, in a move paisley had previously stated would never happen. a year ### ### later, paisley stood down as dup leader and was replaced by his deputy, peter robinson. in the run ### ### up to the 2010 elections, robinson suffered a series of personal setbacks. prior to and during the ### ### mp expenses scandal in 2009, questions were asked about his family's remunerations and expenses. ### ### the following year, ### ### =============================================== v_q ================================================ ### ### ('dodd', 0, 20019) ('paisley', 1, 0) ('nigel', 2, 13030) ('.', 3, 135) ('fein', 4, 29) ### ### ('robinson', 5, 3) ('peter', 6, 8) ('ireland', 7, 24) ('cabinet', 8, 98) ('du', 9, 10) ### ### ('committee', 10, 9548) ('sd', 11, 20343) ('executive', 12, 8977) ('social', 13, 6941) ### ### ('labour', 14, 2256) ('sinn', 15, 39) ('minister', 16, 37) ('irish', 17, 72) ### ### ('nationalist', 18, 1488) ('party', 19, 1894) ### ### =============================================== v_p ================================================ ### ### ('paisley', 1, 0) ('##olved', 12775, 1) ('##ont', 17330, 2) ('robinson', 5, 3) ### ### ('expenses', 9545, 4) ('storm', 6268, 5) ('2010', 6448, 6) ('##eration', 13133, 7) ('peter', 6, 8) ### ### ('martin', 4956, 9) ('du', 9, 10) ('##gui', 22040, 11) ('ian', 4636, 12) ('kingdom', 180, 13) ### ### ('scandal', 7119, 14) ('reformation', 10896, 15) ('##mun', 15644, 16) ('elections', 257, 17) ### ### ('##nne', 17547, 18) ('##backs', 24225, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('paisley', 1, 0) ('robinson', 5, 3) ('peter', 6, 8) ('du', 9, 10) ('fein', 4, 29) ### ### ('ireland', 7, 24) ('minister', 16, 37) ('sinn', 15, 39) ('##p', 34, 28) ('election', 45, 25) ### ### ('kingdom', 180, 13) ('leader', 85, 32) ('ministers', 27, 48) ('mp', 201, 22) ### ### ('elections', 257, 17) ('irish', 17, 72) ('cabinet', 8, 98) ('.', 3, 135) ('government', 31, 53) ### ### ('leaders', 71, 46) ### ############################################################################################################ [2023-11-25 02:07:10,890][root][INFO] - Epoch: 1: Step: 10101/19530, loss[v]=0.365301, lr=0.000010, acc@1[1]=952.0/1024=0.9296875, acc@1[2]=976.0/1024=0.953125 [2023-11-25 02:09:01,065][root][INFO] - Train batch 10200 [2023-11-25 02:09:01,066][root][INFO] - Avg. loss per last 100 batches: 0.260889 [2023-11-25 02:09:01,786][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.05 | max: 3.57 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.04 | max: 5.25 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: where is where is the second album by cetu javu, released in 1992. ### ### [P_TEXT]: por que? " por que? " is a 1991 single by german synthpop band cetu javu. it featured on ### ### the lp and cd versions of their second album " where is where ". ### ### =============================================== v_q ================================================ ### ### ('##tu', 0, 5) ('##vu', 1, 2) ('where', 2, 9) ('ce', 3, 7) ('album', 4, 13) ('1992', 5, 24) ### ### ('ja', 6, 10) ('.', 7, 285) ('fiji', 8, 3001) ('1991', 9, 8) ('second', 10, 16) ('albums', 11, 25) ### ### ('band', 12, 17) ('musician', 13, 69) ('when', 14, 32) ('hendrix', 15, 14510) ('malta', 16, 258) ### ### ('is', 17, 84) ('madagascar', 18, 1577) ('estonia', 19, 15193) ### ### =============================================== v_p ================================================ ### ### ('que', 1644, 0) ('por', 22118, 1) ('##vu', 1, 2) ('synth', 3308, 3) ('?', 1291, 4) ('##tu', 0, 5) ### ### ('lp', 35, 6) ('ce', 3, 7) ('1991', 9, 8) ('where', 2, 9) ('ja', 6, 10) ('##pop', 488, 11) ### ### ('cd', 66, 12) ('album', 4, 13) ('single', 884, 14) ('bands', 34, 15) ('second', 10, 16) ### ### ('band', 12, 17) ('german', 2368, 18) ('1990', 53, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##vu', 1, 2) ('##tu', 0, 5) ('where', 2, 9) ('ce', 3, 7) ('album', 4, 13) ('ja', 6, 10) ### ### ('1991', 9, 8) ('1992', 5, 24) ('lp', 35, 6) ('second', 10, 16) ('band', 12, 17) ('bands', 34, 15) ### ### ('cd', 66, 12) ('que', 1644, 0) ('1990', 53, 19) ('albums', 11, 25) ('?', 1291, 4) ('.', 7, 285) ### ### ('##pop', 488, 11) ('synth', 3308, 3) ### ############################################################################################################ [2023-11-25 02:09:02,286][root][INFO] - Epoch: 1: Step: 10201/19530, loss[v]=0.322628, lr=0.000010, acc@1[1]=953.5/1024=0.93115234375, acc@1[2]=976.5/1024=0.95361328125 [2023-11-25 02:10:52,543][root][INFO] - Train batch 10300 [2023-11-25 02:10:52,544][root][INFO] - Avg. loss per last 100 batches: 0.260052 [2023-11-25 02:10:53,277][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.05 | max: 3.74 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 5.04 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: ' s australasian antarctic expedition in 1911, but left the expedition before it departed ### ### for the antarctic. ### ### [P_TEXT]: ross sea party douglas mawson and the australasian antarctic expedition. to lead the ross ### ### sea party shackleton chose aeneas mackintosh, having first attempted to persuade the admiralty to ### ### provide him with a naval crew. mackintosh, like shackleton, was a former merchant navy officer, who ### ### had been on the nimrod expedition until his participation was cut short by an accident that ### ### resulted in the loss of his right eye. another " nimrod " veteran, ernest joyce, whose antarctic ### ### experiences had begun with captain scott's discovery expedition, was appointed to take charge of ### ### sledging ### ### =============================================== v_q ================================================ ### ### ('antarctic', 0, 9) ('expedition', 1, 15) ('1911', 2, 9251) ('.', 3, 683) ('##tral', 4, 49) ### ### ('aus', 5, 28) ('antarctica', 6, 120) ('sailed', 7, 2974) ('polar', 8, 929) ('1912', 9, 11285) ### ### ('australian', 10, 548) ('1910', 11, 11287) ('absently', 12, 359) ('explorer', 13, 957) ### ### ('glacier', 14, 4643) ('australia', 15, 239) ('survey', 16, 408) ('geological', 17, 1437) ### ### ('alaska', 18, 1811) ('1909', 19, 15303) ### ### =============================================== v_p ================================================ ### ### ('ross', 4696, 0) ('joyce', 17125, 1) ('eye', 11884, 2) ('admiralty', 764, 3) ('scott', 7025, 4) ### ### ('sea', 41, 5) ('##neas', 27977, 6) ('##leton', 23357, 7) ('shack', 10992, 8) ('antarctic', 0, 9) ### ### ('##edging', 20171, 10) ('##ws', 27626, 11) ('##rod', 26024, 12) ('party', 1987, 13) ### ### ('accident', 1498, 14) ('expedition', 1, 15) ('mack', 12689, 16) ('ernest', 2246, 17) ### ### ('ni', 19256, 18) ('##int', 21325, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('antarctic', 0, 9) ('expedition', 1, 15) ('aus', 5, 28) ('##tral', 4, 49) ('sea', 41, 5) ### ### ('discovery', 143, 21) ('admiralty', 764, 3) ('expeditions', 27, 51) ('ross', 4696, 0) ### ### ('accident', 1498, 14) ('party', 1987, 13) ('antarctica', 6, 120) ('captain', 334, 30) ### ### ('ernest', 2246, 17) ('navy', 767, 26) ('.', 3, 683) ('convince', 240, 38) ('scott', 7025, 4) ### ### ('manpower', 22, 72) ('naval', 898, 37) ### ############################################################################################################ [2023-11-25 02:10:53,781][root][INFO] - Epoch: 1: Step: 10301/19530, loss[v]=0.357333, lr=0.000011, acc@1[1]=961.0/1024=0.9384765625, acc@1[2]=983.5/1024=0.96044921875 [2023-11-25 02:12:43,896][root][INFO] - Train batch 10400 [2023-11-25 02:12:43,896][root][INFO] - Avg. loss per last 100 batches: 0.254554 [2023-11-25 02:12:44,625][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.04 | max: 3.44 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 5.14 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.6/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: bunting reached the quarter - finals for the third major event in a row by beating wade ### ### 10 – 8 at the grand slam of darts, but lost 16 – 9 against mervyn king. ### ### [P_TEXT]: mervyn king ( darts player ) the masters to james wade. king showed his best battling ### ### abilities at the uk open to defeat gary anderson 9 – 8 after being 8 – 4 down and surviving 13 ### ### match darts from anderson. he then overturned an 8 – 5 deficit against nathan derry to win 9 – 8 ### ### and emerged successfully from a deciding leg for the third game in a row by beating kyle anderson 9 ### ### – 8 to reach the quarter - finals. he met stephen bunting and the first 12 legs went on throw, ### ### before bunting took four unanswered legs to inflict a ### ### =============================================== v_q ================================================ ### ### ('##ting', 0, 13) ('wade', 1, 1) ('##vyn', 2, 5) ('king', 3, 8) ('darts', 4, 6) ('bun', 5, 9) ### ### ('.', 6, 2352) ('mer', 7, 18) ('quarter', 8, 162) ('golf', 9, 2195) ('snooker', 10, 4553) ### ### ('masters', 11, 20) ('ˈ', 12, 588) ('slam', 13, 14738) ('tournament', 14, 60) ('open', 15, 22) ### ### ('round', 16, 302) ('major', 17, 17373) ('pga', 18, 1123) ('##ding', 19, 634) ### ### =============================================== v_p ================================================ ### ### ('derry', 19288, 0) ('wade', 1, 1) ('anderson', 4274, 2) ('nathan', 5745, 3) ('throw', 5392, 4) ### ### ('##vyn', 2, 5) ('darts', 4, 6) ('kyle', 1823, 7) ('king', 3, 8) ('bun', 5, 9) ('gary', 4828, 10) ### ### ('deficit', 4866, 11) ('##flict', 13654, 12) ('##ting', 0, 13) ('stephen', 547, 14) ### ### ('deciding', 6572, 15) ('uk', 286, 16) ('##wer', 16236, 17) ('mer', 7, 18) ('james', 3822, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('wade', 1, 1) ('##ting', 0, 13) ('##vyn', 2, 5) ('darts', 4, 6) ('king', 3, 8) ('bun', 5, 9) ### ### ('mer', 7, 18) ('masters', 11, 20) ('open', 15, 22) ('matches', 37, 31) ('match', 46, 28) ### ### ('finals', 90, 23) ('uk', 286, 16) ('tennis', 20, 50) ('tournament', 14, 60) ('opponents', 25, 44) ### ### ('stephen', 547, 14) ('kyle', 1823, 7) ('anderson', 4274, 2) ('quarter', 8, 162) ### ############################################################################################################ [2023-11-25 02:12:45,126][root][INFO] - Epoch: 1: Step: 10401/19530, loss[v]=0.183944, lr=0.000011, acc@1[1]=980.5/1024=0.95751953125, acc@1[2]=1000.0/1024=0.9765625 [2023-11-25 02:14:34,797][root][INFO] - Train batch 10500 [2023-11-25 02:14:34,798][root][INFO] - Avg. loss per last 100 batches: 0.245550 [2023-11-25 02:14:35,521][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.60 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 5.03 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in his 2006 state of the union address, bush declared, " america is addicted to oil " and ### ### announced his advanced energy initiative to increase energy development research. ### ### [P_TEXT]: u. s. bank tower ( los angeles ) guinness book of world records for highest - placed ### ### logo. on june 16, 2004, the 9 / 11 commission reported that the original plan for the september 11 ### ### attacks called for the hijacking of ten planes, one of which was to be crashed into the building. ### ### on october 6, 2005, house officials stated that the government had foiled a previously undisclosed ### ### second plot to crash a plane into the building in mid - 2002. in his televised 2006 state of the ### ### union address, president george w. bush asserted that american counterterrorism officials foiled a ### ### plot ### ### =============================================== v_q ================================================ ### ### ('union', 0, 14) ('energy', 1, 489) ('bush', 2, 21) ('advanced', 3, 4094) ('state', 4, 35) ### ### ('america', 5, 240) ('initiative', 6, 2190) ('addicted', 7, 13353) ('oil', 8, 260) ('2006', 9, 18) ### ### ('development', 10, 4322) ('.', 11, 238) ('research', 12, 3321) ('address', 13, 1) ### ### ('addiction', 14, 12971) ('addict', 15, 9314) ('researchers', 16, 6707) ('reagan', 17, 1499) ### ### ('speeches', 18, 8872) ('petroleum', 19, 1108) ### ### =============================================== v_p ================================================ ### ### ('guinness', 26204, 0) ('address', 13, 1) ('commission', 3470, 2) ('tower', 4112, 3) ### ### ('attacks', 5723, 4) ('logo', 5685, 5) ('bank', 18033, 6) ('u', 9758, 7) ('##jack', 27847, 8) ### ### ('planes', 18514, 9) ('angeles', 15305, 10) ('counter', 4377, 11) ('##ror', 28526, 12) ### ### ('plot', 9784, 13) ('union', 0, 14) ('book', 6781, 15) ('hi', 13466, 16) ('officials', 2570, 17) ### ### ('2006', 9, 18) ('11', 12705, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('union', 0, 14) ('bush', 2, 21) ('address', 13, 1) ('state', 4, 35) ('2006', 9, 18) ### ### ('americans', 21, 43) ('president', 33, 32) ('america', 5, 240) ('addresses', 131, 41) ### ### ('oil', 8, 260) ('2005', 116, 44) ('2002', 503, 22) ('energy', 1, 489) ('commission', 3470, 2) ### ### ('usa', 64, 53) ('.', 11, 238) ('tower', 4112, 3) ('addressing', 44, 63) ('house', 1490, 27) ### ### ('addressed', 97, 58) ### ############################################################################################################ [2023-11-25 02:14:36,024][root][INFO] - Epoch: 1: Step: 10501/19530, loss[v]=0.385915, lr=0.000011, acc@1[1]=941.5/1024=0.91943359375, acc@1[2]=964.5/1024=0.94189453125 [2023-11-25 02:16:26,048][root][INFO] - Train batch 10600 [2023-11-25 02:16:26,049][root][INFO] - Avg. loss per last 100 batches: 0.255675 [2023-11-25 02:16:26,788][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.05 | max: 3.64 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 5.07 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: samuel marsden, who also knew of kendall's affair and his close relationship with hongi ### ### hika, returned to new zealand in august 1823 to sack him in person. ### ### [P_TEXT]: samuel marsden view of assisting engaging in missionary work among the maori people in ### ### new zealand. at a meeting in the colony of new south wales, held at sydney, on 20 december 1813, ### ### marsden formed the new south wales society for affording protection to the natives of the south sea ### ### islands, and promoting their civilization, for the protection of south sea islanders who may be ### ### brought to port jackson, and to defend their claims on the masters and owners of the vessels who ### ### mistreat those islanders. thomas kendall and william hall sailed on the " earl ### ### =============================================== v_q ================================================ ### ### ('marsden', 0, 0) ('kendall', 1, 1) ('hong', 2, 5556) ('samuel', 3, 2) ('zealand', 4, 25) ### ### ('hi', 5, 20375) ('affair', 6, 13417) ('##ka', 7, 25031) ('.', 8, 1376) ('1823', 9, 7247) ### ### ('##i', 10, 28043) ('maori', 11, 12) ('1824', 12, 2275) ('sydney', 13, 9) ('australia', 14, 86) ### ### ('tanzania', 15, 2585) ('##ヒ', 16, 926) ('relationship', 17, 7025) ('sack', 18, 16764) ### ### ('queensland', 19, 299) ### ### =============================================== v_p ================================================ ### ### ('marsden', 0, 0) ('kendall', 1, 1) ('samuel', 3, 2) ('jackson', 1386, 3) ('islanders', 3180, 4) ### ### ('hall', 6952, 5) ('civilization', 3564, 6) ('society', 156, 7) ('mist', 18748, 8) ### ### ('sydney', 13, 9) ('masters', 6246, 10) ('earl', 2662, 11) ('maori', 11, 12) ('afford', 15270, 13) ### ### ('colony', 764, 14) ('william', 486, 15) ('port', 12008, 16) ('sea', 1327, 17) ### ### ('islands', 4215, 18) ('missionary', 178, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('marsden', 0, 0) ('kendall', 1, 1) ('samuel', 3, 2) ('zealand', 4, 25) ('maori', 11, 12) ### ### ('sydney', 13, 9) ('sailed', 31, 23) ('1813', 70, 21) ('society', 156, 7) ('missionary', 178, 19) ### ### ('jackson', 1386, 3) ('william', 486, 15) ('colony', 764, 14) ('sea', 1327, 17) ### ### ('islanders', 3180, 4) ('civilization', 3564, 6) ('earl', 2662, 11) ('missionaries', 136, 37) ### ### ('wales', 1447, 22) ('newly', 73, 46) ### ############################################################################################################ [2023-11-25 02:16:27,292][root][INFO] - Epoch: 1: Step: 10601/19530, loss[v]=0.164918, lr=0.000011, acc@1[1]=980.5/1024=0.95751953125, acc@1[2]=1006.0/1024=0.982421875 [2023-11-25 02:18:17,477][root][INFO] - Train batch 10700 [2023-11-25 02:18:17,478][root][INFO] - Avg. loss per last 100 batches: 0.255965 [2023-11-25 02:18:18,193][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.56 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 4.62 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: chris froome and vincenzo nibali, he lost the leader's jersey a day after he had won it. ### ### [P_TEXT]: 2009 tour of california most of team astana, consolidated on the descent. four riders ### ### came clear and were in the lead approaching the last climb of the day - vincenzo nibali, frank ### ### schleck, bauke mollema, and glen chadwick. nibali, 2'21 " behind leipheimer, was a small threat to ### ### the race lead, but the astana - paced leading group kept the time gap at around one minute. after ### ### numerous attacks, only nibali and schleck remained out front. the chase from the astana - led group ### ### proved too slow to catch nibali and schleck, and they were able ### ### =============================================== v_q ================================================ ### ### ('chris', 0, 13411) ('##oom', 1, 26644) ('##bal', 2, 2) ('jersey', 3, 7795) ('fr', 4, 20022) ### ### ('vincenzo', 5, 15) ('ni', 6, 17) ('prix', 7, 3169) ('.', 8, 1582) ('riders', 9, 30) ### ### ('cycling', 10, 1731) ('cyclist', 11, 6462) ('race', 12, 28) ('rally', 13, 6737) ('##e', 14, 21612) ### ### ('winner', 15, 8850) ('racing', 16, 43) ('rider', 17, 78) ('scratched', 18, 2651) ('lap', 19, 6821) ### ### =============================================== v_p ================================================ ### ### ('##tana', 21117, 0) ('tour', 24, 1) ('##bal', 2, 2) ('california', 5171, 3) ('##wick', 16796, 4) ### ### ('##lle', 4490, 5) ('chad', 40, 6) ('2009', 803, 7) ('##uke', 21046, 8) ('##eck', 26281, 9) ### ### ('gap', 101, 10) ('glen', 3404, 11) ('lei', 4991, 12) ('frank', 3048, 13) ('climb', 141, 14) ### ### ('vincenzo', 5, 15) ('##ma', 19400, 16) ('ni', 6, 17) ('##hl', 22379, 18) ('21', 7163, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##bal', 2, 2) ('vincenzo', 5, 15) ('ni', 6, 17) ('tour', 24, 1) ('riders', 9, 30) ### ### ('race', 12, 28) ('chad', 40, 6) ('gap', 101, 10) ('climb', 141, 14) ('racing', 16, 43) ### ### ('chase', 245, 23) ('2009', 803, 7) ('front', 93, 34) ('day', 76, 38) ('team', 219, 33) ### ### ('rider', 17, 78) ('##i', 25, 66) ('attacks', 1485, 21) ('##lle', 4490, 5) ('california', 5171, 3) ### ############################################################################################################ [2023-11-25 02:18:18,695][root][INFO] - Epoch: 1: Step: 10701/19530, loss[v]=0.408012, lr=0.000011, acc@1[1]=940.5/1024=0.91845703125, acc@1[2]=965.0/1024=0.9423828125 [2023-11-25 02:20:08,413][root][INFO] - Train batch 10800 [2023-11-25 02:20:08,414][root][INFO] - Avg. loss per last 100 batches: 0.254221 [2023-11-25 02:20:09,150][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.67 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 4.90 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he is also the subject of the 1989 film " lean on me ", starring morgan freeman. ### ### [P_TEXT]: eastside high school ( paterson, new jersey ) ( against rancocas valley regional high ### ### school in the tournament final ) and in 2015 ( vs. cherry hill high school east ). the team won the ### ### group iv title in 2011 with a 56 - 52 win against rancocas valley in the championship game. the ### ### team won its second group iv title with a 50 - 34 win against cherry hill east in the tournament ### ### final. paterson eastside high is known for its renaissance in the mid - 1980s under the leadership ### ### of joe clark as principal. the school was depicted in the 1989 film " lean on me ", starring morgan ### ### freeman. former student, fetty wap, filmed the music video for his song " wake ### ### =============================================== v_q ================================================ ### ### ('lean', 0, 0) ('morgan', 1, 5) ('freeman', 2, 7) ('me', 3, 13) ('1989', 4, 18) ('film', 5, 38) ### ### ('.', 6, 474) ('1988', 7, 68) ('subject', 8, 6435) ('my', 9, 99) ('1990', 10, 57) ('on', 11, 117) ### ### ('sean', 12, 13755) ('starring', 13, 211) ('films', 14, 60) ('michael', 15, 6584) ### ### ('hollywood', 16, 907) ('1993', 17, 81) ('john', 18, 14313) ('nathan', 19, 5016) ### ### =============================================== v_p ================================================ ### ### ('lean', 0, 0) ('wake', 1743, 1) ('##cas', 28858, 2) ('cherry', 10332, 3) ('paterson', 18018, 4) ### ### ('morgan', 1, 5) ('clark', 2620, 6) ('freeman', 2, 7) ('##side', 19995, 8) ('school', 2307, 9) ### ### ('iv', 9734, 10) ('tournament', 18203, 11) ('student', 9536, 12) ('me', 3, 13) ('joe', 486, 14) ### ### ('renaissance', 9222, 15) ('##tty', 26811, 16) ('filmed', 49, 17) ('1989', 4, 18) ('ran', 5817, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('lean', 0, 0) ('morgan', 1, 5) ('freeman', 2, 7) ('me', 3, 13) ('1989', 4, 18) ('film', 5, 38) ### ### ('filmed', 49, 17) ('filming', 24, 39) ('1988', 7, 68) ('1990', 10, 57) ('wake', 1743, 1) ### ### ('joe', 486, 14) ('films', 14, 60) ('my', 9, 99) ('clark', 2620, 6) ('school', 2307, 9) ### ### ('hill', 1104, 25) ('game', 470, 35) ('1993', 17, 81) ('movie', 28, 70) ### ############################################################################################################ [2023-11-25 02:20:09,653][root][INFO] - Epoch: 1: Step: 10801/19530, loss[v]=0.332002, lr=0.000011, acc@1[1]=948.0/1024=0.92578125, acc@1[2]=976.0/1024=0.953125 [2023-11-25 02:21:59,984][root][INFO] - Train batch 10900 [2023-11-25 02:21:59,985][root][INFO] - Avg. loss per last 100 batches: 0.250094 [2023-11-25 02:22:00,723][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.05 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 5.17 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the club has won eight syrian league titles and eight syrian cup titles. ### ### [P_TEXT]: 2010 – 11 syrian cup the 2010 - 11 version of the syrian cup is the 41st edition to be ### ### played. it is the premier knockout tournament for football teams in syria. al - karamah went into ### ### this edition as the holders once again. the competition started on 21 september 2010 but changed ### ### from the previous year of games over two legs to a one legged affair. the cup winner were ### ### guaranteed a place in the 2012 afc cup. two teams play a knockout tie. one team advance to the next ### ### round. games played over two legs. 26 teams play a knockout ### ### =============================================== v_q ================================================ ### ### ('syrian', 0, 2) ('league', 1, 2455) ('syria', 2, 0) ('.', 3, 257) ('cup', 4, 7) ### ### ('footballer', 5, 187) ('club', 6, 607) ('eight', 7, 317) ('clubs', 8, 247) ('damascus', 9, 32) ### ### ('football', 10, 24) ('yemen', 11, 695) ('egypt', 12, 51) ('fc', 13, 683) ('jordan', 14, 1390) ### ### ('rivals', 15, 310) ('squad', 16, 3563) ('shocked', 17, 52) ('##ァ', 18, 1423) ('syriac', 19, 141) ### ### =============================================== v_p ================================================ ### ### ('syria', 2, 0) ('knockout', 92, 1) ('syrian', 0, 2) ('kara', 9730, 3) ('afc', 188, 4) ### ### ('legged', 3648, 5) ('2010', 2856, 6) ('cup', 4, 7) ('11', 4818, 8) ('affair', 9842, 9) ### ### ('##mah', 29086, 10) ('2012', 716, 11) ('premier', 180, 12) ('legs', 2242, 13) ('al', 178, 14) ### ### ('holders', 1150, 15) ('guaranteed', 20696, 16) ('teams', 24, 17) ('tournament', 310, 18) ### ### ('41st', 23288, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('syrian', 0, 2) ('syria', 2, 0) ('cup', 4, 7) ('knockout', 92, 1) ('football', 10, 24) ### ### ('teams', 24, 17) ('afc', 188, 4) ('damascus', 9, 32) ('premier', 180, 12) ('al', 178, 14) ### ### ('.', 3, 257) ('2012', 716, 11) ('tournament', 310, 18) ('footballer', 5, 187) ('egypt', 12, 51) ### ### ('shocked', 17, 52) ('legged', 3648, 5) ('2010', 2856, 6) ('eight', 7, 317) ('holders', 1150, 15) ### ############################################################################################################ [2023-11-25 02:22:01,226][root][INFO] - Epoch: 1: Step: 10901/19530, loss[v]=0.314261, lr=0.000011, acc@1[1]=956.0/1024=0.93359375, acc@1[2]=982.5/1024=0.95947265625 [2023-11-25 02:23:51,021][root][INFO] - Train batch 11000 [2023-11-25 02:23:51,023][root][INFO] - Avg. loss per last 100 batches: 0.227152 [2023-11-25 02:23:51,736][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.53 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.03 | max: 5.10 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 29.0/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: under the bbc's control, the songs in the second half were by schubert, quilter and parry ### ### rather than ballads from chappell's. ### ### [P_TEXT]: beatrice harrison fritz kreisler in a mendelssohn concert at helsingfors. both may and ### ### beatrice won the gold medal of the associated board for violin and cello respectively. the harrison ### ### family became friends with roger quilter and his circle through the soldiers'concerts in 1916. on ### ### 11 march 1918 beatrice performed dvorak's cello concerto in b minor with the royal philharmonic ### ### orchestra under thomas beecham. hugo becker had spoken to sir henry wood of his admiration for ### ### beatrice harrison's playing even before her debut. both ### ### =============================================== v_q ================================================ ### ### ('schubert', 0, 17174) ('##ppel', 1, 16203) ('quilt', 2, 2) ('parry', 3, 14019) ('bbc', 4, 3579) ### ### ('ballad', 5, 10619) ('ballads', 6, 10009) ('cha', 7, 20699) ('.', 8, 1548) ('##er', 9, 44) ### ### ('melody', 10, 649) ('song', 11, 633) ('##l', 12, 10285) ('vienna', 13, 1122) ### ### ('beethoven', 14, 1165) ('german', 15, 926) ('songs', 16, 526) ('tune', 17, 354) ### ### ('singers', 18, 168) ('itv', 19, 18024) ### ### =============================================== v_p ================================================ ### ### ('beatrice', 7624, 0) ('harrison', 77, 1) ('quilt', 2, 2) ('beech', 16846, 3) ('minor', 1091, 4) ### ### ('wood', 3191, 5) ('fritz', 8709, 6) ('concerto', 1136, 7) ('becker', 22386, 8) ('##vor', 23745, 9) ### ### ('cello', 3335, 10) ('circle', 3913, 11) ('hugo', 8687, 12) ('roger', 605, 13) ('board', 1946, 14) ### ### ('##for', 26822, 15) ('##ak', 26313, 16) ('philharmonic', 2524, 17) ('associated', 9058, 18) ### ### ('##ls', 579, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('quilt', 2, 2) ('##er', 9, 44) ('harrison', 77, 1) ('blanket', 21, 51) ('orchestra', 119, 23) ### ### ('concert', 108, 27) ('violin', 165, 25) ('minor', 1091, 4) ('roger', 605, 13) ('##ls', 579, 19) ### ### ('concerto', 1136, 7) ('wood', 3191, 5) ('board', 1946, 14) ('henry', 809, 29) ### ### ('beatrice', 7624, 0) ('thomas', 816, 31) ('philharmonic', 2524, 17) ('cello', 3335, 10) ### ### ('circle', 3913, 11) ('music', 22, 97) ### ############################################################################################################ [2023-11-25 02:23:52,238][root][INFO] - Epoch: 1: Step: 11001/19530, loss[v]=0.304664, lr=0.000011, acc@1[1]=967.0/1024=0.9443359375, acc@1[2]=982.5/1024=0.95947265625 [2023-11-25 02:25:42,067][root][INFO] - Train batch 11100 [2023-11-25 02:25:42,068][root][INFO] - Avg. loss per last 100 batches: 0.234225 [2023-11-25 02:25:42,798][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.47 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 4.99 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.7/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: i, a woman ( original danish : jeg - en kvinde ) is a 1965 black - and - white danish - ### ### swedish erotic film whose break - through popularity helped initiate a wave of sexploitation film s ### ### into mainstream american theaters in the late 1960s, and inspired andy warhol to write and direct ### ### his feature - length experimental film version " i, a man ". ### ### [P_TEXT]: i, a man i, a man is a 1967 american erotic drama film written, directed and filmed by ### ### andy warhol. it debuted at the hudson theatre in new york city on august 25, 1967. the film depicts ### ### the main character, played by tom baker, in a series of sexual encounters with eight women. warhol ### ### created the movie as a response to the popular erotic scandinavia n film " i, a woman " ( 1965 ) ### ### which had opened in the united states in october 1966. the film featured several ### ### =============================================== v_q ================================================ ### ### ('woman', 0, 9) ('man', 1, 2) ('danish', 2, 70) ('erotic', 3, 4) ('##hol', 4, 10) ('film', 5, 18) ### ### ('break', 6, 3109) ('andy', 7, 11) ('i', 8, 3) ('##pl', 9, 19606) ('##ind', 10, 25963) ### ### ('denmark', 11, 36) ('1965', 12, 16) ('и', 13, 47) ('women', 14, 14) ('##ation', 15, 9055) ### ### ('war', 16, 17) ('sex', 17, 137) ('##oit', 18, 26135) ('swedish', 19, 406) ### ### =============================================== v_p ================================================ ### ### ('hudson', 2980, 0) ('baker', 4580, 1) ('man', 1, 2) ('i', 8, 3) ('erotic', 3, 4) ('tom', 1427, 5) ### ### ('scandinavia', 240, 6) ('theatre', 231, 7) ('1967', 67, 8) ('woman', 0, 9) ('##hol', 4, 10) ### ### ('andy', 7, 11) ('1966', 43, 12) ('filmed', 86, 13) ('women', 14, 14) ('sexual', 41, 15) ### ### ('1965', 12, 16) ('war', 16, 17) ('film', 5, 18) ('encounters', 13763, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('man', 1, 2) ('woman', 0, 9) ('erotic', 3, 4) ('i', 8, 3) ('##hol', 4, 10) ('andy', 7, 11) ### ### ('1965', 12, 16) ('film', 5, 18) ('women', 14, 14) ('war', 16, 17) ('denmark', 11, 36) ### ### ('1966', 43, 12) ('sexual', 41, 15) ('1967', 67, 8) ('и', 13, 47) ('danish', 2, 70) ('men', 34, 23) ### ### ('sweden', 26, 32) ('filmed', 86, 13) ('scandinavia', 240, 6) ### ############################################################################################################ [2023-11-25 02:25:43,300][root][INFO] - Epoch: 1: Step: 11101/19530, loss[v]=0.156002, lr=0.000011, acc@1[1]=990.0/1024=0.966796875, acc@1[2]=1006.0/1024=0.982421875 [2023-11-25 02:27:33,913][root][INFO] - Train batch 11200 [2023-11-25 02:27:33,914][root][INFO] - Avg. loss per last 100 batches: 0.231287 [2023-11-25 02:27:34,655][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.05 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 5.01 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: a blind arch is an arch infilled with solid construction so it cannot function as a ### ### window, door, or passageway. ### ### [P_TEXT]: blind arcade a blind arcade or blind window is an arcade that is composed of a series of ### ### arch es that has no actual openings and that is applied to the surface of a wall as a decorative ### ### element : i. e., the arches are not window s or openings but are part of the masonry face. it is ### ### designed as an ornamental architectural element and has no load - bearing function. whereas a blind ### ### arch is usually a single arch or a series of joined arches as a frieze ( sometimes called ### ### =============================================== v_q ================================================ ### ### ('arch', 0, 2) ('blind', 1, 1) ('door', 2, 1120) ('window', 3, 9) ('solid', 4, 1908) ### ### ('arches', 5, 7) ('corridor', 6, 1653) ('passageway', 7, 23003) ('doors', 8, 3587) ('.', 9, 351) ### ### ('tunnel', 10, 668) ('sparhawk', 11, 88) ('construction', 12, 300) ('arched', 13, 48) ### ### ('is', 14, 8956) ('hallway', 15, 2197) ('passage', 16, 2859) ('windows', 17, 25) ### ### ('##fi', 18, 23531) ('##lled', 19, 23549) ### ### =============================================== v_p ================================================ ### ### ('arcade', 2496, 0) ('blind', 1, 1) ('arch', 0, 2) ('ornamental', 230, 3) ('openings', 659, 4) ### ### ('es', 17165, 5) ('masonry', 713, 6) ('arches', 5, 7) ('fr', 18258, 8) ('window', 3, 9) ### ### ('load', 10735, 10) ('face', 1617, 11) ('surface', 3958, 12) ('decorative', 29, 13) ### ### ('element', 1582, 14) ('architectural', 75, 15) ('bearing', 4043, 16) ('##ie', 26389, 17) ### ### ('single', 3099, 18) ('applied', 1748, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('blind', 1, 1) ('arch', 0, 2) ('window', 3, 9) ('arches', 5, 7) ('decorative', 29, 13) ### ### ('windows', 17, 25) ('ornamental', 230, 3) ('wall', 32, 21) ('architectural', 75, 15) ### ### ('architecture', 31, 24) ('openings', 659, 4) ('arcade', 2496, 0) ('masonry', 713, 6) ### ### ('arched', 13, 48) ('function', 102, 20) ('sparhawk', 11, 88) ('face', 1617, 11) ### ### ('element', 1582, 14) ('door', 2, 1120) ('.', 9, 351) ### ############################################################################################################ [2023-11-25 02:27:35,158][root][INFO] - Epoch: 1: Step: 11201/19530, loss[v]=0.312133, lr=0.000011, acc@1[1]=960.0/1024=0.9375, acc@1[2]=981.0/1024=0.9580078125 [2023-11-25 02:29:25,598][root][INFO] - Train batch 11300 [2023-11-25 02:29:25,599][root][INFO] - Avg. loss per last 100 batches: 0.254196 [2023-11-25 02:29:26,330][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.56 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 5.05 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: as part of the big city plan, digbeth is undergoing a large redevelopment scheme that ### ### will regenerate the old industrial buildings into apartments, retail premises, offices and arts ### ### facilities. ### ### [P_TEXT]: big city plan on a site at the end of digbeth high street, directly opposite birmingham's ### ### already iconic selfridges building ( it will occupy the site of the city's last seven burgage plots ### ### ). complete with state - of - the - art environmental and energy systems, the final development is ### ### likely to consist of a marriott hotel, apartments, green spaces, a niche shopping centre, a public ### ### space intended to house birmingham's historic john f. kennedy memorial and two refurbished cold ### ### storage buildings which will be redeveloped as creative industries business incubation and ### ### innovation hubs. planning permission was granted for the beorma ### ### =============================================== v_q ================================================ ### ### ('dig', 0, 8) ('##bet', 1, 1) ('big', 2, 3) ('##h', 3, 61) ('plan', 4, 6) ('redevelopment', 5, 195) ### ### ('##ه', 6, 94) ('industrial', 7, 74) ('city', 8, 14) ('.', 9, 154) ('renovation', 10, 483) ### ### ('buildings', 11, 44) ('scheme', 12, 190) ('planning', 13, 41) ('building', 14, 27) ### ### ('tenants', 15, 10037) ('dug', 16, 107) ('##ہ', 17, 1000) ('apartment', 18, 57) ### ### ('housing', 19, 273) ### ### =============================================== v_p ================================================ ### ### ('marriott', 14033, 0) ('##bet', 1, 1) ('kennedy', 4905, 2) ('big', 2, 3) ('birmingham', 31, 4) ### ### ('##ridge', 25771, 5) ('plan', 4, 6) ('##uba', 26551, 7) ('dig', 0, 8) ('niche', 8308, 9) ### ### ('##rga', 28826, 10) ('innovation', 1905, 11) ('creative', 894, 12) ('storage', 9087, 13) ### ### ('city', 8, 14) ('self', 14817, 15) ('##or', 20140, 16) ('memorial', 1338, 17) ### ### ('environmental', 1022, 18) ('cold', 6659, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##bet', 1, 1) ('dig', 0, 8) ('big', 2, 3) ('plan', 4, 6) ('city', 8, 14) ('birmingham', 31, 4) ### ### ('building', 14, 27) ('##h', 3, 61) ('planning', 13, 41) ('buildings', 11, 44) ('shopping', 72, 20) ### ### ('apartments', 39, 40) ('centre', 20, 46) ('cities', 25, 48) ('art', 84, 38) ('industrial', 7, 74) ### ### ('development', 48, 45) ('hotel', 123, 34) ('apartment', 18, 57) ('##ه', 6, 94) ### ############################################################################################################ [2023-11-25 02:29:26,831][root][INFO] - Epoch: 1: Step: 11301/19530, loss[v]=0.114661, lr=0.000012, acc@1[1]=993.0/1024=0.9697265625, acc@1[2]=1011.5/1024=0.98779296875 [2023-11-25 02:31:17,275][root][INFO] - Train batch 11400 [2023-11-25 02:31:17,276][root][INFO] - Avg. loss per last 100 batches: 0.237007 [2023-11-25 02:31:18,005][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.57 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 5.04 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: on 1 june in that year, the broad gauge lostwithiel and fowey railway was opened to new ### ### jetties situated above carne point, and in 1873, the standard gauge cornwall minerals railway ( cmr ### ### ) opened a line from newquay and par to further jetties between caffa mill pill and carne point. ### ### [P_TEXT]: fowey railway station fowey railway station was a station in fowey, cornwall from 1874 ### ### until 1965. the rail connection to the docks at carne point remains open for china clay traffic. ### ### the lostwithiel and fowey railway ( l & fr ) had opened as far as carne point in 1869 but was never ### ### completed to the intended terminus at fowey. the cornwall minerals railway ( cmr ) arrived from the ### ### opposite direction in 1874. a passenger service from fowey to via started on 1 june 1874. the ### ### station had two platforms with loading docks and a goods shed at the st ### ### =============================================== v_q ================================================ ### ### ('jett', 0, 3395) ('pill', 1, 19505) ('car', 2, 15) ('cornwall', 3, 1) ('##owe', 4, 0) ### ### ('##quay', 5, 8457) ('##with', 6, 4) ('point', 7, 9) ('railway', 8, 14) ('##fa', 9, 13832) ### ### ('##iel', 10, 8) ('caf', 11, 28501) ('##ne', 12, 17) ('par', 13, 13618) ('gauge', 14, 7839) ### ### ('cornish', 15, 29) ('mill', 16, 5718) ('cm', 17, 3) ('.', 18, 272) ('1873', 19, 53) ### ### =============================================== v_p ================================================ ### ### ('##owe', 4, 0) ('cornwall', 3, 1) ('clay', 6760, 2) ('cm', 17, 3) ('##with', 6, 4) ### ### ('china', 1514, 5) ('minerals', 28, 6) ('docks', 4910, 7) ('##iel', 10, 8) ('point', 7, 9) ### ### ('f', 39, 10) ('lost', 20, 11) ('loading', 10014, 12) ('shed', 7124, 13) ('railway', 8, 14) ### ### ('car', 2, 15) ('connection', 1690, 16) ('##ne', 12, 17) ('remains', 2818, 18) ('goods', 1459, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##owe', 4, 0) ('cornwall', 3, 1) ('##with', 6, 4) ('car', 2, 15) ('point', 7, 9) ('cm', 17, 3) ### ### ('##iel', 10, 8) ('railway', 8, 14) ('##ne', 12, 17) ('minerals', 28, 6) ('lost', 20, 11) ### ### ('cornish', 15, 29) ('f', 39, 10) ('##y', 43, 21) ('station', 38, 26) ('rail', 34, 32) ### ### ('1873', 19, 53) ('mineral', 94, 22) ('newly', 21, 50) ('1872', 37, 40) ### ############################################################################################################ [2023-11-25 02:31:18,509][root][INFO] - Epoch: 1: Step: 11401/19530, loss[v]=0.325818, lr=0.000012, acc@1[1]=954.5/1024=0.93212890625, acc@1[2]=979.0/1024=0.9560546875 [2023-11-25 02:33:09,166][root][INFO] - Train batch 11500 [2023-11-25 02:33:09,167][root][INFO] - Avg. loss per last 100 batches: 0.249955 [2023-11-25 02:33:09,904][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.65 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 4.87 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the puget sound region is a coastal area of the pacific northwest in the u. s. state of ### ### washington, including puget sound, the puget sound lowlands, and the surrounding region roughly ### ### west of the cascade range and east of the olympic mountains. ### ### [P_TEXT]: puget sound salish sea was established by the united states board on geographic names as ### ### the collective waters of puget sound, the strait of juan de fuca, and the strait of georgia. ### ### sometimes the terms " puget sound " and " puget sound and adjacent waters " are used for not only ### ### puget sound proper but also for waters to the north, such as bellingham bay and the san juan ### ### islands region. the term " puget sound " is used not just for the body of water but also the puget ### ### sound region centered on the sound. major cities on the sound include ### ### =============================================== v_q ================================================ ### ### ('puget', 0, 0) ('sound', 1, 3) ('olympic', 2, 112) ('region', 3, 27) ('washington', 4, 1254) ### ### ('seattle', 5, 99) ('.', 6, 653) ('oregon', 7, 108) ('lowlands', 8, 10900) ('mountains', 9, 1777) ### ### ('pacific', 10, 141) ('coastal', 11, 1813) ('tacoma', 12, 47) ('california', 13, 199) ### ### ('northwest', 14, 3628) ('range', 15, 1896) ('olympics', 16, 15629) ('dakota', 17, 12786) ### ### ('regions', 18, 26) ('montana', 19, 8984) ### ### =============================================== v_p ================================================ ### ### ('puget', 0, 0) ('juan', 2357, 1) ('board', 952, 2) ('sound', 1, 3) ('proper', 5557, 4) ### ### ('collective', 2775, 5) ('fu', 15624, 6) ('sal', 7219, 7) ('##ingham', 1723, 8) ('islands', 191, 9) ### ### ('strait', 943, 10) ('##ish', 23973, 11) ('boards', 18488, 12) ('georgia', 94, 13) ### ### ('water', 1888, 14) ('##ca', 15876, 15) ('body', 795, 16) ('cities', 1540, 17) ('bell', 5918, 18) ### ### ('san', 284, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('puget', 0, 0) ('sound', 1, 3) ('region', 3, 27) ('regions', 18, 26) ('olympic', 2, 112) ### ### ('states', 20, 31) ('georgia', 94, 13) ('islands', 191, 9) ('seattle', 5, 99) ('tacoma', 12, 47) ### ### ('board', 952, 2) ('bay', 81, 23) ('oregon', 7, 108) ('sea', 155, 20) ('san', 284, 19) ### ### ('juan', 2357, 1) ('pacific', 10, 141) ('strait', 943, 10) ('##ingham', 1723, 8) ('body', 795, 16) ### ############################################################################################################ [2023-11-25 02:33:10,405][root][INFO] - Epoch: 1: Step: 11501/19530, loss[v]=0.130673, lr=0.000012, acc@1[1]=985.0/1024=0.9619140625, acc@1[2]=1009.0/1024=0.9853515625 [2023-11-25 02:35:00,634][root][INFO] - Train batch 11600 [2023-11-25 02:35:00,635][root][INFO] - Avg. loss per last 100 batches: 0.232781 [2023-11-25 02:35:01,361][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.63 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 4.89 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: guitarist statham collaborated with ex - bauhaus singer peter murphy, successfully ### ### participating as part of murphy's solo backing band, the hundred men. ### ### [P_TEXT]: peter murphy ( musician ) " final solution " that made a minor splash on the club scene. ### ### the followup, " love hysteria ", was not successful in the uk although in the us it performed ### ### better than his previous solo releases. the album also marked the beginning of a long - term ### ### collaboration with songwriter paul statham from b - movie, who co - wrote songs with murphy until ### ### 1995. the resulting singles " all night long " and " indigo eyes " helped garner a wider following, ### ### and the black - and - white video for " all night long " entered rotation on mtv. the pinnacle of ### ### murphy's solo popularity in the ### ### =============================================== v_q ================================================ ### ### ('stat', 0, 1) ('murphy', 1, 0) ('##aus', 2, 27127) ('hundred', 3, 15687) ('##uh', 4, 25295) ### ### ('##ham', 5, 16) ('peter', 6, 5) ('men', 7, 7735) ('band', 8, 74) ('ba', 9, 16111) ### ### ('album', 10, 33) ('thousand', 11, 13072) ('frontman', 12, 41) ('bands', 13, 56) ('.', 14, 556) ### ### ('artists', 15, 240) ('berlin', 16, 2394) ('beatles', 17, 1310) ('hendrix', 18, 9316) ### ### ('punk', 19, 987) ### ### =============================================== v_p ================================================ ### ### ('murphy', 1, 0) ('stat', 0, 1) ('indigo', 1005, 2) ('solution', 10598, 3) ('pinnacle', 14647, 4) ### ### ('peter', 6, 5) ('eyes', 9168, 6) ('hysteria', 18895, 7) ('night', 437, 8) ('long', 7561, 9) ### ### ('mtv', 335, 10) ('final', 2878, 11) ('movie', 7180, 12) ('follow', 2606, 13) ### ### ('solutions', 18960, 14) ('rotation', 22092, 15) ('##ham', 5, 16) ('##up', 1838, 17) ### ### ('paul', 265, 18) ('film', 844, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('murphy', 1, 0) ('stat', 0, 1) ('peter', 6, 5) ('##ham', 5, 16) ('album', 10, 33) ### ### ('frontman', 12, 41) ('indigo', 1005, 2) ('club', 78, 20) ('band', 8, 74) ('bands', 13, 56) ### ### ('mtv', 335, 10) ('night', 437, 8) ('video', 49, 28) ('solo', 24, 37) ('paul', 265, 18) ### ### ('scene', 87, 27) ('musician', 83, 38) ('film', 844, 19) ('musicians', 84, 40) ('follow', 2606, 13) ### ############################################################################################################ [2023-11-25 02:35:01,862][root][INFO] - Epoch: 1: Step: 11601/19530, loss[v]=0.129550, lr=0.000012, acc@1[1]=988.5/1024=0.96533203125, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 02:36:52,393][root][INFO] - Train batch 11700 [2023-11-25 02:36:52,393][root][INFO] - Avg. loss per last 100 batches: 0.240265 [2023-11-25 02:36:53,131][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.63 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 4.73 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the gambler was gunned down in front of the hotel metropole in times square on july 16, ### ### 1912, two days after his story appeared in the newspapers. ### ### [P_TEXT]: sam schepps samuel schepps (? – 12 january 1936 ), also known as schapps, was a new york ### ### city mobster with the monk eastman gang. schepps ran gambling houses for jack zelig as part of his ### ### work with the gang. on july 16, 1912, herman rosenthal was murdered in front of the hotel metropole ### ### in new york city. schepps was identified by coconspirators as the " paymaster " of the plot to kill ### ### rosenthal. within the week, schepps had disappeared and was believed to have left new york. he was ### ### tracked down and, on ### ### =============================================== v_q ================================================ ### ### ('gamble', 0, 248) ('square', 1, 6677) ('times', 2, 11637) ('hotel', 3, 15) ('metro', 4, 12) ### ### ('##pole', 5, 3) ('1912', 6, 19) ('gunn', 7, 11562) ('gambling', 8, 11) ('.', 9, 260) ### ### ('story', 10, 1177) ('newspaper', 11, 7941) ('1913', 12, 41) ('1911', 13, 37) ('1914', 14, 55) ### ### ('july', 15, 155) ('restaurant', 16, 1007) ('stories', 17, 11368) ('plaza', 18, 4697) ### ### ('down', 19, 213) ### ### =============================================== v_p ================================================ ### ### ('rosenthal', 9297, 0) ('eastman', 21525, 1) ('##pps', 27750, 2) ('##pole', 5, 3) ### ### ('herman', 113, 4) ('coco', 20245, 5) ('monk', 6539, 6) ('##master', 9969, 7) ('##he', 23817, 8) ### ### ('ze', 21803, 9) ('pay', 10446, 10) ('gambling', 8, 11) ('metro', 4, 12) ('sam', 10447, 13) ### ### ('plot', 344, 14) ('hotel', 3, 15) ('samuel', 99, 16) ('mob', 249, 17) ('gang', 4877, 18) ### ### ('1912', 6, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##pole', 5, 3) ('metro', 4, 12) ('hotel', 3, 15) ('1912', 6, 19) ('gambling', 8, 11) ### ### ('herman', 113, 4) ('1911', 13, 37) ('samuel', 99, 16) ('plot', 344, 14) ('mob', 249, 17) ### ### ('1913', 12, 41) ('gamble', 0, 248) ('hotels', 20, 39) ('kill', 97, 27) ('1914', 14, 55) ### ### ('rosenthal', 9297, 0) ('murdered', 781, 29) ('monk', 6539, 6) ('1910', 24, 59) ('york', 575, 35) ### ############################################################################################################ [2023-11-25 02:36:53,635][root][INFO] - Epoch: 1: Step: 11701/19530, loss[v]=0.257849, lr=0.000012, acc@1[1]=958.5/1024=0.93603515625, acc@1[2]=989.0/1024=0.9658203125 [2023-11-25 02:38:44,219][root][INFO] - Train batch 11800 [2023-11-25 02:38:44,220][root][INFO] - Avg. loss per last 100 batches: 0.239983 [2023-11-25 02:38:44,943][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 4.98 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he was also a former deputy governor of the hungarian national bank. ### ### [P_TEXT]: gyula szapary lorinc szapary ). his great - grandchild is gyorgy szapary economist, ### ### former deputy governor of the hungarian national bank and ambassador to the united states from ### ### january 2011 to january 2015. he studied law, entered the government service and became viscount ( ### ### " vicecomes " ) of heves county. he spent nine legislative sessions for the liberal party as a ### ### representative in the diet of hungary. he was elected count ( " comes " ) of heves in 1867. in 1870 ### ### he became secretary to the transport ministry, interior minister in 1873, finance minister from ### ### 1878 to 1887 ### ### =============================================== v_q ================================================ ### ### ('bank', 0, 11) ('hungarian', 1, 17) ('governor', 2, 8) ('hungary', 3, 14) ('deputy', 4, 31) ### ### ('national', 5, 24) ('budapest', 6, 106) ('.', 7, 244) ('banks', 8, 68) ('##ʂ', 9, 1210) ### ### ('nacional', 10, 188) ('governors', 11, 50) ('ernest', 12, 1935) ('##bank', 13, 567) ### ### ('hungarians', 14, 155) ('vienna', 15, 651) ('president', 16, 1641) ('##ᵉ', 17, 1468) ### ### ('romania', 18, 507) ('banking', 19, 85) ### ### =============================================== v_p ================================================ ### ### ('##com', 25872, 0) ('##ves', 21100, 1) ('comes', 5036, 2) ('##yu', 14718, 3) ('##par', 7570, 4) ### ### ('diet', 3369, 5) ('viscount', 1628, 6) ('count', 3124, 7) ('governor', 2, 8) ('finance', 163, 9) ### ### ('ambassador', 11174, 10) ('bank', 0, 11) ('liberal', 1858, 12) ('transport', 3800, 13) ### ### ('hungary', 3, 14) ('lori', 21568, 15) ('vice', 1151, 16) ('hungarian', 1, 17) ### ### ('interior', 1429, 18) ('ministry', 10934, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('bank', 0, 11) ('hungarian', 1, 17) ('governor', 2, 8) ('hungary', 3, 14) ('national', 5, 24) ### ### ('deputy', 4, 31) ('governors', 11, 50) ('finance', 163, 9) ('budapest', 6, 106) ('banks', 8, 68) ### ### ('viscount', 1628, 6) ('legislative', 141, 44) ('vice', 1151, 16) ('liberal', 1858, 12) ### ### ('minister', 312, 35) ('.', 7, 244) ('interior', 1429, 18) ('representatives', 166, 52) ### ### ('count', 3124, 7) ('diet', 3369, 5) ### ############################################################################################################ [2023-11-25 02:38:45,443][root][INFO] - Epoch: 1: Step: 11801/19530, loss[v]=0.374284, lr=0.000012, acc@1[1]=949.0/1024=0.9267578125, acc@1[2]=974.5/1024=0.95166015625 [2023-11-25 02:40:35,888][root][INFO] - Train batch 11900 [2023-11-25 02:40:35,889][root][INFO] - Avg. loss per last 100 batches: 0.234003 [2023-11-25 02:40:36,620][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.68 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 5.07 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: located in the city of maceio, the federal university of alagoas (, ufal or ufal ) is the ### ### major university in coastal state alagoas and one of the main research centers in brazilian north ### ### eastern region ( one of the five regions of brazil ). ### ### [P_TEXT]: benedito de lira benedito de lira ( born may 1, 1942 ) is a brazil ian politician from ### ### alagoas, member of progressistas ( pp ), former senator for his state and incumbent mayor of barra ### ### de sao miguel. lira is bachelor of laws by the federal university of alagoas ( ufal ). he began his ### ### political career as city councillor in his hometown between 1966 and 1970, and later as city ### ### councillor of maceio for two consecutive terms, between 1973 and 1982. he was state deputy from ### ### 1983 to 1994. in 2002, lira ### ### =============================================== v_q ================================================ ### ### ('mace', 0, 0) ('ala', 1, 6) ('##fa', 2, 7) ('##io', 3, 23) ('brazil', 4, 4) ('university', 5, 25) ### ### ('##as', 6, 20) ('##l', 7, 39) ('brazilian', 8, 42) ('##go', 9, 9) ('eastern', 10, 13179) ### ### ('federal', 11, 12) ('coastal', 12, 6629) ('##ل', 13, 47) ('u', 14, 30) ('region', 15, 8747) ### ### ('##ァ', 16, 149) ('.', 17, 572) ('universities', 18, 44) ('state', 19, 32) ### ### =============================================== v_p ================================================ ### ### ('mace', 0, 0) ('bachelor', 15489, 1) ('barr', 12572, 2) ('progress', 12210, 3) ('brazil', 4, 4) ### ### ('miguel', 630, 5) ('ala', 1, 6) ('##fa', 2, 7) ('li', 11976, 8) ('##go', 9, 9) ('ian', 1372, 10) ### ### ('##ista', 23435, 11) ('federal', 11, 12) ('councillor', 5070, 13) ('pp', 9916, 14) ### ### ('##ito', 7605, 15) ('senator', 17859, 16) ('senate', 3121, 17) ('sao', 3309, 18) ### ### ('laws', 8294, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('mace', 0, 0) ('ala', 1, 6) ('##fa', 2, 7) ('brazil', 4, 4) ('##io', 3, 23) ('##as', 6, 20) ### ### ('##go', 9, 9) ('university', 5, 25) ('federal', 11, 12) ('u', 14, 30) ('##l', 7, 39) ### ### ('brazilian', 8, 42) ('##ل', 13, 47) ('state', 19, 32) ('city', 24, 34) ('universities', 18, 44) ### ### ('miguel', 630, 5) ('cities', 32, 45) ('mayor', 545, 21) ('ian', 1372, 10) ### ############################################################################################################ [2023-11-25 02:40:37,125][root][INFO] - Epoch: 1: Step: 11901/19530, loss[v]=0.259358, lr=0.000012, acc@1[1]=956.5/1024=0.93408203125, acc@1[2]=988.0/1024=0.96484375 [2023-11-25 02:42:27,565][root][INFO] - Train batch 12000 [2023-11-25 02:42:27,566][root][INFO] - Avg. loss per last 100 batches: 0.241007 [2023-11-25 02:42:28,311][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.60 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 5.07 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: again besieged and captured by the french during the thirty years'war in september 1642, ### ### perpignan was formally ceded by spain 17 years later in the treaty of the pyrenees, and from then ### ### on remained a french possession. ### ### [P_TEXT]: charles de la porte charles de la porte ( 1602 in paris – 8 february 1664 in paris ) was ### ### a french nobleman and general. he was marquis then duke of la meilleraye, duke of rethel and peer ### ### of france, baron of parthenay and of saint - maixent, count of secondigny, seigneur of le boisliet, ### ### la lunardiere, la jobeliniere and villeneuve. in 1639 he became marshal of france after taking ### ### hesdin. in 1642, after a ten - month siege, he conquered perpignan and salses - le - chateau, ### ### completing the ### ### =============================================== v_q ================================================ ### ### ('##gnan', 0, 17) ('pyrenees', 1, 23090) ('thirty', 2, 8016) ('##pi', 3, 25) ('treaty', 4, 4101) ### ### ('per', 5, 53) ('spain', 6, 985) ('1642', 7, 38) ('.', 8, 1803) ('french', 9, 50) ('war', 10, 573) ### ### ('spanish', 11, 1739) ('commune', 12, 22046) ('france', 13, 30) ('ɡ', 14, 184) ('years', 15, 20456) ### ### ('conquest', 16, 73) ('napoleon', 17, 63) ('siege', 18, 9) ('toulouse', 19, 145) ### ### =============================================== v_p ================================================ ### ### ('porte', 9471, 0) ('##igny', 22717, 1) ('##the', 11611, 2) ('lunar', 16811, 3) ('##hen', 14425, 4) ### ### ('##sd', 28520, 5) ('##neuve', 1038, 6) ('##ses', 7824, 7) ('bois', 24058, 8) ('siege', 18, 9) ### ### ('##lie', 18884, 10) ('##elin', 16629, 11) ('##ay', 19726, 12) ('marshal', 5196, 13) ### ### ('count', 1417, 14) ('peer', 27660, 15) ('160', 1494, 16) ('##gnan', 0, 17) ('re', 23929, 18) ### ### ('##ller', 21713, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##gnan', 0, 17) ('##pi', 3, 25) ('1642', 7, 38) ('per', 5, 53) ('siege', 18, 9) ### ### ('france', 13, 30) ('french', 9, 50) ('paris', 27, 36) ('##neuve', 1038, 6) ('conquered', 24, 57) ### ### ('charles', 568, 23) ('napoleon', 17, 63) ('conquest', 16, 73) ('chateau', 556, 35) ### ### ('163', 937, 26) ('duke', 642, 33) ('count', 1417, 14) ('160', 1494, 16) ('besieged', 32, 62) ### ### ('nobility', 315, 47) ### ############################################################################################################ [2023-11-25 02:42:28,812][root][INFO] - Epoch: 1: Step: 12001/19530, loss[v]=0.365692, lr=0.000012, acc@1[1]=947.0/1024=0.9248046875, acc@1[2]=971.5/1024=0.94873046875 [2023-11-25 02:44:19,180][root][INFO] - Train batch 12100 [2023-11-25 02:44:19,180][root][INFO] - Avg. loss per last 100 batches: 0.228492 [2023-11-25 02:44:19,903][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.05 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.03 | max: 5.09 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: " kharaharapriya's " notes when shifted using " graha bhedam ", yields 5 other major " ### ### melakarta " ragams, namely, " kalyani ", " hanumatodi ", " natabhairavi ", " shankarabharanam " and ### ### " harikambhoji ". ### ### [P_TEXT]: natabhairavi this section covers the theoretical and scientific aspect of this ragam. " ### ### natabhairavi's " notes when shifted using " graha bhedam ", yields 5 other major " melakarta " ### ### ragams, namely, " kalyani ", " sankarabharanam ", " hanumatodi ", " kharaharapriya " and " ### ### harikambhoji ". for further details and an illustration of graha bhedam of this ragam refer graha ### ### bhedam on sankarabharanam. even though natabhairavi has quite evenly spaced " swara sthanas " ( ### ### pitch positions, notes ) like the other 5 in this group, it has not found as much importance in ### ### concerts. one is likely to find " kalyani ", " todi ", " sankarabharanam " ### ### =============================================== v_q ================================================ ### ### ('##hair', 0, 0) ('##aha', 1, 14) ('##hara', 2, 15) ('##oj', 3, 13) ('han', 4, 22) ('##ly', 5, 42) ### ### ('hari', 6, 21) ('##ab', 7, 16) ('##iya', 8, 35) ('##bh', 9, 72) ('shankar', 10, 17785) ### ### ('##hed', 11, 11) ('##kam', 12, 18) ('##ani', 13, 24) ('nat', 14, 5) ('##nam', 15, 27409) ### ### ('rag', 16, 7) ('##pr', 17, 45) ('##aka', 18, 44) ('##rta', 19, 30) ### ### =============================================== v_p ================================================ ### ### ('##hair', 0, 0) ('pitch', 3658, 1) ('##hana', 10435, 2) ('##avi', 47, 3) ('tod', 17595, 4) ### ### ('nat', 14, 5) ('sank', 7024, 6) ('rag', 16, 7) ('notes', 44, 8) ('##rana', 17943, 9) ### ### ('##uma', 20, 10) ('##hed', 11, 11) ('spaced', 21810, 12) ('##oj', 3, 13) ('##aha', 1, 14) ### ### ('##hara', 2, 15) ('##ab', 7, 16) ('refer', 15500, 17) ('##kam', 12, 18) ('positions', 7225, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##hair', 0, 0) ('##aha', 1, 14) ('nat', 14, 5) ('##hara', 2, 15) ('##oj', 3, 13) ('rag', 16, 7) ### ### ('##hed', 11, 11) ('##ab', 7, 16) ('han', 4, 22) ('hari', 6, 21) ('##kam', 12, 18) ### ### ('##uma', 20, 10) ('##ani', 13, 24) ('##iya', 8, 35) ('##ly', 5, 42) ('shifted', 23, 23) ### ### ('##rta', 19, 30) ('##avi', 47, 3) ('##am', 26, 25) ('notes', 44, 8) ### ############################################################################################################ [2023-11-25 02:44:20,404][root][INFO] - Epoch: 1: Step: 12101/19530, loss[v]=0.124991, lr=0.000012, acc@1[1]=990.0/1024=0.966796875, acc@1[2]=1010.5/1024=0.98681640625 [2023-11-25 02:46:10,414][root][INFO] - Train batch 12200 [2023-11-25 02:46:10,414][root][INFO] - Avg. loss per last 100 batches: 0.243395 [2023-11-25 02:46:11,156][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.60 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 4.89 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: throughout his six years of reign, he faced an active rebellion by sithu kyawhtin, who ### ### was supported by the shan state of mohnyin. ### ### [P_TEXT]: sithu kyawhtin sithu kyawhtin (, ; also known as narapati sithu (, ) ) was the last king ### ### of ava from 1551 to 1555. he came to power by overthrowing king narapati iii in 1551, the ### ### culmination of his six - year rebellion ( 1545 – 1551 ) with the support of the shan state of ### ### mohnyin. as king, he was able to get all members of the confederation of shan states to unite ### ### against an impending invasion by king bayinnaung of toungoo dynasty. he sought a peace treaty with ### ### bayinnaung but was rebuffed. the ava - led confederation withstood the initial invasion ### ### =============================================== v_q ================================================ ### ### ('##hn', 0, 5) ('sit', 1, 3) ('shan', 2, 8) ('myanmar', 3, 613) ('##ht', 4, 13) ('burma', 5, 1481) ### ### ('##yin', 6, 6) ('reign', 7, 1974) ('##aw', 8, 28) ('taiwan', 9, 673) ('mo', 10, 30) ### ### ('##hu', 11, 17) ('burmese', 12, 909) ('state', 13, 47) ('ky', 14, 22) ('thailand', 15, 489) ### ### ('king', 16, 18) ('emperor', 17, 168) ('.', 18, 763) ('kingdom', 19, 160) ### ### =============================================== v_p ================================================ ### ### ('##inn', 16415, 0) ('confederation', 61, 1) ('##pati', 8911, 2) ('sit', 1, 3) ('ava', 626, 4) ### ### ('##hn', 0, 5) ('##yin', 6, 6) ('nara', 20077, 7) ('shan', 2, 8) ('##ffed', 24014, 9) ### ### ('invasion', 224, 10) ('155', 10279, 11) ('154', 11308, 12) ('##ht', 4, 13) ('impending', 2033, 14) ### ### ('treaty', 1341, 15) ('bay', 9897, 16) ('##hu', 11, 17) ('king', 16, 18) ('rebellion', 24, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##hn', 0, 5) ('sit', 1, 3) ('shan', 2, 8) ('##yin', 6, 6) ('##ht', 4, 13) ('##hu', 11, 17) ### ### ('king', 16, 18) ('ky', 14, 22) ('##aw', 8, 28) ('mo', 10, 30) ('rebellion', 24, 19) ### ### ('confederation', 61, 1) ('dynasty', 23, 23) ('state', 13, 47) ('invasion', 224, 10) ### ### ('ava', 626, 4) ('##in', 38, 43) ('states', 66, 37) ('myanmar', 3, 613) ('treaty', 1341, 15) ### ############################################################################################################ [2023-11-25 02:46:11,661][root][INFO] - Epoch: 1: Step: 12201/19530, loss[v]=0.148987, lr=0.000012, acc@1[1]=987.0/1024=0.9638671875, acc@1[2]=1006.0/1024=0.982421875 [2023-11-25 02:48:01,682][root][INFO] - Train batch 12300 [2023-11-25 02:48:01,683][root][INFO] - Avg. loss per last 100 batches: 0.233647 [2023-11-25 02:48:02,417][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.74 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=99.99% | mean: 0.03 | max: 4.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.3/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: although it released after their second outing together, " guru sishyan " ( 1988 ). ### ### [P_TEXT]: rajinikanth as " naan sigappu manithan " ( 1985 ), " padikkathavan " ( 1985 ), " mr. ### ### bharath " ( 1986 ), " velaikaran " ( 1987 ), " guru sishyan " ( 1988 ) and " dharmathin thalaivan " ### ### ( 1988 ). in 1988, he made his only american film appearance in " bloodstone ", directed by dwight ### ### little, in which he played an english - speaking indian taxi driver. rajinikanth finished the ### ### decade with films including " rajadhi raja ", " siva ", " raja chinna roja " and " mappillai " ### ### while also starring in a few bollywood productions. " raja chinna roja " was the first indian film ### ### to feature ### ### =============================================== v_q ================================================ ### ### ('guru', 0, 12) ('sis', 1, 15) ('##hya', 2, 21) ('1988', 3, 18) ('##n', 4, 786) ('##ன', 5, 125) ### ### ('film', 6, 23) ('.', 7, 2654) ('album', 8, 1047) ('band', 9, 3591) ('1989', 10, 60) ### ### ('bands', 11, 542) ('##ん', 12, 540) ('1987', 13, 51) ('##न', 14, 194) ('##ि', 15, 87) ### ### ('beatles', 16, 11046) ('tamil', 17, 2155) ('artists', 18, 2570) ('starr', 19, 25648) ### ### =============================================== v_p ================================================ ### ### ('##stone', 10632, 0) ('dwight', 22011, 1) ('chin', 7704, 2) ('raj', 154, 3) ('taxi', 13447, 4) ### ### ('little', 12094, 5) ('blood', 1756, 6) ('raja', 315, 7) ('##kan', 16992, 8) ('##harat', 14589, 9) ### ### ('##ini', 14638, 10) ('##thi', 2295, 11) ('guru', 0, 12) ('##lai', 29248, 13) ('dharma', 752, 14) ### ### ('sis', 1, 15) ('##kara', 13766, 16) ('map', 1620, 17) ('1988', 3, 18) ('##havan', 11490, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('guru', 0, 12) ('sis', 1, 15) ('##hya', 2, 21) ('1988', 3, 18) ('film', 6, 23) ('raj', 154, 3) ### ### ('films', 37, 25) ('bollywood', 46, 20) ('raja', 315, 7) ('1987', 13, 51) ('1989', 10, 60) ### ### ('india', 24, 48) ('##ன', 5, 125) ('dharma', 752, 14) ('decade', 449, 24) ('blood', 1756, 6) ### ### ('indian', 113, 42) ('##thi', 2295, 11) ('map', 1620, 17) ('productions', 611, 31) ### ############################################################################################################ [2023-11-25 02:48:02,918][root][INFO] - Epoch: 1: Step: 12301/19530, loss[v]=0.392582, lr=0.000013, acc@1[1]=942.5/1024=0.92041015625, acc@1[2]=971.5/1024=0.94873046875 [2023-11-25 02:49:52,873][root][INFO] - Train batch 12400 [2023-11-25 02:49:52,873][root][INFO] - Avg. loss per last 100 batches: 0.226101 [2023-11-25 02:49:53,596][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.82 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 5.21 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: priceless is the fourth album by american singer kelly price. ### ### [P_TEXT]: kelly price of shirley murdock's 1986 hit. a remix of " love sets you free " was issued a ### ### single in 2000, produced by teddy riley and starring a number of fellow r & b performers. price was ### ### featured in the 2003 film, " bringing down the house ", performing a cover of the 1983 rufus & ### ### chaka khan hit " ain't nobody ". a christmas album, " one family ", was issued in 2001 which had a ### ### piano appearance from gospeljazz legend, ben tankard. two years later, price returned with her ### ### third regular studio album, " priceless ". in october 2005, price recorded her first live gospel ### ### album. the " this is who i ### ### =============================================== v_q ================================================ ### ### ('price', 0, 0) ('##less', 1, 6) ('kelly', 2, 1) ('album', 3, 37) ('##lessly', 4, 95) ### ### ('prices', 5, 78) ('fourth', 6, 1189) ('albums', 7, 63) ('singer', 8, 390) ('##lessness', 9, 242) ### ### ('.', 10, 2062) ('rapper', 11, 1324) ('singers', 12, 94) ('artists', 13, 190) ('song', 14, 90) ### ### ('lp', 15, 834) ('trade', 16, 1972) ('4th', 17, 2429) ('oregon', 18, 7162) ('label', 19, 187) ### ### =============================================== v_p ================================================ ### ### ('price', 0, 0) ('kelly', 2, 1) ('riley', 3095, 2) ('nobody', 4506, 3) ('teddy', 10344, 4) ### ### ('rufus', 3817, 5) ('##less', 1, 6) ('free', 298, 7) ('tank', 637, 8) ('ain', 13573, 9) ### ### ('khan', 7189, 10) ('##ard', 8652, 11) ('gospel', 1651, 12) ('##rdo', 19736, 13) ### ### ('family', 2737, 14) ('1983', 6854, 15) ('christmas', 2535, 16) ('shirley', 6760, 17) ### ### ('ben', 9545, 18) ('##zz', 24096, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('price', 0, 0) ('kelly', 2, 1) ('##less', 1, 6) ('album', 3, 37) ('prices', 5, 78) ### ### ('##lessly', 4, 95) ('albums', 7, 63) ('free', 298, 7) ('tank', 637, 8) ('riley', 3095, 2) ### ### ('nobody', 4506, 3) ('gospel', 1651, 12) ('rufus', 3817, 5) ('piano', 568, 29) ('legend', 1250, 20) ### ### ('ones', 433, 31) ('family', 2737, 14) ('singers', 12, 94) ('hits', 132, 45) ### ### ('christmas', 2535, 16) ### ############################################################################################################ [2023-11-25 02:49:54,099][root][INFO] - Epoch: 1: Step: 12401/19530, loss[v]=0.102508, lr=0.000013, acc@1[1]=991.5/1024=0.96826171875, acc@1[2]=1012.0/1024=0.98828125 [2023-11-25 02:51:44,161][root][INFO] - Train batch 12500 [2023-11-25 02:51:44,162][root][INFO] - Avg. loss per last 100 batches: 0.235590 [2023-11-25 02:51:44,882][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.59 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 4.95 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in march 2012, algoma announced that they would be joining the oua in 2013. ### ### [P_TEXT]: northern ontario and the tournament of hearts separately from its province. hockey is ### ### often played on artificial outdoor rinks, and sometimes on frozen lakes. the north bay battalion, ### ### sault ste. marie greyhounds and sudbury wolves play in the ontario hockey league. the algoma ### ### thunderbirds, lakehead thunderwolves, laurentian voyageurs, and nipissing lakers compete in u ### ### sports as members of ontario university athletics. also, the thunder bay chill soccer teams play in ### ### north america's usl league two. northern ontario has hosted the 1981 canada summer games, 1988 ### ### world junior championships in athletics ### ### =============================================== v_q ================================================ ### ### ('ou', 0, 20659) ('##go', 1, 26) ('al', 2, 46) ('##ma', 3, 48) ('2013', 4, 2402) ('2012', 5, 5048) ### ### ('football', 6, 164) ('league', 7, 50) ('##a', 8, 24416) ('signing', 9, 4220) ('.', 10, 121) ### ### ('##コ', 11, 124) ('rugby', 12, 86) ('commune', 13, 23898) ('soccer', 14, 49) ('fc', 15, 3968) ### ### ('signed', 16, 12146) ('federation', 17, 6477) ('burkina', 18, 12052) ('club', 19, 1045) ### ### =============================================== v_p ================================================ ### ### ('northern', 3218, 0) ('ontario', 458, 1) ('greyhound', 5665, 2) ('battalion', 4119, 3) ### ### ('chill', 12884, 4) ('usl', 7090, 5) ('wolves', 1265, 6) ('frozen', 5512, 7) ('hearts', 26856, 8) ### ### ('lakers', 6393, 9) ('##birds', 6373, 10) ('lakes', 25321, 11) ('saul', 14547, 12) ('u', 1594, 13) ### ### ('university', 81, 14) ('athletics', 699, 15) ('artificial', 2019, 16) ('thunder', 9635, 17) ### ### ('junior', 1724, 18) ('separately', 780, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##go', 1, 26) ('al', 2, 46) ('##ma', 3, 48) ('league', 7, 50) ('university', 81, 14) ### ### ('soccer', 14, 49) ('tournament', 51, 25) ('hockey', 71, 21) ('ontario', 458, 1) ('sports', 67, 29) ### ### ('championships', 141, 41) ('teams', 55, 52) ('athletics', 699, 15) ('wolves', 1265, 6) ### ### ('separately', 780, 19) ('northern', 3218, 0) ('canada', 400, 34) ('u', 1594, 13) ### ### ('football', 6, 164) ('battalion', 4119, 3) ### ############################################################################################################ [2023-11-25 02:51:45,384][root][INFO] - Epoch: 1: Step: 12501/19530, loss[v]=0.298338, lr=0.000013, acc@1[1]=965.5/1024=0.94287109375, acc@1[2]=986.0/1024=0.962890625 [2023-11-25 02:53:35,489][root][INFO] - Train batch 12600 [2023-11-25 02:53:35,491][root][INFO] - Avg. loss per last 100 batches: 0.216990 [2023-11-25 02:53:36,231][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.67 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 4.84 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.3/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 2004, american actor zach braff, a friend of radin since their days at northwestern, ### ### introduced radin's first composition, " winter ", to " scrubs " show creator bill lawrence, who ### ### ultimately used several of radin's songs in various scenes of the television series. ### ### [P_TEXT]: chariot ( song ) it was certified platinum. it was featured prominently in a 2004 episode ### ### of " dead like me ", in which he had prior guest starred. the music video for " chariot " was ### ### directed by zach braff, who also directed music videos for joshua radin. it stars model and actress ### ### jaime king and also features a brief cameo from donald faison, who plays the role of christopher ### ### turk in the tv show " scrubs ", in which braff is also a cast member. ### ### =============================================== v_q ================================================ ### ### ('winter', 0, 17043) ('scrub', 1, 1) ('zach', 2, 5) ('lawrence', 3, 19154) ### ### ('northwestern', 4, 4901) ('##din', 5, 3) ('ra', 6, 21) ('##ff', 7, 11) ('bill', 8, 7759) ### ### ('bra', 9, 16) ('summer', 10, 10036) ('series', 11, 424) ('show', 12, 28) ('2004', 13, 13) ### ### ('northeastern', 14, 15983) ('television', 15, 93) ('tv', 16, 29) ('actor', 17, 207) ### ### ('.', 18, 1558) ('song', 19, 15) ### ### =============================================== v_p ================================================ ### ### ('chariot', 19762, 0) ('scrub', 1, 1) ('jaime', 21476, 2) ('##din', 5, 3) ('turk', 12657, 4) ### ### ('zach', 2, 5) ('fai', 18922, 6) ('joshua', 7328, 7) ('king', 1533, 8) ('model', 4828, 9) ### ### ('christopher', 826, 10) ('##ff', 7, 11) ('dead', 748, 12) ('2004', 13, 13) ('donald', 4294, 14) ### ### ('song', 19, 15) ('bra', 9, 16) ('##son', 16548, 17) ('me', 3566, 18) ('like', 7032, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('scrub', 1, 1) ('zach', 2, 5) ('##din', 5, 3) ('##ff', 7, 11) ('ra', 6, 21) ('bra', 9, 16) ### ### ('2004', 13, 13) ('song', 19, 15) ('show', 12, 28) ('episode', 29, 22) ('tv', 16, 29) ### ### ('directing', 163, 20) ('music', 27, 33) ('songs', 23, 35) ('christopher', 826, 10) ### ### ('dead', 748, 12) ('king', 1533, 8) ('sitcom', 35, 39) ('tune', 51, 34) ('filming', 21, 58) ### ############################################################################################################ [2023-11-25 02:53:36,733][root][INFO] - Epoch: 1: Step: 12601/19530, loss[v]=0.259448, lr=0.000013, acc@1[1]=963.0/1024=0.9404296875, acc@1[2]=984.5/1024=0.96142578125 [2023-11-25 02:55:26,523][root][INFO] - Train batch 12700 [2023-11-25 02:55:26,523][root][INFO] - Avg. loss per last 100 batches: 0.241207 [2023-11-25 02:55:27,257][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.52 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 5.15 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.7/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: quality and smoke from an unprecedented number of wildfires. ### ### [P_TEXT]: 2008 california wildfires number of people with eye and throat irritation. the bad air ### ### quality also forced the cancellation of the western states endurance run, the first in the race's ### ### 31 - year history. air quality began to improve on june 28, followed by decreased smoke and ### ### improved visibility a day later. by june 30, residents in the sacramento valley saw blue skies and ### ### good air quality, as a result of onshore winds and the delta breezes. however, air quality in ### ### oregon degraded as plumes of smoke drifted northward instead of concentrating in the central ### ### valley. hazy conditions ### ### =============================================== v_q ================================================ ### ### ('.', 0, 113) ('smoke', 1, 10) ('wild', 2, 14) ('##fires', 3, 0) ('quality', 4, 15) ('fire', 5, 37) ### ### ('##fire', 6, 22) ('fires', 7, 52) ('bonfire', 8, 205) ('absently', 9, 154) ('manpower', 10, 92) ### ### ('awkward', 11, 71) ('flames', 12, 58) ('lighting', 13, 1945) ('worried', 14, 180) ### ### ('gritted', 15, 242) ('noticed', 16, 62) ('##უ', 17, 231) ('श', 18, 334) ('stove', 19, 112) ### ### =============================================== v_p ================================================ ### ### ('##fires', 3, 0) ('endurance', 2342, 1) ('throat', 6314, 2) ('plume', 6190, 3) ('breeze', 4430, 4) ### ### ('sacramento', 10449, 5) ('skies', 7713, 6) ('irritation', 2227, 7) ('2008', 13259, 8) ### ### ('california', 830, 9) ('smoke', 1, 10) ('delta', 10000, 11) ('oregon', 1103, 12) ('eye', 5036, 13) ### ### ('wild', 2, 14) ('quality', 4, 15) ('##shore', 23942, 16) ('valley', 1752, 17) ### ### ('central', 6393, 18) ('winds', 6920, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##fires', 3, 0) ('smoke', 1, 10) ('wild', 2, 14) ('quality', 4, 15) ('##fire', 6, 22) ### ### ('fire', 5, 37) ('.', 0, 113) ('fires', 7, 52) ('air', 135, 23) ('conditions', 60, 32) ### ### ('california', 830, 9) ('visibility', 378, 21) ('oregon', 1103, 12) ('endurance', 2342, 1) ### ### ('states', 438, 26) ('flames', 12, 58) ('irritation', 2227, 7) ('noticed', 16, 62) ### ### ('race', 1161, 20) ('awkward', 11, 71) ### ############################################################################################################ [2023-11-25 02:55:27,761][root][INFO] - Epoch: 1: Step: 12701/19530, loss[v]=0.169778, lr=0.000013, acc@1[1]=983.0/1024=0.9599609375, acc@1[2]=1005.0/1024=0.9814453125 [2023-11-25 02:57:17,668][root][INFO] - Train batch 12800 [2023-11-25 02:57:17,669][root][INFO] - Avg. loss per last 100 batches: 0.219371 [2023-11-25 02:57:18,393][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 5.13 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 26.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he was the piano teacher of the composer francis poulenc and the pianists marcelle meyer, ### ### joaquin nin - culmell and leo - pol morin. ### ### [P_TEXT]: francis poulenc and orchestra. as the only son of a prosperous manufacturer, poulenc was ### ### expected to follow his father into the family firm, and he was not allowed to enrol at a music ### ### college. largely self - educated musically, he studied with the pianist ricardo vines, who became ### ### his mentor after the composer's parents died. poulenc also made the acquaintance of erik satie, ### ### under whose tutelage he became one of a group of young composers known collectively as " les six ". ### ### in his early works poulenc became known for his high spirits and irreverence. during ### ### =============================================== v_q ================================================ ### ### ('meyer', 0, 15963) ('pol', 1, 4500) ('##ule', 2, 1) ('mori', 3, 16453) ('ni', 4, 16615) ### ### ('francis', 5, 11) ('joaquin', 6, 15129) ('##lm', 7, 28492) ('##le', 8, 15013) ('##nc', 9, 4) ### ### ('marcel', 10, 8213) ('piano', 11, 57) ('po', 12, 12) ('leo', 13, 11081) ('##ell', 14, 22089) ### ### ('pianist', 15, 16) ('cu', 16, 20295) ('composer', 17, 29) ('##n', 18, 23700) ('##ன', 19, 756) ### ### =============================================== v_p ================================================ ### ### ('ricardo', 740, 0) ('##ule', 2, 1) ('vines', 3794, 2) ('spirits', 3876, 3) ('##nc', 9, 4) ### ### ('les', 4257, 5) ('erik', 667, 6) ('##vere', 19011, 7) ('sat', 3312, 8) ('six', 12909, 9) ### ### ('##tel', 13572, 10) ('francis', 5, 11) ('po', 12, 12) ('##ie', 13724, 13) ('composers', 24, 14) ### ### ('firm', 7305, 15) ('pianist', 15, 16) ('orchestra', 66, 17) ('music', 25, 18) ('##nce', 5663, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ule', 2, 1) ('francis', 5, 11) ('##nc', 9, 4) ('po', 12, 12) ('pianist', 15, 16) ### ### ('composer', 17, 29) ('composers', 24, 14) ('music', 25, 18) ('piano', 11, 57) ### ### ('orchestra', 66, 17) ('ricardo', 740, 0) ('erik', 667, 6) ('mentor', 493, 20) ('vines', 3794, 2) ### ### ('spirits', 3876, 3) ('sat', 3312, 8) ('spirit', 737, 22) ('les', 4257, 5) ('son', 1111, 26) ### ### ('works', 355, 39) ### ############################################################################################################ [2023-11-25 02:57:18,894][root][INFO] - Epoch: 1: Step: 12801/19530, loss[v]=0.182065, lr=0.000013, acc@1[1]=989.0/1024=0.9658203125, acc@1[2]=1008.0/1024=0.984375 [2023-11-25 02:59:08,816][root][INFO] - Train batch 12900 [2023-11-25 02:59:08,817][root][INFO] - Avg. loss per last 100 batches: 0.236422 [2023-11-25 02:59:09,541][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.45 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 4.78 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: variations for each number, is as follows : the types of dances upon which dvorak based ### ### his music include the furiant, the dumka, the polka, the sousedska, the skocna, the mazurka, the ### ### odzemek, the spacirka, the kolo and the polonaise. ### ### [P_TEXT]: kolo ( dance ) the waist. the basic steps are easy to learn. experienced dancers ### ### demonstrate virtuosity by adding different ornamental elements, such as syncopated steps. each ### ### region has at least one unique kolo. it is difficult to master the dance and even most experienced ### ### dancers cannot master all of them. kolo is performed at weddings, social, cultural, and religious ### ### ceremonies. some dances require both men and women to dance together, others require only the men ### ### or only the women. the music is generally fast - paced. the dance was used by antonin dvorak in his ### ### " slavonic dances ### ### =============================================== v_q ================================================ ### ### ('##nais', 0, 27138) ('polka', 1, 20634) ('sous', 2, 24520) ('##rka', 3, 25941) ('music', 4, 38) ### ### ('##ian', 5, 19335) ('dance', 6, 9) ('##ʂ', 7, 1695) ('du', 8, 13379) ('##ak', 9, 19) ### ### ('fur', 10, 8751) ('czech', 11, 595) ('##ska', 12, 23600) ('spa', 13, 16891) ('dances', 14, 11) ### ### ('##oc', 15, 24133) ('##me', 16, 19340) ('polo', 17, 14872) ('##vor', 18, 6) ('##lo', 19, 0) ### ### =============================================== v_p ================================================ ### ### ('##lo', 19, 0) ('waist', 2774, 1) ('anton', 8905, 2) ('ornamental', 8524, 3) ('##onic', 10061, 4) ### ### ('##av', 13680, 5) ('##vor', 18, 6) ('##opa', 19561, 7) ('sync', 8836, 8) ('dance', 6, 9) ### ### ('##sity', 6373, 10) ('dances', 14, 11) ('ko', 27, 12) ('paced', 4036, 13) ('steps', 12692, 14) ### ### ('fast', 3993, 15) ('weddings', 16213, 16) ('##rt', 19005, 17) ('##uo', 23619, 18) ('##ak', 9, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('dance', 6, 9) ('##lo', 19, 0) ('dances', 14, 11) ('##vor', 18, 6) ('##ak', 9, 19) ('ko', 27, 12) ### ### ('music', 4, 38) ('dancing', 30, 31) ('dancers', 127, 28) ('d', 46, 49) ('basic', 399, 20) ### ### ('ballet', 35, 73) ('ד', 51, 64) ('waist', 2774, 1) ('used', 460, 30) ('musical', 40, 90) ### ### ('dancer', 84, 52) ('tune', 47, 100) ('czech', 11, 595) ('paced', 4036, 13) ### ############################################################################################################ [2023-11-25 02:59:10,043][root][INFO] - Epoch: 1: Step: 12901/19530, loss[v]=0.158645, lr=0.000013, acc@1[1]=992.0/1024=0.96875, acc@1[2]=1008.5/1024=0.98486328125 [2023-11-25 03:01:00,208][root][INFO] - Train batch 13000 [2023-11-25 03:01:00,209][root][INFO] - Avg. loss per last 100 batches: 0.227731 [2023-11-25 03:01:00,942][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.62 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 4.85 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.7/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.4/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: hired many of the best jazz musicians for his band, including bix beiderbecke, frankie ### ### trumbauer, joe venuti, eddie lang, steve brown, mike pingitore, gussie mueller, wilbur hall ( ### ### billed by whiteman as " willie hall " ), jack teagarden, and bunny berigan. ### ### [P_TEXT]: george marsh ( musician ) george marsh ( c. 1900 – 1962 ) was an american jazz drummer. ### ### in the 1920s, he played with one of paul whiteman's bands, as well as recording with musicians ### ### including frankie trumbauer and eddie lang. from 1932 to 1934, marsh played in an orchestra led by ### ### ferde grofe. he later moved to california, where he worked in film music. he died in los angeles in ### ### april 1962. ### ### =============================================== v_q ================================================ ### ### ('lang', 0, 1) ('wilbur', 1, 23663) ('##nut', 2, 25902) ('##itor', 3, 25679) ('ping', 4, 21273) ### ### ('##umb', 5, 9) ('jazz', 6, 11) ('eddie', 7, 6) ('##auer', 8, 7) ('hall', 9, 9182) ### ### ('tea', 10, 8301) ('bei', 11, 24671) ('bunny', 12, 12297) ('##gard', 13, 23929) ### ### ('mueller', 14, 24274) ('##sie', 15, 23334) ('##ri', 16, 25538) ('##beck', 17, 27338) ### ### ('steve', 18, 4561) ('brown', 19, 1392) ### ### =============================================== v_p ================================================ ### ### ('marsh', 9330, 0) ('lang', 0, 1) ('##rde', 14013, 2) ('##of', 23936, 3) ('george', 145, 4) ### ### ('white', 20, 5) ('eddie', 7, 6) ('##auer', 8, 7) ('frankie', 23, 8) ('##umb', 5, 9) ### ### ('paul', 924, 10) ('jazz', 6, 11) ('fe', 10395, 12) ('1962', 280, 13) ('##man', 34, 14) ### ### ('1920s', 84, 15) ('california', 463, 16) ('1934', 1320, 17) ('bands', 138, 18) ### ### ('musicians', 40, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('lang', 0, 1) ('eddie', 7, 6) ('##umb', 5, 9) ('##auer', 8, 7) ('jazz', 6, 11) ('white', 20, 5) ### ### ('frankie', 23, 8) ('##man', 34, 14) ('musicians', 40, 19) ('george', 145, 4) ('tr', 45, 21) ### ### ('music', 43, 26) ('1920s', 84, 15) ('musician', 59, 23) ('band', 49, 29) ('orchestra', 64, 27) ### ### ('bands', 138, 18) ('1962', 280, 13) ('paul', 924, 10) ('drummer', 220, 22) ### ############################################################################################################ [2023-11-25 03:01:01,446][root][INFO] - Epoch: 1: Step: 13001/19530, loss[v]=0.301530, lr=0.000013, acc@1[1]=952.0/1024=0.9296875, acc@1[2]=979.0/1024=0.9560546875 [2023-11-25 03:02:51,674][root][INFO] - Train batch 13100 [2023-11-25 03:02:51,675][root][INFO] - Avg. loss per last 100 batches: 0.236884 [2023-11-25 03:02:52,408][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.70 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 5.00 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: during the raron affair rebellion in 1414 to 1420, some cantons of the swiss ### ### confederation took sides in the conflict. ### ### [P_TEXT]: raron affair secular offices in the valais. after the death of bishop william i of raron ### ### in 1402 and his father peter of raron in 1413, a rebellious party formed in the upper valais. they ### ### were opposed to the new prince - bishop wilhelm ii of raron and the claims of his uncle witschard ### ### of raron. emperor sigismund had granted witschard sovereignty over the valais in 1414 as a reward ### ### for his military service in lombardy. the rights over the valais had been granted to the office of ### ### the bishop of sion in 999. the ### ### =============================================== v_q ================================================ ### ### ('swiss', 0, 1224) ('##ron', 1, 5) ('affair', 2, 2) ('ra', 3, 10) ('canton', 4, 19516) ### ### ('141', 5, 8) ('142', 6, 43) ('confederation', 7, 4462) ('rebellion', 8, 31) ('switzerland', 9, 86) ### ### ('.', 10, 553) ('war', 11, 4417) ('conflict', 12, 353) ('uprising', 13, 62) ('battle', 14, 5231) ### ### ('francs', 15, 1495) ('revolt', 16, 69) ('##ᵉ', 17, 158) ('alps', 18, 7358) ('empire', 19, 77) ### ### =============================================== v_p ================================================ ### ### ('wits', 26486, 0) ('sigismund', 4560, 1) ('affair', 2, 2) ('wilhelm', 1140, 3) ### ### ('lombardy', 8413, 4) ('##ron', 1, 5) ('val', 577, 6) ('140', 31, 7) ('141', 5, 8) ('si', 9141, 9) ### ### ('ra', 3, 10) ('peter', 4502, 11) ('sovereignty', 1142, 12) ('bishop', 384, 13) ### ### ('##ais', 13996, 14) ('claims', 5910, 15) ('william', 1515, 16) ('upper', 1232, 17) ### ### ('uncle', 10580, 18) ('##on', 1269, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('affair', 2, 2) ('##ron', 1, 5) ('ra', 3, 10) ('141', 5, 8) ('rebellion', 8, 31) ('142', 6, 43) ### ### ('140', 31, 7) ('switzerland', 9, 86) ('bishop', 384, 13) ('val', 577, 6) ('wilhelm', 1140, 3) ### ### ('opposed', 46, 33) ('uprising', 13, 62) ('sovereignty', 1142, 12) ('granted', 446, 26) ### ### ('emperor', 698, 20) ('swiss', 0, 1224) ('sigismund', 4560, 1) ('upper', 1232, 17) ### ### ('william', 1515, 16) ### ############################################################################################################ [2023-11-25 03:02:52,909][root][INFO] - Epoch: 1: Step: 13101/19530, loss[v]=0.120604, lr=0.000013, acc@1[1]=996.0/1024=0.97265625, acc@1[2]=1011.5/1024=0.98779296875 [2023-11-25 03:04:43,246][root][INFO] - Train batch 13200 [2023-11-25 03:04:43,246][root][INFO] - Avg. loss per last 100 batches: 0.237120 [2023-11-25 03:04:43,970][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.54 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 5.01 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the new york pancyprian - freedoms defeated maccabee a. c. in the final game in extra ### ### time by a score of 4 – 3. ### ### [P_TEXT]: history of the u. s. open cup this time included the new york pancyprian - freedoms, who ### ### won three domestic titles between 1979 and 1982, and maccabi los angeles who made it to the ### ### national cup finals seven times in nine years, from 1975 to 1984. in fact, the two teams met twice ### ### in the national cup finals ; in 1980 and 1982 with the pancyprian - freedoms defeating maccabi both ### ### times. starting in 1995, professional clubs would retain their ussf sanctioning if their clubs ### ### agreed to participate in the tournament. the 1995 u. s. open cup featured a mix of ### ### =============================================== v_q ================================================ ### ### ('##bee', 0, 29078) ('freedoms', 1, 2) ('##ca', 2, 21475) ('mac', 3, 12973) ('c', 4, 17314) ### ### ('york', 5, 24) ('pan', 6, 14) ('##cy', 7, 5) ('hockey', 8, 609) ('basketball', 9, 411) ### ### ('fiba', 10, 11381) ('##pr', 11, 4) ('##ian', 12, 19) ('.', 13, 170) ('game', 14, 98) ### ### ('あ', 15, 785) ('nba', 16, 9095) ('##あ', 17, 1291) ('lacrosse', 18, 11910) ('final', 19, 53) ### ### =============================================== v_p ================================================ ### ### ('maccabi', 1230, 0) ('##ction', 10636, 1) ('freedoms', 1, 2) ('open', 146, 3) ('##pr', 11, 4) ### ### ('##cy', 7, 5) ('angeles', 2693, 6) ('uss', 1726, 7) ('1995', 483, 8) ('cup', 53, 9) ### ### ('1975', 2204, 10) ('1982', 2578, 11) ('1980', 1447, 12) ('u', 5340, 13) ('pan', 6, 14) ### ### ('1979', 2464, 15) ('finals', 83, 16) ('clubs', 2499, 17) ('tournament', 21, 18) ('##ian', 12, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('freedoms', 1, 2) ('##cy', 7, 5) ('pan', 6, 14) ('##pr', 11, 4) ('york', 5, 24) ('##ian', 12, 19) ### ### ('tournament', 21, 18) ('cup', 53, 9) ('open', 146, 3) ('finals', 83, 16) ('1995', 483, 8) ### ### ('maccabi', 1230, 0) ('final', 19, 53) ('defeated', 151, 28) ('game', 14, 98) ('uss', 1726, 7) ### ### ('americans', 139, 31) ('awkward', 37, 47) ('ny', 23, 72) ('usa', 386, 25) ### ############################################################################################################ [2023-11-25 03:04:44,472][root][INFO] - Epoch: 1: Step: 13201/19530, loss[v]=0.305828, lr=0.000014, acc@1[1]=950.0/1024=0.927734375, acc@1[2]=983.0/1024=0.9599609375 [2023-11-25 03:06:34,676][root][INFO] - Train batch 13300 [2023-11-25 03:06:34,677][root][INFO] - Avg. loss per last 100 batches: 0.229379 [2023-11-25 03:06:35,400][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.43 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.9/29522=99.99% | mean: 0.03 | max: 4.94 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: their sixth album, " lightning strikes ", was once again produced by max norman and ### ### charted at no. ### ### [P_TEXT]: lightning strikes ( loudness album ) lightning strikes is the sixth studio album by the ### ### japan ese heavy metal band loudness. the album, which was released on july 25, 1986, remained 15 ### ### weeks in the u. s. charts, peaking at # 64 ( " billboard " 200 ). the album was produced by max ### ### norman. lyrics for the album are credited to loudness, even though producer max norman did a lot of ### ### uncredited work on them. the album begins with the most commercial sounding song on the album, " ### ### let it go. " however, the rest of the album is ### ### =============================================== v_q ================================================ ### ### ('lightning', 0, 0) ('max', 1, 3) ('norman', 2, 1) ('strikes', 3, 4) ('sixth', 4, 17) ### ### ('album', 5, 14) ('strike', 6, 11) ('band', 7, 18) ('bands', 8, 22) ('.', 9, 957) ### ### ('struck', 10, 31) ('fifth', 11, 47) ('clashes', 12, 360) ('producers', 13, 24) ### ### ('producer', 14, 25) ('beatles', 15, 457) ('albums', 16, 21) ('freddie', 17, 13418) ### ### ('seventh', 18, 82) ('collective', 19, 3051) ### ### =============================================== v_p ================================================ ### ### ('lightning', 0, 0) ('norman', 2, 1) ('loud', 3862, 2) ('max', 1, 3) ('strikes', 3, 4) ### ### ('es', 9309, 5) ('let', 3089, 6) ('japan', 8924, 7) ('##ness', 21127, 8) ('metal', 82, 9) ### ### ('1986', 5055, 10) ('strike', 6, 11) ('go', 4421, 12) ('uncredited', 15472, 13) ('album', 5, 14) ### ### ('64', 8419, 15) ('sounding', 3225, 16) ('sixth', 4, 17) ('band', 7, 18) ('lyrics', 316, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('lightning', 0, 0) ('norman', 2, 1) ('max', 1, 3) ('strikes', 3, 4) ('album', 5, 14) ### ### ('sixth', 4, 17) ('strike', 6, 11) ('band', 7, 18) ('bands', 8, 22) ('metal', 82, 9) ### ### ('producers', 13, 24) ('struck', 10, 31) ('albums', 16, 21) ('producer', 14, 25) ('charts', 29, 20) ### ### ('fifth', 11, 47) ('loud', 3862, 2) ('billboard', 30, 30) ('frontman', 24, 33) ('lyrics', 316, 19) ### ############################################################################################################ [2023-11-25 03:06:35,903][root][INFO] - Epoch: 1: Step: 13301/19530, loss[v]=0.162417, lr=0.000014, acc@1[1]=992.5/1024=0.96923828125, acc@1[2]=1010.0/1024=0.986328125 [2023-11-25 03:08:26,184][root][INFO] - Train batch 13400 [2023-11-25 03:08:26,185][root][INFO] - Avg. loss per last 100 batches: 0.226693 [2023-11-25 03:08:26,929][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 5.11 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he is both the son of a former colorado football player and the nephew of hale irwin, ### ### also a former football player at colorado, who would eventually go on to win three u. s. open golf ### ### titles and become a world golf hall of fame member. ### ### [P_TEXT]: hale irwin enjoys hunting and fishing and spending time with his grandchildren. he is the ### ### uncle of former cu lineman heath irwin. in 2019, in acknowledgement of his character, sportsmanship ### ### and commitment to charity, irwin received the pga tour's payne stewart award presented by southern ### ### company. 1defeated mike donald with a birdie on the 19th hole after they were tied at 74 ( + 2 ) in ### ### an 18 - hole playoff. " the senior british open was not a senior major until 2003. " ### ### =============================================== v_q ================================================ ### ### ('colorado', 0, 2398) ('hale', 1, 1) ('irwin', 2, 0) ('golf', 3, 59) ('football', 4, 1320) ### ### ('nephew', 5, 47) ('open', 6, 14) ('fame', 7, 1802) ('world', 8, 10468) ('.', 9, 591) ### ### ('hall', 10, 8904) ('denver', 11, 4845) ('utah', 12, 1519) ('uncle', 13, 24) ('donald', 14, 3) ### ### ('player', 15, 1413) ('larry', 16, 3252) ('game', 17, 105) ('jimmy', 18, 7162) ('son', 19, 437) ### ### =============================================== v_p ================================================ ### ### ('irwin', 2, 0) ('hale', 1, 1) ('payne', 7292, 2) ('donald', 14, 3) ('heath', 19861, 4) ### ### ('southern', 3363, 5) ('bird', 8165, 6) ('stewart', 4833, 7) ('sportsman', 7802, 8) ### ### ('company', 7337, 9) ('mike', 32, 10) ('pga', 30, 11) ('tour', 452, 12) ('british', 5382, 13) ### ### ('open', 6, 14) ('senior', 5944, 15) ('cu', 12885, 16) ('lineman', 16457, 17) ('hunting', 2886, 18) ### ### ('charity', 6947, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('hale', 1, 1) ('irwin', 2, 0) ('open', 6, 14) ('donald', 14, 3) ('pga', 30, 11) ('mike', 32, 10) ### ### ('uncle', 13, 24) ('nephew', 5, 47) ('golf', 3, 59) ('tour', 452, 12) ('awards', 327, 25) ### ### ('tied', 547, 23) ('brother', 105, 45) ('southern', 3363, 5) ('playoff', 280, 35) ### ### ('hole', 1184, 22) ('payne', 7292, 2) ('stewart', 4833, 7) ('colorado', 0, 2398) ('game', 17, 105) ### ############################################################################################################ [2023-11-25 03:08:27,431][root][INFO] - Epoch: 1: Step: 13401/19530, loss[v]=0.150784, lr=0.000014, acc@1[1]=984.0/1024=0.9609375, acc@1[2]=1011.0/1024=0.9873046875 [2023-11-25 03:10:17,748][root][INFO] - Train batch 13500 [2023-11-25 03:10:17,750][root][INFO] - Avg. loss per last 100 batches: 0.229832 [2023-11-25 03:10:18,488][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.63 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=99.99% | mean: 0.03 | max: 4.95 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in contrast to the ensemble nature of the series, the episode contained none ( except ### ### iain de caestecker ) of the main cast and was centered around henstridge's character. ### ### [P_TEXT]: afterlife ( agents of s. h. i. e. l. d. ) attractive people with powers. and we thought, ### ### “ hey, let ’ s have some of those as well! ” we were looking for a new character to come on, and ### ### luke just really impressed us. he was a good actor, had a nice quality, and we felt he might be a ### ### good person to sort of usher skye into this other world. " in march 2015, marvel revealed that main ### ### cast members clark gregg, ming - na wen, chloe bennet, iain de caestecker, elizabeth henstridge, ### ### nick blood, and adrianne palicki would star as phil coulson, melinda may, skye, leo fitz, jemma ### ### =============================================== v_q ================================================ ### ### ('hen', 0, 38) ('##est', 1, 29) ('##ridge', 2, 8) ('iain', 3, 16) ('##eck', 4, 18) ### ### ('episode', 5, 2409) ('##st', 6, 44) ('doctor', 7, 4105) ('ensemble', 8, 7256) ('bbc', 9, 9227) ### ### ('cast', 10, 47) ('de', 11, 81) ('.', 12, 543) ('character', 13, 56) ('sitcom', 14, 540) ### ### ('tries', 15, 9698) ('absently', 16, 482) ('characters', 17, 62) ('telling', 18, 1026) ### ### ('eastenders', 19, 16385) ### ### =============================================== v_p ================================================ ### ### ('afterlife', 14167, 0) ('melinda', 1125, 1) ('jem', 4092, 2) ('skye', 1112, 3) ('gregg', 23594, 4) ### ### ('phil', 11329, 5) ('##uls', 29017, 6) ('marvel', 470, 7) ('##ridge', 2, 8) ('usher', 9587, 9) ### ### ('agents', 7211, 10) ('wen', 11026, 11) ('chloe', 77, 12) ('fitz', 15135, 13) ('pali', 20908, 14) ### ### ('blood', 123, 15) ('iain', 3, 16) ('hey', 3774, 17) ('##eck', 4, 18) ('##net', 7666, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ridge', 2, 8) ('iain', 3, 16) ('##eck', 4, 18) ('##est', 1, 29) ('hen', 0, 38) ('##st', 6, 44) ### ### ('chloe', 77, 12) ('cast', 10, 47) ('blood', 123, 15) ('character', 13, 56) ('ca', 30, 43) ### ### ('marvel', 470, 7) ('melinda', 1125, 1) ('revealed', 34, 49) ('skye', 1112, 3) ### ### ('characters', 17, 62) ('nick', 388, 24) ('2015', 265, 33) ('actor', 33, 61) ('de', 11, 81) ### ############################################################################################################ [2023-11-25 03:10:18,991][root][INFO] - Epoch: 1: Step: 13501/19530, loss[v]=0.280328, lr=0.000014, acc@1[1]=964.0/1024=0.94140625, acc@1[2]=983.5/1024=0.96044921875 [2023-11-25 03:12:09,280][root][INFO] - Train batch 13600 [2023-11-25 03:12:09,281][root][INFO] - Avg. loss per last 100 batches: 0.216646 [2023-11-25 03:12:10,007][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.67 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 5.25 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: they began and ran a boys'boarding school, which grew into the first christian college in ### ### china, into cheeloo university, and eventually into shandong university. ### ### [P_TEXT]: shandong university the earliest precursor institutions that would later be fused into ### ### shandong university were founded by american and english mission agencies : in early january 1864, ### ### calvin w. mateer, an american presbyterian missionary, and his wife julia brown mateer, arrived in ### ### the recently opened treaty port of dengzhou ( ) in the area of the present - day city of penglai on ### ### the north - eastern coast of shandong peninsula. their journey had begun in new york on july 3, ### ### 1863, had taken them around the cape of good hope to shanghai, and had ended with a ### ### =============================================== v_q ================================================ ### ### ('##elo', 0, 26828) ('shandong', 1, 0) ('christian', 2, 396) ('university', 3, 5) ('che', 4, 16298) ### ### ('china', 5, 48) ('college', 6, 76) ('##o', 7, 26666) ('boys', 8, 12105) ('school', 9, 519) ### ### ('chinese', 10, 115) ('##ो', 11, 1952) ('boarding', 12, 21310) ('shanghai', 13, 6) ### ### ('guangdong', 14, 1210) ('zhejiang', 15, 61) ('henan', 16, 197) ('.', 17, 89) ('taiwan', 18, 485) ### ### ('，', 19, 507) ### ### =============================================== v_p ================================================ ### ### ('shandong', 1, 0) ('deng', 772, 1) ('peng', 4163, 2) ('mate', 13660, 3) ('##lai', 22089, 4) ### ### ('university', 3, 5) ('shanghai', 13, 6) ('treaty', 19804, 7) ('calvin', 7617, 8) ('cape', 1927, 9) ### ### ('presbyterian', 555, 10) ('brown', 10288, 11) ('hope', 2635, 12) ('julia', 1407, 13) ### ### ('peninsula', 11867, 14) ('agencies', 3389, 15) ('mission', 3734, 16) ('##zhou', 151, 17) ### ### ('missionary', 6092, 18) ('journey', 4507, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('shandong', 1, 0) ('university', 3, 5) ('shanghai', 13, 6) ('china', 5, 48) ('college', 6, 76) ### ### ('deng', 772, 1) ('##zhou', 151, 17) ('universities', 24, 43) ('christian', 2, 396) ### ### ('presbyterian', 555, 10) ('zhejiang', 15, 61) ('peng', 4163, 2) ('tianjin', 20, 67) ### ### ('chinese', 10, 115) ('cape', 1927, 9) ('julia', 1407, 13) ('.', 17, 89) ('anyway', 32, 56) ### ### ('institutions', 241, 33) ('hope', 2635, 12) ### ############################################################################################################ [2023-11-25 03:12:10,508][root][INFO] - Epoch: 1: Step: 13601/19530, loss[v]=0.274286, lr=0.000014, acc@1[1]=983.5/1024=0.96044921875, acc@1[2]=1005.5/1024=0.98193359375 [2023-11-25 03:14:00,691][root][INFO] - Train batch 13700 [2023-11-25 03:14:00,692][root][INFO] - Avg. loss per last 100 batches: 0.217716 [2023-11-25 03:14:01,442][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.64 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 5.11 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: situated north of dundee and south of aberdeen, montrose lies between the mouths of the ### ### north and south esk rivers. ### ### [P_TEXT]: aberdeen to alford. the a92 was the original southerly road to aberdeen prior to the ### ### building of the a90, and is now used as a tourist route, connecting the towns of montrose and ### ### arbroath and on the east coast. the a947 exits the city at dyce and goes on to newmachar, ### ### oldmeldrum and turriff finally ending at banff and macduff. after first being mooted 60 years ago ### ### and being held up for the past five years by a number of ### ### =============================================== v_q ================================================ ### ### ('montrose', 0, 1) ('dundee', 1, 45) ('es', 2, 22972) ('aberdeen', 3, 0) ('##k', 4, 27069) ### ### ('river', 5, 5952) ('rivers', 6, 5698) ('north', 7, 2138) ('scotland', 8, 55) ('.', 9, 117) ### ### ('argyll', 10, 1096) ('angus', 11, 3098) ('scottish', 12, 75) ('robertson', 13, 292) ### ### ('fergus', 14, 276) ('cameron', 15, 944) ('edinburgh', 16, 39) ('glen', 17, 10157) ### ### ('fife', 18, 1962) ('whiskey', 19, 1107) ### ### =============================================== v_p ================================================ ### ### ('aberdeen', 3, 0) ('montrose', 0, 1) ('##90', 4910, 2) ('##yce', 14692, 3) ('alf', 8249, 4) ### ### ('##mac', 25484, 5) ('##drum', 26645, 6) ('##ath', 19283, 7) ('##47', 14047, 8) ### ### ('##erly', 13100, 9) ('##ff', 16319, 10) ('##ord', 3323, 11) ('##rri', 16865, 12) ### ### ('##bro', 17813, 13) ('ban', 15367, 14) ('##mel', 18112, 15) ('##du', 22219, 16) ('mac', 68, 17) ### ### ('##9', 12980, 18) ('tu', 1226, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('montrose', 0, 1) ('aberdeen', 3, 0) ('dundee', 1, 45) ('scotland', 8, 55) ('mac', 68, 17) ### ### ('town', 21, 36) ('edinburgh', 16, 39) ('coast', 341, 21) ('.', 9, 117) ('scottish', 12, 75) ### ### ('city', 102, 30) ('##90', 4910, 2) ('towns', 433, 27) ('tu', 1226, 19) ('##ord', 3323, 11) ### ### ('glasgow', 72, 52) ('tourist', 1593, 24) ('alf', 8249, 4) ('south', 63, 59) ('tourism', 2198, 26) ### ############################################################################################################ [2023-11-25 03:14:01,944][root][INFO] - Epoch: 1: Step: 13701/19530, loss[v]=0.120844, lr=0.000014, acc@1[1]=989.5/1024=0.96630859375, acc@1[2]=1012.0/1024=0.98828125 [2023-11-25 03:15:52,192][root][INFO] - Train batch 13800 [2023-11-25 03:15:52,193][root][INFO] - Avg. loss per last 100 batches: 0.237552 [2023-11-25 03:15:52,934][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.66 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 5.21 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 26.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: by late 19th century, the hudson's bay company expanded into the interior and established ### ### posts along rivers that later developed into the modern cities of winnipeg, calgary and edmonton. ### ### [P_TEXT]: edmonton edmonton ( ) is the capital city of the canadian province of alberta. edmonton ### ### is on the north saskatchewan river and is the centre of the edmonton metropolitan region, which is ### ### surrounded by alberta's central region. the city anchors the north end of what statistics canada ### ### defines as the " calgary – edmonton corridor ". the metro area of edmonton had a population of 1, ### ### 491, 000 at the beginning of 2021, making it alberta's second - largest city ( after calgary ) and ### ### canada's fifth - largest municipality. ### ### =============================================== v_q ================================================ ### ### ('.', 0, 30) ('calgary', 1, 9) ('edmonton', 2, 0) ('winnipeg', 3, 360) ('bay', 4, 12076) ### ### ('alaska', 5, 1626) ('alberta', 6, 5) ('canada', 7, 24) ('manitoba', 8, 169) ('hudson', 9, 4371) ### ### ('company', 10, 7949) ('newfoundland', 11, 1267) ('canadian', 12, 18) ('river', 13, 13) ### ### ('absently', 14, 156) ('immigrants', 15, 1094) ('ontario', 16, 271) ('city', 17, 29) ### ### ('cities', 18, 28) ('saskatchewan', 19, 4) ### ### =============================================== v_p ================================================ ### ### ('edmonton', 2, 0) ('corridor', 942, 1) ('metro', 1932, 2) ('metropolitan', 367, 3) ### ### ('saskatchewan', 19, 4) ('alberta', 6, 5) ('central', 3242, 6) ('region', 78, 7) ### ### ('statistics', 7487, 8) ('calgary', 1, 9) ('regions', 331, 10) ('fifth', 16157, 11) ### ### ('largest', 1211, 12) ('river', 13, 13) ('provinces', 3105, 14) ('capital', 897, 15) ### ### ('2021', 25525, 16) ('municipalities', 1888, 17) ('canadian', 12, 18) ('municipality', 307, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('edmonton', 2, 0) ('calgary', 1, 9) ('alberta', 6, 5) ('saskatchewan', 19, 4) ('river', 13, 13) ### ### ('.', 0, 30) ('canada', 7, 24) ('canadian', 12, 18) ('region', 78, 7) ('rivers', 28, 20) ### ### ('cities', 18, 28) ('city', 17, 29) ('metropolitan', 367, 3) ('regions', 331, 10) ### ### ('corridor', 942, 1) ('province', 45, 27) ('municipality', 307, 19) ('vancouver', 27, 38) ### ### ('metro', 1932, 2) ('population', 250, 25) ### ############################################################################################################ [2023-11-25 03:15:53,438][root][INFO] - Epoch: 1: Step: 13801/19530, loss[v]=0.104198, lr=0.000014, acc@1[1]=986.5/1024=0.96337890625, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 03:17:43,504][root][INFO] - Train batch 13900 [2023-11-25 03:17:43,505][root][INFO] - Avg. loss per last 100 batches: 0.228547 [2023-11-25 03:17:44,237][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.63 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 4.78 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: following amendments to the treaty on european union affected by the lisbon treaty in ### ### 2009, this provision now reads : the protection of life during pregnancy act 2013 replaced the ### ### abortion offences in the offences against the person act 1861 and made statutory provision for ### ### [P_TEXT]: mellet v ireland abortion in cases of fatal foetal abnormality violated the right to be ### ### free from torture, inhuman and degrading treatment or punishment. abortion was banned in nearly all ### ### cases in the republic of ireland until 2018. there was a constitutional protection for the unborn ### ### after the 1983 eighth amendment. the 2013 law, the protection of life during pregnancy act 2013 ### ### allowed for abortion in case of danger to the life of the mother. in june 2017 the citizens ### ### assembly recommended holding a referendum to repeal the 8th amendment. a referendum was held on may ### ### =============================================== v_q ================================================ ### ### ('pregnancy', 0, 0) ('person', 1, 1443) ('protection', 2, 29) ('abortion', 3, 1) ('act', 4, 23) ### ### ('life', 5, 27) ('2013', 6, 21) ('offences', 7, 4493) ('1861', 8, 1715) ('european', 9, 4720) ### ### ('against', 10, 10278) ('.', 11, 93) ('provision', 12, 4753) ('lisbon', 13, 16124) ### ### ('treaty', 14, 582) ('provisions', 15, 1690) ('union', 16, 3674) ('statutory', 17, 208) ### ### ('protect', 18, 50) ('2009', 19, 1359) ### ### =============================================== v_p ================================================ ### ### ('pregnancy', 0, 0) ('abortion', 3, 1) ('amendment', 45, 2) ('ireland', 633, 3) ### ### ('abnormal', 1040, 4) ('inhuman', 1883, 5) ('##let', 22071, 6) ('torture', 1759, 7) ### ### ('referendum', 527, 8) ('citizens', 1569, 9) ('mel', 18950, 10) ('eighth', 16710, 11) ### ### ('banned', 1572, 12) ('punishment', 613, 13) ('##tal', 14956, 14) ('danger', 999, 15) ### ### ('foe', 23990, 16) ('assembly', 2362, 17) ('repeal', 294, 18) ('##born', 26489, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('pregnancy', 0, 0) ('abortion', 3, 1) ('protection', 2, 29) ('act', 4, 23) ('life', 5, 27) ### ### ('2013', 6, 21) ('amendment', 45, 2) ('law', 59, 30) ('ireland', 633, 3) ### ### ('constitutional', 119, 28) ('protect', 18, 50) ('referendum', 527, 8) ('repeal', 294, 18) ### ### ('abnormal', 1040, 4) ('mother', 312, 20) ('fetal', 28, 42) ('punishment', 613, 13) ### ### ('pregnant', 21, 49) ('inhuman', 1883, 5) ('irish', 71, 39) ### ############################################################################################################ [2023-11-25 03:17:44,738][root][INFO] - Epoch: 1: Step: 13901/19530, loss[v]=0.333863, lr=0.000014, acc@1[1]=951.5/1024=0.92919921875, acc@1[2]=983.5/1024=0.96044921875 [2023-11-25 03:19:35,103][root][INFO] - Train batch 14000 [2023-11-25 03:19:35,104][root][INFO] - Avg. loss per last 100 batches: 0.223379 [2023-11-25 03:19:35,830][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.63 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 5.05 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.9/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it is named after the area of edinburgh it is located in, murrayfield. ### ### [P_TEXT]: murrayfield easter belmont road ( a private road ), is one of edinburgh's " millionaire's ### ### rows " and is home to local businessman david murray. aside from sports facilities, there is much ### ### residential land use and a private hospital, spire murrayfield hospital, part of the spire ### ### healthcare group. there are also a few shops, businesses and hotels. murrayfield is known for ### ### murrayfield stadium, home to the scottish national rugby union team and venue for many sporting ### ### events. in the shadow of the stadium is murrayfield ice rink, which currently hosts the edinburgh ### ### capitals ice hockey team and previously played host to the famous murrayfield racers. in the ### ### =============================================== v_q ================================================ ### ### ('murray', 0, 0) ('##field', 1, 4) ('edinburgh', 2, 8) ('scotland', 3, 36) ('scottish', 4, 14) ### ### ('.', 5, 176) ('stuart', 6, 1200) ('catering', 7, 96) ('named', 8, 11191) ('glasgow', 9, 72) ### ### ('naming', 10, 1453) ('suburb', 11, 12579) ('neighbourhood', 12, 13365) ('field', 13, 101) ### ### ('confusion', 14, 343) ('resides', 15, 702) ('scotia', 16, 8616) ('london', 17, 264) ### ### ('dundee', 18, 190) ('surname', 19, 2920) ### ### =============================================== v_p ================================================ ### ### ('murray', 0, 0) ('spire', 6977, 1) ('rows', 9411, 2) ('capitals', 2199, 3) ('##field', 1, 4) ### ### ('hockey', 5320, 5) ('belmont', 12808, 6) ('shadow', 13143, 7) ('edinburgh', 2, 8) ### ### ('racers', 4729, 9) ('millionaire', 20792, 10) ('rugby', 1823, 11) ('stadium', 180, 12) ### ### ('david', 3716, 13) ('scottish', 4, 14) ('healthcare', 6093, 15) ('ice', 6160, 16) ### ### ('hospital', 41, 17) ('easter', 16734, 18) ('private', 1354, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('murray', 0, 0) ('##field', 1, 4) ('edinburgh', 2, 8) ('scottish', 4, 14) ('scotland', 3, 36) ### ### ('hospital', 41, 17) ('stadium', 180, 12) ('road', 63, 25) ('capitals', 2199, 3) ('venue', 122, 27) ### ### ('glasgow', 9, 72) ('spire', 6977, 1) ('rugby', 1823, 11) ('racing', 512, 24) ('private', 1354, 19) ### ### ('.', 5, 176) ('rows', 9411, 2) ('hockey', 5320, 5) ('racers', 4729, 9) ('catering', 7, 96) ### ############################################################################################################ [2023-11-25 03:19:36,332][root][INFO] - Epoch: 1: Step: 14001/19530, loss[v]=0.116291, lr=0.000014, acc@1[1]=990.0/1024=0.966796875, acc@1[2]=1008.5/1024=0.98486328125 [2023-11-25 03:21:26,332][root][INFO] - Train batch 14100 [2023-11-25 03:21:26,333][root][INFO] - Avg. loss per last 100 batches: 0.211114 [2023-11-25 03:21:27,043][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 5.20 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the appalachian state mountaineers football team is the college football program at ### ### appalachian state university in boone, north carolina. ### ### [P_TEXT]: appalachian state – georgia southern football rivalry the appalachian state – georgia ### ### southern football rivalry, also known as deeper than hate, is a fierce college rivalry between the ### ### mountaineers of appalachian state university in boone, north carolina, and the eagles of georgia ### ### southern university in statesboro, georgia. despite being located in separate states, the two ### ### universities have similar academic profiles, both having developed from teachers'colleges, and ### ### having enrollments of approximately 20, 000. in addition, both appalachian and georgia southern ### ### historically held a very strong presence during their time in the ### ### =============================================== v_q ================================================ ### ### ('appalachian', 0, 1) ('##eers', 1, 6) ('football', 2, 8) ('state', 3, 11) ('boone', 4, 0) ### ### ('mountain', 5, 13) ('carolina', 6, 17) ('university', 7, 14) ('college', 8, 47) ('.', 9, 26) ### ### ('georgia', 10, 7) ('north', 11, 172) ('tennessee', 12, 123) ('team', 13, 1182) ### ### ('collegiate', 14, 1936) ('bowl', 15, 1040) ('kentucky', 16, 72) ('quarterback', 17, 707) ### ### ('nfl', 18, 393) ('ohio', 19, 87) ### ### =============================================== v_p ================================================ ### ### ('boone', 4, 0) ('appalachian', 0, 1) ('##boro', 24268, 2) ('eagles', 1540, 3) ('hate', 5163, 4) ### ### ('rivalry', 68, 5) ('##eers', 1, 6) ('georgia', 10, 7) ('football', 2, 8) ('southern', 1332, 9) ### ### ('profiles', 15607, 10) ('state', 3, 11) ('states', 28, 12) ('mountain', 5, 13) ### ### ('university', 7, 14) ('fierce', 1888, 15) ('eagle', 3599, 16) ('carolina', 6, 17) ### ### ('deeper', 12267, 18) ('teachers', 4261, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('appalachian', 0, 1) ('boone', 4, 0) ('##eers', 1, 6) ('football', 2, 8) ('state', 3, 11) ### ### ('mountain', 5, 13) ('carolina', 6, 17) ('university', 7, 14) ('georgia', 10, 7) ('rivalry', 68, 5) ### ### ('states', 28, 12) ('.', 9, 26) ('college', 8, 47) ('eagles', 1540, 3) ('##eer', 47, 22) ### ### ('hate', 5163, 4) ('southern', 1332, 9) ('kentucky', 16, 72) ('tennessee', 12, 123) ### ### ('north', 11, 172) ### ############################################################################################################ [2023-11-25 03:21:27,546][root][INFO] - Epoch: 1: Step: 14101/19530, loss[v]=0.401129, lr=0.000014, acc@1[1]=954.0/1024=0.931640625, acc@1[2]=980.5/1024=0.95751953125 [2023-11-25 03:23:17,164][root][INFO] - Train batch 14200 [2023-11-25 03:23:17,165][root][INFO] - Avg. loss per last 100 batches: 0.213162 [2023-11-25 03:23:17,885][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.70 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 4.93 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in response, konietzko issued a statement : after the group disbanded, schulz and esch ### ### formed the band slick idiot, while konietzko and skold regrouped as mdfmk, adding singer lucia ### ### cifarelli to form a trio. ### ### [P_TEXT]: mdfmk ( album ) mdfmk is an album released by industrial rock band mdfmk on march 28, ### ### 2000. it was the only studio album released by the band before it was discontinued and reformed ### ### once again as kmfdm. in a 2003 interview, band leader sascha konietzko said the album had sold ### ### around 120, 000 copies. ### ### =============================================== v_q ================================================ ### ### ('slick', 0, 3151) ('idiot', 1, 6064) ('##fare', 2, 26413) ('lucia', 3, 23359) ('##hul', 4, 28756) ### ### ('##fm', 5, 2) ('##lli', 6, 24757) ('##nie', 7, 10) ('##ch', 8, 7048) ('es', 9, 18367) ### ### ('##old', 10, 28198) ('md', 11, 0) ('ci', 12, 20417) ('##tz', 13, 5) ('sk', 14, 5827) ### ### ('sc', 15, 21912) ('##ʂ', 16, 892) ('ko', 17, 20) ('##k', 18, 8) ('.', 19, 1681) ### ### =============================================== v_p ================================================ ### ### ('md', 11, 0) ('##dm', 12138, 1) ('##fm', 5, 2) ('industrial', 5400, 3) ('km', 15873, 4) ### ### ('##tz', 13, 5) ('sas', 11307, 6) ('album', 21, 7) ('##k', 18, 8) ('##cha', 16225, 9) ### ### ('##nie', 7, 10) ('2000', 5721, 11) ('discontinued', 268, 12) ('##f', 2054, 13) ('##ko', 24, 14) ### ### ('rock', 166, 15) ('bands', 37, 16) ('2003', 10389, 17) ('band', 25, 18) ('albums', 134, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('md', 11, 0) ('##fm', 5, 2) ('##tz', 13, 5) ('##nie', 7, 10) ('##k', 18, 8) ('album', 21, 7) ### ### ('ko', 17, 20) ('##ko', 24, 14) ('band', 25, 18) ('##ك', 23, 22) ('bands', 37, 16) ### ### ('frontman', 33, 23) ('discontinued', 268, 12) ('rock', 166, 15) ('albums', 134, 19) ('コ', 29, 51) ### ### ('singers', 28, 56) ('industrial', 5400, 3) ('##क', 56, 31) ('musicians', 67, 33) ### ############################################################################################################ [2023-11-25 03:23:18,387][root][INFO] - Epoch: 1: Step: 14201/19530, loss[v]=0.287224, lr=0.000015, acc@1[1]=953.5/1024=0.93115234375, acc@1[2]=990.5/1024=0.96728515625 [2023-11-25 03:25:08,518][root][INFO] - Train batch 14300 [2023-11-25 03:25:08,518][root][INFO] - Avg. loss per last 100 batches: 0.208359 [2023-11-25 03:25:09,248][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.76 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 5.23 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: during the development of " rebelde way ", rojas and his co - stars : luisana lopilato, ### ### felipe colombo and camila bordonaba have become the members of the band erreway. ### ### [P_TEXT]: luisana lopilato morena. she starred with benjamin rojas, felipe colombo and camila ### ### bordonaba, with whom she formed a pop rock band, erreway. erreway released three studio albums — " ### ### senales " ( 2002 ), " tiempo " ( 2003 ) and " memoria " ( 2004 ) — which sold over 5 million ### ### copies. " senales " and " tiempo " earned platinum certification in argentina ( " senales " earned ### ### double platinum ) while " memoria " earned gold. their 2004 film, " erreway : 4 caminos ", was a ### ### box – office success. in 2007, lopilato left erreway and rojas, ### ### =============================================== v_q ================================================ ### ### ('##jas', 0, 2) ('ro', 1, 22) ('er', 2, 13) ('##rew', 3, 16) ('luisa', 4, 0) ('rebel', 5, 5733) ### ### ('colombo', 6, 3) ('way', 7, 13219) ('felipe', 8, 9) ('##ay', 9, 23) ('##pi', 10, 7) ### ### ('cam', 11, 21) ('band', 12, 34) ('##de', 13, 23472) ('philippines', 14, 1850) ('##rdon', 15, 6) ### ### ('##lat', 16, 8) ('bands', 17, 36) ('manila', 18, 2391) ('argentine', 19, 48) ### ### =============================================== v_p ================================================ ### ### ('luisa', 4, 0) ('memo', 20796, 1) ('##jas', 0, 2) ('colombo', 6, 3) ('##ales', 13988, 4) ### ### ('tie', 4023, 5) ('##rdon', 15, 6) ('##pi', 10, 7) ('##lat', 16, 8) ('felipe', 8, 9) ### ### ('benjamin', 3127, 10) ('sen', 16565, 11) ('##mp', 13120, 12) ('er', 2, 13) ('argentina', 44, 14) ### ### ('##ino', 10862, 15) ('##rew', 3, 16) ('##ila', 22, 17) ('more', 12824, 18) ('##aba', 23, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##jas', 0, 2) ('luisa', 4, 0) ('colombo', 6, 3) ('er', 2, 13) ('felipe', 8, 9) ('##pi', 10, 7) ### ### ('##rew', 3, 16) ('ro', 1, 22) ('##ay', 9, 23) ('##rdon', 15, 6) ('##lat', 16, 8) ('cam', 11, 21) ### ### ('##ila', 22, 17) ('band', 12, 34) ('##aba', 23, 19) ('lo', 20, 24) ('argentina', 44, 14) ### ### ('bands', 17, 36) ('##na', 25, 28) ('album', 24, 37) ### ############################################################################################################ [2023-11-25 03:25:09,753][root][INFO] - Epoch: 1: Step: 14301/19530, loss[v]=0.154379, lr=0.000015, acc@1[1]=980.5/1024=0.95751953125, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-25 03:26:59,659][root][INFO] - Train batch 14400 [2023-11-25 03:26:59,659][root][INFO] - Avg. loss per last 100 batches: 0.235186 [2023-11-25 03:27:00,364][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.52 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 4.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in december 2020, marston's took over the running of 156 pubs in wales from welsh brewer ### ### brains. ### ### [P_TEXT]: mild ale some notable examples of milds are : bank's mild, cain's dark mild, highgate ### ### dark mild, brain's dark, moorehouse blackcat, rudgate ruby mild, and theakston traditional mild a ### ### popular drink in the west midlands, " brown and mild " ( also known as a " boilermaker " ) is a ### ### half pint of draught mild served mixed with a half pint of bottled brown ale in a pint glass. in ### ### north west england, a mixture of half a pint of mild and half a ### ### =============================================== v_q ================================================ ### ### ('mars', 0, 20947) ('brains', 1, 860) ('##ton', 2, 27) ('wales', 3, 1813) ('brewer', 4, 5153) ### ### ('welsh', 5, 677) ('pubs', 6, 11087) ('pub', 7, 2483) ('brewery', 8, 198) ('shropshire', 9, 2461) ### ### ('2020', 10, 26868) ('.', 11, 457) ('brain', 12, 6) ('brewing', 13, 2195) ('drink', 14, 9) ### ### ('brew', 15, 975) ('tavern', 16, 2972) ('beer', 17, 57) ('156', 18, 7980) ('cardiff', 19, 2807) ### ### =============================================== v_p ================================================ ### ### ('draught', 78, 0) ('mild', 721, 1) ('##cat', 25570, 2) ('cain', 1756, 3) ('midlands', 1282, 4) ### ### ('ale', 366, 5) ('brain', 12, 6) ('ruby', 16439, 7) ('moore', 4669, 8) ('drink', 14, 9) ### ### ('boiler', 3796, 10) ('##aks', 29348, 11) ('dark', 1292, 12) ('bottled', 6495, 13) ### ### ('pin', 7852, 14) ('##maker', 7041, 15) ('##house', 7668, 16) ('bank', 1062, 17) ### ### ('brown', 5337, 18) ('glass', 177, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('brain', 12, 6) ('##ton', 2, 27) ('drink', 14, 9) ('draught', 78, 0) ('ale', 366, 5) ### ### ('mild', 721, 1) ('glass', 177, 19) ('england', 41, 28) ('midlands', 1282, 4) ('bottle', 25, 35) ### ### ('cain', 1756, 3) ('drinks', 84, 26) ('brains', 1, 860) ('dark', 1292, 12) ('brewery', 8, 198) ### ### ('bank', 1062, 17) ('welsh', 5, 677) ('drinking', 30, 43) ('beer', 17, 57) ('black', 1347, 20) ### ############################################################################################################ [2023-11-25 03:27:00,867][root][INFO] - Epoch: 1: Step: 14401/19530, loss[v]=0.324230, lr=0.000015, acc@1[1]=951.0/1024=0.9287109375, acc@1[2]=976.5/1024=0.95361328125 [2023-11-25 03:28:50,642][root][INFO] - Train batch 14500 [2023-11-25 03:28:50,643][root][INFO] - Avg. loss per last 100 batches: 0.222666 [2023-11-25 03:28:51,374][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.53 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.0/29522=99.99% | mean: 0.03 | max: 4.92 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: with 824, 529 m2 ( 82. 45 hectares [ 203. 7 acres ] ) of space in what was originally the ### ### atlantic forest, the zoo is south of the city of sao paulo. ### ### [P_TEXT]: sao paulo neighborhoods and buildings of historical value. the city has a large number of ### ### museums and art galleries. among the museums in the city are sao paulo museum of art ( masp ), the ### ### ipiranga museum, the museum of sacred art, the museum of the portuguese language, the pinacoteca do ### ### estado de sao paulo, among other renowned institutions. it also houses one of the top five zoos in ### ### the world, the sao paulo zoo. popularly known as " ipiranga museum ", the first monument built to ### ### preserve the memory of the independence of brazil, ### ### =============================================== v_q ================================================ ### ### ('zoo', 0, 2) ('paulo', 1, 1) ('forest', 2, 2261) ('atlantic', 3, 12118) ('sao', 4, 0) ### ### ('brazil', 5, 14) ('.', 6, 41) ('gerais', 7, 1444) ('brazilian', 8, 37) ('city', 9, 29) ### ### ('82', 10, 20166) ('space', 11, 5139) ('museum', 12, 8) ('52', 13, 8971) ('park', 14, 2129) ### ### ('porto', 15, 3019) ('janeiro', 16, 291) ('aquarium', 17, 846) ('absently', 18, 363) ### ### ('municipality', 19, 5277) ### ### =============================================== v_p ================================================ ### ### ('sao', 4, 0) ('paulo', 1, 1) ('zoo', 0, 2) ('ip', 9364, 3) ('sacred', 4762, 4) ('mas', 6936, 5) ### ### ('memory', 1965, 6) ('independence', 3638, 7) ('museum', 12, 8) ('##nga', 22178, 9) ### ### ('portuguese', 22, 10) ('##ira', 3500, 11) ('##aco', 17108, 12) ('language', 4386, 13) ### ### ('brazil', 5, 14) ('art', 2994, 15) ('pin', 11725, 16) ('monument', 428, 17) ('##tec', 28397, 18) ### ### ('neighborhoods', 3875, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('paulo', 1, 1) ('zoo', 0, 2) ('sao', 4, 0) ('brazil', 5, 14) ('museum', 12, 8) ### ### ('portuguese', 22, 10) ('city', 9, 29) ('.', 6, 41) ('brazilian', 8, 37) ('portugal', 27, 26) ### ### ('monument', 428, 17) ('memory', 1965, 6) ('sacred', 4762, 4) ('independence', 3638, 7) ### ### ('##ira', 3500, 11) ('forest', 2, 2261) ('cities', 30, 45) ('neighborhood', 279, 28) ### ### ('art', 2994, 15) ('museums', 1103, 21) ### ############################################################################################################ [2023-11-25 03:28:51,880][root][INFO] - Epoch: 1: Step: 14501/19530, loss[v]=0.151403, lr=0.000015, acc@1[1]=989.5/1024=0.96630859375, acc@1[2]=1008.0/1024=0.984375 [2023-11-25 03:30:41,963][root][INFO] - Train batch 14600 [2023-11-25 03:30:41,963][root][INFO] - Avg. loss per last 100 batches: 0.228288 [2023-11-25 03:30:42,692][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.73 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 5.20 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it also publishes the magazine " geoscientist " for fellows, and has a share in " geology ### ### today ", published by blackwell science. ### ### [P_TEXT]: geoscientist ( magazine ) geoscientist is a monthly magazine produced for the fellow ship ### ### of the geological society of london. the magazine was first published in 1990. it is editorially ### ### independent of the geological society's administration. it has a print run of 10, 000 and is freely ### ### available to all fellows. ### ### =============================================== v_q ================================================ ### ### ('blackwell', 0, 12799) ('##ient', 1, 2) ('geology', 2, 19) ('geo', 3, 3) ('fellows', 4, 1) ### ### ('today', 5, 4557) ('science', 6, 3045) ('##sc', 7, 8) ('##ist', 8, 18) ('magazine', 9, 11) ### ### ('geologist', 10, 98) ('geological', 11, 0) ('.', 12, 2745) ('editor', 13, 367) ('shape', 14, 306) ### ### ('oxford', 15, 74) ('review', 16, 323) ('rocks', 17, 297) ('publishing', 18, 38) ### ### ('shares', 19, 931) ### ### =============================================== v_p ================================================ ### ### ('geological', 11, 0) ('fellows', 4, 1) ('##ient', 1, 2) ('geo', 3, 3) ('administration', 1152, 4) ### ### ('ship', 2776, 5) ('society', 52, 6) ('fellow', 56, 7) ('##sc', 7, 8) ('london', 462, 9) ### ### ('print', 60, 10) ('magazine', 9, 11) ('monthly', 13186, 12) ('ships', 14770, 13) ### ### ('editorial', 311, 14) ('freely', 7089, 15) ('1990', 5736, 16) ('fellowship', 25, 17) ### ### ('##ist', 8, 18) ('geology', 2, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ient', 1, 2) ('fellows', 4, 1) ('geo', 3, 3) ('##sc', 7, 8) ('geological', 11, 0) ### ### ('geology', 2, 19) ('magazine', 9, 11) ('##ist', 8, 18) ('society', 52, 6) ('fellow', 56, 7) ### ### ('print', 60, 10) ('fellowship', 25, 17) ('london', 462, 9) ('administration', 1152, 4) ### ### ('editorial', 311, 14) ('publications', 38, 24) ('published', 20, 34) ('##ists', 39, 25) ### ### ('publishing', 18, 38) ('geologist', 10, 98) ### ############################################################################################################ [2023-11-25 03:30:43,197][root][INFO] - Epoch: 1: Step: 14601/19530, loss[v]=0.113748, lr=0.000015, acc@1[1]=1002.5/1024=0.97900390625, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 03:32:32,806][root][INFO] - Train batch 14700 [2023-11-25 03:32:32,807][root][INFO] - Avg. loss per last 100 batches: 0.212320 [2023-11-25 03:32:33,523][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.58 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.0/29522=99.99% | mean: 0.03 | max: 4.98 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the molucca sea plate is one of many tectonic features that compose the molucca sea ### ### collision complex, which refers to the tectonic relationship of the sangihe plate, halmahera plate, ### ### and the molucca ### ### [P_TEXT]: sorong fault sorong fault also ( sorong fault zone, sfz ) is an active, broad zone of ### ### inferred left lateral shear at the triple junction of the australian plate, eurasian plate, and ### ### pacific plates, where many plate fragments exist, such as the philippine sea plate, bird's head ### ### plate, halmahera plate and the molucca sea plate. it has been implicated in numerous large ### ### earthquakes. it is one of the two major faults created by the australian and pacific plate ### ### convergence, the other being the ramu - markham fault zone. the ### ### =============================================== v_q ================================================ ### ### ('##cca', 0, 13) ('plate', 1, 5) ('sea', 2, 18) ('sang', 3, 12115) ('mo', 4, 31) ('##ih', 5, 28140) ### ### ('##lu', 6, 25) ('collision', 7, 661) ('##cton', 8, 29415) ('hal', 9, 35) ('complex', 10, 4324) ### ### ('.', 11, 39) ('tanzania', 12, 5213) ('geological', 13, 365) ('hawaiian', 14, 1179) ### ### ('te', 15, 20949) ('volcanic', 16, 595) ('##era', 17, 27) ('continent', 18, 409) ### ### ('collided', 19, 319) ### ### =============================================== v_p ================================================ ### ### ('markham', 4662, 0) ('##rong', 29179, 1) ('fault', 821, 2) ('eurasian', 1889, 3) ### ### ('shear', 3132, 4) ('plate', 1, 5) ('convergence', 4121, 6) ('triple', 6809, 7) ### ### ('australian', 112, 8) ('plates', 174, 9) ('ram', 12828, 10) ('fragments', 2216, 11) ### ### ('sf', 20422, 12) ('##cca', 0, 13) ('so', 10561, 14) ('philippine', 206, 15) ('pacific', 1916, 16) ### ### ('bird', 4024, 17) ('sea', 2, 18) ('head', 8752, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('plate', 1, 5) ('##cca', 0, 13) ('sea', 2, 18) ('##lu', 6, 25) ('mo', 4, 31) ('hal', 9, 35) ### ### ('##era', 17, 27) ('australian', 112, 8) ('.', 11, 39) ('earthquake', 29, 26) ('plates', 174, 9) ### ### ('fault', 821, 2) ('philippine', 206, 15) ('australia', 99, 23) ('eurasian', 1889, 3) ### ### ('zone', 381, 22) ('##mah', 120, 30) ('markham', 4662, 0) ('shear', 3132, 4) ### ### ('convergence', 4121, 6) ### ############################################################################################################ [2023-11-25 03:32:34,025][root][INFO] - Epoch: 1: Step: 14701/19530, loss[v]=0.275708, lr=0.000015, acc@1[1]=957.5/1024=0.93505859375, acc@1[2]=989.5/1024=0.96630859375 [2023-11-25 03:34:23,969][root][INFO] - Train batch 14800 [2023-11-25 03:34:23,970][root][INFO] - Avg. loss per last 100 batches: 0.227812 [2023-11-25 03:34:24,707][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.75 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.0/29522=99.99% | mean: 0.03 | max: 5.09 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: was organized and led by charles haskell and elizabeth mcknight, and sponsored in part by ### ### the theodore roosevelt association, the american museum of natural history, the national wildlife ### ### federation and a private trust set up by haskell and mcknight. ### ### [P_TEXT]: american museum of natural history of the fossils on display represent unique and ### ### historic pieces that were collected during the museum's golden era of worldwide expeditions ( 1880s ### ### – 1930s ). examples of some of these expeditions, financed in whole or part by the amnh are : jesup ### ### north pacific expedition, the whitney south seas expedition, the roosevelt – rondon scientific ### ### expedition, the crocker land expedition, and the expeditions to madagascar and new guinea by ### ### richard archbold. on a smaller scale, expeditions continue into the present. the museum also ### ### publishes several peer - reviewed journals, including the " bulletin of the american museum of ### ### natural history ". in 1976, animal rights activist henry spira led a campaign against ### ### =============================================== v_q ================================================ ### ### ('wildlife', 0, 1405) ('##kell', 1, 28832) ('##night', 2, 28784) ('museum', 3, 3) ### ### ('roosevelt', 4, 16) ('elizabeth', 5, 11836) ('federation', 6, 10549) ('theodore', 7, 15001) ### ### ('charles', 8, 9393) ('trust', 9, 2622) ('association', 10, 8595) ('mc', 11, 20396) ### ### ('american', 12, 34) ('history', 13, 35) ('natural', 14, 14) ('national', 15, 10520) ### ### ('has', 16, 14040) ('##k', 17, 20305) ('americans', 18, 37) ('sponsored', 19, 743) ### ### =============================================== v_p ================================================ ### ### ('##bold', 28571, 0) ('whitney', 422, 1) ('##ira', 22556, 2) ('museum', 3, 3) ### ### ('expedition', 772, 4) ('madagascar', 634, 5) ('seas', 17983, 6) ('##nh', 19333, 7) ### ### ('bulletin', 7082, 8) ('golden', 5102, 9) ('##ock', 17326, 10) ('land', 4072, 11) ### ### ('scientific', 491, 12) ('##don', 25516, 13) ('natural', 14, 14) ('arch', 3865, 15) ### ### ('roosevelt', 4, 16) ('expeditions', 16265, 17) ('pacific', 3845, 18) ('##su', 17929, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('museum', 3, 3) ('roosevelt', 4, 16) ('natural', 14, 14) ('american', 12, 34) ('history', 13, 35) ### ### ('americans', 18, 37) ('whitney', 422, 1) ('animal', 63, 27) ('madagascar', 634, 5) ### ### ('scientific', 491, 12) ('expedition', 772, 4) ('rights', 426, 32) ('henry', 450, 33) ### ### ('review', 193, 40) ('wildlife', 0, 1405) ('.', 21, 108) ('era', 1412, 29) ('land', 4072, 11) ### ### ('arch', 3865, 15) ('golden', 5102, 9) ### ############################################################################################################ [2023-11-25 03:34:25,209][root][INFO] - Epoch: 1: Step: 14801/19530, loss[v]=0.113010, lr=0.000015, acc@1[1]=989.0/1024=0.9658203125, acc@1[2]=1011.5/1024=0.98779296875 [2023-11-25 03:36:15,178][root][INFO] - Train batch 14900 [2023-11-25 03:36:15,179][root][INFO] - Avg. loss per last 100 batches: 0.225084 [2023-11-25 03:36:15,907][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.73 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 5.10 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he won the nabf belt with a split decision over the highly ranked orlin norris, and ### ### successfully defended it against future world champion oliver mccall, winning another 12 - round ### ### decision. ### ### [P_TEXT]: orlin norris decked in the 8th round and twisted his knee, unable to continue. in 1990 ### ### norris regained some ground by beating future world champion oliver mccall by split decision, and ### ### in 1991 regained his nabf title with a first - round knockout of lionel washington. he defended the ### ### title against ex - world champion tony tucker ; however, he lost on a controversial split decision, ### ### after it appeared norris had outboxed his giant opponent rather easily. after the tucker ### ### performance, norris struggled to find contenders who would risk fighting him, and was forced to ### ### move down a division, into the cruiserweights. ### ### =============================================== v_q ================================================ ### ### ('mccall', 0, 3) ('oliver', 1, 5) ('##bf', 2, 1) ('norris', 3, 4) ('##lin', 4, 7) ('na', 5, 24) ### ### ('boxing', 6, 35) ('world', 7, 17) ('martial', 8, 17249) ('fight', 9, 92) ('bout', 10, 1148) ### ### ('or', 11, 19) ('wrestling', 12, 26) ('defend', 13, 187) ('decision', 14, 18) ('boxer', 15, 1149) ### ### ('opponents', 16, 55) ('fights', 17, 218) ('against', 18, 80) ('opponent', 19, 31) ### ### =============================================== v_p ================================================ ### ### ('tucker', 593, 0) ('##bf', 2, 1) ('lionel', 20157, 2) ('mccall', 0, 3) ('norris', 3, 4) ### ### ('oliver', 1, 5) ('washington', 1355, 6) ('##lin', 4, 7) ('tony', 177, 8) ('deck', 3133, 9) ### ### ('split', 20, 10) ('ex', 13652, 11) ('knockout', 1856, 12) ('cruiser', 12229, 13) ### ### ('giant', 5205, 14) ('1990', 4452, 15) ('knee', 654, 16) ('world', 7, 17) ('decision', 14, 18) ### ### ('or', 11, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('mccall', 0, 3) ('##bf', 2, 1) ('oliver', 1, 5) ('norris', 3, 4) ('##lin', 4, 7) ('world', 7, 17) ### ### ('na', 5, 24) ('split', 20, 10) ('or', 11, 19) ('decision', 14, 18) ('boxing', 6, 35) ### ### ('wrestling', 12, 26) ('future', 24, 28) ('opponent', 19, 31) ('tony', 177, 8) ('tucker', 593, 0) ### ### ('champion', 32, 38) ('round', 87, 25) ('opponents', 16, 55) ('fight', 9, 92) ### ############################################################################################################ [2023-11-25 03:36:16,411][root][INFO] - Epoch: 1: Step: 14901/19530, loss[v]=0.323951, lr=0.000015, acc@1[1]=952.0/1024=0.9296875, acc@1[2]=985.0/1024=0.9619140625 [2023-11-25 03:38:05,956][root][INFO] - Train batch 15000 [2023-11-25 03:38:05,956][root][INFO] - Avg. loss per last 100 batches: 0.205755 [2023-11-25 03:38:06,683][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.72 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=99.99% | mean: 0.03 | max: 5.15 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he sensed during shooting that the production was " going off the rails ", and announced ### ### that the director, stephen norrington should be " locked up for insanity ". ### ### [P_TEXT]: the league of extraordinary gentlemen ( film ) the league of extraordinary gentlemen, ### ### also promoted as lxg, is a 2003 dieselpunk superhero film loosely based on the first volume of the ### ### comic book series of the same name by alan moore and kevin o'neill. distributed by 20th century ### ### fox, it was released on 11 july 2003 in the united states, and 17 october 2003 in the united ### ### kingdom. it was directed by stephen norrington and starred sean connery, naseeruddin shah, peta ### ### wilson, ### ### =============================================== v_q ================================================ ### ### ('##rington', 0, 1) ('rails', 1, 5395) ('insanity', 2, 14900) ('stephen', 3, 11) ('nor', 4, 7) ### ### ('off', 5, 8806) ('film', 6, 16) ('locked', 7, 5337) ('going', 8, 8578) ('sensed', 9, 991) ### ### ('films', 10, 24) ('shakespeare', 11, 7765) ('theatre', 12, 228) ('production', 13, 174) ### ### ('productions', 14, 310) ('filming', 15, 35) ('musicals', 16, 113) ('broadway', 17, 1909) ### ### ('should', 18, 3447) ('"', 19, 20559) ### ### =============================================== v_p ================================================ ### ### ('neill', 17029, 0) ('##rington', 0, 1) ('gentlemen', 15195, 2) ('diesel', 6290, 3) ### ### ('extraordinary', 6981, 4) ('##uddin', 21244, 5) ('pet', 17576, 6) ('nor', 4, 7) ### ### ('conner', 11630, 8) ('moore', 2666, 9) ('league', 10289, 10) ('stephen', 3, 11) ('kevin', 895, 12) ### ### ('wilson', 2391, 13) ('shah', 17286, 14) ('fox', 2595, 15) ('film', 6, 16) ('superhero', 15249, 17) ### ### ('##eer', 21786, 18) ('alan', 2727, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##rington', 0, 1) ('nor', 4, 7) ('stephen', 3, 11) ('film', 6, 16) ('films', 10, 24) ### ### ('filming', 15, 35) ('movie', 30, 40) ('kevin', 895, 12) ('filmed', 156, 36) ('series', 308, 28) ### ### ('directing', 141, 43) ('dramas', 40, 58) ('sean', 536, 27) ('moore', 2666, 9) ('movies', 98, 54) ### ### ('wilson', 2391, 13) ('neither', 33, 70) ('starred', 284, 42) ('fox', 2595, 15) ### ### ('production', 13, 174) ### ############################################################################################################ [2023-11-25 03:38:07,184][root][INFO] - Epoch: 1: Step: 15001/19530, loss[v]=0.352014, lr=0.000015, acc@1[1]=946.0/1024=0.923828125, acc@1[2]=966.0/1024=0.943359375 [2023-11-25 03:39:56,923][root][INFO] - Train batch 15100 [2023-11-25 03:39:56,924][root][INFO] - Avg. loss per last 100 batches: 0.216652 [2023-11-25 03:39:57,665][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.64 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 4.97 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: she participates in " kof xi " and " kof xiii " with k'and maxima as her allies, having ### ### also joined forces with the ikari team. ### ### [P_TEXT]: the king of fighters xiii 7 and one week later, the final character for k's team, kula ### ### diamond, was introduced in the same way. saiki, the manipulator behind the events of " the king of ### ### fighters xiii ", appears as sub - boss while an alternate form of ash crimson known as evil ash ### ### appears as the final boss. on june 28, 2011, " famitsu " reported the addition of billy kane from " ### ### fatal fury : king of fighters " as a console version exclusive character. on july 29, 2011, the ### ### human form of saiki was also announced to be playable, while the original iori yagami to be ### ### =============================================== v_q ================================================ ### ### ('maxim', 0, 23936) ('ik', 1, 15555) ('xi', 2, 673) ('ko', 3, 9867) ('xiii', 4, 0) ### ### ('##f', 5, 17293) ('##ari', 6, 17796) ('k', 7, 18) ('team', 8, 16) ('！', 9, 297) ('##a', 10, 20608) ### ### ('##士', 11, 251) ('freaking', 12, 415) ('ninja', 13, 6433) ('abilities', 14, 1311) ### ### ('ally', 15, 5361) ('comics', 16, 197) ('weapon', 17, 2061) ('storyline', 18, 2547) ### ### ('titans', 19, 7161) ### ### =============================================== v_p ================================================ ### ### ('xiii', 4, 0) ('fatal', 604, 1) ('fighters', 202, 2) ('crimson', 2371, 3) ('sai', 18406, 4) ### ### ('billy', 5310, 5) ('ash', 11651, 6) ('##mit', 26226, 7) ('fury', 31, 8) ('diamond', 7851, 9) ### ### ('##gami', 9291, 10) ('kane', 74, 11) ('evil', 445, 12) ('io', 2994, 13) ('console', 4439, 14) ### ### ('##lat', 24262, 15) ('team', 8, 16) ('king', 7832, 17) ('k', 7, 18) ('##la', 17175, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('xiii', 4, 0) ('k', 7, 18) ('team', 8, 16) ('fury', 31, 8) ('kane', 74, 11) ('fighters', 202, 2) ### ### ('fatal', 604, 1) ('evil', 445, 12) ('playable', 85, 27) ('characters', 72, 30) ### ### ('fighter', 396, 20) ('crimson', 2371, 3) ('human', 520, 25) ('appears', 66, 44) ### ### ('events', 218, 41) ('io', 2994, 13) ('character', 282, 38) ('exclusive', 999, 28) ### ### ('billy', 5310, 5) ('xi', 2, 673) ### ############################################################################################################ [2023-11-25 03:39:58,168][root][INFO] - Epoch: 1: Step: 15101/19530, loss[v]=0.104751, lr=0.000015, acc@1[1]=989.0/1024=0.9658203125, acc@1[2]=1010.5/1024=0.98681640625 [2023-11-25 03:41:48,177][root][INFO] - Train batch 15200 [2023-11-25 03:41:48,178][root][INFO] - Avg. loss per last 100 batches: 0.197888 [2023-11-25 03:41:48,880][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 5.10 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.5/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the province is divided into the following twelve counties ( shahrestan s ) : aliabad ### ### county, aqqala county, azadshahr county, bandar - e gaz county, gonbad - e qabus county, gorgan ### ### county, kalaleh county, kordkuy county, maraveh tappeh county, minudasht county, ramian county, and ### ### torkaman county. ### ### [P_TEXT]: fenderesk - e shomali rural district fenderesk - e shomali rural district ( ) is a rural ### ### district ( " dehestan " ) in fenderesk district, ramian county, golestan province, iran. at the ### ### 2006 census, its population was 7, 176, in 1, 752 families. the rural district has 8 villages. ### ### =============================================== v_q ================================================ ### ### ('county', 0, 9) ('##kam', 1, 13118) ('##uda', 2, 27508) ('ram', 3, 2) ('province', 4, 56) ### ### ('##rgan', 5, 28526) ('##ads', 6, 28566) ('iran', 7, 88) ('kala', 8, 12330) ('go', 9, 10) ### ### ('##le', 10, 497) ('##ian', 11, 3) ('##rd', 12, 26277) ('tor', 13, 19314) ('district', 14, 8) ### ### ('twelve', 15, 5918) ('##rest', 16, 27864) ('min', 17, 21312) ('##ad', 18, 25739) ### ### ('banda', 19, 25498) ### ### =============================================== v_p ================================================ ### ### ('fender', 16425, 0) ('##mal', 18194, 1) ('ram', 3, 2) ('##ian', 11, 3) ('##les', 701, 4) ### ### ('districts', 905, 5) ('sho', 16515, 6) ('##es', 19805, 7) ('district', 14, 8) ('county', 0, 9) ### ### ('go', 9, 10) ('counties', 34, 11) ('families', 128, 12) ('villages', 10864, 13) ('##k', 15334, 14) ### ### ('population', 721, 15) ('provinces', 240, 16) ('75', 15353, 17) ('village', 157, 18) ### ### ('rural', 26, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ram', 3, 2) ('county', 0, 9) ('##ian', 11, 3) ('go', 9, 10) ('district', 14, 8) ### ### ('counties', 34, 11) ('rural', 26, 19) ('province', 4, 56) ('families', 128, 12) ('##les', 701, 4) ### ### ('iran', 7, 88) ('districts', 905, 5) ('anyway', 46, 22) ('awkward', 47, 23) ('##le', 10, 497) ### ### ('village', 157, 18) ('slumped', 56, 36) ('##セ', 30, 79) ('provinces', 240, 16) ### ### ('absently', 23, 286) ### ############################################################################################################ [2023-11-25 03:41:49,383][root][INFO] - Epoch: 1: Step: 15201/19530, loss[v]=0.359936, lr=0.000016, acc@1[1]=951.0/1024=0.9287109375, acc@1[2]=977.0/1024=0.9541015625 [2023-11-25 03:43:38,703][root][INFO] - Train batch 15300 [2023-11-25 03:43:38,704][root][INFO] - Avg. loss per last 100 batches: 0.222373 [2023-11-25 03:43:39,424][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.73 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 5.27 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: ( consisting of andrew forson, graviton, jude the entropic man, mentallo, yelena belova, ### ### and the ### ### [P_TEXT]: secret avengers after taskmaster agrees to join the secret avengers, he goes on a mission ### ### to infiltrate a. i. m.'s new high council ( consisting of andrew forson, graviton, jude the ### ### entropic man, mentallo, superia, and yelena belova ). when daisy johnson and nick fury jr. were ### ### attending the weapons expo, they meet up with senator robert ralston. daisy notices that one of the ### ### items at the expo is the iron patriot armor. a. i. m. then attacks the weapons expo which leads to ### ### the supposed death of senator robert ralston during a. i. m.'s fight ### ### =============================================== v_q ================================================ ### ### ('mental', 0, 7) ('##lena', 1, 31) ('##lo', 2, 13) ('andrew', 3, 23) ('##son', 4, 18) ('ye', 5, 50) ### ### ('##ova', 6, 11) ('##avi', 7, 9) ('##tro', 8, 36) ('##ton', 9, 35) ('##pic', 10, 14) ### ### ('bel', 11, 33) ('jude', 12, 17) ('absently', 13, 383) ('man', 14, 68) ('mutant', 15, 9343) ### ### ('storyline', 16, 4825) ('.', 17, 212) ('gr', 18, 63) ('comics', 19, 1146) ### ### =============================================== v_p ================================================ ### ### ('avengers', 22, 0) ('##lston', 20169, 1) ('patriot', 12580, 2) ('secret', 338, 3) ### ### ('expo', 15012, 4) ('daisy', 3662, 5) ('fury', 313, 6) ('mental', 0, 7) ('super', 5675, 8) ### ### ('##avi', 7, 9) ('high', 6656, 10) ('##ova', 6, 11) ('##master', 1947, 12) ('##lo', 2, 13) ### ### ('##pic', 10, 14) ('weapons', 5616, 15) ('iron', 5612, 16) ('jude', 12, 17) ('##son', 4, 18) ### ### ('task', 3822, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('mental', 0, 7) ('##lo', 2, 13) ('##son', 4, 18) ('andrew', 3, 23) ('##avi', 7, 9) ### ### ('##lena', 1, 31) ('##ova', 6, 11) ('##pic', 10, 14) ('avengers', 22, 0) ('jude', 12, 17) ### ### ('bel', 11, 33) ('##tro', 8, 36) ('##ton', 9, 35) ('ye', 5, 50) ('secret', 338, 3) ('fury', 313, 6) ### ### ('man', 14, 68) ('infiltrate', 429, 22) ('gr', 18, 63) ('en', 62, 45) ### ############################################################################################################ [2023-11-25 03:43:39,926][root][INFO] - Epoch: 1: Step: 15301/19530, loss[v]=0.283332, lr=0.000016, acc@1[1]=959.5/1024=0.93701171875, acc@1[2]=988.0/1024=0.96484375 [2023-11-25 03:45:29,756][root][INFO] - Train batch 15400 [2023-11-25 03:45:29,757][root][INFO] - Avg. loss per last 100 batches: 0.218537 [2023-11-25 03:45:30,484][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.80 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 5.05 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: command decision was a 1948 play in three acts written by william wister haines, and ### ### formed the basis for his best - selling novel of the same title. ### ### [P_TEXT]: command decision ( film ) command decision is a 1949 war film released by metro - goldwyn ### ### - mayer starring clark gable, walter pidgeon, van johnson, and brian donlevy, and directed by sam ### ### wood, based on the 1948 stage play of the same name written by william wister haines, which he ### ### based on his best - selling 1947 novel. the screenplay for the film was written by george froeschel ### ### and william r. laidlaw. haines'play ran on broadway for almost a year beginning in october ### ### =============================================== v_q ================================================ ### ### ('command', 0, 5) ('decision', 1, 0) ('1948', 2, 15) ('##nes', 3, 19) ('novel', 4, 28) ### ### ('##ster', 5, 24) ('hai', 6, 14) ('play', 7, 27) ('decisions', 8, 26) ('wi', 9, 22) ### ### ('william', 10, 45) ('decide', 11, 181) ('acts', 12, 6997) ('commands', 13, 33) ('1947', 14, 23) ### ### ('playwright', 15, 123) ('theatre', 16, 68) ('plays', 17, 29) ('act', 18, 2633) ('1949', 19, 11) ### ### =============================================== v_p ================================================ ### ### ('decision', 1, 0) ('##lev', 29348, 1) ('gable', 5991, 2) ('##oes', 12225, 3) ('##dge', 23622, 4) ### ### ('command', 0, 5) ('wood', 10370, 6) ('sam', 6881, 7) ('laid', 3569, 8) ('##chel', 27685, 9) ### ### ('johnson', 4146, 10) ('1949', 19, 11) ('van', 17558, 12) ('mayer', 7639, 13) ('hai', 6, 14) ### ### ('1948', 2, 15) ('fr', 17270, 16) ('selling', 23, 17) ('brian', 2962, 18) ('##nes', 3, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('decision', 1, 0) ('command', 0, 5) ('1948', 2, 15) ('hai', 6, 14) ('##nes', 3, 19) ### ### ('##ster', 5, 24) ('novel', 4, 28) ('play', 7, 27) ('decisions', 8, 26) ('wi', 9, 22) ### ### ('1949', 19, 11) ('1947', 14, 23) ('commands', 13, 33) ('plays', 17, 29) ('selling', 23, 17) ### ### ('william', 10, 45) ('film', 84, 21) ('same', 27, 46) ('1946', 25, 54) ('broadway', 42, 42) ### ############################################################################################################ [2023-11-25 03:45:30,989][root][INFO] - Epoch: 1: Step: 15401/19530, loss[v]=0.347291, lr=0.000016, acc@1[1]=950.0/1024=0.927734375, acc@1[2]=977.0/1024=0.9541015625 [2023-11-25 03:47:20,774][root][INFO] - Train batch 15500 [2023-11-25 03:47:20,775][root][INFO] - Avg. loss per last 100 batches: 0.210256 [2023-11-25 03:47:21,485][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.50 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 5.66 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: herfølge is a suburb of the town of køge, located about 5 kilometers south of central ### ### køge, and is a part of køge municipality. ### ### [P_TEXT]: køge a merger of professional football between herfølge boldklub and køge boldklub in ### ### 2009. it currently plays in the danish 1st division. their home ground is named herfølge stadion. ### ### it is used and serves as the primary training fields and facilities of the professional club, ### ### besides the now amateur mother club of herfølge bk. the facilities is situated in the suburb of ### ### herfølge, about 5 kilometres south of central køge. though the same club was in 2016 abolished ### ### after around half of the team screamed the n - word towards the club paris saint ### ### =============================================== v_q ================================================ ### ### ('##ge', 0, 0) ('##ø', 1, 3) ('##lge', 2, 1) ('her', 3, 18) ('k', 4, 11) ('municipality', 5, 4831) ### ### ('##f', 6, 24) ('central', 7, 17) ('suburb', 8, 12) ('absently', 9, 650) ('denmark', 10, 37) ### ### ('norwegian', 11, 99) ('ヘ', 12, 169) ('norway', 13, 96) ('##ʂ', 14, 1693) ('danish', 15, 7) ### ### ('³', 16, 170) ('copenhagen', 17, 60) ('ɡ', 18, 1673) ('ק', 19, 110) ### ### =============================================== v_p ================================================ ### ### ('##ge', 0, 0) ('##lge', 2, 1) ('bold', 4691, 2) ('##ø', 1, 3) ('paris', 1202, 4) ### ### ('##lub', 16112, 5) ('stadion', 9888, 6) ('danish', 15, 7) ('bk', 885, 8) ('1st', 7610, 9) ### ### ('division', 3100, 10) ('k', 4, 11) ('suburb', 8, 12) ('mother', 1737, 13) ('club', 16516, 14) ### ### ('saint', 5222, 15) ('amateur', 8428, 16) ('central', 7, 17) ('her', 3, 18) ('word', 3641, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ge', 0, 0) ('##lge', 2, 1) ('##ø', 1, 3) ('k', 4, 11) ('her', 3, 18) ('suburb', 8, 12) ### ### ('central', 7, 17) ('danish', 15, 7) ('##f', 6, 24) ('denmark', 10, 37) ('suburbs', 84, 26) ### ### ('anyway', 42, 40) ('paris', 1202, 4) ('bk', 885, 8) ('copenhagen', 17, 60) ('norwegian', 11, 99) ### ### ('she', 48, 51) ('bold', 4691, 2) ('mother', 1737, 13) ('norway', 13, 96) ### ############################################################################################################ [2023-11-25 03:47:21,986][root][INFO] - Epoch: 1: Step: 15501/19530, loss[v]=0.095725, lr=0.000016, acc@1[1]=995.5/1024=0.97216796875, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 03:49:11,323][root][INFO] - Train batch 15600 [2023-11-25 03:49:11,324][root][INFO] - Avg. loss per last 100 batches: 0.205496 [2023-11-25 03:49:12,046][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.62 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.6/29522=99.99% | mean: 0.03 | max: 5.09 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: shipitsyno ( ) is an urban locality ( an urban - type settlement ) in kotlassky district ### ### of arkhangelsk oblast, russia, located on the left bank of the northern dvina river from kotlas. ### ### [P_TEXT]: kotlassky district s, one town of district significance ( solvychegodsk ), and two urban ### ### - type settlements with jurisdictional territory ( privodino and shipitsyno ). three other ### ### inhabited localities which previously had work settlement status were downgraded to rural status in ### ### 2005 ; these are cheryomushsky, kharitonovo, and udimsky. the following selsoviets have been ### ### established ( the administrative centers are given in parentheses ) : as a municipal division, the ### ### district is incorporated as kotlassky municipal district. the town of oblast significance of kotlas ### ### is incorporated separately from the district as kotlas urban ### ### =============================================== v_q ================================================ ### ### ('ship', 0, 4) ('##its', 1, 3) ('##tl', 2, 1) ('##vina', 3, 12420) ('##yn', 4, 28) ('##ass', 5, 0) ### ### ('ד', 6, 214) ('oblast', 7, 50) ('##as', 8, 29) ('russia', 9, 179) ('ko', 10, 36) ('##ky', 11, 25) ### ### ('d', 12, 18822) ('ships', 13, 24) ('##hang', 14, 29311) ('##ो', 15, 59) ('##o', 16, 98) ### ### ('ད', 17, 678) ('locality', 18, 76) ('belarus', 19, 8390) ### ### =============================================== v_p ================================================ ### ### ('##ass', 5, 0) ('##tl', 2, 1) ('##hari', 27340, 2) ('##its', 1, 3) ('ship', 0, 4) ### ### ('significance', 80, 5) ('##din', 16864, 6) ('cher', 9732, 7) ('jurisdiction', 6804, 8) ### ### ('ud', 20427, 9) ('sol', 8346, 10) ('parentheses', 20939, 11) ('##grade', 21278, 12) ### ### ('pri', 12953, 13) ('##vo', 2415, 14) ('##mus', 12750, 15) ('##im', 12146, 16) ### ### ('districts', 5856, 17) ('##ovo', 2044, 18) ('municipal', 1085, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##tl', 2, 1) ('##its', 1, 3) ('ship', 0, 4) ('##ass', 5, 0) ('##yn', 4, 28) ('##as', 8, 29) ### ### ('##ky', 11, 25) ('ships', 13, 24) ('oblast', 7, 50) ('ko', 10, 36) ('significance', 80, 5) ### ### ('settlement', 25, 31) ('##sky', 30, 32) ('district', 68, 20) ('urban', 28, 38) ('##ो', 15, 59) ### ### ('ד', 6, 214) ('russia', 9, 179) ('locality', 18, 76) ('コ', 21, 79) ### ############################################################################################################ [2023-11-25 03:49:12,549][root][INFO] - Epoch: 1: Step: 15601/19530, loss[v]=0.142390, lr=0.000016, acc@1[1]=981.5/1024=0.95849609375, acc@1[2]=1008.0/1024=0.984375 [2023-11-25 03:51:02,448][root][INFO] - Train batch 15700 [2023-11-25 03:51:02,449][root][INFO] - Avg. loss per last 100 batches: 0.216369 [2023-11-25 03:51:03,183][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.52 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 4.96 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: white people is the second studio album by handsome boy modeling school. ### ### [P_TEXT]: mike patton eyvind kang and his ensemble playground to play the piece " virginal co ### ### ordinates " at the festival internazionale di musica in bologna. the performance would be release ### ### as an album in 2003. in 2004, patton worked with bjork and the beat boxer rahzel on the album " ### ### medulla ". that same year, patton released the album " romances " with kaada and contributed vocals ### ### to the album " white people " by handsome boy modeling school ( dan the automator and prince paul ### ### ). in 2005, patton collaborated with hip - hop dj trio and turntablists the x - ecutioners to ### ### release ### ### =============================================== v_q ================================================ ### ### ('modeling', 0, 7) ('white', 1, 15) ('handsome', 2, 8) ('people', 3, 18) ('boy', 4, 35) ### ### ('school', 5, 32) ('album', 6, 46) ('model', 7, 64) ('modelling', 8, 30) ('albums', 9, 65) ### ### ('##₂', 10, 345) ('models', 11, 51) ('peoples', 12, 58) ('person', 13, 125) ('persons', 14, 154) ### ### ('absently', 15, 617) ('individuals', 16, 157) ('lp', 17, 268) ('acts', 18, 4979) ('.', 19, 1162) ### ### =============================================== v_p ================================================ ### ### ('patton', 18075, 0) ('bologna', 4492, 1) ('playground', 9823, 2) ('##ulla', 21314, 3) ### ### ('##ada', 22498, 4) ('kang', 13731, 5) ('mike', 708, 6) ('modeling', 0, 7) ('handsome', 2, 8) ### ### ('romance', 5483, 9) ('musica', 4707, 10) ('##dina', 29057, 11) ('##zel', 25512, 12) ### ### ('boxer', 6511, 13) ('med', 22229, 14) ('white', 1, 15) ('virgin', 2417, 16) ('##onale', 24229, 17) ### ### ('people', 3, 18) ('##ution', 25522, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('modeling', 0, 7) ('handsome', 2, 8) ('white', 1, 15) ('people', 3, 18) ('boy', 4, 35) ### ### ('school', 5, 32) ('album', 6, 46) ('modelling', 8, 30) ('model', 7, 64) ('models', 11, 51) ### ### ('schooling', 22, 42) ('peoples', 12, 58) ('albums', 9, 65) ('mike', 708, 6) ('trio', 417, 20) ### ### ('djs', 237, 31) ('dj', 168, 41) ('lyrical', 33, 57) ('bologna', 4492, 1) ('prince', 1377, 22) ### ############################################################################################################ [2023-11-25 03:51:03,687][root][INFO] - Epoch: 1: Step: 15701/19530, loss[v]=0.242940, lr=0.000016, acc@1[1]=961.5/1024=0.93896484375, acc@1[2]=999.5/1024=0.97607421875 [2023-11-25 03:52:53,137][root][INFO] - Train batch 15800 [2023-11-25 03:52:53,138][root][INFO] - Avg. loss per last 100 batches: 0.219881 [2023-11-25 03:52:53,840][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.69 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 5.20 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: and ensure that there are no additional heart arrhythmias present that might require ### ### specific attention, such as ventricular tachycardia. ### ### [P_TEXT]: heart rate turbulence dying from cardiac disease after a patient suffers a myocardial ### ### infarction. hrt can also be used to predict death in patients with congestive heart failure from a ### ### lethal arrhythmia. hrt has nothing to do with turbulence in fluid dynamics. the concept of hrt was ### ### introduced to the medical community by georg schmidt and colleagues from the technical university ### ### of munich in 1999 in the british medical journal the lancet. while studying pvc characteristics, ### ### schmidt and his colleagues had noticed that heart rate ### ### =============================================== v_q ================================================ ### ### ('heart', 0, 16) ('##chy', 1, 28101) ('##card', 2, 27713) ('ar', 3, 46) ('cardiac', 4, 15) ### ### ('.', 5, 774) ('vent', 6, 21722) ('absently', 7, 701) ('##ˢ', 8, 1858) ('specific', 9, 1086) ### ### ('##rh', 10, 26) ('##phobia', 11, 10577) ('diagnosed', 12, 6497) ('attention', 13, 6328) ### ### ('##ular', 14, 14712) ('cancer', 15, 156) ('ta', 16, 22046) ('¹', 17, 245) ('stomach', 18, 686) ### ### ('pmid', 19, 189) ### ### =============================================== v_p ================================================ ### ### ('turbulence', 3102, 0) ('hr', 8742, 1) ('schmidt', 16623, 2) ('dynamics', 5939, 3) ### ### ('lance', 19080, 4) ('georg', 18328, 5) ('##hmi', 41, 6) ('pv', 28446, 7) ('cong', 16483, 8) ### ### ('rate', 674, 9) ('fluid', 1325, 10) ('characteristics', 9960, 11) ('technical', 6584, 12) ### ### ('lethal', 876, 13) ('munich', 7175, 14) ('cardiac', 4, 15) ('heart', 0, 16) ('medical', 29, 17) ### ### ('journal', 1287, 18) ('failure', 5458, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('heart', 0, 16) ('cardiac', 4, 15) ('##rh', 10, 26) ('##hmi', 41, 6) ('ar', 3, 46) ### ### ('medical', 29, 17) ('##yt', 20, 39) ('rate', 674, 9) ('patient', 147, 32) ('turbulence', 3102, 0) ### ### ('noticed', 197, 29) ('lethal', 876, 13) ('predict', 462, 24) ('fluid', 1325, 10) ### ### ('anyway', 40, 51) ('disease', 48, 50) ('dying', 594, 27) ('patients', 196, 41) ### ### ('journal', 1287, 18) ('dynamics', 5939, 3) ### ############################################################################################################ [2023-11-25 03:52:54,342][root][INFO] - Epoch: 1: Step: 15801/19530, loss[v]=0.306729, lr=0.000016, acc@1[1]=958.0/1024=0.935546875, acc@1[2]=981.5/1024=0.95849609375 [2023-11-25 03:54:43,776][root][INFO] - Train batch 15900 [2023-11-25 03:54:43,777][root][INFO] - Avg. loss per last 100 batches: 0.204402 [2023-11-25 03:54:44,501][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.66 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 5.16 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.4/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: on the night of february 26, 2012, in sanford, florida, united states, george zimmerman ### ### fatally shot trayvon martin, a 17 - year - old african - american high school student. ### ### [P_TEXT]: jose baez ( lawyer ) of robyn gardner. aruba's high court released gary giordano on ### ### november 28, 2011. no charges were ever filed against gary giordano. baez and lejuez successfully ### ### defended an appeal filed by aruba's chief prosecutor, taco stein. baez defended giordano on the ### ### american reality prime time court show " you the jury ", which was canceled after two episodes. ### ### baez also defended gerod roth in the show. in 2012, george zimmerman was arrested in the shooting ### ### death of 17 ‐ year ‐ old trayvon martin. after jail calls that were released where george zimmerman ### ### discussed ### ### =============================================== v_q ================================================ ### ### ('zimmerman', 0, 1) ('sanford', 1, 7446) ('##von', 2, 8) ('martin', 3, 24) ('florida', 4, 955) ### ### ('george', 5, 36) ('tray', 6, 10) ('african', 7, 7343) ('2012', 8, 59) ('shooting', 9, 25) ### ### ('shot', 10, 80) ('murder', 11, 107) ('trigger', 12, 12573) ('victims', 13, 185) ### ### ('participants', 14, 541) ('violence', 15, 1240) ('absently', 16, 273) ('sniper', 17, 2641) ### ### ('high', 18, 29) ('lethal', 19, 4974) ### ### =============================================== v_p ================================================ ### ### ('bae', 27137, 0) ('zimmerman', 0, 1) ('stein', 7981, 2) ('##rod', 16694, 3) ('robyn', 22504, 4) ### ### ('jose', 6792, 5) ('roth', 9964, 6) ('prosecutor', 2811, 7) ('##von', 2, 8) ('##ju', 20584, 9) ### ### ('tray', 6, 10) ('##uba', 23343, 11) ('filed', 1573, 12) ('gardner', 17581, 13) ('jury', 9598, 14) ### ### ('gary', 5010, 15) ('##ord', 25230, 16) ('reality', 9125, 17) ('##co', 23846, 18) ### ### ('court', 5908, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('zimmerman', 0, 1) ('##von', 2, 8) ('tray', 6, 10) ('martin', 3, 24) ('george', 5, 36) ### ### ('shooting', 9, 25) ('high', 18, 29) ('2012', 8, 59) ('2011', 61, 42) ('americans', 24, 53) ### ### ('deaths', 130, 39) ('jail', 162, 40) ('arrested', 148, 44) ('shot', 10, 80) ('filed', 1573, 12) ### ### ('prosecutor', 2811, 7) ('florida', 4, 955) ('murder', 11, 107) ('discussed', 197, 52) ### ### ('death', 1034, 37) ### ############################################################################################################ [2023-11-25 03:54:45,004][root][INFO] - Epoch: 1: Step: 15901/19530, loss[v]=0.296097, lr=0.000016, acc@1[1]=965.0/1024=0.9423828125, acc@1[2]=988.5/1024=0.96533203125 [2023-11-25 03:56:34,790][root][INFO] - Train batch 16000 [2023-11-25 03:56:34,791][root][INFO] - Avg. loss per last 100 batches: 0.212211 [2023-11-25 03:56:35,533][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.78 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 5.46 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 2003, pearsall annexed the unincorporated neighborhood of west pearsall, increasing ### ### the population of the town by about 350 people. ### ### [P_TEXT]: west pearsall, texas west pearsall was a census - designated place ( cdp ) in frio ### ### county, texas, united states. the population was 349 at the 2000 census. west pearsall was annexed ### ### into the city of pearsall in late 2002. according to the united states census bureau, the cdp had a ### ### total area of, all of it land. as of the census of 2000, there were 349 people, 107 households, and ### ### 89 families residing in the cdp. the population density was 4, 553. 6 people per square mile ( 1, ### ### 684. 4 / km2 ). there were ### ### =============================================== v_q ================================================ ### ### ('pear', 0, 0) ('##sal', 1, 1) ('west', 2, 6) ('massachusetts', 3, 719) ('##l', 4, 11) ### ### ('unincorporated', 5, 4479) ('neighborhood', 6, 1663) ('maine', 7, 8041) ('brooklyn', 8, 3391) ### ### ('town', 9, 62) ('##ل', 10, 28) ('towns', 11, 352) ('wisconsin', 12, 5571) ('chicago', 13, 1846) ### ### ('annexed', 14, 4) ('alderman', 15, 15384) ('connecticut', 16, 913) ('2003', 17, 86) ### ### ('illinois', 18, 968) ('vermont', 19, 8787) ### ### =============================================== v_p ================================================ ### ### ('pear', 0, 0) ('##sal', 1, 1) ('##io', 23288, 2) ('fr', 23327, 3) ('annexed', 14, 4) ### ### ('texas', 283, 5) ('west', 2, 6) ('cdp', 1122, 7) ('county', 43, 8) ('counties', 1106, 9) ### ### ('city', 20, 10) ('##l', 4, 11) ('34', 11263, 12) ('2002', 251, 13) ('bureau', 1721, 14) ### ### ('census', 935, 15) ('cities', 181, 16) ('families', 2013, 17) ('population', 390, 18) ### ### ('land', 502, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('pear', 0, 0) ('##sal', 1, 1) ('west', 2, 6) ('##l', 4, 11) ('annexed', 14, 4) ('city', 20, 10) ### ### ('county', 43, 8) ('##ل', 10, 28) ('texas', 283, 5) ('town', 9, 62) ('2002', 251, 13) ### ### ('cities', 181, 16) ('cdp', 1122, 7) ('counties', 1106, 9) ('population', 390, 18) ### ### ('massachusetts', 3, 719) ('land', 502, 19) ('usa', 89, 31) ('households', 563, 20) ### ### ('states', 124, 29) ### ############################################################################################################ [2023-11-25 03:56:36,035][root][INFO] - Epoch: 1: Step: 16001/19530, loss[v]=0.176044, lr=0.000016, acc@1[1]=985.0/1024=0.9619140625, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-25 03:58:25,452][root][INFO] - Train batch 16100 [2023-11-25 03:58:25,453][root][INFO] - Avg. loss per last 100 batches: 0.242761 [2023-11-25 03:58:26,167][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.64 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 5.25 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: ten years later, west end games acquired the rights to publish their own version, " the ### ### world of indiana jones ". ### ### [P_TEXT]: the world of indiana jones the world of indiana jones is a role - playing game published ### ### by west end games, based on the indiana jones franchise. west end games acquired the rights to ### ### publish an indiana jones role - playing game ten years after the publication of tsr's " the ### ### adventures of indiana jones role - playing game " ( 1984 ). a role - playing game based on the ### ### indiana jones movies was produced by west end games ( weg ) in 1994, called " the world of indiana ### ### jones ". it was designed by brian sean perry and was offered in a boxed set, and several adventure ### ### =============================================== v_q ================================================ ### ### ('indiana', 0, 3) ('end', 1, 7) ('jones', 2, 0) ('game', 3, 9) ('world', 4, 13) ('games', 5, 6) ### ### ('west', 6, 16) ('version', 7, 811) ('gamer', 8, 176) ('content', 9, 401) ('nintendo', 10, 466) ### ### ('chess', 11, 108) ('edition', 12, 1796) ('gaming', 13, 24) ('versions', 14, 1105) ### ### ('manga', 15, 403) ('freaking', 16, 678) ('publisher', 17, 53) ('publishing', 18, 23) ### ### ('ends', 19, 64) ### ### =============================================== v_p ================================================ ### ### ('jones', 2, 0) ('perry', 4246, 1) ('sean', 7753, 2) ('indiana', 0, 3) ('adventures', 328, 4) ### ### ('ts', 794, 5) ('games', 5, 6) ('end', 1, 7) ('we', 11399, 8) ('game', 3, 9) ('boxed', 9176, 10) ### ### ('movies', 1223, 11) ('brian', 10034, 12) ('world', 4, 13) ('movie', 2622, 14) ### ### ('franchise', 98, 15) ('west', 6, 16) ('##g', 17826, 17) ('gameplay', 39, 18) ### ### ('designed', 2940, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('indiana', 0, 3) ('jones', 2, 0) ('end', 1, 7) ('game', 3, 9) ('games', 5, 6) ('world', 4, 13) ### ### ('west', 6, 16) ('gaming', 13, 24) ('publishing', 18, 23) ('gameplay', 39, 18) ### ### ('adventures', 328, 4) ('franchise', 98, 15) ('ts', 794, 5) ('publisher', 17, 53) ### ### ('published', 34, 38) ('movies', 1223, 11) ('ends', 19, 64) ('chess', 11, 108) ### ### ('adventure', 429, 21) ('perry', 4246, 1) ### ############################################################################################################ [2023-11-25 03:58:26,668][root][INFO] - Epoch: 1: Step: 16101/19530, loss[v]=0.115763, lr=0.000016, acc@1[1]=990.5/1024=0.96728515625, acc@1[2]=1007.0/1024=0.9833984375 [2023-11-25 04:00:15,956][root][INFO] - Train batch 16200 [2023-11-25 04:00:15,957][root][INFO] - Avg. loss per last 100 batches: 0.222544 [2023-11-25 04:00:16,677][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.54 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 5.12 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: there were three previous ryder cup captains : raymond floyd, tom lehman and davis love ### ### iii together with rickie fowler, jim furyk, phil mickelson, steve stricker and tiger woods. ### ### [P_TEXT]: davis love iii davis milton love iii ( born april 13, 1964 ) is an american professional ### ### golfer who has won 21 events on the pga tour, including one major championship : the 1997 pga ### ### championship. he won the players championship in 1992 and 2003. he was in the top 10 of the ### ### official world golf ranking for over 450 weeks, reaching a high ranking of 2nd. he captained the u. ### ### s. ryder cup teams in 2012 and 2016. love was inducted ### ### =============================================== v_q ================================================ ### ### ('lehman', 0, 11869) ('floyd', 1, 8782) ('fowler', 2, 8153) ('##rick', 3, 22799) ('fury', 4, 1773) ### ### ('raymond', 5, 9714) ('rick', 6, 3351) ('mick', 7, 17346) ('iii', 8, 5) ('##els', 9, 28283) ### ### ('.', 10, 443) ('tom', 11, 16426) ('##ie', 12, 27189) ('davis', 13, 1) ('golf', 14, 12) ### ### ('jim', 15, 4384) ('tiger', 16, 12908) ('ryder', 17, 0) ('love', 18, 2) ('absently', 19, 358) ### ### =============================================== v_p ================================================ ### ### ('ryder', 17, 0) ('davis', 13, 1) ('love', 18, 2) ('pga', 46, 3) ('milton', 16001, 4) ('iii', 8, 5) ### ### ('tour', 43, 6) ('players', 30, 7) ('1997', 1343, 8) ('rankings', 473, 9) ('1992', 2839, 10) ### ### ('2012', 790, 11) ('golf', 14, 12) ('ranking', 79, 13) ('2003', 864, 14) ('championship', 44, 15) ### ### ('u', 19837, 16) ('official', 2750, 17) ('americans', 3192, 18) ('cup', 29, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('davis', 13, 1) ('iii', 8, 5) ('ryder', 17, 0) ('love', 18, 2) ('golf', 14, 12) ('pga', 46, 3) ### ### ('players', 30, 7) ('tour', 43, 6) ('golfer', 24, 21) ('cup', 29, 19) ('championship', 44, 15) ### ### ('ranking', 79, 13) ('tournament', 22, 34) ('rankings', 473, 9) ('player', 31, 36) ### ### ('anyway', 47, 33) ('captained', 179, 22) ('event', 62, 30) ('2012', 790, 11) ('1997', 1343, 8) ### ############################################################################################################ [2023-11-25 04:00:17,180][root][INFO] - Epoch: 1: Step: 16201/19530, loss[v]=0.124173, lr=0.000017, acc@1[1]=990.5/1024=0.96728515625, acc@1[2]=1010.5/1024=0.98681640625 [2023-11-25 04:02:06,793][root][INFO] - Train batch 16300 [2023-11-25 04:02:06,794][root][INFO] - Avg. loss per last 100 batches: 0.206088 [2023-11-25 04:02:07,520][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.62 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 5.48 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: on 23 june 1904, a breakout attempt by the russian squadron, now under the command of ### ### admiral wilgelm vitgeft, failed. ### ### [P_TEXT]: russian cruiser diana diana ( ) was the second of three protected cruiser s built for the ### ### imperial russian navy. the cruiser served during the russo - japanese war and took part in the ### ### battle of the yellow sea on 10 august 1904. later, she served as part of the russian baltic fleet ### ### during world war i. " diana " was the second of the three vessels in the " pallada " class, built ### ### at the admiralty shipyard in st petersburg, russia. the class was intended to reinforce the russian ### ### presence in the ### ### =============================================== v_q ================================================ ### ### ('wil', 0, 17887) ('1904', 1, 13) ('##gel', 2, 22471) ('russian', 3, 6) ('vi', 4, 16020) ### ### ('squadron', 5, 59) ('##ft', 6, 23901) ('admiral', 7, 945) ('##ge', 8, 26457) ('russia', 9, 20) ### ### ('attempt', 10, 9927) ('.', 11, 224) ('evacuation', 12, 8437) ('1905', 13, 63) ### ### ('absently', 14, 282) ('##m', 15, 25366) ('war', 16, 23) ('naval', 17, 87) ('breakout', 18, 25190) ### ### ('battle', 19, 10) ### ### =============================================== v_p ================================================ ### ### ('diana', 6236, 0) ('admiralty', 107, 1) ('cruiser', 757, 2) ('##lad', 21988, 3) ### ### ('protected', 1659, 4) ('baltic', 638, 5) ('russian', 3, 6) ('yellow', 7492, 7) ### ### ('shipyard', 1965, 8) ('cruisers', 1845, 9) ('battle', 19, 10) ('fleet', 27, 11) ('pal', 22457, 12) ### ### ('1904', 1, 13) ('presence', 10286, 14) ('japanese', 964, 15) ('russo', 269, 16) ### ### ('class', 1182, 17) ('petersburg', 78, 18) ('vessels', 1548, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('russian', 3, 6) ('1904', 1, 13) ('russia', 9, 20) ('battle', 19, 10) ('fleet', 27, 11) ### ### ('admiralty', 107, 1) ('war', 16, 23) ('sea', 25, 21) ('navy', 21, 25) ('squadron', 5, 59) ### ### ('petersburg', 78, 18) ('1903', 28, 29) ('cruiser', 757, 2) ('baltic', 638, 5) ('russo', 269, 16) ### ### ('ship', 30, 41) ('protected', 1659, 4) ('1905', 13, 63) ('1902', 22, 47) ('1900', 33, 46) ### ############################################################################################################ [2023-11-25 04:02:08,029][root][INFO] - Epoch: 1: Step: 16301/19530, loss[v]=0.336640, lr=0.000017, acc@1[1]=965.0/1024=0.9423828125, acc@1[2]=983.5/1024=0.96044921875 [2023-11-25 04:03:57,551][root][INFO] - Train batch 16400 [2023-11-25 04:03:57,552][root][INFO] - Avg. loss per last 100 batches: 0.220439 [2023-11-25 04:03:58,271][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.44 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 5.11 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in february 2013, sfx acquired beatport, an online music store specializing in electronic ### ### music. ### ### [P_TEXT]: beatport their chart performance. in december 2016, sfx emerged from bankruptcy under new ### ### leadership, and renamed livestyle. the company's ceo, randy phillips ( formerly of aeg live ) ### ### stated that beatport had returned to profitability. in october 2017, robb mcdaniels was announced ### ### as ceo of beatport. in may 2019, beatport announced two new subscription services tailored towards ### ### professional djs : beatport link is a streaming service allowing music from the platform's library ### ### to be streamed directly into supported dj software. pioneer dj served as a launch partner, offering ### ### integration through its newly - developed mobile app wedj, and announcing that rekordbox ### ### =============================================== v_q ================================================ ### ### ('sf', 0, 13) ('##port', 1, 2) ('beat', 2, 8) ('##x', 3, 30) ('electronic', 4, 1713) ### ### ('music', 5, 42) ('store', 6, 2804) ('online', 7, 594) ('shopping', 8, 10870) ('2013', 9, 3118) ### ### ('stores', 10, 1678) ('website', 11, 2313) ('brand', 12, 3127) ('web', 13, 856) ### ### ('absently', 14, 253) ('musician', 15, 143) ('users', 16, 2974) ('likes', 17, 4860) ### ### ('album', 18, 276) ('ceo', 19, 23) ### ### =============================================== v_p ================================================ ### ### ('wed', 9714, 0) ('##tyle', 1134, 1) ('##port', 1, 2) ('phillips', 8732, 3) ('dj', 240, 4) ### ### ('randy', 846, 5) ('ae', 16667, 6) ('djs', 105, 7) ('beat', 2, 8) ('##db', 26172, 9) ### ### ('mobile', 99, 10) ('robb', 23910, 11) ('streaming', 678, 12) ('sf', 0, 13) ('pioneer', 3933, 14) ### ### ('bankruptcy', 131, 15) ('##iel', 25274, 16) ('software', 150, 17) ('subscription', 595, 18) ### ### ('launch', 25, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##port', 1, 2) ('sf', 0, 13) ('beat', 2, 8) ('##x', 3, 30) ('music', 5, 42) ('launch', 25, 19) ### ### ('ceo', 19, 23) ('djs', 105, 7) ('mobile', 99, 10) ('dj', 240, 4) ('bankruptcy', 131, 15) ### ### ('software', 150, 17) ('##tyle', 1134, 1) ('randy', 846, 5) ('streaming', 678, 12) ### ### ('subscription', 595, 18) ('live', 891, 20) ('library', 578, 27) ('2017', 170, 38) ### ### ('profit', 651, 28) ### ############################################################################################################ [2023-11-25 04:03:58,773][root][INFO] - Epoch: 1: Step: 16401/19530, loss[v]=0.126658, lr=0.000017, acc@1[1]=988.5/1024=0.96533203125, acc@1[2]=1009.0/1024=0.9853515625 [2023-11-25 04:05:48,338][root][INFO] - Train batch 16500 [2023-11-25 04:05:48,339][root][INFO] - Avg. loss per last 100 batches: 0.192108 [2023-11-25 04:05:49,067][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.60 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 5.17 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.3/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it occurs within hydrothermal cobalt - bearing mineral deposits as a rare phase ### ### associated with roselite, erythrite, annabergite and cobalt rich calcite and dolomite. ### ### [P_TEXT]: cobalt cobalt in compound form occurs in copper and nickel minerals. it is the major ### ### metallic component that combines with sulfur and arsenic in the sulfidic cobaltite ( coass ), ### ### safflorite ( coas2 ), glaucodot ( ), and skutterudite ( coas3 ) minerals. the mineral cattierite is ### ### similar to pyrite and occurs together with vaesite in the copper deposits of katanga province. when ### ### it reaches the atmosphere, weathering occurs ; the sulfide minerals oxidize and form pink erythrite ### ### ( " cobalt glance " : co3 ( aso4 ) 2 · 8h2o ) and spherocobaltite ( coco3 ). cobalt is also a ### ### constituent of tobacco smoke. the tobacco plant readily absorbs and ### ### =============================================== v_q ================================================ ### ### ('cobalt', 0, 0) ('rose', 1, 1744) ('##lite', 2, 1428) ('anna', 3, 8405) ('.', 4, 77) ### ### ('##berg', 5, 18065) ('hydro', 6, 12114) ('##lom', 7, 22449) ('##ite', 8, 21) ('er', 9, 51) ### ### ('##hri', 10, 19) ('do', 11, 15327) ('absently', 12, 253) ('##thermal', 13, 2891) ('##yt', 14, 61) ### ### ('deposits', 15, 41) ('mineral', 16, 30) ('rare', 17, 2030) ('geologist', 18, 2773) ### ### ('geological', 19, 80) ### ### =============================================== v_p ================================================ ### ### ('cobalt', 0, 0) ('##tier', 22689, 1) ('##utter', 13470, 2) ('arsenic', 2374, 3) ('coa', 23872, 4) ### ### ('coco', 13707, 5) ('##lau', 28591, 6) ('va', 20058, 7) ('smoke', 1691, 8) ('##nga', 14413, 9) ### ### ('##dot', 24170, 10) ('##udi', 28825, 11) ('tobacco', 4756, 12) ('##tite', 52, 13) ### ### ('kata', 23560, 14) ('pink', 2504, 15) ('##dic', 12032, 16) ('cat', 12377, 17) ('##fide', 2586, 18) ### ### ('##hri', 10, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('cobalt', 0, 0) ('##ite', 8, 21) ('##hri', 10, 19) ('mineral', 16, 30) ('er', 9, 51) ### ### ('deposits', 15, 41) ('##tite', 52, 13) ('##yt', 14, 61) ('.', 4, 77) ('nickel', 164, 22) ### ### ('sulfur', 201, 26) ('copper', 214, 27) ('province', 273, 31) ('occurs', 69, 55) ('anyway', 24, 65) ### ### ('minerals', 88, 54) ('##₂', 84, 56) ('geological', 19, 80) ('plant', 529, 24) ('arsenic', 2374, 3) ### ############################################################################################################ [2023-11-25 04:05:49,577][root][INFO] - Epoch: 1: Step: 16501/19530, loss[v]=0.331157, lr=0.000017, acc@1[1]=957.0/1024=0.9345703125, acc@1[2]=987.5/1024=0.96435546875 [2023-11-25 04:07:39,714][root][INFO] - Train batch 16600 [2023-11-25 04:07:39,715][root][INFO] - Avg. loss per last 100 batches: 0.195468 [2023-11-25 04:07:40,436][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.80 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 5.20 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.3/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the prevalence of phantom pain after limb amputation ranged from 50 % to 78 %. ### ### [P_TEXT]: amputation burn, feel tense, dry or wet, locked in or trapped or they can feel as if they ### ### are moving. some scientists believe it has to do with a kind of neural map that the brain has of ### ### the body, which sends information to the rest of the brain about limbs regardless of their ### ### existence. phantom sensations and phantom pain may also occur after the removal of body parts other ### ### than the limbs, e. g. after amputation of the breast, extraction of a tooth ( phantom tooth pain ) ### ### or removal of an eye ( phantom eye syndrome ). a similar phenomenon ### ### =============================================== v_q ================================================ ### ### ('phantom', 0, 1) ('limb', 1, 54) ('amp', 2, 3) ('pain', 3, 11) ('##utation', 4, 0) ('.', 5, 78) ### ### ('diagnosed', 6, 16123) ('absently', 7, 302) ('78', 8, 18805) ('prevalence', 9, 253) ### ### ('limbs', 10, 10) ('disability', 11, 299) ('paralysis', 12, 1817) ('50', 13, 7450) ### ### ('mumbled', 14, 74) ('%', 15, 7996) ('congenital', 16, 1652) ('intently', 17, 512) ### ### ('anyway', 18, 50) ('worried', 19, 44) ### ### =============================================== v_p ================================================ ### ### ('##utation', 4, 0) ('phantom', 0, 1) ('trapped', 5853, 2) ('amp', 2, 3) ('wet', 1426, 4) ### ### ('tooth', 596, 5) ('extraction', 1073, 6) ('locked', 11120, 7) ('map', 6937, 8) ('breast', 6234, 9) ### ### ('limbs', 10, 10) ('pain', 3, 11) ('eye', 3446, 12) ('removal', 7998, 13) ('syndrome', 162, 14) ### ### ('tense', 1916, 15) ('neural', 1992, 16) ('burn', 3014, 17) ('dry', 17163, 18) ('brain', 71, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##utation', 4, 0) ('phantom', 0, 1) ('amp', 2, 3) ('pain', 3, 11) ('limbs', 10, 10) ### ### ('limb', 1, 54) ('brain', 71, 19) ('body', 32, 20) ('syndrome', 162, 14) ('.', 5, 78) ### ### ('tooth', 596, 5) ('occur', 26, 36) ('extraction', 1073, 6) ('wet', 1426, 4) ('worried', 19, 44) ### ### ('phenomenon', 450, 21) ('anyway', 18, 50) ('parts', 780, 22) ('sensation', 584, 27) ### ### ('tense', 1916, 15) ### ############################################################################################################ [2023-11-25 04:07:40,937][root][INFO] - Epoch: 1: Step: 16601/19530, loss[v]=0.146808, lr=0.000017, acc@1[1]=989.5/1024=0.96630859375, acc@1[2]=1009.5/1024=0.98583984375 [2023-11-25 04:09:30,560][root][INFO] - Train batch 16700 [2023-11-25 04:09:30,561][root][INFO] - Avg. loss per last 100 batches: 0.198327 [2023-11-25 04:09:31,279][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.61 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 5.29 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: highways serving hamilton are us highway 127, state route 128, state route 129, state ### ### route 130, and state route 4. ### ### [P_TEXT]: twin oaks ( wyoming, ohio ) street ) rises to a gable and is pierced by four windows, ### ### while the left bay stops at the normal roofline and features the main entrance. a side entrance ### ### with portico is placed farther to the rear, underneath another subsidiary gable. good ### ### transportation is a leading reason for wyoming's prosperity. the city lies near the old pre - ### ### statehood road that connected cincinnati with locations farther north, such as fort hamilton and ### ### fallen timbers. curves in the road were cut off in 1806, forming a new road that is today followed ### ### by springfield pike through central wyoming. improvements in the 1830s only enhanced its ### ### importance. by this time, another mode of transportation ### ### =============================================== v_q ================================================ ### ### ('hamilton', 0, 3) ('129', 1, 2773) ('128', 2, 7272) ('127', 3, 7755) ('130', 4, 11488) ### ### ('highway', 5, 61) ('4', 6, 15901) ('route', 7, 82) ('us', 8, 4605) ('.', 9, 204) ### ### ('maine', 10, 7552) ('states', 11, 7679) ('state', 12, 2279) ('arkansas', 13, 1687) ### ### ('carolina', 14, 14895) ('oregon', 15, 1293) ('massachusetts', 16, 1626) ('idaho', 17, 439) ### ### ('dakota', 18, 3293) ('connecticut', 19, 1316) ### ### =============================================== v_p ================================================ ### ### ('oaks', 15420, 0) ('gable', 4757, 1) ('wyoming', 230, 2) ('hamilton', 0, 3) ### ### ('springfield', 6031, 4) ('fallen', 10452, 5) ('twin', 12327, 6) ('timbers', 11279, 7) ### ### ('pike', 96, 8) ('cincinnati', 58, 9) ('portico', 12797, 10) ('fort', 604, 11) ### ### ('statehood', 339, 12) ('transportation', 285, 13) ('curves', 20776, 14) ('prosperity', 4573, 15) ### ### ('pierced', 624, 16) ('farther', 488, 17) ('ohio', 40, 18) ('normal', 1340, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('hamilton', 0, 3) ('cincinnati', 58, 9) ('highway', 5, 61) ('ohio', 40, 18) ('pike', 96, 8) ### ### ('wyoming', 230, 2) ('route', 7, 82) ('transportation', 285, 13) ('statehood', 339, 12) ### ### ('fort', 604, 11) ('farther', 488, 17) ('road', 168, 29) ('pierced', 624, 16) ('.', 9, 204) ### ### ('colorado', 21, 74) ('roadway', 118, 38) ('gable', 4757, 1) ('location', 106, 42) ### ### ('normal', 1340, 19) ('roads', 204, 36) ### ############################################################################################################ [2023-11-25 04:09:31,784][root][INFO] - Epoch: 1: Step: 16701/19530, loss[v]=0.238350, lr=0.000017, acc@1[1]=962.5/1024=0.93994140625, acc@1[2]=991.5/1024=0.96826171875 [2023-11-25 04:11:21,487][root][INFO] - Train batch 16800 [2023-11-25 04:11:21,488][root][INFO] - Avg. loss per last 100 batches: 0.215043 [2023-11-25 04:11:22,217][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.64 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 5.25 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the work includes evidence of browne's adherence to the baconian method of empirical ### ### observation of nature, and was in the vanguard of work - in - progress scientific journalism during ### ### the 17th - century scientific revolution. ### ### [P_TEXT]: francis bacon divided into three great branches : bacon's seminal work " novum organum " ### ### was influential in the 1630s and 1650s among scholars, in particular sir thomas browne, who in his ### ### encyclopedia " pseudodoxia epidemica " ( 1646 – 72 ) frequently adheres to a baconian approach to ### ### his scientific enquiries. this book entails the basis of the scientific method as a means of ### ### observation and induction. according to francis bacon, learning and knowledge all derive from the ### ### basis of inductive reasoning. through his belief of experimental encounters, he theorized that all ### ### the knowledge that was necessary to fully ### ### =============================================== v_q ================================================ ### ### ('browne', 0, 2) ('scientific', 1, 13) ('bacon', 2, 0) ('journalism', 3, 1785) ### ### ('revolution', 4, 5805) ('empirical', 5, 80) ('progress', 6, 7481) ('.', 7, 327) ('##ian', 8, 31) ### ### ('observation', 9, 11) ('sparhawk', 10, 912) ('theory', 11, 628) ('journalist', 12, 12717) ### ### ('treatise', 13, 253) ('nature', 14, 2564) ('science', 15, 302) ('century', 16, 11261) ### ### ('method', 17, 20) ('work', 18, 69) ('journalists', 19, 9632) ### ### =============================================== v_p ================================================ ### ### ('bacon', 2, 0) ('induction', 5648, 1) ('browne', 0, 2) ('epidemic', 3853, 3) ### ### ('encyclopedia', 56, 4) ('nov', 17206, 5) ('organ', 934, 6) ('pseudo', 2732, 7) ### ### ('francis', 1527, 8) ('experimental', 180, 9) ('learning', 2084, 10) ('observation', 9, 11) ### ### ('thomas', 575, 12) ('scientific', 1, 13) ('1630', 4160, 14) ('knowledge', 321, 15) ### ### ('seminal', 8604, 16) ('##tive', 7135, 17) ('encounter', 2037, 18) ('branches', 11802, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('bacon', 2, 0) ('browne', 0, 2) ('scientific', 1, 13) ('observation', 9, 11) ('##ian', 8, 31) ### ### ('method', 17, 20) ('encyclopedia', 56, 4) ('experimental', 180, 9) ('scholars', 49, 28) ### ### ('empirical', 5, 80) ('knowledge', 321, 15) ('organ', 934, 6) ('methods', 76, 38) ### ### ('derive', 303, 25) ('thomas', 575, 12) ('observations', 21, 51) ('francis', 1527, 8) ### ### ('work', 18, 69) ('anyway', 64, 46) ('epidemic', 3853, 3) ### ############################################################################################################ [2023-11-25 04:11:22,720][root][INFO] - Epoch: 1: Step: 16801/19530, loss[v]=0.136570, lr=0.000017, acc@1[1]=990.5/1024=0.96728515625, acc@1[2]=1012.0/1024=0.98828125 [2023-11-25 04:13:12,693][root][INFO] - Train batch 16900 [2023-11-25 04:13:12,693][root][INFO] - Avg. loss per last 100 batches: 0.221875 [2023-11-25 04:13:13,419][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.53 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 5.26 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.6/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: " all we need is love " was written by ricki - lee coulter, stuart crichton and ilan ### ### kidron. ### ### [P_TEXT]: ricki - lee coulter hit on the " billboard " japan hot 100 chart and earned coulter her ### ### first aria music award nomination for song of the year. her fourth album " dance in the rain " ( ### ### 2014 ) became her second top - fifteen album, and featured the top - forty single " all we need is ### ### love ". coulter has branched out from recording music, pursuing careers in television and radio. ### ### she returned to " australian idol " for seasons six and seven, as a presenter and reporter ### ### alongside james mathison and andrew gunsberg. the role earned her ### ### =============================================== v_q ================================================ ### ### ('rick', 0, 1) ('coulter', 1, 0) ('##ron', 2, 23728) ('##ich', 3, 24711) ('kid', 4, 10766) ### ### ('il', 5, 27127) ('need', 6, 10) ('cr', 7, 22318) ('stuart', 8, 6238) ('lee', 9, 17) ('we', 10, 19) ### ### ('love', 11, 22) ('absently', 12, 315) ('##ton', 13, 18227) ('song', 14, 34) ('##an', 15, 25618) ### ### ('"', 16, 115) ('」', 17, 59) ('rihanna', 18, 13971) ('##i', 19, 40) ### ### =============================================== v_p ================================================ ### ### ('coulter', 1, 0) ('rick', 0, 1) ('idol', 21, 2) ('##ison', 11794, 3) ('math', 15786, 4) ### ### ('seasons', 10435, 5) ('dance', 171, 6) ('rain', 151, 7) ('aria', 240, 8) ('andrew', 7558, 9) ### ### ('need', 6, 10) ('forty', 15430, 11) ('six', 9338, 12) ('guns', 5327, 13) ('##berg', 16441, 14) ### ### ('japan', 4695, 15) ('australian', 281, 16) ('lee', 9, 17) ('2014', 118, 18) ('we', 10, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('coulter', 1, 0) ('rick', 0, 1) ('need', 6, 10) ('lee', 9, 17) ('we', 10, 19) ('love', 11, 22) ### ### ('idol', 21, 2) ('song', 14, 34) ('album', 24, 26) ('rain', 151, 7) ('dance', 171, 6) ### ### ('single', 31, 25) ('##i', 19, 40) ('aria', 240, 8) ('」', 17, 59) ('″', 22, 45) ('2014', 118, 18) ### ### ('singles', 32, 33) ('awards', 73, 23) ('australian', 281, 16) ### ############################################################################################################ [2023-11-25 04:13:13,921][root][INFO] - Epoch: 1: Step: 16901/19530, loss[v]=0.141328, lr=0.000017, acc@1[1]=986.5/1024=0.96337890625, acc@1[2]=1006.5/1024=0.98291015625 [2023-11-25 04:15:03,465][root][INFO] - Train batch 17000 [2023-11-25 04:15:03,466][root][INFO] - Avg. loss per last 100 batches: 0.207464 [2023-11-25 04:15:04,181][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.52 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 5.18 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: rare surviving examples of beating the bounds are at the nearby city church of all ### ### hallows - by - the - tower ( every three years ), and at the queen's chapel of the savoy, ### ### westminster. ### ### [P_TEXT]: savoy chapel st mary - at - hill ). the new congregation's first pastor, irenaeus crusius ### ### ( previously an associate at holy trinity ), dedicated the chapel on the 19th sunday after trinity ### ### 1694 as the " marienkirche " or the german church of st mary - le - savoy. archibald cameron of ### ### lochiel, the last jacobite leader to be executed for treason, was buried there in 1753. an anglican ### ### place of worship, the chapel was noted in the 18th and 19th centuries as a place where marriages ### ### without banns might occur outside of the usual parameters of ### ### =============================================== v_q ================================================ ### ### ('bounds', 0, 14156) ('##ows', 1, 25266) ('savoy', 2, 0) ('queen', 3, 12051) ('tower', 4, 3172) ### ### ('beating', 5, 18641) ('hall', 6, 11419) ('vimes', 7, 194) ('church', 8, 25) ### ### ('westminster', 9, 6728) ('chapel', 10, 3) ('tolkien', 11, 26915) ('city', 12, 15774) ### ### ('absently', 13, 543) ('sparhawk', 14, 196) ('entertained', 15, 7415) ('commemorated', 16, 3329) ### ### ('all', 17, 19880) ('covent', 18, 16121) ('##isation', 19, 12986) ### ### =============================================== v_p ================================================ ### ### ('savoy', 2, 0) ('treason', 2614, 1) ('jacobite', 1687, 2) ('chapel', 10, 3) ('archibald', 9786, 4) ### ### ('loch', 6072, 5) ('169', 5512, 6) ('cameron', 5362, 7) ('##iel', 14375, 8) ('trinity', 2632, 9) ### ### ('german', 5797, 10) ('sunday', 1858, 11) ('ban', 15662, 12) ('##nk', 27765, 13) ('mary', 1057, 14) ### ### ('congregation', 1914, 15) ('associate', 14593, 16) ('pastor', 5712, 17) ('worship', 565, 18) ### ### ('19th', 1563, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('savoy', 2, 0) ('chapel', 10, 3) ('church', 8, 25) ('jacobite', 1687, 2) ('centuries', 169, 35) ### ### ('worship', 565, 18) ('treason', 2614, 1) ('medieval', 40, 50) ('holy', 577, 29) ('mary', 1057, 14) ### ### ('vimes', 7, 194) ('trinity', 2632, 9) ('sunday', 1858, 11) ('19th', 1563, 19) ### ### ('congregation', 1914, 15) ('anglican', 1246, 32) ('169', 5512, 6) ('sparhawk', 14, 196) ### ### ('cameron', 5362, 7) ('anyway', 110, 55) ### ############################################################################################################ [2023-11-25 04:15:04,684][root][INFO] - Epoch: 1: Step: 17001/19530, loss[v]=0.134110, lr=0.000017, acc@1[1]=982.0/1024=0.958984375, acc@1[2]=1007.0/1024=0.9833984375 [2023-11-25 04:16:54,849][root][INFO] - Train batch 17100 [2023-11-25 04:16:54,849][root][INFO] - Avg. loss per last 100 batches: 0.198966 [2023-11-25 04:16:55,586][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.71 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 5.55 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: styles defeated roh world tag team champions christopher daniels and frankie kazarian. ### ### [P_TEXT]: frankie kazarian part of evolution he and nova were defeated by edge and christian. he ### ### used to freestyle rap along with his close friends john cena and samoa joe while the three of them ### ### were on the road in upw. he toured the uk, australia and new zealand with the now defunct world ### ### wrestling all - stars promotion in may 2003, where he met glenn gilberti and scott d'amore, leading ### ### to his being offered a job by total nonstop action wrestling. kazarian also worked for pro ### ### wrestling guerrilla, and won a tournament to become the first ever pwg champion on august ### ### =============================================== v_q ================================================ ### ### ('styles', 0, 434) ('frankie', 1, 0) ('christopher', 2, 11608) ('daniels', 3, 7614) ('.', 4, 120) ### ### ('absently', 5, 227) ('ro', 6, 21528) ('##zar', 7, 7) ('##h', 8, 21751) ('match', 9, 52) ### ### ('style', 10, 2777) ('tag', 11, 2250) ('wrestling', 12, 12) ('wwe', 13, 135) ('intently', 14, 164) ### ### ('world', 15, 49) ('promotions', 16, 149) ('tagged', 17, 104) ('！', 18, 122) ('team', 19, 342) ### ### =============================================== v_p ================================================ ### ### ('frankie', 1, 0) ('amore', 690, 1) ('gilbert', 1068, 2) ('guerrilla', 5026, 3) ('nova', 11719, 4) ### ### ('evolution', 2298, 5) ('##wg', 15837, 6) ('##zar', 7, 7) ('glenn', 16593, 8) ### ### ('christian', 8122, 9) ('edge', 3083, 10) ('samoa', 2080, 11) ('wrestling', 12, 12) ### ### ('freestyle', 1379, 13) ('stars', 2539, 14) ('rap', 5807, 15) ('zealand', 6579, 16) ### ### ('##ian', 44, 17) ('ce', 12206, 18) ('scott', 3217, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('frankie', 1, 0) ('##zar', 7, 7) ('wrestling', 12, 12) ('##ian', 44, 17) ('match', 9, 52) ### ### ('ka', 38, 30) ('promotion', 27, 32) ('amore', 690, 1) ('world', 15, 49) ('.', 4, 120) ### ### ('gilbert', 1068, 2) ('styles', 0, 434) ('anyway', 24, 51) ('wrestlers', 37, 45) ### ### ('defeated', 55, 39) ('absently', 5, 227) ('freestyle', 1379, 13) ('champion', 76, 43) ### ### ('2003', 797, 20) ('evolution', 2298, 5) ### ############################################################################################################ [2023-11-25 04:16:56,091][root][INFO] - Epoch: 1: Step: 17101/19530, loss[v]=0.114229, lr=0.000018, acc@1[1]=997.5/1024=0.97412109375, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 04:18:45,962][root][INFO] - Train batch 17200 [2023-11-25 04:18:45,963][root][INFO] - Avg. loss per last 100 batches: 0.214990 [2023-11-25 04:18:46,679][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.69 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 5.39 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: skye, islay and mull are the three largest, and also have the highest populations. ### ### [P_TEXT]: islay islay ( ;, ) is the southernmost island of the inner hebrides of scotland. known as ### ### " the queen of the hebrides ", it lies in argyll just south west of jura and around north of the ### ### northern irish coast. the island's capital is bowmore where the distinctive round kilarrow parish ### ### church and a distillery are located. port ellen is the main port. islay is the fifth - largest ### ### scottish island and the eighth - largest island of the british isles, with a total area of almost ### ### =============================================== v_q ================================================ ### ### ('isla', 0, 0) ('mu', 1, 16490) ('##ll', 2, 25054) ('skye', 3, 9053) ('.', 4, 445) ### ### ('scotland', 5, 25) ('absently', 6, 180) ('island', 7, 16) ('##y', 8, 17) ('isle', 9, 879) ### ### ('##ɨ', 10, 44) ('shetland', 11, 13404) ('scottish', 12, 9) ('##kney', 13, 8860) ### ### ('islands', 14, 34) ('intently', 15, 108) ('isles', 16, 13) ('ireland', 17, 29) ### ### ('estuary', 18, 3908) ('argyll', 19, 4) ### ### =============================================== v_p ================================================ ### ### ('isla', 0, 0) ('ellen', 14818, 1) ('ju', 24642, 2) ('bow', 10012, 3) ('argyll', 19, 4) ### ### ('##bri', 11302, 5) ('distillery', 7549, 6) ('##row', 20990, 7) ('##more', 2938, 8) ### ### ('scottish', 12, 9) ('irish', 33, 10) ('##lar', 24895, 11) ('##des', 15870, 12) ('isles', 16, 13) ### ### ('##ra', 17973, 14) ('queen', 1848, 15) ('island', 7, 16) ('##y', 8, 17) ('inner', 10390, 18) ### ### ('eighth', 18964, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('isla', 0, 0) ('island', 7, 16) ('##y', 8, 17) ('scottish', 12, 9) ('argyll', 19, 4) ### ### ('scotland', 5, 25) ('isles', 16, 13) ('irish', 33, 10) ('ireland', 17, 29) ('largest', 32, 24) ### ### ('islands', 14, 34) ('##ɨ', 10, 44) ('parish', 37, 27) ('port', 216, 21) ('worried', 28, 36) ### ### ('northern', 392, 23) ('coast', 199, 28) ('.', 4, 445) ('absently', 6, 180) ('british', 118, 35) ### ############################################################################################################ [2023-11-25 04:18:47,181][root][INFO] - Epoch: 1: Step: 17201/19530, loss[v]=0.106659, lr=0.000018, acc@1[1]=996.0/1024=0.97265625, acc@1[2]=1012.5/1024=0.98876953125 [2023-11-25 04:20:37,012][root][INFO] - Train batch 17300 [2023-11-25 04:20:37,013][root][INFO] - Avg. loss per last 100 batches: 0.203351 [2023-11-25 04:20:37,741][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.68 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.5/29522=99.99% | mean: 0.03 | max: 4.97 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the potsdam conference, via substantive recognition of the pro - soviet polish committee ### ### of national liberation, implicitly consented to the deportation of polish people from " kresy " ( ### ### see polish population transfers ( 1944 – 1946 ) ). ### ### [P_TEXT]: walter priesnitz he was among the millions driven west by the ethnic cleansing of 1945 ### ### which had seen the german population of silesia forcibly replaced by ethnic poles, themselves ### ### driven out of the larger ( but less densely populated ) area which before 1939 had been ### ### internationally recognised as the eastern third of poland. during 1950 / 51 priesnitz undertook an ### ### apprenticeship with the zwickau er. instead of following his father into a banking career, ### ### priesnitz now spent the next four years obtaining a university - level education, studying ### ### jurisprudence and at berlin ( free university ), munster and cologne. while still an undergraduate ### ### student, in 1954, walter priesnitz joined ### ### =============================================== v_q ================================================ ### ### ('polish', 0, 28) ('##res', 1, 13309) ('soviet', 2, 907) ('##ʂ', 3, 254) ('potsdam', 4, 12208) ### ### ('liberation', 5, 2491) ('transfers', 6, 21022) ('committee', 7, 5637) ('##ɨ', 8, 4344) ### ### ('poland', 9, 20) ('population', 10, 12) ('deportation', 11, 675) ('implicit', 12, 3508) ### ### ('pro', 13, 11667) ('transfer', 14, 6803) ('k', 15, 22864) ('1946', 16, 91) ('.', 17, 109) ### ### ('people', 18, 858) ('##y', 19, 27817) ### ### =============================================== v_p ================================================ ### ### ('##nitz', 23434, 0) ('silesia', 296, 1) ('walter', 1303, 2) ('munster', 7063, 3) ### ### ('cleansing', 483, 4) ('pri', 6076, 5) ('##wick', 10539, 6) ('er', 12408, 7) ('banking', 24208, 8) ### ### ('berlin', 29, 9) ('ethnic', 116, 10) ('apprenticeship', 16515, 11) ('population', 10, 12) ### ### ('##au', 14814, 13) ('poles', 78, 14) ('university', 3819, 15) ('eastern', 2045, 16) ### ### ('jurisprudence', 19254, 17) ('51', 5944, 18) ('cologne', 3597, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('polish', 0, 28) ('population', 10, 12) ('poland', 9, 20) ('berlin', 29, 9) ('1945', 27, 24) ### ### ('poles', 78, 14) ('silesia', 296, 1) ('ethnic', 116, 10) ('1944', 20, 58) ('cleansing', 483, 4) ### ### ('german', 100, 23) ('walter', 1303, 2) ('hitler', 33, 50) ('populations', 70, 40) ('1946', 16, 91) ### ### ('.', 17, 109) ('##ʂ', 3, 254) ('worried', 64, 51) ('eastern', 2045, 16) ('1954', 498, 37) ### ############################################################################################################ [2023-11-25 04:20:38,245][root][INFO] - Epoch: 1: Step: 17301/19530, loss[v]=0.243608, lr=0.000018, acc@1[1]=966.5/1024=0.94384765625, acc@1[2]=994.0/1024=0.970703125 [2023-11-25 04:22:28,365][root][INFO] - Train batch 17400 [2023-11-25 04:22:28,366][root][INFO] - Avg. loss per last 100 batches: 0.209910 [2023-11-25 04:22:29,104][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.50 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.2/29522=99.99% | mean: 0.03 | max: 4.94 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the standard aspect ratios are 1. 66, 1. 85, and 2. 39 ( anamorphic ). ### ### [P_TEXT]: aspect ratio ( image ) 4 : 3 ( 1. : 1 ) ( generally read as four - three, four - by - ### ### three, or four - to - three ) for standard television has been in use since the invention of moving ### ### picture cameras and many computer monitor s used to employ the same aspect ratio. 4 : 3 was the ### ### aspect ratio used for 35 mm films in the silent era. it is also very close to the 1. 375 : 1 ### ### academy ratio, defined by the academy of motion picture arts and sciences as a standard after the ### ### advent of optical sound - on - film. by having tv match this ### ### =============================================== v_q ================================================ ### ### ('aspect', 0, 1) ('##morphic', 1, 18165) ('ana', 2, 15906) ('.', 3, 90) ('absently', 4, 426) ### ### ('66', 5, 12602) ('ratio', 6, 0) ('intently', 7, 215) ('85', 8, 11407) ('erebidae', 9, 1962) ### ### ('39', 10, 1607) ('standard', 11, 14) ('image', 12, 12) ('manual', 13, 2638) ### ### ('definition', 14, 2203) ('ந', 15, 140) ('##₂', 16, 233) ('aesthetic', 17, 209) ### ### ('cosmetic', 18, 1504) ('used', 19, 69) ### ### =============================================== v_p ================================================ ### ### ('ratio', 6, 0) ('aspect', 0, 1) ('monitor', 2923, 2) ('375', 15970, 3) ('optical', 617, 4) ### ### ('academy', 2267, 5) ('match', 6827, 6) ('silent', 3094, 7) ('picture', 147, 8) ('sound', 1251, 9) ### ### ('computer', 696, 10) ('read', 14983, 11) ('image', 12, 12) ('mm', 877, 13) ('standard', 11, 14) ### ### ('35', 3674, 15) ('sciences', 9038, 16) ('era', 3394, 17) ('moving', 8911, 18) ('camera', 65, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('aspect', 0, 1) ('ratio', 6, 0) ('standard', 11, 14) ('image', 12, 12) ('optical', 617, 4) ### ### ('picture', 147, 8) ('camera', 65, 19) ('.', 3, 90) ('aspects', 79, 28) ('ratios', 95, 31) ### ### ('film', 169, 29) ('computer', 696, 10) ('monitor', 2923, 2) ('mm', 877, 13) ('academy', 2267, 5) ### ### ('anyway', 66, 39) ('sound', 1251, 9) ('standards', 58, 44) ('absently', 4, 426) ### ### ('noticed', 34, 53) ### ############################################################################################################ [2023-11-25 04:22:29,606][root][INFO] - Epoch: 1: Step: 17401/19530, loss[v]=0.103569, lr=0.000018, acc@1[1]=998.0/1024=0.974609375, acc@1[2]=1015.0/1024=0.9912109375 [2023-11-25 04:24:19,568][root][INFO] - Train batch 17500 [2023-11-25 04:24:19,568][root][INFO] - Avg. loss per last 100 batches: 0.210156 [2023-11-25 04:24:20,285][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.65 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 5.29 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in a part of the brewer's large gardens on what is today the site of the new, and ### ### temporary, courtyard theatre. ### ### [P_TEXT]: tim minchin courtyard theatre in stratford - upon - avon, running from 9 november 2010 to ### ### 30 january 2011, and it began its west end run at the cambridge theatre on 25 october 2011 to great ### ### critical acclaim. in 2013, " matilda " opened on broadway at the shubert theatre, and earned 12 ### ### tony award nominations. the musical has subsequently been presented worldwide winning numerous ### ### awards. minchin was cast in the role of judas in the 2012 uk and ireland arena tour of the tim rice ### ### and andrew lloyd webber rock opera " jesus ### ### =============================================== v_q ================================================ ### ### ('courtyard', 0, 10) ('brewer', 1, 21580) ('theatre', 2, 11) ('temporary', 3, 9445) ### ### ('gardens', 4, 5472) ('brewery', 5, 6570) ('london', 6, 508) ('garden', 7, 252) ### ### ('absently', 8, 226) ('.', 9, 765) ('brewers', 10, 22618) ('intently', 11, 157) ('circus', 12, 521) ### ### ('hertfordshire', 13, 9850) ('demolished', 14, 2569) ('vimes', 15, 1575) ('brewing', 16, 19564) ### ### ('backyard', 17, 2277) ('thames', 18, 12212) ('theater', 19, 34) ### ### =============================================== v_p ================================================ ### ### ('matilda', 2118, 0) ('judas', 21407, 1) ('##chin', 22983, 2) ('rice', 7874, 3) ### ### ('cambridge', 128, 4) ('min', 17861, 5) ('jesus', 8426, 6) ('tim', 12631, 7) ('arena', 550, 8) ### ### ('shu', 16552, 9) ('courtyard', 0, 10) ('theatre', 2, 11) ('avon', 8046, 12) ('stratford', 26, 13) ### ### ('##bert', 17008, 14) ('broadway', 367, 15) ('tour', 8790, 16) ('webber', 8763, 17) ### ### ('ireland', 904, 18) ('tony', 2079, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('courtyard', 0, 10) ('theatre', 2, 11) ('stratford', 26, 13) ('cambridge', 128, 4) ### ### ('theatres', 29, 31) ('opera', 127, 20) ('theater', 19, 34) ('arena', 550, 8) ('uk', 162, 25) ### ### ('broadway', 367, 15) ('matilda', 2118, 0) ('oxford', 28, 42) ('″', 666, 21) ('musicals', 687, 23) ### ### ('ireland', 904, 18) ('noticed', 75, 44) ('garden', 7, 252) ('absently', 8, 226) ### ### ('intently', 11, 157) ('musical', 959, 27) ### ############################################################################################################ [2023-11-25 04:24:20,788][root][INFO] - Epoch: 1: Step: 17501/19530, loss[v]=0.269173, lr=0.000018, acc@1[1]=950.0/1024=0.927734375, acc@1[2]=987.5/1024=0.96435546875 [2023-11-25 04:26:10,755][root][INFO] - Train batch 17600 [2023-11-25 04:26:10,755][root][INFO] - Avg. loss per last 100 batches: 0.209712 [2023-11-25 04:26:11,493][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.76 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 5.51 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the terms " tip " and " ring " originated in the early days of telephony when telephone ### ### operators used plugs to connect customer calls. ### ### [P_TEXT]: phone connector ( audio ) three - contact versions are known as " trs connectors ", where ### ### " t " stands for " tip ", " r " stands for " ring " and " s " stands for " sleeve ". ring contacts ### ### are typically the same diameter as the sleeve, the long shank. similarly, two -, four - and five - ### ### contact versions are called " ts ", " trrs " and " trrrs connectors " respectively. the outside ### ### diameter of the " sleeve " conductor is. the " mini " connector has a diameter of and the " sub - ### ### mini " connector has a diameter of. specific models, and connectors used in specific applications, ### ### may be termed e. g. " stereo plug ", " headphone jack ", " microphone jack ", " aux input ", etc. ### ### =============================================== v_q ================================================ ### ### ('tip', 0, 10) ('plug', 1, 12) ('ring', 2, 18) ('tel', 3, 22511) ('##ep', 4, 28529) ### ### ('##hony', 5, 17926) ('customer', 6, 4793) ('.', 7, 759) ('telephone', 8, 85) ('phone', 9, 14) ### ### ('telecommunications', 10, 1553) ('customers', 11, 10964) ('absently', 12, 1147) ### ### ('intently', 13, 896) ('operators', 14, 11504) ('use', 15, 141) ('tips', 16, 47) ('used', 17, 61) ### ### ('operator', 18, 2627) ('tipped', 19, 70) ### ### =============================================== v_p ================================================ ### ### ('connector', 835, 0) ('shan', 8980, 1) ('sleeve', 11739, 2) ('mini', 9349, 3) ### ### ('microphone', 1463, 4) ('tr', 16327, 5) ('aux', 13636, 6) ('audio', 421, 7) ('##phone', 616, 8) ### ### ('contact', 1125, 9) ('tip', 0, 10) ('##rr', 28912, 11) ('plug', 1, 12) ('conductor', 5715, 13) ### ### ('phone', 9, 14) ('stereo', 16543, 15) ('diameter', 754, 16) ('ts', 7986, 17) ('ring', 2, 18) ### ### ('contacts', 2913, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('tip', 0, 10) ('plug', 1, 12) ('ring', 2, 18) ('phone', 9, 14) ('tips', 16, 47) ### ### ('connector', 835, 0) ('audio', 421, 7) ('##phone', 616, 8) ('telephone', 8, 85) ### ### ('diameter', 754, 16) ('microphone', 1463, 4) ('contact', 1125, 9) ('used', 17, 61) ### ### ('applications', 38, 46) ('typically', 75, 45) ('″', 248, 31) ('rings', 28, 60) ('tipped', 19, 70) ### ### ('versions', 310, 34) ('application', 54, 62) ### ############################################################################################################ [2023-11-25 04:26:11,997][root][INFO] - Epoch: 1: Step: 17601/19530, loss[v]=0.082831, lr=0.000018, acc@1[1]=998.0/1024=0.974609375, acc@1[2]=1016.0/1024=0.9921875 [2023-11-25 04:28:02,079][root][INFO] - Train batch 17700 [2023-11-25 04:28:02,080][root][INFO] - Avg. loss per last 100 batches: 0.211147 [2023-11-25 04:28:02,811][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.59 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 5.27 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: british sign language ( bsl ) is a sign language used in the united kingdom ( uk ), and ### ### is the first or preferred language of some deaf people in the uk. ### ### [P_TEXT]: sign language and dumb mans tutor ", in which he presented his own method of deaf ### ### education, including an " arthrological " alphabet, where letters are indicated by pointing to ### ### different joints of the fingers and palm of the left hand. arthrological systems had been in use by ### ### hearing people for some time ; some have speculated that they can be traced to early ogham manual ### ### alphabets. the vowel s of this alphabet have survived in the modern alphabets used in british sign ### ### language, auslan and new zealand sign language. the earliest known printed ### ### =============================================== v_q ================================================ ### ### ('language', 0, 11) ('bs', 1, 20742) ('sign', 2, 0) ('deaf', 3, 4) ('british', 4, 29) ### ### ('uk', 5, 105) ('##l', 6, 23939) ('##ل', 7, 2379) ('absently', 8, 373) ('.', 9, 148) ### ### ('preferred', 10, 1815) ('intently', 11, 367) ('britain', 12, 38) ('kingdom', 13, 3965) ### ### ('london', 14, 149) ('preference', 15, 14624) ('communication', 16, 396) ### ### ('hertfordshire', 17, 13804) ('bbc', 18, 244) ('whilst', 19, 360) ### ### =============================================== v_p ================================================ ### ### ('sign', 2, 0) ('alphabet', 304, 1) ('manual', 213, 2) ('og', 21925, 3) ('deaf', 3, 4) ### ### ('mans', 20629, 5) ('vowel', 1920, 6) ('aus', 17293, 7) ('joints', 17955, 8) ('zealand', 402, 9) ### ### ('##lan', 25655, 10) ('language', 0, 11) ('tutor', 1471, 12) ('dumb', 8470, 13) ('hearing', 24, 14) ### ### ('##ological', 3700, 15) ('palm', 3985, 16) ('fingers', 574, 17) ('education', 2713, 18) ### ### ('##hr', 27557, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('sign', 2, 0) ('language', 0, 11) ('deaf', 3, 4) ('british', 4, 29) ('hearing', 24, 14) ### ### ('manual', 213, 2) ('languages', 29, 23) ('britain', 12, 38) ('alphabet', 304, 1) ### ### ('signs', 116, 21) ('zealand', 402, 9) ('used', 31, 37) ('uk', 5, 105) ('digits', 80, 30) ### ### ('fingers', 574, 17) ('anyway', 25, 49) ('vowel', 1920, 6) ('tutor', 1471, 12) ### ### ('earliest', 178, 32) ('letters', 259, 33) ### ############################################################################################################ [2023-11-25 04:28:03,312][root][INFO] - Epoch: 1: Step: 17701/19530, loss[v]=0.105932, lr=0.000018, acc@1[1]=993.5/1024=0.97021484375, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 04:29:53,589][root][INFO] - Train batch 17800 [2023-11-25 04:29:53,590][root][INFO] - Avg. loss per last 100 batches: 0.194464 [2023-11-25 04:29:54,304][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.70 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 5.13 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.8/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: a reduced instruction set computer, or risc ( ), is a computer with a small, highly ### ### optimized set of instructions, rather than the more specialized set often found in other types of ### ### architecture, such as in a complex instruction set computer ( cisc ). ### ### [P_TEXT]: system on a chip for a specific type of workload. multiprocessor socs have more than one ### ### processor core by definition. whether single - core, multi - core or manycore, soc processor cores ### ### typically use risc instruction set architectures. risc architectures are advantageous over cisc ### ### processors for socs because they require less digital logic, and therefore less power and area on ### ### board, and in the embedded and mobile computing markets, area and power are often highly ### ### constrained. in particular, soc processor cores often use the arm architecture because it is a soft ### ### processor specified as an ip core and is more power efficient than x86. socs must have ### ### =============================================== v_q ================================================ ### ### ('reduced', 0, 6060) ('cis', 1, 4) ('instruction', 2, 13) ('computer', 3, 53) ('set', 4, 29) ### ### ('##sc', 5, 23) ('ri', 6, 15) ('instructions', 7, 70) ('reduction', 8, 2696) ('absently', 9, 259) ### ### ('complex', 10, 5327) ('.', 11, 82) ('reduce', 12, 1742) ('opt', 13, 18423) ('intently', 14, 204) ### ### ('##c', 15, 36) ('computers', 16, 49) ('reducing', 17, 6777) ('used', 18, 101) ('device', 19, 431) ### ### =============================================== v_p ================================================ ### ### ('chip', 237, 0) ('##load', 12776, 1) ('cores', 681, 2) ('soc', 7420, 3) ('cis', 1, 4) ### ### ('arm', 6101, 5) ('embedded', 91, 6) ('core', 74, 7) ('##core', 14839, 8) ('ip', 1491, 9) ### ### ('soft', 3541, 10) ('mobile', 1035, 11) ('computing', 26, 12) ('instruction', 2, 13) ### ### ('architecture', 40, 14) ('ri', 6, 15) ('##86', 5613, 16) ('multi', 7731, 17) ('power', 229, 18) ### ### ('processor', 104, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('cis', 1, 4) ('instruction', 2, 13) ('ri', 6, 15) ('##sc', 5, 23) ('set', 4, 29) ### ### ('computing', 26, 12) ('core', 74, 7) ('embedded', 91, 6) ('computer', 3, 53) ### ### ('architecture', 40, 14) ('chip', 237, 0) ('##c', 15, 36) ('system', 56, 25) ('processor', 104, 19) ### ### ('instructions', 7, 70) ('sets', 25, 39) ('cores', 681, 2) ('power', 229, 18) ('computers', 16, 49) ### ### ('logic', 315, 21) ### ############################################################################################################ [2023-11-25 04:29:54,808][root][INFO] - Epoch: 1: Step: 17801/19530, loss[v]=0.253215, lr=0.000018, acc@1[1]=960.0/1024=0.9375, acc@1[2]=996.5/1024=0.97314453125 [2023-11-25 04:31:45,208][root][INFO] - Train batch 17900 [2023-11-25 04:31:45,208][root][INFO] - Avg. loss per last 100 batches: 0.216489 [2023-11-25 04:31:45,935][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.04 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.8/29522=99.99% | mean: 0.02 | max: 4.85 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the lake, known as a wonderful spot for largemouth bass, is located within greenbo lake ### ### state resort park. ### ### [P_TEXT]: greenbo lake state resort park greenbo lake state resort park in kentucky is a resort ### ### park in the northeastern part of the commonwealth, close to the town of greenup, kentucky in ### ### greenup county on kentucky state route 1. it features a 36 - room lodge named for greenup county ### ### resident and writer jesse hilton stuart, a 63 - site campground with 35 primitive sites, a swimming ### ### pool with slides, two tennis courts, an 18 - hole miniature golf course, an amphitheater and a ### ### scuba refuge area. the lodge contains a 232 - seat dining room. it is centered on the greenbo lake ### ### =============================================== v_q ================================================ ### ### ('##bo', 0, 6) ('green', 1, 8) ('lake', 2, 20) ('bass', 3, 9407) ('resort', 4, 16) ### ### ('##bos', 5, 430) ('park', 6, 40) ('state', 7, 26) ('large', 8, 4997) ('##mouth', 9, 24982) ### ### ('absently', 10, 174) ('intently', 11, 171) ('alaska', 12, 4512) ('utah', 13, 1138) ### ### ('lakes', 14, 95) ('michigan', 15, 528) ('colorado', 16, 1284) ('unincorporated', 17, 17350) ### ### ('river', 18, 1134) ('creek', 19, 1390) ### ### =============================================== v_p ================================================ ### ### ('##up', 7702, 0) ('stuart', 5556, 1) ('scuba', 1629, 2) ('kentucky', 261, 3) ('hilton', 5989, 4) ### ### ('commonwealth', 281, 5) ('##bo', 0, 6) ('miniature', 7020, 7) ('green', 1, 8) ('refuge', 588, 9) ### ### ('primitive', 1149, 10) ('jesse', 10075, 11) ('room', 1761, 12) ('1', 20307, 13) ### ### ('dining', 10359, 14) ('lodge', 140, 15) ('resort', 4, 16) ('route', 401, 17) ('63', 13636, 18) ### ### ('##heater', 509, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##bo', 0, 6) ('green', 1, 8) ('lake', 2, 20) ('resort', 4, 16) ('state', 7, 26) ('park', 6, 40) ### ### ('kentucky', 261, 3) ('commonwealth', 281, 5) ('lodge', 140, 15) ('campground', 32, 32) ### ### ('refuge', 588, 9) ('county', 113, 22) ('scuba', 1629, 2) ('route', 401, 17) ('##bos', 5, 430) ### ### ('primitive', 1149, 10) ('##heater', 509, 19) ('site', 176, 36) ('room', 1761, 12) ### ### ('anyway', 88, 49) ### ############################################################################################################ [2023-11-25 04:31:46,438][root][INFO] - Epoch: 1: Step: 17901/19530, loss[v]=0.126709, lr=0.000018, acc@1[1]=989.0/1024=0.9658203125, acc@1[2]=1008.5/1024=0.98486328125 [2023-11-25 04:33:36,855][root][INFO] - Train batch 18000 [2023-11-25 04:33:36,855][root][INFO] - Avg. loss per last 100 batches: 0.189143 [2023-11-25 04:33:37,593][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.72 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 5.39 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: during the last deglaciation, numerous glacial lake outburst flood s were caused by the ### ### collapse of either ice sheets or glaciers that formed the dams of proglacial lake s. examples of ### ### older outburst floods are known from the geological past of the earth and inferred from ### ### geomorphological evidence on mars. ### ### [P_TEXT]: glacial lake outburst flood a glacial lake outburst flood ( glof ) is a type of outburst ### ### flood caused by the failure of a dam containing a glacial lake. an event similar to a glof, where a ### ### body of water contained by a glacier melts or overflows the glacier, is called a jokulhlaup. the ### ### dam can consist of glacier ice or a terminal moraine. failure can happen due to erosion, a buildup ### ### of water pressure, an avalanche of rock or heavy snow, an earthquake or cryoseism, volcanic ### ### eruptions under the ice, or massive displacement of water in a glacial lake when ### ### =============================================== v_q ================================================ ### ### ('flood', 0, 3) ('lake', 1, 21) ('glacial', 2, 0) ('##gl', 3, 24283) ('mars', 4, 15337) ### ### ('pro', 5, 16548) ('outburst', 6, 2) ('geological', 7, 136) ('absently', 8, 208) ### ### ('##ial', 9, 23034) ('dams', 10, 49) ('collapse', 11, 1094) ('past', 12, 9928) ('de', 13, 24987) ### ### ('.', 14, 189) ('##ac', 15, 26534) ('glacier', 16, 9) ('geologic', 17, 3382) ('earth', 18, 6214) ### ### ('floods', 19, 38) ### ### =============================================== v_p ================================================ ### ### ('glacial', 2, 0) ('##ose', 24448, 1) ('outburst', 6, 2) ('flood', 0, 3) ('cry', 2670, 4) ### ### ('earthquake', 124, 5) ('##hl', 13546, 6) ('volcanic', 72, 7) ('displacement', 3228, 8) ### ### ('glacier', 16, 9) ('##lo', 19685, 10) ('##flow', 2541, 11) ('mora', 4661, 12) ('erosion', 197, 13) ### ### ('snow', 123, 14) ('glaciers', 87, 15) ('avalanche', 785, 16) ('terminal', 13529, 17) ### ### ('##f', 25064, 18) ('ice', 20, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('flood', 0, 3) ('glacial', 2, 0) ('outburst', 6, 2) ('lake', 1, 21) ('glacier', 16, 9) ### ### ('ice', 20, 19) ('floods', 19, 38) ('volcanic', 72, 7) ('dams', 10, 49) ('earthquake', 124, 5) ### ### ('glaciers', 87, 15) ('snow', 123, 14) ('eruption', 80, 27) ('erosion', 197, 13) ### ### ('flooding', 54, 35) ('geological', 7, 136) ('dam', 303, 25) ('body', 179, 36) ('bodies', 32, 51) ### ### ('caused', 33, 52) ### ############################################################################################################ [2023-11-25 04:33:38,095][root][INFO] - Epoch: 1: Step: 18001/19530, loss[v]=0.133129, lr=0.000018, acc@1[1]=991.0/1024=0.9677734375, acc@1[2]=1005.5/1024=0.98193359375 [2023-11-25 04:35:28,363][root][INFO] - Train batch 18100 [2023-11-25 04:35:28,364][root][INFO] - Avg. loss per last 100 batches: 0.197019 [2023-11-25 04:35:29,090][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.76 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 5.60 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: as candidate for lehendakari they chose pilar zabala, who had already been chosen by the ### ### bases of podemos / ahal dugu in primary elections. ### ### [P_TEXT]: pilar zabala artano maria pilar zabala artano ( born 1968 in tolosa ), commonly known as ### ### pili zabala is an odontologist, professor and spanish politician, as well as elkarrekin podemos ### ### candidate for lehendakari in the 2016 basque parliamentary election. born in tolosa in 1968, she ### ### resides in zarautz. she is married and mother to a son and a daughter. the disappearance of her ### ### brother joxi zabala in 1983 marked her life ; joxi, an alleged member of eta, was kidnapped, ### ### tortured and murdered by the gal. since then, she became an ### ### =============================================== v_q ================================================ ### ### ('##bala', 0, 1) ('ah', 1, 11544) ('##aka', 2, 9) ('pi', 3, 24) ('za', 4, 15) ('dug', 5, 9788) ### ### ('##lar', 6, 11) ('pod', 7, 19) ('##hend', 8, 5) ('##em', 9, 31) ('le', 10, 56) ('##ri', 11, 45) ### ### ('bases', 12, 8108) ('elections', 13, 32) ('primary', 14, 2895) ('##u', 15, 19772) ### ### ('primaries', 16, 840) ('##al', 17, 22801) ('base', 18, 8242) ('##ு', 19, 2255) ### ### =============================================== v_p ================================================ ### ### ('eta', 3692, 0) ('##bala', 0, 1) ('gal', 20252, 2) ('##utz', 19652, 3) ('##los', 395, 4) ### ### ('##hend', 8, 5) ('basque', 1663, 6) ('##rek', 28669, 7) ('##don', 25747, 8) ('##aka', 2, 9) ### ### ('elk', 21018, 10) ('##lar', 6, 11) ('2016', 87, 12) ('##ano', 22229, 13) ('##xi', 26394, 14) ### ### ('za', 4, 15) ('art', 22535, 16) ('parliamentary', 7470, 17) ('jo', 24194, 18) ('pod', 7, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##bala', 0, 1) ('##aka', 2, 9) ('##hend', 8, 5) ('##lar', 6, 11) ('za', 4, 15) ('pi', 3, 24) ### ### ('pod', 7, 19) ('##em', 9, 31) ('elections', 13, 32) ('##ri', 11, 45) ('##os', 20, 34) ### ### ('election', 26, 26) ('le', 10, 56) ('2016', 87, 12) ('##los', 395, 4) ('spanish', 85, 25) ### ### ('basque', 1663, 6) ('spain', 161, 44) ('eta', 3692, 0) ('resides', 702, 21) ### ############################################################################################################ [2023-11-25 04:35:29,593][root][INFO] - Epoch: 1: Step: 18101/19530, loss[v]=0.083111, lr=0.000019, acc@1[1]=995.5/1024=0.97216796875, acc@1[2]=1016.0/1024=0.9921875 [2023-11-25 04:37:19,623][root][INFO] - Train batch 18200 [2023-11-25 04:37:19,624][root][INFO] - Avg. loss per last 100 batches: 0.203290 [2023-11-25 04:37:20,350][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.58 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 5.53 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: jobst of moravia ( or " jost lucembursky " ; or " jodokus von mahren " ; c. 1354 – 18 ### ### january 1411 ), a member of the house of luxembourg, was margrave of moravia from 1375, duke of ### ### luxembourg and elector of brandenburg from 1388 as well as elected king of germany ( king of the ### ### romans ) from 1410 until his death. ### ### [P_TEXT]: john henry, margrave of moravia banned both louis and margaret ; nevertheless, the ### ### luxembourg rule over tyrol was terminated. in 1346 john died in the battle of crecy and was ### ### succeeded by his eldest son, charles. after john henry's marriage was conclusively terminated ### ### according to canon law in 1349, he married margaret, daughter of duke nicholas ii of opava, and ### ### charles iv gave him the margraviate of moravia as appanage. in turn, john henry had to renounce all ### ### rights to the bohemian throne. his second marriage produced several sons, including the future ### ### margrave jobst of moravia. after margaret of opava died in 1363, john henry married margaret, the ### ### daughter of duke ### ### =============================================== v_q ================================================ ### ### ('mora', 0, 8) ('luxembourg', 1, 10) ('jobs', 2, 16) ('luce', 3, 14159) ('jo', 4, 23133) ### ### ('##hre', 5, 28287) ('##via', 6, 7) ('##urs', 7, 25093) ('germany', 8, 702) ('141', 9, 940) ### ### ('king', 10, 1657) ('house', 11, 12721) ('brandenburg', 12, 13479) ('absently', 13, 278) ### ### ('princely', 14, 761) ('duke', 15, 33) ('sovereign', 16, 1467) ('135', 17, 124) ('mar', 18, 30) ### ### ('.', 19, 225) ### ### =============================================== v_p ================================================ ### ### ('nicholas', 471, 0) ('tyrol', 822, 1) ('john', 1588, 2) ('henry', 230, 3) ('bohemian', 3364, 4) ### ### ('margaret', 6109, 5) ('##ec', 17423, 6) ('##via', 6, 7) ('mora', 0, 8) ('op', 17692, 9) ### ### ('luxembourg', 1, 10) ('##grave', 49, 11) ('canon', 4391, 12) ('banned', 7607, 13) ### ### ('battle', 525, 14) ('##ava', 18697, 15) ('jobs', 2, 16) ('iv', 6538, 17) ('app', 16825, 18) ### ### ('charles', 639, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('mora', 0, 8) ('luxembourg', 1, 10) ('jobs', 2, 16) ('##via', 6, 7) ('mar', 18, 30) ### ### ('duke', 15, 33) ('##grave', 49, 11) ('medieval', 42, 24) ('henry', 230, 3) ('nicholas', 471, 0) ### ### ('134', 157, 22) ('tyrol', 822, 1) ('throne', 158, 28) ('battle', 525, 14) ('nobility', 66, 40) ### ### ('law', 416, 26) ('136', 564, 21) ('john', 1588, 2) ('charles', 639, 19) ('##t', 26, 88) ### ############################################################################################################ [2023-11-25 04:37:20,853][root][INFO] - Epoch: 1: Step: 18201/19530, loss[v]=0.269806, lr=0.000019, acc@1[1]=966.0/1024=0.943359375, acc@1[2]=991.5/1024=0.96826171875 [2023-11-25 04:39:10,812][root][INFO] - Train batch 18300 [2023-11-25 04:39:10,813][root][INFO] - Avg. loss per last 100 batches: 0.203972 [2023-11-25 04:39:11,553][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.04 | max: 3.67 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 5.10 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: chodron teaches the traditional " yarne " retreat at gampo abbey each winter and the " ### ### guide to the bodhisattva's way of life " in berkeley each summer. ### ### [P_TEXT]: gampo abbey of tibetan buddhism. residents of gampo abbey include monks and nuns who have ### ### taken life ordination, monks and nuns who have taken temporary ordination, and laymen and laywomen. ### ### the life monastics are all ordained in the mulasarvastivadin lineages of the vinaya, or in the case ### ### of the bhiksunis, a combination of the mulasarvastivadin and dharmaguptaka lineages. gampo abbey's ### ### guiding teacher is the well - known author, buddhist nun, and teacher pema chodron. the abbey also ### ### has ties to the local cape breton shambhala sangha. the ### ### =============================================== v_q ================================================ ### ### ('berkeley', 0, 9985) ('yarn', 1, 13289) ('##dron', 2, 3) ('cho', 3, 18) ('##mp', 4, 1) ### ### ('ga', 5, 21) ('abbey', 6, 9) ('buddhist', 7, 23) ('guide', 8, 66) ('california', 9, 4331) ### ### ('retreat', 10, 2951) ('winter', 11, 13663) ('absently', 12, 323) ('life', 13, 25) ### ### ('##sat', 14, 22338) ('buddha', 15, 484) ('teaches', 16, 934) ('nepal', 17, 190) ('.', 18, 198) ### ### ('bhutan', 19, 864) ### ### =============================================== v_p ================================================ ### ### ('breton', 11697, 0) ('##mp', 4, 1) ('vin', 17419, 2) ('##dron', 2, 3) ('lay', 169, 4) ### ### ('tibetan', 21, 5) ('sang', 4108, 6) ('##pta', 24098, 7) ('##las', 14658, 8) ('abbey', 6, 9) ### ### ('monastic', 550, 10) ('##tiv', 22032, 11) ('sham', 6610, 12) ('##aya', 2709, 13) ### ### ('ordination', 1081, 14) ('ordained', 2494, 15) ('guiding', 2576, 16) ('temporary', 22450, 17) ### ### ('cho', 3, 18) ('cape', 2294, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##dron', 2, 3) ('##mp', 4, 1) ('cho', 3, 18) ('abbey', 6, 9) ('ga', 5, 21) ('buddhist', 7, 23) ### ### ('tibetan', 21, 5) ('life', 13, 25) ('tibet', 24, 31) ('dharma', 45, 20) ('buddhism', 28, 33) ### ### ('monks', 32, 37) ('##o', 34, 38) ('lay', 169, 4) ('guide', 8, 66) ('monastery', 43, 40) ### ### ('monastic', 550, 10) ('ordination', 1081, 14) ('lineage', 619, 26) ('chinese', 27, 84) ### ############################################################################################################ [2023-11-25 04:39:12,056][root][INFO] - Epoch: 1: Step: 18301/19530, loss[v]=0.107557, lr=0.000019, acc@1[1]=995.5/1024=0.97216796875, acc@1[2]=1014.5/1024=0.99072265625 [2023-11-25 04:41:02,065][root][INFO] - Train batch 18400 [2023-11-25 04:41:02,065][root][INFO] - Avg. loss per last 100 batches: 0.208418 [2023-11-25 04:41:02,802][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.04 | max: 3.59 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=99.99% | mean: 0.03 | max: 5.34 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the fa is a member of both uefa and fifa and holds a permanent seat on the international ### ### football association board ( ifab ) which is responsible for the laws of the game. ### ### [P_TEXT]: laws of the game ( association football ) date back to 1863 where a ruleset was formally ### ### adopted by the newly formed football association. over time, the laws have been amended, and since ### ### 1886 they have been maintained by the international football association board. they are the only ### ### rules of association football fifa permits its members to use. the laws allow some minor optional ### ### variations which can be implemented by national football associations, including some for play at ### ### the lowest levels, but otherwise almost all organised football worldwide is played under the same ### ### ruleset. the current laws of the game ### ### =============================================== v_q ================================================ ### ### ('##ab', 0, 24864) ('board', 1, 2) ('fa', 2, 6455) ('uefa', 3, 1527) ('international', 4, 14) ### ### ('football', 5, 1) ('absently', 6, 299) ('laws', 7, 0) ('if', 8, 3014) ('association', 9, 3) ### ### ('fifa', 10, 10) ('.', 11, 124) ('intently', 12, 313) ('footballer', 13, 35) ('law', 14, 30) ### ### ('league', 15, 506) ('federation', 16, 1308) ('footballers', 17, 45) ('governing', 18, 533) ### ### ('icao', 19, 834) ### ### =============================================== v_p ================================================ ### ### ('laws', 7, 0) ('football', 5, 1) ('board', 1, 2) ('association', 9, 3) ('1863', 11269, 4) ### ### ('rules', 27, 5) ('game', 20, 6) ('associations', 265, 7) ('organised', 25, 8) ('##et', 21536, 9) ### ### ('fifa', 10, 10) ('boards', 1109, 11) ('worldwide', 1007, 12) ('amended', 1156, 13) ### ### ('international', 4, 14) ('variations', 5709, 15) ('levels', 1025, 16) ('membership', 162, 17) ### ### ('1886', 20013, 18) ('optional', 6852, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('board', 1, 2) ('football', 5, 1) ('laws', 7, 0) ('association', 9, 3) ('international', 4, 14) ### ### ('fifa', 10, 10) ('game', 20, 6) ('rules', 27, 5) ('organised', 25, 8) ('law', 14, 30) ### ### ('footballer', 13, 35) ('associations', 265, 7) ('footballers', 17, 45) ('membership', 162, 17) ### ### ('.', 11, 124) ('shirt', 39, 37) ('absently', 6, 299) ('worldwide', 1007, 12) ('boards', 1109, 11) ### ### ('lowest', 553, 20) ### ############################################################################################################ [2023-11-25 04:41:03,306][root][INFO] - Epoch: 1: Step: 18401/19530, loss[v]=0.137918, lr=0.000019, acc@1[1]=990.0/1024=0.966796875, acc@1[2]=1007.0/1024=0.9833984375 [2023-11-25 04:42:53,050][root][INFO] - Train batch 18500 [2023-11-25 04:42:53,050][root][INFO] - Avg. loss per last 100 batches: 0.218551 [2023-11-25 04:42:53,766][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.04 | max: 3.54 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 5.13 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.6/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: muddy fork is a tributary, about long, of the sandy river in the u. s. state of oregon. ### ### [P_TEXT]: sandy glacier sandy glacier is an alpine glacier located on the west slope of mount hood ### ### in the u. s. state of oregon. it extends in elevation from about. the glacier is the source of ### ### muddy fork, a tributary of the sandy river. the upper extent of the glacier is known for extensive ### ### crevasse s. the lower extent was marked by the sandy glacier caves, possibly the largest glacier ### ### cave system in the lower 48 u. s. states. the glacier is a remnant of the massive glaciers ### ### =============================================== v_q ================================================ ### ### ('muddy', 0, 6) ('sandy', 1, 0) ('oregon', 2, 7) ('fork', 3, 4) ('river', 4, 19) ('forks', 5, 47) ### ### ('tributary', 6, 18) ('.', 7, 163) ('rivers', 8, 23) ('long', 9, 6780) ('states', 10, 25) ### ### ('sand', 11, 55) ('portland', 12, 252) ('washington', 13, 2326) ('creek', 14, 997) ### ### ('columbia', 15, 1571) ('headwaters', 16, 248) ('maine', 17, 457) ('proponents', 18, 820) ### ### ('absently', 19, 273) ### ### =============================================== v_p ================================================ ### ### ('sandy', 1, 0) ('hood', 3699, 1) ('glacier', 114, 2) ('cave', 453, 3) ('fork', 3, 4) ### ### ('slope', 2177, 5) ('muddy', 0, 6) ('oregon', 2, 7) ('##sse', 19064, 8) ('glaciers', 18237, 9) ### ### ('massive', 865, 10) ('48', 4690, 11) ('mount', 3134, 12) ('alpine', 2647, 13) ('##eva', 25612, 14) ### ### ('caves', 2460, 15) ('tributaries', 24, 16) ('remnant', 2720, 17) ('tributary', 6, 18) ### ### ('river', 4, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('sandy', 1, 0) ('muddy', 0, 6) ('oregon', 2, 7) ('fork', 3, 4) ('river', 4, 19) ### ### ('tributary', 6, 18) ('glacier', 114, 2) ('rivers', 8, 23) ('tributaries', 24, 16) ### ### ('states', 10, 25) ('cave', 453, 3) ('forks', 5, 47) ('americans', 47, 20) ('massive', 865, 10) ### ### ('hood', 3699, 1) ('slope', 2177, 5) ('sand', 11, 55) ('extent', 196, 24) ('alpine', 2647, 13) ### ### ('caves', 2460, 15) ### ############################################################################################################ [2023-11-25 04:42:54,268][root][INFO] - Epoch: 1: Step: 18501/19530, loss[v]=0.265798, lr=0.000019, acc@1[1]=965.5/1024=0.94287109375, acc@1[2]=997.0/1024=0.9736328125 [2023-11-25 04:44:44,350][root][INFO] - Train batch 18600 [2023-11-25 04:44:44,351][root][INFO] - Avg. loss per last 100 batches: 0.202552 [2023-11-25 04:44:45,074][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.60 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 5.43 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in water, hexafluorosilicic acid readily hydrolyzes to hydrofluoric acid and various ### ### forms of amorphous and hydrated silica ( " sio2 " ). ### ### [P_TEXT]: silicon dioxide onto crystal surface from silane had been used using nitrogen as a ### ### carrier gas at 200 – 500 °c. fluorine reacts with silicon dioxide to form sif4 and o2 whereas the ### ### other halogen gases ( cl2, br2, i2 ) are essentially unreactive. silicon dioxide is attacked by ### ### hydrofluoric acid ( hf ) to produce hexafluorosilicic acid : hf is used to remove or pattern ### ### silicon dioxide in the semiconductor industry. under normal conditions, silicon does not react with ### ### most acids but is dissolved by hydrofluoric acid. silicon is attacked by bases such as aqueous ### ### sodium hydroxide to give silicates. silicon dioxide acts ### ### =============================================== v_q ================================================ ### ### ('si', 0, 14) ('amor', 1, 8407) ('acid', 2, 28) ('##xa', 3, 10) ('##fl', 4, 41) ### ### ('absently', 5, 287) ('.', 6, 81) ('intently', 7, 153) ('##ili', 8, 30) ('##cic', 9, 6) ### ### ('chemist', 10, 5942) ('##lica', 11, 20) ('hydra', 12, 3125) ('##ros', 13, 46) ('##uo', 14, 40) ### ### ('##ric', 15, 44) ('##ो', 16, 876) ('chemical', 17, 70) ('chemicals', 18, 173) ('##⁄₄', 19, 133) ### ### =============================================== v_p ================================================ ### ### ('silicon', 643, 0) ('dioxide', 77, 1) ('carrier', 2946, 2) ('pattern', 779, 3) ('onto', 1292, 4) ### ### ('halo', 9818, 5) ('##cic', 9, 6) ('##orin', 8071, 7) ('##lane', 6894, 8) ('hydro', 28, 9) ### ### ('##xa', 3, 10) ('sodium', 1085, 11) ('##f', 10058, 12) ('br', 23662, 13) ('si', 0, 14) ### ### ('nitrogen', 83, 15) ('semiconductor', 503, 16) ('o', 9337, 17) ('carriers', 16032, 18) ### ### ('crystal', 1206, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('si', 0, 14) ('##xa', 3, 10) ('acid', 2, 28) ('##cic', 9, 6) ('##lica', 11, 20) ('##ili', 8, 30) ### ### ('##fl', 4, 41) ('dioxide', 77, 1) ('hydro', 28, 9) ('##uo', 14, 40) ('##ros', 13, 46) ### ### ('##ric', 15, 44) ('##₂', 29, 33) ('##2', 30, 31) ('silicon', 643, 0) ('nitrogen', 83, 15) ### ### ('surface', 43, 32) ('.', 6, 81) ('gas', 146, 27) ('crystals', 81, 38) ### ############################################################################################################ [2023-11-25 04:44:45,577][root][INFO] - Epoch: 1: Step: 18601/19530, loss[v]=0.189681, lr=0.000019, acc@1[1]=986.5/1024=0.96337890625, acc@1[2]=1008.5/1024=0.98486328125 [2023-11-25 04:46:35,490][root][INFO] - Train batch 18700 [2023-11-25 04:46:35,491][root][INFO] - Avg. loss per last 100 batches: 0.199011 [2023-11-25 04:46:36,206][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.61 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.4/29522=99.99% | mean: 0.03 | max: 5.01 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: on february 21, 2019, two days after the game's release, players discovered a " fulu " ### ### talisman decorating a wall in the game contained the words " xi jinping winnie the pooh " ( ) in ### ### chinese seal script, referencing a recent chinese internet meme that compared the chinese general ### ### secretary to the disney character. ### ### [P_TEXT]: review bomb review bomb the game. the game " devotion " by the taiwanese studio red ### ### candle games was review bombed by chinese players after they discovered an in - game poster that ### ### referenced the meme of chinese communist party general secretary xi jinping censoring the character ### ### winnie the pooh. while red candle stated the poster was not meant to be in the final game and took ### ### it out on the next immediate patch, the game was still hit with negative user reviews, forcing red ### ### candle to terminate their publishing deals and pull the game ### ### =============================================== v_q ================================================ ### ### ('disney', 0, 10466) ('##lu', 1, 28354) ('seal', 2, 5959) ('##me', 3, 8) ('game', 4, 25) ### ### ('##isman', 5, 26962) ('xi', 6, 22) ('chinese', 7, 14) ('winnie', 8, 2) ('internet', 9, 4988) ### ### ('##ping', 10, 15) ('secretary', 11, 7) ('tal', 12, 17030) ('fu', 13, 19617) ('script', 14, 924) ### ### ('po', 15, 23) ('absently', 16, 213) ('wall', 17, 3555) ('me', 18, 29) ('2019', 19, 21035) ### ### =============================================== v_p ================================================ ### ### ('devotion', 4155, 0) ('candle', 3292, 1) ('winnie', 8, 2) ('bombed', 28993, 3) ('bomb', 5623, 4) ### ### ('taiwanese', 2392, 5) ('poster', 3158, 6) ('secretary', 11, 7) ('##me', 3, 8) ('review', 12015, 9) ### ### ('taiwan', 32, 10) ('##nsor', 19840, 11) ('red', 6615, 12) ('##oh', 20, 13) ('chinese', 7, 14) ### ### ('##ping', 10, 15) ('communist', 8307, 16) ('immediate', 14277, 17) ('pull', 5858, 18) ### ### ('games', 25, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('winnie', 8, 2) ('##me', 3, 8) ('secretary', 11, 7) ('chinese', 7, 14) ('##ping', 10, 15) ### ### ('xi', 6, 22) ('game', 4, 25) ('##oh', 20, 13) ('po', 15, 23) ('general', 21, 26) ### ### ('taiwan', 32, 10) ('games', 25, 19) ('me', 18, 29) ('players', 35, 20) ('china', 23, 33) ### ### ('jin', 24, 32) ('″', 33, 34) ('player', 31, 41) ('generals', 58, 35) ('"', 40, 49) ### ############################################################################################################ [2023-11-25 04:46:36,709][root][INFO] - Epoch: 1: Step: 18701/19530, loss[v]=0.241007, lr=0.000019, acc@1[1]=965.0/1024=0.9423828125, acc@1[2]=988.5/1024=0.96533203125 [2023-11-25 04:48:26,300][root][INFO] - Train batch 18800 [2023-11-25 04:48:26,301][root][INFO] - Avg. loss per last 100 batches: 0.208900 [2023-11-25 04:48:27,027][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.04 | max: 3.74 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 5.38 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: they were often involved in street fighting with the protestant billy boys of bridgeton, ### ### but were on friendlier terms with the ### ### [P_TEXT]: gangs in the united kingdom their weapon of choice. one of glasgow's most notorious gangs ### ### were the billy boys, a sectarian anti - catholic gang, who were formed in 1924 by william fullerton ### ### after he was attacked by a group of catholic youths. many gangs in the east end of glasgow ( such ### ### as the billy boys'rivals, the norman conks ) were both sectarian and territorial, whereas in other ### ### districts they were primarily territorial. the gang culture prevalent in the older, central areas ### ### of the city such as the gorbals which became overcrowded and substandard in living conditions, did ### ### not ### ### =============================================== v_q ================================================ ### ### ('bridget', 0, 20932) ('billy', 1, 8) ('boys', 2, 10) ('fighting', 3, 2055) ('##on', 4, 21202) ### ### ('protestant', 5, 114) ('fight', 6, 1298) ('.', 7, 68) ('absently', 8, 192) ('vimes', 9, 338) ### ### ('militia', 10, 315) ('street', 11, 4400) ('fought', 12, 13680) ('dublin', 13, 90) ### ### ('glasgow', 14, 3) ('boy', 15, 53) ('ireland', 16, 314) ('belfast', 17, 96) ('girls', 18, 58) ### ### ('bonfire', 19, 7620) ### ### =============================================== v_p ================================================ ### ### ('gangs', 75, 0) ('territorial', 12443, 1) ('fuller', 4246, 2) ('glasgow', 14, 3) ### ### ('##crow', 1404, 4) ('choice', 6611, 5) ('rivals', 2897, 6) ('norman', 2561, 7) ('billy', 1, 8) ### ### ('##rba', 24948, 9) ('boys', 2, 10) ('con', 12055, 11) ('##ks', 10551, 12) ('youths', 2255, 13) ### ### ('1924', 5950, 14) ('end', 14646, 15) ('##stand', 24814, 16) ('gang', 532, 17) ('##ls', 19466, 18) ### ### ('rival', 1234, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('billy', 1, 8) ('boys', 2, 10) ('glasgow', 14, 3) ('gangs', 75, 0) ('catholic', 33, 20) ### ### ('.', 7, 68) ('boy', 15, 53) ('girls', 18, 58) ('attacked', 221, 27) ('protestant', 5, 114) ### ### ('gang', 532, 17) ('##crow', 1404, 4) ('dublin', 13, 90) ('william', 523, 22) ('uk', 91, 42) ### ### ('protestants', 118, 40) ('belfast', 17, 96) ('norman', 2561, 7) ('sect', 782, 26) ### ### ('scotland', 86, 48) ### ############################################################################################################ [2023-11-25 04:48:27,529][root][INFO] - Epoch: 1: Step: 18801/19530, loss[v]=0.192820, lr=0.000019, acc@1[1]=972.5/1024=0.94970703125, acc@1[2]=998.0/1024=0.974609375 [2023-11-25 04:50:17,485][root][INFO] - Train batch 18900 [2023-11-25 04:50:17,486][root][INFO] - Avg. loss per last 100 batches: 0.190417 [2023-11-25 04:50:18,221][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.57 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 5.70 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: during most of recorded history froson was the regional centre of jamtland, and it is the ### ### location of the froso runestone, the northernmost in the world. ### ### [P_TEXT]: froso runestone frosostenen ( j rs1928 ; 66 $ ) is the northernmost raised runestone in ### ### scandinavia and jamtland's only runestone. it originally stood at the tip of ferry terminal on the ### ### sound between the island of froson and ostersund. the following old norse person and place names ### ### appear in the inscription : the stone is also unique in that it was done in memory of austmaðr's ### ### christianization of jamtland and bridge building, rather than as a cenotaph. the stone dates to ### ### between 1030 ### ### =============================================== v_q ================================================ ### ### ('fr', 0, 21) ('##tland', 1, 11) ('##oso', 2, 0) ('jam', 3, 3) ('runes', 4, 4) ('norse', 5, 31) ### ### ('##n', 6, 48) ('rune', 7, 66) ('absently', 8, 217) ('swedish', 9, 63) ('northernmost', 10, 28) ### ### ('iceland', 11, 46) ('denmark', 12, 62) ('intently', 13, 169) ('sweden', 14, 43) ### ### ('copenhagen', 15, 197) ('.', 16, 154) ('danish', 17, 71) ('cumbria', 18, 6992) ('##tone', 19, 27) ### ### =============================================== v_p ================================================ ### ### ('##oso', 2, 0) ('sound', 3324, 1) ('ferry', 3245, 2) ('jam', 3, 3) ('runes', 4, 4) ### ### ('##tma', 26124, 5) ('##not', 8947, 6) ('$', 26824, 7) ('raised', 1707, 8) ('os', 1647, 9) ### ### ('bridge', 15122, 10) ('##tland', 1, 11) ('terminal', 16802, 12) ('aus', 14074, 13) ### ### ('rs', 8284, 14) ('##ters', 14553, 15) ('103', 6168, 16) ('building', 4931, 17) ### ### ('##ization', 16786, 18) ('christian', 1795, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##oso', 2, 0) ('##tland', 1, 11) ('jam', 3, 3) ('runes', 4, 4) ('fr', 0, 21) ('norse', 5, 31) ### ### ('northernmost', 10, 28) ('##n', 6, 48) ('##tone', 19, 27) ('iceland', 11, 46) ('sweden', 14, 43) ### ### ('##ð', 20, 37) ('swedish', 9, 63) ('rune', 7, 66) ('denmark', 12, 62) ('icelandic', 24, 51) ### ### ('scandinavia', 28, 55) ('stone', 34, 54) ('medieval', 30, 56) ('danish', 17, 71) ### ############################################################################################################ [2023-11-25 04:50:18,725][root][INFO] - Epoch: 1: Step: 18901/19530, loss[v]=0.102393, lr=0.000019, acc@1[1]=993.0/1024=0.9697265625, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 04:52:08,224][root][INFO] - Train batch 19000 [2023-11-25 04:52:08,225][root][INFO] - Avg. loss per last 100 batches: 0.191506 [2023-11-25 04:52:08,935][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.98 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 5.26 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: andrei andreyevich eshpai (, born 18 april 1956, moscow, russia ) is a ( russia n and ### ### soviet ) film director, screenwriter and producer of mari descent. ### ### [P_TEXT]: yakov eshpai yakov andreyevich eshpai ( ;, kokshamary – 20 february 1963, moscow ) was a ### ### russian composer and music teacher. he studied under georgi conus at the moscow conservatory. he ### ### was ethnic mari. he is partly noted for his work concerning the folk music of the mari people. he ### ### was the father of the better - known composer andrei eshpai, and the grandfather of the filmmaker ### ### andrei andreyevich eshpai. ### ### =============================================== v_q ================================================ ### ### ('##hp', 0, 3) ('andrei', 1, 5) ('mari', 2, 0) ('##ai', 3, 14) ('es', 4, 17) ('film', 5, 27) ### ### ('moscow', 6, 6) ('##evich', 7, 19) ('andrey', 8, 18) ('russia', 9, 41) ('soviet', 10, 312) ### ### ('screenwriter', 11, 1168) ('director', 12, 1217) ('descent', 13, 4224) ('russian', 14, 23) ### ### ('directors', 15, 747) ('directing', 16, 456) ('filmmakers', 17, 24) ('cinema', 18, 81) ### ### ('intently', 19, 225) ### ### =============================================== v_p ================================================ ### ### ('mari', 2, 0) ('##ksha', 26521, 1) ('conservatory', 3319, 2) ('##hp', 0, 3) ('##mar', 21926, 4) ### ### ('andrei', 1, 5) ('moscow', 6, 6) ('ethnic', 326, 7) ('georg', 9827, 8) ('folk', 3481, 9) ### ### ('grandfather', 713, 10) ('##kov', 724, 11) ('con', 13804, 12) ('ya', 10884, 13) ('##ai', 3, 14) ### ### ('people', 1333, 15) ('concerning', 894, 16) ('es', 4, 17) ('andrey', 8, 18) ('##evich', 7, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('mari', 2, 0) ('##hp', 0, 3) ('andrei', 1, 5) ('##ai', 3, 14) ('moscow', 6, 6) ('es', 4, 17) ### ### ('##evich', 7, 19) ('andrey', 8, 18) ('film', 5, 27) ('russian', 14, 23) ('russia', 9, 41) ### ### ('filmmakers', 17, 24) ('##い', 25, 34) ('ethnic', 326, 7) ('films', 22, 50) ### ### ('grandfather', 713, 10) ('##kov', 724, 11) ('filmmaker', 50, 37) ('composer', 142, 22) ### ### ('cinema', 18, 81) ### ############################################################################################################ [2023-11-25 04:52:09,438][root][INFO] - Epoch: 1: Step: 19001/19530, loss[v]=0.327639, lr=0.000019, acc@1[1]=959.5/1024=0.93701171875, acc@1[2]=987.0/1024=0.9638671875 [2023-11-25 04:53:59,183][root][INFO] - Train batch 19100 [2023-11-25 04:53:59,184][root][INFO] - Avg. loss per last 100 batches: 0.188899 [2023-11-25 04:53:59,919][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 5.27 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.7/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: for most of 2009, alex has been co - writing, arranging and drumming on a new pipettes ### ### album as well as assisting london - based band official secrets act. ### ### [P_TEXT]: the pipettes are credited to the pipettes as a whole. generally one member has a concept ### ### and basic outline for a song, on which the other members subsequently have an input as to ### ### arrangement or lyric changes. as a result, all members share credit on their original songs. the ### ### line - up consisting of rose ( " rosay " ), julia ( " the duchess ( of darkness ) " or " julia ### ### caesar " ), riotbecki, jon, seb, joe ( " robin of loxley " ) and bobby settled in mid - 2004, after ### ### earlier line - ups included members of the electric soft parade and marc beatty of brakes. in april ### ### 2005 founder member julia left to ### ### =============================================== v_q ================================================ ### ### ('##ttes', 0, 1) ('alex', 1, 3975) ('pipe', 2, 0) ('secrets', 3, 14640) ('act', 4, 618) ### ### ('official', 5, 10616) ('london', 6, 4173) ('band', 7, 45) ('album', 8, 49) ('drumming', 9, 217) ### ### ('uk', 10, 7982) ('acts', 11, 487) ('dex', 12, 81) ('whilst', 13, 359) ('pipes', 14, 39) ### ### ('bands', 15, 43) ('.', 16, 925) ('dogg', 17, 13281) ('musician', 18, 276) ### ### ('officially', 19, 12944) ### ### =============================================== v_p ================================================ ### ### ('pipe', 2, 0) ('##ttes', 0, 1) ('beatty', 16595, 2) ('parade', 12576, 3) ('rosa', 13815, 4) ### ### ('julia', 3805, 5) ('rose', 7527, 6) ('caesar', 17719, 7) ('darkness', 620, 8) ('robin', 1658, 9) ### ### ('brakes', 19809, 10) ('##xley', 17468, 11) ('bobby', 2932, 12) ('marc', 1374, 13) ('jon', 54, 14) ### ### ('joe', 2360, 15) ('duchess', 7574, 16) ('settled', 4331, 17) ('##beck', 6185, 18) ### ### ('riot', 3356, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ttes', 0, 1) ('pipe', 2, 0) ('band', 7, 45) ('jon', 54, 14) ('album', 8, 49) ### ### ('frontman', 26, 30) ('pipes', 14, 39) ('bands', 15, 43) ('##tte', 21, 41) ('darkness', 620, 8) ### ### ('tour', 173, 28) ('marc', 1374, 13) ('robin', 1658, 9) ('dex', 12, 81) ('act', 4, 618) ### ### ('joe', 2360, 15) ('artists', 29, 75) ('bobby', 2932, 12) ('music', 34, 70) ('julia', 3805, 5) ### ############################################################################################################ [2023-11-25 04:54:00,420][root][INFO] - Epoch: 1: Step: 19101/19530, loss[v]=0.281216, lr=0.000020, acc@1[1]=986.5/1024=0.96337890625, acc@1[2]=1006.5/1024=0.98291015625 [2023-11-25 04:55:50,503][root][INFO] - Train batch 19200 [2023-11-25 04:55:50,504][root][INFO] - Avg. loss per last 100 batches: 0.218286 [2023-11-25 04:55:51,226][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.72 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 5.30 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he was re - elected again following elections in 1964 and 1967. ### ### [P_TEXT]: 1970 western samoan general election to vote in elections. prior to 1967 elections, the ### ### number of matais was increased significantly, as candidates seeking to increase their vote were ### ### able to bestow titles on people to create new matai. following the elections, a law was introduced ### ### that prohibited conferring of a matai title on anyone younger than 21. although prime minister ### ### fiame mata'afa had been in favour of universal suffrage during the 1954 and 1960 constitutional ### ### conventions, he opposed the motion, arguing that voters had approved limiting suffrage to matais in ### ### the 1961 constitutional referendum, which had been ### ### =============================================== v_q ================================================ ### ### ('1967', 0, 9) ('1964', 1, 337) ('election', 2, 15) ('.', 3, 99) ('absently', 4, 297) ### ### ('1965', 5, 204) ('intently', 6, 233) ('anyway', 7, 64) ('67', 8, 142) ('1968', 9, 124) ### ### ('liberal', 10, 1115) ('1963', 11, 278) ('elections', 12, 10) ('electoral', 13, 515) ### ### ('party', 14, 769) ('minister', 15, 43) ('1969', 16, 36) ('reelection', 17, 2050) ### ### ('littered', 18, 272) ('##hyllum', 19, 763) ### ### =============================================== v_p ================================================ ### ### ('mata', 22611, 0) ('referendum', 581, 1) ('1970', 237, 2) ('samoa', 7304, 3) ### ### ('constitutional', 1797, 4) ('af', 18268, 5) ('suffrage', 6862, 6) ('1960', 235, 7) ### ### ('motion', 4172, 8) ('1967', 0, 9) ('elections', 12, 10) ('1954', 2428, 11) ### ### ('convention', 2013, 12) ('western', 6125, 13) ('1961', 997, 14) ('election', 2, 15) ### ### ('fia', 9148, 16) ('titles', 2182, 17) ('conventions', 19433, 18) ('reelected', 2005, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('1967', 0, 9) ('election', 2, 15) ('elections', 12, 10) ('1970', 237, 2) ('.', 3, 99) ### ### ('1960', 235, 7) ('referendum', 581, 1) ('1969', 16, 36) ('cabinet', 25, 32) ('anyway', 7, 64) ### ### ('minister', 15, 43) ('1964', 1, 337) ('absently', 4, 297) ('constitutional', 1797, 4) ### ### ('1968', 9, 124) ('elected', 36, 56) ('1961', 997, 14) ('67', 8, 142) ('favour', 412, 27) ### ### ('1965', 5, 204) ### ############################################################################################################ [2023-11-25 04:55:51,731][root][INFO] - Epoch: 1: Step: 19201/19530, loss[v]=0.267326, lr=0.000020, acc@1[1]=967.0/1024=0.9443359375, acc@1[2]=994.5/1024=0.97119140625 [2023-11-25 04:57:41,148][root][INFO] - Train batch 19300 [2023-11-25 04:57:41,148][root][INFO] - Avg. loss per last 100 batches: 0.197204 [2023-11-25 04:57:41,869][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.74 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 5.44 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: with xu guangqi, he participated in modifying the chinese calendar and compiling what is ### ### known as the chongzhen calendar. ### ### [P_TEXT]: apus in chinese astronomy the modern constellation apus is not included in the three ### ### enclosures and twenty - eight mansions system of traditional chinese uranography because its stars ### ### are too far south for observers in china to know about them prior to the introduction of western ### ### star charts. based on the work of xu guangqi and the german jesuit missionary johann adam schall ### ### von bell in the late ming dynasty, this constellation has been classified as one of the 23 southern ### ### asterisms ( 南 星, " jinnanjixingou " ) under the name exotic bird (, " yique " ). the name of the ### ### western constellation in ### ### =============================================== v_q ================================================ ### ### ('calendar', 0, 4715) ('##qi', 1, 8) ('chong', 2, 26766) ('xu', 3, 15) ('##ang', 4, 14) ### ### ('gu', 5, 41) ('##zh', 6, 24319) ('chinese', 7, 11) ('absently', 8, 547) ('depicts', 9, 1926) ### ### ('intently', 10, 361) ('china', 11, 48) ('garion', 12, 236) ('##馬', 13, 633) ('emperor', 14, 1791) ### ### ('##en', 15, 24823) ('vimes', 16, 2149) ('##子', 17, 218) ('celebrated', 18, 1210) ### ### ('##igraphy', 19, 4986) ### ### =============================================== v_p ================================================ ### ### ('ap', 19116, 0) ('constellation', 41, 1) ('##hall', 16825, 2) ('exotic', 13051, 3) ### ### ('enclosure', 1699, 4) ('astronomy', 509, 5) ('mansions', 20557, 6) ('bell', 9482, 7) ### ### ('##qi', 1, 8) ('##que', 6307, 9) ('charts', 6492, 10) ('chinese', 7, 11) ('##graphy', 4532, 12) ### ### ('##xing', 2858, 13) ('##ang', 4, 14) ('xu', 3, 15) ('##us', 3250, 16) ('ming', 1052, 17) ### ### ('stars', 6375, 18) ('western', 2955, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##qi', 1, 8) ('xu', 3, 15) ('##ang', 4, 14) ('chinese', 7, 11) ('gu', 5, 41) ### ### ('constellation', 41, 1) ('china', 11, 48) ('astronomy', 509, 5) ('dynasty', 154, 26) ### ### ('johann', 274, 20) ('anyway', 30, 51) ('yi', 384, 23) ('enclosure', 1699, 4) ('jin', 387, 36) ### ### ('ming', 1052, 17) ('traditional', 439, 43) ('##xing', 2858, 13) ('system', 731, 42) ### ### ('star', 2059, 22) ('western', 2955, 19) ### ############################################################################################################ [2023-11-25 04:57:42,372][root][INFO] - Epoch: 1: Step: 19301/19530, loss[v]=0.268489, lr=0.000020, acc@1[1]=957.5/1024=0.93505859375, acc@1[2]=989.0/1024=0.9658203125 [2023-11-25 04:59:32,139][root][INFO] - Train batch 19400 [2023-11-25 04:59:32,140][root][INFO] - Avg. loss per last 100 batches: 0.210358 [2023-11-25 04:59:32,879][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.67 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 5.07 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: one james martin & co built example was purchased second - hand from the tarrawingee ### ### tramway. ### ### [P_TEXT]: wagr g class in large numbers as the silverton tramway y class, south australian railways ### ### y class and tasmanian government railways c class, and also in queensland and on the emu bay ### ### railway and north australia railway. they were designed by beyer, peacock & co who built seven, ### ### with james martin & co building 29 and neilson & co 12. during world war ii, 13 were loaned to the ### ### commonwealth railways for use on the north australia railway as their nfc and nga classes. others ### ### were sold for further use by timber mill operators in ### ### =============================================== v_q ================================================ ### ### ('martin', 0, 6) ('##wing', 1, 21277) ('tramway', 2, 10) ('james', 3, 32) ('tar', 4, 16753) ### ### ('##ee', 5, 26464) ('##ra', 6, 21825) ('example', 7, 5748) ('&', 8, 47) ('australia', 9, 16) ### ### ('railway', 10, 18) ('queensland', 11, 25) ('absently', 12, 215) ('tram', 13, 71) ### ### ('australian', 14, 23) ('examples', 15, 5345) ('##ラ', 16, 936) ('melbourne', 17, 103) ### ### ('locomotives', 18, 38) ('intently', 19, 200) ### ### =============================================== v_p ================================================ ### ### ('nfc', 26883, 0) ('bey', 26469, 1) ('em', 8057, 2) ('silver', 9900, 3) ('commonwealth', 5678, 4) ### ### ('peacock', 1413, 5) ('martin', 0, 6) ('neil', 9117, 7) ('class', 297, 8) ('railways', 133, 9) ### ### ('tramway', 2, 10) ('ng', 4027, 11) ('tasmanian', 84, 12) ('building', 62, 13) ('##gr', 19314, 14) ### ### ('y', 8288, 15) ('australia', 9, 16) ('29', 15070, 17) ('railway', 10, 18) ('bay', 9573, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('martin', 0, 6) ('tramway', 2, 10) ('james', 3, 32) ('australia', 9, 16) ('railway', 10, 18) ### ### ('queensland', 11, 25) ('australian', 14, 23) ('&', 8, 47) ('locomotives', 18, 38) ### ### ('building', 62, 13) ('tasmania', 30, 27) ('tasmanian', 84, 12) ('railways', 133, 9) ('co', 21, 42) ### ### ('built', 23, 45) ('class', 297, 8) ('tram', 13, 71) ('anyway', 38, 39) ('trams', 35, 40) ### ### ('timber', 269, 21) ### ############################################################################################################ [2023-11-25 04:59:33,382][root][INFO] - Epoch: 1: Step: 19401/19530, loss[v]=0.230163, lr=0.000020, acc@1[1]=961.5/1024=0.93896484375, acc@1[2]=997.0/1024=0.9736328125 [2023-11-25 05:01:23,434][root][INFO] - Train batch 19500 [2023-11-25 05:01:23,435][root][INFO] - Avg. loss per last 100 batches: 0.189547 [2023-11-25 05:01:24,165][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.72 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 5.40 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he played on donovan's " barabajagal " ( 1968 ), as well as on jeff beck's " beck - ola " ### ### ( 1969 ). ### ### [P_TEXT]: micky waller style. the following year, a second album from the band was " beck - ola ". ### ### but by this point waller had already departed the band, to be replaced by tony newman. the ### ### mercurial beck had decided he wanted a " heavier " drummer, and on 11 february 1969 manager peter ### ### grant fired waller and wood - although wood was retained per beck's wishes. in 1969, waller flew to ### ### venice beach los angeles to form the band " silver metre " with guitarist leigh stephens of blue ### ### cheer fame, fellow brit pete sears on bass ### ### =============================================== v_q ================================================ ### ### ('donovan', 0, 9865) ('beck', 1, 7) ('##aba', 2, 28900) ('##gal', 3, 28773) ('jeff', 4, 11990) ### ### ('ol', 5, 13) ('##ja', 6, 26469) ('bar', 7, 16971) ('1969', 8, 20) ('album', 9, 36) ### ### ('1968', 10, 75) ('.', 11, 1413) ('absently', 12, 329) ('beatles', 13, 62) ('intently', 14, 182) ### ### ('hendrix', 15, 1154) ('pmid', 16, 767) ('blues', 17, 2168) ('reissue', 18, 2424) ### ### ('guitar', 19, 45) ### ### =============================================== v_p ================================================ ### ### ('waller', 799, 0) ('newman', 1352, 1) ('cheer', 2932, 2) ('sears', 4155, 3) ('stephens', 11439, 4) ### ### ('mick', 707, 5) ('leigh', 5180, 6) ('beck', 1, 7) ('##rial', 23364, 8) ('grant', 4128, 9) ### ### ('metre', 19137, 10) ('venice', 5330, 11) ('silver', 4646, 12) ('ol', 5, 13) ('heavier', 11855, 14) ### ### ('beach', 2951, 15) ('wood', 3232, 16) ('pete', 554, 17) ('tony', 297, 18) ('fame', 1774, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('beck', 1, 7) ('ol', 5, 13) ('1969', 8, 20) ('album', 9, 36) ('waller', 799, 0) ('band', 21, 32) ### ### ('drummer', 57, 27) ('mick', 707, 5) ('newman', 1352, 1) ('1968', 10, 75) ('guitar', 19, 45) ### ### ('tony', 297, 18) ('beatles', 13, 62) ('pete', 554, 17) ('cheer', 2932, 2) ('peter', 263, 26) ### ### ('bands', 168, 34) ('bass', 42, 46) ('sears', 4155, 3) ('anyway', 94, 41) ### ############################################################################################################ [2023-11-25 05:01:24,669][root][INFO] - Epoch: 1: Step: 19501/19530, loss[v]=0.277857, lr=0.000020, acc@1[1]=958.0/1024=0.935546875, acc@1[2]=976.0/1024=0.953125 [2023-11-25 05:01:56,453][root][INFO] - rank=0; last iteration 19530 [2023-11-25 05:01:56,453][root][INFO] - rank=0; Multi set iteration finished: iteration per set: [9765, 9765] [2023-11-25 05:01:56,453][root][INFO] - Finished iterating, iteration=9765, shard=0 [2023-11-25 05:01:56,454][root][INFO] - rank=2; last iteration 19530 [2023-11-25 05:01:56,455][root][INFO] - rank=2; Multi set iteration finished: iteration per set: [9765, 9765] [2023-11-25 05:01:56,455][root][INFO] - Finished iterating, iteration=9765, shard=2 [2023-11-25 05:01:56,455][root][INFO] - rank=1; last iteration 19530 [2023-11-25 05:01:56,455][root][INFO] - rank=1; Multi set iteration finished: iteration per set: [9765, 9765] [2023-11-25 05:01:56,455][root][INFO] - Finished iterating, iteration=9765, shard=1 [2023-11-25 05:01:56,455][root][INFO] - rank=3; last iteration 19530 [2023-11-25 05:01:56,455][root][INFO] - rank=3; Multi set iteration finished: iteration per set: [9765, 9765] [2023-11-25 05:01:56,456][root][INFO] - Finished iterating, iteration=9765, shard=3 [2023-11-25 05:01:56,620][root][INFO] - Finished iterating, iteration=9765, shard=2 [2023-11-25 05:01:56,628][root][INFO] - Finished iterating, iteration=9765, shard=0 [2023-11-25 05:01:56,636][root][INFO] - Finished iterating, iteration=9765, shard=1 [2023-11-25 05:01:56,642][root][INFO] - Finished iterating, iteration=9765, shard=3 [2023-11-25 05:01:56,722][root][INFO] - rank=2; Multi set iteration finished after next: iteration per set: [0, 0] [2023-11-25 05:01:56,722][root][INFO] - Epoch finished on 2 [2023-11-25 05:01:56,742][root][INFO] - rank=0; Multi set iteration finished after next: iteration per set: [0, 0] [2023-11-25 05:01:56,742][root][INFO] - Epoch finished on 0 [2023-11-25 05:01:56,752][root][INFO] - rank=3; Multi set iteration finished after next: iteration per set: [0, 0] [2023-11-25 05:01:56,753][root][INFO] - Epoch finished on 3 [2023-11-25 05:01:56,754][root][INFO] - rank=1; Multi set iteration finished after next: iteration per set: [0, 0] [2023-11-25 05:01:56,754][root][INFO] - Epoch finished on 1 [2023-11-25 05:03:20,126][root][INFO] - Saved checkpoint at ./vdr_1 [2023-11-25 05:03:20,126][root][INFO] - Av Loss per epoch=13.030566 [2023-11-25 05:03:20,126][root][INFO] - epoch total (1) correct predictions=18019762 [2023-11-25 05:03:20,126][root][INFO] - epoch total (2) correct predictions=18719683 [2023-11-25 05:03:20,126][root][INFO] - Saved checkpoint at ./vdr_1 [2023-11-25 05:03:20,127][root][INFO] - Saved checkpoint at ./vdr_1 [2023-11-25 05:03:20,127][root][INFO] - Av Loss per epoch=13.030566 [2023-11-25 05:03:20,127][root][INFO] - epoch total (1) correct predictions=18019762 [2023-11-25 05:03:20,127][root][INFO] - epoch total (2) correct predictions=18719683 [2023-11-25 05:03:20,127][root][INFO] - Av Loss per epoch=13.030566 [2023-11-25 05:03:20,127][root][INFO] - epoch total (1) correct predictions=18019762 [2023-11-25 05:03:20,127][root][INFO] - epoch total (2) correct predictions=18719683 [2023-11-25 05:03:20,128][root][INFO] - Saved checkpoint at ./vdr_1 [2023-11-25 05:03:20,129][root][INFO] - Av Loss per epoch=13.030566 [2023-11-25 05:03:20,129][root][INFO] - epoch total (1) correct predictions=18019762 [2023-11-25 05:03:20,129][root][INFO] - epoch total (2) correct predictions=18719683 [2023-11-25 05:03:20,129][root][INFO] - ***** Epoch 2 ***** [2023-11-25 05:03:20,129][root][INFO] - ***** Epoch 2 ***** [2023-11-25 05:03:20,130][root][INFO] - ***** Epoch 2 ***** [2023-11-25 05:03:20,131][root][INFO] - rank=1; Iteration start [2023-11-25 05:03:20,131][root][INFO] - rank=1; Multi set iteration: iteration ptr per set: [0, 0] [2023-11-25 05:03:20,131][root][INFO] - rank=1; Multi set iteration: source 0, batches to be taken: 9765 [2023-11-25 05:03:20,132][root][INFO] - rank=1; Multi set iteration: source 1, batches to be taken: 9765 [2023-11-25 05:03:20,132][root][INFO] - rank=2; Iteration start [2023-11-25 05:03:20,132][root][INFO] - rank=2; Multi set iteration: iteration ptr per set: [0, 0] [2023-11-25 05:03:20,132][root][INFO] - rank=0; Iteration start [2023-11-25 05:03:20,132][root][INFO] - rank=2; Multi set iteration: source 0, batches to be taken: 9765 [2023-11-25 05:03:20,132][root][INFO] - rank=0; Multi set iteration: iteration ptr per set: [0, 0] [2023-11-25 05:03:20,132][root][INFO] - rank=2; Multi set iteration: source 1, batches to be taken: 9765 [2023-11-25 05:03:20,132][root][INFO] - rank=0; Multi set iteration: source 0, batches to be taken: 9765 [2023-11-25 05:03:20,132][root][INFO] - rank=0; Multi set iteration: source 1, batches to be taken: 9765 [2023-11-25 05:03:20,132][root][INFO] - ***** Epoch 2 ***** [2023-11-25 05:03:20,134][root][INFO] - rank=3; Iteration start [2023-11-25 05:03:20,134][root][INFO] - rank=3; Multi set iteration: iteration ptr per set: [0, 0] [2023-11-25 05:03:20,134][root][INFO] - rank=3; Multi set iteration: source 0, batches to be taken: 9765 [2023-11-25 05:03:20,134][root][INFO] - rank=3; Multi set iteration: source 1, batches to be taken: 9765 [2023-11-25 05:03:20,137][root][INFO] - rank=1; data_src_indices len=19530 [2023-11-25 05:03:20,137][root][INFO] - rank=2; data_src_indices len=19530 [2023-11-25 05:03:20,137][root][INFO] - rank=0; data_src_indices len=19530 [2023-11-25 05:03:20,139][root][INFO] - rank=3; data_src_indices len=19530 [2023-11-25 05:03:26,433][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.61 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 5.23 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: label, instant ( which was initially called'valiant'before they discovered the name was ### ### already in use ), he began to work with the pianist and arranger allen toussaint. ### ### [P_TEXT]: allen toussaint with allen orange in the early 1960s. in 1960, joe banashak, of minit ### ### records and later instant records, hired toussaint as an a & r man and record producer. he did ### ### freelance work for other labels, such as fury. toussaint played piano, wrote, arranged and produced ### ### a string of hits in the early and mid - 1960s for new orleans r & b artists such as ernie k - doe, ### ### chris kenner, irma thomas ( including " it's raining " ), art and aaron neville ### ### =============================================== v_q ================================================ ### ### ('instant', 0, 0) ('##aint', 1, 5) ('valiant', 2, 5190) ('allen', 3, 9) ('##uss', 4, 16) ### ### ('ט', 5, 26) ('arranger', 6, 112) ('label', 7, 31) ('.', 8, 303) ('benny', 9, 4446) ### ### ('pianist', 10, 166) ('absently', 11, 459) ('jazz', 12, 138) ('ellington', 13, 6111) ### ### ('allmusic', 14, 716) ('1960s', 15, 35) ('rca', 16, 9115) ('beatles', 17, 57) ('intently', 18, 242) ### ### ('blues', 19, 1062) ### ### =============================================== v_p ================================================ ### ### ('instant', 0, 0) ('mini', 4157, 1) ('fury', 2693, 2) ('neville', 4480, 3) ('raining', 28108, 4) ### ### ('##aint', 1, 5) ('doe', 22503, 6) ('##ash', 23196, 7) ('ernie', 2181, 8) ('allen', 3, 9) ### ### ('##ner', 15458, 10) ('orange', 3338, 11) ('aaron', 3070, 12) ('ban', 19456, 13) ('ir', 9727, 14) ### ### ('##ak', 7672, 15) ('##uss', 4, 16) ('orleans', 1568, 17) ('thomas', 1431, 18) ('piano', 55, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('instant', 0, 0) ('##aint', 1, 5) ('allen', 3, 9) ('##uss', 4, 16) ('ט', 5, 26) ('label', 7, 31) ### ### ('labels', 52, 20) ('1960s', 15, 35) ('piano', 55, 19) ('records', 30, 30) ('1960', 157, 23) ### ### ('producer', 62, 32) ('arranger', 6, 112) ('beatles', 17, 57) ('artists', 46, 37) ('ken', 531, 21) ### ### ('fury', 2693, 2) ('&', 317, 29) ('producers', 53, 42) ('instantly', 29, 53) ### ############################################################################################################ [2023-11-25 05:03:26,934][root][INFO] - Epoch: 2: Step: 1/19530, loss[v]=0.252988, lr=0.000020, acc@1[1]=953.0/1024=0.9306640625, acc@1[2]=994.0/1024=0.970703125 [2023-11-25 05:05:22,621][root][INFO] - Train batch 100 [2023-11-25 05:05:22,622][root][INFO] - Avg. loss per last 100 batches: 0.200315 [2023-11-25 05:05:23,338][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.58 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.8/29522=99.99% | mean: 0.03 | max: 5.04 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: jennings has since followed with six more full - length studio albums : " electric rodeo ### ### " ( 2006 ), " the wolf " ( 2007 ), " black ribbons " ( 2009 ), " family man " ( 2012 ), " the other ### ### life " ( 2013 ) and " countach ( for giorgio ) " ( 2016 ), in addition to a live album, a ### ### compilation, and numerous eps. ### ### [P_TEXT]: stephen king king collaborated with musician shooter jennings and his band hierophant, ### ### providing the narration for their album, " black ribbons ". king played guitar for the rock band ### ### rock bottom remainders, several of whose members are authors. other members include dave barry, ### ### ridley pearson, scott turow, amy tan, james mcbride, mitch albom, roy blount, jr., matt groening, ### ### kathi kamen goldmark, sam barry, and greg iles. king and the other band members collaborated to ### ### release an e - book called " hard listening : the greatest ### ### =============================================== v_q ================================================ ### ### ('jennings', 0, 18) ('ribbons', 1, 0) ('rodeo', 2, 8538) ('wolf', 3, 12886) ('electric', 4, 3929) ### ### ('##ach', 5, 27795) ('life', 6, 11354) ('count', 7, 21037) ('else', 8, 890) ('giorgio', 9, 24233) ### ### ('family', 10, 15514) ('album', 11, 46) ('ribbon', 12, 45) ('man', 13, 8381) ('black', 14, 30) ### ### ('2013', 15, 7291) ('other', 16, 24632) ('.', 17, 526) ('absently', 18, 364) ### ### ('compilation', 19, 1386) ### ### =============================================== v_p ================================================ ### ### ('ribbons', 1, 0) ('stephen', 2953, 1) ('ile', 18560, 2) ('mcbride', 17380, 3) ### ### ('remainder', 4647, 4) ('##mark', 14639, 5) ('barry', 7431, 6) ('##row', 24524, 7) ### ### ('bottom', 2268, 8) ('king', 2839, 9) ('mitch', 5979, 10) ('ridley', 19041, 11) ('##ero', 7816, 12) ### ### ('tan', 8370, 13) ('authors', 2791, 14) ('sam', 3856, 15) ('pearson', 12114, 16) ('dave', 146, 17) ### ### ('jennings', 0, 18) ('##oe', 13923, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ribbons', 1, 0) ('jennings', 0, 18) ('black', 14, 30) ('album', 11, 46) ('ribbon', 12, 45) ### ### ('bands', 23, 41) ('band', 25, 42) ('dave', 146, 17) ('matt', 121, 32) ('guitar', 55, 47) ### ### ('artists', 22, 84) ('rapper', 33, 59) ('frontman', 40, 55) ('amy', 577, 22) ('dex', 36, 64) ### ### ('stephen', 2953, 1) ('greg', 667, 31) ('albums', 31, 87) ('bottom', 2268, 8) ('else', 8, 890) ### ############################################################################################################ [2023-11-25 05:05:23,843][root][INFO] - Epoch: 2: Step: 101/19530, loss[v]=0.351199, lr=0.000020, acc@1[1]=955.5/1024=0.93310546875, acc@1[2]=979.5/1024=0.95654296875 [2023-11-25 05:07:12,994][root][INFO] - Train batch 200 [2023-11-25 05:07:12,995][root][INFO] - Avg. loss per last 100 batches: 0.194814 [2023-11-25 05:07:13,699][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.75 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.5/29522=99.99% | mean: 0.03 | max: 5.32 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: is the third installment in sunrise's long running " gundam " franchise and the last tv ### ### series in the franchise released in japan's showa period. ### ### [P_TEXT]: kunihisa sugishima sugishima started his career on sunrise in the 1980s, having worked as ### ### the setting producer for " heavy metal l - gaim " s episodes 1 to 18, as well as the director and ### ### storyboarder for other episodes. another major work in his early career was his involvement with ### ### the " gundam " series ; he directed several episodes of " mobile suit zeta gundam " and " mobile ### ### suit gundam zz ". his first work as main director was in 1994 when he directed at sunrise a single ### ### - episode special titled " shinizokonai ### ### =============================================== v_q ================================================ ### ### ('gundam', 0, 0) ('sunrise', 1, 7) ('period', 2, 4670) ('##a', 3, 28067) ('show', 4, 71) ### ### ('absently', 5, 198) ('.', 6, 283) ('series', 7, 28) ('anime', 8, 150) ('″', 9, 24) ### ### ('franchise', 10, 78) ('japan', 11, 239) ('tv', 12, 602) ('intently', 13, 112) ('・', 14, 123) ### ### ('voiced', 15, 10325) ('jedi', 16, 19863) ('"', 17, 34) ('installment', 18, 3802) ### ### ('japanese', 19, 594) ### ### =============================================== v_p ================================================ ### ### ('gundam', 0, 0) ('##board', 3714, 1) ('##isa', 18757, 2) ('##im', 23055, 3) ('##nai', 25258, 4) ### ### ('##hima', 7686, 5) ('zeta', 4420, 6) ('sunrise', 1, 7) ('##ih', 16944, 8) ('kun', 10854, 9) ### ### ('##gis', 14099, 10) ('l', 7858, 11) ('metal', 328, 12) ('##boards', 18764, 13) ### ### ('##oko', 25853, 14) ('1994', 1090, 15) ('mobile', 3327, 16) ('z', 8117, 17) ('director', 7491, 18) ### ### ('episode', 98, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('gundam', 0, 0) ('sunrise', 1, 7) ('series', 7, 28) ('″', 9, 24) ('"', 17, 34) ('show', 4, 71) ### ### ('episode', 98, 19) ('metal', 328, 12) ('franchise', 10, 78) ('##子', 36, 40) ('absently', 5, 198) ### ### ('anime', 8, 150) ('anyway', 40, 43) ('.', 6, 283) ('intently', 13, 112) ('episodes', 699, 20) ### ### ('dramas', 32, 52) ('setting', 593, 21) ('・', 14, 123) ('story', 290, 36) ### ############################################################################################################ [2023-11-25 05:07:14,202][root][INFO] - Epoch: 2: Step: 201/19530, loss[v]=0.258557, lr=0.000020, acc@1[1]=961.5/1024=0.93896484375, acc@1[2]=991.5/1024=0.96826171875 [2023-11-25 05:09:03,725][root][INFO] - Train batch 300 [2023-11-25 05:09:03,726][root][INFO] - Avg. loss per last 100 batches: 0.192030 [2023-11-25 05:09:04,442][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.49 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.2/29522=99.99% | mean: 0.03 | max: 5.11 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: according to cisco's product literature, ios xr shares very little infrastructure with ### ### the other ios trains, and is instead built upon a " preemptive, memory protected, multitasking, ### ### microkernel - based operating system ". ### ### [P_TEXT]: cisco ios microkernel ( qnx ), so a large part of the ios source code was re - written to ### ### take advantage of the features offered by the kernel. in 2005 cisco introduced the cisco ios xr ### ### network operating system on the 12000 series of network routers, extending the microkernel ### ### architecture from the crs - 1 routers to cisco's widely deployed core router s. in 2006 cisco ### ### introduced ios software modularity, which extends the microkernel architecture into the ios ### ### environment, while still providing the software upgrade capabilities. ### ### =============================================== v_q ================================================ ### ### ('##ker', 0, 7) ('x', 1, 15) ('ios', 2, 1) ('train', 3, 3547) ('cisco', 4, 0) ('trains', 5, 1865) ### ### ('##tas', 6, 22092) ('##king', 7, 3265) ('##nel', 8, 6) ('##r', 9, 27) ('protected', 10, 8333) ### ### ('micro', 11, 19) ('unicode', 12, 686) ('operating', 13, 36) ('absently', 14, 268) ### ### ('iphone', 15, 127) ('intently', 16, 153) ('connectivity', 17, 820) ('system', 18, 48) ### ### ('programming', 19, 304) ### ### =============================================== v_p ================================================ ### ### ('cisco', 4, 0) ('ios', 2, 1) ('##nx', 22276, 2) ('modular', 1856, 3) ('core', 39, 4) ### ### ('1200', 2039, 5) ('##nel', 8, 6) ('##ker', 0, 7) ('kernel', 869, 8) ('route', 9103, 9) ### ### ('cr', 3503, 10) ('software', 24, 11) ('network', 58, 12) ('architecture', 160, 13) ### ### ('routing', 4974, 14) ('x', 1, 15) ('routes', 20687, 16) ('##rs', 324, 17) ('q', 1535, 18) ### ### ('micro', 11, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('cisco', 4, 0) ('ios', 2, 1) ('##ker', 0, 7) ('x', 1, 15) ('##nel', 8, 6) ('micro', 11, 19) ### ### ('##r', 9, 27) ('core', 39, 4) ('software', 24, 11) ('network', 58, 12) ('operating', 13, 36) ### ### ('architecture', 160, 13) ('modular', 1856, 3) ('##rs', 324, 17) ('system', 18, 48) ### ### ('kernel', 869, 8) ('upgrade', 125, 24) ('##র', 27, 54) ('1200', 2039, 5) ('code', 80, 34) ### ############################################################################################################ [2023-11-25 05:09:04,947][root][INFO] - Epoch: 2: Step: 301/19530, loss[v]=0.241476, lr=0.000020, acc@1[1]=957.5/1024=0.93505859375, acc@1[2]=990.0/1024=0.966796875 [2023-11-25 05:10:54,593][root][INFO] - Train batch 400 [2023-11-25 05:10:54,593][root][INFO] - Avg. loss per last 100 batches: 0.180662 [2023-11-25 05:10:55,305][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.65 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.0/29522=99.99% | mean: 0.03 | max: 5.46 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the debut of the winter classic, an outdoor regular season nhl game held on new year's ### ### day ### ### [P_TEXT]: mirabito outdoor classic goal in ahl history. the game was played in front of a then ahl ### ### record crowd of 21, 508, on a rink constructed at the new york state fairgrounds on a dirt auto ### ### racing track. crunch owner, howard dolgon, pursued the ideal of having his team in an outdoor game, ### ### but unlike the national hockey league's ( nhl ) winter classic the crunch were responsible for ### ### funding the entire event. in order to afford the cost of hosting an outdoor game sponsorship deals ### ### with several independent organizations were made including mirabito energy products, which ### ### =============================================== v_q ================================================ ### ### ('classic', 0, 8) ('winter', 1, 12) ('outdoor', 2, 1) ('nhl', 3, 13) ('intently', 4, 304) ### ### ('absently', 5, 418) ('season', 6, 83) ('.', 7, 929) ('hockey', 8, 18) ('toronto', 9, 245) ### ### ('game', 10, 35) ('new', 11, 264) ('arena', 12, 412) ('goaltender', 13, 50) ('year', 14, 14091) ### ### ('moose', 15, 920) ('ice', 16, 231) ('canada', 17, 149) ('newly', 18, 138) ('freaking', 19, 351) ### ### =============================================== v_p ================================================ ### ### ('crunch', 294, 0) ('outdoor', 2, 1) ('##bit', 23294, 2) ('mira', 13990, 3) ('##gon', 5844, 4) ### ### ('##grounds', 3307, 5) ('dirt', 11480, 6) ('howard', 5064, 7) ('classic', 0, 8) ('energy', 2700, 9) ### ### ('ahl', 2854, 10) ('sponsors', 23995, 11) ('winter', 1, 12) ('nhl', 3, 13) ('fair', 3137, 14) ### ### ('products', 7035, 15) ('auto', 8958, 16) ('do', 16390, 17) ('hockey', 8, 18) ('leaned', 210, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('outdoor', 2, 1) ('classic', 0, 8) ('winter', 1, 12) ('nhl', 3, 13) ('hockey', 8, 18) ### ### ('game', 10, 35) ('crunch', 294, 0) ('season', 6, 83) ('outdoors', 46, 33) ('goaltender', 13, 50) ### ### ('leaned', 210, 19) ('games', 42, 47) ('rink', 489, 20) ('event', 36, 61) ('intently', 4, 304) ### ### ('##grounds', 3307, 5) ('sponsorship', 592, 29) ('energy', 2700, 9) ('absently', 5, 418) ### ### ('anyway', 93, 49) ### ############################################################################################################ [2023-11-25 05:10:55,810][root][INFO] - Epoch: 2: Step: 401/19530, loss[v]=0.260836, lr=0.000020, acc@1[1]=963.5/1024=0.94091796875, acc@1[2]=996.5/1024=0.97314453125 [2023-11-25 05:12:45,049][root][INFO] - Train batch 500 [2023-11-25 05:12:45,049][root][INFO] - Avg. loss per last 100 batches: 0.203480 [2023-11-25 05:12:45,770][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.56 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.6/29522=99.98% | mean: 0.02 | max: 5.15 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.8/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: calvert served as leader of the saskatchewan new democratic party from 2001 to june 6, ### ### 2009, when he was succeeded by dwain lingenfelter. ### ### [P_TEXT]: saskatchewan new democratic party of canada mp. romanow and his government formed a ### ### coalition government with the three elected liberal mlas ; one, jack hillson, subsequently left ### ### cabinet to sit as an independent liberal in opposition. jim melenchuk and ron osika remained in the ### ### coalition and ran under the ndp banner in the 2003 provincial election, where both were defeated. ### ### romanow retired in 2001 and was succeeded by lorne calvert, who led the party into the 2003 general ### ### election. in an upset, the ndp not only retained power, but was able to ### ### =============================================== v_q ================================================ ### ### ('saskatchewan', 0, 0) ('calvert', 1, 2) ('##wai', 2, 28411) ('democratic', 3, 18) ### ### ('ling', 4, 10611) ('new', 5, 28) ('leader', 6, 1307) ('newly', 7, 35) ('party', 8, 29) ### ### ('canada', 9, 24) ('d', 10, 12390) ('election', 11, 17) ('proponents', 12, 268) ### ### ('canadian', 13, 63) ('saskatoon', 14, 149) ('angrily', 15, 120) ('leadership', 16, 711) ### ### ('ד', 17, 206) ('2009', 18, 2663) ('hillary', 19, 23000) ### ### =============================================== v_p ================================================ ### ### ('saskatchewan', 0, 0) ('romano', 15514, 1) ('calvert', 1, 2) ('##chuk', 14878, 3) ### ### ('##ika', 25944, 4) ('##rne', 17643, 5) ('hills', 11105, 6) ('ndp', 30, 7) ('2003', 877, 8) ### ### ('ron', 890, 9) ('banner', 4633, 10) ('jack', 3025, 11) ('jim', 97, 12) ('os', 21614, 13) ### ### ('2001', 23, 14) ('liberals', 186, 15) ('upset', 718, 16) ('election', 11, 17) ### ### ('democratic', 3, 18) ('##w', 9676, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('saskatchewan', 0, 0) ('calvert', 1, 2) ('democratic', 3, 18) ('new', 5, 28) ('party', 8, 29) ### ### ('canada', 9, 24) ('election', 11, 17) ('newly', 7, 35) ('ndp', 30, 7) ('2001', 23, 14) ### ### ('liberal', 36, 23) ('jim', 97, 12) ('mla', 60, 20) ('cabinet', 59, 22) ('liberals', 186, 15) ### ### ('provincial', 122, 26) ('succeeded', 161, 21) ('canadian', 13, 63) ('2003', 877, 8) ### ### ('ron', 890, 9) ### ############################################################################################################ [2023-11-25 05:12:46,273][root][INFO] - Epoch: 2: Step: 501/19530, loss[v]=0.148376, lr=0.000020, acc@1[1]=980.5/1024=0.95751953125, acc@1[2]=1007.5/1024=0.98388671875 [2023-11-25 05:14:35,819][root][INFO] - Train batch 600 [2023-11-25 05:14:35,820][root][INFO] - Avg. loss per last 100 batches: 0.206872 [2023-11-25 05:14:36,543][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.73 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.8/29522=99.99% | mean: 0.03 | max: 5.32 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the 2009 volta a catalunya was the 89th edition of the " volta ". ### ### [P_TEXT]: alejandro valverde alejandro valverde belmonte ( born 25 april 1980 ) is a spanish road ### ### racing cyclist, who currently rides for uci worldteam. valverde's biggest wins have been the vuelta ### ### a espana in 2009, criterium du dauphine in 2008 and 2009, tour of the basque country in 2017, volta ### ### a catalunya in 2009, 2017 and 2018, liege – bastogne – liege in 2006, 2008, 2015 ### ### =============================================== v_q ================================================ ### ### ('volta', 0, 6) ('2009', 1, 16) ('##ya', 2, 15) ('cat', 3, 10) ('edition', 4, 3308) ('89', 5, 1909) ### ### ('##al', 6, 30) ('intently', 7, 235) ('cycling', 8, 76) ('spain', 9, 33) ('absently', 10, 385) ### ### ('commune', 11, 22403) ('rally', 12, 308) ('##un', 13, 41) ('catalonia', 14, 1645) ### ### ('spanish', 15, 25) ('barcelona', 16, 439) ('cyclist', 17, 58) ('³', 18, 212) ('race', 19, 31) ### ### =============================================== v_p ================================================ ### ### ('basque', 23, 0) ('liege', 1006, 1) ('##rde', 4704, 2) ('belmont', 14146, 3) ('valve', 1550, 4) ### ### ('dauphin', 24147, 5) ('volta', 0, 6) ('##iter', 25077, 7) ('alejandro', 9675, 8) ('uci', 57, 9) ### ### ('cat', 3, 10) ('espana', 536, 11) ('vuelta', 34, 12) ('##gne', 4245, 13) ('2017', 191, 14) ### ### ('##ya', 2, 15) ('2009', 1, 16) ('cyclists', 3961, 17) ('2018', 1616, 18) ('tour', 80, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('volta', 0, 6) ('2009', 1, 16) ('cat', 3, 10) ('##ya', 2, 15) ('basque', 23, 0) ('##al', 6, 30) ### ### ('spanish', 15, 25) ('vuelta', 34, 12) ('spain', 9, 33) ('uci', 57, 9) ('2008', 27, 21) ### ### ('race', 19, 31) ('##un', 13, 41) ('racing', 50, 26) ('tour', 80, 19) ('2017', 191, 14) ### ### ('liege', 1006, 1) ('espana', 536, 11) ('##я', 22, 50) ('cycling', 8, 76) ### ############################################################################################################ [2023-11-25 05:14:37,048][root][INFO] - Epoch: 2: Step: 601/19530, loss[v]=0.208396, lr=0.000020, acc@1[1]=971.0/1024=0.9482421875, acc@1[2]=995.5/1024=0.97216796875 [2023-11-25 05:16:26,568][root][INFO] - Train batch 700 [2023-11-25 05:16:26,568][root][INFO] - Avg. loss per last 100 batches: 0.190192 [2023-11-25 05:16:27,275][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.62 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.2/29522=99.99% | mean: 0.02 | max: 5.23 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he was also a member of the supreme council of the armed forces which governed egypt de ### ### facto after the egyptian revolution of 2011. ### ### [P_TEXT]: supreme council of the armed forces the following day after mubarak's resignation. the ### ### military junta was headed by field marshal mohamed hussein tantawi who served as the minister of ### ### defense under mubarak, and included the service heads and other senior commanders of the egyptian ### ### armed forces, namely lt. gen. sami hafez anan, armed forces chief of staff ; air marshal reda ### ### mahmoud hafez mohamed, air force commander ; lt. gen. abd el aziz seif - eldeen, commander of air ### ### defense ; and vice admiral mohab mamish, navy commander in chief. the scaf ### ### =============================================== v_q ================================================ ### ### ('armed', 0, 11) ('egypt', 1, 32) ('egyptian', 2, 5) ('supreme', 3, 8) ('2011', 4, 16057) ### ### ('revolution', 5, 1257) ('council', 6, 6) ('.', 7, 169) ('absently', 8, 180) ('mohamed', 9, 21) ### ### ('facto', 10, 19214) ('forces', 11, 28) ('military', 12, 34) ('intently', 13, 264) ### ### ('ahmed', 14, 134) ('littered', 15, 142) ('！', 16, 225) ('icao', 17, 175) ('³', 18, 167) ### ### ('egyptians', 19, 220) ### ### =============================================== v_p ================================================ ### ### ('##bara', 19301, 0) ('##mis', 13705, 1) ('junta', 302, 2) ('defense', 2915, 3) ('air', 857, 4) ### ### ('egyptian', 2, 5) ('council', 6, 6) ('marshal', 2573, 7) ('supreme', 3, 8) ('lt', 2068, 9) ### ### ('tan', 1229, 10) ('armed', 0, 11) ('heads', 2484, 12) ('chief', 2389, 13) ('hussein', 90, 14) ### ### ('sc', 13413, 15) ('##af', 8834, 16) ('gen', 5008, 17) ('##dee', 19068, 18) ('admiral', 2242, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('egyptian', 2, 5) ('supreme', 3, 8) ('armed', 0, 11) ('council', 6, 6) ('egypt', 1, 32) ### ### ('mohamed', 9, 21) ('forces', 11, 28) ('military', 12, 34) ('junta', 302, 2) ('hussein', 90, 14) ### ### ('cabinet', 44, 35) ('minister', 97, 29) ('force', 45, 47) ('air', 857, 4) ('staff', 245, 30) ### ### ('navy', 393, 23) ('mahmoud', 310, 27) ('abd', 92, 43) ('tan', 1229, 10) ('.', 7, 169) ### ############################################################################################################ [2023-11-25 05:16:27,778][root][INFO] - Epoch: 2: Step: 701/19530, loss[v]=0.116563, lr=0.000020, acc@1[1]=996.5/1024=0.97314453125, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 05:18:17,073][root][INFO] - Train batch 800 [2023-11-25 05:18:17,073][root][INFO] - Avg. loss per last 100 batches: 0.191321 [2023-11-25 05:18:17,788][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.73 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.1/29522=99.99% | mean: 0.02 | max: 5.26 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: she and gail edwards, paul kreppel and marian mercer were the only actors from " it's a ### ### living " to last all the way through the show's network and ### ### [P_TEXT]: gail edwards days ", " lou grant ", " m * a * s * h " and " taxi ". in 1979, edwards ### ### auditioned for a witt / thomas production, in which she landed the role of dot higgins on abc's " ### ### it ’ s a living ". the series ran on abc from 1980 to 1982 and was revived in first - run ### ### syndication from 1985 to 1989. edwards, along with barrie youngfellow, paul kreppel, and the late ### ### marian mercer, were the only four members of the " it's a living " cast who lasted during the ### ### network ### ### =============================================== v_q ================================================ ### ### ('living', 0, 2) ('gail', 1, 1) ('marian', 2, 7) ('mercer', 3, 12) ('edwards', 4, 3) ### ### ('##ppel', 5, 10) ('″', 6, 17) ('paul', 7, 21) ('##re', 8, 32) ('"', 9, 35) ('k', 10, 38) ### ### ('absently', 11, 205) ('あ', 12, 50) ('network', 13, 51) ('it', 14, 39) ('films', 15, 305) ### ### ('intently', 16, 219) ('ק', 17, 42) ('##レ', 18, 172) ('hollywood', 19, 1713) ### ### =============================================== v_p ================================================ ### ### ('taxi', 16616, 0) ('gail', 1, 1) ('living', 0, 2) ('edwards', 4, 3) ('higgins', 5430, 4) ### ### ('wit', 17471, 5) ('##fell', 13588, 6) ('marian', 2, 7) ('barrie', 8472, 8) ('dot', 2345, 9) ### ### ('##ppel', 5, 10) ('grant', 4548, 11) ('mercer', 3, 12) ('thomas', 6083, 13) ('*', 3166, 14) ### ### ('lou', 6242, 15) ('syndication', 648, 16) ('″', 6, 17) ('days', 4137, 18) ('abc', 113, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('gail', 1, 1) ('living', 0, 2) ('edwards', 4, 3) ('marian', 2, 7) ('mercer', 3, 12) ### ### ('##ppel', 5, 10) ('″', 6, 17) ('paul', 7, 21) ('##re', 8, 32) ('"', 9, 35) ('k', 10, 38) ### ### ('it', 14, 39) ('あ', 12, 50) ('network', 13, 51) ('abc', 113, 19) ('ק', 17, 42) ('cast', 63, 23) ### ### ('series', 60, 25) ('syndication', 648, 16) ('dot', 2345, 9) ### ############################################################################################################ [2023-11-25 05:18:18,292][root][INFO] - Epoch: 2: Step: 801/19530, loss[v]=0.146438, lr=0.000020, acc@1[1]=989.0/1024=0.9658203125, acc@1[2]=1012.5/1024=0.98876953125 [2023-11-25 05:20:07,911][root][INFO] - Train batch 900 [2023-11-25 05:20:07,912][root][INFO] - Avg. loss per last 100 batches: 0.197641 [2023-11-25 05:20:08,646][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.72 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.4/29522=99.99% | mean: 0.03 | max: 5.55 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.8/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: rossi held off ducati's andrea dovizioso to cross the line ### ### [P_TEXT]: andrea dovizioso would be joining ducati for 2017. a few weeks later ducati announced ### ### that dovizioso was to stay at ducati to partner lorenzo, while iannone signed a contract with the ### ### suzuki team. in the inaugural austrian grand prix, where ducati were favourites to win, he finished ### ### 2nd to his teammate iannone, which left disappointed. however he finished the season strongly and ### ### took only his 2nd motogp win in the malaysian grand prix ahead of valentino rossi and jorge ### ### lorenzo, his first win in 7 years. he finished the season fifth in the championship. once again, ### ### dovizioso started the season ### ### =============================================== v_q ================================================ ### ### ('rossi', 0, 5) ('##zio', 1, 8) ('##ati', 2, 17) ('andrea', 3, 0) ('##so', 4, 12) ### ### ('absently', 5, 457) ('##vi', 6, 14) ('.', 7, 142) ('prix', 8, 18) ('intently', 9, 221) ### ### ('do', 10, 21) ('duc', 11, 13) ('race', 12, 15) ('motorcycle', 13, 43) ('start', 14, 78) ### ### ('rally', 15, 207) ('italian', 16, 266) ('finished', 17, 29) ('ferrari', 18, 46) ('car', 19, 139) ### ### =============================================== v_p ================================================ ### ### ('andrea', 3, 0) ('suzuki', 5053, 1) ('lorenzo', 673, 2) ('jorge', 2681, 3) ('##non', 23422, 4) ### ### ('rossi', 0, 5) ('austria', 1573, 6) ('malaysian', 20657, 7) ('##zio', 1, 8) ('ian', 6018, 9) ### ### ('malaysia', 13270, 10) ('austrian', 1896, 11) ('##so', 4, 12) ('duc', 11, 13) ('##vi', 6, 14) ### ### ('race', 12, 15) ('2017', 620, 16) ('##ati', 2, 17) ('prix', 8, 18) ('valentin', 3301, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('rossi', 0, 5) ('andrea', 3, 0) ('##zio', 1, 8) ('##so', 4, 12) ('##ati', 2, 17) ('##vi', 6, 14) ### ### ('duc', 11, 13) ('prix', 8, 18) ('race', 12, 15) ('do', 10, 21) ('finished', 17, 29) ### ### ('finish', 27, 25) ('motorcycle', 13, 43) ('season', 23, 34) ('team', 59, 22) ('lorenzo', 673, 2) ### ### ('racing', 37, 32) ('ferrari', 18, 46) ('.', 7, 142) ('crash', 32, 40) ### ############################################################################################################ [2023-11-25 05:20:09,151][root][INFO] - Epoch: 2: Step: 901/19530, loss[v]=0.093201, lr=0.000020, acc@1[1]=991.0/1024=0.9677734375, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 05:21:58,719][root][INFO] - Train batch 1000 [2023-11-25 05:21:58,720][root][INFO] - Avg. loss per last 100 batches: 0.193944 [2023-11-25 05:21:59,426][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.60 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.0/29522=99.99% | mean: 0.03 | max: 5.34 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: on april 29, 2010, at the time of earnhardt's induction to the nascar hall of fame, dei ### ### announced a partnership with jr motorsports, richard childress racing, and wrangler jeans where, ### ### for one race, as a tribute to his father, dale earnhardt jr. would drive the no. ### ### [P_TEXT]: richard childress racing ended up winning his second championship and was the first ### ### championship that was won back - to - back while driving for different teams ( 2018 was in jr ### ### motorsports ). on december 10, 2020, it was announced that myatt snider will drive the no. 2 car ### ### full - time. the no. 3 car was initially run in the then - busch series by dale earnhardt, inc. ### ### beginning in 1989, winning two back - to - back championships with dale earnhardt, jr. in 1998 and ### ### 1999. the no. 3 made its debut as part of the rcr stable in 2002 at the eas / gnc live well 300 at ### ### daytona ### ### =============================================== v_q ================================================ ### ### ('dei', 0, 22728) ('jeans', 1, 3486) ('jr', 2, 4) ('dale', 3, 19) ('nascar', 4, 25) ### ### ('motorsports', 5, 7) ('##hardt', 6, 15) ('earn', 7, 11) ('race', 8, 35) ('absently', 9, 366) ### ### ('richard', 10, 6) ('fame', 11, 1821) ('##ang', 12, 23745) ('child', 13, 8) ('racing', 14, 31) ### ### ('2010', 15, 5272) ('wr', 16, 17135) ('.', 17, 211) ('antics', 18, 427) ('##ler', 19, 27231) ### ### =============================================== v_p ================================================ ### ### ('ea', 11698, 0) ('##att', 26994, 1) ('rc', 9696, 2) ('stable', 51, 3) ('jr', 2, 4) ### ### ('##nc', 25755, 5) ('richard', 10, 6) ('motorsports', 5, 7) ('child', 13, 8) ('##ress', 236, 9) ### ### ('live', 13158, 10) ('earn', 7, 11) ('300', 654, 12) ('##ni', 19806, 13) ('daytona', 333, 14) ### ### ('##hardt', 6, 15) ('3', 6209, 16) ('##der', 13898, 17) ('2002', 1373, 18) ('dale', 3, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('jr', 2, 4) ('motorsports', 5, 7) ('richard', 10, 6) ('dale', 3, 19) ('earn', 7, 11) ### ### ('child', 13, 8) ('##hardt', 6, 15) ('nascar', 4, 25) ('race', 8, 35) ('racing', 14, 31) ### ### ('stable', 51, 3) ('car', 28, 29) ('cars', 63, 27) ('##ress', 236, 9) ('sr', 27, 44) ### ### ('daytona', 333, 14) ('drivers', 24, 59) ('races', 80, 37) ('drive', 33, 51) ('300', 654, 12) ### ############################################################################################################ [2023-11-25 05:21:59,929][root][INFO] - Epoch: 2: Step: 1001/19530, loss[v]=0.108585, lr=0.000020, acc@1[1]=994.0/1024=0.970703125, acc@1[2]=1009.5/1024=0.98583984375 [2023-11-25 05:23:49,258][root][INFO] - Train batch 1100 [2023-11-25 05:23:49,259][root][INFO] - Avg. loss per last 100 batches: 0.196994 [2023-11-25 05:23:49,987][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.54 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.6/29522=99.99% | mean: 0.03 | max: 5.23 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 1463 he collaborated with francesco di antonio del chierico on a choir - book for the ### ### cathedral of florence, now at the laurentian library ( nos. ### ### [P_TEXT]: francesco di antonio del chierico 1457 when he illustrated the " exorcism of st zeno " in ### ### choir - books for pistoia cathedral commissioned by bishop donato medici. he continued his ### ### employment by the medici family when piero de ’ medici commissioned him in 1458 to illustrate pliny ### ### ’ s natural history. he later illuminated cosimo il vecchio's choir - books of the badia of fiesole ### ### in 1461. in 1463, in collaboration with zanobi strozzi, he helped illuminate a choir - book for ### ### cathedral of florence, now in the laurentian library ( nos. 149, 150, 151 ). strozzi was eleven ### ### years older ### ### =============================================== v_q ================================================ ### ### ('##eric', 0, 1) ('florence', 1, 21) ('laurent', 2, 0) ('cathedral', 3, 2) ('antonio', 4, 32) ### ### ('book', 5, 39) ('chi', 6, 78) ('francesco', 7, 14) ('choir', 8, 20) ('146', 9, 11) ('del', 10, 61) ### ### ('sparhawk', 11, 158) ('library', 12, 18) ('garion', 13, 209) ('leonardo', 14, 3493) ### ### ('pupils', 15, 1318) ('giovanni', 16, 2141) ('di', 17, 82) ('147', 18, 57) ('augustine', 19, 870) ### ### =============================================== v_p ================================================ ### ### ('laurent', 2, 0) ('##eric', 0, 1) ('cathedral', 3, 2) ('zen', 2178, 3) ('medici', 219, 4) ### ### ('dona', 3688, 5) ('##cis', 21343, 6) ('##ole', 19048, 7) ('##iny', 6486, 8) ('fi', 10869, 9) ### ### ('pier', 2868, 10) ('146', 9, 11) ('bad', 13762, 12) ('##sto', 9425, 13) ('francesco', 7, 14) ### ### ('##cchi', 13287, 15) ('##nob', 9754, 16) ('##zzi', 4703, 17) ('library', 12, 18) ### ### ('illuminated', 1469, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##eric', 0, 1) ('laurent', 2, 0) ('cathedral', 3, 2) ('florence', 1, 21) ('146', 9, 11) ### ### ('francesco', 7, 14) ('choir', 8, 20) ('antonio', 4, 32) ('library', 12, 18) ('book', 5, 39) ### ### ('nos', 23, 28) ('del', 10, 61) ('##ian', 22, 41) ('medici', 219, 4) ('chi', 6, 78) ### ### ('books', 45, 35) ('147', 18, 57) ('145', 172, 24) ('bishop', 120, 30) ('150', 205, 23) ### ############################################################################################################ [2023-11-25 05:23:50,492][root][INFO] - Epoch: 2: Step: 1101/19530, loss[v]=0.133046, lr=0.000020, acc@1[1]=990.5/1024=0.96728515625, acc@1[2]=1009.5/1024=0.98583984375 [2023-11-25 05:25:40,259][root][INFO] - Train batch 1200 [2023-11-25 05:25:40,260][root][INFO] - Avg. loss per last 100 batches: 0.196154 [2023-11-25 05:25:40,992][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.71 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.8/29522=99.99% | mean: 0.03 | max: 5.27 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.3/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the incumbent home secretary, alan johnson, called grayling's use of crime statistics " ### ### dodgy " and said that the british crime survey clearly showed that violent crime had reduced by 41 ### ### % over the same period. ### ### [P_TEXT]: jonathan arnott 2005 he stood as a ukip candidate for the sheffield attercliffe ### ### constituency, coming fourth overall with 4. 5 % of the vote, losing his deposit. arnott was then ### ### selected to contest the sheffield south east constituency for the 2010 general election. just ### ### before the election, he took part in the second half of a radio 5 live debate on crime with alan ### ### johnson, chris grayling and chris huhne. he spoke four times, calling for a reduction in police ### ### bureaucracy, abolition of jail sentence remission and ### ### =============================================== v_q ================================================ ### ### ('alan', 0, 6) ('johnson', 1, 5) ('crime', 2, 12) ('statistics', 3, 12064) ('gray', 4, 9) ### ### ('##ling', 5, 19) ('survey', 6, 3112) ('violent', 7, 2486) ('dod', 8, 15282) ('secretary', 9, 2387) ### ### ('british', 10, 515) ('home', 11, 14535) ('crimes', 12, 45) ('absently', 13, 263) ### ### ('criticised', 14, 74) ('violence', 15, 304) ('surveys', 16, 6459) ('criminal', 17, 55) ### ### ('.', 18, 1349) ('britain', 19, 160) ### ### =============================================== v_p ================================================ ### ### ('huh', 15910, 0) ('##not', 24110, 1) ('sheffield', 205, 2) ('##mission', 13588, 3) ### ### ('##cliffe', 7101, 4) ('johnson', 1, 5) ('alan', 0, 6) ('jonathan', 2481, 7) ('deposit', 5851, 8) ### ### ('gray', 4, 9) ('constituencies', 4687, 10) ('chris', 6266, 11) ('crime', 2, 12) ### ### ('radio', 1950, 13) ('bureaucracy', 3298, 14) ('ar', 19084, 15) ('constituency', 113, 16) ### ### ('2010', 926, 17) ('jail', 2022, 18) ('##ling', 5, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('johnson', 1, 5) ('alan', 0, 6) ('crime', 2, 12) ('gray', 4, 9) ('##ling', 5, 19) ### ### ('police', 20, 20) ('uk', 23, 23) ('crimes', 12, 45) ('sheffield', 205, 2) ('election', 29, 32) ### ### ('reduction', 57, 24) ('constituency', 113, 16) ('criminal', 17, 55) ('criticised', 14, 74) ### ### ('prison', 70, 43) ('sentence', 355, 28) ('2010', 926, 17) ('reduce', 45, 52) ('worried', 414, 34) ### ### ('jonathan', 2481, 7) ### ############################################################################################################ [2023-11-25 05:25:41,496][root][INFO] - Epoch: 2: Step: 1201/19530, loss[v]=0.204584, lr=0.000020, acc@1[1]=976.0/1024=0.953125, acc@1[2]=998.5/1024=0.97509765625 [2023-11-25 05:27:31,082][root][INFO] - Train batch 1300 [2023-11-25 05:27:31,083][root][INFO] - Avg. loss per last 100 batches: 0.208410 [2023-11-25 05:27:31,791][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.63 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.3/29522=99.99% | mean: 0.02 | max: 5.17 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: posthumous releases " oaxaca " ( recorded 1970 – 71 ), " live on the air " ( recorded ### ### february 1974 ) and " north beach " feature both studio and live performances recorded during this ### ### period of transition. ### ### [P_TEXT]: oaxaca ( album ) oaxaca is a compilation album by american jazz pianist vince guaraldi ### ### released by d & d records ( guaraldi's briefly resurrected label ) in 2004. the album is a mix of ### ### previously unreleased studio and live recordings taped in 1970 and 1971. in the mid - 2000s, vince ### ### guaraldi's son, david guaraldi, worked with audio archivist michael graves at his atlanta, georgia ### ### - based osiris studio, in an effort to restore a wealth of unreleased recorded material from his ### ### father's archives. the recordings on " oaxaca " come from a variety of sources : some songs were ### ### performed live at in your ear, a jazz ### ### =============================================== v_q ================================================ ### ### ('oaxaca', 0, 0) ('beach', 1, 9555) ('air', 2, 7796) ('north', 3, 21379) ('live', 4, 74) ### ### ('1974', 5, 2811) ('mexico', 6, 56) ('album', 7, 21) ('1970', 8, 25) ('absently', 9, 148) ### ### ('71', 10, 391) ('reissue', 11, 4123) ('band', 12, 251) ('onto', 13, 1366) ('concert', 14, 215) ### ### ('cover', 15, 4351) ('transition', 16, 3176) ('mexican', 17, 84) ('jazz', 18, 16) ### ### ('recordings', 19, 37) ### ### =============================================== v_p ================================================ ### ### ('oaxaca', 0, 0) ('graves', 9499, 1) ('##iri', 27589, 2) ('vince', 2005, 3) ('audio', 905, 4) ### ### ('ear', 8231, 5) ('resurrected', 5747, 6) ('atlanta', 378, 7) ('michael', 449, 8) ### ### ('##ld', 28870, 9) ('your', 1675, 10) ('gu', 22261, 11) ('##ara', 25075, 12) ('d', 17873, 13) ### ### ('arch', 12543, 14) ('son', 7335, 15) ('jazz', 18, 16) ('david', 1063, 17) ('archives', 1229, 18) ### ### ('os', 19175, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('oaxaca', 0, 0) ('album', 7, 21) ('1970', 8, 25) ('jazz', 18, 16) ('live', 4, 74) ### ### ('mexico', 6, 56) ('records', 33, 23) ('recordings', 19, 37) ('1971', 46, 30) ### ### ('compilation', 79, 24) ('puebla', 47, 39) ('atlanta', 378, 7) ('michael', 449, 8) ### ### ('audio', 905, 4) ('pianist', 174, 22) ('″', 26, 61) ('studio', 158, 28) ('vince', 2005, 3) ### ### ('beatles', 23, 70) ('musicians', 25, 68) ### ############################################################################################################ [2023-11-25 05:27:32,293][root][INFO] - Epoch: 2: Step: 1301/19530, loss[v]=0.100002, lr=0.000020, acc@1[1]=993.0/1024=0.9697265625, acc@1[2]=1012.5/1024=0.98876953125 [2023-11-25 05:29:21,654][root][INFO] - Train batch 1400 [2023-11-25 05:29:21,655][root][INFO] - Avg. loss per last 100 batches: 0.193027 [2023-11-25 05:29:22,372][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.65 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.4/29522=99.99% | mean: 0.02 | max: 5.08 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: while performing shows with the second city in 1997, fey submitted several scripts to ### ### nbc's variety show " saturday night live " ( " snl " ), at the request of its head writer adam ### ### mckay, a former performer at second city. ### ### [P_TEXT]: saturday night live fallon, and tina fey proved promising. the show focused on ### ### performers, and writers were forced to supply material for the cast's existing characters before ### ### they could write original sketches. by 1997, ohlmeyer renewed his focus on limiting michaels's ### ### independence, forcing the removal of writer jim downey and cast member norm macdonald. the original ### ### 1975 cast of " snl ", titled " the not ready for prime - time players, " a term coined by writer ### ### herb sargent, included laraine newman, john belushi, jane curtin, gilda radner, dan aykroyd, ### ### garrett morris, and chevy chase. radner was the ### ### =============================================== v_q ================================================ ### ### ('fey', 0, 14) ('mckay', 1, 5654) ('second', 2, 6527) ('city', 3, 8208) ('adam', 4, 7119) ### ### ('saturday', 5, 2) ('##₂', 6, 552) ('2nd', 7, 3479) ('absently', 8, 228) ('cities', 9, 14853) ### ### ('show', 10, 43) ('live', 11, 15) ('shows', 12, 97) ('antics', 13, 109) ('.', 14, 78) ### ### ('performer', 15, 174) ('performers', 16, 115) ('！', 17, 445) ('1997', 18, 41) ### ### ('variety', 19, 4244) ### ### =============================================== v_p ================================================ ### ### ('sargent', 16844, 0) ('macdonald', 7152, 1) ('saturday', 5, 2) ('norm', 7790, 3) ### ### ('garrett', 1549, 4) ('newman', 2019, 5) ('lara', 6215, 6) ('herb', 16981, 7) ('curt', 1077, 8) ### ### ('michaels', 79, 9) ('##eyer', 24402, 10) ('##dner', 13565, 11) ('morris', 4353, 12) ### ### ('bel', 16750, 13) ('fey', 0, 14) ('live', 11, 15) ('##ush', 13781, 16) ('night', 27, 17) ### ### ('ready', 4808, 18) ('gil', 10752, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('fey', 0, 14) ('saturday', 5, 2) ('live', 11, 15) ('night', 27, 17) ('show', 10, 43) ### ### ('michaels', 79, 9) ('1997', 18, 41) ('##nl', 48, 29) ('chase', 258, 21) ('cast', 90, 34) ### ### ('curt', 1077, 8) ('garrett', 1549, 4) ('jim', 523, 25) ('convinces', 22, 61) ('jane', 767, 24) ### ### ('.', 14, 78) ('newman', 2019, 5) ('hissed', 29, 57) ('writers', 58, 47) ('shows', 12, 97) ### ############################################################################################################ [2023-11-25 05:29:22,878][root][INFO] - Epoch: 2: Step: 1401/19530, loss[v]=0.141004, lr=0.000020, acc@1[1]=993.0/1024=0.9697265625, acc@1[2]=1011.0/1024=0.9873046875 [2023-11-25 05:31:12,876][root][INFO] - Train batch 1500 [2023-11-25 05:31:12,877][root][INFO] - Avg. loss per last 100 batches: 0.196205 [2023-11-25 05:31:13,611][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.7/29522=99.99% | mean: 0.03 | max: 5.47 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: barrackpore police commissionerate or barrackpore city police, established on 20 january ### ### 2012, is a city police force with primary responsibilities in law enforcement and investigation ### ### within the area which roughly constitutes the barrackpore subdivision, and is located within the ### ### northern part of kolkata metropolitan area. ### ### [P_TEXT]: bandipur, north 24 parganas 96 % of the population of barrackpore subdivision ( partly ### ### presented in the map alongside, the places marked on the map are linked in full screen map ) lives ### ### in urban areas. in 2011, it had a density of population of 10, 967 per km2. the subdivision has 16 ### ### municipalities and 24 census towns. khardaha police station under barrackpore police ### ### commissionerate has jurisdiction over khardaha municipal area and barrackpore ii cd block. bandipur ### ### has a delivery branch post office, with pin 700119 in the north presidency division of north 24 ### ### parganas district in calcutta region. other ### ### =============================================== v_q ================================================ ### ### ('##pore', 0, 4) ('barr', 1, 9) ('police', 2, 12) ('kolkata', 3, 47) ('##ack', 4, 27) ### ### ('commissioner', 5, 17) ('absently', 6, 206) ('calcutta', 7, 0) ('city', 8, 362) ### ### ('metropolitan', 9, 1903) ('india', 10, 63) ('##ate', 11, 58) ('.', 12, 274) ### ### ('subdivision', 13, 87) ('enforcement', 14, 648) ('force', 15, 3179) ('arrested', 16, 1251) ### ### ('bangladesh', 17, 492) ('bengal', 18, 61) ('bengali', 19, 57) ### ### =============================================== v_p ================================================ ### ### ('calcutta', 7, 0) ('##hard', 28230, 1) ('delivery', 2010, 2) ('##ip', 11201, 3) ('##pore', 0, 4) ### ### ('presidency', 786, 5) ('pin', 4416, 6) ('##11', 12504, 7) ('band', 9972, 8) ('barr', 1, 9) ### ### ('##ur', 7734, 10) ('700', 11146, 11) ('police', 2, 12) ('branch', 547, 13) ### ### ('municipalities', 4963, 14) ('24', 8575, 15) ('##aha', 14953, 16) ('commissioner', 5, 17) ### ### ('screen', 3280, 18) ('north', 2785, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##pore', 0, 4) ('barr', 1, 9) ('police', 2, 12) ('calcutta', 7, 0) ('commissioner', 5, 17) ### ### ('##ack', 4, 27) ('kolkata', 3, 47) ('##ate', 11, 58) ('india', 10, 63) ('district', 30, 33) ### ### ('bengal', 18, 61) ('bengali', 19, 57) ('presidency', 786, 5) ('municipal', 264, 20) ### ### ('areas', 36, 46) ('subdivision', 13, 87) ('branch', 547, 13) ('block', 220, 25) ### ### ('delivery', 2010, 2) ('absently', 6, 206) ### ############################################################################################################ [2023-11-25 05:31:14,115][root][INFO] - Epoch: 2: Step: 1501/19530, loss[v]=0.341794, lr=0.000020, acc@1[1]=961.5/1024=0.93896484375, acc@1[2]=988.0/1024=0.96484375 [2023-11-25 05:33:03,781][root][INFO] - Train batch 1600 [2023-11-25 05:33:03,782][root][INFO] - Avg. loss per last 100 batches: 0.200278 [2023-11-25 05:33:04,487][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.78 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 5.59 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: kalisz ( is a city in central poland with 100, 246 inhabitants ( december 2019 ) making ### ### it the second - largest city in the greater poland voivodeship. ### ### [P_TEXT]: greater poland voivodeship of this historic region, except for some western parts. ### ### greater poland voivodeship is second in area and third in population among poland's sixteen ### ### voivodeships, with an area of and a population of close to 3. 5 million. its capital city is poznan ### ### ; other important cities include kalisz, konin, piła, ostrow wielkopolski, gniezno ( an early ### ### capital of poland ) and leszno. it is bordered by seven other voivodeships : west pomeranian to the ### ### northwest, pomeranian to the north, kuyavian - pomeranian to the north - east, łodz to the south - ### ### east, opole to the south, lower silesian ### ### =============================================== v_q ================================================ ### ### ('kali', 0, 9) ('##ʂ', 1, 58) ('##sz', 2, 108) ('greater', 3, 0) ('poland', 4, 3) ### ### ('central', 5, 7862) ('absently', 6, 142) ('voivodeship', 7, 2) ('city', 8, 43) ('polish', 9, 19) ### ### ('.', 10, 53) ('largest', 11, 1036) ('cities', 12, 18) ('2019', 13, 13515) ('intently', 14, 575) ### ### ('##州', 15, 94) ('cradled', 16, 173) ('ק', 17, 95) ('populous', 18, 24) ('colloquially', 19, 18711) ### ### =============================================== v_p ================================================ ### ### ('greater', 3, 0) ('##nin', 21823, 1) ('voivodeship', 7, 2) ('poland', 4, 3) ('les', 20323, 4) ### ### ('##pol', 9088, 5) ('poznan', 160, 6) ('łodz', 253, 7) ('pomeranian', 2267, 8) ('kali', 0, 9) ### ### ('pi', 15172, 10) ('##nie', 10760, 11) ('##no', 2836, 12) ('capital', 24, 13) ('silesian', 242, 14) ### ### ('population', 86, 15) ('##tro', 16379, 16) ('op', 15090, 17) ('cities', 12, 18) ('polish', 9, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('greater', 3, 0) ('kali', 0, 9) ('poland', 4, 3) ('voivodeship', 7, 2) ('polish', 9, 19) ### ### ('cities', 12, 18) ('##ʂ', 1, 58) ('capital', 24, 13) ('city', 8, 43) ('##sz', 2, 108) ### ### ('.', 10, 53) ('populous', 18, 24) ('poznan', 160, 6) ('population', 86, 15) ('absently', 6, 142) ### ### ('łodz', 253, 7) ('silesian', 242, 14) ('##州', 15, 94) ('anyway', 59, 32) ('disbelief', 110, 29) ### ############################################################################################################ [2023-11-25 05:33:04,989][root][INFO] - Epoch: 2: Step: 1601/19530, loss[v]=0.149239, lr=0.000020, acc@1[1]=983.5/1024=0.96044921875, acc@1[2]=1002.5/1024=0.97900390625 [2023-11-25 05:34:54,854][root][INFO] - Train batch 1700 [2023-11-25 05:34:54,855][root][INFO] - Avg. loss per last 100 batches: 0.188576 [2023-11-25 05:34:55,585][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.63 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.4/29522=99.99% | mean: 0.03 | max: 5.36 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: her debut album, " public warning ", was released, featuring " random ", " 9 to 5 ", " ### ### hoodie " and single, " love me or hate me ", which was also released on the same day. ### ### [P_TEXT]: vertically challenged ( ep ) vertically challenged is the debut ep from uk hip hop artist ### ### lady sovereign. it was only released in the u. s. and australia. the ep features the uk hit single ### ### " random " and a remix of her limited edition white - label single " ch ching ( cheque 1 2 ) ". in ### ### the us, a limited edition was manufactured containing a bonus dvd. the limited edition was not ### ### released in australia. the " detroit metro times " wrote that the ep " suitably reps lady sovereign ### ### ’ s ample cockiness and lyrical zing. " " slate " commended sovereign's " tunefully bratty ### ### delivery. " " billboard ### ### =============================================== v_q ================================================ ### ### ('random', 0, 1) ('warning', 1, 5191) ('public', 2, 15959) ('9', 3, 7097) ('##ie', 4, 18873) ### ### ('hate', 5, 6798) ('absently', 6, 223) ('hood', 7, 13493) ('5', 8, 6763) ('nine', 9, 6594) ### ### ('adele', 10, 6547) ('ט', 11, 854) ('her', 12, 67) ('love', 13, 6254) ('me', 14, 16938) ### ### ('debut', 15, 25) ('9th', 16, 11713) ('or', 17, 10990) ('album', 18, 66) ('.', 19, 2543) ### ### =============================================== v_p ================================================ ### ### ('sovereign', 5690, 0) ('random', 0, 1) ('ching', 26872, 2) ('slate', 11012, 3) ### ### ('vertically', 10390, 4) ('challenged', 10299, 5) ('lyrical', 37, 6) ('detroit', 3149, 7) ### ### ('lady', 4577, 8) ('##iness', 16554, 9) ('brat', 12150, 10) ('ep', 199, 11) ('##que', 19221, 12) ### ### ('tune', 791, 13) ('che', 17034, 14) ('ch', 26346, 15) ('times', 11904, 16) ('metro', 4373, 17) ### ### ('delivery', 1226, 18) ('ample', 13721, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('random', 0, 1) ('lyrical', 37, 6) ('debut', 15, 25) ('single', 39, 32) ('ep', 199, 11) ### ### ('song', 22, 49) ('herself', 20, 56) ('her', 12, 67) ('album', 18, 66) ('hop', 191, 31) ### ### ('singles', 92, 39) ('″', 42, 54) ('tune', 791, 13) ('remix', 362, 26) ('absently', 6, 223) ### ### ('uk', 269, 34) ('track', 132, 48) ('billboard', 219, 41) ('delivery', 1226, 18) ('"', 32, 86) ### ############################################################################################################ [2023-11-25 05:34:56,091][root][INFO] - Epoch: 2: Step: 1701/19530, loss[v]=0.264200, lr=0.000020, acc@1[1]=967.5/1024=0.94482421875, acc@1[2]=995.5/1024=0.97216796875 [2023-11-25 05:36:46,416][root][INFO] - Train batch 1800 [2023-11-25 05:36:46,416][root][INFO] - Avg. loss per last 100 batches: 0.191323 [2023-11-25 05:36:47,146][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.63 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.2/29522=99.99% | mean: 0.03 | max: 5.20 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: rural tourism, focusing on folklore and traditions, has become an important alternative, ### ### and is targeted to promote such sites as bran and its dracula's castle, the painted churches of ### ### northern moldavia, and the wooden churches of maramures, or the villages with fortified churches in ### ### transylvania. ### ### [P_TEXT]: saint parascheva church, poienile izei saint parascheva church is a romanian orthodox ### ### church in poienile izei commune, maramures county, romania. built in 1700, it is one of eight ### ### buildings that make up the wooden churches of maramures unesco world heritage site, and is also ### ### listed as a historic monument by the country's ministry of culture and religious affairs. ### ### =============================================== v_q ================================================ ### ### ('bran', 0, 15046) ('##mur', 1, 2) ('mold', 2, 10075) ('dracula', 3, 2896) ### ### ('transylvania', 4, 1537) ('romania', 5, 24) ('.', 6, 337) ('##avia', 7, 28720) ('churches', 8, 23) ### ### ('tourism', 9, 140) ('mara', 10, 3) ('absently', 11, 127) ('##es', 12, 29) ('painted', 13, 4370) ### ### ('romanian', 14, 6) ('northern', 15, 9020) ('rural', 16, 4358) ('castle', 17, 5815) ### ### ('folklore', 18, 3978) ('village', 19, 1003) ### ### =============================================== v_p ================================================ ### ### ('##sche', 27861, 0) ('wooden', 22, 1) ('##mur', 1, 2) ('mara', 10, 3) ('ministry', 5289, 4) ### ### ('##ile', 15547, 5) ('romanian', 14, 6) ('##ien', 24701, 7) ('orthodox', 1990, 8) ('para', 4864, 9) ### ### ('culture', 105, 10) ('##ze', 13141, 11) ('saint', 325, 12) ('worried', 387, 13) ('1700', 3563, 14) ### ### ('unesco', 54, 15) ('church', 20, 16) ('monument', 887, 17) ('county', 1167, 18) ### ### ('heritage', 308, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##mur', 1, 2) ('mara', 10, 3) ('romanian', 14, 6) ('romania', 5, 24) ('wooden', 22, 1) ### ### ('churches', 8, 23) ('##es', 12, 29) ('church', 20, 16) ('site', 29, 20) ('unesco', 54, 15) ### ### ('culture', 105, 10) ('leaned', 56, 25) ('sites', 89, 27) ('saint', 325, 12) ('tourism', 9, 140) ### ### ('religious', 93, 32) ('worried', 387, 13) ('absently', 11, 127) ('heritage', 308, 19) ### ### ('monuments', 203, 28) ### ############################################################################################################ [2023-11-25 05:36:47,649][root][INFO] - Epoch: 2: Step: 1801/19530, loss[v]=0.245786, lr=0.000020, acc@1[1]=965.0/1024=0.9423828125, acc@1[2]=991.5/1024=0.96826171875 [2023-11-25 05:38:37,735][root][INFO] - Train batch 1900 [2023-11-25 05:38:37,736][root][INFO] - Avg. loss per last 100 batches: 0.184910 [2023-11-25 05:38:38,442][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.74 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=100.00% | mean: 0.03 | max: 5.99 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in the book, zychowicz quotes a number of historians and publicists, such as paweł ### ### wieczorkiewicz, andrzej wielowieyski, adolf bochenski, stanisław mackiewicz, władysław studnicki, ### ### jerzy łojek, grzegorz gorski, rafał ziemkiewicz, stanisław zerko, mieczysław pruszynski, ### ### [P_TEXT]: władysław studnicki and communism. another historian, tomasz gabis wrote in his book " ### ### imperial games " ( 2008 ) that if, in the summer of 1939, poland had come to an agreement with ### ### germany, the war would have never taken place, as france and the united kingdom would have had no ### ### reason to declare it on germany. instead of the molotov – ribbentrop pact, a central european anti ### ### - soviet bloc would have been created. another supporter of studnicki's views is piotr zychowicz in ### ### his book " pact ribbentrop - beck ". ### ### =============================================== v_q ================================================ ### ### ('##ych', 0, 8) ('##icz', 1, 4) ('##ʂ', 2, 130) ('paw', 3, 8088) ('adolf', 4, 3019) ### ### ('raf', 5, 5524) ('mack', 6, 15773) ('z', 7, 55) ('##chen', 8, 21722) ('##wicz', 9, 2532) ### ### ('polish', 10, 24) ('##nick', 11, 0) ('##elo', 12, 25404) ('stud', 13, 6) ('##dy', 14, 11) ### ### ('absently', 15, 278) ('stanisław', 16, 326) ('poland', 17, 17) ('##oj', 18, 28229) ### ### ('##kiewicz', 19, 9312) ### ### =============================================== v_p ================================================ ### ### ('##nick', 11, 0) ('##bis', 27733, 1) ('tomas', 8120, 2) ('imperial', 7028, 3) ('##icz', 1, 4) ### ### ('games', 18358, 5) ('stud', 13, 6) ('pact', 20671, 7) ('##ych', 0, 8) ('##sław', 35, 9) ### ### ('##lot', 25968, 10) ('##dy', 14, 11) ('anti', 17627, 12) ('rib', 22680, 13) ('communism', 420, 14) ### ### ('##tro', 19178, 15) ('bloc', 14632, 16) ('poland', 17, 17) ('beck', 10499, 18) ('ga', 19893, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##icz', 1, 4) ('##ych', 0, 8) ('##nick', 11, 0) ('stud', 13, 6) ('##dy', 14, 11) ### ### ('polish', 10, 24) ('poland', 17, 17) ('##sław', 35, 9) ('z', 7, 55) ('book', 32, 40) ### ### ('##ʂ', 2, 130) ('historians', 27, 63) ('leaned', 82, 43) ('historian', 36, 62) ('##ła', 212, 22) ### ### ('##ow', 39, 66) ('communism', 420, 14) ('anyway', 69, 58) ('warsaw', 118, 49) ('w', 96, 53) ### ############################################################################################################ [2023-11-25 05:38:38,946][root][INFO] - Epoch: 2: Step: 1901/19530, loss[v]=0.099616, lr=0.000020, acc@1[1]=996.0/1024=0.97265625, acc@1[2]=1011.0/1024=0.9873046875 [2023-11-25 05:40:29,267][root][INFO] - Train batch 2000 [2023-11-25 05:40:29,268][root][INFO] - Avg. loss per last 100 batches: 0.179890 [2023-11-25 05:40:29,990][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.74 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.7/29522=99.99% | mean: 0.03 | max: 5.42 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: mike mcalary, a " new york daily news " journalist, investigated and reported an expose ### ### of the brutalization of louima by nypd ### ### [P_TEXT]: mike mcalary michael james mcalary ( december 15, 1957 – december 25, 1998 ) was an ### ### american journalist and columnist who worked at the " new york daily news " for 12 years, beginning ### ### with the police beat. he won a pulitzer prize in 1998 for his columns exposing police brutality ### ### against haitian immigrant abner louima. he was sued for libel by a woman he had accused of lying in ### ### her claim that she had been raped. he also wrote five books inspired by cases he had covered. ### ### mcalary died of colon cancer ### ### =============================================== v_q ================================================ ### ### ('##ima', 0, 2) ('mca', 1, 0) ('##lary', 2, 1) ('lou', 3, 8) ('mike', 4, 4) ('daily', 5, 15) ### ### ('ny', 6, 94) ('york', 7, 29) ('brutal', 8, 71) ('news', 9, 17) ('police', 10, 6) ('fbi', 11, 261) ### ### ('##pd', 12, 6982) ('absently', 13, 151) ('rumors', 14, 214) ('abuse', 15, 144) ('files', 16, 1439) ### ### ('″', 17, 88) ('witnesses', 18, 5097) ('expose', 19, 38) ### ### =============================================== v_p ================================================ ### ### ('mca', 1, 0) ('##lary', 2, 1) ('##ima', 0, 2) ('haitian', 8706, 3) ('mike', 4, 4) ### ### ('colon', 12219, 5) ('police', 10, 6) ('beat', 8749, 7) ('lou', 3, 8) ('cancer', 1620, 9) ### ### ('libel', 22459, 10) ('rape', 226, 11) ('##ner', 20825, 12) ('pulitzer', 1814, 13) ### ### ('brutality', 76, 14) ('daily', 5, 15) ('ab', 20265, 16) ('news', 9, 17) ('columns', 3484, 18) ### ### ('immigrants', 4030, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('mca', 1, 0) ('##lary', 2, 1) ('##ima', 0, 2) ('lou', 3, 8) ('mike', 4, 4) ('daily', 5, 15) ### ### ('news', 9, 17) ('police', 10, 6) ('york', 7, 29) ('brutality', 76, 14) ('columnist', 35, 26) ### ### ('brutal', 8, 71) ('rape', 226, 11) ('expose', 19, 38) ('ny', 6, 94) ('journalism', 33, 41) ### ### ('journalist', 28, 47) ('1998', 338, 23) ('haiti', 290, 25) ('cancer', 1620, 9) ### ############################################################################################################ [2023-11-25 05:40:30,496][root][INFO] - Epoch: 2: Step: 2001/19530, loss[v]=0.140487, lr=0.000019, acc@1[1]=994.5/1024=0.97119140625, acc@1[2]=1009.0/1024=0.9853515625 [2023-11-25 05:42:20,934][root][INFO] - Train batch 2100 [2023-11-25 05:42:20,934][root][INFO] - Avg. loss per last 100 batches: 0.168850 [2023-11-25 05:42:21,657][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 5.77 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the station is owned by good karma broadcasting, llc, and runs a sports talk radio format ### ### as a network affiliate of espn radio, sharing most programming with milwaukee based wkti and wauk ### ### identifying jointly on air as espn wisconsin. ### ### [P_TEXT]: milwaukee bucks radio network spectrum news 1, has been part of the bmo bucks radio ### ### network since 1996. davis does play - by - play for every game, with krause providing color ### ### commentary at home games. occasionally, davis may have a broadcasting assignment for a national ### ### radio network, and krause performs play - by - play without a partner. the post - game show is made ### ### up of two parts, the first half hour being the radio. com post game show, and the last hour being ### ### bucks and - one ( branded as bucks and - one presented by noble vines ), which airs only on espn ### ### milwaukee and espn madison. both parts are hosted by ### ### =============================================== v_q ================================================ ### ### ('##kti', 0, 27983) ('espn', 1, 6) ('karma', 2, 25692) ('wisconsin', 3, 48) ('##uk', 4, 24293) ### ### ('milwaukee', 5, 2) ('good', 6, 2709) ('radio', 7, 14) ('llc', 8, 15405) ('absently', 9, 173) ### ### ('w', 10, 11074) ('.', 11, 510) ('wa', 12, 10311) ('sports', 13, 258) ('talk', 14, 2253) ### ### ('network', 15, 16) ('broadcasting', 16, 75) ('##®', 17, 114) ('coverage', 18, 2033) ### ### ('intently', 19, 638) ### ### =============================================== v_p ================================================ ### ### ('bucks', 10575, 0) ('##raus', 27430, 1) ('milwaukee', 5, 2) ('vines', 9871, 3) ('noble', 8911, 4) ### ### ('madison', 1419, 5) ('espn', 1, 6) ('post', 7803, 7) ('spectrum', 693, 8) ('##mo', 23376, 9) ### ### ('davis', 1896, 10) ('com', 6299, 11) ('assignment', 5527, 12) ('news', 72, 13) ('radio', 7, 14) ### ### ('color', 335, 15) ('network', 15, 16) ('leaned', 229, 17) ('hour', 427, 18) ('ones', 2678, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('espn', 1, 6) ('milwaukee', 5, 2) ('radio', 7, 14) ('wisconsin', 3, 48) ('network', 15, 16) ### ### ('news', 72, 13) ('spectrum', 693, 8) ('game', 137, 21) ('color', 335, 15) ('leaned', 229, 17) ### ### ('madison', 1419, 5) ('games', 125, 28) ('absently', 9, 173) ('hour', 427, 18) ### ### ('broadcasting', 16, 75) ('anyway', 161, 34) ('host', 30, 61) ('ק', 27, 65) ('show', 183, 32) ### ### ('davis', 1896, 10) ### ############################################################################################################ [2023-11-25 05:42:22,160][root][INFO] - Epoch: 2: Step: 2101/19530, loss[v]=0.290727, lr=0.000019, acc@1[1]=961.5/1024=0.93896484375, acc@1[2]=990.5/1024=0.96728515625 [2023-11-25 05:44:12,255][root][INFO] - Train batch 2200 [2023-11-25 05:44:12,256][root][INFO] - Avg. loss per last 100 batches: 0.179571 [2023-11-25 05:44:12,982][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.6/29522=99.99% | mean: 0.03 | max: 5.46 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: jacopo da carrara was elected lord of padua in 1318, at that point the city was home to ### ### 40, 000 ### ### [P_TEXT]: marsilio da carrara marsilio da carrara ( 1294 – march 1338 ) was lord of padua after his ### ### uncle jacopo i. he was a member of the carraresi family. he successfully faced a plot against him ### ### in the city. however, after treason of his nephew nicolo da carrara who had sided with the scaliger ### ### of verona, marsilio was forced to relinquish padua to cangrande della scala in 1328. he retained ### ### the title of vicar for the city, and managed to marry jacopo's daughter taddea to mastino ii della ### ### scala. in the war against the papal states, he warred alongside the latter at brescia in 1330 / ### ### 1331, conquering the ### ### =============================================== v_q ================================================ ### ### ('padua', 0, 1) ('carr', 1, 3) ('131', 2, 72) ('##ara', 3, 22) ('ja', 4, 27) ('##co', 5, 21) ### ### ('italy', 6, 83) ('venice', 7, 444) ('.', 8, 534) ('city', 9, 47) ('da', 10, 31) ('lord', 11, 42) ### ### ('##po', 12, 14) ('absently', 13, 206) ('garion', 14, 145) ('sparhawk', 15, 385) ### ### ('calabria', 16, 4066) ('##¨', 17, 3806) ('##州', 18, 167) ('naples', 19, 301) ### ### =============================================== v_p ================================================ ### ### ('##ares', 8806, 0) ('padua', 0, 1) ('tad', 12096, 2) ('carr', 1, 3) ('verona', 38, 4) ### ### ('treason', 4845, 5) ('vicar', 1767, 6) ('mast', 17820, 7) ('mars', 15234, 8) ('nico', 25189, 9) ### ### ('scala', 1364, 10) ('##ili', 24216, 11) ('##ali', 21845, 12) ('133', 244, 13) ('##po', 12, 14) ### ### ('##gra', 25474, 15) ('della', 192, 16) ('##de', 14248, 17) ('129', 73, 18) ('br', 17167, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('padua', 0, 1) ('carr', 1, 3) ('##ara', 3, 22) ('##co', 5, 21) ('ja', 4, 27) ('##po', 12, 14) ### ### ('verona', 38, 4) ('da', 10, 31) ('papal', 34, 20) ('131', 2, 72) ('lord', 11, 42) ('city', 9, 47) ### ### ('129', 73, 18) ('italy', 6, 83) ('della', 192, 16) ('medieval', 21, 50) ('133', 244, 13) ### ### ('132', 59, 34) ('anyway', 83, 30) ('134', 81, 44) ### ############################################################################################################ [2023-11-25 05:44:13,489][root][INFO] - Epoch: 2: Step: 2201/19530, loss[v]=0.274545, lr=0.000019, acc@1[1]=958.5/1024=0.93603515625, acc@1[2]=992.0/1024=0.96875 [2023-11-25 05:46:03,824][root][INFO] - Train batch 2300 [2023-11-25 05:46:03,825][root][INFO] - Avg. loss per last 100 batches: 0.183726 [2023-11-25 05:46:04,559][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 4.09 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.7/29522=99.99% | mean: 0.03 | max: 5.44 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 2008 the ninth general election saw a return to power for sheikh hasina and the awami ### ### league led grand alliance in a landslide victory. ### ### [P_TEXT]: chandpur - 2 chandpur - 2 is a constituency represented in the jatiya sangsad ( national ### ### parliament ) of bangladesh since 2019 by nurul amin ruhul of the awami league. the constituency ### ### encompasses matlab dakshin and matlab uttar upazilas. the constituency was created in 1984 from the ### ### comilla - 20 constituency when the former comilla district was split into three districts : ### ### brahmanbaria, comilla, and chandpur. ahead of the 2008 general election, the election commission ### ### redrew constituency boundaries to reflect population changes revealed ### ### =============================================== v_q ================================================ ### ### ('2008', 0, 8) ('sheikh', 1, 15733) ('pakistan', 2, 227) ('bangladesh', 3, 2) ('ninth', 4, 15628) ### ### ('9th', 5, 7185) ('absently', 6, 125) ('##ina', 7, 20411) ('alliance', 8, 415) ('.', 9, 206) ### ### ('grand', 10, 21015) ('general', 11, 46) ('somali', 12, 1585) ('league', 13, 30) ### ### ('election', 14, 20) ('afghanistan', 15, 237) ('oman', 16, 12672) ('intently', 17, 211) ### ### ('dubai', 18, 3787) ('！', 19, 182) ### ### =============================================== v_p ================================================ ### ### ('##hul', 15451, 0) ('##illa', 5869, 1) ('bangladesh', 3, 2) ('chan', 7232, 3) ('##lab', 24076, 4) ### ### ('mat', 5401, 5) ('aw', 21, 6) ('com', 15306, 7) ('2008', 0, 8) ('##hman', 8388, 9) ### ### ('##bari', 627, 10) ('##kshi', 5439, 11) ('parliament', 194, 12) ('sang', 12199, 13) ### ### ('districts', 1018, 14) ('2', 16944, 15) ('district', 297, 16) ('##dp', 17730, 17) ### ### ('nur', 1871, 18) ('##ur', 4901, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('bangladesh', 3, 2) ('2008', 0, 8) ('aw', 21, 6) ('election', 14, 20) ('league', 13, 30) ### ### ('general', 11, 46) ('constituency', 41, 28) ('##ami', 30, 43) ('elections', 89, 24) ### ### ('parliament', 194, 12) ('anyway', 73, 41) ('district', 297, 16) ('amin', 274, 23) ### ### ('absently', 6, 125) ('##bari', 627, 10) ('worried', 261, 27) ('leaned', 156, 42) ### ### ('pakistan', 2, 227) ('population', 218, 44) ('.', 9, 206) ### ############################################################################################################ [2023-11-25 05:46:05,065][root][INFO] - Epoch: 2: Step: 2301/19530, loss[v]=0.286498, lr=0.000019, acc@1[1]=954.0/1024=0.931640625, acc@1[2]=993.0/1024=0.9697265625 [2023-11-25 05:47:55,257][root][INFO] - Train batch 2400 [2023-11-25 05:47:55,258][root][INFO] - Avg. loss per last 100 batches: 0.196628 [2023-11-25 05:47:55,976][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.72 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.7/29522=99.99% | mean: 0.03 | max: 5.54 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: before the geico 500 race the next day, the drivers and crews pushed wallace's car to the ### ### front of pit road in a show of solidarity, a gesture that drivers jimmie johnson and kevin harvick ### ### proposed. ### ### [P_TEXT]: jimmie johnson history. other drivers and teammates have also lauded johnson for his ### ### leadership and helpfulness, including younger drivers corey lajoie ( for whom johnson advocated to ### ### get a seat in a cup car ), chase elliott, william byron, bubba wallace, alex bowman, and ryan ### ### blaney as well as veterans including brad keselowski, kevin harvick, jeff gordon, and more. in ### ### 2000, " people " recognized johnson as one of their " men in the fast lane ". johnson has ### ### =============================================== v_q ================================================ ### ### ('##ico', 0, 23789) ('wallace', 1, 9) ('jimmie', 2, 0) ('kevin', 3, 19) ('ge', 4, 22740) ### ### ('absently', 5, 192) ('500', 6, 1417) ('##vic', 7, 12) ('crews', 8, 1762) ('.', 9, 70) ### ### ('nascar', 10, 28) ('ha', 11, 24) ('johnson', 12, 5) ('pit', 13, 1627) ('race', 14, 25) ### ### ('season', 15, 273) ('racing', 16, 36) ('intently', 17, 491) ('retirement', 18, 796) ### ### ('waived', 19, 12385) ### ### =============================================== v_p ================================================ ### ### ('jimmie', 2, 0) ('bowman', 785, 1) ('byron', 11652, 2) ('##lane', 9963, 3) ('bubba', 758, 4) ### ### ('johnson', 12, 5) ('elliott', 2024, 6) ('chase', 1148, 7) ('corey', 5533, 8) ('wallace', 1, 9) ### ### ('##owski', 8018, 10) ('lane', 2298, 11) ('##vic', 7, 12) ('gordon', 2228, 13) ('alex', 381, 14) ### ### ('##jo', 18987, 15) ('2000', 145, 16) ('##sel', 28154, 17) ('fast', 3392, 18) ('kevin', 3, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('jimmie', 2, 0) ('wallace', 1, 9) ('johnson', 12, 5) ('kevin', 3, 19) ('##vic', 7, 12) ### ### ('ha', 11, 24) ('nascar', 10, 28) ('race', 14, 25) ('racing', 16, 36) ('.', 9, 70) ### ### ('bowman', 785, 1) ('2000', 145, 16) ('cars', 34, 33) ('bubba', 758, 4) ('anyway', 79, 22) ### ### ('car', 26, 43) ('drivers', 41, 32) ('team', 31, 45) ('motorsports', 32, 44) ('alex', 381, 14) ### ############################################################################################################ [2023-11-25 05:47:56,479][root][INFO] - Epoch: 2: Step: 2401/19530, loss[v]=0.109382, lr=0.000019, acc@1[1]=994.0/1024=0.970703125, acc@1[2]=1007.5/1024=0.98388671875 [2023-11-25 05:49:46,220][root][INFO] - Train batch 2500 [2023-11-25 05:49:46,221][root][INFO] - Avg. loss per last 100 batches: 0.182848 [2023-11-25 05:49:46,948][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.4/29522=99.99% | mean: 0.03 | max: 5.23 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.3/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the early 1980s by hep records, including " memories of you " ( 1983 ). ### ### [P_TEXT]: hep records hep records is a jazz record company and label founded by alastair robertson ### ### in edinburgh, scotland, in 1974. hep started as a reissue label for material from radio ### ### transcription discs, mainly big band music from the 1940s. other reissue material includes fletcher ### ### henderson, andy kirk, jimmie lunceford, and don redman. when it began to issue new recordings, it ### ### added buddy defranco, don lanphere, and eddie thompson. other musicians on the catalogue include ### ### slim gaillard ### ### =============================================== v_q ================================================ ### ### ('##p', 0, 3) ('memories', 1, 2863) ('you', 2, 6139) ('1983', 3, 5434) ('records', 4, 8) ### ### ('he', 5, 24) ('1980s', 6, 3745) ('##ps', 7, 60) ('##ப', 8, 281) ('1984', 9, 6159) ### ### ('memory', 10, 1433) ('album', 11, 157) ('intently', 12, 232) ('##ping', 13, 913) ('label', 14, 45) ### ### ('absently', 15, 130) ('artists', 16, 147) ('##پ', 17, 3012) ('」', 18, 1209) ('music', 19, 41) ### ### =============================================== v_p ================================================ ### ### ('transcription', 3778, 0) ('##pher', 12563, 1) ('robertson', 3958, 2) ('##p', 0, 3) ### ### ('thompson', 1770, 4) ('fletcher', 17617, 5) ('eddie', 567, 6) ('kirk', 11667, 7) ('records', 4, 8) ### ### ('lan', 8486, 9) ('edinburgh', 19506, 10) ('red', 3692, 11) ('##ran', 13545, 12) ### ### ('gail', 12797, 13) ('jimmie', 8176, 14) ('##lard', 26229, 15) ('don', 12069, 16) ('andy', 867, 17) ### ### ('henderson', 1611, 18) ('buddy', 12396, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##p', 0, 3) ('records', 4, 8) ('he', 5, 24) ('##ps', 7, 60) ('label', 14, 45) ### ### ('littered', 40, 34) ('music', 19, 41) ('recordings', 25, 43) ('eddie', 567, 6) ### ### ('reissue', 194, 29) ('anyway', 208, 32) ('jazz', 117, 35) ('radio', 118, 36) ('record', 29, 50) ### ### ('andy', 867, 17) ('musicians', 56, 46) ('thompson', 1770, 4) ('band', 31, 55) ### ### ('transcription', 3778, 0) ('discs', 949, 22) ### ############################################################################################################ [2023-11-25 05:49:47,453][root][INFO] - Epoch: 2: Step: 2501/19530, loss[v]=0.092913, lr=0.000019, acc@1[1]=992.0/1024=0.96875, acc@1[2]=1012.5/1024=0.98876953125 [2023-11-25 05:51:37,416][root][INFO] - Train batch 2600 [2023-11-25 05:51:37,416][root][INFO] - Avg. loss per last 100 batches: 0.184338 [2023-11-25 05:51:38,149][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.63 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.7/29522=99.99% | mean: 0.03 | max: 5.46 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in the year after the recording of their second album, " talk talk talk " ( 1981 ), the ### ### psychedelic furs underwent a tumultuous time of change and, in the words of the guitarist john ### ### ashton, " a bit of a dark period in the band's history ". ### ### [P_TEXT]: john ashton ( musician ) kilburn and roger morris. their first album release was a hit in ### ### the uk and europe, and they had added john ashton on guitar and vince ely on drums. their second ### ### release " talk talk talk " placed them on the us charts. before the release of " forever now " in ### ### late 1982, morris and kilburn left the band, and ashton and the butler brothers went on to release ### ### " mirror moves " and other albums. the psychedelic furs went on hiatus after their 1992 tour, but ### ### reformed in 2001. since then, ashton ### ### =============================================== v_q ================================================ ### ### ('ashton', 0, 0) ('talk', 1, 5) ('fur', 2, 7) ('psychedelic', 3, 13) ('john', 4, 12) ### ### ('1981', 5, 78) ('album', 6, 21) ('band', 7, 22) ('.', 8, 2600) ('bands', 9, 27) ### ### ('absently', 10, 122) ('1980', 11, 403) ('beatles', 12, 48) ('talking', 13, 128) ### ### ('hendrix', 14, 336) ('bowie', 15, 919) ('##₂', 16, 332) ('##s', 17, 22994) ('punk', 18, 469) ### ### ('frontman', 19, 57) ### ### =============================================== v_p ================================================ ### ### ('ashton', 0, 0) ('ely', 5613, 1) ('moves', 1016, 2) ('vince', 2233, 3) ('morris', 3799, 4) ### ### ('talk', 1, 5) ('forever', 1974, 6) ('fur', 2, 7) ('##urn', 16092, 8) ('butler', 5564, 9) ### ### ('mirror', 8905, 10) ('roger', 506, 11) ('john', 4, 12) ('psychedelic', 3, 13) ('now', 13605, 14) ### ### ('1992', 1687, 15) ('##lb', 19308, 16) ('brothers', 3942, 17) ('tour', 239, 18) ### ### ('worried', 3342, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ashton', 0, 0) ('talk', 1, 5) ('fur', 2, 7) ('psychedelic', 3, 13) ('john', 4, 12) ### ### ('album', 6, 21) ('band', 7, 22) ('bands', 9, 27) ('1981', 5, 78) ('1982', 67, 23) ### ### ('beatles', 12, 48) ('albums', 51, 29) ('moves', 1016, 2) ('roger', 506, 11) ('guitar', 38, 31) ### ### ('tour', 239, 18) ('hissed', 20, 52) ('anyway', 116, 26) ('″', 43, 38) ('frontman', 19, 57) ### ############################################################################################################ [2023-11-25 05:51:38,652][root][INFO] - Epoch: 2: Step: 2601/19530, loss[v]=0.258074, lr=0.000019, acc@1[1]=961.5/1024=0.93896484375, acc@1[2]=992.5/1024=0.96923828125 [2023-11-25 05:53:28,706][root][INFO] - Train batch 2700 [2023-11-25 05:53:28,707][root][INFO] - Avg. loss per last 100 batches: 0.187405 [2023-11-25 05:53:29,433][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.75 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.2/29522=99.99% | mean: 0.02 | max: 5.35 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: designer robin walker and executive erik johnson, became fans of the mod and wanted to ### ### build a modern sequel. ### ### [P_TEXT]: robin walker ( game designer ) and later on " team fortress 2 ". walker has played ### ### development roles in various valve games, including " half - life 2 " and " dota 2 ". more ### ### recently, walker has been focused on the collision of economics and game design, in an attempt to ### ### transform " team fortress 2 " into a free - to - play, microtransaction - based game. walker worked ### ### on valve's flagship virtual reality game, " ", released on march 23, 2020. walker has started to ### ### use " team fortress 2 " updates to research what additional features are and ### ### =============================================== v_q ================================================ ### ### ('mod', 0, 10312) ('erik', 1, 4576) ('robin', 2, 4) ('walker', 3, 2) ('johnson', 4, 10485) ### ### ('sequel', 5, 93) ('sequels', 6, 10706) ('game', 7, 11) ('"', 8, 26) ('！', 9, 455) ('″', 10, 25) ### ### ('absently', 11, 125) ('playable', 12, 684) ('modern', 13, 8528) ('design', 14, 14) ### ### ('gamer', 15, 115) ('freaking', 16, 226) ('.', 17, 1637) ('prequel', 18, 5158) ('designer', 19, 19) ### ### =============================================== v_p ================================================ ### ### ('dot', 4936, 0) ('collision', 20622, 1) ('walker', 3, 2) ('fortress', 3232, 3) ('robin', 2, 4) ### ### ('economics', 3698, 5) ('valve', 10568, 6) ('life', 2313, 7) ('half', 8178, 8) ('##act', 23572, 9) ### ### ('reality', 1677, 10) ('game', 7, 11) ('micro', 2082, 12) ('team', 1446, 13) ('design', 14, 14) ### ### ('2', 5097, 15) ('valves', 25037, 16) ('leaned', 3225, 17) ('virtual', 881, 18) ### ### ('designer', 19, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('walker', 3, 2) ('robin', 2, 4) ('game', 7, 11) ('design', 14, 14) ('"', 8, 26) ('″', 10, 25) ### ### ('designer', 19, 19) ('games', 20, 21) ('sequel', 5, 93) ('developers', 27, 34) ### ### ('gameplay', 77, 24) ('designers', 55, 39) ('development', 147, 22) ('fortress', 3232, 3) ### ### ('dot', 4936, 0) ('gaming', 167, 35) ('life', 2313, 7) ('economics', 3698, 5) ('absently', 11, 125) ### ### ('update', 180, 37) ### ############################################################################################################ [2023-11-25 05:53:29,938][root][INFO] - Epoch: 2: Step: 2701/19530, loss[v]=0.125811, lr=0.000019, acc@1[1]=992.5/1024=0.96923828125, acc@1[2]=1011.0/1024=0.9873046875 [2023-11-25 05:55:20,044][root][INFO] - Train batch 2800 [2023-11-25 05:55:20,045][root][INFO] - Avg. loss per last 100 batches: 0.187623 [2023-11-25 05:55:20,760][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.65 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.0/29522=99.99% | mean: 0.03 | max: 5.31 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: quality of american production and was " miserable until [ maurice ] tourneur came along. ### ### " ### ### [P_TEXT]: the rail rider the rail rider is a 1916 american silent drama film directed by maurice ### ### tourneur and starring house peters, bertram marburgh, and henry west. the film's sets were designed ### ### by the art director ben carre. fragmentary prints of " the rail rider " survive in the library of ### ### congress and the dawson city collection of the library and archives canada. ### ### =============================================== v_q ================================================ ### ### ('maurice', 0, 3) ('tour', 1, 6) ('##ne', 2, 34) ('absently', 3, 102) ('##ur', 4, 23) ### ### ('production', 5, 631) ('miserable', 6, 17643) ('intently', 7, 303) ('″', 8, 185) ### ### ('productions', 9, 536) ('americans', 10, 24) ('bernstein', 11, 5640) ('convinces', 12, 60) ### ### ('american', 13, 389) ('silent', 14, 13) ('.', 15, 3299) ('quality', 16, 6552) ('"', 17, 353) ### ### ('insists', 18, 1107) ('titanic', 19, 11303) ### ### =============================================== v_p ================================================ ### ### ('rider', 6255, 0) ('dawson', 6743, 1) ('bertram', 436, 2) ('maurice', 0, 3) ('carr', 8287, 4) ### ### ('peters', 13815, 5) ('tour', 1, 6) ('ben', 15306, 7) ('henry', 3512, 8) ('rail', 15310, 9) ### ### ('##burgh', 26809, 10) ('archives', 5043, 11) ('library', 17097, 12) ('silent', 14, 13) ### ### ('fragment', 2494, 14) ('1916', 947, 15) ('congress', 8737, 16) ('house', 6140, 17) ### ### ('canada', 80, 18) ('west', 12724, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('maurice', 0, 3) ('tour', 1, 6) ('##ur', 4, 23) ('##ne', 2, 34) ('silent', 14, 13) ### ### ('americans', 10, 24) ('bertram', 436, 2) ('film', 20, 31) ('canada', 80, 18) ('absently', 3, 102) ### ### ('##titles', 43, 30) ('anyway', 52, 35) ('convinces', 12, 60) ('worried', 72, 41) ### ### ('shocked', 34, 48) ('1916', 947, 15) ('canadian', 131, 40) ('convincing', 282, 26) ('″', 8, 185) ### ### ('rider', 6255, 0) ### ############################################################################################################ [2023-11-25 05:55:21,265][root][INFO] - Epoch: 2: Step: 2801/19530, loss[v]=0.257730, lr=0.000019, acc@1[1]=966.0/1024=0.943359375, acc@1[2]=992.0/1024=0.96875 [2023-11-25 05:57:11,420][root][INFO] - Train batch 2900 [2023-11-25 05:57:11,421][root][INFO] - Avg. loss per last 100 batches: 0.186892 [2023-11-25 05:57:12,167][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.60 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.1/29522=99.99% | mean: 0.03 | max: 5.31 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: their closest rivals are sheffield wednesday, with whom they contest the steel city ### ### derby. ### ### [P_TEXT]: steven maclean ( footballer ) part of the season to score two penalties for wednesday, ### ### first at home to rival club sheffield united in a 2 – 1 defeat and then against league champions ### ### reading, which earned the owls a 1 – 1 draw. maclean sustained another injury setback in august ### ### 2006 after scoring in a 1 – 1 draw against burnley, but made a comeback a month later in the defeat ### ### to derby county. he finished the 2006 – 07 season with 13 goals from 22 first team starts and 22 ### ### substitute appearances, making him ### ### =============================================== v_q ================================================ ### ### ('sheffield', 0, 4) ('steel', 1, 8714) ('wednesday', 2, 8) ('derby', 3, 2) ('.', 4, 45) ### ### ('city', 5, 7013) ('absently', 6, 116) ('rivals', 7, 13) ('№', 8, 140) ('yorkshire', 9, 1052) ### ### ('shirt', 10, 12) ('rugby', 11, 243) ('club', 12, 14) ('clubs', 13, 24) ('rovers', 14, 1526) ### ### ('intently', 15, 663) ('league', 16, 42) ('relegation', 17, 85) ('vimes', 18, 482) ### ### ('thursday', 19, 80) ### ### =============================================== v_p ================================================ ### ### ('maclean', 20287, 0) ('owls', 5890, 1) ('derby', 3, 2) ('steven', 5405, 3) ('sheffield', 0, 4) ### ### ('reading', 1313, 5) ('burnley', 195, 6) ('07', 4865, 7) ('wednesday', 2, 8) ('2006', 4417, 9) ### ### ('rival', 40, 10) ('season', 29, 11) ('shirt', 10, 12) ('rivals', 7, 13) ('club', 12, 14) ### ### ('worried', 274, 15) ('rivalry', 56, 16) ('penalty', 348, 17) ('injury', 867, 18) ### ### ('penalties', 950, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('sheffield', 0, 4) ('derby', 3, 2) ('wednesday', 2, 8) ('rivals', 7, 13) ('shirt', 10, 12) ### ### ('club', 12, 14) ('.', 4, 45) ('season', 29, 11) ('rival', 40, 10) ('clubs', 13, 24) ### ### ('footballer', 20, 23) ('burnley', 195, 6) ('rivalry', 56, 16) ('league', 16, 42) ### ### ('absently', 6, 116) ('knockout', 41, 39) ('matches', 53, 37) ('reading', 1313, 5) ### ### ('match', 62, 38) ('goals', 90, 30) ### ############################################################################################################ [2023-11-25 05:57:12,671][root][INFO] - Epoch: 2: Step: 2901/19530, loss[v]=0.206306, lr=0.000019, acc@1[1]=970.5/1024=0.94775390625, acc@1[2]=994.0/1024=0.970703125 [2023-11-25 05:59:02,824][root][INFO] - Train batch 3000 [2023-11-25 05:59:02,825][root][INFO] - Avg. loss per last 100 batches: 0.182015 [2023-11-25 05:59:03,540][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.54 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.3/29522=99.98% | mean: 0.02 | max: 5.04 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in players championship 8 in march, brown saw off dave chisnall, christian kist, benito ### ### van de pas, cristo reyes and andy boulton, before he was defeated 6 – 7 by raymond van barneveld ### ### [P_TEXT]: toni alcinas defeating brian woods 5 – 3 in his final match. in march 2014, alcinas beat ### ### glenn spearing and matthew edgar to reach the third round of the uk open for the first time in his ### ### career where he was thrashed 9 – 2 by christian kist. he advanced to the last 16 of an event for ### ### the first time in over a year at the seventh players championship courtesy of victories over nigel ### ### heydon, mark barilli and stuart kellett, before losing 6 – 3 to keegan brown ### ### =============================================== v_q ================================================ ### ### ('brown', 0, 18) ('players', 1, 17) ('##ulton', 2, 21522) ('chess', 3, 91) ('ki', 4, 23) ### ### ('chi', 5, 18782) ('pas', 6, 24564) ('##isto', 7, 29176) ('##st', 8, 12) ('benito', 9, 26248) ### ### ('##nall', 10, 24741) ('darts', 11, 172) ('andy', 12, 2934) ('snooker', 13, 71) ### ### ('championship', 14, 45) ('8', 15, 2082) ('absently', 16, 156) ('reyes', 17, 9348) ('.', 18, 230) ### ### ('christian', 19, 16) ### ### =============================================== v_p ================================================ ### ### ('##cina', 28945, 0) ('toni', 9218, 1) ('woods', 2020, 2) ('bari', 16161, 3) ('edgar', 9503, 4) ### ### ('##llet', 25047, 5) ('##egan', 24159, 6) ('spear', 12849, 7) ('hey', 6773, 8) ('matthew', 754, 9) ### ### ('stuart', 3721, 10) ('uk', 930, 11) ('##st', 8, 12) ('glenn', 14493, 13) ('round', 49, 14) ### ### ('nigel', 105, 15) ('christian', 19, 16) ('players', 1, 17) ('brown', 0, 18) ('event', 140, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('brown', 0, 18) ('players', 1, 17) ('##st', 8, 12) ('ki', 4, 23) ('christian', 19, 16) ### ### ('open', 32, 21) ('round', 49, 14) ('littered', 29, 27) ('championship', 14, 45) ('nigel', 105, 15) ### ### ('snooker', 13, 71) ('chess', 3, 91) ('event', 140, 19) ('match', 57, 30) ('tennis', 67, 31) ### ### ('ɐ', 40, 37) ('matthew', 754, 9) ('woods', 2020, 2) ('uk', 930, 11) ('anyway', 141, 29) ### ############################################################################################################ [2023-11-25 05:59:04,051][root][INFO] - Epoch: 2: Step: 3001/19530, loss[v]=0.221007, lr=0.000019, acc@1[1]=967.5/1024=0.94482421875, acc@1[2]=991.5/1024=0.96826171875 [2023-11-25 06:00:54,176][root][INFO] - Train batch 3100 [2023-11-25 06:00:54,177][root][INFO] - Avg. loss per last 100 batches: 0.179237 [2023-11-25 06:00:54,893][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.5/29522=99.99% | mean: 0.03 | max: 5.63 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: a roman road crossed the area roughly on the line of the present a6 road through walkden ### ### and little hulton. ### ### [P_TEXT]: little hulton'walkden yard ', south of walkden high street, close to the ellesmere ### ### colliery, was partly in little hulton. the workshops were built in 1878 by the bridgewater ### ### collieries as a central works depot providing engineering services for its collieries and the ### ### locomotives used on its colliery railway system. it closed as a british coal workshop in 1986 and ### ### is now the site of a housing estate. uk coal was granted planning permission to surface mine 900, ### ### 000 tonnes of coal and rework the cutacre spoil tip in 2001. the ### ### =============================================== v_q ================================================ ### ### ('walk', 0, 2) ('##6', 1, 19231) ('##den', 2, 6) ('hu', 3, 4) ('##lton', 4, 0) ('little', 5, 7) ### ### ('road', 6, 471) ('あ', 7, 1434) ('roman', 8, 10317) ('absently', 9, 92) ('vimes', 10, 898) ### ### ('somerset', 11, 4021) ('england', 12, 102) ('walks', 13, 130) ('walking', 14, 70) ### ### ('bristol', 15, 149) ('sparhawk', 16, 1676) ('##shire', 17, 23332) ('kent', 18, 5479) ### ### ('hampshire', 19, 8224) ### ### =============================================== v_p ================================================ ### ### ('##lton', 4, 0) ('spoil', 1111, 1) ('walk', 0, 2) ('coal', 857, 3) ('hu', 3, 4) ### ### ('surface', 9987, 5) ('##den', 2, 6) ('little', 5, 7) ('elle', 23721, 8) ('depot', 13811, 9) ### ### ('cut', 12421, 10) ('tip', 4155, 11) ('colliery', 114, 12) ('1878', 6922, 13) ('housing', 2142, 14) ### ### ('works', 2779, 15) ('locomotives', 7224, 16) ('##water', 599, 17) ('yard', 4580, 18) ### ### ('bridge', 3589, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('walk', 0, 2) ('##lton', 4, 0) ('hu', 3, 4) ('##den', 2, 6) ('little', 5, 7) ('colliery', 114, 12) ### ### ('spoil', 1111, 1) ('uk', 230, 20) ('street', 208, 24) ('coal', 857, 3) ('walking', 14, 70) ### ### ('##water', 599, 17) ('absently', 9, 92) ('leaned', 219, 38) ('england', 12, 102) ### ### ('britain', 121, 49) ('estate', 130, 47) ('british', 115, 51) ('worried', 408, 33) ### ### ('walks', 13, 130) ### ############################################################################################################ [2023-11-25 06:00:55,396][root][INFO] - Epoch: 2: Step: 3101/19530, loss[v]=0.132549, lr=0.000019, acc@1[1]=988.5/1024=0.96533203125, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 06:02:45,328][root][INFO] - Train batch 3200 [2023-11-25 06:02:45,329][root][INFO] - Avg. loss per last 100 batches: 0.186712 [2023-11-25 06:02:46,066][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.83 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.1/29522=99.99% | mean: 0.03 | max: 5.52 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 29.0/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: by the end of the 1960s, the moog synthesizer took a leading place in the sound of ### ### emerging progressive rock with bands including pink floyd, yes, emerson, lake & palmer, and genesis ### ### making ### ### [P_TEXT]: pink floyd that touched on hard rock, blues, country, folk, and electronic music. " ### ### released in 1968, the song " careful with that axe, eugene " helped galvanise their reputation as ### ### an art rock group. other genres attributed to the band are space rock, experimental rock, acid ### ### rock, proto - prog, experimental pop ( while under barrett ), psychedelic pop, and psychedelic ### ### rock. o'neill surber comments on the music of pink floyd : rarely will you ### ### =============================================== v_q ================================================ ### ### ('genesis', 0, 9350) ('progressive', 1, 4113) ('synthesizer', 2, 1219) ('##og', 3, 16849) ### ### ('absently', 4, 88) ('.', 5, 82) ('emerson', 6, 10222) ('mo', 7, 22104) ('rock', 8, 17) ### ### ('pink', 9, 3) ('floyd', 10, 0) ('palmer', 11, 12656) ('used', 12, 6675) ('lake', 13, 16255) ### ### ('intently', 14, 395) ('yes', 15, 1966) ('1960s', 16, 287) ('band', 17, 35) ('psychedelic', 18, 15) ### ### ('genre', 19, 34) ### ### =============================================== v_p ================================================ ### ### ('floyd', 10, 0) ('careful', 877, 1) ('axe', 789, 2) ('pink', 9, 3) ('barrett', 6503, 4) ### ### ('eugene', 10456, 5) ('acid', 706, 6) ('proto', 5078, 7) ('##vani', 25484, 8) ('space', 474, 9) ### ### ('neill', 13575, 10) ('experimental', 47, 11) ('pro', 10639, 12) ('sur', 23175, 13) ### ### ('##g', 8155, 14) ('psychedelic', 18, 15) ('1968', 124, 16) ('rock', 8, 17) ('art', 2317, 18) ### ### ('gal', 21724, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('floyd', 10, 0) ('pink', 9, 3) ('rock', 8, 17) ('psychedelic', 18, 15) ('experimental', 47, 11) ### ### ('band', 17, 35) ('genre', 19, 34) ('.', 5, 82) ('absently', 4, 88) ('folk', 39, 28) ### ### ('1968', 124, 16) ('axe', 789, 2) ('bands', 49, 29) ('careful', 877, 1) ('electronic', 48, 30) ### ### ('space', 474, 9) ('acid', 706, 6) ('anyway', 91, 27) ('music', 66, 32) ('shocked', 28, 49) ### ############################################################################################################ [2023-11-25 06:02:46,570][root][INFO] - Epoch: 2: Step: 3201/19530, loss[v]=0.109752, lr=0.000019, acc@1[1]=996.0/1024=0.97265625, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 06:04:36,564][root][INFO] - Train batch 3300 [2023-11-25 06:04:36,564][root][INFO] - Avg. loss per last 100 batches: 0.191311 [2023-11-25 06:04:37,286][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.03 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.2/29522=99.99% | mean: 0.03 | max: 5.31 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: vaclav klaus, ### ### [P_TEXT]: valtr komarek 1984 he was director of the forecasting institute of the czechoslovak ### ### academy, heading a team handling proposals for possible amendments. other members of the institute ### ### included vaclav klaus, milos zeman, tomas jezek, vladimir dlouhy, and miloslav ransdorf. in ### ### november 1989 he worked with civic forum, and was appointed first vice - president on 10 december ### ### 1989 in marian calfa government of national unity. from april until june 1990 he was deputy prime ### ### minister. initially one of the most popular economic ministers in ### ### =============================================== v_q ================================================ ### ### ('klaus', 0, 1) ('czech', 1, 50) ('##cl', 2, 22) ('##av', 3, 56) ('va', 4, 54) ('.', 5, 1528) ### ### ('absently', 6, 149) ('prague', 7, 116) ('##ʂ', 8, 176) ('intently', 9, 275) ('slovenia', 10, 779) ### ### ('slavic', 11, 413) ('austria', 12, 300) ('slovakia', 13, 154) ('slovak', 14, 59) ### ### ('polish', 15, 91) ('##さ', 16, 223) ('scraped', 17, 697) ('¹', 18, 150) ('³', 19, 111) ### ### =============================================== v_p ================================================ ### ### ('##mare', 28505, 0) ('klaus', 0, 1) ('czechoslovak', 43, 2) ('calf', 9005, 3) ('##zek', 15304, 4) ### ### ('forecast', 27162, 5) ('milos', 2416, 6) ('forum', 2582, 7) ('marian', 6692, 8) ### ### ('civic', 16122, 9) ('##tr', 10150, 10) ('unity', 8917, 11) ('##sd', 24775, 12) ### ### ('academy', 7429, 13) ('institute', 8196, 14) ('val', 1952, 15) ('##orf', 27103, 16) ### ### ('czechoslovakia', 86, 17) ('vladimir', 67, 18) ('dl', 14955, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('klaus', 0, 1) ('##cl', 2, 22) ('czech', 1, 50) ('va', 4, 54) ('##av', 3, 56) ### ### ('czechoslovak', 43, 2) ('vladimir', 67, 18) ('czechoslovakia', 86, 17) ('littered', 80, 25) ### ### ('slovak', 14, 59) ('prague', 7, 116) ('absently', 6, 149) ('anyway', 85, 39) ('##ʂ', 8, 176) ### ### ('he', 325, 33) ('polish', 15, 91) ('milos', 2416, 6) ('intently', 9, 275) ('1989', 699, 27) ### ### ('forum', 2582, 7) ### ############################################################################################################ [2023-11-25 06:04:37,793][root][INFO] - Epoch: 2: Step: 3301/19530, loss[v]=0.270203, lr=0.000019, acc@1[1]=959.0/1024=0.9365234375, acc@1[2]=987.5/1024=0.96435546875 [2023-11-25 06:06:27,576][root][INFO] - Train batch 3400 [2023-11-25 06:06:27,577][root][INFO] - Avg. loss per last 100 batches: 0.183728 [2023-11-25 06:06:28,299][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.71 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 5.78 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.6/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: this work was extensively plagiarised by richard anthony salisbury, who had memorised ### ### much of the linnean reading and then inserted it in joseph knight's 1809 publication " on the ### ### cultivation of the plants belonging to the natural order of proteeae ". ### ### [P_TEXT]: mimetes cucullatus scolymocephalus, angustiori folio, apicibus tridentatis " [ african ### ### white tree with artichoke head, narrow leafs with tips having three teeth ]. in 1753 carl linnaeus ### ### provided the first valid description and called it " leucadendron cucullatum ", but he changed his ### ### opinion and in 1771 collapsed " leucadendron " into " protea " creating the new combination " ### ### protea cucullata ". richard anthony salisbury called the species " mimetes lyrigera " in 1809, but ### ### his is an illegitimate name because he says it is identical to " leucadendron cucullatum ", and he ### ### thus should have adopted the older species epithet. this was corrected in 1810 by robert brown, who ### ### created the correct combination " mimetes cucullatus ". carl meissner, who contributed a section on ### ### =============================================== v_q ================================================ ### ### ('joseph', 0, 9688) ('knight', 1, 16756) ('salisbury', 2, 0) ('anthony', 3, 14) ('##tee', 4, 16580) ### ### ('pro', 5, 29) ('richard', 6, 42) ('lin', 7, 15879) ('reading', 8, 4796) ('order', 9, 4246) ### ### ('sparhawk', 10, 649) ('##nea', 11, 9454) ('##ae', 12, 16153) ('1809', 13, 31) ('garion', 14, 187) ### ### ('cultivated', 15, 590) ('described', 16, 154) ('##aris', 17, 15364) ('examined', 18, 748) ### ### ('belonging', 19, 14612) ### ### =============================================== v_p ================================================ ### ### ('salisbury', 2, 0) ('trident', 11962, 1) ('##tea', 8483, 2) ('##olio', 2870, 3) ### ### ('linnaeus', 55, 4) ('angus', 10546, 5) ('mei', 2508, 6) ('brown', 4930, 7) ('tips', 20555, 8) ### ### ('illegitimate', 2459, 9) ('leafs', 24966, 10) ('combination', 4809, 11) ('carl', 10206, 12) ### ### ('valid', 1784, 13) ('anthony', 3, 14) ('##oke', 25172, 15) ('##tum', 14550, 16) ### ### ('##iger', 12085, 17) ('1753', 2081, 18) ('##uca', 17330, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('salisbury', 2, 0) ('anthony', 3, 14) ('pro', 5, 29) ('richard', 6, 42) ('1809', 13, 31) ### ### ('linnaeus', 55, 4) ('species', 38, 41) ('1810', 44, 46) ('plant', 22, 85) ('tree', 394, 22) ### ### ('plants', 25, 92) ('trees', 317, 40) ('description', 181, 57) ('##olio', 2870, 3) ### ### ('robert', 1022, 25) ('garion', 14, 187) ('mei', 2508, 6) ('vegetation', 31, 114) ### ### ('described', 16, 154) ('valid', 1784, 13) ### ############################################################################################################ [2023-11-25 06:06:28,801][root][INFO] - Epoch: 2: Step: 3401/19530, loss[v]=0.250558, lr=0.000019, acc@1[1]=969.5/1024=0.94677734375, acc@1[2]=995.0/1024=0.9716796875 [2023-11-25 06:08:19,122][root][INFO] - Train batch 3500 [2023-11-25 06:08:19,123][root][INFO] - Avg. loss per last 100 batches: 0.170328 [2023-11-25 06:08:19,856][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.78 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.3/29522=99.99% | mean: 0.03 | max: 5.42 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the county is subdivided into three districts : the central district, seydun district, ### ### and meydavud district. ### ### [P_TEXT]: chahar darreh chahar darreh (, also romanized as chahar darreh ; also known as chahar ### ### darreh - ye bala ) is a village in saroleh rural district, meydavud district, bagh - e malek ### ### county, khuzestan province, iran. at the 2006 census, its population was 114, in 22 families. ### ### =============================================== v_q ================================================ ### ### ('county', 0, 11) ('district', 1, 12) ('absently', 2, 78) ('iran', 3, 722) ('##yd', 4, 0) ### ### ('intently', 5, 162) ('##av', 6, 7) ('rural', 7, 361) ('¹', 8, 392) ('³', 9, 207) ('##ʂ', 10, 367) ### ### ('##un', 11, 24798) ('central', 12, 6458) ('se', 13, 17890) ('province', 14, 851) ('##ud', 15, 10) ### ### ('.', 16, 3826) ('ந', 17, 137) ('貴', 18, 59) ('me', 19, 15) ### ### =============================================== v_p ================================================ ### ### ('##yd', 4, 0) ('sar', 3711, 1) ('bag', 3412, 2) ('##ole', 26910, 3) ('male', 2226, 4) ### ### ('districts', 200, 5) ('counties', 394, 6) ('##av', 6, 7) ('leaned', 54, 8) ('##hu', 16402, 9) ### ### ('##ud', 15, 10) ('county', 0, 11) ('district', 1, 12) ('worried', 99, 13) ('##zes', 23948, 14) ### ### ('me', 19, 15) ('romanized', 593, 16) ('anyway', 26, 17) ('##h', 9462, 18) ('males', 1193, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##yd', 4, 0) ('county', 0, 11) ('district', 1, 12) ('##av', 6, 7) ('##ud', 15, 10) ('me', 19, 15) ### ### ('leaned', 54, 8) ('anyway', 26, 17) ('absently', 2, 78) ('worried', 99, 13) ('districts', 200, 5) ### ### ('shocked', 21, 22) ('intently', 5, 162) ('counties', 394, 6) ('shrug', 42, 32) ### ### ('disbelief', 58, 25) ('貴', 18, 59) ('iranian', 51, 36) ('³', 9, 207) ('gritted', 38, 47) ### ############################################################################################################ [2023-11-25 06:08:20,361][root][INFO] - Epoch: 2: Step: 3501/19530, loss[v]=0.286498, lr=0.000019, acc@1[1]=957.0/1024=0.9345703125, acc@1[2]=990.5/1024=0.96728515625 [2023-11-25 06:10:10,373][root][INFO] - Train batch 3600 [2023-11-25 06:10:10,374][root][INFO] - Avg. loss per last 100 batches: 0.182691 [2023-11-25 06:10:11,092][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.4/29522=99.99% | mean: 0.03 | max: 5.26 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: mount peacock is a high peak standing directly at the head of kelly glacier, southwest of ### ### mount herschel. ### ### [P_TEXT]: ironside glacier ironside glacier is a glacier, about long, originating at the south side ### ### of mount minto in the admiralty mountains and draining southeast between mount whewell and mount ### ### herschel into moubray bay, victoria land, antarctica. at its mouth it is joined by the honeycomb ### ### glacier flowing in from the north. the name is suggested by an association of ideas involved in the ### ### name admiralty mountains, and by the impression of power given by the great icefall in the lower ### ### portion of ### ### =============================================== v_q ================================================ ### ### ('peacock', 0, 15576) ('kelly', 1, 13319) ('mount', 2, 19) ('hers', 3, 3) ('##chel', 4, 6) ### ### ('glacier', 5, 9) ('peak', 6, 405) ('.', 7, 2914) ('antarctic', 8, 21) ('mountain', 9, 42) ### ### ('everest', 10, 2590) ('mountains', 11, 17) ('antarctica', 12, 35) ('expedition', 13, 75) ### ### ('snow', 14, 384) ('mapping', 15, 4465) ('winter', 16, 781) ('height', 17, 1310) ### ### ('absently', 18, 113) ('geologist', 19, 4266) ### ### =============================================== v_p ================================================ ### ### ('admiralty', 4923, 0) ('##fall', 759, 1) ('mint', 9534, 2) ('hers', 3, 3) ('##side', 2828, 4) ### ### ('iron', 2144, 5) ('##chel', 4, 6) ('##well', 11504, 7) ('honey', 12506, 8) ('glacier', 5, 9) ### ### ('##comb', 5709, 10) ('bay', 379, 11) ('##ub', 26336, 12) ('##ray', 7635, 13) ('##he', 25446, 14) ### ### ('glaciers', 545, 15) ('ice', 249, 16) ('mountains', 11, 17) ('victoria', 11224, 18) ### ### ('mount', 2, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('hers', 3, 3) ('##chel', 4, 6) ('mount', 2, 19) ('glacier', 5, 9) ('antarctic', 8, 21) ### ### ('mountains', 11, 17) ('antarctica', 12, 35) ('mountain', 9, 42) ('##fall', 759, 1) ### ### ('bay', 379, 11) ('ice', 249, 16) ('glaciers', 545, 15) ('admiralty', 4923, 0) ('iron', 2144, 5) ### ### ('vicinity', 75, 31) ('##side', 2828, 4) ('expedition', 13, 75) ('land', 231, 27) ### ### ('nearby', 178, 29) ('glacial', 24, 52) ### ############################################################################################################ [2023-11-25 06:10:11,598][root][INFO] - Epoch: 2: Step: 3601/19530, loss[v]=0.198703, lr=0.000019, acc@1[1]=969.5/1024=0.94677734375, acc@1[2]=1000.5/1024=0.97705078125 [2023-11-25 06:12:01,148][root][INFO] - Train batch 3700 [2023-11-25 06:12:01,148][root][INFO] - Avg. loss per last 100 batches: 0.186275 [2023-11-25 06:12:01,872][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.70 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.3/29522=99.99% | mean: 0.03 | max: 5.54 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: namely, c - e is narrowed to a justly intonated ratio of and at the same time e - g is ### ### widened to the just ratio of the drawback is that the fifths a - e and e - b, by flattening e, ### ### become almost as dissonant as the pythagorean wolf fifth. ### ### [P_TEXT]: just intonation is an interval whose tuning is too far from its just - tuned equivalent, ### ### usually perceived as discordant and undesirable. commas are very small intervals that result from ### ### minute differences between pairs of just intervals. for example, the 5 : 4 ratio is different from ### ### the pythagorean ( 3 - limit ) major third ( 81 : 64 ) by a difference of 81 : 80, called the ### ### syntonic comma. cents are a measure of interval size. in 12 - tone equal temperament, every half ### ### step is 100 cents. pythagorean tuning has been attributed to both pythagoras and eratosthenes by ### ### later writers, ### ### =============================================== v_q ================================================ ### ### ('##nated', 0, 760) ('fifth', 1, 387) ('just', 2, 6) ('##sson', 3, 21255) ('##gor', 4, 13) ### ### ('flat', 5, 1845) ('wolf', 6, 14364) ('absently', 7, 258) ('intently', 8, 507) ('widened', 9, 2407) ### ### ('ratio', 10, 19) ('##ean', 11, 25) ('e', 12, 17057) ('##yt', 13, 15) ('##¨', 14, 2745) ### ### ('.', 15, 186) ('draw', 16, 5611) ('##へ', 17, 355) ('##nation', 18, 1) ('c', 19, 19763) ### ### =============================================== v_p ================================================ ### ### ('cents', 6744, 0) ('##nation', 18, 1) ('limit', 1473, 2) ('##rdan', 14157, 3) ('tuning', 181, 4) ### ### ('temperament', 48, 5) ('just', 2, 6) ('intervals', 1827, 7) ('era', 4616, 8) ('syn', 7363, 9) ### ### ('81', 12103, 10) ('interval', 386, 11) ('tone', 89, 12) ('##gor', 4, 13) ('size', 316, 14) ### ### ('##yt', 13, 15) ('disco', 20927, 16) ('##the', 5672, 17) ('##nes', 21145, 18) ('ratio', 10, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('just', 2, 6) ('##nation', 18, 1) ('##gor', 4, 13) ('ratio', 10, 19) ('##yt', 13, 15) ### ### ('##ean', 11, 25) ('temperament', 48, 5) ('p', 25, 41) ('tuning', 181, 4) ('tone', 89, 12) ### ### ('anyway', 35, 49) ('tuned', 146, 20) ('size', 316, 14) ('interval', 386, 11) ('##tonic', 174, 34) ### ### ('difference', 218, 35) ('tune', 188, 40) ('limit', 1473, 2) ('equal', 335, 31) ('##ha', 30, 67) ### ############################################################################################################ [2023-11-25 06:12:02,374][root][INFO] - Epoch: 2: Step: 3701/19530, loss[v]=0.086946, lr=0.000019, acc@1[1]=1001.0/1024=0.9775390625, acc@1[2]=1015.5/1024=0.99169921875 [2023-11-25 06:13:52,403][root][INFO] - Train batch 3800 [2023-11-25 06:13:52,403][root][INFO] - Avg. loss per last 100 batches: 0.183971 [2023-11-25 06:13:53,129][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.69 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.1/29522=99.99% | mean: 0.03 | max: 5.39 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: two of the founding members, drummer brian downey and bass guitarist and lead vocalist ### ### phil lynott, met while still in school. ### ### [P_TEXT]: a man and a woman ( song ) taken from another completely different song, they chopped it ### ### up and connected together with the new one. it was inspiring to bono to keep working on that. bono ### ### cites thin lizzy founding member phil lynott as the influence for the style in which he sings the ### ### song. bono, who has been interested in the distance that lies between men and women, wrote this ### ### song about " rediscovering a kind of flirtatious and romantic love ". the edge classes " a man and ### ### a woman " as the wild card on " how to dismantle an atomic bomb ". ### ### =============================================== v_q ================================================ ### ### ('##ott', 0, 1) ('phil', 1, 13) ('##ey', 2, 27837) ('##yn', 3, 41) ('brian', 4, 9270) ### ### ('down', 5, 7254) ('absently', 6, 194) ('l', 7, 42) ('.', 8, 1018) ('intently', 9, 433) ### ### ('band', 10, 121) ('smashing', 11, 2614) ('beatles', 12, 72) ('¹', 13, 494) ('uk', 14, 5288) ### ### ('bands', 15, 71) ('nick', 16, 5526) ('riff', 17, 2366) ('metal', 18, 1078) ('³', 19, 370) ### ### =============================================== v_p ================================================ ### ### ('woman', 8503, 0) ('##ott', 0, 1) ('bono', 1327, 2) ('atomic', 506, 3) ('bomb', 492, 4) ### ### ('card', 12223, 5) ('flirt', 11381, 6) ('man', 2319, 7) ('wild', 3664, 8) ('chopped', 9624, 9) ### ### ('how', 1454, 10) ('liz', 4416, 11) ('classes', 10088, 12) ('phil', 1, 13) ('edge', 434, 14) ### ### ('style', 2321, 15) ('thin', 4221, 16) ('ones', 4396, 17) ('##sman', 26694, 18) ### ### ('influences', 109, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ott', 0, 1) ('phil', 1, 13) ('##yn', 3, 41) ('l', 7, 42) ('atomic', 506, 3) ('anyway', 48, 33) ### ### ('influences', 109, 19) ('bomb', 492, 4) ('″', 116, 21) ('absently', 6, 194) ('beatles', 12, 72) ### ### ('bands', 15, 71) ('edge', 434, 14) ('bono', 1327, 2) ('frontman', 25, 62) ('shocked', 27, 58) ### ### ('album', 34, 53) ('worried', 250, 25) ('band', 10, 121) ('member', 83, 48) ### ############################################################################################################ [2023-11-25 06:13:53,635][root][INFO] - Epoch: 2: Step: 3801/19530, loss[v]=0.277771, lr=0.000019, acc@1[1]=959.0/1024=0.9365234375, acc@1[2]=984.5/1024=0.96142578125 [2023-11-25 06:15:43,842][root][INFO] - Train batch 3900 [2023-11-25 06:15:43,843][root][INFO] - Avg. loss per last 100 batches: 0.188365 [2023-11-25 06:15:44,556][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.69 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.5/29522=99.99% | mean: 0.03 | max: 5.24 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: edgar buchanan, too, gives an excellent performance as a good - old - charlie friend, and ### ### beulah bondi is sensible as an orphanage matron. ### ### [P_TEXT]: edgar buchanan the age of 36, after which he turned his dentistry practice over to his ### ### wife. he was a member of theta chi fraternity and a freemason. buchanan appeared in more than 100 ### ### films, including " texas " ( 1941 ), in which he played a dentist and appeared with william holden ### ### and glenn ford and later in " penny serenade " ( 1941 ) with irene dunne and cary grant, " ### ### tombstone, the town too tough to die " ( 1942 ), " the talk of the town ### ### =============================================== v_q ================================================ ### ### ('edgar', 0, 2) ('buchanan', 1, 0) ('##ula', 2, 8920) ('bond', 3, 14715) ('charlie', 4, 7758) ### ### ('″', 5, 33) ('orphanage', 6, 4333) ('"', 7, 39) ('##h', 8, 28007) ('##i', 9, 28004) ('ヘ', 10, 195) ### ### ('mat', 11, 21850) ('intently', 12, 254) ('##ron', 13, 20001) ('be', 14, 21126) ('films', 15, 32) ### ### ('##い', 16, 618) ('picture', 17, 642) ('film', 18, 27) ('1946', 19, 780) ### ### =============================================== v_p ================================================ ### ### ('buchanan', 1, 0) ('dunne', 9655, 1) ('edgar', 0, 2) ('holden', 8506, 3) ('tombstone', 15291, 4) ### ### ('irene', 3919, 5) ('texas', 969, 6) ('serena', 11483, 7) ('glenn', 24249, 8) ('ford', 3484, 9) ### ### ('tough', 3362, 10) ('dentistry', 18299, 11) ('penny', 1108, 12) ('theta', 12038, 13) ### ### ('grant', 7268, 14) ('talk', 2259, 15) ('cary', 22710, 16) ('1941', 69, 17) ('dentist', 20670, 18) ### ### ('town', 11839, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('edgar', 0, 2) ('buchanan', 1, 0) ('″', 5, 33) ('"', 7, 39) ('film', 18, 27) ('films', 15, 32) ### ### ('1941', 69, 17) ('1942', 52, 26) ('1940', 21, 40) ('movie', 46, 37) ('texas', 969, 6) ### ### ('anyway', 128, 29) ('penny', 1108, 12) ('movies', 27, 54) ('ヘ', 10, 195) ('irene', 3919, 5) ### ### ('##titles', 115, 41) ('talk', 2259, 15) ('convincing', 436, 31) ('ford', 3484, 9) ### ############################################################################################################ [2023-11-25 06:15:45,060][root][INFO] - Epoch: 2: Step: 3901/19530, loss[v]=0.125408, lr=0.000019, acc@1[1]=999.0/1024=0.9755859375, acc@1[2]=1012.0/1024=0.98828125 [2023-11-25 06:17:34,815][root][INFO] - Train batch 4000 [2023-11-25 06:17:34,816][root][INFO] - Avg. loss per last 100 batches: 0.171038 [2023-11-25 06:17:35,526][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.75 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.1/29522=99.98% | mean: 0.02 | max: 5.23 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.5/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it lies approximately south - east of sobkow, east of jedrzejow, and south of the ### ### regional capital kielce. ### ### [P_TEXT]: gmina sobkow brzegi, brzezno, choiny, chomentow, jawor, karsy, korytnica, lipa, miasowa, ### ### mokrsko dolne, mokrsko gorne, mzurowa, niziny, nowe kotlice, osowa, sobkow, sokołow dolny, sokołow ### ### gorny, staniowice, stare kotlice, szczepanow, wierzbica, wolka kawecka and zerniki. gmina sobkow is ### ### bordered by the gminas of checiny, imielno, ### ### =============================================== v_q ================================================ ### ### ('sob', 0, 0) ('##kow', 1, 1) ('gmina', 2, 2) ('##iec', 3, 5196) ('jed', 4, 24572) ### ### ('absently', 5, 163) ('##ʂ', 6, 78) ('.', 7, 118) ('##rz', 8, 24952) ('##ej', 9, 25094) ### ### ('№', 10, 3384) ('##ce', 11, 24188) ('within', 12, 4632) ('intently', 13, 323) ('kiel', 14, 10449) ### ### ('³', 15, 40) ('##ow', 16, 1845) ('paused', 17, 45) ('¹', 18, 180) ('poland', 19, 7) ### ### =============================================== v_p ================================================ ### ### ('sob', 0, 0) ('##kow', 1, 1) ('gmina', 2, 2) ('##rnik', 2672, 3) ('nearby', 109, 4) ### ### ('##mento', 16813, 5) ('##gi', 24426, 6) ('poland', 19, 7) ('it', 6232, 8) ('anyway', 32, 9) ### ### ('polish', 26, 10) ('annexed', 91, 11) ('##bic', 8237, 12) ('ze', 4460, 13) ('##we', 12013, 14) ### ### ('neighboring', 1273, 15) ('che', 13440, 16) ('##tl', 6420, 17) ('warsaw', 71, 18) ### ### ('nearest', 3639, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('sob', 0, 0) ('##kow', 1, 1) ('gmina', 2, 2) ('poland', 19, 7) ('polish', 26, 10) ### ### ('anyway', 32, 9) ('nearby', 109, 4) ('##ʂ', 6, 78) ('annexed', 91, 11) ('absently', 5, 163) ### ### ('.', 7, 118) ('warsaw', 71, 18) ('³', 15, 40) ('paused', 17, 45) ('##rnik', 2672, 3) ### ### ('desperately', 39, 31) ('ம', 36, 36) ('##さ', 22, 62) ('shocked', 185, 21) ('ந', 20, 68) ### ############################################################################################################ [2023-11-25 06:17:36,029][root][INFO] - Epoch: 2: Step: 4001/19530, loss[v]=0.233333, lr=0.000019, acc@1[1]=970.0/1024=0.947265625, acc@1[2]=1000.5/1024=0.97705078125 [2023-11-25 06:19:26,239][root][INFO] - Train batch 4100 [2023-11-25 06:19:26,240][root][INFO] - Avg. loss per last 100 batches: 0.180462 [2023-11-25 06:19:26,969][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.68 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.5/29522=99.98% | mean: 0.03 | max: 5.45 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the major tributaries to the truckee river in california from the lake tahoe outlet and ### ### heading downstream include : bear creek, squaw creek, cabin creek, pole creek, donner creek, trout ### ### creek, martis creek, prosser creek, the little truckee river, gray creek, and bronco creek. ### ### [P_TEXT]: donner creek donner creeek is a stream in nevada county, california, that flows into and ### ### out of donner lake. it is a tributary to the truckee river. a significant tributary to donner creek ### ### is cold creek, which drains coldstream canyon, a 12. 5 square mile area south of donner lake, in ### ### and near truckee. ### ### =============================================== v_q ================================================ ### ### ('##ee', 0, 4) ('truck', 1, 2) ('california', 2, 9) ('ta', 3, 21909) ('bear', 4, 9474) ### ### ('creek', 5, 8) ('river', 6, 19) ('##hoe', 7, 27046) ('absently', 8, 52) ('pole', 9, 11922) ### ### ('sq', 10, 4205) ('.', 11, 906) ('cabin', 12, 1553) ('gray', 13, 1177) ('intently', 14, 168) ### ### ('trout', 15, 2589) ('mart', 16, 22153) ('##nco', 17, 21240) ('##ner', 18, 3) ('rivers', 19, 21) ### ### =============================================== v_p ================================================ ### ### ('##stream', 12843, 0) ('nevada', 76, 1) ('truck', 1, 2) ('##ner', 18, 3) ('##ee', 0, 4) ### ### ('cold', 1198, 5) ('cree', 1075, 6) ('don', 20, 7) ('creek', 5, 8) ('california', 2, 9) ### ### ('worried', 68, 10) ('stream', 193, 11) ('tributary', 27, 12) ('leaned', 322, 13) ### ### ('##ek', 14534, 14) ('lake', 34, 15) ('canyon', 205, 16) ('trucks', 29, 17) ('county', 127, 18) ### ### ('river', 6, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('truck', 1, 2) ('##ee', 0, 4) ('california', 2, 9) ('creek', 5, 8) ('##ner', 18, 3) ### ### ('river', 6, 19) ('don', 20, 7) ('tributary', 27, 12) ('nevada', 76, 1) ('rivers', 19, 21) ### ### ('lake', 34, 15) ('trucks', 29, 17) ('worried', 68, 10) ('absently', 8, 52) ('tributaries', 70, 22) ### ### ('county', 127, 18) ('stream', 193, 11) ('canyon', 205, 16) ('leaned', 322, 13) ('cree', 1075, 6) ### ############################################################################################################ [2023-11-25 06:19:27,474][root][INFO] - Epoch: 2: Step: 4101/19530, loss[v]=0.077510, lr=0.000019, acc@1[1]=995.5/1024=0.97216796875, acc@1[2]=1015.0/1024=0.9912109375 [2023-11-25 06:21:17,580][root][INFO] - Train batch 4200 [2023-11-25 06:21:17,581][root][INFO] - Avg. loss per last 100 batches: 0.186863 [2023-11-25 06:21:18,315][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.3/29522=99.98% | mean: 0.02 | max: 5.36 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: she and her teammate carol habben, who hit a total of 16 homers ( one in postseason ), ### ### were dubbed the " home run twins " by terrorizing ### ### [P_TEXT]: jane stoll of recruiting new players, the lassies posted a fourth - place record of 48 – ### ### 49. in the first round of the playoffs, kalamazoo defeated the second - place south bend ( 48 - 44 ### ### ), two to one games, behind the strong pitching of nancy warren and elaine roth. in the final ### ### round, kalamazoo surprised the heavily favored fort wayne, three to two games, thanks to a clutch ### ### hitting of chris ballingall and carol habben, appropriately dubbed the " home run twins ", and a ### ### nice hurling performance by peppas in decisive game 5, during what turned out to be the ### ### =============================================== v_q ================================================ ### ### ('carol', 0, 17) ('twins', 1, 3) ('##bbe', 2, 6) ('absently', 3, 159) ('ha', 4, 47) ### ### ('postseason', 5, 84) ('run', 6, 27) ('home', 7, 35) ('intently', 8, 193) ('homer', 9, 2322) ### ### ('terror', 10, 4287) ('##子', 11, 226) ('baseball', 12, 44) ('##n', 13, 80) ('！', 14, 1132) ### ### ('jenna', 15, 6276) ('jennifer', 16, 1008) ('tina', 17, 1265) ('alicia', 18, 8120) ### ### ('series', 19, 356) ### ### =============================================== v_p ================================================ ### ### ('##oll', 28546, 0) ('elaine', 184, 1) ('warren', 3029, 2) ('twins', 1, 3) ('jane', 563, 4) ### ### ('bend', 21302, 5) ('##bbe', 2, 6) ('pep', 8551, 7) ('clutch', 95, 8) ('roth', 11789, 9) ### ### ('##pas', 27542, 10) ('##zoo', 29337, 11) ('##inga', 26449, 12) ('nancy', 214, 13) ### ### ('kala', 11033, 14) ('hurling', 1875, 15) ('nice', 5826, 16) ('carol', 0, 17) ('lass', 2983, 18) ### ### ('recruiting', 18312, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('twins', 1, 3) ('##bbe', 2, 6) ('carol', 0, 17) ('run', 6, 27) ('home', 7, 35) ('ha', 4, 47) ### ### ('clutch', 95, 8) ('she', 20, 33) ('baseball', 12, 44) ('elaine', 184, 1) ('playoffs', 80, 23) ### ### ('nancy', 214, 13) ('game', 42, 41) ('hitting', 85, 30) ('jane', 563, 4) ('anyway', 73, 34) ### ### ('postseason', 5, 84) ('absently', 3, 159) ('##n', 13, 80) ('littered', 528, 24) ### ############################################################################################################ [2023-11-25 06:21:18,819][root][INFO] - Epoch: 2: Step: 4201/19530, loss[v]=0.324070, lr=0.000019, acc@1[1]=956.0/1024=0.93359375, acc@1[2]=984.5/1024=0.96142578125 [2023-11-25 06:23:08,845][root][INFO] - Train batch 4300 [2023-11-25 06:23:08,845][root][INFO] - Avg. loss per last 100 batches: 0.190947 [2023-11-25 06:23:09,562][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.62 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.6/29522=99.98% | mean: 0.02 | max: 5.35 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: initially, atoms are energized from the ground state to the excited state by a process ### ### called " pumping ", described below. ### ### [P_TEXT]: q - switching optical resonator. a high q factor corresponds to low resonator losses per ### ### roundtrip, and vice versa. the variable attenuator is commonly called a " q - switch ", when used ### ### for this purpose. initially the laser medium is pumped while the q - switch is set to prevent ### ### feedback of light into the gain medium ( producing an optical resonator with low q ). this produces ### ### a population inversion, but laser operation cannot yet occur since there is no feedback from the ### ### resonator. since the rate of stimulated emission is dependent on the amount of light entering the ### ### medium, ### ### =============================================== v_q ================================================ ### ### ('pumping', 0, 18) ('pump', 1, 22) ('ground', 2, 9457) ('nuclear', 3, 1796) ('intently', 4, 735) ### ### ('state', 5, 15310) ('absently', 6, 342) ('##¨', 7, 1241) ('.', 8, 117) ('##gi', 9, 26241) ### ### ('quantum', 10, 1803) ('excited', 11, 185) ('pumped', 12, 12) ('##₈', 13, 327) ### ### ('chemist', 14, 19427) ('atomic', 15, 11754) ('atom', 16, 11493) ('atoms', 17, 3668) ### ### ('isotope', 18, 9189) ('element', 19, 343) ### ### =============================================== v_p ================================================ ### ### ('inversion', 378, 0) ('switching', 9875, 1) ('gain', 1301, 2) ('##trip', 2737, 3) ### ### ('feedback', 229, 4) ('emission', 160, 5) ('stimulated', 843, 6) ('q', 8299, 7) ('switch', 6790, 8) ### ### ('laser', 100, 9) ('optical', 1195, 10) ('factor', 1533, 11) ('pumped', 12, 12) ### ### ('population', 10790, 13) ('##ten', 23135, 14) ('res', 13720, 15) ('medium', 262, 16) ### ### ('populations', 15664, 17) ('pumping', 0, 18) ('##ua', 24652, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('pumping', 0, 18) ('pump', 1, 22) ('pumped', 12, 12) ('emission', 160, 5) ('inversion', 378, 0) ### ### ('laser', 100, 9) ('feedback', 229, 4) ('stimulated', 843, 6) ('medium', 262, 16) ('gain', 1301, 2) ### ### ('variable', 172, 23) ('anyway', 131, 26) ('optical', 1195, 10) ('##trip', 2737, 3) ### ### ('factor', 1533, 11) ('emissions', 548, 24) ('rate', 836, 21) ('.', 8, 117) ('worried', 415, 29) ### ### ('prevent', 457, 33) ### ############################################################################################################ [2023-11-25 06:23:10,066][root][INFO] - Epoch: 2: Step: 4301/19530, loss[v]=0.240399, lr=0.000019, acc@1[1]=964.5/1024=0.94189453125, acc@1[2]=995.5/1024=0.97216796875 [2023-11-25 06:25:00,031][root][INFO] - Train batch 4400 [2023-11-25 06:25:00,031][root][INFO] - Avg. loss per last 100 batches: 0.176255 [2023-11-25 06:25:00,749][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.57 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.6/29522=99.99% | mean: 0.02 | max: 5.15 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: space empires is a series of 4x turn - based strategy games by malfador machinations that ### ### allow the player to assume the role of the leader of a space - faring civilization. ### ### [P_TEXT]: malfador machinations malfador machinations is a small game company based in santa rosa, ### ### california. they are best known for their " space empires " series. malfador machinations was ### ### created by aaron hall to publish " space empires ii ", the sequel to his self - produced 4x game. ### ### on september 6, 2002 malfador machinations released a hack - n - slash rpg entitled " dungeon ### ### odyssey " but this was not as successful as the " space empires " games. on april 5, 2006, malfador ### ### machinations was acquired by strategy first. ### ### =============================================== v_q ================================================ ### ### ('space', 0, 8) ('empires', 1, 1) ('game', 2, 14) ('##dor', 3, 10) ('turn', 4, 7703) ### ### ('strategy', 5, 4) ('##fa', 6, 23) ('games', 7, 26) ('civilization', 8, 390) ('4', 9, 15) ### ### ('empire', 10, 38) ('mal', 11, 19) ('mach', 12, 18) ('##x', 13, 13) ('far', 14, 16110) ### ### ('##inations', 15, 5) ('intently', 16, 197) ('！', 17, 233) ('series', 18, 41) ('rpg', 19, 29) ### ### =============================================== v_p ================================================ ### ### ('dungeon', 91, 0) ('empires', 1, 1) ('rosa', 9800, 2) ('aaron', 10864, 3) ('strategy', 5, 4) ### ### ('##inations', 15, 5) ('odyssey', 188, 6) ('hall', 9169, 7) ('space', 0, 8) ('slash', 100, 9) ### ### ('##dor', 3, 10) ('hack', 3302, 11) ('santa', 17877, 12) ('##x', 13, 13) ('game', 2, 14) ### ### ('4', 9, 15) ('worried', 3726, 16) ('sequel', 44, 17) ('mach', 12, 18) ('mal', 11, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('empires', 1, 1) ('space', 0, 8) ('strategy', 5, 4) ('##dor', 3, 10) ('game', 2, 14) ('4', 9, 15) ### ### ('##fa', 6, 23) ('##inations', 15, 5) ('mal', 11, 19) ('##x', 13, 13) ('games', 7, 26) ### ### ('mach', 12, 18) ('dungeon', 91, 0) ('empire', 10, 38) ('rpg', 19, 29) ('gaming', 28, 24) ### ### ('sequel', 44, 17) ('slash', 100, 9) ('odyssey', 188, 6) ('series', 18, 41) ### ############################################################################################################ [2023-11-25 06:25:01,254][root][INFO] - Epoch: 2: Step: 4401/19530, loss[v]=0.093958, lr=0.000019, acc@1[1]=995.5/1024=0.97216796875, acc@1[2]=1017.0/1024=0.9931640625 [2023-11-25 06:26:51,360][root][INFO] - Train batch 4500 [2023-11-25 06:26:51,361][root][INFO] - Avg. loss per last 100 batches: 0.176259 [2023-11-25 06:26:52,109][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.65 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.2/29522=99.99% | mean: 0.03 | max: 5.58 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.8/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: from being lost in the urine ( proteinuria ), there will be a reduction in oncotic ### ### pressure and an increase in filtration across the capillary, resulting in excess fluid buildup in ### ### the tissues ( edema ). ### ### [P_TEXT]: edema parvovirus b19 infection may cause generalized edemas. although a low plasma ### ### oncotic pressure is widely cited for the edema of nephrotic syndrome, most physicians note that the ### ### edema may occur before there is any significant protein in the urine ( proteinuria ) or fall in ### ### plasma protein level. most forms of nephrotic syndrome are due to biochemical and structural ### ### changes in the basement membrane of capillaries in the kidney glomeruli, and these changes occur, ### ### if to a lesser degree, in the vessels of most other tissues of the body. thus the resulting ### ### increase in permeability ### ### =============================================== v_q ================================================ ### ### ('##uria', 0, 5) ('protein', 1, 13) ('ed', 2, 1) ('##cot', 3, 2) ('##ema', 4, 0) ('##illa', 5, 20) ### ### ('##ration', 6, 23564) ('absently', 7, 144) ('pressure', 8, 12) ('intently', 9, 265) ('.', 10, 78) ### ### ('urine', 11, 21) ('##lt', 12, 27001) ('fi', 13, 21949) ('fluid', 14, 73) ('cap', 15, 38) ### ### ('kidney', 16, 16) ('blood', 17, 98) ('pmid', 18, 154) ('increases', 19, 643) ### ### =============================================== v_p ================================================ ### ### ('##ema', 4, 0) ('ed', 2, 1) ('##cot', 3, 2) ('plasma', 307, 3) ('basement', 627, 4) ### ### ('##uria', 0, 5) ('##ability', 1964, 6) ('membrane', 896, 7) ('vessels', 9546, 8) ('b1', 17809, 9) ### ### ('##uli', 21076, 10) ('generalized', 3828, 11) ('pressure', 8, 12) ('protein', 1, 13) ### ### ('##lom', 16836, 14) ('##vo', 15448, 15) ('kidney', 16, 16) ('##rot', 9327, 17) ('##9', 17839, 18) ### ### ('##virus', 9201, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ema', 4, 0) ('ed', 2, 1) ('##uria', 0, 5) ('##cot', 3, 2) ('protein', 1, 13) ### ### ('pressure', 8, 12) ('##illa', 5, 20) ('urine', 11, 21) ('kidney', 16, 16) ('cap', 15, 38) ### ### ('symptoms', 39, 27) ('##ic', 21, 36) ('plasma', 307, 3) ('.', 10, 78) ('fluid', 14, 73) ### ### ('syndrome', 168, 22) ('basement', 627, 4) ('occur', 37, 45) ('anyway', 101, 33) ### ### ('worried', 141, 30) ### ############################################################################################################ [2023-11-25 06:26:52,611][root][INFO] - Epoch: 2: Step: 4501/19530, loss[v]=0.105642, lr=0.000019, acc@1[1]=995.0/1024=0.9716796875, acc@1[2]=1012.0/1024=0.98828125 [2023-11-25 06:28:42,591][root][INFO] - Train batch 4600 [2023-11-25 06:28:42,591][root][INFO] - Avg. loss per last 100 batches: 0.196173 [2023-11-25 06:28:43,306][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.58 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.8/29522=99.99% | mean: 0.03 | max: 5.17 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: caterina va in citta ( english title : " caterina in the big city " ) is a 2003 italian ### ### comedy - drama film directed by paolo virzi and written by virzi and francesco bruni. ### ### [P_TEXT]: margherita buy ozpetek's " the ignorant fairies ", in which she portrayed a widow who ### ### discovered her husband had been having an affair with a man for the last seven years won her a ### ### silver ribbon as best lead. cristina comencini's " the best day of my life " won her another silver ### ### ribbon, this time as best supporting actress. paolo virzi's " caterina in the big city " won her ### ### third david and her third silver ribbon as best supporting actress. for her work in ### ### =============================================== v_q ================================================ ### ### ('##rz', 0, 5) ('2003', 1, 18295) ('cater', 2, 0) ('″', 3, 30) ('va', 4, 24922) ('##ina', 5, 11) ### ### ('##uni', 6, 26351) ('ci', 7, 16611) ('paolo', 8, 10) ('italian', 9, 360) ('vi', 10, 25) ### ### ('##tta', 11, 15313) ('big', 12, 7) ('br', 13, 21728) ('city', 14, 22) ('francesco', 15, 205) ### ### ('film', 16, 174) ('italy', 17, 127) ('intently', 18, 195) ('"', 19, 60) ### ### =============================================== v_p ================================================ ### ### ('cater', 2, 0) ('ribbon', 9423, 1) ('fairies', 11981, 2) ('cristina', 12603, 3) ### ### ('ignorant', 16347, 4) ('##rz', 0, 5) ('buy', 3652, 6) ('big', 12, 7) ('oz', 7431, 8) ### ### ('##pet', 17989, 9) ('paolo', 8, 10) ('##ina', 5, 11) ('david', 4593, 12) ('##eri', 706, 13) ### ### ('##ek', 22287, 14) ('silver', 2206, 15) ('day', 10119, 16) ('##nc', 20513, 17) ('my', 4023, 18) ### ### ('come', 5363, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('cater', 2, 0) ('##rz', 0, 5) ('##ina', 5, 11) ('paolo', 8, 10) ('big', 12, 7) ('vi', 10, 25) ### ### ('″', 3, 30) ('city', 14, 22) ('bigger', 30, 40) ('"', 19, 60) ('dramas', 22, 55) ### ### ('##eri', 706, 13) ('italy', 17, 127) ('film', 16, 174) ('cities', 32, 53) ('francesco', 15, 205) ### ### ('anyway', 160, 27) ('italian', 9, 360) ('hissed', 51, 46) ('##i', 21, 99) ### ############################################################################################################ [2023-11-25 06:28:43,813][root][INFO] - Epoch: 2: Step: 4601/19530, loss[v]=0.237448, lr=0.000019, acc@1[1]=969.0/1024=0.9462890625, acc@1[2]=997.5/1024=0.97412109375 [2023-11-25 06:30:33,547][root][INFO] - Train batch 4700 [2023-11-25 06:30:33,548][root][INFO] - Avg. loss per last 100 batches: 0.175365 [2023-11-25 06:30:34,292][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.71 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.2/29522=99.99% | mean: 0.03 | max: 5.15 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: to the west, kirkdale to the north, and anfield to the north - east. ### ### [P_TEXT]: william connolly ( vc ) in great homer street everton and had a deposit at great homer ### ### street post office savings bank. he died on 31 december 1891 of bronchitis at the johnson family's ### ### residence at 14 westminster road kirkdale, liverpool. he was 75 years old. mrs johnson's daughter, ### ### emma catterall, was present at his passing and informed the kirkdale registrar, robert henry ### ### webster, the next day. he was buried on 4 january 1892 in section ce 17 grave 220 at kirkdale ### ### cemetery. the mode of burial was'public'so the grave was unmarked. most people were buried ### ### in'public ' ### ### =============================================== v_q ================================================ ### ### ('##field', 0, 18739) ('##dale', 1, 9) ('kirk', 2, 8) ('an', 3, 27891) ('##安', 4, 379) ### ### ('intently', 5, 214) ('absently', 6, 156) ('nearby', 7, 53) ('neighbouring', 8, 94) ('.', 9, 3630) ### ### ('yorkshire', 10, 324) ('##fields', 11, 22206) ('ragged', 12, 742) ('shillings', 13, 16682) ### ### ('scotland', 14, 265) ('local', 15, 3837) ('aged', 16, 7703) ('areas', 17, 360) ### ### ('southampton', 18, 175) ('¹', 19, 281) ### ### =============================================== v_p ================================================ ### ### ('connolly', 5556, 0) ('everton', 1253, 1) ('homer', 14212, 2) ('##itis', 24643, 3) ### ### ('registrar', 6505, 4) ('savings', 7173, 5) ('##tera', 28390, 6) ('johnson', 7011, 7) ### ### ('kirk', 2, 8) ('##dale', 1, 9) ('westminster', 1512, 10) ('##nch', 25048, 11) ### ### ('webster', 8301, 12) ('liverpool', 36, 13) ('emma', 2343, 14) ('mode', 9430, 15) ### ### ('220', 14796, 16) ('vc', 26847, 17) ('deposit', 1972, 18) ('william', 1954, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##dale', 1, 9) ('kirk', 2, 8) ('liverpool', 36, 13) ('road', 21, 39) ('nearby', 7, 53) ### ### ('everton', 1253, 1) ('anyway', 147, 34) ('worried', 281, 30) ('westminster', 1512, 10) ### ### ('residence', 502, 22) ('neighbouring', 8, 94) ('family', 546, 25) ('cemetery', 856, 20) ### ### ('accommodation', 46, 61) ('intently', 5, 214) ('street', 679, 26) ('absently', 6, 156) ### ### ('connolly', 5556, 0) ('##安', 4, 379) ('leaned', 287, 44) ### ############################################################################################################ [2023-11-25 06:30:34,796][root][INFO] - Epoch: 2: Step: 4701/19530, loss[v]=0.234623, lr=0.000019, acc@1[1]=969.5/1024=0.94677734375, acc@1[2]=998.0/1024=0.974609375 [2023-11-25 06:32:24,957][root][INFO] - Train batch 4800 [2023-11-25 06:32:24,957][root][INFO] - Avg. loss per last 100 batches: 0.200586 [2023-11-25 06:32:25,667][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.66 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.7/29522=99.99% | mean: 0.03 | max: 5.30 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the peoria party to attempt to colonizing oregon. ### ### [P_TEXT]: peoria party the peoria party was a group of men from peoria in the u. s. state of ### ### illinois, who set out about may 1, 1839, with the intention to colonize the oregon country on ### ### behalf of the united states and drive out the english fur trading companies operating there. the ### ### men of the peoria party were among the first pioneer s to blaze the oregon trail. the fall before ### ### the peoria party began the long journey to oregon, rev. jason lee had visited peoria on a national ### ### speaking tour about the oregon country and moving westward. he recruited sixteen men ### ### =============================================== v_q ================================================ ### ### ('oregon', 0, 3) ('##oria', 1, 0) ('pe', 2, 6) ('party', 3, 18) ('colon', 4, 35) ### ### ('intently', 5, 128) ('.', 6, 472) ('napoleon', 7, 691) ('whig', 8, 2146) ('roosevelt', 9, 2164) ### ### ('lafayette', 10, 5498) ('colonization', 11, 133) ('absently', 12, 111) ('empire', 13, 3049) ### ### ('maine', 14, 1607) ('##yler', 15, 15360) ('arizona', 16, 546) ('longed', 17, 389) ### ### ('washington', 18, 1567) ('statehood', 19, 5182) ### ### =============================================== v_p ================================================ ### ### ('##oria', 1, 0) ('pioneer', 1291, 1) ('trail', 1772, 2) ('oregon', 0, 3) ('jason', 13118, 4) ### ### ('blaze', 10967, 5) ('pe', 2, 6) ('journey', 1773, 7) ('lee', 6914, 8) ('illinois', 217, 9) ### ### ('worried', 807, 10) ('speaking', 11922, 11) ('1839', 1106, 12) ('rev', 15488, 13) ### ### ('fur', 2194, 14) ('country', 1165, 15) ('trading', 2312, 16) ('fall', 3357, 17) ('party', 3, 18) ### ### ('leaned', 2276, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##oria', 1, 0) ('oregon', 0, 3) ('pe', 2, 6) ('party', 3, 18) ('colon', 4, 35) ### ### ('illinois', 217, 9) ('americans', 80, 25) ('pioneer', 1291, 1) ('anyway', 187, 21) ### ### ('1840', 123, 24) ('worried', 807, 10) ('trail', 1772, 2) ('1839', 1106, 12) ('journey', 1773, 7) ### ### ('westward', 613, 20) ('intently', 5, 128) ('country', 1165, 15) ('men', 521, 27) ('1840s', 25, 58) ### ### ('pioneers', 530, 29) ### ############################################################################################################ [2023-11-25 06:32:26,171][root][INFO] - Epoch: 2: Step: 4801/19530, loss[v]=0.099357, lr=0.000019, acc@1[1]=992.0/1024=0.96875, acc@1[2]=1011.5/1024=0.98779296875 [2023-11-25 06:34:16,204][root][INFO] - Train batch 4900 [2023-11-25 06:34:16,205][root][INFO] - Avg. loss per last 100 batches: 0.178420 [2023-11-25 06:34:16,932][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.68 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.8/29522=99.99% | mean: 0.03 | max: 5.60 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: south of the scoria cone is lake tauanui. ### ### [P_TEXT]: lake tauanui lake tauanui is a lake in the northland region of new zealand. north of the ### ### lake is tauanui volcanic scoria cone. ### ### =============================================== v_q ================================================ ### ### ('##anu', 0, 0) ('tau', 1, 2) ('##oria', 2, 1) ('lake', 3, 6) ('cone', 4, 3) ('sc', 5, 15) ### ### ('intently', 6, 167) ('absently', 7, 109) ('##i', 8, 45) ('.', 9, 1716) ('¹', 10, 94) ### ### ('endemic', 11, 4475) ('ecuador', 12, 1845) ('##nesia', 13, 24058) ('³', 14, 468) ### ### ('hawaii', 15, 92) ('##秀', 16, 444) ('alps', 17, 3781) ('volcano', 18, 22) ('##¨', 19, 8911) ### ### =============================================== v_p ================================================ ### ### ('##anu', 0, 0) ('##oria', 2, 1) ('tau', 1, 2) ('cone', 4, 3) ('volcanic', 31, 4) ### ### ('zealand', 41, 5) ('lake', 3, 6) ('##land', 13304, 7) ('worried', 271, 8) ('regions', 1703, 9) ### ### ('cones', 1768, 10) ('leaned', 1208, 11) ('region', 405, 12) ('north', 2319, 13) ### ### ('volcanoes', 879, 14) ('sc', 5, 15) ('anyway', 39, 16) ('auckland', 209, 17) ('lakes', 315, 18) ### ### ('countryside', 4949, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##anu', 0, 0) ('##oria', 2, 1) ('tau', 1, 2) ('lake', 3, 6) ('cone', 4, 3) ('sc', 5, 15) ### ### ('volcanic', 31, 4) ('zealand', 41, 5) ('anyway', 39, 16) ('worried', 271, 8) ('##i', 8, 45) ### ### ('volcano', 18, 22) ('region', 405, 12) ('absently', 7, 109) ('intently', 6, 167) ### ### ('cradled', 24, 27) ('maori', 97, 20) ('¹', 10, 94) ('auckland', 209, 17) ('##い', 44, 24) ### ############################################################################################################ [2023-11-25 06:34:17,438][root][INFO] - Epoch: 2: Step: 4901/19530, loss[v]=0.135833, lr=0.000019, acc@1[1]=992.0/1024=0.96875, acc@1[2]=1008.0/1024=0.984375 [2023-11-25 06:36:07,911][root][INFO] - Train batch 5000 [2023-11-25 06:36:07,911][root][INFO] - Avg. loss per last 100 batches: 0.186496 [2023-11-25 06:36:08,639][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.73 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.8/29522=99.99% | mean: 0.03 | max: 5.41 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the city has a velodrome, the velodrome de saint - quentin - en - yvelines. ### ### [P_TEXT]: cuba at the uci track cycling world championships 2015 uci track cycling world ### ### championships in saint - quentin - en - yvelines at the velodrome de saint - quentin - en - ### ### yvelines from 18 – 22 february 2015. a team of 6 cyclists ( 6 women, 0 men ) was announced to ### ### represent the country in the event. cuba competed at the 2016 uci track cycling world championships ### ### at the lee valley velopark in london, united kingdom from 2 – 4 march 2016. a team of 3 cyclists ( ### ### 3 women, 0 men ) was announced to represent the country in the event. ### ### =============================================== v_q ================================================ ### ### ('##drome', 0, 8) ('quentin', 1, 1) ('##ines', 2, 9) ('absently', 3, 178) ('y', 4, 17) ### ### ('##vel', 5, 16) ('intently', 6, 295) ('ve', 7, 21) ('saint', 8, 7) ('quebec', 9, 4577) ### ### ('commune', 10, 9449) ('.', 11, 386) ('brussels', 12, 3407) ('france', 13, 596) ### ### ('cyclist', 14, 180) ('##lo', 15, 20) ('paris', 16, 920) ('¹', 17, 123) ('³', 18, 177) ### ### ('strasbourg', 19, 8327) ### ### =============================================== v_p ================================================ ### ### ('cuba', 8418, 0) ('quentin', 1, 1) ('##park', 1278, 2) ('cyclists', 11618, 3) ('lee', 13158, 4) ### ### ('uci', 1683, 5) ('valley', 2464, 6) ('saint', 8, 7) ('##drome', 0, 8) ('##ines', 2, 9) ### ### ('championships', 170, 10) ('cycling', 21, 11) ('leaned', 469, 12) ('2016', 1768, 13) ### ### ('track', 92, 14) ('london', 2872, 15) ('##vel', 5, 16) ('y', 4, 17) ('worried', 189, 18) ### ### ('2015', 884, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('quentin', 1, 1) ('##drome', 0, 8) ('##ines', 2, 9) ('saint', 8, 7) ('##vel', 5, 16) ('y', 4, 17) ### ### ('cycling', 21, 11) ('ve', 7, 21) ('##lo', 15, 20) ('track', 92, 14) ('championships', 170, 10) ### ### ('worried', 189, 18) ('en', 67, 25) ('leaned', 469, 12) ('absently', 3, 178) ('anyway', 138, 22) ### ### ('race', 45, 33) ('##park', 1278, 2) ('park', 37, 45) ('intently', 6, 295) ### ############################################################################################################ [2023-11-25 06:36:09,142][root][INFO] - Epoch: 2: Step: 5001/19530, loss[v]=0.212961, lr=0.000019, acc@1[1]=976.5/1024=0.95361328125, acc@1[2]=994.0/1024=0.970703125 [2023-11-25 06:37:59,106][root][INFO] - Train batch 5100 [2023-11-25 06:37:59,107][root][INFO] - Avg. loss per last 100 batches: 0.186543 [2023-11-25 06:37:59,818][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.68 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.4/29522=99.99% | mean: 0.03 | max: 5.51 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: they take various forms ( with various names ) throughout the body ( e. g., histiocyte s, ### ### kupffer cell s, alveolar macrophage s, microglia, and others ), but all are part of the mononuclear ### ### phagocyte system. ### ### [P_TEXT]: lymphatic system cell s and macrophage s while promoting tissue healing. the spleen is a ### ### center of activity of the mononuclear phagocyte system and can be considered analogous to a large ### ### lymph node, as its absence causes a predisposition to certain infection s. like the thymus, the ### ### spleen has only efferent lymphatic vessels. both the short gastric arteries and the splenic artery ### ### supply it with blood. the germinal centers are supplied by arteriole s called " penicilliary ### ### radicles ". until the fifth month of prenatal development ### ### =============================================== v_q ================================================ ### ### ('##ti', 0, 24864) ('absently', 1, 245) ('##pf', 2, 28451) ('##ve', 3, 22754) ('##ago', 4, 4) ### ### ('micro', 5, 1257) ('.', 6, 175) ('intently', 7, 657) ('##glia', 8, 29198) ('##oc', 9, 21282) ### ### ('##fer', 10, 6132) ('ku', 11, 24970) ('##yte', 12, 28482) ('##cle', 13, 11) ('macro', 14, 27) ### ### ('pmid', 15, 152) ('##cy', 16, 20) ('¹', 17, 241) ('cell', 18, 30) ('body', 19, 1226) ### ### =============================================== v_p ================================================ ### ### ('##ffer', 1240, 0) ('arteries', 18043, 1) ('healing', 1539, 2) ('##rio', 21685, 3) ('##ago', 4, 4) ### ### ('##dic', 11825, 5) ('arte', 6668, 6) ('thy', 8618, 7) ('infection', 2481, 8) ('month', 3607, 9) ### ### ('##rmin', 14657, 10) ('##cle', 13, 11) ('##nat', 16344, 12) ('##leen', 18262, 13) ### ### ('node', 176, 14) ('vessels', 13363, 15) ('artery', 305, 16) ('absence', 1415, 17) ### ### ('##ym', 16123, 18) ('##nu', 23, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ago', 4, 4) ('##cle', 13, 11) ('##cy', 16, 20) ('macro', 14, 27) ('cell', 18, 30) ### ### ('##nu', 23, 19) ('##pha', 26, 23) ('##ge', 43, 22) ('system', 38, 38) ('##te', 22, 62) ### ### ('mono', 32, 51) ('node', 176, 14) ('worried', 87, 32) ('artery', 305, 16) ('anyway', 93, 49) ### ### ('##ffer', 1240, 0) ('ph', 70, 61) ('absently', 1, 245) ('.', 6, 175) ('cells', 177, 41) ### ############################################################################################################ [2023-11-25 06:38:00,323][root][INFO] - Epoch: 2: Step: 5101/19530, loss[v]=0.243180, lr=0.000019, acc@1[1]=962.5/1024=0.93994140625, acc@1[2]=991.5/1024=0.96826171875 [2023-11-25 06:39:50,375][root][INFO] - Train batch 5200 [2023-11-25 06:39:50,376][root][INFO] - Avg. loss per last 100 batches: 0.175738 [2023-11-25 06:39:51,114][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.8/29522=99.99% | mean: 0.03 | max: 5.41 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: following deep reconnaissance missions in the region by the 3rd cavalry group and units ### ### of the 1139 engineer combat group, and after heavy fighting in and around the city, chartres was ### ### liberated, on 18 august 1944, by the u. s. 5th infantry and 7th armored divisions belonging to the ### ### xx corps of the u. s. third army commanded by general george s. patton. ### ### [P_TEXT]: george s. patton visiting paris, rennes, chartres, brussels, metz, reims, luxembourg, and ### ### verdun. then he went to stockholm, where he reunited with other athletes from the 1912 olympics. ### ### patton decided that he would leave his post at the 15th army and not return to europe once he left ### ### on december 10 for christmas leave. he intended to discuss with his wife whether he would continue ### ### in a stateside post or retire from the army. patton's chief of ### ### =============================================== v_q ================================================ ### ### ('chart', 0, 4) ('absently', 1, 176) ('patton', 2, 0) ('##res', 3, 17) ('third', 4, 10483) ### ### ('xx', 5, 26354) ('armored', 6, 1899) ('george', 7, 8) ('113', 8, 7997) ('liberated', 9, 1140) ### ### ('cavalry', 10, 99) ('5th', 11, 229) ('reconnaissance', 12, 8891) ('7th', 13, 2785) ### ### ('intently', 14, 690) ('group', 15, 3213) ('3rd', 16, 5949) ('.', 17, 95) ('engineer', 18, 16939) ### ### ('infantry', 19, 113) ### ### =============================================== v_p ================================================ ### ### ('patton', 2, 0) ('metz', 5238, 1) ('luxembourg', 2723, 2) ('brussels', 1082, 3) ('chart', 0, 4) ### ### ('15th', 7562, 5) ('rei', 19813, 6) ('1912', 3710, 7) ('george', 7, 8) ('ren', 19446, 9) ### ### ('army', 105, 10) ('stockholm', 4529, 11) ('##nes', 4568, 12) ('olympics', 8377, 13) ### ### ('##un', 18779, 14) ('##rd', 23152, 15) ('olympic', 13255, 16) ('##res', 3, 17) ### ### ('armies', 2347, 18) ('leave', 8714, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('patton', 2, 0) ('chart', 0, 4) ('george', 7, 8) ('##res', 3, 17) ('army', 105, 10) ### ### ('cavalry', 10, 99) ('absently', 1, 176) ('brussels', 1082, 3) ('.', 17, 95) ('anyway', 121, 32) ### ### ('infantry', 19, 113) ('paris', 574, 20) ('luxembourg', 2723, 2) ('5th', 11, 229) ('metz', 5238, 1) ### ### ('nearby', 558, 27) ('troops', 410, 30) ('1912', 3710, 7) ('charts', 53, 88) ('bothering', 26, 195) ### ############################################################################################################ [2023-11-25 06:39:51,618][root][INFO] - Epoch: 2: Step: 5201/19530, loss[v]=0.114902, lr=0.000019, acc@1[1]=995.5/1024=0.97216796875, acc@1[2]=1009.0/1024=0.9853515625 [2023-11-25 06:41:42,084][root][INFO] - Train batch 5300 [2023-11-25 06:41:42,085][root][INFO] - Avg. loss per last 100 batches: 0.176186 [2023-11-25 06:41:42,803][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.82 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.6/29522=99.99% | mean: 0.03 | max: 5.55 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the name " satoyama " comes from a japanese term for landscape s that include both human ### ### production activities, such as agriculture, forestry and animal husbandry, and natural habitats, ### ### where human influence is an essential aspect of the local ecosystem ### ### [P_TEXT]: satoyama loss of a habitat for the butterflies. " n. fusca " is listed in the japan red ### ### list as endangered – there has been around a 39 % decrease in their records in prefectures. ### ### throughout the 80s and 90s, the satoyama conservation movement was implemented in japan. as of ### ### 2001, there are more than 500 environmental groups that work for the conservation of satoyama. ### ### because of their efforts, satoyama has become more prevalent in japanese landscapes. the satoyama ### ### initiative was established at unesco headquarters in paris in 2009 as a global effort to realize " ### ### societies in harmony with nature " ### ### =============================================== v_q ================================================ ### ### ('sato', 0, 0) ('landscape', 1, 18) ('##yama', 2, 1) ('influence', 3, 9578) ('absently', 4, 202) ### ### ('japanese', 5, 21) ('production', 6, 3688) ('.', 7, 1015) ('japan', 8, 5) ('intently', 9, 1209) ### ### ('natural', 10, 239) ('。', 11, 121) ('agriculture', 12, 373) ('husband', 13, 18328) ### ### ('animal', 14, 2772) ('vegetation', 15, 79) ('¹', 16, 186) ('environment', 17, 53) ### ### ('kyoto', 18, 5480) ('ecosystem', 19, 390) ### ### =============================================== v_p ================================================ ### ### ('sato', 0, 0) ('##yama', 2, 1) ('butterflies', 18114, 2) ('paris', 12628, 3) ('list', 1812, 4) ### ### ('japan', 8, 5) ('##sca', 20860, 6) ('conservation', 640, 7) ('nature', 53, 8) ### ### ('headquarters', 2966, 9) ('harmony', 3316, 10) ('initiative', 7005, 11) ('90s', 23258, 12) ### ### ('societies', 3999, 13) ('lists', 24187, 14) ('listing', 18628, 15) ('red', 12422, 16) ### ### ('unesco', 51, 17) ('landscape', 1, 18) ('n', 17087, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('sato', 0, 0) ('##yama', 2, 1) ('landscape', 1, 18) ('japan', 8, 5) ('japanese', 5, 21) ### ### ('nature', 53, 8) ('landscapes', 37, 22) ('unesco', 51, 17) ('habitat', 29, 27) ('worried', 33, 30) ### ### ('environmental', 76, 26) ('habitats', 30, 40) ('environment', 17, 53) ('anyway', 59, 39) ### ### ('prefecture', 57, 41) ('conservation', 640, 7) ('species', 107, 35) ('prevalent', 63, 45) ### ### ('absently', 4, 202) ('vegetation', 15, 79) ### ############################################################################################################ [2023-11-25 06:41:43,306][root][INFO] - Epoch: 2: Step: 5301/19530, loss[v]=0.119522, lr=0.000019, acc@1[1]=994.5/1024=0.97119140625, acc@1[2]=1012.0/1024=0.98828125 [2023-11-25 06:43:33,307][root][INFO] - Train batch 5400 [2023-11-25 06:43:33,308][root][INFO] - Avg. loss per last 100 batches: 0.158461 [2023-11-25 06:43:34,044][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.5/29522=99.99% | mean: 0.03 | max: 5.62 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: bears survive over the coming years, and the club was provided with priority draft picks ### ### and special recruiting zones to give it access to some of the nation's best talent, which over the ### ### next few years allowed the club to recruit future stars such as michael voss, jason akermanis, ### ### clark keating, steven lawrence and darryl white. ### ### [P_TEXT]: jason akermanis he was not the son of his mother's former partner, john akermanis. his ### ### father was denis dezdjek. he was educated at wavell state high school then at st. joseph's nudgee ### ### college and played football with the mayne football club in the qafl where he was spotted by ### ### brisbane bears talent scouts. he completed year 12 at high school in 1995, during his first year ### ### with the brisbane bears. akermanis was a solid contributor to the brisbane bears midfield ( wing ), ### ### with his pace and skills evident even early in his career. akermanis made his debut ### ### =============================================== v_q ================================================ ### ### ('voss', 0, 21194) ('lawrence', 1, 6099) ('steven', 2, 3233) ('clark', 3, 11634) ### ### ('keating', 4, 18329) ('darryl', 5, 24305) ('##erman', 6, 0) ('jason', 7, 1) ('bears', 8, 7) ### ### ('white', 9, 16825) ('michael', 10, 6069) ('ak', 11, 14) ('.', 12, 1264) ('##is', 13, 31) ### ### ('draft', 14, 3790) ('melbourne', 15, 97) ('absently', 16, 139) ('priority', 17, 9543) ### ### ('season', 18, 77) ('zone', 19, 9745) ### ### =============================================== v_p ================================================ ### ### ('##erman', 6, 0) ('jason', 7, 1) ('##dge', 25974, 2) ('##zd', 27814, 3) ('denis', 21093, 4) ### ### ('wave', 919, 5) ('scouts', 40, 6) ('bears', 8, 7) ('brisbane', 49, 8) ('##af', 27201, 9) ### ### ('joseph', 1014, 10) ('worried', 121, 11) ('##ne', 22814, 12) ('may', 11481, 13) ('ak', 11, 14) ### ### ('##ll', 14128, 15) ('##je', 16658, 16) ('1995', 583, 17) ('midfield', 8035, 18) ('high', 8350, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##erman', 6, 0) ('jason', 7, 1) ('bears', 8, 7) ('ak', 11, 14) ('##is', 13, 31) ('scouts', 40, 6) ### ### ('brisbane', 49, 8) ('club', 28, 27) ('football', 29, 33) ('worried', 121, 11) ('shirt', 109, 25) ### ### ('talent', 134, 24) ('season', 18, 77) ('melbourne', 15, 97) ('wave', 919, 5) ('anyway', 170, 36) ### ### ('joseph', 1014, 10) ('1995', 583, 17) ('pace', 515, 20) ('clubs', 150, 42) ### ############################################################################################################ [2023-11-25 06:43:34,550][root][INFO] - Epoch: 2: Step: 5401/19530, loss[v]=0.131142, lr=0.000019, acc@1[1]=991.5/1024=0.96826171875, acc@1[2]=1012.0/1024=0.98828125 [2023-11-25 06:45:25,076][root][INFO] - Train batch 5500 [2023-11-25 06:45:25,077][root][INFO] - Avg. loss per last 100 batches: 0.172709 [2023-11-25 06:45:25,813][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.2/29522=99.99% | mean: 0.03 | max: 5.46 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: from 2003 through to 2009, parker pearson directed the stonehenge riverside project. ### ### [P_TEXT]: stonehenge riverside project stonehenge was built 500 years earlier than previously ### ### thought. the monument is believed to have been built to unify the peoples of britain. it also found ### ### a previously unknown stone circle, bluestonehenge. the project was directed by mike parker pearson ### ### ( sheffield university ), julian thomas ( manchester university ), colin richards ( manchester ### ### university ), kate welham ( bournemouth university ), joshua pollard ( university of southampton ), ### ### and chris tilley ( university college london ). the main aims of the project were to test the ### ### hypotheses of earlier studies ### ### =============================================== v_q ================================================ ### ### ('pearson', 0, 4) ('parker', 1, 10) ('##hen', 2, 15) ('riverside', 3, 1) ('stone', 4, 18) ### ### ('##ge', 5, 20) ('project', 6, 25) ('2009', 7, 16572) ('2003', 8, 16579) ('directed', 9, 97) ### ### ('directing', 10, 51) ('workshop', 11, 793) ('director', 12, 133) ('creativity', 13, 92) ### ### ('isbn', 14, 63) ('projects', 15, 65) ('improvisation', 16, 3430) ('artistic', 17, 1665) ### ### ('film', 18, 1133) ('filmmakers', 19, 265) ### ### =============================================== v_p ================================================ ### ### ('pollard', 6371, 0) ('riverside', 3, 1) ('##lham', 28718, 2) ('richards', 8215, 3) ### ### ('pearson', 0, 4) ('till', 21683, 5) ('bournemouth', 18350, 6) ('manchester', 2961, 7) ### ### ('julian', 7744, 8) ('southampton', 2905, 9) ('parker', 1, 10) ('circle', 8737, 11) ### ### ('blues', 13401, 12) ('kate', 494, 13) ('colin', 4220, 14) ('##hen', 2, 15) ('joshua', 1500, 16) ### ### ('thomas', 12695, 17) ('stone', 4, 18) ('sheffield', 7517, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('pearson', 0, 4) ('riverside', 3, 1) ('parker', 1, 10) ('##hen', 2, 15) ('stone', 4, 18) ### ### ('##ge', 5, 20) ('project', 6, 25) ('directing', 10, 51) ('isbn', 14, 63) ('projects', 15, 65) ### ### ('directed', 9, 97) ('kate', 494, 13) ('hissed', 28, 62) ('creativity', 13, 92) ('chris', 518, 22) ### ### ('littered', 414, 31) ('director', 12, 133) ('joshua', 1500, 16) ('directs', 43, 69) ### ### ('pollard', 6371, 0) ### ############################################################################################################ [2023-11-25 06:45:26,318][root][INFO] - Epoch: 2: Step: 5501/19530, loss[v]=0.127958, lr=0.000019, acc@1[1]=989.0/1024=0.9658203125, acc@1[2]=1010.5/1024=0.98681640625 [2023-11-25 06:47:16,797][root][INFO] - Train batch 5600 [2023-11-25 06:47:16,798][root][INFO] - Avg. loss per last 100 batches: 0.185718 [2023-11-25 06:47:17,524][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.5/29522=99.98% | mean: 0.03 | max: 5.38 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the municipality has a population of ( ) and is by far the third largest municipality in ### ### finland after nurmijarvi and kirkkonummi that doesn't use the town or city title by itself. ### ### [P_TEXT]: nurmijarvi area. the municipality is located along the middle course of the river vantaa ### ### and there are two large rapids, the nukarinkoski and the myllykoski. the river vantaa is joined by ### ### the palojoki river from hyvinkaa and tuusula, near the palojoki village. the lepsama river and the ### ### luhtajoki river flowing in the southern part of the municipality join the vantaa town side near the ### ### keimola village before joining the river vantaa. the largest lake in nurmijarvi is the lake saaksi ### ### ( which is partly situated on the northwest side of hyvinkaa ), which ### ### =============================================== v_q ================================================ ### ### ('finland', 0, 243) ('nur', 1, 0) ('##kon', 2, 17698) ('##mi', 3, 25) ('kirk', 4, 12279) ### ### ('##um', 5, 17391) ('finnish', 6, 641) ('##vi', 7, 20) ('municipality', 8, 36) ('##jar', 9, 10) ### ### ('absently', 10, 134) ('.', 11, 1880) ('intently', 12, 197) ('city', 13, 2223) ('¹', 14, 178) ### ### ('cities', 15, 451) ('estonia', 16, 832) ('³', 17, 87) ('largest', 18, 90) ('towns', 19, 149) ### ### =============================================== v_p ================================================ ### ### ('nur', 1, 0) ('palo', 20250, 1) ('##aks', 21688, 2) ('kei', 4003, 3) ('van', 2285, 4) ### ### ('##ps', 9445, 5) ('##ama', 3962, 6) ('##vin', 917, 7) ('##hta', 8502, 8) ('rapids', 8058, 9) ### ### ('##jar', 9, 10) ('##mo', 2380, 11) ('##ula', 21315, 12) ('##ta', 6268, 13) ('##nko', 19596, 14) ### ### ('##kos', 910, 15) ('lake', 328, 16) ('##jo', 5070, 17) ('river', 1902, 18) ('tu', 1462, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('nur', 1, 0) ('##mi', 3, 25) ('##jar', 9, 10) ('##vi', 7, 20) ('municipality', 8, 36) ### ### ('village', 40, 22) ('municipalities', 28, 31) ('finland', 0, 243) ('absently', 10, 134) ### ### ('anyway', 88, 32) ('town', 26, 60) ('³', 17, 87) ('intently', 12, 197) ('##vin', 917, 7) ### ### ('lake', 328, 16) ('finnish', 6, 641) ('largest', 18, 90) ('van', 2285, 4) ('ம', 33, 68) ### ### ('sweden', 24, 88) ### ############################################################################################################ [2023-11-25 06:47:18,028][root][INFO] - Epoch: 2: Step: 5601/19530, loss[v]=0.078185, lr=0.000019, acc@1[1]=1000.5/1024=0.97705078125, acc@1[2]=1017.0/1024=0.9931640625 [2023-11-25 06:49:08,295][root][INFO] - Train batch 5700 [2023-11-25 06:49:08,296][root][INFO] - Avg. loss per last 100 batches: 0.172478 [2023-11-25 06:49:09,032][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.4/29522=99.98% | mean: 0.03 | max: 5.26 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: violinist caitlin cary remembers getting drunk and riding a horse bareback across a ### ### nearby field, while other band members played around with firecrackers - which inspired adams to ### ### later remark : " out in the parking lot across the street... the roman candles and black cats ### ### sounded a lot like i'd hope we'd one day sound - pretty little things all set on fire waiting to ### ### get destroyed. " ### ### [P_TEXT]: ryan adams in 1994 after releasing a split 7 " single containing two songs ( the patty ### ### duke syndrome's song on one side, and a song from a band called glamourpuss on the other ). ### ### following the breakup of his high school band, the patty duke syndrome, adams helped found ### ### whiskeytown with caitlin cary, eric " skillet " gilmore, steve grothmann and phil wandscher. ### ### whiskeytown saw adams move to alt - country, describing punk rock as " too hard to sing " in the ### ### title track of whiskeytown's debut album " faithless street ". whiskeytown was ### ### =============================================== v_q ================================================ ### ### ('cary', 0, 4) ('adams', 1, 2) ('caitlin', 2, 3) ('cats', 3, 10348) ('roman', 4, 5362) ### ### ('candles', 5, 14872) ('fire', 6, 10912) ('##acker', 7, 25837) ('##cr', 8, 28111) ### ### ('black', 9, 11995) ('parking', 10, 2669) ('bare', 11, 14261) ('field', 12, 16736) ### ### ('sound', 13, 883) ('absently', 14, 159) ('destruction', 15, 6522) ('destroyed', 16, 2886) ### ### ('.', 17, 2882) ('hope', 18, 1621) ('##back', 19, 23151) ### ### =============================================== v_p ================================================ ### ### ('ryan', 842, 0) ('patty', 7294, 1) ('adams', 1, 2) ('caitlin', 2, 3) ('cary', 0, 4) ### ### ('glamour', 3839, 5) ('whiskey', 1076, 6) ('syndrome', 10950, 7) ('wand', 19742, 8) ### ### ('##sche', 26342, 9) ('gilmore', 8488, 10) ('##less', 12596, 11) ('sing', 5498, 12) ### ### ('skill', 889, 13) ('hard', 4521, 14) ('duke', 9980, 15) ('faith', 3268, 16) ('##pus', 11499, 17) ### ### ('##town', 8259, 18) ('##oth', 20258, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('cary', 0, 4) ('adams', 1, 2) ('caitlin', 2, 3) ('band', 21, 38) ('ryan', 842, 0) ### ### ('bands', 47, 41) ('album', 36, 46) ('street', 121, 28) ('hissed', 52, 42) ('″', 22, 79) ### ### ('whiskey', 1076, 6) ('punk', 176, 27) ('littered', 339, 21) ('skill', 889, 13) ('artists', 44, 62) ### ### ('absently', 14, 159) ('beatles', 87, 49) ('lyrical', 289, 32) ('split', 611, 23) ### ### ('tracks', 103, 48) ### ############################################################################################################ [2023-11-25 06:49:09,538][root][INFO] - Epoch: 2: Step: 5701/19530, loss[v]=0.240864, lr=0.000019, acc@1[1]=960.5/1024=0.93798828125, acc@1[2]=987.5/1024=0.96435546875 [2023-11-25 06:50:59,979][root][INFO] - Train batch 5800 [2023-11-25 06:50:59,980][root][INFO] - Avg. loss per last 100 batches: 0.192046 [2023-11-25 06:51:00,726][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.57 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.5/29522=99.99% | mean: 0.03 | max: 5.28 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in mathematics, an algebraic structure consists of a nonempty set " a " ( called the ### ### underlying set, carrier set or domain ), a collection of operation s on " a " of finite arity ( ### ### typically binary operation s ), and a finite set of identities, known as axioms, that these ### ### operations must satisfy. ### ### [P_TEXT]: exponentiation is then readily shown to be true for every integer. the following ### ### identities hold for all integer exponents, provided that the base is non - zero : the powers of a ### ### sum can normally be computed from the powers of the summands by the binomial formula however, this ### ### formula is true only if the summands commute ( i. e. that ), which is implied if they belong to a ### ### structure that is commutative. otherwise, if and are, say, square matrices of the same size, this ### ### formula cannot be used. it follows that in computer algebra ### ### =============================================== v_q ================================================ ### ### ('algebraic', 0, 89) ('structure', 1, 4) ('underlying', 2, 1950) ('mathematics', 3, 63) ### ### ('absently', 4, 468) ('.', 5, 216) ('carrier', 6, 6260) ('operation', 7, 8456) ('domain', 8, 1279) ### ### ('binary', 9, 424) ('mathematician', 10, 10828) ('intently', 11, 795) ('logic', 12, 149) ### ### ('##mpt', 13, 28211) ('ax', 14, 14328) ('identity', 15, 32) ('##ty', 16, 24386) ### ### ('denoted', 17, 22530) ('ari', 18, 21844) ('collection', 19, 4669) ### ### =============================================== v_p ================================================ ### ### ('##nent', 18639, 0) ('##mute', 10929, 1) ('##iation', 13118, 2) ('base', 790, 3) ### ### ('structure', 1, 4) ('algebra', 75, 5) ('computer', 863, 6) ('matrices', 3535, 7) ('sum', 904, 8) ### ### ('belong', 3372, 9) ('powers', 4042, 10) ('bases', 2208, 11) ('##omi', 17593, 12) ### ### ('implied', 4915, 13) ('expo', 14961, 14) ('##ds', 14276, 15) ('bin', 19178, 16) ### ### ('square', 3985, 17) ('computers', 2774, 18) ('formula', 27, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('structure', 1, 4) ('identity', 15, 32) ('algebra', 75, 5) ('formula', 27, 19) ### ### ('mathematics', 3, 63) ('algebraic', 0, 89) ('if', 26, 30) ('structures', 78, 23) ### ### ('integer', 43, 36) ('anyway', 214, 24) ('zero', 205, 26) ('base', 790, 3) ('matrix', 173, 31) ### ### ('implies', 97, 35) ('computer', 863, 6) ('worried', 70, 38) ('sum', 904, 8) ('cannot', 243, 29) ### ### ('.', 5, 216) ('absently', 4, 468) ### ############################################################################################################ [2023-11-25 06:51:01,229][root][INFO] - Epoch: 2: Step: 5801/19530, loss[v]=0.211508, lr=0.000019, acc@1[1]=974.0/1024=0.951171875, acc@1[2]=1001.0/1024=0.9775390625 [2023-11-25 06:52:51,790][root][INFO] - Train batch 5900 [2023-11-25 06:52:51,791][root][INFO] - Avg. loss per last 100 batches: 0.179177 [2023-11-25 06:52:52,510][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.63 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.9/29522=99.99% | mean: 0.03 | max: 5.50 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: its facilities, supported by the orangutan foundation international, are a part of the ### ### tanjung puting national park. ### ### [P_TEXT]: tanjung puting tanjung puting national park is a national park in indonesia located in ### ### the southeast part of west kotawaringin regency in the indonesian province of central kalimantan ( ### ### central borneo ). the nearest main town is the capital of the regency, pangkalan bun. the park is ### ### famous for its orangutan conservation. the park is composed of 416, 040 hectares of dryland ### ### dipterocarp forest, peat swamp forest, heath forest, mangrove and coastal beach forest, and ### ### secondary forest. the park may be accessed from ### ### =============================================== v_q ================================================ ### ### ('##uta', 0, 4) ('putin', 1, 0) ('tan', 2, 29) ('park', 3, 27) ('##ang', 4, 26) ### ### ('foundation', 5, 8134) ('##jun', 6, 12) ('intently', 7, 357) ('indonesia', 8, 28) ### ### ('international', 9, 3393) ('singapore', 10, 775) ('national', 11, 34) ('infrastructure', 12, 147) ### ### ('absently', 13, 164) ('¹', 14, 137) ('³', 15, 196) ('##g', 16, 58) ('malaysia', 17, 228) ### ### ('cradled', 18, 90) ('museum', 19, 856) ### ### =============================================== v_p ================================================ ### ### ('putin', 1, 0) ('regency', 929, 1) ('borneo', 142, 2) ('##kala', 20583, 3) ('##uta', 0, 4) ### ### ('##wari', 17728, 5) ('dry', 11247, 6) ('bun', 223, 7) ('kota', 11397, 8) ('mangrove', 3645, 9) ### ### ('kali', 2025, 10) ('worried', 286, 11) ('##jun', 6, 12) ('forest', 75, 13) ('dip', 3906, 14) ### ### ('nearest', 11499, 15) ('heath', 11969, 16) ('indonesian', 26, 17) ('coastal', 3023, 18) ### ### ('provinces', 12840, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('putin', 1, 0) ('##uta', 0, 4) ('tan', 2, 29) ('park', 3, 27) ('##ang', 4, 26) ('##jun', 6, 12) ### ### ('indonesia', 8, 28) ('national', 11, 34) ('indonesian', 26, 17) ('borneo', 142, 2) ### ### ('forest', 75, 13) ('bun', 223, 7) ('##g', 16, 58) ('##ng', 60, 35) ('regency', 929, 1) ### ### ('worried', 286, 11) ('anyway', 105, 33) ('parks', 61, 44) ('vegetation', 97, 42) ### ### ('rainforest', 38, 55) ### ############################################################################################################ [2023-11-25 06:52:53,015][root][INFO] - Epoch: 2: Step: 5901/19530, loss[v]=0.098859, lr=0.000018, acc@1[1]=994.0/1024=0.970703125, acc@1[2]=1015.0/1024=0.9912109375 [2023-11-25 06:54:43,052][root][INFO] - Train batch 6000 [2023-11-25 06:54:43,053][root][INFO] - Avg. loss per last 100 batches: 0.189506 [2023-11-25 06:54:43,773][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.68 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.1/29522=99.99% | mean: 0.03 | max: 5.36 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.2/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the equivalent title in the scottish peerage was duke of albany. ### ### [P_TEXT]: duke of york and albany duke of york and albany was a title of nobility in the peerage of ### ### great britain. the title was created three times during the 18th century and was usually given to ### ### the second son of british monarchs. the predecessor titles in the english and scottish peerages ### ### were duke of york and duke of albany. the individual dukedoms of york and of albany had previously ### ### each been created several times in the peerages of england and of scotland respectively. each had ### ### become a traditional title ### ### =============================================== v_q ================================================ ### ### ('albany', 0, 0) ('duke', 1, 2) ('scottish', 2, 11) ('peerage', 3, 3) ('.', 4, 65) ### ### ('vimes', 5, 1807) ('scotland', 6, 6) ('##ʂ', 7, 5144) ('styled', 8, 502) ('duchy', 9, 71) ### ### ('¹', 10, 129) ('№', 11, 6887) ('titles', 12, 12) ('absently', 13, 173) ('title', 14, 16) ### ### ('intently', 15, 357) ('³', 16, 338) ('duchess', 17, 363) ('royal', 18, 876) ('hound', 19, 6840) ### ### =============================================== v_p ================================================ ### ### ('albany', 0, 0) ('nobility', 36, 1) ('duke', 1, 2) ('peerage', 3, 3) ('york', 46, 4) ### ### ('britain', 156, 5) ('scotland', 6, 6) ('monarch', 196, 7) ('sons', 788, 8) ('individual', 315, 9) ### ### ('son', 1445, 10) ('scottish', 2, 11) ('titles', 12, 12) ('anyway', 551, 13) ('dukes', 38, 14) ### ### ('18th', 702, 15) ('title', 14, 16) ('##dom', 15247, 17) ('traditional', 2040, 18) ### ### ('british', 219, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('albany', 0, 0) ('duke', 1, 2) ('peerage', 3, 3) ('scottish', 2, 11) ('scotland', 6, 6) ### ### ('nobility', 36, 1) ('titles', 12, 12) ('title', 14, 16) ('york', 46, 4) ('.', 4, 65) ### ### ('dukes', 38, 14) ('britain', 156, 5) ('monarch', 196, 7) ('individual', 315, 9) ('duchy', 9, 71) ### ### ('aristocratic', 31, 39) ('medieval', 21, 45) ('british', 219, 19) ('sons', 788, 8) ### ### ('anyway', 551, 13) ### ############################################################################################################ [2023-11-25 06:54:44,277][root][INFO] - Epoch: 2: Step: 6001/19530, loss[v]=0.320490, lr=0.000018, acc@1[1]=954.5/1024=0.93212890625, acc@1[2]=975.5/1024=0.95263671875 [2023-11-25 06:56:34,263][root][INFO] - Train batch 6100 [2023-11-25 06:56:34,264][root][INFO] - Avg. loss per last 100 batches: 0.180960 [2023-11-25 06:56:35,014][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.68 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.5/29522=99.99% | mean: 0.03 | max: 5.57 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: there has been recent modernization on this route, with the old spessart ramp and the ### ### schwarzkopf tunnel being replaced with a series of new tunnels between laufach and heigenbrucken. ### ### [P_TEXT]: laufach laufach is a community in the aschaffenburg district in the " regierungsbezirk " ### ### of lower franconia ( " unterfranken " ) in bavaria, germany. the community lies in northwest ### ### bavaria east of aschaffenburg. laufach lies in the heart of the vorspessart ( range ) and stretches ### ### through its outlying centre of hain into the high spessart. the community's elevation ranges from ### ### 170 to 500 m above sea level. laufach's " ortsteile " are frohnhofen ( 898 inhabitants ), hain ( 1, ### ### 177 inhabitants ) and laufach ( 3, 204 inhabitants ). in 1084, laufach, which was ### ### =============================================== v_q ================================================ ### ### ('lau', 0, 0) ('tunnel', 1, 4575) ('sp', 2, 6) ('##igen', 3, 25037) ('##art', 4, 17) ### ### ('##pf', 5, 26580) ('ramp', 6, 22740) ('schwarz', 7, 17802) ('##bahn', 8, 4891) ### ### ('##bruck', 9, 8901) ('##ess', 10, 9) ('##fa', 11, 5) ('##ch', 12, 18) ('##en', 13, 124) ### ### ('germany', 14, 38) ('##ko', 15, 27081) ('##ʂ', 16, 1901) ('german', 17, 62) ('motorway', 18, 1224) ### ### ('zurich', 19, 4781) ### ### =============================================== v_p ================================================ ### ### ('lau', 0, 0) ('hai', 7229, 1) ('##ffen', 4505, 2) ('franco', 4705, 3) ('##pes', 27671, 4) ### ### ('##fa', 11, 5) ('sp', 2, 6) ('bavaria', 129, 7) ('##cha', 4780, 8) ('##ess', 10, 9) ### ### ('##nh', 18813, 10) ('range', 5353, 11) ('##ile', 14784, 12) ('##sar', 21847, 13) ### ### ('worried', 125, 14) ('lower', 314, 15) ('vo', 6264, 16) ('##art', 4, 17) ('##ch', 12, 18) ### ### ('##fra', 8166, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('lau', 0, 0) ('sp', 2, 6) ('##fa', 11, 5) ('##art', 4, 17) ('##ess', 10, 9) ('##ch', 12, 18) ### ### ('germany', 14, 38) ('bavaria', 129, 7) ('leaned', 57, 28) ('worried', 125, 14) ('german', 17, 62) ### ### ('lower', 314, 15) ('##en', 13, 124) ('anyway', 247, 33) ('##burg', 479, 24) ('munich', 30, 109) ### ### ('bavarian', 370, 37) ('¹', 32, 113) ('infrastructure', 55, 76) ('##ffen', 4505, 2) ### ############################################################################################################ [2023-11-25 06:56:35,516][root][INFO] - Epoch: 2: Step: 6101/19530, loss[v]=0.137308, lr=0.000018, acc@1[1]=987.0/1024=0.9638671875, acc@1[2]=1009.0/1024=0.9853515625 [2023-11-25 06:58:25,368][root][INFO] - Train batch 6200 [2023-11-25 06:58:25,369][root][INFO] - Avg. loss per last 100 batches: 0.176836 [2023-11-25 06:58:26,096][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.5/29522=99.99% | mean: 0.03 | max: 5.71 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: according to legend, she performed miraculous healings and converted the georgian queen, ### ### nana, and eventually the pagan king mirian iii of iberia, who, lost in darkness and blinded on a ### ### hunting trip, found his way only after he prayed to " nino ’ s god ". ### ### [P_TEXT]: nana of iberia dedicated to " the most high god " on the occasion of the building of the ### ### jewish " prayer house ". alternatively, christian settipani identifies nana as a younger daughter ### ### of theothorses, a bosporan king. the medieval georgian sources relate that nana had been a staunch ### ### pagan and despised christian preaching until she was miraculously cured of a terrible disease, and ### ### subsequently converted, by a cappadocia n christian missionary, nino. the roman scholar tyrannius ### ### rufinus, writing his history half a century after the iberian conversion on the basis of ### ### =============================================== v_q ================================================ ### ### ('nino', 0, 1) ('nana', 1, 0) ('georgian', 2, 14) ('mir', 3, 25944) ('miraculous', 4, 8) ### ### ('iii', 5, 16359) ('healing', 6, 72) ('ib', 7, 12) ('queen', 8, 3264) ('absently', 9, 224) ### ### ('healer', 10, 2589) ('god', 11, 21) ('pagan', 12, 15) ('.', 13, 163) ('georgia', 14, 101) ### ### ('##ian', 15, 6567) ('darkness', 16, 4563) ('intently', 17, 490) ('legend', 18, 4909) ### ### ('journey', 19, 2187) ### ### =============================================== v_p ================================================ ### ### ('nana', 1, 0) ('nino', 0, 1) ('##thor', 27765, 2) ('##fin', 21344, 3) ('##tip', 22110, 4) ### ### ('##ses', 16691, 5) ('##eria', 26, 6) ('##oran', 25935, 7) ('miraculous', 4, 8) ('##sp', 14942, 9) ### ### ('##nius', 22056, 10) ('conversion', 3294, 11) ('ib', 7, 12) ('iberian', 40, 13) ### ### ('georgian', 2, 14) ('pagan', 12, 15) ('theo', 3240, 16) ('prayer', 41, 17) ('cured', 114, 18) ### ### ('##ani', 3469, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('nino', 0, 1) ('nana', 1, 0) ('georgian', 2, 14) ('miraculous', 4, 8) ('ib', 7, 12) ### ### ('pagan', 12, 15) ('god', 11, 21) ('##eria', 26, 6) ('iberian', 40, 13) ('prayer', 41, 17) ### ### ('miracles', 22, 49) ('healing', 6, 72) ('cured', 114, 18) ('tbilisi', 23, 69) ('georgia', 14, 101) ### ### ('worried', 382, 24) ('christianity', 193, 41) ('missionary', 312, 32) ('medieval', 425, 28) ### ### ('she', 96, 56) ### ############################################################################################################ [2023-11-25 06:58:26,602][root][INFO] - Epoch: 2: Step: 6201/19530, loss[v]=0.109403, lr=0.000018, acc@1[1]=992.5/1024=0.96923828125, acc@1[2]=1012.0/1024=0.98828125 [2023-11-25 07:00:16,515][root][INFO] - Train batch 6300 [2023-11-25 07:00:16,516][root][INFO] - Avg. loss per last 100 batches: 0.175154 [2023-11-25 07:00:17,236][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.67 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.7/29522=99.99% | mean: 0.03 | max: 5.29 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: his west germany career started with a youth call - up in 1961 that prompted sepp ### ### herberger to take him to the senior level a year later. ### ### [P_TEXT]: robert bernard ( footballer ) against norway eliminated the host nation from the ### ### tournament and got coach otto nerz sacked. nerz'successor sepp herberger, who later handed his son ### ### gunter caps for germany, did not see use in playing robert bernard any further. ### ### =============================================== v_q ================================================ ### ### ('herb', 0, 5) ('sep', 1, 9) ('##ger', 2, 11) ('1961', 3, 9782) ('germany', 4, 12) ('ɐ', 5, 155) ### ### ('youth', 6, 1881) ('west', 7, 17214) ('german', 8, 35) ('intently', 9, 387) ('¹', 10, 202) ### ### ('hockey', 11, 901) ('##ʂ', 12, 211) ('³', 13, 192) ('1960', 14, 7314) ('tennis', 15, 448) ### ### ('handball', 16, 1392) ('##ப', 17, 102) ('football', 18, 48) ('##¨', 19, 5702) ### ### =============================================== v_p ================================================ ### ### ('gunter', 620, 0) ('bernard', 1763, 1) ('caps', 6178, 2) ('robert', 6353, 3) ('otto', 2014, 4) ### ### ('herb', 0, 5) ('norway', 1099, 6) ('sacked', 5191, 7) ('littered', 20, 8) ('sep', 1, 9) ### ### ('##rz', 10107, 10) ('##ger', 2, 11) ('germany', 4, 12) ('ne', 13521, 13) ('sons', 14946, 14) ### ### ('tournament', 88, 15) ('coach', 46, 16) ('footballer', 38, 17) ('shirt', 218, 18) ### ### ('son', 5513, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('herb', 0, 5) ('sep', 1, 9) ('##ger', 2, 11) ('germany', 4, 12) ('littered', 20, 8) ### ### ('german', 8, 35) ('footballer', 38, 17) ('coach', 46, 16) ('##p', 37, 21) ('tournament', 88, 15) ### ### ('gunter', 620, 0) ('football', 18, 48) ('manager', 51, 27) ('ɐ', 5, 155) ('player', 23, 61) ### ### ('shirt', 218, 18) ('players', 27, 51) ('bernard', 1763, 1) ('norway', 1099, 6) ('team', 44, 54) ### ############################################################################################################ [2023-11-25 07:00:17,740][root][INFO] - Epoch: 2: Step: 6301/19530, loss[v]=0.204769, lr=0.000018, acc@1[1]=968.0/1024=0.9453125, acc@1[2]=992.0/1024=0.96875 [2023-11-25 07:02:07,897][root][INFO] - Train batch 6400 [2023-11-25 07:02:07,898][root][INFO] - Avg. loss per last 100 batches: 0.177199 [2023-11-25 07:02:08,628][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.2/29522=99.99% | mean: 0.03 | max: 5.58 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 2006 she published a standard work on the figurative coffins of the ga people. ### ### [P_TEXT]: fantasy coffin the christians the use of custom coffins is relatively new and began in ### ### the greater accra region around 1950. these coffins were formerly used only by the ga chiefs and ### ### priests, but since around 1960 figurative coffins have become an integral part of the local funeral ### ### culture. the christians had taken over the figurative coffins from the traditional ga who were ### ### using figurative palanquins and coffins already since the beginning of the 20th century. the ### ### invention of these figurative coffins used to be attributed to seth kane kwei, though the ### ### anthropologist roberta bonetti and especially regula tschumi question this myth. the idea of making ### ### and using custom coffins was inspired by the figurative palanquins in which ### ### =============================================== v_q ================================================ ### ### ('ga', 0, 8) ('coffin', 1, 0) ('figurative', 2, 13) ('people', 3, 1270) ('ghana', 4, 591) ### ### ('peoples', 5, 788) ('##子', 6, 160) ('work', 7, 8158) ('motif', 8, 287) ('intently', 9, 319) ### ### ('monique', 10, 797) ('2006', 11, 21163) ('.', 12, 1622) ('hissed', 13, 136) ('authors', 14, 16569) ### ### ('herself', 15, 49) ('vanessa', 16, 7880) ('¹', 17, 313) ('myanmar', 18, 11488) ### ### ('absently', 19, 272) ### ### =============================================== v_p ================================================ ### ### ('coffin', 1, 0) ('fantasy', 1654, 1) ('roberta', 1589, 2) ('acc', 17786, 3) ('chiefs', 1087, 4) ### ### ('custom', 1537, 5) ('kane', 14437, 6) ('greater', 7244, 7) ('ga', 0, 8) ('reg', 21388, 9) ### ### ('##quin', 19237, 10) ('seth', 4344, 11) ('funeral', 56, 12) ('figurative', 2, 13) ### ### ('##chu', 11825, 14) ('pal', 26995, 15) ('bone', 4426, 16) ('kw', 3484, 17) ('priests', 3697, 18) ### ### ('##ei', 17238, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('coffin', 1, 0) ('ga', 0, 8) ('figurative', 2, 13) ('funeral', 56, 12) ('she', 23, 45) ### ### ('culture', 75, 30) ('traditional', 54, 36) ('herself', 15, 49) ('anthropologist', 82, 32) ### ### ('chiefs', 1087, 4) ('fantasy', 1654, 1) ('roberta', 1589, 2) ('anyway', 208, 33) ### ### ('custom', 1537, 5) ('anthropology', 428, 26) ('##子', 6, 160) ('littered', 529, 34) ### ### ('decorative', 37, 68) ('region', 511, 39) ('hissed', 13, 136) ### ############################################################################################################ [2023-11-25 07:02:09,132][root][INFO] - Epoch: 2: Step: 6401/19530, loss[v]=0.088092, lr=0.000018, acc@1[1]=1001.0/1024=0.9775390625, acc@1[2]=1015.5/1024=0.99169921875 [2023-11-25 07:03:59,236][root][INFO] - Train batch 6500 [2023-11-25 07:03:59,237][root][INFO] - Avg. loss per last 100 batches: 0.168246 [2023-11-25 07:03:59,957][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.69 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.8/29522=99.99% | mean: 0.03 | max: 5.36 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the area in general is regarded by its own populace as one of the most prestigious ### ### locations in the birmingham area and even in central england ; a 2007 report by the website ### ### mouseprice. com placed two sutton coldfield streets amongst the 20 most expensive in ### ### [P_TEXT]: birmingham meaning " woodland clearing ". birmingham is dominated by the birmingham ### ### fault, which runs diagonally through the city from the lickey hills in the south west, passing ### ### through edgbaston and the bull ring, to erdington and sutton coldfield in the north east. to the ### ### south and east of the fault the ground is largely softer mercia mudstone, interspersed with beds of ### ### bunter pebbles and crossed by the valleys of the rivers tame, rea and cole and their tributaries. ### ### to ### ### =============================================== v_q ================================================ ### ### ('birmingham', 0, 0) ('mouse', 1, 5811) ('sutton', 2, 15) ('##field', 3, 34) ('##pr', 4, 28232) ### ### ('cold', 5, 19) ('england', 6, 1081) ('street', 7, 5895) ('expensive', 8, 2689) ('##ice', 9, 28050) ### ### ('com', 10, 24773) ('london', 11, 217) ('absently', 12, 124) ('streets', 13, 4523) ### ### ('central', 14, 10217) ('prestigious', 15, 11223) ('intently', 16, 142) ('staffordshire', 17, 607) ### ### ('midlands', 18, 109) ('mayfair', 19, 16055) ### ### =============================================== v_p ================================================ ### ### ('birmingham', 0, 0) ('fault', 18577, 1) ('##dington', 1251, 2) ('lick', 16596, 3) ### ### ('cole', 9544, 4) ('pebbles', 13510, 5) ('tame', 11280, 6) ('bull', 15195, 7) ('clearing', 3921, 8) ### ### ('ring', 9800, 9) ('diagonal', 920, 10) ('re', 20149, 11) ('##gb', 12567, 12) ('mud', 6155, 13) ### ### ('bun', 8636, 14) ('sutton', 2, 15) ('##stone', 1146, 16) ('woodland', 638, 17) ### ### ('##ast', 23890, 18) ('cold', 5, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('birmingham', 0, 0) ('sutton', 2, 15) ('cold', 5, 19) ('##field', 3, 34) ('##dington', 1251, 2) ### ### ('.', 21, 67) ('anyway', 319, 23) ('diagonal', 920, 10) ('woodland', 638, 17) ### ### ('neighbouring', 298, 29) ('absently', 12, 124) ('nearby', 186, 39) ('outskirts', 76, 51) ### ### ('midlands', 18, 109) ('##stone', 1146, 16) ('countryside', 328, 35) ('intently', 16, 142) ### ### ('london', 11, 217) ('clearing', 3921, 8) ('city', 289, 50) ### ############################################################################################################ [2023-11-25 07:04:00,463][root][INFO] - Epoch: 2: Step: 6501/19530, loss[v]=0.071115, lr=0.000018, acc@1[1]=1004.0/1024=0.98046875, acc@1[2]=1021.0/1024=0.9970703125 [2023-11-25 07:05:50,337][root][INFO] - Train batch 6600 [2023-11-25 07:05:50,338][root][INFO] - Avg. loss per last 100 batches: 0.163323 [2023-11-25 07:05:51,071][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.70 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29516.0/29522=99.98% | mean: 0.02 | max: 5.25 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the freikorps were also called the " black reichswehr " ( black army ) for they were ### ### a'secret'army outside the bounds of the versailles treaty. ### ### [P_TEXT]: black reichswehr black reichswehr ( ) was the name for the extra - legal paramilitary ### ### formations promoted by the german reichswehr army during the time of the weimar republic ; it was ### ### raised despite restrictions imposed by the versailles treaty. the secret organization was dissolved ### ### in 1923 upon the failed kustrin putsch. the versailles treaty restricted the numbers of the german ### ### army to seven divisions of infantry and three of cavalry, for a total of 100, 000 men, and no more ### ### than 4, 000 officers. conscription was prohibited, and civilian employees engaged in forest ### ### protection, customs inspection and ### ### =============================================== v_q ================================================ ### ### ('black', 0, 3) ('army', 1, 21) ('##eh', 2, 9) ('reich', 3, 14) ('secret', 4, 10) ('##ʂ', 5, 1982) ### ### ('fr', 6, 19764) ('##sw', 7, 1) ('versailles', 8, 13) ('absently', 9, 227) ('treaty', 10, 5) ### ### ('##kor', 11, 28425) ('german', 12, 22) ('intently', 13, 273) ('hitler', 14, 64) ### ### ('germany', 15, 31) ('##ei', 16, 15016) ('nuremberg', 17, 1833) ('.', 18, 84) ### ### ('wehrmacht', 19, 6695) ### ### =============================================== v_p ================================================ ### ### ('customs', 4373, 0) ('##sw', 7, 1) ('conscription', 791, 2) ('black', 0, 3) ### ### ('inspection', 9042, 4) ('treaty', 10, 5) ('puts', 11066, 6) ('dissolved', 2568, 7) ### ### ('cavalry', 261, 8) ('##eh', 2, 9) ('secret', 4, 10) ('1923', 76, 11) ('protection', 3734, 12) ### ### ('versailles', 8, 13) ('reich', 3, 14) ('restrictions', 4288, 15) ('weimar', 30, 16) ### ### ('##rin', 16152, 17) ('worried', 156, 18) ('leaned', 2731, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('black', 0, 3) ('##sw', 7, 1) ('##eh', 2, 9) ('army', 1, 21) ('reich', 3, 14) ('treaty', 10, 5) ### ### ('secret', 4, 10) ('versailles', 8, 13) ('german', 12, 22) ('germany', 15, 31) ('weimar', 30, 16) ### ### ('1923', 76, 11) ('hitler', 14, 64) ('worried', 156, 18) ('##r', 31, 39) ('anyway', 75, 29) ### ### ('cavalry', 261, 8) ('infantry', 113, 28) ('conscription', 791, 2) ('armies', 99, 35) ### ############################################################################################################ [2023-11-25 07:05:51,574][root][INFO] - Epoch: 2: Step: 6601/19530, loss[v]=0.081986, lr=0.000018, acc@1[1]=998.5/1024=0.97509765625, acc@1[2]=1015.5/1024=0.99169921875 [2023-11-25 07:07:42,131][root][INFO] - Train batch 6700 [2023-11-25 07:07:42,132][root][INFO] - Avg. loss per last 100 batches: 0.163314 [2023-11-25 07:07:42,850][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.83 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.4/29522=99.99% | mean: 0.03 | max: 5.58 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: they control the micronesian games, the nation's football championship and the micronesia ### ### national football team. ### ### [P_TEXT]: micronesian games the micronesia n games ( or " micro games ", " microgames " ) are a ### ### quadrennial international multi - sport event within the micronesian region. the games were first ### ### held in 1969 in saipan ( northern mariana islands ). the 2010 micronesian games were initially due ### ### to be held in majuro ( marshall islands ), until the hosts withdrew. the 2010 games were hosted by ### ### palau. the federated states of micronesia won the bidding to host the 2014 micronesian games in ### ### pohnpei state, and later ### ### =============================================== v_q ================================================ ### ### ('football', 0, 977) ('##nesia', 1, 31) ('##nesian', 2, 4) ('micro', 3, 3) ('games', 4, 1) ### ### ('intently', 5, 135) ('absently', 6, 154) ('national', 7, 9156) ('samoa', 8, 108) ('.', 9, 387) ### ### ('¹', 10, 250) ('ɐ', 11, 159) ('footballer', 12, 2598) ('pacific', 13, 167) ('oceania', 14, 482) ### ### ('islands', 15, 7) ('hawaii', 16, 58) ('seychelles', 17, 9920) ('papua', 18, 938) ('³', 19, 246) ### ### =============================================== v_p ================================================ ### ### ('##games', 850, 0) ('games', 4, 1) ('##uro', 23748, 2) ('micro', 3, 3) ('##nesian', 2, 4) ### ### ('quad', 6452, 5) ('marshall', 3042, 6) ('islands', 15, 7) ('leaned', 195, 8) ('worried', 135, 9) ### ### ('##hn', 12251, 10) ('maj', 22153, 11) ('bidding', 13001, 12) ('sai', 9003, 13) ('mariana', 83, 14) ### ### ('withdrew', 680, 15) ('##pan', 1739, 16) ('2010', 2309, 17) ('pal', 8223, 18) ('2014', 3453, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##nesian', 2, 4) ('micro', 3, 3) ('games', 4, 1) ('##nesia', 1, 31) ('islands', 15, 7) ### ### ('mariana', 83, 14) ('worried', 135, 9) ('island', 25, 39) ('leaned', 195, 8) ('intently', 5, 135) ### ### ('absently', 6, 154) ('hawaiian', 23, 47) ('anyway', 116, 24) ('hawaii', 16, 58) ### ### ('olympics', 57, 37) ('##games', 850, 0) ('samoa', 8, 108) ('football', 0, 977) ('cradled', 34, 56) ### ### ('game', 232, 29) ### ############################################################################################################ [2023-11-25 07:07:43,356][root][INFO] - Epoch: 2: Step: 6701/19530, loss[v]=0.141373, lr=0.000018, acc@1[1]=990.0/1024=0.966796875, acc@1[2]=1006.0/1024=0.982421875 [2023-11-25 07:09:33,686][root][INFO] - Train batch 6800 [2023-11-25 07:09:33,687][root][INFO] - Avg. loss per last 100 batches: 0.176122 [2023-11-25 07:09:34,401][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.1/29522=99.99% | mean: 0.03 | max: 5.55 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: commanders take place and precedence immediately after members and honorary members of ### ### the order of jamaica. ### ### [P_TEXT]: migonette patricia durrant post of president for years. she died on november 26, 2019 ### ### shortly after having been diagnosed with pancreatic cancer. durrant is the recipient of several ### ### awards for her career achievements. these are : the order of distinction of jamaica in the rank of ### ### commander ( c. d. ) in 1992 ; the order of jamaica ( o. j. ) by the government of jamaica for her ### ### unique achievements in respect of international affairs in 2000 ( the ideal of the order is " for a ### ### covenant of the people " }. ; and honoured as one of the university of the west indies's ### ### distinguished graduates on the occasion of ### ### =============================================== v_q ================================================ ### ### ('jamaica', 0, 2) ('order', 1, 10) ('jamaican', 2, 75) ('honorary', 3, 492) ('precedence', 4, 991) ### ### ('.', 5, 109) ('intently', 6, 546) ('commander', 7, 14) ('caribbean', 8, 77) ('absently', 9, 404) ### ### ('¹', 10, 408) ('monarch', 11, 370) ('honour', 12, 187) ('awarded', 13, 70) ('haiti', 14, 183) ### ### ('³', 15, 957) ('insignia', 16, 3953) ('barbados', 17, 149) ('trinidad', 18, 1188) ('rank', 19, 23) ### ### =============================================== v_p ================================================ ### ### ('covenant', 8517, 0) ('##rran', 28289, 1) ('jamaica', 0, 2) ('indies', 558, 3) ('mig', 29059, 4) ### ### ('patricia', 1917, 5) ('ideal', 3203, 6) ('##tte', 17460, 7) ('government', 377, 8) ### ### ('cancer', 16555, 9) ('order', 1, 10) ('university', 2215, 11) ('distinction', 287, 12) ### ### ('##one', 18713, 13) ('commander', 7, 14) ('awards', 407, 15) ('honoured', 174, 16) ### ### ('1992', 17153, 17) ('##atic', 3424, 18) ('distinguished', 1566, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('jamaica', 0, 2) ('order', 1, 10) ('commander', 7, 14) ('rank', 19, 23) ('jamaican', 2, 75) ### ### ('caribbean', 8, 77) ('indies', 558, 3) ('.', 5, 109) ('commanders', 36, 37) ('orders', 34, 43) ### ### ('honoured', 174, 16) ('government', 377, 8) ('awarded', 13, 70) ('distinction', 287, 12) ### ### ('recipient', 101, 27) ('awards', 407, 15) ('worried', 286, 25) ('president', 615, 22) ### ### ('patricia', 1917, 5) ('anyway', 207, 42) ### ############################################################################################################ [2023-11-25 07:09:34,906][root][INFO] - Epoch: 2: Step: 6801/19530, loss[v]=0.226192, lr=0.000018, acc@1[1]=970.0/1024=0.947265625, acc@1[2]=997.5/1024=0.97412109375 [2023-11-25 07:11:25,191][root][INFO] - Train batch 6900 [2023-11-25 07:11:25,192][root][INFO] - Avg. loss per last 100 batches: 0.165940 [2023-11-25 07:11:25,940][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.9/29522=99.99% | mean: 0.03 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: this town includes with some colonies like batika, berela, vivekananda pally ( formerly ### ### garoan para ), charabagan, nunia danga, telcopa, halder dighi etc. ### ### [P_TEXT]: batika batika is a census town in pandua cd block in chinsurah subdivision of hooghly ### ### district in the india n state of west bengal. batika is generated by combining shape of some ### ### colonies. they are purbapalli ( formerly goalapara ), roypara and majher para or muslim para. ### ### batika is located at 23. 1209035 n and 88. 2012919 e. it has an average elevation of 21 m ( 69 ft ### ### ). howrah - bardhaman main line passed through the edge of this town. road density is very good ### ### here. road conditions are excellent. boinchi - kalna road ( wb state highway no 11 ) also passed ### ### through the edge of this town. nearest railway ### ### =============================================== v_q ================================================ ### ### ('##ika', 0, 0) ('##rel', 1, 28195) ('para', 2, 10) ('##ro', 3, 26970) ('##aba', 4, 16858) ### ### ('tel', 5, 22383) ('bat', 6, 1) ('dig', 7, 14598) ('##gan', 8, 22896) ('char', 9, 7264) ### ### ('ga', 10, 26473) ('##co', 11, 26975) ('nun', 12, 25405) ('##pa', 13, 12681) ('nepal', 14, 143) ### ### ('colony', 15, 38) ('intently', 16, 222) ('##ve', 17, 23460) ('##der', 18, 23463) ### ### ('hal', 19, 24147) ### ### =============================================== v_p ================================================ ### ### ('##ika', 0, 0) ('bat', 6, 1) ('bard', 18887, 2) ('roy', 10145, 3) ('##ln', 25695, 4) ### ### ('goal', 7202, 5) ('subdivisions', 4726, 6) ('##hly', 25099, 7) ('##du', 17672, 8) ### ### ('##rah', 7380, 9) ('para', 2, 10) ('wb', 15148, 11) ('##para', 810, 12) ('chin', 9956, 13) ### ### ('colonies', 106, 14) ('##apa', 6704, 15) ('##pal', 7919, 16) ('##rba', 10104, 17) ### ### ('nearest', 1356, 18) ('##sur', 14560, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ika', 0, 0) ('bat', 6, 1) ('para', 2, 10) ('colony', 15, 38) ('india', 33, 29) ### ### ('leaned', 54, 31) ('colonies', 106, 14) ('worried', 173, 22) ('district', 60, 50) ('town', 31, 87) ### ### ('##ा', 58, 61) ('nepal', 14, 143) ('bengal', 79, 51) ('##para', 810, 12) ('##か', 41, 94) ### ### ('##a', 96, 62) ('स', 69, 75) ('anyway', 280, 39) ('odisha', 63, 84) ('intently', 16, 222) ### ############################################################################################################ [2023-11-25 07:11:26,443][root][INFO] - Epoch: 2: Step: 6901/19530, loss[v]=0.119368, lr=0.000018, acc@1[1]=995.0/1024=0.9716796875, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 07:13:17,157][root][INFO] - Train batch 7000 [2023-11-25 07:13:17,157][root][INFO] - Avg. loss per last 100 batches: 0.171415 [2023-11-25 07:13:17,890][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.9/29522=99.99% | mean: 0.02 | max: 5.28 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: kidney function, metformin, a biguanide, has become the most commonly used agent for type ### ### 2 diabetes in children and teenagers. ### ### [P_TEXT]: acquired generalized lipodystrophy excluding the possible lipodystrophy, but there is no ### ### well - established standardized leptin ranges. initial and general approach for agl patients are to ### ### treat the metabolic complications such as leptin - replacement therapy and / or to control the ### ### abnormal levels of lipid s or glucose levels. anti - diabetic medication s such as insulin, ### ### metformin, or thiazolidinedione s are used for insulin - resistance or high glucose levels, or ### ### statin s or fibrate s are used for hyperlipidemia. if symptoms persist, metreleptin can ### ### =============================================== v_q ================================================ ### ### ('##form', 0, 2) ('##uan', 1, 26796) ('diabetes', 2, 575) ('met', 3, 17) ('type', 4, 2037) ### ### ('kidney', 5, 1843) ('absently', 6, 203) ('big', 7, 16679) ('2', 8, 22305) ('intently', 9, 176) ### ### ('##ide', 10, 63) ('¹', 11, 207) ('##in', 12, 25) ('.', 13, 107) ('diagnosed', 14, 2415) ### ### ('tested', 15, 6516) ('pmid', 16, 563) ('drug', 17, 255) ('health', 18, 2510) ('³', 19, 431) ### ### =============================================== v_p ================================================ ### ### ('stat', 17437, 0) ('##brate', 3087, 1) ('##form', 0, 2) ('acquired', 2405, 3) ('lip', 6091, 4) ### ### ('generalized', 6684, 5) ('##lip', 26877, 6) ('ag', 15001, 7) ('symptoms', 351, 8) ### ### ('##bet', 3783, 9) ('##brates', 1221, 10) ('##yst', 26733, 11) ('resistance', 5761, 12) ### ### ('insulin', 105, 13) ('complications', 809, 14) ('metre', 11216, 15) ('metabolic', 513, 16) ### ### ('met', 3, 17) ('abnormal', 3315, 18) ('##pt', 28399, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##form', 0, 2) ('met', 3, 17) ('##in', 12, 25) ('medication', 32, 29) ('##ide', 10, 63) ### ### ('insulin', 105, 13) ('glucose', 30, 37) ('symptoms', 351, 8) ('levels', 214, 31) ### ### ('worried', 115, 45) ('anyway', 132, 43) ('absently', 6, 203) ('metabolic', 513, 16) ### ### ('intently', 9, 176) ('.', 13, 107) ('therapy', 312, 35) ('complications', 809, 14) ### ### ('diabetes', 2, 575) ('patient', 223, 46) ('##brates', 1221, 10) ### ############################################################################################################ [2023-11-25 07:13:18,398][root][INFO] - Epoch: 2: Step: 7001/19530, loss[v]=0.317757, lr=0.000018, acc@1[1]=953.5/1024=0.93115234375, acc@1[2]=992.0/1024=0.96875 [2023-11-25 07:15:08,645][root][INFO] - Train batch 7100 [2023-11-25 07:15:08,646][root][INFO] - Avg. loss per last 100 batches: 0.187598 [2023-11-25 07:15:09,363][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.56 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.2/29522=99.99% | mean: 0.03 | max: 5.41 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it is usually classified as a central algonquian language, with languages such as ojibwe, ### ### cree, menominee, miami - illinois, shawnee and fox. ### ### [P_TEXT]: list of place names of native american origin in michigan many places throughout the ### ### state of michigan take their names from native american indigenous languages. this list includes ### ### counties, townships, & settlements whose names are derived from indigenous languages in michigan. ### ### the primary native american languages in michigan are ojibwe, odawa, & potawatomi, all of which are ### ### dialects of algonquin. some other places names in michigan are found to be derived from sauk, ### ### oneida, wyandot, abenaki, shawnee, mohawk, seneca, seminole, iroquois ### ### =============================================== v_q ================================================ ### ### ('.', 0, 81) ('cree', 1, 12306) ('##ee', 2, 29) ('absently', 3, 119) ('miami', 4, 129) ### ### ('illinois', 5, 977) ('##qui', 6, 7417) ('intently', 7, 122) ('shawn', 8, 7) ('language', 9, 26) ### ### ('fox', 10, 14117) ('##gon', 11, 20) ('##nee', 12, 14895) ('al', 13, 31) ('¹', 14, 131) ### ### ('##omi', 15, 27679) ('##we', 16, 13) ('central', 17, 18773) ('³', 18, 427) ('guatemala', 19, 734) ### ### =============================================== v_p ================================================ ### ### ('seminole', 211, 0) ('##ida', 12039, 1) ('indigenous', 105, 2) ('michigan', 29, 3) ### ### ('##quin', 10911, 4) ('##uk', 13643, 5) ('mohawk', 146, 6) ('shawn', 8, 7) ('iroquois', 21, 8) ### ### ('abe', 19522, 9) ('##yan', 22202, 10) ('seneca', 56, 11) ('dialects', 983, 12) ('##we', 16, 13) ### ### ('##tom', 9451, 14) ('sa', 17616, 15) ('pot', 1827, 16) ('dialect', 202, 17) ('languages', 30, 18) ### ### ('##dot', 16454, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('shawn', 8, 7) ('##ee', 2, 29) ('##we', 16, 13) ('language', 9, 26) ('##gon', 11, 20) ### ### ('iroquois', 21, 8) ('michigan', 29, 3) ('al', 13, 31) ('languages', 30, 18) ('indigenous', 105, 2) ### ### ('seneca', 56, 11) ('native', 34, 25) ('.', 0, 81) ('seminole', 211, 0) ('##b', 23, 41) ### ### ('mohawk', 146, 6) ('##ji', 27, 44) ('anyway', 86, 22) ('absently', 3, 119) ('miami', 4, 129) ### ############################################################################################################ [2023-11-25 07:15:09,867][root][INFO] - Epoch: 2: Step: 7101/19530, loss[v]=0.235803, lr=0.000018, acc@1[1]=972.0/1024=0.94921875, acc@1[2]=1001.0/1024=0.9775390625 [2023-11-25 07:17:00,065][root][INFO] - Train batch 7200 [2023-11-25 07:17:00,065][root][INFO] - Avg. loss per last 100 batches: 0.177502 [2023-11-25 07:17:00,814][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 5.69 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: spinach is a good source ( 10 - 19 % of dv ) of the b vitamins riboflavin and vitamin b6, ### ### vitamin e, calcium, potassium, and dietary fiber ### ### [P_TEXT]: quinoa quinoa ( " chenopodium quinoa " ; or, from quechua'or') is a flowering plant in ### ### the amaranth family. it is a herbaceous annual plant grown as a crop primarily for its edible seed ### ### s ; the seeds are rich in protein, dietary fiber, b vitamins, and dietary mineral s in amounts ### ### greater than in many grains. quinoa is not a grass, but rather a pseudocereal botanically related ### ### to spinach and amaranth ( " amaranthus spp. " ), and originated in the andean region of ### ### northwestern south america. it ### ### =============================================== v_q ================================================ ### ### ('##ach', 0, 6) ('spin', 1, 5) ('ד', 2, 1653) ('b', 3, 28) ('##of', 4, 13627) ('absently', 5, 83) ### ### ('potassium', 6, 3649) ('##lav', 7, 29297) ('intently', 8, 263) ('ingredient', 9, 328) ### ### ('calcium', 10, 1201) ('rib', 11, 22704) ('fiber', 12, 51) ('¹', 13, 193) ('.', 14, 228) ### ### ('d', 15, 8621) ('vitamin', 16, 18) ('##6', 17, 18188) ('dietary', 18, 1) ('##v', 19, 25935) ### ### =============================================== v_p ================================================ ### ### ('amara', 3071, 0) ('dietary', 18, 1) ('##ean', 14715, 2) ('qui', 21844, 3) ('##chua', 18864, 4) ### ### ('spin', 1, 5) ('##ach', 0, 6) ('seed', 614, 7) ('edible', 114, 8) ('##aceous', 208, 9) ### ### ('##eal', 6233, 10) ('mineral', 1922, 11) ('pseudo', 13452, 12) ('grains', 1950, 13) ### ### ('##cer', 26226, 14) ('chen', 7335, 15) ('##no', 25782, 16) ('grain', 259, 17) ('vitamin', 16, 18) ### ### ('crop', 13113, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ach', 0, 6) ('spin', 1, 5) ('dietary', 18, 1) ('b', 3, 28) ('vitamin', 16, 18) ### ### ('fiber', 12, 51) ('edible', 114, 8) ('##aceous', 208, 9) ('absently', 5, 83) ('worried', 158, 26) ### ### ('seed', 614, 7) ('grain', 259, 17) ('food', 37, 44) ('anyway', 144, 32) ('protein', 210, 27) ### ### ('##®', 30, 63) ('amara', 3071, 0) ('¹', 13, 193) ('plant', 630, 23) ('seeds', 191, 40) ### ############################################################################################################ [2023-11-25 07:17:01,319][root][INFO] - Epoch: 2: Step: 7201/19530, loss[v]=0.269058, lr=0.000018, acc@1[1]=957.0/1024=0.9345703125, acc@1[2]=990.0/1024=0.966796875 [2023-11-25 07:18:51,994][root][INFO] - Train batch 7300 [2023-11-25 07:18:51,995][root][INFO] - Avg. loss per last 100 batches: 0.173701 [2023-11-25 07:18:52,728][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.63 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.8/29522=99.99% | mean: 0.03 | max: 5.51 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: dark energy and the casimir effect shows any such energy to be exceptionally weak. ### ### [P_TEXT]: casimir effect electrodynamic vacuum, it is seen that the plates do affect the virtual ### ### photons which constitute the field, and generate a net force – either an attraction or a repulsion ### ### depending on the specific arrangement of the two plates. although the casimir effect can be ### ### expressed in terms of virtual particles interacting with the objects, it is best described and more ### ### easily calculated in terms of the zero - point energy of a quantized field in the intervening space ### ### between the objects. this force has been measured and is a striking example of an effect ### ### =============================================== v_q ================================================ ### ### ('dark', 0, 6513) ('casimir', 1, 0) ('energy', 2, 11) ('weak', 3, 591) ('effect', 4, 3) ### ### ('.', 5, 81) ('intently', 6, 404) ('absently', 7, 354) ('physics', 8, 67) ('darker', 9, 10994) ### ### ('##¨', 10, 2481) ('¹', 11, 125) ('exceptional', 12, 767) ('atom', 13, 1706) ### ### ('electromagnetic', 14, 192) ('experiment', 15, 2638) ('radiation', 16, 43) ('universe', 17, 1008) ### ### ('extremely', 18, 3949) ('electron', 19, 98) ### ### =============================================== v_p ================================================ ### ### ('casimir', 1, 0) ('attraction', 181, 1) ('quan', 3135, 2) ('effect', 4, 3) ('vacuum', 36, 4) ### ### ('##ulsion', 10368, 5) ('zero', 411, 6) ('net', 3294, 7) ('virtual', 10620, 8) ('photon', 84, 9) ### ### ('##tized', 10911, 10) ('energy', 2, 11) ('field', 119, 12) ('interaction', 412, 13) ### ### ('particles', 203, 14) ('electro', 85, 15) ('plate', 2738, 16) ('rep', 19158, 17) ### ### ('point', 295, 18) ('interactions', 5933, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('casimir', 1, 0) ('effect', 4, 3) ('energy', 2, 11) ('vacuum', 36, 4) ('attraction', 181, 1) ### ### ('photon', 84, 9) ('electro', 85, 15) ('.', 5, 81) ('field', 119, 12) ('zero', 411, 6) ### ### ('radiation', 16, 43) ('particle', 71, 24) ('particles', 203, 14) ('anyway', 61, 32) ### ### ('point', 295, 18) ('physics', 8, 67) ('interaction', 412, 13) ('worried', 87, 35) ### ### ('space', 98, 39) ('fields', 501, 23) ### ############################################################################################################ [2023-11-25 07:18:53,234][root][INFO] - Epoch: 2: Step: 7301/19530, loss[v]=0.086459, lr=0.000018, acc@1[1]=998.0/1024=0.974609375, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 07:20:43,245][root][INFO] - Train batch 7400 [2023-11-25 07:20:43,246][root][INFO] - Avg. loss per last 100 batches: 0.178603 [2023-11-25 07:20:43,972][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.1/29522=99.99% | mean: 0.03 | max: 5.37 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.1/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the contract is under the nextstep2 program. ### ### [P_TEXT]: deep space habitat. early preliminary concepts considered 60 - day and 500 - day mission ### ### configurations, composed of international space station - derived hardware, the orion crew capsule ### ### and various support craft. the habitat would be equipped with at least one international docking ### ### system standard ( idss ) docking system. developing a deep space habitat would allow a crew to live ### ### and work safely in space for about one year on missions to explore cislunar space, mars, and some ### ### near - earth asteroids. in 2015 nasa funded studies for several types of deep space habitat ### ### concepts under the next space technologies for exploration partnerships ( nextstep ). lockheed ### ### martin, the main contractor of the orion capsule, has also ### ### =============================================== v_q ================================================ ### ### ('##ste', 0, 7) ('next', 1, 8) ('##p', 2, 52) ('##2', 3, 20380) ('intently', 4, 142) ('¹', 5, 236) ### ### ('！', 6, 1530) ('absently', 7, 214) ('contract', 8, 2120) ('program', 9, 1308) ('³', 10, 409) ### ### ('consortium', 11, 586) ('##₈', 12, 264) ('2020', 13, 19914) ('airbus', 14, 1666) ('ร', 15, 973) ### ### ('.', 16, 121) ('／', 17, 779) ('project', 18, 1596) ('erebidae', 19, 854) ### ### =============================================== v_p ================================================ ### ### ('habitat', 9643, 0) ('docking', 14228, 1) ('deep', 13437, 2) ('orion', 6161, 3) ### ### ('capsule', 1080, 4) ('##lun', 27670, 5) ('asteroids', 13394, 6) ('##ste', 0, 7) ('next', 1, 8) ### ### ('id', 13698, 9) ('space', 351, 10) ('martin', 19039, 11) ('lockheed', 375, 12) ### ### ('technologies', 901, 13) ('cis', 7156, 14) ('partnerships', 2959, 15) ('crew', 2419, 16) ### ### ('mars', 14971, 17) ('concepts', 5592, 18) ('##ss', 14494, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ste', 0, 7) ('next', 1, 8) ('##p', 2, 52) ('anyway', 38, 29) ('system', 61, 30) ### ### ('space', 351, 10) ('nasa', 65, 39) ('worried', 154, 24) ('capsule', 1080, 4) ('2015', 93, 36) ### ### ('contractor', 51, 43) ('lockheed', 375, 12) ('intently', 4, 142) ('partnership', 345, 21) ### ### ('technologies', 901, 13) ('contractors', 289, 37) ('¹', 5, 236) ('software', 22, 72) ### ### ('absently', 7, 214) ('mission', 122, 54) ### ############################################################################################################ [2023-11-25 07:20:44,475][root][INFO] - Epoch: 2: Step: 7401/19530, loss[v]=0.233831, lr=0.000018, acc@1[1]=966.0/1024=0.943359375, acc@1[2]=994.5/1024=0.97119140625 [2023-11-25 07:22:34,728][root][INFO] - Train batch 7500 [2023-11-25 07:22:34,728][root][INFO] - Avg. loss per last 100 batches: 0.181564 [2023-11-25 07:22:35,476][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.70 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.6/29522=99.99% | mean: 0.03 | max: 5.32 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the study of roman religion and myth is complicated by the early influence of greek ### ### religion on the italian peninsula during rome's protohistory, and by the later artistic imitation ### ### of greek literary models by roman authors. ### ### [P_TEXT]: star of the more prominent individual stars were also given names, particularly with ### ### arabic or latin designations. as well as certain constellations and the sun itself, individual ### ### stars have their own myths. to the ancient greeks, some " stars ", known as planet s ( greek ### ### πλανητης ( planetes ), meaning " wanderer " ), represented various important deities, from which ### ### the names of the planets mercury, venus, mars, jupiter and saturn were taken. ( uranus and neptune ### ### were also greek and roman gods, but neither planet was known in antiquity because of their low ### ### brightness. their names were assigned by later ### ### =============================================== v_q ================================================ ### ### ('roman', 0, 29) ('greek', 1, 14) ('religion', 2, 281) ('##his', 3, 25307) ('mythology', 4, 103) ### ### ('absently', 5, 369) ('.', 6, 34) ('intently', 7, 110) ('italian', 8, 340) ('proto', 9, 15626) ### ### ('peninsula', 10, 5582) ('literary', 11, 7773) ('imitation', 12, 2867) ('rome', 13, 69) ### ### ('greece', 14, 65) ('¹', 15, 108) ('myth', 16, 9) ('##tory', 17, 19398) ('italy', 18, 1174) ### ### ('literature', 19, 3642) ### ### =============================================== v_p ================================================ ### ### ('star', 5099, 0) ('wander', 10567, 1) ('deities', 6424, 2) ('planets', 17966, 3) ### ### ('myths', 5445, 4) ('neptune', 2800, 5) ('arabic', 1569, 6) ('brightness', 15287, 7) ### ### ('planet', 2238, 8) ('myth', 16, 9) ('mercury', 910, 10) ('greeks', 30, 11) ('stars', 13452, 12) ### ### ('names', 1166, 13) ('greek', 1, 14) ('##anus', 10968, 15) ('venus', 1479, 16) ('anyway', 98, 17) ### ### ('ur', 2228, 18) ('designations', 5041, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('greek', 1, 14) ('roman', 0, 29) ('myth', 16, 9) ('.', 6, 34) ('greeks', 30, 11) ### ### ('ancient', 21, 27) ('antiquity', 22, 31) ('god', 39, 28) ('rome', 13, 69) ('greece', 14, 65) ### ### ('worried', 69, 22) ('mythology', 4, 103) ('anyway', 98, 17) ('gods', 102, 21) ('intently', 7, 110) ### ### ('romans', 97, 24) ('religion', 2, 281) ('¹', 15, 108) ('arabic', 1569, 6) ('deity', 388, 20) ### ############################################################################################################ [2023-11-25 07:22:35,982][root][INFO] - Epoch: 2: Step: 7501/19530, loss[v]=0.233650, lr=0.000018, acc@1[1]=968.0/1024=0.9453125, acc@1[2]=987.5/1024=0.96435546875 [2023-11-25 07:24:26,271][root][INFO] - Train batch 7600 [2023-11-25 07:24:26,271][root][INFO] - Avg. loss per last 100 batches: 0.174253 [2023-11-25 07:24:26,986][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.75 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.4/29522=99.99% | mean: 0.03 | max: 5.52 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.6/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: due to his focus on gathering empirical documentation of unwritten languages through ### ### fieldwork and training of native linguists, kaufman's list of publications is less extensive than ### ### those of other scholars in the field though many of his articles were co - authored with other ### ### scholars such as lyle campbell, sarah thomason and john justeson. ### ### [P_TEXT]: quechumaran languages quechumaran or kechumaran is a language - family proposal that ### ### unites quechua and aymara. quechuan languages, especially those of the south, share a large amount ### ### of vocabulary with aymara. terrence kaufman finds the proposal reasonably convincing, but willem ### ### adelaar, a quechua specialist, believes the similarities to be caused by borrowing during long - ### ### term contact. lyle campbell suspects that the proposal is valid but does not consider it to have ### ### been conclusively proved. moulian " et al. " ( 2015 ) posits the puquina language of the tiwanaku ### ### empire as a possible source for ### ### =============================================== v_q ================================================ ### ### ('kaufman', 0, 0) ('##eson', 1, 28828) ('lyle', 2, 4) ('campbell', 3, 7) ('##written', 4, 6822) ### ### ('just', 5, 16814) ('sarah', 6, 13067) ('thomas', 7, 12103) ('linguist', 8, 173) ('##on', 9, 19588) ### ### ('john', 10, 15166) ('native', 11, 1447) ('linguistics', 12, 116) ('languages', 13, 13) ### ### ('language', 14, 27) ('documentation', 15, 2834) ('absently', 16, 322) ('scholars', 17, 784) ### ### ('anthropologist', 18, 17997) ('indigenous', 19, 96) ### ### =============================================== v_p ================================================ ### ### ('kaufman', 0, 0) ('##wana', 15224, 1) ('##mara', 25201, 2) ('##quin', 11051, 3) ('lyle', 2, 4) ### ### ('willem', 25030, 5) ('##ulia', 25466, 6) ('campbell', 3, 7) ('##chua', 395, 8) ### ### ('contact', 1171, 9) ('que', 10839, 10) ('pu', 23132, 11) ('terre', 11979, 12) ### ### ('languages', 13, 13) ('##ym', 15843, 14) ('##ara', 23024, 15) ('##ku', 20679, 16) ### ### ('term', 309, 17) ('vocabulary', 339, 18) ('##ela', 24229, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('kaufman', 0, 0) ('lyle', 2, 4) ('campbell', 3, 7) ('languages', 13, 13) ('language', 14, 27) ### ### ('linguistics', 12, 116) ('linguist', 8, 173) ('##chua', 395, 8) ('term', 309, 17) ### ### ('vocabulary', 339, 18) ('indigenous', 19, 96) ('intrigued', 21, 95) ('contact', 1171, 9) ### ### ('anyway', 509, 37) ('specialist', 1285, 21) ('varieties', 81, 65) ('linguistic', 41, 103) ### ### ('absently', 16, 322) ('littered', 179, 55) ('experts', 271, 50) ### ############################################################################################################ [2023-11-25 07:24:27,491][root][INFO] - Epoch: 2: Step: 7601/19530, loss[v]=0.294186, lr=0.000018, acc@1[1]=948.0/1024=0.92578125, acc@1[2]=985.0/1024=0.9619140625 [2023-11-25 07:26:17,375][root][INFO] - Train batch 7700 [2023-11-25 07:26:17,376][root][INFO] - Avg. loss per last 100 batches: 0.184628 [2023-11-25 07:26:18,102][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.46 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.4/29522=99.99% | mean: 0.03 | max: 5.32 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.2/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: after the rule of pachacuti, when an inca died, his title went to one son and his ### ### property was given to a corporation controlled by his other relatives ( split inheritance ). ### ### [P_TEXT]: manco capac remained in the city until the reign of pachacuti, who ordered its move to ### ### the tiwanaku temple in lake titicaca. in cusco there remained only a statue erected in his honor. ### ### manco capac is the protagonist of the two main legends that explain the origin of the inca empire. ### ### both legends state that he was the founder of the city of cusco and that his wife was mama uqllu. ### ### in this legend, manco capac ( ayar manco ) was the son of viracocha of paqariq tampu ( six league s ### ### or 25 km south of cusco ). he and his brothers ( ayar auca, ayar ### ### =============================================== v_q ================================================ ### ### ('inca', 0, 6) ('split', 1, 16293) ('##cut', 2, 0) ('corporation', 3, 14974) ### ### ('inheritance', 4, 4121) ('intently', 5, 158) ('absently', 6, 228) ('¹', 7, 139) ('pac', 8, 17) ### ### ('##ha', 9, 52) ('son', 10, 46) ('.', 11, 2121) ('³', 12, 549) ('peru', 13, 114) ### ### ('kinship', 14, 311) ('rituals', 15, 309) ('guatemala', 16, 255) ('ந', 17, 168) ('garion', 18, 232) ### ### ('ecuador', 19, 222) ### ### =============================================== v_p ================================================ ### ### ('##cut', 2, 0) ('##wana', 10515, 1) ('##yar', 15985, 2) ('##tica', 17290, 3) ('league', 19167, 4) ### ### ('temple', 4995, 5) ('inca', 0, 6) ('lake', 6506, 7) ('##co', 6044, 8) ('cu', 10886, 9) ### ### ('au', 17478, 10) ('##ac', 22422, 11) ('##pu', 4123, 12) ('mama', 11683, 13) ('##qa', 9298, 14) ### ### ('##och', 11489, 15) ('man', 5222, 16) ('pac', 8, 17) ('##ll', 16446, 18) ('##rac', 2544, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##cut', 2, 0) ('inca', 0, 6) ('pac', 8, 17) ('son', 10, 46) ('##ha', 9, 52) ('sons', 42, 35) ### ### ('anyway', 105, 29) ('intently', 5, 158) ('¹', 7, 139) ('spanish', 36, 50) ('peruvian', 22, 57) ### ### ('absently', 6, 228) ('shrug', 27, 55) ('peru', 13, 114) ('worried', 184, 34) ('empire', 379, 31) ### ### ('aztec', 24, 76) ('legend', 436, 33) ('mexico', 102, 64) ('legends', 1461, 25) ### ############################################################################################################ [2023-11-25 07:26:18,606][root][INFO] - Epoch: 2: Step: 7701/19530, loss[v]=0.224027, lr=0.000018, acc@1[1]=966.0/1024=0.943359375, acc@1[2]=1001.0/1024=0.9775390625 [2023-11-25 07:28:08,865][root][INFO] - Train batch 7800 [2023-11-25 07:28:08,866][root][INFO] - Avg. loss per last 100 batches: 0.184705 [2023-11-25 07:28:09,610][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.62 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.5/29522=99.99% | mean: 0.03 | max: 5.31 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in october 2018 the title was bestowed upon lakshman kiriella, the minister of public ### ### enterprises and kandy development and current leader of the house, for his service to buddhism and ### ### society. ### ### [P_TEXT]: lakshman kiriella he has been a member ever since. he has served as the minister of ### ### tourism and sports and the deputy minister of foreign affairs. in 2015 he was appointed as the ### ### leader of the house of the united national party government under prime minister ranil ### ### wickramasinghe. he is married to charmaine jayawardena and they have two daughters, kishanee and ### ### chami, and one son, dinuk l. b. in october 2018 the asgiriya chapter of siam nikaya conferred the ### ### title of shasanamamaka jana prasadini on kiriella, in recognition of his service to buddhism ### ### =============================================== v_q ================================================ ### ### ('house', 0, 14) ('ki', 1, 18) ('leader', 2, 61) ('buddhism', 3, 8) ('enterprises', 4, 2110) ### ### ('##lla', 5, 15) ('##rie', 6, 5) ('kan', 7, 7873) ('##hman', 8, 4) ('##dy', 9, 24120) ### ### ('##ks', 10, 26) ('awarded', 11, 243) ('absently', 12, 409) ('minister', 13, 45) ### ### ('intently', 14, 348) ('development', 15, 2421) ('madagascar', 16, 228) ('bestowed', 17, 366) ### ### ('leaders', 18, 81) ('myanmar', 19, 416) ### ### =============================================== v_p ================================================ ### ### ('##giri', 11965, 0) ('rani', 17654, 1) ('din', 13469, 2) ('united', 14393, 3) ('##hman', 8, 4) ### ### ('##rie', 6, 5) ('nik', 10981, 6) ('##ward', 16588, 7) ('buddhism', 3, 8) ('##shan', 9082, 9) ### ### ('jana', 5596, 10) ('siam', 68, 11) ('prasad', 2458, 12) ('foreign', 2670, 13) ('house', 0, 14) ### ### ('##lla', 5, 15) ('##aya', 13234, 16) ('charm', 12095, 17) ('ki', 1, 18) ('littered', 2160, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('house', 0, 14) ('buddhism', 3, 8) ('##rie', 6, 5) ('##hman', 8, 4) ('ki', 1, 18) ('##lla', 5, 15) ### ### ('##ks', 10, 26) ('minister', 13, 45) ('leader', 2, 61) ('buddhist', 26, 28) ('thailand', 29, 32) ### ### ('conferred', 35, 35) ('siam', 68, 11) ('leaders', 18, 81) ('anyway', 104, 41) ('la', 24, 84) ### ### ('2018', 34, 71) ('thai', 115, 40) ('government', 334, 25) ('awarded', 11, 243) ### ############################################################################################################ [2023-11-25 07:28:10,117][root][INFO] - Epoch: 2: Step: 7801/19530, loss[v]=0.165994, lr=0.000018, acc@1[1]=983.5/1024=0.96044921875, acc@1[2]=1003.0/1024=0.9794921875 [2023-11-25 07:30:00,512][root][INFO] - Train batch 7900 [2023-11-25 07:30:00,513][root][INFO] - Avg. loss per last 100 batches: 0.173914 [2023-11-25 07:30:01,234][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.80 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.6/29522=99.99% | mean: 0.02 | max: 5.16 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in may 1946 it was one of the parties that established the bloc of democratic parties, ### ### alongside the romanian social democratic party, the national liberal party – tatarescu, the ### ### ploughmen's front, the romanian communist party and the national popular party. ### ### [P_TEXT]: 1946 romanian general election breaking with the traditional universal male suffrage ### ### confirmed by the 1923 constitution, it was the first national election to feature women's suffrage, ### ### and the first to allow active public officials and army personnel the right to vote. the bpd, ### ### representing the incumbent leftist government formed around prime minister petru groza, was an ### ### electoral alliance comprising the pcr, the social democratic party ( psd ), the ploughmen's front, ### ### the national liberal party – tatarescu ( pnl – tatarescu ), the national peasants'party – ### ### alexandrescu ( pnt – alexandrescu ) and the national popular party. according to official results, ### ### the bpd won 69. 8 % of the vote, enough for an overwhelming ### ### =============================================== v_q ================================================ ### ### ('romanian', 0, 4) ('romania', 1, 10) ('party', 2, 22) ('1946', 3, 1) ('##ough', 4, 24) ### ### ('liberal', 5, 25) ('bloc', 6, 3791) ('.', 7, 111) ('intently', 8, 196) ('absently', 9, 245) ### ### ('front', 10, 29) ('parties', 11, 163) ('tatar', 12, 3) ('popular', 13, 12) ('social', 14, 33) ### ### ('communist', 15, 99) ('##¨', 16, 675) ('¹', 17, 118) ('democratic', 18, 197) ('national', 19, 27) ### ### =============================================== v_p ================================================ ### ### ('##oza', 23136, 0) ('1946', 3, 1) ('bp', 12441, 2) ('tatar', 12, 3) ('romanian', 0, 4) ### ### ('peasants', 326, 5) ('1923', 103, 6) ('alexandre', 1750, 7) ('constitution', 1254, 8) ### ### ('littered', 597, 9) ('romania', 1, 10) ('personnel', 1200, 11) ('popular', 13, 12) ### ### ('suffrage', 3479, 13) ('anyway', 62, 14) ('leftist', 71, 15) ('women', 3162, 16) ### ### ('elections', 74, 17) ('ps', 1342, 18) ('##nl', 13285, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('romanian', 0, 4) ('1946', 3, 1) ('romania', 1, 10) ('tatar', 12, 3) ('party', 2, 22) ### ### ('##ough', 4, 24) ('liberal', 5, 25) ('popular', 13, 12) ('front', 10, 29) ('social', 14, 33) ### ### ('national', 19, 27) ('1945', 26, 20) ('1923', 103, 6) ('##escu', 30, 36) ('anyway', 62, 14) ### ### ('##men', 25, 43) ('pl', 41, 28) ('leftist', 71, 15) ('elections', 74, 17) ('1947', 27, 47) ### ############################################################################################################ [2023-11-25 07:30:01,738][root][INFO] - Epoch: 2: Step: 7901/19530, loss[v]=0.247156, lr=0.000018, acc@1[1]=960.0/1024=0.9375, acc@1[2]=989.5/1024=0.96630859375 [2023-11-25 07:31:51,740][root][INFO] - Train batch 8000 [2023-11-25 07:31:51,741][root][INFO] - Avg. loss per last 100 batches: 0.179769 [2023-11-25 07:31:52,469][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.5/29522=99.99% | mean: 0.03 | max: 5.61 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: by the early 17th century, the island and its smaller neighbors ( notably tortuga ) ### ### became regular stopping points for caribbean pirates. ### ### [P_TEXT]: tortuga ( cocktail ) a tortuga cocktail is a non - alcoholic beverage named for the haiti ### ### an island tortuga of northern hispaniola that was used by pirates as launching ground for piracy ### ### activities. it is made with iced tea and brown sugar, garnished with cinnamon and a lime wedge. it ### ### is similar to the popular american drink sweet tea. an unrelated drink of the same name was served ### ### by trader vic's in the 1930s. it was a variation on the daiquiri, made using overproof guyanese and ### ### cuban rums and small amounts ### ### =============================================== v_q ================================================ ### ### ('tor', 0, 8) ('##tu', 1, 12) ('##ga', 2, 14) ('caribbean', 3, 157) ('pirates', 4, 3) ### ### ('absently', 5, 242) ('.', 6, 159) ('island', 7, 29) ('##ᵉ', 8, 918) ('intently', 9, 168) ### ### ('pirate', 10, 21) ('17th', 11, 7075) ('nautical', 12, 1166) ('¹', 13, 294) ('mosquito', 14, 1697) ### ### ('stopping', 15, 11633) ('antigua', 16, 212) ('criminals', 17, 2546) ('dangers', 18, 2299) ### ### ('jamaica', 19, 199) ### ### =============================================== v_p ================================================ ### ### ('vic', 16207, 0) ('wedge', 10947, 1) ('##proof', 11794, 2) ('pirates', 4, 3) ('iced', 26352, 4) ### ### ('piracy', 24, 5) ('tea', 6418, 6) ('cocktail', 3027, 7) ('tor', 0, 8) ('sweet', 2291, 9) ### ### ('haiti', 69, 10) ('guy', 18132, 11) ('##tu', 1, 12) ('cuban', 207, 13) ('##ga', 2, 14) ### ### ('brown', 3906, 15) ('cinnamon', 813, 16) ('lime', 5816, 17) ('worried', 93, 18) ### ### ('##qui', 5975, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('tor', 0, 8) ('##tu', 1, 12) ('##ga', 2, 14) ('pirates', 4, 3) ('pirate', 10, 21) ### ### ('island', 7, 29) ('piracy', 24, 5) ('haiti', 69, 10) ('worried', 93, 18) ('cuba', 26, 40) ### ### ('cuban', 207, 13) ('anyway', 56, 33) ('islands', 23, 57) ('sugar', 211, 24) ('caribbean', 3, 157) ### ### ('traders', 224, 34) ('cinnamon', 813, 16) ('.', 6, 159) ('activities', 363, 30) ('rum', 783, 20) ### ############################################################################################################ [2023-11-25 07:31:52,975][root][INFO] - Epoch: 2: Step: 8001/19530, loss[v]=0.259787, lr=0.000018, acc@1[1]=961.5/1024=0.93896484375, acc@1[2]=991.5/1024=0.96826171875 [2023-11-25 07:33:43,395][root][INFO] - Train batch 8100 [2023-11-25 07:33:43,396][root][INFO] - Avg. loss per last 100 batches: 0.179724 [2023-11-25 07:33:44,132][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.9/29522=99.99% | mean: 0.03 | max: 5.46 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: barrackpore is bounded by north barrackpur and jafarpur on the north, mohanpur, chak ### ### kanthalia, ruiya and patulia on the east, titagarh on the south, and the hooghly and barrackpur ### ### cantonment on the west. ### ### [P_TEXT]: titagarh in 2011, it had a population density of 10, 967 per km2. the subdivision has 16 ### ### municipalities and 24 census towns. titagarh is bounded by barrackpore on the north, khardaha on ### ### the east and south and hooghly river on the west. as per the 2011 census of india, titagarh had a ### ### total population of 116, 541, of which 62, 735 ( 54 % ) were males and 53, 806 ( 46 % ) were ### ### females. population below 6 years was 11, 510. the total number of literates in titagarh was 84, ### ### 160 ( 80. 13 % of the population over 6 years ). india census ### ### =============================================== v_q ================================================ ### ### ('barr', 0, 1) ('##ack', 1, 14) ('##pur', 2, 1366) ('absently', 3, 109) ('##pore', 4, 8) ### ### ('intently', 5, 117) ('india', 6, 17) ('ru', 7, 24824) ('.', 8, 155) ('##tag', 9, 0) ('¹', 10, 123) ### ### ('pakistan', 11, 70) ('kolkata', 12, 55) ('##ulia', 13, 27267) ('ந', 14, 104) ('ti', 15, 4) ### ### ('mumbai', 16, 110) ('গ', 17, 189) ('bangladesh', 18, 95) ('cha', 19, 21743) ### ### =============================================== v_p ================================================ ### ### ('##tag', 9, 0) ('barr', 0, 1) ('##hard', 24109, 2) ('##aha', 20841, 3) ('ti', 15, 4) ### ### ('subdivisions', 1785, 5) ('municipalities', 10809, 6) ('census', 1839, 7) ('##pore', 4, 8) ### ### ('ho', 25, 9) ('##ar', 167, 10) ('rivers', 5991, 11) ('anyway', 236, 12) ('worried', 118, 13) ### ### ('##ack', 1, 14) ('river', 2267, 15) ('bengali', 77, 16) ('india', 6, 17) ('males', 1278, 18) ### ### ('##hly', 233, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('barr', 0, 1) ('##tag', 9, 0) ('##ack', 1, 14) ('##pore', 4, 8) ('ti', 15, 4) ('india', 6, 17) ### ### ('ho', 25, 9) ('outskirts', 22, 23) ('absently', 3, 109) ('kolkata', 12, 55) ('bengali', 77, 16) ### ### ('##ar', 167, 10) ('calcutta', 36, 30) ('worried', 118, 13) ('intently', 5, 117) ('shrug', 38, 31) ### ### ('pakistan', 11, 70) ('##og', 31, 40) ('₹', 35, 39) ('##h', 84, 22) ### ############################################################################################################ [2023-11-25 07:33:44,638][root][INFO] - Epoch: 2: Step: 8101/19530, loss[v]=0.090506, lr=0.000018, acc@1[1]=997.0/1024=0.9736328125, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 07:35:34,856][root][INFO] - Train batch 8200 [2023-11-25 07:35:34,857][root][INFO] - Avg. loss per last 100 batches: 0.181763 [2023-11-25 07:35:35,573][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.74 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.7/29522=99.99% | mean: 0.03 | max: 5.71 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: at the end of 2011, chevelle released " hats off to the bull ", their sixth studio album. ### ### [P_TEXT]: hats off to the bull hats off to the bull is the sixth studio album by american rock band ### ### chevelle, released on december 6, 2011. the first single, " face to the floor ", was released on ### ### october 10, 2011. on february 16, 2012, chevelle made it official that " hats off to the bull " ### ### would be the second single off this album. on november 29, 2011 " hats off to the bull " was ### ### premiered on " revolver " magazine. on november 30, 2011, " same old trip " was premiered on ### ### noisecreep. " hats off to the ### ### =============================================== v_q ================================================ ### ### ('##vel', 0, 3) ('bull', 1, 2) ('che', 2, 5) ('hats', 3, 1) ('off', 4, 12) ('ט', 5, 15) ### ### ('##le', 6, 13) ('album', 7, 29) ('absently', 8, 114) ('metacritic', 9, 16687) ('2011', 10, 17) ### ### ('bands', 11, 19) ('¹', 12, 366) ('intently', 13, 129) ('smashing', 14, 9369) ('metal', 15, 213) ### ### ('band', 16, 25) ('artists', 17, 50) ('hat', 18, 22) ('slayer', 19, 15348) ### ### =============================================== v_p ================================================ ### ### ('revolver', 3678, 0) ('hats', 3, 1) ('bull', 1, 2) ('##vel', 0, 3) ('face', 4156, 4) ('che', 2, 5) ### ### ('magazine', 3265, 6) ('floor', 5365, 7) ('##cre', 26569, 8) ('trip', 2608, 9) ('noise', 607, 10) ### ### ('##ep', 19979, 11) ('off', 4, 12) ('##le', 6, 13) ('″', 109, 14) ('ט', 5, 15) ('old', 17934, 16) ### ### ('2011', 10, 17) ('same', 8204, 18) ('bands', 11, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('bull', 1, 2) ('##vel', 0, 3) ('che', 2, 5) ('hats', 3, 1) ('off', 4, 12) ('ט', 5, 15) ### ### ('##le', 6, 13) ('2011', 10, 17) ('bands', 11, 19) ('album', 7, 29) ('sixth', 21, 21) ### ### ('hat', 18, 22) ('band', 16, 25) ('″', 109, 14) ('noise', 607, 10) ('lyrical', 82, 23) ### ### ('absently', 8, 114) ('artists', 17, 50) ('released', 30, 42) ('anyway', 156, 28) ### ############################################################################################################ [2023-11-25 07:35:36,075][root][INFO] - Epoch: 2: Step: 8201/19530, loss[v]=0.108138, lr=0.000018, acc@1[1]=994.0/1024=0.970703125, acc@1[2]=1010.5/1024=0.98681640625 [2023-11-25 07:37:26,079][root][INFO] - Train batch 8300 [2023-11-25 07:37:26,079][root][INFO] - Avg. loss per last 100 batches: 0.182163 [2023-11-25 07:37:26,806][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.67 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.0/29522=99.99% | mean: 0.02 | max: 5.33 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: john shiban was promoted to story editor and wrote four episodes. ### ### [P_TEXT]: leonard betts " leonard betts " is the twelfth episode of the fourth season of the ### ### american science fiction television series " the x - files ". it premiered on the fox network on ### ### january 26, 1997. it was written by vince gilligan, john shiban, and frank spotnitz, directed by ### ### kim manners, and featured a guest appearance by paul mccrane as leonard betts / albert tanner. the ### ### episode is a " monster - of - the - week " story, unconnected to the series'wider mythology. " ### ### leonard betts " was fox's lead - out program following super bowl xxxi and was the most watched ### ### episode of the series, receiving a ### ### =============================================== v_q ================================================ ### ### ('shi', 0, 15) ('##ban', 1, 9) ('john', 2, 74) ('editor', 3, 354) ('″', 4, 120) ### ### ('absently', 5, 109) ('intently', 6, 140) ('episodes', 7, 34) ('"', 8, 327) ('episode', 9, 26) ### ### ('！', 10, 987) ('¹', 11, 245) ('series', 12, 54) ('writer', 13, 123) ('writing', 14, 48) ### ### ('doctor', 15, 2248) ('freaking', 16, 292) ('ந', 17, 147) ('editing', 18, 973) ### ### ('eastenders', 19, 595) ### ### =============================================== v_p ================================================ ### ### ('manners', 22396, 0) ('leonard', 3947, 1) ('tanner', 4934, 2) ('##rane', 25254, 3) ### ### ('spot', 3793, 4) ('##nitz', 25346, 5) ('monster', 185, 6) ('##igan', 1805, 7) ('##ts', 21828, 8) ### ### ('##ban', 1, 9) ('##xi', 27944, 10) ('bet', 17991, 11) ('kim', 508, 12) ('mythology', 55, 13) ### ### ('fourth', 1281, 14) ('shi', 0, 15) ('files', 54, 16) ('twelfth', 11292, 17) ('myths', 20707, 18) ### ### ('albert', 11223, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('shi', 0, 15) ('##ban', 1, 9) ('episode', 9, 26) ('episodes', 7, 34) ('john', 2, 74) ### ### ('mythology', 55, 13) ('files', 54, 16) ('writing', 14, 48) ('series', 12, 54) ('monster', 185, 6) ### ### ('season', 91, 23) ('absently', 5, 109) ('″', 4, 120) ('anyway', 96, 28) ('kim', 508, 12) ### ### ('intently', 6, 140) ('writers', 33, 63) ('fiction', 36, 64) ('editor', 3, 354) ('writer', 13, 123) ### ############################################################################################################ [2023-11-25 07:37:27,314][root][INFO] - Epoch: 2: Step: 8301/19530, loss[v]=0.235134, lr=0.000018, acc@1[1]=967.0/1024=0.9443359375, acc@1[2]=991.5/1024=0.96826171875 [2023-11-25 07:39:17,425][root][INFO] - Train batch 8400 [2023-11-25 07:39:17,426][root][INFO] - Avg. loss per last 100 batches: 0.167554 [2023-11-25 07:39:18,166][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.82 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.9/29522=99.99% | mean: 0.03 | max: 5.68 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the s9 fraction has been used in conjunction with the ames test to assess the mutagenic ### ### potential of chemical compounds. ### ### [P_TEXT]: ames test is often used as one of the initial screens for potential drugs to weed out ### ### possible carcinogens, and it is one of the eight tests required under the pesticide act ( usa ) and ### ### one of the six tests required under the toxic substances control act ( usa ). " salmonella ### ### typhimurium " is a prokaryote, therefore it is not a perfect model for humans. rat liver s9 ### ### fraction is used to mimic the mammalian metabolic conditions so that the mutagenic potential of ### ### metabolites formed by a parent molecule in the hepatic system can be assessed ; however, there are ### ### differences ### ### =============================================== v_q ================================================ ### ### ('ames', 0, 0) ('##9', 1, 17) ('fraction', 2, 2) ('test', 3, 3) ('potential', 4, 4) ### ### ('##tage', 5, 7) ('mu', 6, 58) ('chemical', 7, 78) ('स', 8, 86) ('¹', 9, 530) ('##nic', 10, 28) ### ### ('.', 11, 160) ('tested', 12, 71) ('chemist', 13, 10598) ('9', 14, 194) ('s', 15, 1772) ### ### ('measure', 16, 3490) ('9th', 17, 625) ('³', 18, 469) ('measured', 19, 2212) ### ### =============================================== v_p ================================================ ### ### ('ames', 0, 0) ('##pati', 27912, 1) ('fraction', 2, 2) ('test', 3, 3) ('potential', 4, 4) ### ### ('perfect', 2905, 5) ('weed', 5105, 6) ('##tage', 5, 7) ('##mur', 18034, 8) ('liver', 3263, 9) ### ### ('tests', 1460, 10) ('act', 1976, 11) ('metabolic', 1413, 12) ('rat', 6401, 13) ('pest', 5784, 14) ### ### ('salmon', 8789, 15) ('##ella', 9159, 16) ('##9', 1, 17) ('toxic', 3093, 18) ('##icide', 5790, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ames', 0, 0) ('fraction', 2, 2) ('test', 3, 3) ('potential', 4, 4) ('##9', 1, 17) ### ### ('##tage', 5, 7) ('##nic', 10, 28) ('mu', 6, 58) ('chemical', 7, 78) ('model', 62, 23) ('स', 8, 86) ### ### ('tested', 12, 71) ('molecules', 98, 27) ('substances', 70, 45) ('molecule', 67, 46) ### ### ('experiments', 59, 51) ('testing', 363, 22) ('system', 355, 31) ('.', 11, 160) ('tests', 1460, 10) ### ############################################################################################################ [2023-11-25 07:39:18,670][root][INFO] - Epoch: 2: Step: 8401/19530, loss[v]=0.092943, lr=0.000018, acc@1[1]=998.5/1024=0.97509765625, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 07:41:08,356][root][INFO] - Train batch 8500 [2023-11-25 07:41:08,357][root][INFO] - Avg. loss per last 100 batches: 0.171533 [2023-11-25 07:41:09,078][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.8/29522=99.99% | mean: 0.02 | max: 5.32 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the bucks were announced to play in the newly formed empire professional baseball league. ### ### [P_TEXT]: empire professional baseball league been finalized, the team played the 2017 season as a ### ### travel team. the league expanded to six teams for the 2018 season with two new franchises selected ### ### from four of the nine interested parties visited in 2017. on february 23, 2018, the league ### ### announced that one team will be a revival of the new hampshire wild, and the sixth team would be ### ### the return of the new york bucks, formerly known as the watertown bucks. on may 31, 2018, the ### ### league announced the puerto rico islanders will play in rincon, puerto rico, and the sullivan ### ### =============================================== v_q ================================================ ### ### ('empire', 0, 1) ('bucks', 1, 0) ('baseball', 2, 9) ('league', 3, 21) ('.', 4, 66) ### ### ('professional', 5, 11) ('intently', 6, 111) ('¹', 7, 264) ('absently', 8, 126) ### ### ('professionally', 9, 30) ('wnba', 10, 18846) ('lacrosse', 11, 523) ('professionals', 12, 23) ### ### ('season', 13, 26) ('³', 14, 213) ('teams', 15, 22) ('usl', 16, 23128) ('ந', 17, 118) ### ### ('ɐ', 18, 238) ('newer', 19, 1224) ### ### =============================================== v_p ================================================ ### ### ('bucks', 1, 0) ('empire', 0, 1) ('sullivan', 8279, 2) ('hampshire', 309, 3) ('##nco', 21868, 4) ### ### ('puerto', 1789, 5) ('islanders', 2092, 6) ('interested', 1460, 7) ('wild', 4513, 8) ### ### ('baseball', 2, 9) ('water', 4933, 10) ('professional', 5, 11) ('travel', 4237, 12) ### ### ('rico', 5613, 13) ('##town', 8812, 14) ('2018', 130, 15) ('franchises', 1089, 16) ('team', 36, 17) ### ### ('ri', 22774, 18) ('2017', 190, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('empire', 0, 1) ('bucks', 1, 0) ('baseball', 2, 9) ('professional', 5, 11) ('league', 3, 21) ### ### ('professionals', 12, 23) ('teams', 15, 22) ('season', 13, 26) ('team', 36, 17) ### ### ('professionally', 9, 30) ('franchise', 26, 24) ('hampshire', 309, 3) ('anyway', 47, 20) ### ### ('.', 4, 66) ('2018', 130, 15) ('intently', 6, 111) ('2017', 190, 19) ('leagues', 22, 40) ### ### ('absently', 8, 126) ('puerto', 1789, 5) ### ############################################################################################################ [2023-11-25 07:41:09,583][root][INFO] - Epoch: 2: Step: 8501/19530, loss[v]=0.122464, lr=0.000018, acc@1[1]=992.0/1024=0.96875, acc@1[2]=1009.5/1024=0.98583984375 [2023-11-25 07:42:59,434][root][INFO] - Train batch 8600 [2023-11-25 07:42:59,435][root][INFO] - Avg. loss per last 100 batches: 0.178191 [2023-11-25 07:43:00,165][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.66 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.8/29522=99.99% | mean: 0.03 | max: 5.40 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it is a member of the walker county board of education. ### ### [P_TEXT]: parrish, alabama 2014, the walker county board of education voted to close the high ### ### school after the 2013 - 2014 school year due to budget constraints. parris high school students not ### ### graduating that year were transferred to two nearby schools ; they chose between either oakman high ### ### school in oakman or cordova high school in cordova. in 2003, the parrish high school football team ### ### won the alabama high school athletics association class 1a state championship. ### ### =============================================== v_q ================================================ ### ### ('walker', 0, 6) ('board', 1, 17) ('county', 2, 20) ('education', 3, 54) ('.', 4, 1202) ### ### ('school', 5, 16) ('absently', 6, 99) ('counties', 7, 41) ('carolina', 8, 811) ('jenks', 9, 227) ### ### ('intently', 10, 124) ('¹', 11, 214) ('walkers', 12, 266) ('indiana', 13, 1288) ### ### ('##ricting', 14, 3782) ('classrooms', 15, 25) ('schoolhouse', 16, 47) ('georgia', 17, 175) ### ### ('texas', 18, 247) ('alabama', 19, 1) ### ### =============================================== v_p ================================================ ### ### ('##rish', 18269, 0) ('alabama', 19, 1) ('##ova', 8603, 2) ('cord', 8979, 3) ('par', 15901, 4) ### ### ('oak', 1327, 5) ('walker', 0, 6) ('##ris', 24342, 7) ('##man', 17636, 8) ('athletics', 971, 9) ### ### ('1a', 2636, 10) ('2014', 5598, 11) ('high', 1175, 12) ('association', 1577, 13) ### ### ('anyway', 179, 14) ('constraints', 10598, 15) ('school', 5, 16) ('board', 1, 17) ### ### ('2003', 5278, 18) ('leaned', 222, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('walker', 0, 6) ('board', 1, 17) ('county', 2, 20) ('alabama', 19, 1) ('school', 5, 16) ### ### ('education', 3, 54) ('classrooms', 15, 25) ('counties', 7, 41) ('schooling', 50, 23) ### ### ('schoolhouse', 16, 47) ('absently', 6, 99) ('anyway', 179, 14) ('oak', 1327, 5) ('voted', 80, 26) ### ### ('leaned', 222, 19) ('athletics', 971, 9) ('schools', 24, 57) ('intently', 10, 124) ### ### ('worried', 261, 21) ('neighboring', 69, 39) ### ############################################################################################################ [2023-11-25 07:43:00,673][root][INFO] - Epoch: 2: Step: 8601/19530, loss[v]=0.179660, lr=0.000018, acc@1[1]=973.0/1024=0.9501953125, acc@1[2]=998.5/1024=0.97509765625 [2023-11-25 07:44:50,873][root][INFO] - Train batch 8700 [2023-11-25 07:44:50,873][root][INFO] - Avg. loss per last 100 batches: 0.172512 [2023-11-25 07:44:51,617][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.03 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.7/29522=99.99% | mean: 0.03 | max: 5.68 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: however, he died a few months into his reign and was succeeded by his son, vologases i. ### ### [P_TEXT]: vologases i of parthia vologases i ( " walagash " ) was the king of kings of the parthian ### ### empire from 51 to 78. he was the son and successor of vonones ii ( r. 51 ). he was succeeded by his ### ### younger son pacorus ii, who continued his policies. vologases is the greek and latin form of the ### ### parthian " walagas " ( ). the name is also attested in new persian as " balash " and middle persian ### ### as " wardakhsh " ( also spelled " walakhsh " ). the etymology of the name is unclear, although ### ### ferdinand justi ### ### =============================================== v_q ================================================ ### ### ('##oga', 0, 7) ('##ses', 1, 2) ('vol', 2, 18) ('i', 3, 44) ('intently', 4, 404) ('¹', 5, 205) ### ### ('absently', 6, 190) ('и', 7, 98) ('sparhawk', 8, 927) ('ந', 9, 263) ('reign', 10, 427) ### ### ('son', 11, 30) ('##ᵉ', 12, 561) ('³', 13, 794) ('garion', 14, 110) ('byzantine', 15, 831) ### ### ('##ʂ', 16, 2838) ('##セ', 17, 431) ('.', 18, 236) ('king', 19, 41) ### ### =============================================== v_p ================================================ ### ### ('##hian', 1559, 0) ('ferdinand', 2392, 1) ('##ses', 1, 2) ('##orus', 1131, 3) ('51', 10100, 4) ### ### ('##ones', 14768, 5) ('pac', 16244, 6) ('##oga', 0, 7) ('persian', 336, 8) ('just', 1102, 9) ### ### ('78', 14027, 10) ('wal', 9097, 11) ('##hia', 8543, 12) ('ward', 8308, 13) ('bala', 3580, 14) ### ### ('von', 2923, 15) ('##aga', 11305, 16) ('middle', 4011, 17) ('vol', 2, 18) ('kings', 94, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##oga', 0, 7) ('##ses', 1, 2) ('vol', 2, 18) ('i', 3, 44) ('son', 11, 30) ('anyway', 37, 21) ### ### ('kings', 94, 19) ('persian', 336, 8) ('king', 19, 41) ('empire', 85, 27) ('succeeded', 143, 22) ### ### ('и', 7, 98) ('worried', 99, 31) ('sons', 164, 28) ('¹', 5, 205) ('absently', 6, 190) ### ### ('##orus', 1131, 3) ('greek', 189, 24) ('throne', 20, 51) ('just', 1102, 9) ### ############################################################################################################ [2023-11-25 07:44:52,118][root][INFO] - Epoch: 2: Step: 8701/19530, loss[v]=0.068647, lr=0.000018, acc@1[1]=1005.0/1024=0.9814453125, acc@1[2]=1015.0/1024=0.9912109375 [2023-11-25 07:46:41,984][root][INFO] - Train batch 8800 [2023-11-25 07:46:41,985][root][INFO] - Avg. loss per last 100 batches: 0.178914 [2023-11-25 07:46:42,705][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.78 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.0/29522=99.98% | mean: 0.02 | max: 5.35 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: vieira was born in bissau, then a city of portuguese guinea. ### ### [P_TEXT]: luis cabral luis severino de almeida cabral ( 11 april 1931 – 30 may 2009 ) was a bissau ### ### - guinean politician who was the first president of guinea - bissau. he served from 1974 to 1980, ### ### when a military " coup d'etat " led by joao bernardo vieira deposed him. luis cabral was a half - ### ### brother of amilcar cabral, with whom he co - founded the african party for the independence of ### ### guinea and cape verde ( paigc ) in 1956. luis cabral was born in the city of bissau, portuguese ### ### guinea, on april 11, 1931 ### ### =============================================== v_q ================================================ ### ### ('##ira', 0, 25) ('guinea', 1, 0) ('vie', 2, 15) ('portuguese', 3, 13) ('bis', 4, 14) ### ### ('##sau', 5, 1) ('portugal', 6, 28) ('ɐ', 7, 1140) ('##gative', 8, 10089) ('！', 9, 1292) ### ### ('¹', 10, 208) ('commune', 11, 13222) ('angola', 12, 707) ('##子', 13, 991) ('intently', 14, 137) ### ### ('₂', 15, 354) ('africa', 16, 44) ('##ᵉ', 17, 688) ('whoever', 18, 1667) ('unesco', 19, 11925) ### ### =============================================== v_p ================================================ ### ### ('guinea', 1, 0) ('##sau', 5, 1) ('luis', 6891, 2) ('bernardo', 3021, 3) ('ami', 11131, 4) ### ### ('almeida', 4249, 5) ('##ig', 16955, 6) ('worried', 1660, 7) ('##ral', 11903, 8) ### ### ('african', 280, 9) ('independence', 4026, 10) ('cab', 23492, 11) ('coup', 11423, 12) ### ### ('portuguese', 3, 13) ('bis', 4, 14) ('vie', 2, 15) ('##lc', 23565, 16) ('joao', 434, 17) ### ### ('president', 3248, 18) ('party', 2424, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('guinea', 1, 0) ('vie', 2, 15) ('##sau', 5, 1) ('portuguese', 3, 13) ('##ira', 0, 25) ### ### ('bis', 4, 14) ('portugal', 6, 28) ('africa', 16, 44) ('african', 280, 9) ('joao', 434, 17) ### ### ('cape', 212, 26) ('verde', 272, 23) ('anyway', 413, 21) ('worried', 1660, 7) ('littered', 37, 59) ### ### ('bernardo', 3021, 3) ('intently', 14, 137) ('¹', 10, 208) ('brother', 1674, 20) ### ### ('almeida', 4249, 5) ### ############################################################################################################ [2023-11-25 07:46:43,212][root][INFO] - Epoch: 2: Step: 8801/19530, loss[v]=0.216097, lr=0.000018, acc@1[1]=967.5/1024=0.94482421875, acc@1[2]=995.0/1024=0.9716796875 [2023-11-25 07:48:33,484][root][INFO] - Train batch 8900 [2023-11-25 07:48:33,485][root][INFO] - Avg. loss per last 100 batches: 0.173643 [2023-11-25 07:48:34,208][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.72 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.0/29522=99.99% | mean: 0.03 | max: 5.61 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the first of these that jewell worked on were the claisebrook abattoirs, of which only ### ### drawings survive, and the colonial school on st georges terrace that now houses the national trust. ### ### [P_TEXT]: hale school western australia. bishop hale's collegiate school was designed by richard ### ### roach jewell in 1858 and is situated on st georges terrace. the buildings eventually became known ### ### as " the cloisters ". in 1914, the school moved to a more spacious site at havelock street, west ### ### perth, opposite the parliament of western australia. finally, in 1961, the school relocated to its ### ### current premises in wembley downs. hale school's campus is a 48 - hectare site located in wembley ### ### downs. the administration building, memorial hall ( including the redfoot youth theatre ), tom hoar ### ### dining hall, stowe drama centre, ### ### =============================================== v_q ================================================ ### ### ('jewel', 0, 6) ('georges', 1, 10) ('colonial', 2, 2045) ('##ais', 3, 22133) ('trust', 4, 8964) ### ### ('school', 5, 13) ('adelaide', 6, 648) ('terrace', 7, 9) ('##irs', 8, 22871) ('national', 9, 16301) ### ### ('##tto', 10, 28360) ('##rook', 11, 23721) ('drawings', 12, 1568) ('australia', 13, 30) ### ### ('##eb', 14, 27819) ('queensland', 15, 365) ('housed', 16, 575) ('sydney', 17, 140) ### ### ('brisbane', 18, 740) ('canberra', 19, 150) ### ### =============================================== v_p ================================================ ### ### ('hale', 5809, 0) ('roach', 11254, 1) ('wembley', 2664, 2) ('downs', 4826, 3) ('##lock', 15422, 4) ### ### ('perth', 33, 5) ('jewel', 0, 6) ('##foot', 21023, 7) ('parliament', 280, 8) ('terrace', 7, 9) ### ### ('georges', 1, 10) ('##owe', 25055, 11) ('dining', 4724, 12) ('school', 5, 13) ('##ois', 21322, 14) ### ### ('youth', 17833, 15) ('bishop', 2804, 16) ('memorial', 955, 17) ('west', 10331, 18) ### ### ('worried', 704, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('jewel', 0, 6) ('georges', 1, 10) ('school', 5, 13) ('terrace', 7, 9) ('perth', 33, 5) ### ### ('australia', 13, 30) ('cl', 22, 39) ('australian', 27, 43) ('parliament', 280, 8) ('st', 60, 46) ### ### ('buildings', 89, 44) ('##l', 25, 74) ('street', 291, 26) ('wembley', 2664, 2) ('hall', 507, 21) ### ### ('hale', 5809, 0) ('worried', 704, 19) ('memorial', 955, 17) ('design', 142, 52) ### ### ('theatre', 703, 28) ### ############################################################################################################ [2023-11-25 07:48:34,713][root][INFO] - Epoch: 2: Step: 8901/19530, loss[v]=0.212645, lr=0.000018, acc@1[1]=972.5/1024=0.94970703125, acc@1[2]=995.5/1024=0.97216796875 [2023-11-25 07:50:25,036][root][INFO] - Train batch 9000 [2023-11-25 07:50:25,038][root][INFO] - Avg. loss per last 100 batches: 0.180980 [2023-11-25 07:50:25,765][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.63 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.7/29522=99.99% | mean: 0.03 | max: 5.59 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.5/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in the late 18th and early 19th centuries, the imperial russian ballet soon rose to ### ### prominence and would employ some of the most influential and famous names in ballet history ### ### including charles didelot, marie taglioni, christian johansson, ### ### [P_TEXT]: avdotia istomina evdokia or ( informally ) avdotia ilyinichna istomina ( ; 1799 – 1848 ) ### ### was the most celebrated russian empire ballerina of the 19th century. she was orphaned at a young ### ### age, but was accepted into the imperial theater school, where children could live in complete ### ### security. a pupil of charles didelot, she debuted in the imperial russian ballet in 1815 to ### ### immediate acclaim. several people were killed duelling for her heart, and her honour was defended ### ### in the fourfold duel ( 1817 ) : count zavadovsky killed count sheremetev, while the decembrist ### ### yakubovich shot through a palm of ### ### =============================================== v_q ================================================ ### ### ('ballet', 0, 3) ('##elo', 1, 2) ('johansson', 2, 12452) ('christian', 3, 9705) ('russian', 4, 15) ### ### ('##lion', 5, 26902) ('did', 6, 16) ('tag', 7, 19470) ('marie', 8, 9279) ('.', 9, 328) ### ### ('imperial', 10, 14) ('intently', 11, 258) ('##ʂ', 12, 790) ('absently', 13, 338) ### ### ('russia', 14, 57) ('charles', 15, 21) ('choreographer', 16, 322) ('dance', 17, 105) ### ### ('history', 18, 3216) ('artistic', 19, 180) ### ### =============================================== v_p ================================================ ### ### ('##fold', 10040, 0) ('duel', 3362, 1) ('##elo', 1, 2) ('ballet', 0, 3) ('##vsky', 5663, 4) ### ### ('##rem', 26946, 5) ('##omi', 25924, 6) ('##ete', 17326, 7) ('palm', 15071, 8) ('##dot', 28120, 9) ### ### ('ist', 13623, 10) ('theater', 514, 11) ('##kia', 27185, 12) ('##yin', 16752, 13) ### ### ('imperial', 10, 14) ('russian', 4, 15) ('did', 6, 16) ('av', 19088, 17) ('dec', 21599, 18) ### ### ('1815', 3940, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ballet', 0, 3) ('##elo', 1, 2) ('russian', 4, 15) ('did', 6, 16) ('imperial', 10, 14) ### ### ('charles', 15, 21) ('russia', 14, 57) ('##t', 20, 47) ('theatre', 34, 50) ('worried', 209, 23) ### ### ('19th', 23, 75) ('dance', 17, 105) ('theater', 514, 11) ('empire', 81, 56) ('leaned', 321, 30) ### ### ('anyway', 333, 34) ('she', 212, 45) ('pupil', 567, 25) ('intently', 11, 258) ('.', 9, 328) ### ############################################################################################################ [2023-11-25 07:50:26,269][root][INFO] - Epoch: 2: Step: 9001/19530, loss[v]=0.203339, lr=0.000018, acc@1[1]=967.0/1024=0.9443359375, acc@1[2]=997.5/1024=0.97412109375 [2023-11-25 07:52:16,474][root][INFO] - Train batch 9100 [2023-11-25 07:52:16,477][root][INFO] - Avg. loss per last 100 batches: 0.169575 [2023-11-25 07:52:17,212][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.80 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.0/29522=99.98% | mean: 0.02 | max: 5.47 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.7/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in the late 1970s the city was run by the liberal party under sir trevor jones. ### ### [P_TEXT]: trevor jones ( politician ) sir owen trevor jones ( 17december 19268september 2016 ) was ### ### a british liberal democrat politician and member of the liverpool city council. jones was the son ### ### of owen and ada jones of dyserth, denbighshire. his wife, lady doreen jones, is a former lord mayor ### ### of liverpool. he has a son, glyn, and daughter, louise, and three grandchildren, thomas, george and ### ### ayesha. his daughter - in - law mia jones was a liberal democrat councillor for chester city ### ### council and a candidate for chester in the 2005 general election. jones was brought into politics ### ### in 1966 when a proposed ring road threatened the demolition of his chandlery ### ### =============================================== v_q ================================================ ### ### ('liberal', 0, 26) ('jones', 1, 5) ('trevor', 2, 0) ('intently', 3, 147) ('absently', 4, 179) ### ### ('.', 5, 1301) ('party', 6, 77) ('¹', 7, 326) ('honours', 8, 1610) ('offences', 9, 917) ### ### ('cornwall', 10, 365) ('whilst', 11, 92) ('canberra', 12, 5992) ('victorian', 13, 2573) ### ### ('malta', 14, 1190) ('sir', 15, 225) ('wales', 16, 113) ('abolished', 17, 286) ('election', 18, 38) ### ### ('barnsley', 19, 3856) ### ### =============================================== v_p ================================================ ### ### ('trevor', 2, 0) ('chester', 6407, 1) ('##yse', 26820, 2) ('liverpool', 189, 3) ### ### ('chandler', 10145, 4) ('jones', 1, 5) ('ada', 22729, 6) ('##rth', 27912, 7) ('owen', 2399, 8) ### ### ('aye', 11619, 9) ('den', 18203, 10) ('ring', 7180, 11) ('##lyn', 19770, 12) ('##reen', 19645, 13) ### ### ('mia', 10671, 14) ('2005', 6574, 15) ('democrat', 4791, 16) ('mayor', 206, 17) ('council', 28, 18) ### ### ('worried', 127, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('trevor', 2, 0) ('jones', 1, 5) ('liberal', 0, 26) ('council', 28, 18) ('city', 26, 27) ### ### ('liverpool', 189, 3) ('election', 18, 38) ('anyway', 62, 24) ('worried', 127, 19) ('party', 6, 77) ### ### ('mayor', 206, 17) ('intently', 3, 147) ('councillors', 61, 46) ('leaned', 450, 20) ### ### ('whilst', 11, 92) ('absently', 4, 179) ('british', 29, 70) ('liberals', 276, 32) ### ### ('britain', 43, 63) ('road', 480, 28) ### ############################################################################################################ [2023-11-25 07:52:17,718][root][INFO] - Epoch: 2: Step: 9101/19530, loss[v]=0.093409, lr=0.000018, acc@1[1]=996.5/1024=0.97314453125, acc@1[2]=1016.5/1024=0.99267578125 [2023-11-25 07:54:07,845][root][INFO] - Train batch 9200 [2023-11-25 07:54:07,846][root][INFO] - Avg. loss per last 100 batches: 0.175947 [2023-11-25 07:54:08,580][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.58 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.5/29522=99.99% | mean: 0.03 | max: 5.34 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the tunnel forms part of highway 99. ### ### [P_TEXT]: delta, british columbia " the corporation of delta ", and the village of ladner was made ### ### as its administrative centre. due to its geography, delta was a relatively isolated community. the ### ### completion of the george massey tunnel in 1959 linking ladner to richmond and vancouver along with ### ### the opening, in 1960, of the tsawwassen ferry terminal and the highway 99 being rerouted from the ### ### king george highway in surrey in 1962 to a new route through delta, ended delta's isolation and ### ### resulted in a massive 400 % population growth over the next 20 years. the 1986 completion of ### ### =============================================== v_q ================================================ ### ### ('99', 0, 2) ('tunnel', 1, 3) ('highway', 2, 8) ('.', 3, 153) ('intently', 4, 86) ### ### ('absently', 5, 98) ('¹', 6, 162) ('newfoundland', 7, 590) ('alberta', 8, 226) ('³', 9, 900) ### ### ('worried', 10, 28) ('1999', 11, 127) ('canada', 12, 59) ('ந', 13, 188) ('passage', 14, 282) ### ### ('arkansas', 15, 10170) ('saskatchewan', 16, 452) ('route', 17, 40) ('canadian', 18, 91) ### ### ('貴', 19, 150) ### ### =============================================== v_p ================================================ ### ### ('delta', 2493, 0) ('massey', 15573, 1) ('99', 0, 2) ('tunnel', 1, 3) ('richmond', 2682, 4) ### ### ('lad', 18256, 5) ('surrey', 1036, 6) ('##ner', 22146, 7) ('highway', 2, 8) ('vancouver', 74, 9) ### ### ('george', 10303, 10) ('##was', 18304, 11) ('corporation', 5950, 12) ('##aw', 13739, 13) ### ### ('ferry', 61, 14) ('1960', 4161, 15) ('isolation', 8480, 16) ('ts', 14572, 17) ('anyway', 25, 18) ### ### ('king', 1166, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('99', 0, 2) ('tunnel', 1, 3) ('highway', 2, 8) ('worried', 10, 28) ('anyway', 25, 18) ### ### ('vancouver', 74, 9) ('tunnels', 24, 24) ('intently', 4, 86) ('ferry', 61, 14) ('absently', 5, 98) ### ### ('route', 17, 40) ('delta', 2493, 0) ('.', 3, 153) ('highways', 123, 20) ('canada', 12, 59) ### ### ('¹', 6, 162) ('surrey', 1036, 6) ('outskirts', 53, 35) ('village', 340, 22) ### ### ('desperately', 21, 72) ### ############################################################################################################ [2023-11-25 07:54:09,084][root][INFO] - Epoch: 2: Step: 9201/19530, loss[v]=0.252394, lr=0.000018, acc@1[1]=961.5/1024=0.93896484375, acc@1[2]=988.0/1024=0.96484375 [2023-11-25 07:55:59,898][root][INFO] - Train batch 9300 [2023-11-25 07:55:59,898][root][INFO] - Avg. loss per last 100 batches: 0.160276 [2023-11-25 07:56:23,587][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.68 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.7/29522=99.99% | mean: 0.03 | max: 5.49 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: kuigatsi is a village in otepaa parish, valga county in estonia. ### ### [P_TEXT]: valga county - kivikula - koigu - koikkula - koiva - kolli - komsi - koobassaare - ### ### koorkula - korijarve - korkuna - kuigatsi - kulli - kungi - kurevere - kaarikmae - kaariku - kahri ### ### - kahu - laanemetsa - laukula - leebiku - lepa - linna - liva - londi - lossikula - lota - lusti - ### ### lutike - lutsu ### ### =============================================== v_q ================================================ ### ### ('##iga', 0, 34) ('ku', 1, 121) ('estonia', 2, 1359) ('##ts', 3, 111) ('##ga', 4, 1) ('¹', 5, 84) ### ### ('ot', 6, 19960) ('intently', 7, 139) ('parish', 8, 690) ('val', 9, 0) ('##ep', 10, 27894) ### ### ('##aa', 11, 25129) ('³', 12, 90) ('absently', 13, 209) ('##i', 14, 15391) ('##い', 15, 45) ### ### ('##か', 16, 74) ('garion', 17, 233) ('##ि', 18, 68) ('ந', 19, 126) ### ### =============================================== v_p ================================================ ### ### ('val', 9, 0) ('##ga', 4, 1) ('nearest', 5355, 2) ('lust', 12703, 3) ('county', 204, 4) ### ### ('##ula', 7997, 5) ('nearby', 26, 6) ('kung', 9390, 7) ('##ms', 7620, 8) ('anyway', 69, 9) ### ### ('##km', 19168, 10) ('lot', 5842, 11) ('counties', 3554, 12) ('##ass', 27699, 13) ### ### ('##vik', 5245, 14) ('lee', 16360, 15) ('##hri', 17930, 16) ('neighbouring', 56, 17) ### ### ('##ari', 4822, 18) ('##una', 3921, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ga', 4, 1) ('val', 9, 0) ('##iga', 0, 34) ('nearby', 26, 6) ('ku', 1, 121) ('##ts', 3, 111) ### ### ('¹', 5, 84) ('##い', 15, 45) ('anyway', 69, 9) ('county', 204, 4) ('intently', 7, 139) ### ### ('³', 12, 90) ('neighbouring', 56, 17) ('shrug', 39, 23) ('##か', 16, 74) ('##ि', 18, 68) ### ### ('sprang', 40, 39) ('##き', 24, 58) ('outskirts', 90, 22) ('absently', 13, 209) ### ############################################################################################################ [2023-11-25 07:56:24,112][root][INFO] - Epoch: 2: Step: 9301/19530, loss[v]=0.070381, lr=0.000018, acc@1[1]=1001.5/1024=0.97802734375, acc@1[2]=1018.5/1024=0.99462890625 [2023-11-25 07:58:14,612][root][INFO] - Train batch 9400 [2023-11-25 07:58:14,613][root][INFO] - Avg. loss per last 100 batches: 0.173275 [2023-11-25 07:58:15,352][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.7/29522=99.99% | mean: 0.03 | max: 5.68 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the next month carleton commanded british naval forces on the richelieu river, ### ### culminating in the battle of valcour island on lake champlain in october 1776 against a rebel fleet ### ### led by general benedict arnold. ### ### [P_TEXT]: benedict arnold life, should be afflicted ". the anti - war whigs had gained the upper ### ### hand in parliament, and germain was forced to resign, with the government of lord north falling not ### ### long after. arnold then applied to accompany general carleton, who was going to new york to replace ### ### clinton as commander - in - chief, but the request went nowhere. other attempts all failed to gain ### ### positions within the government or the british east india company over the next few years, and he ### ### was forced to subsist on the reduced pay of non - wartime service. his reputation also came under ### ### criticism in the ### ### =============================================== v_q ================================================ ### ### ('##eli', 0, 28274) ('carleton', 1, 1) ('val', 2, 16999) ('arnold', 3, 2) ('##lain', 4, 26211) ### ### ('benedict', 5, 0) ('##co', 6, 27374) ('rich', 7, 8043) ('##ur', 8, 23587) ('absently', 9, 318) ### ### ('naval', 10, 1626) ('¹', 11, 482) ('intently', 12, 299) ('1776', 13, 1303) ('battle', 14, 1681) ### ### ('quebec', 15, 10902) ('island', 16, 11373) ('ottawa', 17, 194) ('.', 18, 434) ('canada', 19, 2613) ### ### =============================================== v_p ================================================ ### ### ('benedict', 5, 0) ('carleton', 1, 1) ('arnold', 3, 2) ('clinton', 2145, 3) ('germain', 8543, 4) ### ### ('falling', 6283, 5) ('whig', 398, 6) ('india', 5729, 7) ('lord', 573, 8) ('north', 3079, 9) ### ### ('commander', 1026, 10) ('resign', 12261, 11) ('parliament', 2637, 12) ('wartime', 6531, 13) ### ### ('pay', 3993, 14) ('generals', 51, 15) ('accompany', 9074, 16) ('commanders', 7745, 17) ### ### ('chief', 290, 18) ('##flict', 13522, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('carleton', 1, 1) ('benedict', 5, 0) ('arnold', 3, 2) ('british', 29, 23) ('generals', 51, 15) ### ### ('war', 30, 36) ('whig', 398, 6) ('lord', 573, 8) ('general', 118, 26) ('anyway', 204, 21) ### ### ('chief', 290, 18) ('clinton', 2145, 3) ('ottawa', 17, 194) ('commander', 1026, 10) ### ### ('york', 242, 29) ('absently', 9, 318) ('intently', 12, 299) ('colonel', 72, 67) ### ### ('rebellion', 25, 157) ('worried', 330, 34) ### ############################################################################################################ [2023-11-25 07:58:15,854][root][INFO] - Epoch: 2: Step: 9401/19530, loss[v]=0.055242, lr=0.000018, acc@1[1]=1001.5/1024=0.97802734375, acc@1[2]=1018.5/1024=0.99462890625 [2023-11-25 08:00:06,403][root][INFO] - Train batch 9500 [2023-11-25 08:00:06,404][root][INFO] - Avg. loss per last 100 batches: 0.164591 [2023-11-25 08:00:07,121][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.8/29522=99.99% | mean: 0.03 | max: 5.57 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the road starts at the roundabout near mariehamn hospital, which also originates on ### ### highway 1 and highway 2. ### ### [P_TEXT]: mariehamn airlines including finnair. at only three of aland's four highways, mariehamn ### ### serves as a starting point for main road 1 to eckero, main road 2 to sund and main road 3 to ### ### lumparland. the city is an important centre for aland media ; both of the local newspapers ( ### ### alandstidningen and nya aland ), several radio stations and the local tv channels ( tv aland and ### ### aland24 ) operate out of the city. the islanders are ### ### =============================================== v_q ================================================ ### ### ('marie', 0, 1) ('##ham', 1, 10) ('highway', 2, 29) ('hospital', 3, 5259) ('1', 4, 42) ('2', 5, 89) ### ### ('¹', 6, 287) ('##n', 7, 37) ('##₁', 8, 21) ('absently', 9, 179) ('intently', 10, 170) ### ### ('road', 11, 15) ('##₂', 12, 28) ('³', 13, 457) ('##ն', 14, 117) ('##ன', 15, 100) ### ### ('alberta', 16, 8956) ('.', 17, 919) ('newfoundland', 18, 373) ('canada', 19, 2036) ### ### =============================================== v_p ================================================ ### ### ('lump', 9630, 0) ('marie', 0, 1) ('islanders', 3339, 2) ('alan', 19097, 3) ('finn', 2714, 4) ### ### ('##air', 7786, 5) ('ny', 12060, 6) ('##ker', 19518, 7) ('ec', 12148, 8) ('sun', 7431, 9) ### ### ('##ham', 1, 10) ('##ning', 6541, 11) ('##24', 2521, 12) ('##d', 7421, 13) ('main', 5385, 14) ### ### ('road', 11, 15) ('airline', 19562, 16) ('airlines', 14098, 17) ('highways', 287, 18) ### ### ('##land', 6071, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('marie', 0, 1) ('##ham', 1, 10) ('highway', 2, 29) ('road', 11, 15) ('##₁', 8, 21) ('1', 4, 42) ### ### ('##n', 7, 37) ('##₂', 12, 28) ('2', 5, 89) ('roads', 28, 31) ('motorway', 39, 30) ### ### ('anyway', 81, 23) ('##ன', 15, 100) ('intently', 10, 170) ('##ն', 14, 117) ('absently', 9, 179) ### ### ('highways', 287, 18) ('¹', 6, 287) ('worried', 62, 46) ('centre', 126, 33) ### ############################################################################################################ [2023-11-25 08:00:07,625][root][INFO] - Epoch: 2: Step: 9501/19530, loss[v]=0.190540, lr=0.000018, acc@1[1]=974.0/1024=0.951171875, acc@1[2]=997.5/1024=0.97412109375 [2023-11-25 08:01:57,820][root][INFO] - Train batch 9600 [2023-11-25 08:01:57,820][root][INFO] - Avg. loss per last 100 batches: 0.166146 [2023-11-25 08:01:58,551][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.0/29522=99.99% | mean: 0.03 | max: 5.53 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the overnight stopping point was in uxbridge. ### ### [P_TEXT]: blackstone river greenway uxbridge and blackstone. the path roughly parallels the right - ### ### of - way of the active providence & worcester railroad and its predecessor, the blackstone canal, ### ### running alongside the blackstone river. in some places, the bike path follows the old canal ### ### towpath, with the long - abandoned canal running along one side and the river on the other. once ### ### finished, the greenway will be linked to the east bay bike path, for an additional of trail to ### ### bristol, rhode island. the blackstone river greenway is also a ### ### =============================================== v_q ================================================ ### ### ('##bridge', 0, 5) ('u', 1, 17) ('##x', 2, 24) ('stopping', 3, 3115) ('¹', 4, 140) ### ### ('intently', 5, 110) ('absently', 6, 126) ('overnight', 7, 422) ('stop', 8, 5274) ### ### ('stops', 9, 4461) ('³', 10, 226) ('##ʂ', 11, 2172) ('massachusetts', 12, 68) ('sprang', 13, 51) ### ### ('hampshire', 14, 1005) ('halt', 15, 13984) ('ந', 16, 267) ('discontinued', 17, 756) ### ### ('twinned', 18, 196) ('.', 19, 3093) ### ### =============================================== v_p ================================================ ### ### ('blacks', 15381, 0) ('bristol', 1494, 1) ('worcester', 557, 2) ('path', 3064, 3) ### ### ('providence', 320, 4) ('##bridge', 0, 5) ('##way', 1756, 6) ('##tone', 9855, 7) ('canal', 278, 8) ### ### ('railroad', 43, 9) ('tow', 9425, 10) ('trail', 710, 11) ('bike', 6535, 12) ('bikes', 23127, 13) ### ### ('right', 6213, 14) ('rhode', 2007, 15) ('##path', 5744, 16) ('u', 1, 17) ('anyway', 196, 18) ### ### ('river', 760, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##bridge', 0, 5) ('u', 1, 17) ('##x', 2, 24) ('railroad', 43, 9) ('worcester', 557, 2) ### ### ('providence', 320, 4) ('canal', 278, 8) ('intently', 5, 110) ('sprang', 13, 51) ('¹', 4, 140) ### ### ('bristol', 1494, 1) ('worried', 150, 20) ('anyway', 196, 18) ('absently', 6, 126) ### ### ('massachusetts', 12, 68) ('trail', 710, 11) ('abandoned', 392, 25) ('##way', 1756, 6) ### ### ('route', 60, 43) ('river', 760, 19) ### ############################################################################################################ [2023-11-25 08:01:59,056][root][INFO] - Epoch: 2: Step: 9601/19530, loss[v]=0.181424, lr=0.000018, acc@1[1]=972.5/1024=0.94970703125, acc@1[2]=1000.5/1024=0.97705078125 [2023-11-25 08:03:49,283][root][INFO] - Train batch 9700 [2023-11-25 08:03:49,284][root][INFO] - Avg. loss per last 100 batches: 0.168490 [2023-11-25 08:03:50,023][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.69 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.8/29522=99.99% | mean: 0.03 | max: 5.50 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.5/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it is the capital of malbork county and the population was estimated at 38, 723, as of ### ### 2018. ### ### [P_TEXT]: lipinka, pomeranian voivodeship lipinka ( german " lindenau " ) is a village in the ### ### administrative district of gmina nowy staw, within malbork county, pomeranian voivodeship, in ### ### northern poland. it lies approximately east of nowy staw, north - east of malbork, and south - east ### ### of the regional capital gdansk. before 1772 the area was part of kingdom of poland, 1772 - 1919 ### ### prussia and germany, 1920 - 1939 free city of danzig, 1939 - february 1945 nazi germany. for the ### ### history of the region, see " ### ### =============================================== v_q ================================================ ### ### ('mal', 0, 3) ('##bor', 1, 4) ('##k', 2, 23) ('county', 3, 20) ('¹', 4, 67) ('intently', 5, 79) ### ### ('absently', 6, 109) ('estonia', 7, 3446) ('³', 8, 303) ('capital', 9, 26) ('##ք', 10, 75) ### ### ('##क', 11, 32) ('##ʂ', 12, 62) ('prefecture', 13, 4459) ('ந', 14, 160) ('cambodia', 15, 10646) ### ### ('korea', 16, 2402) ('.', 17, 2266) ('cradled', 18, 33) ('endemic', 19, 3424) ### ### =============================================== v_p ================================================ ### ### ('##ink', 20437, 0) ('##aw', 21194, 1) ('lip', 14841, 2) ('mal', 0, 3) ('##bor', 1, 4) ### ### ('leaned', 40, 5) ('linden', 24272, 6) ('worried', 308, 7) ('danzig', 12746, 8) ('now', 8068, 9) ### ### ('1772', 3376, 10) ('prussia', 4543, 11) ('##y', 3399, 12) ('seat', 1106, 13) ('gdansk', 1536, 14) ### ### ('anyway', 53, 15) ('st', 14801, 16) ('##a', 9549, 17) ('pomeranian', 21331, 18) ### ### ('poland', 354, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('mal', 0, 3) ('##bor', 1, 4) ('county', 3, 20) ('##k', 2, 23) ('leaned', 40, 5) ('capital', 9, 26) ### ### ('##क', 11, 32) ('¹', 4, 67) ('anyway', 53, 15) ('counties', 61, 21) ('cradled', 18, 33) ### ### ('intently', 5, 79) ('worried', 308, 7) ('##ք', 10, 75) ('absently', 6, 109) ('##ʂ', 12, 62) ### ### ('poland', 354, 19) ('desperately', 55, 48) ('³', 8, 303) ('population', 38, 60) ### ############################################################################################################ [2023-11-25 08:03:50,526][root][INFO] - Epoch: 2: Step: 9701/19530, loss[v]=0.243272, lr=0.000018, acc@1[1]=958.0/1024=0.935546875, acc@1[2]=988.5/1024=0.96533203125 [2023-11-25 08:05:40,721][root][INFO] - Train batch 9800 [2023-11-25 08:05:40,722][root][INFO] - Avg. loss per last 100 batches: 0.152010 [2023-11-25 08:05:41,438][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 4.05 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.9/29522=99.99% | mean: 0.02 | max: 5.57 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he recovered in time to lead a brigade in jeb stuart's third ride around the union army, ### ### in the early days of the gettysburg campaign, with his most significant contribution being at the ### ### battle of carlisle. ### ### [P_TEXT]: battle of carlisle behind in york county to guard a train of 125 captured federal supply ### ### wagons. instead of finding ewell, stuart encountered smith's militiamen. despite having a large ### ### numerical advantage, stuart's troopers were too exhausted from a month of campaigning to attack the ### ### town outright, and stuart initially feared that the enemy troops were veterans from the army of the ### ### potomac. after learning that smith's men were only militia, stuart sent maj. gen. fitzhugh lee into ### ### carlisle with a white flag, telling smith to either evacuate the town or clear out the women and ### ### children. ### ### =============================================== v_q ================================================ ### ### ('carlisle', 0, 0) ('gettysburg', 1, 537) ('stuart', 2, 9) ('je', 3, 28430) ('##b', 4, 14649) ### ### ('battle', 5, 7) ('bellowed', 6, 64) ('james', 7, 3118) ('campaign', 8, 58) ('1812', 9, 3195) ### ### ('generals', 10, 107) ('seemed', 11, 154) ('angrily', 12, 118) ('battles', 13, 53) ('¹', 14, 238) ### ### ('union', 15, 2322) ('recovered', 16, 8168) ('henry', 17, 5230) ('lincoln', 18, 1674) ### ### ('virginia', 19, 272) ### ### =============================================== v_p ================================================ ### ### ('carlisle', 0, 0) ('wagons', 9135, 1) ('##well', 7839, 2) ('york', 256, 3) ('fitz', 5041, 4) ### ### ('flag', 4031, 5) ('125', 3290, 6) ('battle', 5, 7) ('potomac', 25, 8) ('stuart', 2, 9) ### ### ('lee', 492, 10) ('anyway', 434, 11) ('evacuated', 1763, 12) ('e', 19321, 13) ('militia', 633, 14) ### ### ('##hu', 23608, 15) ('flags', 5175, 16) ('army', 40, 17) ('white', 4879, 18) ### ### ('outright', 11412, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('carlisle', 0, 0) ('stuart', 2, 9) ('battle', 5, 7) ('potomac', 25, 8) ('army', 40, 17) ### ### ('gen', 30, 26) ('york', 256, 3) ('bellowed', 6, 64) ('campaign', 8, 58) ('battles', 13, 53) ### ### ('guard', 63, 30) ('lee', 492, 10) ('anyway', 434, 11) ('soldiers', 62, 45) ('littered', 33, 56) ### ### ('gettysburg', 1, 537) ('troops', 95, 38) ('enemy', 114, 35) ('militia', 633, 14) ### ### ('generals', 10, 107) ### ############################################################################################################ [2023-11-25 08:05:41,943][root][INFO] - Epoch: 2: Step: 9801/19530, loss[v]=0.085867, lr=0.000017, acc@1[1]=1000.0/1024=0.9765625, acc@1[2]=1014.5/1024=0.99072265625 [2023-11-25 08:07:32,283][root][INFO] - Train batch 9900 [2023-11-25 08:07:32,284][root][INFO] - Avg. loss per last 100 batches: 0.168644 [2023-11-25 08:07:33,001][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.5/29522=99.99% | mean: 0.03 | max: 5.82 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the other honorees included john horvath, janos aczel, akos csaszar and laszlo fuchs. ### ### [P_TEXT]: akos csaszar east of budapest. an infectious illness spread in the camp, and his brother ### ### and father died, but akos survived. he is a member of the group of five students of the late ### ### professor lipot fejer who called them " the big five ". the other four are john horvath, janos ### ### aczel, steven gaal and laszlo fuchs, all of whom are now retired mathematics professors in north ### ### america. only csaszar became a university professor in budapest. between 1952 and 1992 he was head ### ### of the department of analysis at the ### ### =============================================== v_q ================================================ ### ### ('##zel', 0, 5) ('##chs', 1, 7) ('csa', 2, 1) ('fu', 3, 12) ('##th', 4, 26) ('ho', 5, 24) ### ### ('ak', 6, 4) ('##rva', 7, 9) ('john', 8, 34) ('laszlo', 9, 6) ('ac', 10, 18) ('jan', 11, 21) ### ### ('##sz', 12, 29) ('##ʂ', 13, 476) ('hungarian', 14, 27) ('##ar', 15, 28) ('honor', 16, 12357) ### ### ('intently', 17, 180) ('hungary', 18, 19) ('budapest', 19, 17) ### ### =============================================== v_p ================================================ ### ### ('gaa', 15531, 0) ('csa', 2, 1) ('##jer', 14936, 2) ('steven', 2879, 3) ('ak', 6, 4) ### ### ('##zel', 0, 5) ('laszlo', 9, 6) ('##chs', 1, 7) ('infectious', 4728, 8) ('##rva', 7, 9) ### ### ('analysis', 431, 10) ('lip', 15055, 11) ('fu', 3, 12) ('mathematics', 1077, 13) ### ### ('five', 12453, 14) ('littered', 1228, 15) ('big', 6052, 16) ('budapest', 19, 17) ('ac', 10, 18) ### ### ('hungary', 18, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('csa', 2, 1) ('##zel', 0, 5) ('##chs', 1, 7) ('ak', 6, 4) ('laszlo', 9, 6) ('fu', 3, 12) ### ### ('##rva', 7, 9) ('ac', 10, 18) ('ho', 5, 24) ('##th', 4, 26) ('jan', 11, 21) ('john', 8, 34) ### ### ('##sz', 12, 29) ('hungarian', 14, 27) ('budapest', 19, 17) ('hungary', 18, 19) ('##os', 24, 22) ### ### ('##ar', 15, 28) ('he', 34, 31) ('analysis', 431, 10) ### ############################################################################################################ [2023-11-25 08:07:33,506][root][INFO] - Epoch: 2: Step: 9901/19530, loss[v]=0.153917, lr=0.000017, acc@1[1]=987.5/1024=0.96435546875, acc@1[2]=1009.5/1024=0.98583984375 [2023-11-25 08:09:23,653][root][INFO] - Train batch 10000 [2023-11-25 08:09:23,654][root][INFO] - Avg. loss per last 100 batches: 0.171383 [2023-11-25 08:09:24,386][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.70 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.7/29522=99.99% | mean: 0.03 | max: 5.34 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: nh 14 originates from its junction with nh 12 at morgram in murshidabad district and ### ### passes through lohapur ( a little off the highway ), nalhati, rampurhat, mallarpur, gonpur, deucha, ### ### mohammad bazar, tilpara barrage across mayurakshi river, suri, ### ### [P_TEXT]: lohapur is off national highway 14, running from morgram ( in murshidabad district ) to ### ### kharagpur ( in paschim medinipur district ) and state highway 7 running from rajgram ( in birbhum ### ### district ) to midnapore ( in paschim medinipur district ) ( it is a common route in this section ). ### ### there is a station at lohapur on the nalhati - azimganj branch line. lohapur has a delivery sub ### ### post office, with pin 731237, under rampurhat head office. branch offices using the same pin are ### ### situated at bandhkhola, bara, bhadrapur, gopalchak, jesthabhanipur, kantagoria, krishnapur, ### ### noapara, salisand, sitalpur and ujirpur. ### ### =============================================== v_q ================================================ ### ### ('14', 0, 1) ('absently', 1, 127) ('nh', 2, 17053) ('.', 3, 300) ('intently', 4, 86) ('lo', 5, 3) ### ### ('mor', 6, 11) ('mohammad', 7, 10470) ('##uch', 8, 23666) ('ramp', 9, 14) ('sur', 10, 18012) ### ### ('¹', 11, 89) ('##zar', 12, 28047) ('pakistan', 13, 694) ('mall', 14, 8847) ('india', 15, 58) ### ### ('pradesh', 16, 5247) ('##hid', 17, 5) ('barrage', 18, 1142) ('12', 19, 452) ### ### =============================================== v_p ================================================ ### ### ('##gram', 32, 0) ('14', 0, 1) ('##gpur', 468, 2) ('lo', 5, 3) ('sit', 4352, 4) ('##hid', 17, 5) ### ### ('7', 6257, 6) ('##isan', 22251, 7) ('kant', 11455, 8) ('##ha', 31, 9) ('nearest', 664, 10) ### ### ('mor', 6, 11) ('##ola', 9915, 12) ('##pur', 20, 13) ('ramp', 9, 14) ('##hara', 2245, 15) ### ### ('##hum', 18563, 16) ('highway', 70, 17) ('delivery', 7078, 18) ('mid', 5129, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('14', 0, 1) ('lo', 5, 3) ('##hid', 17, 5) ('mor', 6, 11) ('##gram', 32, 0) ('ramp', 9, 14) ### ### ('##pur', 20, 13) ('##ha', 31, 9) ('fourteen', 21, 29) ('##rs', 33, 32) ('##abad', 41, 26) ### ### ('india', 15, 58) ('absently', 1, 127) ('intently', 4, 86) ('highway', 70, 17) ('¹', 11, 89) ### ### ('outskirts', 96, 30) ('##gpur', 468, 2) ('district', 52, 52) ('mu', 57, 56) ### ############################################################################################################ [2023-11-25 08:09:24,889][root][INFO] - Epoch: 2: Step: 10001/19530, loss[v]=0.098188, lr=0.000017, acc@1[1]=997.5/1024=0.97412109375, acc@1[2]=1011.5/1024=0.98779296875 [2023-11-25 08:11:15,274][root][INFO] - Train batch 10100 [2023-11-25 08:11:15,275][root][INFO] - Avg. loss per last 100 batches: 0.173948 [2023-11-25 08:11:16,012][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.72 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.3/29522=99.98% | mean: 0.03 | max: 5.46 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: premier league viewership has increased rapidly, with nbc and nbcsn averaging a record ### ### 479, 000 viewers in the 2014 – 15 season, up 118 % from 2012 – 13 when coverage still aired on fox ### ### soccer and espn / espn2 ( 220, 000 viewers ), and nbc sports has been widely praised for its ### ### coverage. ### ### [P_TEXT]: 2014 – 15 premier league the 2014 – 15 premier league ( known as the barclays premier ### ### league for sponsorship reasons ) was the 23rd season of the premier league, the top english ### ### professional league for association football clubs, since its establishment in 1992. the fixtures ### ### were announced on 18 june 2014. the season started on 16 august 2014 and concluded on 24 may 2015. ### ### manchester city came into the season as defending champions of the 2013 – 14 season. leicester ### ### city, burnley and queens park rangers entered as the three promoted teams. on 3 ### ### =============================================== v_q ================================================ ### ### ('premier', 0, 9) ('absently', 1, 96) ('intently', 2, 90) ('15', 3, 2) ('¹', 4, 95) ### ### ('soccer', 5, 383) ('nbc', 6, 17094) ('2014', 7, 3) ('.', 8, 60) ('2012', 9, 445) ('13', 10, 549) ### ### ('fox', 11, 14204) ('³', 12, 177) ('espn', 13, 1870) ('##n', 14, 22983) ('league', 15, 13) ### ### ('##ன', 16, 2114) ('2013', 17, 16) ('season', 18, 12) ('貴', 19, 101) ### ### =============================================== v_p ================================================ ### ### ('barclay', 3182, 0) ('manchester', 151, 1) ('15', 3, 2) ('2014', 7, 3) ('leicester', 333, 4) ### ### ('burnley', 947, 5) ('14', 223, 6) ('1992', 5183, 7) ('rangers', 1043, 8) ('premier', 0, 9) ### ### ('23rd', 12343, 10) ('shirt', 79, 11) ('season', 18, 12) ('league', 15, 13) ('anyway', 45, 14) ### ### ('clubs', 167, 15) ('2013', 17, 16) ('fifteen', 250, 17) ('fixtures', 41, 18) ### ### ('defending', 5201, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('premier', 0, 9) ('15', 3, 2) ('2014', 7, 3) ('league', 15, 13) ('season', 18, 12) ### ### ('2013', 17, 16) ('manchester', 151, 1) ('absently', 1, 96) ('shirt', 79, 11) ('anyway', 45, 14) ### ### ('.', 8, 60) ('fixtures', 41, 18) ('2015', 26, 27) ('intently', 2, 90) ('sponsorship', 24, 32) ### ### ('14', 223, 6) ('¹', 4, 95) ('footballer', 21, 40) ('leicester', 333, 4) ('worried', 38, 29) ### ############################################################################################################ [2023-11-25 08:11:16,518][root][INFO] - Epoch: 2: Step: 10101/19530, loss[v]=0.130215, lr=0.000017, acc@1[1]=992.5/1024=0.96923828125, acc@1[2]=1008.5/1024=0.98486328125 [2023-11-25 08:13:06,925][root][INFO] - Train batch 10200 [2023-11-25 08:13:06,926][root][INFO] - Avg. loss per last 100 batches: 0.173148 [2023-11-25 08:13:07,645][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.0/29522=99.99% | mean: 0.03 | max: 5.84 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the game is the first in its series to feature an original score, composed by a team of ### ### producers collaborating over several years. ### ### [P_TEXT]: music of grand theft auto v and selections from the in - game radio ; " the ### ### cinematographic score — gta 5 ", an electronic album released in march 2014, comprises tracks ### ### composed and produced by tangerine dream founder edgar froese ; and " welcome to los santos ", ### ### released with the microsoft windows version of the game, features songs from the in - game radio ### ### station " the lab ", produced by the alchemist and oh no. critical reception to the soundtracks was ### ### positive, as reviewers felt that the music connected appropriately with the gameplay. " grand theft ### ### auto v " is the first game in its series to use ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 145) ('.', 1, 2098) ('score', 2, 21) ('¹', 3, 294) ('game', 4, 16) ### ### ('intently', 5, 169) ('original', 6, 4578) ('³', 7, 535) ('series', 8, 70) ('″', 9, 45) ### ### ('"', 10, 40) ('godzilla', 11, 2618) ('anyway', 12, 20) ('nintendo', 13, 158) ('貴', 14, 282) ### ### ('music', 15, 30) ('worried', 16, 44) ('freaking', 17, 181) ('soundtrack', 18, 23) ('ந', 19, 251) ### ### =============================================== v_p ================================================ ### ### ('lab', 4431, 0) ('##oese', 22864, 1) ('##chemist', 4532, 2) ('oh', 2345, 3) ('santos', 21485, 4) ### ### ('gt', 3555, 5) ('edgar', 7494, 6) ('welcome', 10183, 7) ('theft', 1390, 8) ('auto', 4784, 9) ### ### ('los', 18002, 10) ('radio', 7332, 11) ('v', 6449, 12) ('cinema', 1904, 13) ('fr', 22311, 14) ### ### ('no', 3439, 15) ('game', 4, 16) ('##erine', 3732, 17) ('2014', 648, 18) ('##graphic', 5904, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('game', 4, 16) ('score', 2, 21) ('anyway', 12, 20) ('music', 15, 30) ('soundtrack', 18, 23) ### ### ('″', 9, 45) ('"', 10, 40) ('worried', 16, 44) ('absently', 0, 145) ('games', 22, 46) ### ### ('series', 8, 70) ('intently', 5, 169) ('producers', 20, 75) ('producer', 26, 58) ### ### ('electronic', 74, 32) ('gameplay', 154, 26) ('¹', 3, 294) ('dream', 113, 43) ('shocked', 36, 61) ### ### ('hissed', 82, 48) ### ############################################################################################################ [2023-11-25 08:13:08,149][root][INFO] - Epoch: 2: Step: 10201/19530, loss[v]=0.100559, lr=0.000017, acc@1[1]=997.0/1024=0.9736328125, acc@1[2]=1012.5/1024=0.98876953125 [2023-11-25 08:14:58,533][root][INFO] - Train batch 10300 [2023-11-25 08:14:58,534][root][INFO] - Avg. loss per last 100 batches: 0.172936 [2023-11-25 08:14:59,278][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.62 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.0/29522=99.99% | mean: 0.02 | max: 5.39 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 29.0/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: , south uist, and barra. ### ### [P_TEXT]: mgm - 5 corporal there were six battalions in germany, two in italy and four in the ### ### united states. live - fire training for germany - based us forces took place at fort bliss but ### ### later the british royal artillery guided weapons range on the scottish island of south uist in the ### ### outer hebrides was used. missiles were fired toward designated target coordinates in the atlantic ### ### ocean. radar on hirta ( the main island of the st kilda archipelago of scotland ) identified ### ### missile landing points. frequently, soviet spy ship'fishing trawlers'would intrude into the target ### ### area. the uk ### ### =============================================== v_q ================================================ ### ### ('barr', 0, 21491) ('ui', 1, 2) ('##a', 2, 26627) ('##st', 3, 11) ('absently', 4, 133) ### ### ('¹', 5, 202) ('intently', 6, 173) ('south', 7, 48) ('.', 8, 287) ('³', 9, 316) ### ### ('shetland', 10, 19345) ('irish', 11, 124) ('ந', 12, 356) ('##sts', 13, 75) ('scotland', 14, 37) ### ### ('anyway', 15, 20) ('schooner', 16, 1172) ('hms', 17, 482) ('ireland', 18, 340) ('##а', 19, 13753) ### ### =============================================== v_p ================================================ ### ### ('##rta', 23833, 0) ('kilda', 3613, 1) ('ui', 1, 2) ('mgm', 12686, 3) ('bliss', 13538, 4) ### ### ('corporal', 1204, 5) ('guided', 20000, 6) ('hi', 22928, 7) ('fort', 435, 8) ('missiles', 24668, 9) ### ### ('archipelago', 429, 10) ('##st', 3, 11) ('spy', 19566, 12) ('missile', 16580, 13) ### ### ('range', 4509, 14) ('radar', 3178, 15) ('fishing', 277, 16) ('artillery', 1742, 17) ### ### ('weapons', 19950, 18) ('##aw', 20028, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ui', 1, 2) ('##st', 3, 11) ('south', 7, 48) ('anyway', 15, 20) ('scotland', 14, 37) ### ### ('scottish', 35, 22) ('worried', 34, 23) ('battalion', 33, 29) ('island', 69, 36) ### ### ('absently', 4, 133) ('islands', 20, 64) ('##sts', 13, 75) ('fishing', 277, 16) ('fort', 435, 8) ### ### ('archipelago', 429, 10) ('intently', 6, 173) ('¹', 5, 202) ('corporal', 1204, 5) ### ### ('atlantic', 315, 34) ('leaned', 79, 56) ### ############################################################################################################ [2023-11-25 08:14:59,782][root][INFO] - Epoch: 2: Step: 10301/19530, loss[v]=0.270862, lr=0.000017, acc@1[1]=955.0/1024=0.9326171875, acc@1[2]=995.5/1024=0.97216796875 [2023-11-25 08:16:50,060][root][INFO] - Train batch 10400 [2023-11-25 08:16:50,061][root][INFO] - Avg. loss per last 100 batches: 0.179936 [2023-11-25 08:16:50,812][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.66 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.5/29522=99.99% | mean: 0.03 | max: 5.43 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: excluded are a range of related diagnoses, including dysthymia, which involves a chronic ### ### but milder mood disturbance ; recurrent brief depression, consisting of briefer depressive episodes ### ### ; minor depressive disorder, whereby only some symptoms of major depression are present ; and ### ### adjustment disorder with depressed mood, which denotes low mood resulting from a psychological ### ### response to an identifiable event or stressor. ### ### [P_TEXT]: john hinckley jr. has schizophrenia. park dietz, a forensic psychiatrist who testified ### ### for the prosecution, diagnosed hinckley with narcissistic and schizoid personality disorder s and ### ### dysthymia, as well as borderline and passive - aggressive features. at the hospital hinckley was ### ### treated for narcissistic and schizotypal personality disorder and major depressive disorder. ### ### hinckley was confined at st. elizabeths hospital in washington, d. c. after hinckley was admitted, ### ### tests found that he was an " unpredictably dangerous " man who might harm himself or any third ### ### party. in ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 175) ('##yst', 1, 10) ('.', 2, 157) ('adjustment', 3, 5898) ('##pressive', 4, 24) ### ### ('intently', 5, 228) ('mood', 6, 104) ('stress', 7, 5504) ('¹', 8, 216) ('minor', 9, 776) ### ### ('depressed', 10, 1517) ('##hy', 11, 19) ('diagnosed', 12, 566) ('depression', 13, 115) ### ### ('excluded', 14, 6133) ('de', 15, 333) ('brief', 16, 13596) ('worried', 17, 26) ('##mia', 18, 39) ### ### ('ד', 19, 318) ### ### =============================================== v_p ================================================ ### ### ('passive', 10828, 0) ('diet', 3963, 1) ('elizabeth', 5265, 2) ('border', 20911, 3) ### ### ('##line', 8676, 4) ('##istic', 3077, 5) ('##kley', 25526, 6) ('##zo', 23885, 7) ### ### ('aggressive', 995, 8) ('hospital', 46, 9) ('##yst', 1, 10) ('##nc', 27104, 11) ### ### ('personality', 1643, 12) ('forensic', 6260, 13) ('symptoms', 203, 14) ('disorder', 20, 15) ### ### ('washington', 19481, 16) ('##iss', 21377, 17) ('prosecution', 23188, 18) ('##hy', 11, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##yst', 1, 10) ('##pressive', 4, 24) ('##hy', 11, 19) ('disorder', 20, 15) ('worried', 17, 26) ### ### ('hospital', 46, 9) ('##mia', 18, 39) ('anyway', 26, 32) ('absently', 0, 175) ('symptoms', 203, 14) ### ### ('schizophrenia', 136, 23) ('d', 21, 65) ('mood', 6, 104) ('.', 2, 157) ('major', 42, 51) ### ### ('depression', 13, 115) ('intently', 5, 228) ('aggressive', 995, 8) ('leaned', 405, 28) ### ### ('¹', 8, 216) ### ############################################################################################################ [2023-11-25 08:16:51,319][root][INFO] - Epoch: 2: Step: 10401/19530, loss[v]=0.229274, lr=0.000017, acc@1[1]=969.5/1024=0.94677734375, acc@1[2]=987.5/1024=0.96435546875 [2023-11-25 08:18:41,397][root][INFO] - Train batch 10500 [2023-11-25 08:18:41,397][root][INFO] - Avg. loss per last 100 batches: 0.178841 [2023-11-25 08:18:42,122][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.73 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.5/29522=99.99% | mean: 0.03 | max: 5.55 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: haut - lomami, haut - katanga, lualaba, and tanganyika provinces are the result of the ### ### dismemberment of the former katanga province. ### ### [P_TEXT]: lukushi river the lukushi river is a tributary of the luvua river. it runs from south to ### ### north through the malemba - nkulu territory of haut - lomami province and the manono territory of ### ### tanganyika province, passing the twin tin - mining towns of kitotolo and manono shortly before ### ### entering the luvua. in the early 1800s, king kumwimbe of the luba kingdom created a client state ### ### that united the hemba villages of the lukushi river valley, and that played an important role in ### ### preserving luba dominance over other small states in the region. in ### ### =============================================== v_q ================================================ ### ### ('##nga', 0, 21670) ('province', 1, 14) ('tanzania', 2, 173) ('absently', 3, 138) ('kenya', 4, 249) ### ### ('zimbabwe', 5, 84) ('¹', 6, 97) ('.', 7, 1106) ('rwanda', 8, 786) ('##ala', 9, 12344) ### ### ('tang', 10, 31) ('zambia', 11, 1791) ('malawi', 12, 1121) ('intently', 13, 133) ### ### ('africa', 14, 284) ('leone', 15, 17957) ('leaned', 16, 61) ('##ut', 17, 32) ('ha', 18, 42) ### ### ('##mi', 19, 35) ### ### =============================================== v_p ================================================ ### ### ('hem', 16031, 0) ('##kus', 10128, 1) ('##vu', 736, 2) ('##ono', 7045, 3) ('##ulu', 2132, 4) ### ### ('##oto', 8774, 5) ('lu', 27, 6) ('kit', 3266, 7) ('client', 12269, 8) ('male', 1273, 9) ### ### ('dominance', 5425, 10) ('man', 17387, 11) ('##mbe', 410, 12) ('provinces', 89, 13) ### ### ('province', 1, 14) ('##hi', 13395, 15) ('##ba', 24, 16) ('1800s', 18013, 17) ('river', 6720, 18) ### ### ('rivers', 2333, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('province', 1, 14) ('lu', 27, 6) ('##ba', 24, 16) ('tang', 10, 31) ('##ut', 17, 32) ('lo', 23, 24) ### ### ('##ika', 26, 27) ('##mi', 19, 35) ('ha', 18, 42) ('provinces', 89, 13) ('zimbabwe', 5, 84) ### ### ('leaned', 16, 61) ('##any', 54, 36) ('¹', 6, 97) ('worried', 58, 40) ('nigeria', 22, 78) ### ### ('territory', 101, 33) ('absently', 3, 138) ('##vu', 736, 2) ('fiji', 21, 94) ### ############################################################################################################ [2023-11-25 08:18:42,627][root][INFO] - Epoch: 2: Step: 10501/19530, loss[v]=0.189174, lr=0.000017, acc@1[1]=976.5/1024=0.95361328125, acc@1[2]=1002.0/1024=0.978515625 [2023-11-25 08:20:32,801][root][INFO] - Train batch 10600 [2023-11-25 08:20:32,801][root][INFO] - Avg. loss per last 100 batches: 0.173434 [2023-11-25 08:20:33,563][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.03 | max: 3.63 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.5/29522=99.99% | mean: 0.03 | max: 5.71 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: with barrett o'neill he made foundational contributions to the study of convex functions ### ### and convex sets in riemannian geometry and their applications in the study of negative sectional ### ### curvature, including to the geometry of warped products. ### ### [P_TEXT]: bishop – gromov inequality in mathematics, the bishop – gromov inequality is a comparison ### ### theorem in riemannian geometry, named after richard l. bishop and mikhail gromov. it is closely ### ### related to myers'theorem, and is the key point in the proof of gromov's compactness theorem. let ### ### formula _ 1 be a complete " n " - dimensional riemannian manifold whose ricci curvature satisfies ### ### the lower bound for a constant formula _ 3. let formula _ 4 be the complete " n " - dimensional ### ### simply connected space of constant sectional curvature formula _ 5 ( and hence of constant ricci ### ### curvature formula _ 6 ) ; thus formula _ 4 is the " n " - sphere of ### ### =============================================== v_q ================================================ ### ### ('barrett', 0, 12297) ('warped', 1, 20967) ('neill', 2, 19954) ('convex', 3, 388) ### ### ('geometry', 4, 31) ('sectional', 5, 7) ('mathematician', 6, 1524) ('o', 7, 9460) ### ### ('negative', 8, 6389) ('littered', 9, 32) ('curvature', 10, 4) ('ri', 11, 30) ('pmid', 12, 726) ### ### ('distortion', 13, 789) ('mathematics', 14, 21) ('sets', 15, 3088) ('sparhawk', 16, 428) ### ### ('turing', 17, 11363) ('¹', 18, 92) ('contributions', 19, 19867) ### ### =============================================== v_p ================================================ ### ### ('myers', 8687, 0) ('##omo', 25328, 1) ('bishop', 7437, 2) ('compact', 4007, 3) ### ### ('curvature', 10, 4) ('ric', 11442, 5) ('sphere', 255, 6) ('sectional', 5, 7) ('richard', 367, 8) ### ### ('theorem', 38, 9) ('comparison', 2891, 10) ('inequality', 3354, 11) ('mikhail', 735, 12) ### ### ('manifold', 65, 13) ('##emann', 83, 14) ('complete', 2985, 15) ('constant', 844, 16) ### ### ('connected', 12808, 17) ('##ness', 6613, 18) ('proof', 2180, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('sectional', 5, 7) ('curvature', 10, 4) ('geometry', 4, 31) ('ri', 11, 30) ('littered', 9, 32) ### ### ('mathematics', 14, 21) ('theorem', 38, 9) ('manifold', 65, 13) ('##emann', 83, 14) ### ### ('##ian', 20, 40) ('sphere', 255, 6) ('richard', 367, 8) ('space', 157, 28) ('mikhail', 735, 12) ### ### ('constant', 844, 16) ('formula', 200, 38) ('contrast', 89, 47) ('convex', 3, 388) ### ### ('anyway', 891, 24) ('compact', 4007, 3) ### ############################################################################################################ [2023-11-25 08:20:34,068][root][INFO] - Epoch: 2: Step: 10601/19530, loss[v]=0.253393, lr=0.000017, acc@1[1]=962.5/1024=0.93994140625, acc@1[2]=994.0/1024=0.970703125 [2023-11-25 08:22:24,097][root][INFO] - Train batch 10700 [2023-11-25 08:22:24,097][root][INFO] - Avg. loss per last 100 batches: 0.168164 [2023-11-25 08:22:24,814][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.61 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.2/29522=99.98% | mean: 0.02 | max: 5.30 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: master's degree ( 1970 ) and phd ( 1974 ) from the university of chicago. ### ### [P_TEXT]: university of chicago david graeber and donald johanson, who is best known for ### ### discovering the fossil of a female hominid australopithecine known as " lucy " in the afar triangle ### ### region, psychologist john b. watson, american psychologist who established the psychological school ### ### of behaviorism, communication theorist harold innis, chess grandmaster samuel reshevsky, and ### ### conservative international relations scholar and white house coordinator of security planning for ### ### the national security council samuel p. huntington. american civil rights movement leaders vernon ### ### johns, considered by some to be the founder ### ### =============================================== v_q ================================================ ### ### ('chicago', 0, 1) ('.', 1, 227) ('¹', 2, 161) ('university', 3, 13) ('intently', 4, 156) ### ### ('isbn', 5, 171) ('absently', 6, 89) ('pmid', 7, 71) ('anyway', 8, 28) ('1970', 9, 13936) ### ### ('proponents', 10, 103) ('³', 11, 870) ('argues', 12, 7288) ('##¨', 13, 980) ('argued', 14, 4236) ### ### ('ethan', 15, 3762) ('littered', 16, 25) ('discusses', 17, 470) ('kramer', 18, 6513) ### ### ('phd', 19, 4559) ### ### =============================================== v_p ================================================ ### ### ('afar', 26215, 0) ('chicago', 0, 1) ('watson', 1743, 2) ('lucy', 21464, 3) ('triangle', 3801, 4) ### ### ('##ecin', 20295, 5) ('res', 697, 6) ('vernon', 6006, 7) ('johns', 1258, 8) ('huntington', 993, 9) ### ### ('behavior', 2024, 10) ('inn', 8213, 11) ('johan', 9751, 12) ('university', 3, 13) ### ### ('##vsky', 6174, 14) ('fossil', 5384, 15) ('psychologists', 23918, 16) ('relations', 3315, 17) ### ### ('chess', 1512, 18) ('harold', 315, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('chicago', 0, 1) ('university', 3, 13) ('anyway', 8, 28) ('littered', 16, 25) ('pmid', 7, 71) ### ### ('absently', 6, 89) ('res', 697, 6) ('harold', 315, 19) ('¹', 2, 161) ('john', 23, 64) ### ### ('intently', 4, 156) ('proponents', 10, 103) ('.', 1, 227) ('huntington', 993, 9) ('isbn', 5, 171) ### ### ('watson', 1743, 2) ('johns', 1258, 8) ('david', 35, 74) ('psychological', 417, 31) ### ### ('donald', 311, 35) ### ############################################################################################################ [2023-11-25 08:22:25,317][root][INFO] - Epoch: 2: Step: 10701/19530, loss[v]=0.079028, lr=0.000017, acc@1[1]=1000.0/1024=0.9765625, acc@1[2]=1017.5/1024=0.99365234375 [2023-11-25 08:24:14,764][root][INFO] - Train batch 10800 [2023-11-25 08:24:14,765][root][INFO] - Avg. loss per last 100 batches: 0.166209 [2023-11-25 08:24:15,494][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.71 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.8/29522=99.98% | mean: 0.02 | max: 5.32 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: on january 17, 2016, future released another mixtape, titled purple reign, with ### ### productions from metro boomin, southside, zaytoven, and more. ### ### [P_TEXT]: super slimey super slimey is a collaborative commercial mixtape by american rappers ### ### future and young thug. it was released on october 20, 2017, by 300 entertainment, atlantic records, ### ### epic records, freebandz and ysl records. the mixtape features a sole guest appearance from offset. ### ### meanwhile, the production includes southside, richie souf, wheezy and fuse, among others. in 2016, ### ### there were rumors that was circulating that future and young thug ### ### =============================================== v_q ================================================ ### ### ('purple', 0, 2834) ('future', 1, 9) ('reign', 2, 6752) ('metro', 3, 11915) ('boom', 4, 6358) ### ### ('mixtape', 5, 16) ('##yt', 6, 28003) ('##side', 7, 3) ('##oven', 8, 26250) ('absently', 9, 97) ### ### ('rapper', 10, 25) ('za', 11, 20070) ('2016', 12, 34) ('¹', 13, 185) ('hop', 14, 1014) ### ### ('south', 15, 48) ('.', 16, 89) ('intently', 17, 107) ('##in', 18, 23722) ('awesome', 19, 1810) ### ### =============================================== v_p ================================================ ### ### ('offset', 2187, 0) ('fuse', 1841, 1) ('slim', 42, 2) ('##side', 7, 3) ('300', 7360, 4) ### ### ('thug', 45, 5) ('##sl', 18037, 6) ('##uf', 16815, 7) ('##band', 16766, 8) ('future', 1, 9) ### ### ('richie', 5012, 10) ('##ey', 18316, 11) ('y', 1728, 12) ('rappers', 161, 13) ('rumors', 244, 14) ### ### ('super', 3941, 15) ('mixtape', 5, 16) ('young', 1550, 17) ('##hee', 24848, 18) ### ### ('atlantic', 6847, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('future', 1, 9) ('##side', 7, 3) ('mixtape', 5, 16) ('rapper', 10, 25) ('slim', 42, 2) ### ### ('thug', 45, 5) ('2016', 12, 34) ('entertainment', 25, 27) ('rappers', 161, 13) ('2017', 84, 21) ### ### ('south', 15, 48) ('artists', 30, 35) ('rumors', 244, 14) ('label', 69, 26) ('records', 88, 24) ### ### ('rap', 21, 43) ('absently', 9, 97) ('fuse', 1841, 1) ('anyway', 224, 23) ### ### ('collaborations', 95, 29) ### ############################################################################################################ [2023-11-25 08:24:15,998][root][INFO] - Epoch: 2: Step: 10801/19530, loss[v]=0.236386, lr=0.000017, acc@1[1]=969.0/1024=0.9462890625, acc@1[2]=1002.5/1024=0.97900390625 [2023-11-25 08:26:06,053][root][INFO] - Train batch 10900 [2023-11-25 08:26:06,054][root][INFO] - Avg. loss per last 100 batches: 0.152498 [2023-11-25 08:26:06,817][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.76 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.5/29522=99.98% | mean: 0.02 | max: 5.42 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it has a precision - made replica renault 70 hp engine. ### ### [P_TEXT]: royal aircraft factory f. e. 1 tailplane lifted to the top longeron s. the nacelle was by ### ### now deeper and more spacious, while the mainplanes were identical to those of the b. e. 2a. the ### ### gnome was replaced by a 70 hp ( 52 kw ) air cooled renault v - 8 engine. effectively, although the ### ### factory now routinely constructed original aircraft, it was another case of a new design reusing ### ### the designation of an older one. it was lost in a crash near wittering on 23 february 1914 when the ### ### pilot, r. kemp lost control while in a dive, kemp being unable to ### ### =============================================== v_q ================================================ ### ### ('renault', 0, 3) ('70', 1, 40) ('replica', 2, 588) ('precision', 3, 2847) ('engine', 4, 19) ### ### ('absently', 5, 213) ('hp', 6, 23) ('.', 7, 1978) ('¹', 8, 223) ('intently', 9, 248) ### ### ('airbus', 10, 7196) ('ɐ', 11, 2712) ('vehicles', 12, 738) ('peugeot', 13, 362) ('model', 14, 685) ### ### ('³', 15, 399) ('貴', 16, 203) ('auto', 17, 8991) ('dealers', 18, 18321) ('supplier', 19, 5915) ### ### =============================================== v_p ================================================ ### ### ('##tering', 1211, 0) ('gnome', 2691, 1) ('kemp', 23523, 2) ('renault', 0, 3) ('2a', 6077, 4) ### ### ('wit', 19889, 5) ('longer', 9085, 6) ('cooling', 6335, 7) ('cooled', 14825, 8) ('royal', 1118, 9) ### ### ('e', 8626, 10) ('factory', 244, 11) ('##cel', 14962, 12) ('dive', 16300, 13) ('##on', 13868, 14) ### ### ('v', 12153, 15) ('8', 7629, 16) ('crash', 427, 17) ('engines', 111, 18) ('engine', 4, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('renault', 0, 3) ('70', 1, 40) ('engine', 4, 19) ('hp', 6, 23) ('anyway', 42, 30) ### ### ('engines', 111, 18) ('##tering', 1211, 0) ('factory', 244, 11) ('littered', 191, 29) ### ### ('worried', 93, 39) ('crash', 427, 17) ('absently', 5, 213) ('¹', 8, 223) ('aircraft', 339, 24) ### ### ('gnome', 2691, 1) ('air', 114, 47) ('royal', 1118, 9) ('replica', 2, 588) ('intently', 9, 248) ### ### ('identical', 507, 33) ### ############################################################################################################ [2023-11-25 08:26:07,323][root][INFO] - Epoch: 2: Step: 10901/19530, loss[v]=0.242649, lr=0.000017, acc@1[1]=963.5/1024=0.94091796875, acc@1[2]=996.5/1024=0.97314453125 [2023-11-25 08:27:57,157][root][INFO] - Train batch 11000 [2023-11-25 08:27:57,157][root][INFO] - Avg. loss per last 100 batches: 0.162265 [2023-11-25 08:27:57,871][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.82 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.9/29522=99.99% | mean: 0.03 | max: 5.60 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: then created his own political party poland together and started cooperation with law and ### ### justice party as part of united right. ### ### [P_TEXT]: poland comes first deal after the election due in october. however, before the election, ### ### kluzik - rostkowska left the party for civic platform. co - founder adam bielan mep also left the ### ### pjn to sit as an independent, eventually joining jarosław gowin's new poland together party, whilst ### ### remaining within the ecr group in the eu parliament alongside the pis meps. at the october 2011 ### ### election, the party performed poorly, coming sixth in the sejm vote with 2. 2 %. this being below ### ### the 5 % threshold, the party failed to hold or win any seats in the sejm. the party ### ### =============================================== v_q ================================================ ### ### ('together', 0, 17) ('poland', 1, 2) ('right', 2, 3689) ('justice', 3, 2412) ('polish', 4, 31) ### ### ('party', 5, 24) ('united', 6, 8739) ('¹', 7, 135) ('##ʂ', 8, 339) ('！', 9, 1489) ('law', 10, 7606) ### ### ('intently', 11, 210) ('cooperation', 12, 1306) ('political', 13, 871) ('³', 14, 335) ### ### ('parties', 15, 67) ('politics', 16, 195) ('icao', 17, 1330) ('spokesperson', 18, 85) ### ### ('absently', 19, 172) ### ### =============================================== v_p ================================================ ### ### ('comes', 9292, 0) ('eu', 2106, 1) ('poland', 1, 2) ('2011', 612, 3) ('##win', 22639, 4) ### ### ('pi', 2642, 5) ('platform', 1708, 6) ('civic', 1167, 7) ('october', 2300, 8) ('##jm', 17650, 9) ### ### ('adam', 895, 10) ('##jn', 10364, 11) ('ec', 9373, 12) ('election', 208, 13) ('go', 10204, 14) ### ### ('littered', 27, 15) ('##ela', 13035, 16) ('together', 0, 17) ('elections', 74, 18) ### ### ('jar', 745, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('poland', 1, 2) ('together', 0, 17) ('party', 5, 24) ('polish', 4, 31) ('littered', 27, 15) ### ### ('elections', 74, 18) ('election', 208, 13) ('criticised', 33, 44) ('¹', 7, 135) ('2011', 612, 3) ### ### ('parties', 15, 67) ('anyway', 301, 21) ('hissed', 26, 66) ('adam', 895, 10) ('civic', 1167, 7) ### ### ('##kow', 123, 43) ('eu', 2106, 1) ('spokesperson', 18, 85) ('jar', 745, 19) ('##ʂ', 8, 339) ### ############################################################################################################ [2023-11-25 08:27:58,375][root][INFO] - Epoch: 2: Step: 11001/19530, loss[v]=0.230632, lr=0.000017, acc@1[1]=970.5/1024=0.94775390625, acc@1[2]=999.5/1024=0.97607421875 [2023-11-25 08:29:47,835][root][INFO] - Train batch 11100 [2023-11-25 08:29:47,835][root][INFO] - Avg. loss per last 100 batches: 0.165464 [2023-11-25 08:29:48,567][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.4/29522=99.99% | mean: 0.02 | max: 5.55 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the producers were unable to acquire the rights to use popeye, tom and jerry, little ### ### lulu, casper, or the terrytoons for appearances from their respective owners ( king features, ### ### turner, western publishing, harvey comics, and ### ### [P_TEXT]: willoughby ( looney tunes ) career was essentially over before the end of world war ii. ### ### these dogs were mostly similar character design in " porky's bear facts " ( 1941 ), " a corny ### ### concerto " ( 1943 ), " hare ribbin'" ( 1944 ), " the goofy gophers " ( 1946 ), and possibly " inki ### ### at the circus " ( 1947 ). willoughby was planned to be made as a cameo in the deleted scene " ### ### acme's funeral " from the 1988 film " who framed roger rabbit ". he also appears with other ### ### animated characters scared when casper appears ### ### =============================================== v_q ================================================ ### ### ('western', 0, 9621) ('casper', 1, 1) ('lulu', 2, 13816) ('absently', 3, 112) ('pope', 4, 21107) ### ### ('jerry', 5, 3229) ('turner', 6, 11191) ('harvey', 7, 7961) ('publishing', 8, 10215) ### ### ('comics', 9, 376) ('terry', 10, 12715) ('##ye', 11, 25503) ('king', 12, 12646) ('tom', 13, 6061) ### ### ('.', 14, 159) ('little', 15, 15809) ('¹', 16, 135) ('″', 17, 13) ('intently', 18, 165) ### ### ('cartoons', 19, 71) ### ### =============================================== v_p ================================================ ### ### ('willoughby', 5454, 0) ('casper', 1, 1) ('facts', 1216, 2) ('concerto', 25503, 3) ('ink', 7622, 4) ### ### ('##pher', 21100, 5) ('circus', 236, 6) ('funeral', 1748, 7) ('rabbit', 850, 8) ('goofy', 2273, 9) ### ### ('roger', 7702, 10) ('##me', 8275, 11) ('ac', 17457, 12) ('″', 17, 13) ('pork', 7857, 14) ### ### ('′', 2097, 15) ('hare', 3845, 16) ('##bin', 22644, 17) ('rib', 27538, 18) ('1944', 668, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('casper', 1, 1) ('″', 17, 13) ('"', 26, 33) ('circus', 236, 6) ('1946', 92, 23) ### ### ('absently', 3, 112) ('facts', 1216, 2) ('animated', 35, 39) ('rabbit', 850, 8) ### ### ('cartoons', 19, 71) ('anyway', 172, 24) ('willoughby', 5454, 0) ('antics', 23, 77) ### ### ('1941', 270, 25) ('character', 51, 43) ('animation', 145, 31) ('1944', 668, 19) ### ### ('funeral', 1748, 7) ('.', 14, 159) ('characters', 104, 38) ### ############################################################################################################ [2023-11-25 08:29:49,071][root][INFO] - Epoch: 2: Step: 11101/19530, loss[v]=0.218618, lr=0.000017, acc@1[1]=966.0/1024=0.943359375, acc@1[2]=998.5/1024=0.97509765625 [2023-11-25 08:31:39,077][root][INFO] - Train batch 11200 [2023-11-25 08:31:39,078][root][INFO] - Avg. loss per last 100 batches: 0.155654 [2023-11-25 08:31:39,814][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.03 | max: 3.67 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.3/29522=99.99% | mean: 0.03 | max: 5.29 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: ( see list of air ministry specifications ). ### ### [P_TEXT]: blackburn b. 44 the blackburn b. 44 was a british single - engined fighter aircraft ### ### designed by blackburn aircraft in 1942. it was notable as a rare example of a flying boat fighter, ### ### featuring blackburn's unique retractable hull, and designed to meet the requirements of air ### ### ministry specification n. 2 / 42. after japan's initial successes in the pacific during world war ### ### ii, the need for a fighter aircraft capable of operating from austere island sites with minimal ### ### infrastructure was regarded as a high priority. n. 2 / 42 called for a retractable - hull flying - ### ### boat ### ### =============================================== v_q ================================================ ### ### ('air', 0, 23) ('absently', 1, 114) ('ministry', 2, 5) ('.', 3, 885) ('intently', 4, 261) ### ### ('¹', 5, 378) ('list', 6, 11809) ('anyway', 7, 28) ('specification', 8, 18) ### ### ('specifications', 9, 11) ('standards', 10, 212) ('貴', 11, 179) ('aerial', 12, 171) ### ### ('regulations', 13, 186) ('##さ', 14, 289) ('standard', 15, 784) ('specified', 16, 1711) ### ### ('ம', 17, 96) ('aircraft', 18, 10) ('ந', 19, 413) ### ### =============================================== v_p ================================================ ### ### ('blackburn', 2020, 0) ('44', 10705, 1) ('aus', 6479, 2) ('##ctable', 18385, 3) ('hull', 6051, 4) ### ### ('ministry', 2, 5) ('42', 6639, 6) ('n', 22357, 7) ('boat', 5561, 8) ('boats', 18894, 9) ### ### ('aircraft', 18, 10) ('specifications', 9, 11) ('priority', 9319, 12) ('japan', 1768, 13) ### ### ('fighter', 249, 14) ('flying', 25, 15) ('1942', 611, 16) ('island', 9081, 17) ### ### ('specification', 8, 18) ('worried', 21, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ministry', 2, 5) ('air', 0, 23) ('specifications', 9, 11) ('anyway', 7, 28) ### ### ('specification', 8, 18) ('aircraft', 18, 10) ('flying', 25, 15) ('worried', 21, 19) ### ### ('absently', 1, 114) ('requirements', 23, 40) ('aviation', 22, 45) ('fighter', 249, 14) ### ### ('blackburn', 2020, 0) ('intently', 4, 261) ('bomber', 36, 44) ('¹', 5, 378) ('design', 76, 39) ### ### ('british', 88, 41) ('1942', 611, 16) ('ம', 17, 96) ### ############################################################################################################ [2023-11-25 08:31:40,324][root][INFO] - Epoch: 2: Step: 11201/19530, loss[v]=0.244476, lr=0.000017, acc@1[1]=968.0/1024=0.9453125, acc@1[2]=992.5/1024=0.96923828125 [2023-11-25 08:33:30,149][root][INFO] - Train batch 11300 [2023-11-25 08:33:30,150][root][INFO] - Avg. loss per last 100 batches: 0.174034 [2023-11-25 08:33:30,859][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.76 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.6/29522=99.99% | mean: 0.03 | max: 5.42 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: horobin was the son of j. c. horobin, the principal of homerton college, cambridge, and ### ### his wife, maud adeline nee ford ( known after her second marriage as m. a. cloudesley brereton ### ### [P_TEXT]: m. a. cloudesley brereton in 1893, then from 1894 as headmistress of baroness burdett - ### ### coutts school in highgate, both girls'secondary schools. in 1897 she became resident tutor at ### ### homerton college, cambridge, marrying the principal, john horobin, shortly thereafter. by the time ### ### of horobin's early death in 1902, they had two daughters and a son. one of their daughters, norah ### ### horobin, went on to become a teacher and ended her career as headmistress of roedean school. their ### ### son, ian, would become a conservative member of parliament. maud briefly served as acting principal ### ### at ### ### =============================================== v_q ================================================ ### ### ('##bin', 0, 4) ('ho', 1, 14) ('##ro', 2, 19) ('homer', 3, 0) ('maud', 4, 3) ('¹', 5, 131) ### ### ('cambridge', 6, 27) ('##elin', 7, 22820) ('intently', 8, 194) ('ford', 9, 10301) ### ### ('absently', 10, 233) ('cloud', 11, 1) ('##ere', 12, 7) ('j', 13, 17267) ('c', 14, 10874) ### ### ('college', 15, 60) ('##ton', 16, 23) ('seemed', 17, 176) ('oxford', 18, 199) ('ad', 19, 22556) ### ### =============================================== v_p ================================================ ### ### ('homer', 3, 0) ('cloud', 11, 1) ('nora', 13603, 2) ('maud', 4, 3) ('##bin', 0, 4) ### ### ('roe', 11792, 5) ('##rde', 22190, 6) ('##ere', 12, 7) ('##de', 12396, 8) ('##gate', 13521, 9) ### ### ('principal', 4369, 10) ('ian', 7803, 11) ('baroness', 5941, 12) ('##mist', 21626, 13) ### ### ('ho', 1, 14) ('tutor', 63, 15) ('##ut', 18302, 16) ('littered', 1223, 17) ('anyway', 141, 18) ### ### ('##ro', 2, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##bin', 0, 4) ('homer', 3, 0) ('maud', 4, 3) ('ho', 1, 14) ('##ro', 2, 19) ('cloud', 11, 1) ### ### ('##ere', 12, 7) ('cambridge', 6, 27) ('##ton', 16, 23) ('pupils', 20, 36) ('tutor', 63, 15) ### ### ('college', 15, 60) ('m', 72, 21) ('¹', 5, 131) ('##ley', 30, 41) ('##es', 24, 47) ('son', 32, 42) ### ### ('anyway', 141, 18) ('sons', 81, 33) ('intently', 8, 194) ### ############################################################################################################ [2023-11-25 08:33:31,362][root][INFO] - Epoch: 2: Step: 11301/19530, loss[v]=0.156445, lr=0.000017, acc@1[1]=990.5/1024=0.96728515625, acc@1[2]=1008.5/1024=0.98486328125 [2023-11-25 08:35:21,250][root][INFO] - Train batch 11400 [2023-11-25 08:35:21,251][root][INFO] - Avg. loss per last 100 batches: 0.178556 [2023-11-25 08:35:21,980][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.63 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.4/29522=99.98% | mean: 0.03 | max: 5.29 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: yarrie or yarrie station is a pastoral lease and cattle station that once operated as a ### ### sheep station, located approximately north east of marble bar and south east of port hedland in the ### ### pilbara region of western australia. ### ### [P_TEXT]: mount edgar station mount edgar station, often referred to as mount edgar, is a pastoral ### ### lease that once operated as a sheep station but now operates as a cattle station. it is located ### ### about south east of marble bar and south east of port hedland and in the pilbara region of western ### ### australia. the property is bounded by yarrie station. the property was established in 1890 by ### ### george dudley ( known as dudley ) and alfred howden drake - brockman ( known as howden ), who both ### ### died early in the 1900s. ### ### =============================================== v_q ================================================ ### ### ('ya', 0, 15) ('##rrie', 1, 1) ('bar', 2, 4) ('marble', 3, 0) ('sheep', 4, 9) ('##lb', 5, 8) ### ### ('cattle', 6, 11) ('station', 7, 23) ('¹', 8, 156) ('perth', 9, 49) ('nearby', 10, 53) ### ### ('port', 11, 21) ('##dl', 12, 5) ('australia', 13, 29) ('pastoral', 14, 31) ('pi', 15, 18) ### ### ('##ara', 16, 16) ('##and', 17, 17) ('luckily', 18, 182) ('died', 19, 230) ### ### =============================================== v_p ================================================ ### ### ('marble', 3, 0) ('##rrie', 1, 1) ('edgar', 4202, 2) ('dudley', 6672, 3) ('bar', 2, 4) ### ### ('##dl', 12, 5) ('how', 8761, 6) ('brock', 17950, 7) ('##lb', 5, 8) ('sheep', 4, 9) ### ### ('lease', 43, 10) ('cattle', 6, 11) ('drake', 4264, 12) ('mount', 170, 13) ('##den', 1101, 14) ### ### ('ya', 0, 15) ('##ara', 16, 16) ('##and', 17, 17) ('pi', 15, 18) ('alfred', 619, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##rrie', 1, 1) ('marble', 3, 0) ('ya', 0, 15) ('bar', 2, 4) ('sheep', 4, 9) ('##lb', 5, 8) ### ### ('cattle', 6, 11) ('##dl', 12, 5) ('station', 7, 23) ('##ara', 16, 16) ('pi', 15, 18) ### ### ('port', 11, 21) ('##and', 17, 17) ('australia', 13, 29) ('lease', 43, 10) ('pastoral', 14, 31) ### ### ('mount', 170, 13) ('australian', 20, 35) ('property', 109, 24) ('perth', 9, 49) ### ############################################################################################################ [2023-11-25 08:35:22,486][root][INFO] - Epoch: 2: Step: 11401/19530, loss[v]=0.273783, lr=0.000017, acc@1[1]=962.5/1024=0.93994140625, acc@1[2]=988.5/1024=0.96533203125 [2023-11-25 08:37:12,816][root][INFO] - Train batch 11500 [2023-11-25 08:37:12,817][root][INFO] - Avg. loss per last 100 batches: 0.165432 [2023-11-25 08:37:13,536][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.2/29522=99.99% | mean: 0.03 | max: 5.39 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: stolle won ten grand slam doubles titles, partnering with compatriots bob hewitt ( 4 ### ### titles ), roy emerson ( 4 titles ) and ken rosewall ( 2 titles ). ### ### [P_TEXT]: mike davies ( tennis ) dan maskell. he played on the british davis cup team with bobby ### ### wilson, billy knight and roger becker. in 1952 davies went to australia for the first of three ### ### winter visits to work with harry hopman, the australian davis cup coach, and australian players ### ### like lew hoad, ken rosewall, roy emerson, fred stolle. it was there that davies developed his game. ### ### from 1958 to 1960 davies was ranked number 1 in great britain. he ### ### =============================================== v_q ================================================ ### ### ('##oll', 0, 8) ('st', 1, 54) ('rose', 2, 13) ('hewitt', 3, 4145) ('absently', 4, 94) ('¹', 5, 128) ### ### ('intently', 6, 103) ('roy', 7, 15) ('bob', 8, 5250) ('ken', 9, 21) ('##wall', 10, 7) ### ### ('##e', 11, 430) ('tennis', 12, 22) ('emerson', 13, 4) ('##ʂ', 14, 837) ('chess', 15, 69) ### ### ('##ɛ', 16, 100) ('幸', 17, 102) ('doubles', 18, 147) ('open', 19, 37) ### ### =============================================== v_p ================================================ ### ### ('davies', 2676, 0) ('becker', 1257, 1) ('lew', 10426, 2) ('knight', 1727, 3) ('emerson', 13, 4) ### ### ('mike', 797, 5) ('hop', 22040, 6) ('##wall', 10, 7) ('##oll', 0, 8) ('wilson', 4046, 9) ### ### ('mask', 2273, 10) ('bobby', 1386, 11) ('billy', 1224, 12) ('rose', 2, 13) ('harry', 2624, 14) ### ### ('roy', 7, 15) ('roger', 236, 16) ('##ad', 23940, 17) ('littered', 126, 18) ('##ell', 8275, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##oll', 0, 8) ('rose', 2, 13) ('##wall', 10, 7) ('roy', 7, 15) ('emerson', 13, 4) ('ken', 9, 21) ### ### ('tennis', 12, 22) ('st', 1, 54) ('open', 19, 37) ('absently', 4, 94) ('intently', 6, 103) ### ### ('¹', 5, 128) ('littered', 126, 18) ('chess', 15, 69) ('roger', 236, 16) ('becker', 1257, 1) ### ### ('mike', 797, 5) ('hissed', 32, 52) ('anyway', 278, 20) ('round', 64, 44) ### ############################################################################################################ [2023-11-25 08:37:14,041][root][INFO] - Epoch: 2: Step: 11501/19530, loss[v]=0.226658, lr=0.000017, acc@1[1]=969.0/1024=0.9462890625, acc@1[2]=996.0/1024=0.97265625 [2023-11-25 08:39:04,037][root][INFO] - Train batch 11600 [2023-11-25 08:39:04,038][root][INFO] - Avg. loss per last 100 batches: 0.172018 [2023-11-25 08:39:04,749][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.63 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.5/29522=99.99% | mean: 0.03 | max: 5.64 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: between 1906 and 1909 he founded an artistic community " l'abbaye de creteil " with ### ### charles vildrac ( who would become his brother in law ). ### ### [P_TEXT]: hiroatsu takata, aristide maillol, and antoine bourdelle, exchanged correspondence with ### ### romain rolland, and even made a sketch of mahatma gandhi. takata refused to return to japan as ### ### scheduled, and remained in paris for the next 27 years. his circle of acquaintances included paul ### ### signac, emile chartier, charles vildrac, georges duhamel, jules romains, georges rouault, and jean ### ### cocteau and partially supported himself by sending their works to japan. in 1937, ### ### =============================================== v_q ================================================ ### ### ('##rac', 0, 4) ('##ld', 1, 14) ('crete', 2, 9798) ('vi', 3, 22) ('##ᵉ', 4, 216) ('charles', 5, 53) ### ### ('¹', 6, 117) ('ab', 7, 25425) ('##bay', 8, 27976) ('##¨', 9, 254) ('francs', 10, 62) ### ### ('intently', 11, 139) ('french', 12, 68) ('georges', 13, 23) ('##il', 14, 21609) ('##ʂ', 15, 2104) ### ### ('auguste', 16, 2547) ('paris', 17, 25) ('belgian', 18, 758) ('picasso', 19, 1561) ### ### =============================================== v_p ================================================ ### ### ('##dell', 23675, 0) ('chart', 15447, 1) ('sign', 5547, 2) ('antoine', 26, 3) ('##rac', 0, 4) ### ### ('roma', 10247, 5) ('##ats', 18480, 6) ('mail', 18083, 7) ('roll', 7978, 8) ('##ult', 24375, 9) ### ### ('##ua', 13517, 10) ('##ata', 19731, 11) ('tak', 11686, 12) ('gandhi', 12338, 13) ('##ld', 1, 14) ### ### ('##eau', 3083, 15) ('du', 1433, 16) ('##ham', 24671, 17) ('ro', 14317, 18) ### ### ('correspondence', 4617, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##rac', 0, 4) ('##ld', 1, 14) ('vi', 3, 22) ('antoine', 26, 3) ('georges', 13, 23) ### ### ('paris', 17, 25) ('charles', 5, 53) ('emile', 38, 26) ('francs', 10, 62) ('hissed', 27, 49) ### ### ('french', 12, 68) ('jean', 47, 51) ('¹', 6, 117) ('littered', 118, 44) ('##ᵉ', 4, 216) ### ### ('anyway', 301, 29) ('intently', 11, 139) ('sketches', 228, 41) ('napoleon', 52, 63) ### ### ('france', 34, 75) ### ############################################################################################################ [2023-11-25 08:39:05,253][root][INFO] - Epoch: 2: Step: 11601/19530, loss[v]=0.221213, lr=0.000017, acc@1[1]=971.5/1024=0.94873046875, acc@1[2]=999.5/1024=0.97607421875 [2023-11-25 08:40:55,226][root][INFO] - Train batch 11700 [2023-11-25 08:40:55,227][root][INFO] - Avg. loss per last 100 batches: 0.165717 [2023-11-25 08:40:55,949][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.54 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.6/29522=99.99% | mean: 0.03 | max: 5.50 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: and produced the top ten singles " love like this " and " all night long " ( released ### ### march 30, 1999 ) prompting evans to start an 18 - city theater tour with dru hill and total the ### ### following year. ### ### [P_TEXT]: black rob his first rap group, the schizophrenics, though they did not release any ### ### albums. he initially rapped under the alias " bacardi rob ". he began associating with the bad boy ### ### label as early as 1996, appearing on the bad boy remix to 112's " come see me ". he was then ### ### featured on several bad boy releases, including remixes to total's " what about us " ( 1997 ) and ### ### faith evans's " love like this " ( 1998 ), the song " 24 hrs. to live " ( 1997 ) from mase, and ### ### albums by puff ### ### =============================================== v_q ================================================ ### ### ('hill', 0, 10320) ('evans', 1, 6) ('##u', 2, 22223) ('dr', 3, 17271) ('like', 4, 5) ### ### ('total', 5, 3) ('night', 6, 6107) ('love', 7, 12) ('long', 8, 22221) ('1999', 9, 209) ### ### ('absently', 10, 114) ('tour', 11, 63) ('¹', 12, 154) ('this', 13, 29) ('intently', 14, 121) ### ### ('theater', 15, 6713) ('song', 16, 73) ('carey', 17, 10831) ('rihanna', 18, 1908) ('##ு', 19, 940) ### ### =============================================== v_p ================================================ ### ### ('rob', 1824, 0) ('black', 870, 1) ('hr', 22034, 2) ('total', 5, 3) ('mas', 5179, 4) ('like', 4, 5) ### ### ('evans', 1, 6) ('what', 4345, 7) ('##card', 21948, 8) ('112', 2687, 9) ('puff', 5293, 10) ### ### ('faith', 1431, 11) ('love', 7, 12) ('us', 1883, 13) ('##ics', 13060, 14) ('rap', 72, 15) ### ### ('come', 9420, 16) ('24', 5853, 17) ('about', 17781, 18) ('see', 11169, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('evans', 1, 6) ('like', 4, 5) ('total', 5, 3) ('love', 7, 12) ('this', 13, 29) ('rap', 72, 15) ### ### ('tour', 11, 63) ('1998', 44, 24) ('liked', 22, 43) ('lyrical', 41, 33) ('rapper', 24, 45) ### ### ('artists', 20, 50) ('boy', 175, 22) ('absently', 10, 114) ('remix', 227, 20) ('black', 870, 1) ### ### ('single', 77, 31) ('1997', 253, 21) ('rob', 1824, 0) ('song', 16, 73) ### ############################################################################################################ [2023-11-25 08:40:56,456][root][INFO] - Epoch: 2: Step: 11701/19530, loss[v]=0.271998, lr=0.000017, acc@1[1]=957.5/1024=0.93505859375, acc@1[2]=990.0/1024=0.966796875 [2023-11-25 08:42:46,954][root][INFO] - Train batch 11800 [2023-11-25 08:42:46,955][root][INFO] - Avg. loss per last 100 batches: 0.169411 [2023-11-25 08:42:47,695][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 5.70 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it followed the 1845 u. s. annexation of texas, which mexico considered mexican territory ### ### since the government did not recognize the velasco treaty signed by mexican general antonio lopez ### ### de santa anna when he was a prisoner of the texian army during the 1836 texas revolution ### ### [P_TEXT]: missouri territory contiguous united states, although the cession of lands north of the ### ### 49th parallel would turn out to be the only " permanent " cession of u. s. territory ( the ### ### territories ceded to spain would be re - taken by the u. s. following the texas annexation and the ### ### mexican – american war ). on march 2, 1819, all of the missouri territory south of the parallel ### ### 36°30'north, except the missouri bootheel between the mississippi river and the saint francis river ### ### north of the 36th parallel north, was designated the new territory of arkansaw. ### ### =============================================== v_q ================================================ ### ### ('texas', 0, 19) ('##las', 1, 24834) ('annexation', 2, 20) ('absently', 3, 162) ### ### ('revolution', 4, 1933) ('mexico', 5, 31) ('lopez', 6, 5178) ('mexican', 7, 16) ('.', 8, 123) ### ### ('¹', 9, 165) ('##ian', 10, 22844) ('1845', 11, 1512) ('ve', 12, 23478) ('1836', 13, 1549) ### ### ('intently', 14, 183) ('##co', 15, 25458) ('santa', 16, 10580) ('treaty', 17, 61) ### ### ('antonio', 18, 14128) ('territory', 19, 2) ### ### =============================================== v_p ================================================ ### ### ('missouri', 183, 0) ('francis', 10413, 1) ('territory', 19, 2) ('49th', 24048, 3) ('ark', 9327, 4) ### ### ('##ssion', 340, 5) ('booth', 12631, 6) ('36th', 10038, 7) ('parallel', 14355, 8) ('##°', 5721, 9) ### ### ('##ans', 8551, 10) ('##aw', 19142, 11) ('##30', 13385, 12) ('1819', 5664, 13) ('annexed', 136, 14) ### ### ('permanent', 2967, 15) ('mexican', 7, 16) ('saint', 6062, 17) ('river', 1389, 18) ('texas', 0, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('texas', 0, 19) ('annexation', 2, 20) ('territory', 19, 2) ('mexican', 7, 16) ('mexico', 5, 31) ### ### ('war', 21, 27) ('missouri', 183, 0) ('americans', 43, 24) ('##ssion', 340, 5) ('annexed', 136, 14) ### ### ('spanish', 39, 30) ('treaty', 17, 61) ('absently', 3, 162) ('territories', 140, 26) ### ### ('anyway', 176, 25) ('.', 8, 123) ('u', 66, 37) ('states', 28, 57) ('¹', 9, 165) ### ### ('mississippi', 539, 21) ### ############################################################################################################ [2023-11-25 08:42:48,199][root][INFO] - Epoch: 2: Step: 11801/19530, loss[v]=0.212280, lr=0.000017, acc@1[1]=959.0/1024=0.9365234375, acc@1[2]=1000.0/1024=0.9765625 [2023-11-25 08:44:38,298][root][INFO] - Train batch 11900 [2023-11-25 08:44:38,299][root][INFO] - Avg. loss per last 100 batches: 0.166016 [2023-11-25 08:44:39,016][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.94 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.7/29522=99.99% | mean: 0.02 | max: 5.53 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.2/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: landscape architect alfred caldwell was hired by the works progress administration to ### ### completely redesign this area of lincoln park. ### ### [P_TEXT]: alfred caldwell jensen and had a two - year private practice thereafter. in 1933 he was ### ### appointed superintendent of parks for dubuque, iowa, where he created eagle point park. from 1936 ### ### to 1939 he was a landscape designer for the chicago park district and was the designer and ### ### architect of the alfred caldwell lily pool at lincoln park. in 1938 the lily pool project was ### ### nearing completion and the chicago park district, in an effort to save money, decided to cut the ### ### major wildflower plantings expenditure. caldwell cashed in his $ 5000 life insurance ### ### =============================================== v_q ================================================ ### ### ('lincoln', 0, 1) ('alfred', 1, 3) ('park', 2, 10) ('caldwell', 3, 0) ('landscape', 4, 24) ### ### ('progress', 5, 7706) ('neighborhood', 6, 458) ('administration', 7, 1290) ('yorker', 8, 12588) ### ### ('chicago', 9, 6) ('¹', 10, 175) ('renovations', 11, 12524) ('architect', 12, 27) ### ### ('rebuilding', 13, 9316) ('lawn', 14, 1848) ('nrhp', 15, 17737) ('pasadena', 16, 7573) ### ### ('atlanta', 17, 1090) ('works', 18, 3114) ('.', 19, 2324) ### ### =============================================== v_p ================================================ ### ### ('caldwell', 3, 0) ('lincoln', 0, 1) ('eagle', 10527, 2) ('alfred', 1, 3) ### ### ('superintendent', 130, 4) ('lily', 6942, 5) ('chicago', 9, 6) ('dub', 25751, 7) ### ### ('insurance', 2331, 8) ('pool', 8346, 9) ('park', 2, 10) ('point', 5387, 11) ('worried', 2891, 12) ### ### ('parks', 103, 13) ('5000', 2833, 14) ('iowa', 20, 15) ('##que', 19598, 16) ('jensen', 15986, 17) ### ### ('##flower', 13080, 18) ('expenditure', 18682, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('lincoln', 0, 1) ('alfred', 1, 3) ('caldwell', 3, 0) ('park', 2, 10) ('chicago', 9, 6) ### ### ('landscape', 4, 24) ('iowa', 20, 15) ('architect', 12, 27) ('superintendent', 130, 4) ### ### ('parks', 103, 13) ('district', 247, 21) ('leaned', 462, 23) ('1933', 382, 26) ### ### ('insurance', 2331, 8) ('architects', 330, 31) ('worried', 2891, 12) ('absently', 22, 62) ### ### ('architecture', 27, 59) ('¹', 10, 175) ('designed', 44, 52) ### ############################################################################################################ [2023-11-25 08:44:39,519][root][INFO] - Epoch: 2: Step: 11901/19530, loss[v]=0.085193, lr=0.000017, acc@1[1]=991.0/1024=0.9677734375, acc@1[2]=1016.5/1024=0.99267578125 [2023-11-25 08:46:29,905][root][INFO] - Train batch 12000 [2023-11-25 08:46:29,905][root][INFO] - Avg. loss per last 100 batches: 0.168512 [2023-11-25 08:46:30,637][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.3/29522=99.99% | mean: 0.03 | max: 5.40 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the last of sheila is a 1973 american neo noir mystery film directed by herbert ross and ### ### written directly for the screen by anthony perkins and stephen sondheim. ### ### [P_TEXT]: stephen sondheim's play " sleuth " was inspired by the composer. according to a rumor ( ### ### denied by shaffer in a march 10, 1996 " new york times " interview ), " sleuth " had the working ### ### title " who's afraid of stephen sondheim? " his love of puzzles and mysteries is evident in " the ### ### last of sheila ", an intricate whodunit written with longtime friend anthony perkins. the 1973 ### ### film, directed by herbert ross, featured dyan cannon, raquel welch, james mason, james coburn and ### ### =============================================== v_q ================================================ ### ### ('sheila', 0, 0) ('last', 1, 29) ('ross', 2, 7) ('1973', 3, 19) ('perkins', 4, 1) ('neo', 5, 18605) ### ### ('″', 6, 40) ('herbert', 7, 2) ('noir', 8, 17555) ('anthony', 9, 14) ('##eim', 10, 6) ('"', 11, 74) ### ### ('¹', 12, 166) ('stephen', 13, 15) ('mystery', 14, 59) ('##dh', 15, 21) ('films', 16, 51) ### ### ('sequels', 17, 3185) ('movies', 18, 128) ('film', 19, 34) ### ### =============================================== v_p ================================================ ### ### ('sheila', 0, 0) ('perkins', 4, 1) ('herbert', 7, 2) ('cannon', 3991, 3) ('puzzles', 16508, 4) ### ### ('welch', 3923, 5) ('##eim', 10, 6) ('ross', 2, 7) ('mason', 10882, 8) ('##dun', 28548, 9) ### ### ('##ffer', 13563, 10) ('##eu', 15538, 11) ('##yan', 23382, 12) ('##quel', 6718, 13) ### ### ('anthony', 9, 14) ('stephen', 13, 15) ('rumor', 1651, 16) ('##burn', 21142, 17) ('son', 23, 18) ### ### ('1973', 3, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('sheila', 0, 0) ('perkins', 4, 1) ('ross', 2, 7) ('herbert', 7, 2) ('1973', 3, 19) ### ### ('##eim', 10, 6) ('anthony', 9, 14) ('last', 1, 29) ('stephen', 13, 15) ('″', 6, 40) ### ### ('##dh', 15, 21) ('son', 23, 18) ('film', 19, 34) ('films', 16, 51) ('mystery', 14, 59) ### ### ('"', 11, 74) ('convinces', 27, 54) ('¹', 12, 166) ('anyway', 350, 24) ('mysteries', 166, 37) ### ############################################################################################################ [2023-11-25 08:46:31,143][root][INFO] - Epoch: 2: Step: 12001/19530, loss[v]=0.089989, lr=0.000017, acc@1[1]=994.0/1024=0.970703125, acc@1[2]=1005.5/1024=0.98193359375 [2023-11-25 08:48:21,712][root][INFO] - Train batch 12100 [2023-11-25 08:48:21,713][root][INFO] - Avg. loss per last 100 batches: 0.181251 [2023-11-25 08:48:22,448][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.4/29522=99.99% | mean: 0.03 | max: 5.54 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: with petter øien at the 2012 melodi grand prix to choose norway's entry to the 2012 ### ### eurovision song contest to be held in baku, azerbaijan, in may. ### ### [P_TEXT]: norway in the eurovision song contest 2012 the nine semi - final qualifiers alongside the ### ### jury wildcard, lise karlsnes, competed during the final on 11 february 2012. the winner was ### ### selected over two rounds of voting. in the first round, the top four entries were selected by ### ### public televoting to proceed to the second round, the gold final : " stay " performed by tooji, " ### ### ola nordmann " performed by plumbo, " somewhere beautiful " performed by nora foss al - jabri and " ### ### things change " performed by petter øien & bobby bare. in the gold final, three regional ### ### =============================================== v_q ================================================ ### ### ('2012', 0, 13) ('##ien', 1, 17) ('##od', 2, 27994) ('¹', 3, 142) ('contest', 4, 18) ### ### ('norway', 5, 5) ('pet', 6, 33) ('mel', 7, 18455) ('absently', 8, 130) ('song', 9, 34) ### ### ('prix', 10, 2895) ('##ter', 11, 37) ('ø', 12, 19) ('eurovision', 13, 47) ('intently', 14, 115) ### ### ('azerbaijan', 15, 14717) ('whoever', 16, 109) ('norwegian', 17, 31) ('ɐ', 18, 1068) ### ### ('baku', 19, 9767) ### ### =============================================== v_p ================================================ ### ### ('bare', 4022, 0) ('plum', 18643, 1) ('stay', 1748, 2) ('nora', 3246, 3) ('somewhere', 12259, 4) ### ### ('norway', 5, 5) ('change', 2003, 6) ('things', 4128, 7) ('bobby', 4384, 8) ('nord', 280, 9) ### ### ('##oss', 17787, 10) ('##ji', 12484, 11) ('beautiful', 731, 12) ('2012', 0, 13) ### ### ('##card', 17048, 14) ('too', 7775, 15) ('##bri', 26653, 16) ('##ien', 1, 17) ('contest', 4, 18) ### ### ('ø', 12, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('2012', 0, 13) ('##ien', 1, 17) ('norway', 5, 5) ('contest', 4, 18) ('ø', 12, 19) ('pet', 6, 33) ### ### ('song', 9, 34) ('##ter', 11, 37) ('norwegian', 17, 31) ('eurovision', 13, 47) ('hissed', 22, 44) ### ### ('nord', 280, 9) ('¹', 3, 142) ('littered', 56, 39) ('absently', 8, 130) ('intently', 14, 115) ### ### ('anyway', 180, 27) ('whoever', 16, 109) ('beautiful', 731, 12) ('stay', 1748, 2) ### ############################################################################################################ [2023-11-25 08:48:22,950][root][INFO] - Epoch: 2: Step: 12101/19530, loss[v]=0.096348, lr=0.000017, acc@1[1]=998.5/1024=0.97509765625, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 08:50:12,604][root][INFO] - Train batch 12200 [2023-11-25 08:50:12,605][root][INFO] - Avg. loss per last 100 batches: 0.159154 [2023-11-25 08:50:13,308][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: on december 16, general jeremiah sullivan sent colonel robert g. ingersoll from jackson, ### ### tennessee to lexington ; at lexington, ingersoll had 670 men ( mostly raw recruits ) and two 3 - ### ### inch guns ; part of his force was a contingent of cavalry under future governor alvin hawkins. ### ### [P_TEXT]: robert g. ingersoll of shiloh. ingersoll was later captured in a skirmis h with the ### ### confederates near lexington, tennessee on december 18, 1862, then paroled - i. e. released on his ### ### oath that he would not fight again against the confederate states of america until formally ### ### exchanged for a captured confederate soldier or officer of like rank ( who was often under parole ### ### himself, making the practice a matter of honor and formality, which could be extended to ### ### individuals or even entire regiments " en masse " ). this was an old practice which was still ### ### commonly observed early in the war, ### ### =============================================== v_q ================================================ ### ### ('hawkins', 0, 3611) ('alvin', 1, 15670) ('lexington', 2, 0) ('sullivan', 3, 3079) ('##oll', 4, 8) ### ### ('jeremiah', 5, 1554) ('jackson', 6, 1200) ('ing', 7, 12) ('absently', 8, 267) ('g', 9, 34) ### ### ('tennessee', 10, 11) ('robert', 11, 14) ('¹', 12, 343) ('##ers', 13, 24) ('battle', 14, 159) ### ### ('cavalry', 15, 171) ('guns', 16, 588) ('intently', 17, 310) ('governor', 18, 2402) ('.', 19, 260) ### ### =============================================== v_p ================================================ ### ### ('lexington', 2, 0) ('##rmi', 26172, 1) ('parole', 4980, 2) ('##loh', 27426, 3) ('oath', 10899, 4) ### ### ('ski', 20668, 5) ('confederates', 5368, 6) ('formal', 5244, 7) ('##oll', 4, 8) ### ### ('confederate', 46, 9) ('confederacy', 597, 10) ('tennessee', 10, 11) ('ing', 7, 12) ### ### ('worried', 171, 13) ('robert', 11, 14) ('mass', 5075, 15) ('rank', 3646, 16) ('shi', 23471, 17) ### ### ('honor', 1323, 18) ('##ity', 16620, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('lexington', 2, 0) ('##oll', 4, 8) ('ing', 7, 12) ('tennessee', 10, 11) ('robert', 11, 14) ### ### ('##ers', 13, 24) ('g', 9, 34) ('confederate', 46, 9) ('regiments', 25, 28) ('war', 21, 38) ### ### ('soldiers', 31, 30) ('kentucky', 20, 56) ('regiment', 50, 31) ('worried', 171, 13) ### ### ('troops', 34, 40) ('1862', 112, 26) ('december', 40, 49) ('anyway', 247, 27) ### ### ('confederacy', 597, 10) ('absently', 8, 267) ### ############################################################################################################ [2023-11-25 08:50:13,813][root][INFO] - Epoch: 2: Step: 12201/19530, loss[v]=0.113262, lr=0.000017, acc@1[1]=996.5/1024=0.97314453125, acc@1[2]=1013.5/1024=0.98974609375 [2023-11-25 08:52:03,621][root][INFO] - Train batch 12300 [2023-11-25 08:52:03,622][root][INFO] - Avg. loss per last 100 batches: 0.171238 [2023-11-25 08:52:04,347][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.3/29522=99.99% | mean: 0.03 | max: 5.42 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: davos seaworth, dolorous edd tollett, and other brothers of the watch loyal to jon ### ### barricade themselves in a room with ghost and jon's body, and an attack by thorne and his men is ### ### thwarted ### ### [P_TEXT]: davos seaworth during their march, ramsay bolton and his men infiltrate the camp and ### ### destroy their supplies. davos is sent back to the wall to ask for more supplies, unaware that ### ### stannis plans to sacrifice shireen to assure his victory. lord commander jon snow, though ### ### sympathetic, is unable to offer help. soon after, melisandre arrives at castle black, having fled ### ### in the aftermath of shireen's sacrifice. davos realises that stannis has been defeated and shireen ### ### is dead, though melisandre does not reveal her role. davos is the first to discover jon snow's body ### ### after his assassination, and ### ### =============================================== v_q ================================================ ### ### ('thorne', 0, 1675) ('ghost', 1, 3858) ('jon', 2, 4) ('##ett', 3, 26560) ('watch', 4, 7692) ### ### ('toll', 5, 22384) ('tolkien', 6, 20017) ('absently', 7, 212) ('ed', 8, 24675) ('da', 9, 26) ### ### ('##worth', 10, 9) ('sea', 11, 15) ('¹', 12, 324) ('"', 13, 19214) ('intently', 14, 323) ### ### ('″', 15, 2076) ('barr', 16, 22578) ('##ica', 17, 24684) ('character', 18, 3365) ### ### ('stark', 19, 2597) ### ### =============================================== v_p ================================================ ### ### ('ramsay', 1990, 0) ('##vos', 57, 1) ('black', 2930, 2) ('snow', 3290, 3) ('jon', 2, 4) ### ### ('bolton', 1823, 5) ('shire', 2953, 6) ('stan', 9261, 7) ('castle', 213, 8) ('##worth', 10, 9) ### ### ('sacrificed', 24768, 10) ('##dre', 13580, 11) ('sacrifice', 127, 12) ('##isan', 5340, 13) ### ### ('wall', 3256, 14) ('sea', 11, 15) ('commander', 2305, 16) ('assassination', 3670, 17) ### ### ('##nis', 15750, 18) ('marched', 21232, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('jon', 2, 4) ('##worth', 10, 9) ('sea', 11, 15) ('da', 9, 26) ('##vos', 57, 1) ### ### ('sacrifice', 127, 12) ('castle', 213, 8) ('lord', 200, 20) ('absently', 7, 212) ### ### ('tricked', 28, 88) ('anyway', 308, 21) ('him', 89, 44) ('worried', 148, 38) ('ramsay', 1990, 0) ### ### ('appears', 63, 68) ('revealed', 259, 36) ('vimes', 44, 99) ('¹', 12, 324) ('bolton', 1823, 5) ### ### ('intently', 14, 323) ### ############################################################################################################ [2023-11-25 08:52:04,851][root][INFO] - Epoch: 2: Step: 12301/19530, loss[v]=0.075205, lr=0.000017, acc@1[1]=1000.0/1024=0.9765625, acc@1[2]=1018.5/1024=0.99462890625 [2023-11-25 08:53:54,999][root][INFO] - Train batch 12400 [2023-11-25 08:53:55,000][root][INFO] - Avg. loss per last 100 batches: 0.155578 [2023-11-25 08:53:55,722][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.80 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.8/29522=99.99% | mean: 0.03 | max: 5.59 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: italian football was ### ### [P_TEXT]: football in italy parma but now napoli are known as the seven sisters of italian ### ### football. italian managers are the most successful in european football, especially in competitions ### ### such as the champions league. more players have won the coveted " ballon d'or " award while playing ### ### at a serie a club than any other league in the world, except la liga. other forms of football were ### ### played in italy in ancient times, the earliest of which was " harpastum ", played during the times ### ### of ### ### =============================================== v_q ================================================ ### ### ('football', 0, 8) ('italian', 1, 11) ('.', 2, 81) ('intently', 3, 182) ('absently', 4, 183) ### ### ('italy', 5, 1) ('¹', 6, 270) ('soccer', 7, 55) ('footballer', 8, 37) ('italians', 9, 160) ### ### ('anyway', 10, 21) ('貴', 11, 156) ('footballers', 12, 61) ('italia', 13, 1737) ('ɐ', 14, 388) ### ### ('fifa', 15, 335) ('ball', 16, 6) ('ம', 17, 91) ('worried', 18, 18) ('sports', 19, 884) ### ### =============================================== v_p ================================================ ### ### ('sisters', 23468, 0) ('italy', 5, 1) ('harp', 26032, 2) ('parma', 882, 3) ('napoli', 414, 4) ### ### ('##ast', 29004, 5) ('ball', 16, 6) ('managers', 14396, 7) ('football', 0, 8) ### ### ('managerial', 21886, 9) ('seven', 12719, 10) ('italian', 1, 11) ('earliest', 269, 12) ### ### ('##on', 23446, 13) ('serie', 737, 14) ('##um', 18480, 15) ('ancient', 261, 16) ('liga', 1024, 17) ### ### ('worried', 18, 18) ('champions', 16817, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('football', 0, 8) ('italy', 5, 1) ('italian', 1, 11) ('ball', 16, 6) ('anyway', 10, 21) ### ### ('worried', 18, 18) ('.', 2, 81) ('footballer', 8, 37) ('soccer', 7, 55) ('napoli', 414, 4) ### ### ('footballers', 12, 61) ('intently', 3, 182) ('absently', 4, 183) ('parma', 882, 3) ### ### ('earliest', 269, 12) ('shirt', 125, 22) ('ancient', 261, 16) ('¹', 6, 270) ('ம', 17, 91) ### ### ('european', 230, 24) ### ############################################################################################################ [2023-11-25 08:53:56,225][root][INFO] - Epoch: 2: Step: 12401/19530, loss[v]=0.095466, lr=0.000017, acc@1[1]=998.5/1024=0.97509765625, acc@1[2]=1012.0/1024=0.98828125 [2023-11-25 08:55:46,077][root][INFO] - Train batch 12500 [2023-11-25 08:55:46,078][root][INFO] - Avg. loss per last 100 batches: 0.171802 [2023-11-25 08:55:46,804][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.72 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.1/29522=99.98% | mean: 0.02 | max: 5.28 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: sebastian philip bierk ( born april 3, 1968 ), known professionally as sebastian bach, is ### ### a canadian singer - songwriter who achieved mainstream success as the frontman of heavy metal band ### ### skid row from 1987 to 1996. ### ### [P_TEXT]: subhuman race subhuman race ( stylized subhuman race ) is the third studio album by ### ### american heavy metal band skid row, released on march 28, 1995, by atlantic records. this is the ### ### last skid row album with singer sebastian bach and drummer rob affuso, and the last one to be ### ### released on atlantic. despite receiving positive reviews, " subhuman race " was not as successful ### ### as the band's first two albums. certain tracks from the album were remixed for the band's ### ### compilation " ", given more tender, slightly cleaner mixes to fit better with ### ### =============================================== v_q ================================================ ### ### ('sebastian', 0, 4) ('bach', 1, 2) ('bi', 2, 23322) ('philip', 3, 9488) ('##k', 4, 21678) ### ### ('ski', 5, 5) ('##er', 6, 23172) ('row', 7, 1) ('¹', 8, 262) ('metal', 9, 23) ('absently', 10, 45) ### ### ('frontman', 11, 56) ('band', 12, 20) ('##ك', 13, 3617) ('intently', 14, 115) ('bands', 15, 15) ### ### ('sick', 16, 1536) ('##ք', 17, 942) ('distortion', 18, 949) ('yeah', 19, 32) ### ### =============================================== v_p ================================================ ### ### ('##hum', 22987, 0) ('row', 7, 1) ('bach', 1, 2) ('worried', 3092, 3) ('sebastian', 0, 4) ### ### ('ski', 5, 5) ('race', 5675, 6) ('af', 14433, 7) ('sub', 15736, 8) ('1995', 352, 9) ### ### ('tender', 19899, 10) ('atlantic', 6595, 11) ('##fus', 28847, 12) ('rob', 250, 13) ### ### ('littered', 98, 14) ('bands', 15, 15) ('leaned', 7878, 16) ('compilation', 1359, 17) ### ### ('anyway', 677, 18) ('##d', 29, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('bach', 1, 2) ('sebastian', 0, 4) ('row', 7, 1) ('ski', 5, 5) ('bands', 15, 15) ('metal', 9, 23) ### ### ('band', 12, 20) ('lyrical', 21, 22) ('##d', 29, 19) ('littered', 98, 14) ('absently', 10, 45) ### ### ('album', 44, 21) ('frontman', 11, 56) ('1995', 352, 9) ('rob', 250, 13) ('yeah', 19, 32) ### ### ('¹', 8, 262) ('intently', 14, 115) ('albums', 226, 24) ('anyway', 677, 18) ### ############################################################################################################ [2023-11-25 08:55:47,311][root][INFO] - Epoch: 2: Step: 12501/19530, loss[v]=0.217932, lr=0.000017, acc@1[1]=968.5/1024=0.94580078125, acc@1[2]=995.5/1024=0.97216796875 [2023-11-25 08:57:37,140][root][INFO] - Train batch 12600 [2023-11-25 08:57:37,141][root][INFO] - Avg. loss per last 100 batches: 0.170520 [2023-11-25 08:57:37,875][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.73 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.4/29522=99.99% | mean: 0.03 | max: 5.65 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 29.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it flows west, passing through the village of cornish flat, where it is joined by notch ### ### brook, then flows northwest into the town of plainfield, where it takes a zig - zag ### ### [P_TEXT]: cornish flat, new hampshire cornish flat is an unincorporated community in the town of ### ### cornish in sullivan county, new hampshire, united states. the village is located in the ### ### northeastern corner of cornish, at the southern end of a valley floor which is bordered westerly by ### ### cornish stage road, easterly by new hampshire route 120, and whose northerly end is in the town of ### ### plainfield. the headwaters of blow - me - down brook gather in a swampy lowland at the valley's ### ### north end formed by the confluence of notch, leavitt, wine, and penniman ### ### =============================================== v_q ================================================ ### ### ('plain', 0, 5) ('flat', 1, 9) ('notch', 2, 4) ('##field', 3, 11) ('brook', 4, 18) ### ### ('cornish', 5, 0) ('river', 6, 115) ('maine', 7, 198) ('¹', 8, 98) ('connecticut', 9, 79) ### ### ('massachusetts', 10, 46) ('rivers', 11, 188) ('absently', 12, 144) ('intently', 13, 126) ### ### ('york', 14, 1870) ('z', 15, 9120) ('pond', 16, 474) ('.', 17, 655) ('cornwall', 18, 26) ### ### ('wisconsin', 19, 207) ### ### =============================================== v_p ================================================ ### ### ('cornish', 5, 0) ('##vitt', 7351, 1) ('120', 7668, 2) ('##iman', 20155, 3) ('notch', 2, 4) ### ### ('plain', 0, 5) ('sullivan', 3274, 6) ('blow', 16590, 7) ('wine', 690, 8) ('flat', 1, 9) ### ### ('lea', 4653, 10) ('##field', 3, 11) ('penn', 1868, 12) ('stage', 3766, 13) ('hampshire', 21, 14) ### ### ('lowland', 10858, 15) ('swamp', 979, 16) ('##erly', 730, 17) ('brook', 4, 18) ('worried', 218, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('cornish', 5, 0) ('plain', 0, 5) ('notch', 2, 4) ('flat', 1, 9) ('##field', 3, 11) ### ### ('brook', 4, 18) ('hampshire', 21, 14) ('cornwall', 18, 26) ('village', 22, 32) ('town', 26, 27) ### ### ('massachusetts', 10, 46) ('stream', 24, 35) ('route', 43, 20) ('headwaters', 56, 21) ### ### ('river', 6, 115) ('connecticut', 9, 79) ('¹', 8, 98) ('tributary', 25, 54) ('county', 65, 30) ### ### ('worried', 218, 19) ### ############################################################################################################ [2023-11-25 08:57:38,378][root][INFO] - Epoch: 2: Step: 12601/19530, loss[v]=0.090595, lr=0.000017, acc@1[1]=1000.5/1024=0.97705078125, acc@1[2]=1016.0/1024=0.9921875 [2023-11-25 08:59:28,432][root][INFO] - Train batch 12700 [2023-11-25 08:59:28,433][root][INFO] - Avg. loss per last 100 batches: 0.158190 [2023-11-25 08:59:29,142][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.70 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.9/29522=99.99% | mean: 0.02 | max: 5.27 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the name of kardam is first encountered in the byzantine sources in 791, when emperor ### ### constantine vi embarked on an expedition against bulgaria, in retaliation for bulgarian incursions ### ### in the struma valley since 789. ### ### [P_TEXT]: battle of marcellae after the end of the rule of the dulo. the khans telerig and kardam ### ### managed to consolidate the central authority and put an end of the quarrels among the nobility. the ### ### bulgarians finally had the opportunity to intensify their campaigns in slavic - populated ### ### macedonia. in 789 they penetrated deep into the valley of the struma river and heavily defeated the ### ### byzantines, killing the " strategos " of thrace filites. in order to distract the bulgarian ### ### attention from macedonia, the byzantine emperor constantine vi started a ### ### =============================================== v_q ================================================ ### ### ('##rda', 0, 3) ('ka', 1, 48) ('##rum', 2, 1) ('bulgaria', 3, 69) ('absently', 4, 305) ### ### ('vi', 5, 10) ('intently', 6, 251) ('¹', 7, 184) ('byzantine', 8, 15) ('bulgarian', 9, 5) ### ### ('##m', 10, 93) ('constantine', 11, 2) ('st', 12, 34) ('79', 13, 172) ('##ם', 14, 258) ### ### ('balkans', 15, 2719) ('##ʂ', 16, 1937) ('romani', 17, 8467) ('valley', 18, 65) ### ### ('expedition', 19, 106) ### ### =============================================== v_p ================================================ ### ### ('marcel', 7288, 0) ('##rum', 2, 1) ('constantine', 11, 2) ('##rda', 0, 3) ('slavic', 115, 4) ### ### ('bulgarian', 9, 5) ('##lite', 20931, 6) ('nobility', 1356, 7) ('##rate', 6283, 8) ### ### ('##lo', 15543, 9) ('vi', 5, 10) ('du', 9830, 11) ('78', 34, 12) ('fi', 14386, 13) ### ### ('battle', 267, 14) ('byzantine', 8, 15) ('anyway', 287, 16) ('##la', 10236, 17) ('khan', 1282, 18) ### ### ('macedonia', 21, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##rda', 0, 3) ('##rum', 2, 1) ('vi', 5, 10) ('bulgarian', 9, 5) ('constantine', 11, 2) ### ### ('byzantine', 8, 15) ('ka', 1, 48) ('78', 34, 12) ('macedonia', 21, 19) ('bulgaria', 3, 69) ### ### ('st', 12, 34) ('slavic', 115, 4) ('##m', 10, 93) ('valley', 18, 65) ('battle', 267, 14) ### ### ('macedonian', 89, 29) ('anyway', 287, 16) ('emperor', 39, 42) ('¹', 7, 184) ('.', 22, 78) ### ############################################################################################################ [2023-11-25 08:59:29,647][root][INFO] - Epoch: 2: Step: 12701/19530, loss[v]=0.205081, lr=0.000017, acc@1[1]=977.0/1024=0.9541015625, acc@1[2]=1001.5/1024=0.97802734375 [2023-11-25 09:01:19,498][root][INFO] - Train batch 12800 [2023-11-25 09:01:19,499][root][INFO] - Avg. loss per last 100 batches: 0.176977 [2023-11-25 09:01:20,216][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.63 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29516.6/29522=99.98% | mean: 0.02 | max: 5.20 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: 1969, before finally claiming a medal in the 1969 – 70 inter - cities fairs cup ; he ### ### scored arsenal's winning goal in their 4 – 3 victory on aggregate over rsc anderlecht, after ### ### arsenal had trailed 3 – 1 after the first leg. ### ### [P_TEXT]: 1969 football league cup final defender ian ure and goalkeeper bob wilson over a backpass ### ### led to confusion in the arsenal defence, and allowed smart to capitalise and score. swindon ### ### maintained their 1 – 0 lead until half - time. the second half was again one - sided, with ### ### swindon's tactics being those of all - out defence to protect their lead. swindon goalkeeper peter ### ### downsborough prevented numerous arsenal players from scoring, including saving efforts that looked ### ### unstoppable from jon sammels and bob mcnab. until the 86th minute, it seemed that a great football ### ### upset was sure to occur ; ### ### =============================================== v_q ================================================ ### ### ('fairs', 0, 1015) ('arsenal', 1, 7) ('70', 2, 834) ('cities', 3, 8402) ('rs', 4, 22614) ### ### ('##le', 5, 20627) ('¹', 6, 89) ('1969', 7, 1) ('littered', 8, 13) ('intently', 9, 153) ### ### ('##c', 10, 22453) ('##cht', 11, 26477) ('c', 12, 16073) ('absently', 13, 111) ('inter', 14, 6753) ### ### ('№', 15, 953) ('cup', 16, 16) ('ɐ', 17, 308) ('##er', 18, 18195) ('1970', 19, 28) ### ### =============================================== v_p ================================================ ### ### ('downs', 20078, 0) ('1969', 7, 1) ('swindon', 3718, 2) ('wilson', 2633, 3) ('smart', 5676, 4) ### ### ('ur', 9322, 5) ('##pass', 25859, 6) ('arsenal', 1, 7) ('##mel', 13840, 8) ('bob', 203, 9) ### ### ('final', 53, 10) ('ian', 49, 11) ('anyway', 174, 12) ('littered', 8, 13) ('jon', 428, 14) ### ### ('##borough', 22840, 15) ('cup', 16, 16) ('##sto', 12370, 17) ('##ppa', 12116, 18) ### ### ('##b', 9520, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('arsenal', 1, 7) ('1969', 7, 1) ('littered', 8, 13) ('cup', 16, 16) ('1970', 19, 28) ### ### ('final', 53, 10) ('ian', 49, 11) ('bob', 203, 9) ('football', 42, 27) ('1968', 40, 30) ### ### ('anyway', 174, 12) ('¹', 6, 89) ('shirt', 76, 24) ('goalkeeper', 69, 25) ('peter', 51, 34) ### ### ('hissed', 25, 43) ('jon', 428, 14) ('intently', 9, 153) ('club', 41, 40) ('absently', 13, 111) ### ############################################################################################################ [2023-11-25 09:01:20,722][root][INFO] - Epoch: 2: Step: 12801/19530, loss[v]=0.104824, lr=0.000017, acc@1[1]=997.0/1024=0.9736328125, acc@1[2]=1011.5/1024=0.98779296875 [2023-11-25 09:03:10,441][root][INFO] - Train batch 12900 [2023-11-25 09:03:10,442][root][INFO] - Avg. loss per last 100 batches: 0.171850 [2023-11-25 09:03:11,175][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.3/29522=99.99% | mean: 0.03 | max: 5.64 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in the first, on 7 august, major friedrich joseph, count of nauendorf and two squadrons ### ### of the " wurmser " hussars, surprised a prussian convoy, which surrendered 240 wagons of flour and ### ### [P_TEXT]: friedrich joseph, count of nauendorf joseph encouraged successful raids against the ### ### prussian troops. on 7 august 1778, with two squadrons of his regiment, nauendorf led a raid against ### ### a prussian convoy at biebersdorf in the county of kladsko. the surprised convoy surrendered, and ### ### nauendorf captured its officers, 110 men, 476 horses, 240 wagons of flour, and 13 transport wagons. ### ### in another raid, on 17 – 18 january 1779, nauendorf's commander, dagobert von wurmser advanced into ### ### the county of glatz in five columns, surrounded habelschwerdt, stormed the village. in a subsequent ### ### assault on the so - called swedish blockhouse in oberschwedeldorf ( now szalejow gorny ), it and ### ### the village of habelschwerdt were set on fire by howitzers. in total, the raid resulted in ### ### =============================================== v_q ================================================ ### ### ('joseph', 0, 24) ('prussian', 1, 10) ('##uen', 2, 2) ('absently', 3, 268) ('battle', 4, 130) ### ### ('##rm', 5, 18) ('##ʂ', 6, 207) ('##dorf', 7, 8) ('flour', 8, 21) ('friedrich', 9, 12) ### ### ('##ssar', 10, 27125) ('intently', 11, 339) ('##ser', 12, 7) ('na', 13, 54) ('convoy', 14, 27) ### ### ('wu', 15, 26) ('¹', 16, 236) ('.', 17, 1300) ('august', 18, 125) ('hu', 19, 18129) ### ### =============================================== v_p ================================================ ### ### ('##lat', 27436, 0) ('##itzer', 1364, 1) ('##uen', 2, 2) ('##eber', 23247, 3) ('wagons', 431, 4) ### ### ('##zal', 22052, 5) ('##sko', 26486, 6) ('##ser', 12, 7) ('##dorf', 7, 8) ('##house', 17306, 9) ### ### ('prussian', 1, 10) ('1779', 2503, 11) ('friedrich', 9, 12) ('##rn', 3949, 13) ('1778', 1053, 14) ### ### ('##lad', 27530, 15) ('##orf', 1888, 16) ('##bert', 14425, 17) ('##rm', 5, 18) ('##dt', 14726, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##uen', 2, 2) ('prussian', 1, 10) ('joseph', 0, 24) ('##dorf', 7, 8) ('friedrich', 9, 12) ### ### ('##ser', 12, 7) ('##rm', 5, 18) ('flour', 8, 21) ('convoy', 14, 27) ('wu', 15, 26) ('na', 13, 54) ### ### ('prussia', 20, 45) ('battle', 4, 130) ('regiment', 31, 57) ('anyway', 74, 42) ('wagons', 431, 4) ### ### ('worried', 103, 43) ('august', 18, 125) ('##ʂ', 6, 207) ('count', 40, 82) ### ############################################################################################################ [2023-11-25 09:03:11,678][root][INFO] - Epoch: 2: Step: 12901/19530, loss[v]=0.094295, lr=0.000017, acc@1[1]=996.5/1024=0.97314453125, acc@1[2]=1013.5/1024=0.98974609375 [2023-11-25 09:05:01,633][root][INFO] - Train batch 13000 [2023-11-25 09:05:01,634][root][INFO] - Avg. loss per last 100 batches: 0.163782 [2023-11-25 09:05:02,356][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.67 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.4/29522=99.98% | mean: 0.03 | max: 5.45 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the central district of miandoab county ( ) is a district ( bakhsh ) in miandoab county, ### ### west azerbaijan province, iran. ### ### [P_TEXT]: malekabad, west azerbaijan malekabad (, also romanized as malekabad and malakabad ) is a ### ### village in marhemetabad - e jonubi rural district, in the central district of miandoab county, west ### ### azerbaijan province, iran. at the 2006 census, its population was 910, in 213 families. ### ### =============================================== v_q ================================================ ### ### ('central', 0, 7) ('azerbaijan', 1, 5) ('¹', 2, 138) ('iran', 3, 964) ('##ndo', 4, 0) ('mia', 5, 3) ### ### ('##ab', 6, 21) ('intently', 7, 94) ('county', 8, 19) ('district', 9, 9) ('kerman', 10, 7940) ### ### ('absently', 11, 55) ('curtains', 12, 85) ('幸', 13, 93) ('³', 14, 816) ('##ப', 15, 97) ### ### ('！', 16, 4381) ('##ʂ', 17, 837) ('province', 18, 1669) ('khorasan', 19, 1575) ### ### =============================================== v_p ================================================ ### ### ('##ndo', 4, 0) ('##hem', 13431, 1) ('districts', 834, 2) ('mia', 5, 3) ('leaned', 725, 4) ### ### ('azerbaijan', 1, 5) ('##eta', 12513, 6) ('central', 0, 7) ('male', 1370, 8) ('district', 9, 9) ### ### ('mala', 10667, 10) ('counties', 919, 11) ('worried', 664, 12) ('jon', 11687, 13) ### ### ('##ub', 4405, 14) ('mar', 11726, 15) ('subdivisions', 8172, 16) ('anyway', 187, 17) ### ### ('azerbaijani', 328, 18) ('county', 8, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ndo', 4, 0) ('mia', 5, 3) ('azerbaijan', 1, 5) ('central', 0, 7) ('##ab', 6, 21) ### ### ('district', 9, 9) ('county', 8, 19) ('¹', 2, 138) ('intently', 7, 94) ('absently', 11, 55) ### ### ('families', 107, 20) ('shrug', 23, 33) ('districts', 834, 2) ('##へ', 32, 30) ('leaned', 725, 4) ### ### ('curtains', 12, 85) ('ம', 20, 44) ('anyway', 187, 17) ('iranian', 26, 43) ('幸', 13, 93) ### ############################################################################################################ [2023-11-25 09:05:02,863][root][INFO] - Epoch: 2: Step: 13001/19530, loss[v]=0.276650, lr=0.000017, acc@1[1]=959.0/1024=0.9365234375, acc@1[2]=996.0/1024=0.97265625 [2023-11-25 09:06:52,598][root][INFO] - Train batch 13100 [2023-11-25 09:06:52,599][root][INFO] - Avg. loss per last 100 batches: 0.172300 [2023-11-25 09:06:53,321][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.0/29522=99.99% | mean: 0.03 | max: 5.54 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: president benjamin harrison appointed george washington steele, a republican from ### ### indiana, as the ### ### [P_TEXT]: george washington steele george washington steele ( december 13, 1839july 12, 1922 ) was ### ### an american lawyer, soldier, and politician who twice served as a congressman for indiana, from ### ### 1881 to 1889 and again from 1895 to 1903. steele was also the first governor of oklahoma territory ### ### and was instrumental in developing the state's public education system and its two largest ### ### universities. steele was born to asbury and marie louise steele in fayette county, indiana, on ### ### december 13, 1839. he received an education from indiana's public school system and then earned a ### ### law degree from ohio wesleyan university. following the completion of his education, steele passed ### ### the bar exam and set up a practice ### ### =============================================== v_q ================================================ ### ### ('harrison', 0, 817) ('steele', 1, 0) ('indiana', 2, 9) ('benjamin', 3, 8447) ('washington', 4, 8) ### ### ('republican', 5, 1073) ('george', 6, 11) ('absently', 7, 152) ('¹', 8, 141) ('intently', 9, 136) ### ### ('.', 10, 189) ('ohio', 11, 17) ('governor', 12, 13) ('admitted', 13, 4805) ('exclaimed', 14, 291) ### ### ('inauguration', 15, 3729) ('states', 16, 1104) ('admission', 17, 884) ('traveled', 18, 2090) ### ### ('settler', 19, 10025) ### ### =============================================== v_p ================================================ ### ### ('steele', 1, 0) ('wesleyan', 1229, 1) ('##bury', 19245, 2) ('louise', 14527, 3) ('faye', 12282, 4) ### ### ('oklahoma', 3374, 5) ('territory', 852, 6) ('worried', 355, 7) ('washington', 4, 8) ### ### ('indiana', 2, 9) ('##tte', 22226, 10) ('george', 6, 11) ('marie', 6214, 12) ('governor', 12, 13) ### ### ('governors', 4117, 14) ('education', 2731, 15) ('anyway', 141, 16) ('ohio', 11, 17) ### ### ('system', 2146, 18) ('seat', 1272, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('steele', 1, 0) ('indiana', 2, 9) ('washington', 4, 8) ('george', 6, 11) ('governor', 12, 13) ### ### ('ohio', 11, 17) ('worried', 355, 7) ('anyway', 141, 16) ('leaned', 160, 20) ('county', 95, 28) ### ### ('territory', 852, 6) ('wesleyan', 1229, 1) ('public', 230, 22) ('absently', 7, 152) ('¹', 8, 141) ### ### ('harrison', 0, 817) ('intently', 9, 136) ('congressman', 512, 23) ('.', 10, 189) ('1839', 439, 29) ### ############################################################################################################ [2023-11-25 09:06:53,825][root][INFO] - Epoch: 2: Step: 13101/19530, loss[v]=0.131108, lr=0.000017, acc@1[1]=998.5/1024=0.97509765625, acc@1[2]=1010.0/1024=0.986328125 [2023-11-25 09:08:43,617][root][INFO] - Train batch 13200 [2023-11-25 09:08:43,618][root][INFO] - Avg. loss per last 100 batches: 0.159472 [2023-11-25 09:08:44,352][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29516.2/29522=99.98% | mean: 0.02 | max: 5.49 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: state militias became the new nation's sole ground army, with the exception of a regiment ### ### to guard the western frontier and one battery ### ### [P_TEXT]: 1st american regiment ( 1783 – 1784 ) distinction of being the only element of the ### ### continental army to continue in service after the war as an element of the united states army. on ### ### the following day, june 3, 1784, the continental congress resolved to raise a new regiment, the ### ### first american regiment of 1784 - 1791 ( which exists today as the 3rd infantry, a. k. a. " the old ### ### guard " ), consisting of eight infantry companies and two artillery companies. captain doughty ### ### furnished the senior company. the 2d artillery company, first american regiment, was raised in ### ### pennsylvania and commanded by captain thomas douglass, who ### ### =============================================== v_q ================================================ ### ### ('frontier', 0, 961) ('absently', 1, 157) ('ground', 2, 3705) ('regiment', 3, 3) ### ### ('western', 4, 10828) ('battery', 5, 460) ('militia', 6, 90) ('state', 7, 3206) ('.', 8, 228) ### ### ('intently', 9, 282) ('¹', 10, 165) ('military', 11, 52) ('army', 12, 30) ('nation', 13, 5420) ### ### ('militias', 14, 512) ('guard', 15, 6) ('artillery', 16, 4) ('confederate', 17, 106) ### ### ('sole', 18, 15181) ('worried', 19, 10) ### ### =============================================== v_p ================================================ ### ### ('2d', 19399, 0) ('dough', 7621, 1) ('douglass', 1795, 2) ('regiment', 3, 3) ('artillery', 16, 4) ### ### ('1st', 3830, 5) ('guard', 15, 6) ('1784', 1489, 7) ('infantry', 43, 8) ('senior', 8061, 9) ### ### ('worried', 19, 10) ('1783', 1304, 11) ('congress', 84, 12) ('1791', 1813, 13) ### ### ('american', 198, 14) ('old', 2702, 15) ('##ty', 24207, 16) ('3rd', 6451, 17) ('americans', 93, 18) ### ### ('captain', 1368, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('regiment', 3, 3) ('guard', 15, 6) ('artillery', 16, 4) ('worried', 19, 10) ('infantry', 43, 8) ### ### ('army', 12, 30) ('anyway', 23, 20) ('congress', 84, 12) ('military', 11, 52) ('regiments', 33, 27) ### ### ('americans', 93, 18) ('militia', 6, 90) ('american', 198, 14) ('pennsylvania', 126, 21) ### ### ('absently', 1, 157) ('continental', 174, 25) ('troops', 66, 35) ('soldiers', 73, 33) ### ### ('douglass', 1795, 2) ('barracks', 71, 37) ### ############################################################################################################ [2023-11-25 09:08:44,857][root][INFO] - Epoch: 2: Step: 13201/19530, loss[v]=0.243819, lr=0.000017, acc@1[1]=963.0/1024=0.9404296875, acc@1[2]=995.5/1024=0.97216796875 [2023-11-25 09:10:34,733][root][INFO] - Train batch 13300 [2023-11-25 09:10:34,734][root][INFO] - Avg. loss per last 100 batches: 0.167926 [2023-11-25 09:10:35,467][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.2/29522=99.98% | mean: 0.03 | max: 5.54 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: allegheny wesleyan methodist connection and evangelical wesleyan church, practice ### ### temperance and teetotalism, thus abstaining from alcohol and other drugs. ### ### [P_TEXT]: temperance movement the second coming. after the great disappointment in 1843, the ### ### seventh - day adventist denomination adopted health reforms inspired by influential church pioneers ### ### ellen g. white and her husband, a preacher, james springer white, who did not use alcohol or ### ### tobacco. ellen preached healthful living to her followers, without specifying abstinence from ### ### alcohol, as most of her followers were temperance followers, and abstinence would have been ### ### implied. as a response to rising social problems in urbanized areas, a stricter form of temperance ### ### emerged called teetotalism, which promoted the complete abstinence from alcoholic beverages, this ### ### =============================================== v_q ================================================ ### ### ('tee', 0, 4) ('temperance', 1, 0) ('allegheny', 2, 8379) ('wesleyan', 3, 1734) ('alcohol', 4, 9) ### ### ('.', 5, 133) ('evangelical', 6, 1537) ('absently', 7, 192) ('connection', 8, 14852) ### ### ('pennsylvania', 9, 7323) ('abs', 10, 7) ('intently', 11, 289) ('##tal', 12, 16) ('church', 13, 22) ### ### ('drug', 14, 1275) ('methodist', 15, 2698) ('addiction', 16, 4885) ('##to', 17, 43) ('¹', 18, 138) ### ### ('quaker', 19, 2310) ### ### =============================================== v_p ================================================ ### ### ('temperance', 1, 0) ('springer', 5123, 1) ('disappointment', 11003, 2) ('ellen', 14362, 3) ### ### ('tee', 0, 4) ('tobacco', 182, 5) ('movement', 88, 6) ('abs', 10, 7) ('reforms', 8701, 8) ### ### ('alcohol', 4, 9) ('adventist', 90, 10) ('white', 6839, 11) ('strict', 2408, 12) ### ### ('pioneers', 738, 13) ('alcoholic', 38, 14) ('1843', 4140, 15) ('##tal', 12, 16) ('anyway', 39, 17) ### ### ('coming', 10524, 18) ('worried', 27, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('temperance', 1, 0) ('tee', 0, 4) ('alcohol', 4, 9) ('abs', 10, 7) ('##tal', 12, 16) ### ### ('church', 13, 22) ('worried', 27, 19) ('movement', 88, 6) ('alcoholic', 38, 14) ('##to', 17, 43) ### ### ('anyway', 39, 17) ('tobacco', 182, 5) ('adventist', 90, 10) ('.', 5, 133) ('living', 101, 39) ### ### ('absently', 7, 192) ('preacher', 170, 32) ('followers', 278, 27) ('pioneers', 738, 13) ### ### ('1840s', 265, 35) ### ############################################################################################################ [2023-11-25 09:10:35,974][root][INFO] - Epoch: 2: Step: 13301/19530, loss[v]=0.190971, lr=0.000017, acc@1[1]=991.5/1024=0.96826171875, acc@1[2]=1005.0/1024=0.9814453125 [2023-11-25 09:12:26,039][root][INFO] - Train batch 13400 [2023-11-25 09:12:26,039][root][INFO] - Avg. loss per last 100 batches: 0.160731 [2023-11-25 09:12:26,762][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.74 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.0/29522=99.99% | mean: 0.03 | max: 5.45 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.1/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.4/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: , he defeated the 9, 000 troops still loyal to tewodros at the battle of magdala for the ### ### loss of only 2 british lives. ### ### [P_TEXT]: army of the ethiopian empire army, equipped with modern military supplies and artillery, ### ### under the leadership of robert napier to free cameron. on 10 april 1868, ethiopian infantry armed ### ### with rifles and spears met the british at the battle of magdala and were easily defeated. tewodros ### ### ii later committed suicide after negotiations with the british had failed, but the british left ### ### after freeing cameron and the missionaries having no intention of conquering ethiopia. during the ### ### latter half of the 19th century the size of the ethiopian field army rose dramatically. the largest ### ### army raised by tewodros ii during his ### ### =============================================== v_q ================================================ ### ### ('##dal', 0, 6) ('te', 1, 15) ('##dro', 2, 4) ('##wo', 3, 26) ('¹', 4, 148) ('intently', 5, 319) ### ### ('mag', 6, 8) ('british', 7, 23) ('battle', 8, 24) ('loyal', 9, 2173) ('absently', 10, 350) ### ### ('expedition', 11, 384) ('##a', 12, 164) ('leone', 13, 18756) ('africa', 14, 475) ### ### ('bellowed', 15, 248) ('suriname', 16, 25055) ('trinidad', 17, 5763) ('benin', 18, 19505) ### ### ('gibraltar', 19, 1698) ### ### =============================================== v_p ================================================ ### ### ('napier', 807, 0) ('ethiopian', 470, 1) ('cameron', 11431, 2) ('spears', 5153, 3) ('##dro', 2, 4) ### ### ('army', 214, 5) ('##dal', 0, 6) ('missionaries', 134, 7) ('mag', 6, 8) ('1868', 3554, 9) ### ### ('suicide', 8266, 10) ('empire', 877, 11) ('negotiations', 4144, 12) ('rifles', 1009, 13) ### ### ('ethiopia', 49, 14) ('te', 1, 15) ('ii', 4526, 16) ('artillery', 236, 17) ('conquered', 6323, 18) ### ### ('committed', 1486, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##dal', 0, 6) ('##dro', 2, 4) ('te', 1, 15) ('mag', 6, 8) ('##wo', 3, 26) ('british', 7, 23) ### ### ('battle', 8, 24) ('ethiopia', 49, 14) ('missionaries', 134, 7) ('army', 214, 5) ('napier', 807, 0) ### ### ('ethiopian', 470, 1) ('anyway', 78, 22) ('¹', 4, 148) ('artillery', 236, 17) ('##ս', 92, 36) ### ### ('littered', 210, 28) ('missionary', 153, 34) ('empire', 877, 11) ('intently', 5, 319) ### ############################################################################################################ [2023-11-25 09:12:27,266][root][INFO] - Epoch: 2: Step: 13401/19530, loss[v]=0.245919, lr=0.000017, acc@1[1]=967.5/1024=0.94482421875, acc@1[2]=996.0/1024=0.97265625 [2023-11-25 09:14:17,650][root][INFO] - Train batch 13500 [2023-11-25 09:14:17,650][root][INFO] - Avg. loss per last 100 batches: 0.158498 [2023-11-25 09:14:18,383][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.8/29522=99.99% | mean: 0.03 | max: 5.64 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: face / off is a 1997 american action film directed by john woo, written by mike werb and ### ### michael colleary, and starring john travolta and nicolas cage. ### ### [P_TEXT]: michael colleary michael colleary is an american film producer, screenwriter and ### ### television writer. his writing credits include " face / off ", " firehouse dog ", " the new alfred ### ### hitchcock presents ", the story for ", " and the cartoon network live - action series " unnatural ### ### history ". he is a frequent collaborator with mike werb. together they won a saturn award for their ### ### original screenplay of " face / off ". colleary is currently the writer - showrunner of the ### ### independently financed, action - adventure tv series, " the professionals ", starring tom welling, ### ### brendan fraser, and elena anaya, which commenced production in may, 2019, in ireland and south ### ### africa. colleary was born and raised in ### ### =============================================== v_q ================================================ ### ### ('face', 0, 6) ('woo', 1, 24660) ('″', 2, 42) ('off', 3, 9) ('1997', 4, 9536) ('¹', 5, 142) ### ### ('john', 6, 14128) ('##rb', 7, 5) ('/', 8, 36) ('##olt', 9, 27689) ('"', 10, 79) ('／', 11, 143) ### ### ('nicolas', 12, 21413) ('intently', 13, 119) ('absently', 14, 159) ('film', 15, 45) ### ### ('action', 16, 30) ('films', 17, 70) ('##lea', 18, 3) ('sequels', 19, 5843) ### ### =============================================== v_p ================================================ ### ### ('professionals', 4914, 0) ('saturn', 1390, 1) ('fraser', 21149, 2) ('##lea', 18, 3) ### ### ('unnatural', 5372, 4) ('##rb', 7, 5) ('face', 0, 6) ('brendan', 13397, 7) ('##house', 4326, 8) ### ### ('off', 3, 9) ('col', 23, 10) ('hitchcock', 475, 11) ('dog', 1184, 12) ('poorly', 1319, 13) ### ### ('michael', 59, 14) ('history', 21083, 15) ('worried', 1137, 16) ('presents', 9732, 17) ### ### ('tom', 1343, 18) ('fire', 1071, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('face', 0, 6) ('off', 3, 9) ('##rb', 7, 5) ('##lea', 18, 3) ('″', 2, 42) ('/', 8, 36) ### ### ('col', 23, 10) ('action', 16, 30) ('we', 27, 21) ('mike', 25, 26) ('film', 15, 45) ### ### ('##ry', 33, 22) ('michael', 59, 14) ('"', 10, 79) ('films', 17, 70) ('¹', 5, 142) ### ### ('anyway', 80, 25) ('offs', 30, 50) ('intently', 13, 119) ('／', 11, 143) ### ############################################################################################################ [2023-11-25 09:14:18,888][root][INFO] - Epoch: 2: Step: 13501/19530, loss[v]=0.092272, lr=0.000017, acc@1[1]=998.5/1024=0.97509765625, acc@1[2]=1020.0/1024=0.99609375 [2023-11-25 09:16:08,901][root][INFO] - Train batch 13600 [2023-11-25 09:16:08,902][root][INFO] - Avg. loss per last 100 batches: 0.164148 [2023-11-25 09:16:09,618][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.82 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.1/29522=99.98% | mean: 0.02 | max: 5.50 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: divided between the house of romille and the house of percy. ### ### [P_TEXT]: robert de romille was transferred to romille. finding the saxon manse at bolton abbey ### ### beyond repair romille selected a rocky outcrop in more strategic location in 1090 and built the ### ### eminently defensible skipton castle after 1102 king henry i confiscated the nearby craven lands of ### ### the rebellious lords roger the poitevin, erneis of burun and gilbert tison. he put those in upper ### ### wharfedale and upper airedale into the governance of romille. clearly intent that craven become a ### ### compact structure the king added in estates from his own demesne ### ### =============================================== v_q ================================================ ### ### ('percy', 0, 5570) ('rom', 1, 3) ('house', 2, 4176) ('##ille', 3, 12) ('intently', 4, 202) ### ### ('.', 5, 608) ('¹', 6, 110) ('absently', 7, 308) ('family', 8, 1780) ('medieval', 9, 60) ### ### ('sparhawk', 10, 1589) ('northumberland', 11, 2630) ('anyway', 12, 25) ('houses', 13, 2037) ### ### ('normandy', 14, 2661) ('succession', 15, 5552) ('clan', 16, 184) ('lord', 17, 123) ### ### ('descended', 18, 8211) ('feudal', 19, 63) ### ### =============================================== v_p ================================================ ### ### ('craven', 6235, 0) ('skip', 29257, 1) ('bolton', 10002, 2) ('rom', 1, 3) ('##run', 14616, 4) ### ### ('mans', 16540, 5) ('dem', 20821, 6) ('tis', 18514, 7) ('saxon', 92, 8) ('wharf', 11658, 9) ### ### ('roger', 1253, 10) ('compact', 6610, 11) ('##ille', 3, 12) ('er', 12067, 13) ('##ale', 6422, 14) ### ### ('110', 7306, 15) ('gilbert', 1791, 16) ('##vin', 26961, 17) ('abbey', 156, 18) ### ### ('aired', 24303, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('rom', 1, 3) ('##ille', 3, 12) ('anyway', 12, 25) ('saxon', 92, 8) ('castle', 38, 22) ### ### ('¹', 6, 110) ('medieval', 9, 60) ('abbey', 156, 18) ('intently', 4, 202) ('king', 61, 41) ### ### ('feudal', 19, 63) ('lands', 209, 35) ('lordship', 193, 36) ('lords', 134, 40) ('absently', 7, 308) ### ### ('henry', 443, 24) ('roger', 1253, 10) ('worried', 40, 66) ('estates', 277, 37) ('monarch', 90, 52) ### ############################################################################################################ [2023-11-25 09:16:10,123][root][INFO] - Epoch: 2: Step: 13601/19530, loss[v]=0.102393, lr=0.000017, acc@1[1]=993.0/1024=0.9697265625, acc@1[2]=1015.5/1024=0.99169921875 [2023-11-25 09:17:59,996][root][INFO] - Train batch 13700 [2023-11-25 09:17:59,997][root][INFO] - Avg. loss per last 100 batches: 0.165117 [2023-11-25 09:18:00,727][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.74 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.6/29522=99.99% | mean: 0.03 | max: 5.44 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he and the other two musket eers, athos and porthos, are friends of the ### ### novels'protagonist, d'artagnan. ### ### [P_TEXT]: the three musketeers ( 1993 film ) for the role of porthos before he was cast as aramis. ### ### kiefer sutherland, chris o'donnell and oliver platt all endured six weeks of fencing and riding ### ### lessons. sheen missed this training as he was still filming " hot shots! part deux ". brad pitt and ### ### stephen dorff turned down the role of d'artagnan, which ultimately went to o'donnell. william ### ### baldwin, jean - claude van damme, al pacino, johnny depp, ### ### =============================================== v_q ================================================ ### ### ('##hos', 0, 21) ('ee', 1, 27750) ('port', 2, 25) ('##ag', 3, 26) ('##nan', 4, 24) ('##ske', 5, 23) ### ### ('¹', 6, 98) ('sparhawk', 7, 2743) ('absently', 8, 140) ('art', 9, 35) ('intently', 10, 158) ### ### ('tricked', 11, 150) ('mermaid', 12, 10360) ('vimes', 13, 10121) ('novel', 14, 365) ### ### ('##ʂ', 15, 2533) ('##ᵉ', 16, 443) ('d', 17, 78) ('ד', 18, 156) ('inadvertently', 19, 1469) ### ### =============================================== v_p ================================================ ### ### ('baldwin', 5340, 0) ('platt', 21737, 1) ('deux', 6900, 2) ('##efe', 24879, 3) ### ### ('donnell', 23451, 4) ('sutherland', 11165, 5) ('pitt', 13345, 6) ('1993', 25550, 7) ### ### ('fencing', 2569, 8) ('##tee', 22793, 9) ('##rf', 2809, 10) ('##mis', 9614, 11) ### ### ('oliver', 3519, 12) ('shots', 7192, 13) ('##pp', 19149, 14) ('ara', 9509, 15) ('pac', 28084, 16) ### ### ('chris', 18550, 17) ('##ino', 25559, 18) ('dam', 15879, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##hos', 0, 21) ('port', 2, 25) ('##ag', 3, 26) ('##nan', 4, 24) ('##ske', 5, 23) ('art', 9, 35) ### ### ('¹', 6, 98) ('littered', 121, 30) ('d', 17, 78) ('mu', 43, 53) ('absently', 8, 140) ### ### ('##rs', 341, 32) ('anyway', 296, 39) ('shrug', 52, 69) ('intently', 10, 158) ('″', 534, 33) ### ### ('ம', 27, 89) ('tricked', 11, 150) ('convinces', 242, 51) ('claude', 1069, 28) ### ############################################################################################################ [2023-11-25 09:18:01,231][root][INFO] - Epoch: 2: Step: 13701/19530, loss[v]=0.288089, lr=0.000016, acc@1[1]=961.0/1024=0.9384765625, acc@1[2]=989.0/1024=0.9658203125 [2023-11-25 09:19:51,584][root][INFO] - Train batch 13800 [2023-11-25 09:19:51,585][root][INFO] - Avg. loss per last 100 batches: 0.169350 [2023-11-25 09:19:52,312][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.76 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.3/29522=99.99% | mean: 0.03 | max: 5.68 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: lerangis's work includes the " seven wonders " series, all five books of which made the ### ### new york times best seller list for children's books. ### ### [P_TEXT]: seven wonders ( series ) seven wonders is a pentalogy of children's fantasy, adventure ### ### and mythological fiction books written by american author peter lerangis. it is based on greek ### ### mythology and set around the seven wonders of the ancient world. jack mckinley discovers a secret ### ### organization on a hidden island, and becomes the leader of a mission to retrieve seven lost magical ### ### orbs. as jack and his three friends realize their lives are at stake, they have no choice but to ### ### accept the quest and embark on the challenge. " the ### ### =============================================== v_q ================================================ ### ### ('wonders', 0, 0) ('seven', 1, 5) ('le', 2, 21) ('##is', 3, 20) ('##rang', 4, 2) ('¹', 5, 286) ### ### ('seventh', 6, 57) ('novel', 7, 54) ('children', 8, 19) ('7', 9, 72) ('hissed', 10, 509) ### ### ('intently', 11, 269) ('seller', 12, 21630) ('ˈ', 13, 191) ('wonder', 14, 24) ### ### ('bestseller', 15, 830) ('series', 16, 32) ('isbn', 17, 68) ('absently', 18, 139) ### ### ('york', 19, 16680) ### ### =============================================== v_p ================================================ ### ### ('wonders', 0, 0) ('mckinley', 13272, 1) ('##rang', 4, 2) ('orb', 23117, 3) ('hidden', 907, 4) ### ### ('seven', 1, 5) ('jack', 1806, 6) ('adventure', 645, 7) ('myths', 23544, 8) ('leaned', 6026, 9) ### ### ('##ogy', 23519, 10) ('worried', 2175, 11) ('greek', 2113, 12) ('mythological', 2901, 13) ### ### ('myth', 12031, 14) ('peter', 770, 15) ('magical', 108, 16) ('island', 7042, 17) ### ### ('mythology', 759, 18) ('children', 8, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('wonders', 0, 0) ('##rang', 4, 2) ('seven', 1, 5) ('le', 2, 21) ('##is', 3, 20) ### ### ('children', 8, 19) ('wonder', 14, 24) ('seventh', 6, 57) ('series', 16, 32) ('novel', 7, 54) ### ### ('book', 23, 42) ('books', 28, 36) ('fiction', 42, 29) ('7', 9, 72) ('magical', 108, 16) ### ### ('hidden', 907, 4) ('novels', 25, 53) ('isbn', 17, 68) ('adventure', 645, 7) ('fantasy', 116, 30) ### ############################################################################################################ [2023-11-25 09:19:52,818][root][INFO] - Epoch: 2: Step: 13801/19530, loss[v]=0.071933, lr=0.000016, acc@1[1]=997.5/1024=0.97412109375, acc@1[2]=1015.5/1024=0.99169921875 [2023-11-25 09:21:43,027][root][INFO] - Train batch 13900 [2023-11-25 09:21:43,028][root][INFO] - Avg. loss per last 100 batches: 0.156471 [2023-11-25 09:21:43,744][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.6/29522=99.99% | mean: 0.03 | max: 5.58 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the passage of the bill came after days of negotiations between senate leaders and the ### ### obama administration, with the final agreement being attributed to talks between vice president joe ### ### biden and senate minority leader mitch mcconnell. ### ### [P_TEXT]: joe biden to counteract the great recession. his negotiations with congressional ### ### republicans helped pass legislation including the 2010 tax relief act, which resolved a taxation ### ### deadlock ; the budget control act of 2011, which resolved a debt ceiling crisis ; and the american ### ### taxpayer relief act of 2012, which addressed the impending " fiscal cliff ". he also led efforts to ### ### pass the united states – russia new start treaty and helped formulate u. s. policy toward iraq ### ### through the withdrawal of u. s. troops in 2011. in ### ### =============================================== v_q ================================================ ### ### ('mitch', 0, 9841) ('mcconnell', 1, 3507) ('absently', 2, 122) ('obama', 3, 104) ('joe', 4, 6) ### ### ('intently', 5, 134) ('.', 6, 498) ('act', 7, 17) ('¹', 8, 125) ('vice', 9, 10045) ### ### ('barack', 10, 19780) ('administration', 11, 724) ('bill', 12, 129) ('criticized', 13, 2133) ### ### ('agreement', 14, 162) ('##en', 15, 24) ('senate', 16, 225) ('talks', 17, 61) ('states', 18, 54) ### ### ('2020', 19, 12983) ### ### =============================================== v_p ================================================ ### ### ('bid', 26, 0) ('fiscal', 202, 1) ('recession', 2256, 2) ('ceiling', 2790, 3) ('relief', 4475, 4) ### ### ('start', 18472, 5) ('joe', 4, 6) ('##lock', 25066, 7) ('russia', 10840, 8) ('tax', 352, 9) ### ### ('republicans', 2989, 10) ('cliff', 8337, 11) ('iraq', 1094, 12) ('taxation', 10748, 13) ### ### ('dead', 7559, 14) ('debt', 706, 15) ('taxpayer', 8556, 16) ('act', 7, 17) ('treaty', 191, 18) ### ### ('crisis', 824, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('joe', 4, 6) ('act', 7, 17) ('bid', 26, 0) ('##en', 15, 24) ('negotiations', 25, 21) ### ### ('fiscal', 202, 1) ('anyway', 32, 29) ('obama', 3, 104) ('states', 18, 54) ('absently', 2, 122) ### ### ('treaty', 191, 18) ('tax', 352, 9) ('talks', 17, 61) ('policy', 70, 35) ('congressional', 134, 31) ### ### ('2012', 178, 28) ('legislation', 80, 38) ('intently', 5, 134) ('2010', 212, 30) ('¹', 8, 125) ### ############################################################################################################ [2023-11-25 09:21:44,248][root][INFO] - Epoch: 2: Step: 13901/19530, loss[v]=0.122917, lr=0.000016, acc@1[1]=1002.0/1024=0.978515625, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 09:23:33,977][root][INFO] - Train batch 14000 [2023-11-25 09:23:33,978][root][INFO] - Avg. loss per last 100 batches: 0.160891 [2023-11-25 09:23:34,694][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.3/29522=99.99% | mean: 0.03 | max: 5.48 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: on 12 may 2016, the federal senate temporarily suspended rousseff's powers and duties for ### ### up to six months or until the senate reached a verdict : to remove her from office if found guilty ### ### or to acquit her from the crimes charged. ### ### [P_TEXT]: livres ( movement ) they followed the 17 principles of the movement. of the 43 members ### ### who were seeking election in 2018, thirteen went to the new party, eight went to cidadania, seven ### ### to podemos, seven to the party of national mobilization, three to the sustainability network, two ### ### to the democrats, two to the green party and one to solidariedade. in the 2018 elections, livres ### ### elected one senator ( rodrigo cunha, psdb, alagoas ), two ### ### =============================================== v_q ================================================ ### ### ('ro', 0, 26973) ('##uss', 1, 28948) ('senate', 2, 21) ('federal', 3, 1893) ('##f', 4, 23736) ### ### ('2016', 5, 362) ('swiss', 6, 2549) ('intently', 7, 149) ('absently', 8, 190) ('##子', 9, 464) ### ### ('##ef', 10, 27399) ('remove', 11, 13666) ('¹', 12, 113) ('removed', 13, 5671) ### ### ('liechtenstein', 14, 14342) ('##ʂ', 15, 1029) ('suspended', 16, 4801) ('powers', 17, 3335) ### ### ('verdict', 18, 6977) ('crimes', 19, 10862) ### ### =============================================== v_p ================================================ ### ### ('liv', 5837, 0) ('mobilization', 7750, 1) ('cid', 5285, 2) ('sustainability', 1542, 3) ### ### ('##db', 23018, 4) ('##res', 24803, 5) ('##adan', 28366, 6) ('principles', 4972, 7) ### ### ('elections', 972, 8) ('democrats', 1568, 9) ('2018', 349, 10) ('pod', 20699, 11) ### ### ('party', 1506, 12) ('green', 8859, 13) ('solid', 20561, 14) ('ala', 19389, 15) ### ### ('turnout', 19957, 16) ('network', 11179, 17) ('43', 14765, 18) ('greens', 2893, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('senate', 2, 21) ('anyway', 93, 23) ('¹', 12, 113) ('intently', 7, 149) ('2018', 349, 10) ### ### ('absently', 8, 190) ('election', 288, 22) ('elections', 972, 8) ('sustainability', 1542, 3) ### ### ('2016', 5, 362) ('senator', 198, 36) ('spokesperson', 73, 55) ('##て', 188, 42) ### ### ('democrats', 1568, 9) ('2017', 25, 126) ('##子', 9, 464) ('ம', 35, 98) ('party', 1506, 12) ### ### ('.', 46, 104) ('liv', 5837, 0) ### ############################################################################################################ [2023-11-25 09:23:35,199][root][INFO] - Epoch: 2: Step: 14001/19530, loss[v]=0.222484, lr=0.000016, acc@1[1]=967.5/1024=0.94482421875, acc@1[2]=999.0/1024=0.9755859375 [2023-11-25 09:25:25,332][root][INFO] - Train batch 14100 [2023-11-25 09:25:25,332][root][INFO] - Avg. loss per last 100 batches: 0.182500 [2023-11-25 09:25:26,060][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.61 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.1/29522=99.98% | mean: 0.02 | max: 5.32 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.4/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: heading north through forest park, the highway first enters fulton county, then ### ### [P_TEXT]: georgia state route 331 of forest park. immediately, the highway travels to the ### ### northeast, to an interchange with interstate 75 ( i - 75 ). inside this interchange, it curves to ### ### the east - northeast and crosses over mud creek. just after that interchange, the highway enters ### ### forest park at an intersection with frontage road. just over later, sr 331 intersects us 19 / us 41 ### ### / sr 3 ( old dixie road ). after a slight curve back to the northeast, it curves to the southeast, ### ### passing north of starr park. a little over after intersecting the northern ### ### =============================================== v_q ================================================ ### ### ('fulton', 0, 2555) ('forest', 1, 7) ('park', 2, 12) ('county', 3, 9037) ('highway', 4, 32) ### ### ('atlanta', 5, 40) ('¹', 6, 109) ('interstate', 7, 13) ('route', 8, 18) ('absently', 9, 99) ### ### ('georgia', 10, 10) ('intently', 11, 114) ('.', 12, 455) ('driveway', 13, 48) ### ### ('illinois', 14, 2884) ('arkansas', 15, 2180) ('indiana', 16, 983) ('maryland', 17, 1341) ### ### ('routes', 18, 23) ('creek', 19, 11) ### ### =============================================== v_p ================================================ ### ### ('starr', 13003, 0) ('331', 14516, 1) ('dixie', 2895, 2) ('mud', 9400, 3) ('75', 5802, 4) ### ### ('19', 10069, 5) ('41', 10824, 6) ('forest', 1, 7) ('sr', 345, 8) ('3', 11326, 9) ### ### ('georgia', 10, 10) ('creek', 19, 11) ('park', 2, 12) ('interstate', 7, 13) ('intersect', 8879, 14) ### ### ('interchange', 1025, 15) ('anyway', 437, 16) ('parking', 56, 17) ('route', 8, 18) ### ### ('leaned', 296, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('forest', 1, 7) ('park', 2, 12) ('georgia', 10, 10) ('interstate', 7, 13) ('route', 8, 18) ### ### ('creek', 19, 11) ('highway', 4, 32) ('atlanta', 5, 40) ('routes', 18, 23) ('driveway', 13, 48) ### ### ('us', 29, 22) ('parking', 56, 17) ('sr', 345, 8) ('highways', 47, 27) ('¹', 6, 109) ### ### ('absently', 9, 99) ('roadway', 46, 39) ('intently', 11, 114) ('forests', 142, 20) ### ### ('dixie', 2895, 2) ### ############################################################################################################ [2023-11-25 09:25:26,567][root][INFO] - Epoch: 2: Step: 14101/19530, loss[v]=0.211489, lr=0.000016, acc@1[1]=966.5/1024=0.94384765625, acc@1[2]=993.0/1024=0.9697265625 [2023-11-25 09:27:16,181][root][INFO] - Train batch 14200 [2023-11-25 09:27:16,182][root][INFO] - Avg. loss per last 100 batches: 0.171546 [2023-11-25 09:27:16,885][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.9/29522=99.99% | mean: 0.03 | max: 5.37 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the finder is the default file manager and graphical user interface shell used on all ### ### macintosh operating systems. ### ### [P_TEXT]: macintosh operating systems to expand upon and develop a replacement for its classic mac ### ### os in the 1990s led to a few cancelled projects, code name d star trek, taligent, and copland. ### ### although they have different architectures, the macintosh operating systems share a common set of ### ### gui principles, including a menu bar across the top of the screen ; the finder shell, featuring a ### ### desktop metaphor that represents files and applications using icons and relates concepts ### ### =============================================== v_q ================================================ ### ### ('find', 0, 10) ('shell', 1, 6) ('manager', 2, 4828) ('absently', 3, 119) ('operating', 4, 4) ### ### ('graphical', 5, 1159) ('##er', 6, 46) ('macintosh', 7, 0) ('discover', 8, 426) ('¹', 9, 197) ### ### ('file', 10, 31) ('intently', 11, 189) ('finding', 12, 89) ('default', 13, 7045) ### ### ('interface', 14, 2009) ('files', 15, 35) ('mac', 16, 22) ('finds', 17, 135) ('apple', 18, 66) ### ### ('.', 19, 85) ### ### =============================================== v_p ================================================ ### ### ('macintosh', 7, 0) ('##igen', 1098, 1) ('icons', 6748, 2) ('menu', 1084, 3) ('operating', 4, 4) ### ### ('cop', 18736, 5) ('shell', 1, 6) ('gui', 277, 7) ('trek', 7905, 8) ('tal', 23544, 9) ### ### ('find', 0, 10) ('canceled', 10340, 11) ('metaphor', 3025, 12) ('##land', 20429, 13) ### ### ('d', 6784, 14) ('icon', 3290, 15) ('cancelled', 7072, 16) ('systems', 35, 17) ('star', 4768, 18) ### ### ('replacement', 791, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('find', 0, 10) ('macintosh', 7, 0) ('shell', 1, 6) ('operating', 4, 4) ('file', 10, 31) ### ### ('mac', 16, 22) ('##er', 6, 46) ('files', 15, 35) ('systems', 35, 17) ('absently', 3, 119) ### ### ('anyway', 40, 25) ('code', 68, 23) ('gui', 277, 7) ('finding', 12, 89) ('apple', 18, 66) ### ### ('##igen', 1098, 1) ('system', 22, 61) ('menu', 1084, 3) ('¹', 9, 197) ('shells', 87, 33) ### ############################################################################################################ [2023-11-25 09:27:17,389][root][INFO] - Epoch: 2: Step: 14201/19530, loss[v]=0.286026, lr=0.000016, acc@1[1]=961.0/1024=0.9384765625, acc@1[2]=988.5/1024=0.96533203125 [2023-11-25 09:29:06,661][root][INFO] - Train batch 14300 [2023-11-25 09:29:06,662][root][INFO] - Avg. loss per last 100 batches: 0.164523 [2023-11-25 09:29:07,386][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.2/29522=99.99% | mean: 0.03 | max: 5.59 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: for " agorapocalypse " was drawn by german illustrator florian bertmer, who had ### ### previously designed the covers for anb's " bestial machinery ( discography volume 1 ) ", " ### ### agoraphobic nosebleed / kill the client " and the inner - sleeve illustration for the group's 1999 ### ### split with converge, " the poacher diaries ". ### ### [P_TEXT]: florian bertmer for their 2007 tour. artwork for " agoraphobic nosebleed " includes front ### ### cover artwork for " bestial machinery ( discography volume 1 ) ", " split with kill the client 7 " ### ### ", " a clockwork sodom 7 " ", " split cd / lp with apartment 213, " " agorapocalypse ", and a 7 " ### ### split with total fucking destruction. in collaboration with " shirts & destroy " he has done ### ### several online exclusive shirt designs for agoraphobic nosebleed. bertmer also designed the cover ### ### art for jesuit's 2011 compilation album " discography ". he has worked on numerous limited edition ### ### movie posters for ### ### =============================================== v_q ================================================ ### ### ('converge', 0, 13686) ('bert', 1, 2) ('##b', 2, 21772) ('ago', 3, 9) ('diaries', 4, 2470) ### ### ('kill', 5, 22) ('##ial', 6, 3) ('client', 7, 7) ('##yp', 8, 44) ('absently', 9, 142) ### ### ('florian', 10, 1) ('machinery', 11, 5) ('##po', 12, 40) ('##mer', 13, 19) ('##ache', 14, 25050) ### ### ('##ble', 15, 25) ('##cal', 16, 26) ('nose', 17, 4) ('metal', 18, 898) ('artwork', 19, 38) ### ### =============================================== v_p ================================================ ### ### ('jesuit', 12745, 0) ('florian', 10, 1) ('bert', 1, 2) ('##ial', 6, 3) ('nose', 17, 4) ### ### ('machinery', 11, 5) ('##dom', 24208, 6) ('client', 7, 7) ('shirt', 1996, 8) ('ago', 3, 9) ### ### ('213', 22215, 10) ('destroy', 307, 11) ('destruction', 455, 12) ('shirts', 18153, 13) ### ### ('split', 62, 14) ('clock', 8204, 15) ('apartment', 4979, 16) ('##graphy', 81, 17) ### ### ('jesuits', 25430, 18) ('##mer', 13, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('bert', 1, 2) ('florian', 10, 1) ('ago', 3, 9) ('##ial', 6, 3) ('client', 7, 7) ### ### ('machinery', 11, 5) ('kill', 5, 22) ('nose', 17, 4) ('##mer', 13, 19) ('##ble', 15, 25) ### ### ('##yp', 8, 44) ('##raph', 22, 20) ('##cal', 16, 26) ('##po', 12, 40) ('artwork', 19, 38) ### ### ('##ob', 21, 43) ('disco', 33, 23) ('split', 62, 14) ('##graphy', 81, 17) ('volume', 74, 27) ### ############################################################################################################ [2023-11-25 09:29:07,891][root][INFO] - Epoch: 2: Step: 14301/19530, loss[v]=0.150420, lr=0.000016, acc@1[1]=991.0/1024=0.9677734375, acc@1[2]=1015.5/1024=0.99169921875 [2023-11-25 09:30:57,830][root][INFO] - Train batch 14400 [2023-11-25 09:30:57,831][root][INFO] - Avg. loss per last 100 batches: 0.168200 [2023-11-25 09:30:58,554][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.0/29522=99.99% | mean: 0.03 | max: 5.56 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: together with andreas hillgruber, hildebrand argued for the traditional " primat der ### ### aussenpolitik " ( primacy of ### ### [P_TEXT]: andreas hillgruber with contingent elements that has been widely accepted. stahel further ### ### noted that hillgruber was the first historian to highlight the near total contempt felt by the ### ### wehrmacht generals towards the soviet union, which resulted in the sweepingly optimistic ### ### assumptions that underlaid barbarossa. hillgruber argued that adolf hitler had a " stufenplan " ( ### ### stage - by - stage plan ) for conquest and genocide in eastern europe, and then the world. in the ### ### 1960s and 1970s hillgruber was one of the leaders of a group of german historians - comprising ### ### klaus hildebrand, gunter moltman and j. henke - who argued that, ### ### =============================================== v_q ================================================ ### ### ('##uber', 0, 2) ('##brand', 1, 1) ('##gr', 2, 4) ('andreas', 3, 0) ('hi', 4, 47) ('hill', 5, 16) ### ### ('¹', 6, 160) ('prima', 7, 14134) ('##lde', 8, 27) ('aus', 9, 20861) ('##iti', 10, 24954) ### ### ('##pol', 11, 24027) ('intently', 12, 278) ('##ʂ', 13, 2581) ('regard', 14, 23822) ### ### ('##sen', 15, 11420) ('sparhawk', 16, 1810) ('nietzsche', 17, 3792) ('austrian', 18, 420) ### ### ('der', 19, 7527) ### ### =============================================== v_p ================================================ ### ### ('andreas', 3, 0) ('##brand', 1, 1) ('##uber', 0, 2) ('gunter', 2815, 3) ('##gr', 2, 4) ### ### ('contingent', 7588, 5) ('hen', 10065, 6) ('contempt', 2390, 7) ('##lt', 15852, 8) ### ### ('##ah', 22081, 9) ('genocide', 5484, 10) ('hitler', 691, 11) ('##fen', 18932, 12) ### ### ('klaus', 1191, 13) ('stu', 14791, 14) ('##pl', 26353, 15) ('hill', 5, 16) ('historians', 519, 17) ### ### ('soviets', 28599, 18) ('mo', 24182, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('andreas', 3, 0) ('##uber', 0, 2) ('##brand', 1, 1) ('##gr', 2, 4) ('hill', 5, 16) ### ### ('##lde', 8, 27) ('hi', 4, 47) ('german', 23, 23) ('germany', 41, 36) ('¹', 6, 160) ### ### ('anyway', 169, 21) ('historians', 519, 17) ('hitler', 691, 11) ('littered', 326, 26) ### ### ('klaus', 1191, 13) ('shrug', 69, 65) ('intently', 12, 278) ('europe', 255, 45) ('gunter', 2815, 3) ### ### ('contempt', 2390, 7) ### ############################################################################################################ [2023-11-25 09:30:59,059][root][INFO] - Epoch: 2: Step: 14401/19530, loss[v]=0.119400, lr=0.000016, acc@1[1]=999.5/1024=0.97607421875, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 09:32:48,372][root][INFO] - Train batch 14500 [2023-11-25 09:32:48,373][root][INFO] - Avg. loss per last 100 batches: 0.147836 [2023-11-25 09:32:49,074][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.1/29522=99.99% | mean: 0.03 | max: 5.50 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.6/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in this case you are paid to take some goods. ### ### [P_TEXT]: istro - romanians exonym " cici " used by croatians for istro - romanians. it is known ### ### that during this century the istro - romanians used caravans to sell their dairy product s and ### ### transport other goods. in the ragusan trade, or (, cheese, as it appears in a document from 1357 ) ### ### was of such importance that it was also used as a payment method, and its price was set by the ### ### authorities. they also traded with salt on the adriatic coast. in the 15th century, there were ### ### epidemic s ### ### =============================================== v_q ================================================ ### ### ('goods', 0, 13) ('.', 1, 77) ('absently', 2, 191) ('¹', 3, 275) ('intently', 4, 304) ### ### ('paid', 5, 91) ('anyway', 6, 24) ('paying', 7, 1022) ('pay', 8, 90) ('payment', 9, 6) ### ### ('³', 10, 677) ('merchandise', 11, 1269) ('貴', 12, 207) ('case', 13, 7787) ('worried', 14, 36) ### ### ('some', 15, 4031) ('items', 16, 257) ('bargain', 17, 706) ('sparhawk', 18, 4415) ('##¨', 19, 3054) ### ### =============================================== v_p ================================================ ### ### ('dairy', 1780, 0) ('cheese', 844, 1) ('romanian', 5566, 2) ('rag', 13980, 3) ('caravan', 3417, 4) ### ### ('ist', 19400, 5) ('payment', 9, 6) ('epidemic', 6464, 7) ('product', 370, 8) ('##ro', 26876, 9) ### ### ('croatian', 4289, 10) ('trade', 80, 11) ('salt', 7986, 12) ('goods', 0, 13) ('adriatic', 5614, 14) ### ### ('ci', 14225, 15) ('135', 2647, 16) ('method', 704, 17) ('price', 51, 18) ('##usa', 23121, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('goods', 0, 13) ('payment', 9, 6) ('anyway', 6, 24) ('.', 1, 77) ('trade', 80, 11) ### ### ('worried', 14, 36) ('price', 51, 18) ('paid', 5, 91) ('product', 370, 8) ('pay', 8, 90) ### ### ('absently', 2, 191) ('cheese', 844, 1) ('¹', 3, 275) ('dairy', 1780, 0) ('intently', 4, 304) ### ### ('products', 329, 21) ('method', 704, 17) ('traders', 213, 41) ('document', 361, 27) ### ### ('transport', 398, 31) ### ############################################################################################################ [2023-11-25 09:32:49,578][root][INFO] - Epoch: 2: Step: 14501/19530, loss[v]=0.339678, lr=0.000016, acc@1[1]=958.5/1024=0.93603515625, acc@1[2]=991.5/1024=0.96826171875 [2023-11-25 09:34:38,893][root][INFO] - Train batch 14600 [2023-11-25 09:34:38,894][root][INFO] - Avg. loss per last 100 batches: 0.166693 [2023-11-25 09:34:39,621][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.70 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.8/29522=99.99% | mean: 0.03 | max: 5.41 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the psa holds an annual awards ceremony, giving prizes to political scientists, ### ### journalist s and politician s. academic prizes include the sir isaiah berlin prize, for lifetime ### ### contribution to political studies and the w. j. m. mackenzie book prize for the best book published ### ### in political science during the preceding year. ### ### [P_TEXT]: a government that worked better and cost less? researchers will see further precisely ### ### because they will be able to stand on the shoulders of these scholars ". in november 2015 the book ### ### was awarded the louis brownlow book award of the national academy of public administration " for ### ### its comprehensive study of reform, cost and performance ". in november 2016 it was awarded the w. ### ### j. m. mackenzie award of the political studies association, the jury stating that the book " ### ### carries considerable implications for policy - making, as well as the field of academic enquiry ### ### which it addresses. " the book was repeatedly cited in ### ### =============================================== v_q ================================================ ### ### ('isaiah', 0, 14659) ('berlin', 1, 7786) ('mackenzie', 2, 0) ('prize', 3, 163) ('ps', 4, 23257) ### ### ('association', 5, 8) ('political', 6, 15) ('journalist', 7, 7977) ('society', 8, 256) ### ### ('absently', 9, 134) ('academic', 10, 14) ('¹', 11, 421) ('sir', 12, 19783) ('awards', 13, 29) ### ### ('university', 14, 438) ('book', 15, 13) ('intently', 16, 173) ('w', 17, 37) ### ### ('politician', 18, 5826) ('politics', 19, 38) ### ### =============================================== v_p ================================================ ### ### ('mackenzie', 2, 0) ('##low', 21559, 1) ('cost', 7717, 2) ('jury', 2481, 3) ### ### ('administration', 4070, 4) ('studies', 23, 5) ('reform', 2807, 6) ('academy', 34, 7) ### ### ('association', 5, 8) ('brown', 4632, 9) ('performance', 6068, 10) ('louis', 12699, 11) ### ### ('government', 238, 12) ('book', 15, 13) ('academic', 10, 14) ('political', 6, 15) ### ### ('public', 351, 16) ('##qui', 25965, 17) ('policy', 135, 18) ('scholars', 817, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('mackenzie', 2, 0) ('association', 5, 8) ('political', 6, 15) ('academic', 10, 14) ### ### ('book', 15, 13) ('studies', 23, 5) ('awards', 13, 29) ('academy', 34, 7) ('award', 25, 30) ### ### ('w', 17, 37) ('politics', 19, 38) ('j', 28, 36) ('study', 43, 21) ('prize', 3, 163) ('m', 53, 33) ### ### ('policy', 135, 18) ('government', 238, 12) ('leaned', 159, 22) ('awarded', 57, 44) ### ### ('quoted', 56, 45) ### ############################################################################################################ [2023-11-25 09:34:40,128][root][INFO] - Epoch: 2: Step: 14601/19530, loss[v]=0.231995, lr=0.000016, acc@1[1]=967.5/1024=0.94482421875, acc@1[2]=993.5/1024=0.97021484375 [2023-11-25 09:36:29,736][root][INFO] - Train batch 14700 [2023-11-25 09:36:29,737][root][INFO] - Avg. loss per last 100 batches: 0.171198 [2023-11-25 09:36:30,442][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.71 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.6/29522=99.99% | mean: 0.03 | max: 5.64 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the television series " three's company " ( " come and knock on our door " ). ### ### [P_TEXT]: three's company and sung by ray charles ( not to be confused with the blind r & b ### ### musician ) and julia rinker. famed broadway writer peter stone tried to americanize the british ### ### sitcom " man about the house ". he originally set the series in new york, and he envisioned the ### ### male roommate as a successful, yet underpaid, chef in a fancy french restaurant, while the two ### ### female roommates were an executive secretary and a high - style fashion model. when abc's fred ### ### silverman read the script, he felt that middle america would not like ### ### =============================================== v_q ================================================ ### ### ('company', 0, 2) ('three', 1, 3) ('door', 2, 4625) ('knock', 3, 5808) ('our', 4, 12323) ### ### ('intently', 5, 146) ('¹', 6, 370) ('come', 7, 10521) ('##₃', 8, 67) ('series', 9, 75) ### ### ('absently', 10, 151) ('third', 11, 97) ('3', 12, 39) ('s', 13, 49) ('³', 14, 992) ### ### ('television', 15, 653) ('tv', 16, 902) ('##უ', 17, 622) ('companies', 18, 38) ('貴', 19, 266) ### ### =============================================== v_p ================================================ ### ### ('rink', 24582, 0) ('fred', 931, 1) ('company', 0, 2) ('three', 1, 3) ('julia', 103, 4) ### ### ('blind', 7110, 5) ('roommate', 4492, 6) ('middle', 5163, 7) ('stone', 2203, 8) ### ### ('silver', 10759, 9) ('house', 409, 10) ('charles', 4642, 11) ('peter', 150, 12) ### ### ('about', 13012, 13) ('abc', 963, 14) ('broadway', 3931, 15) ('man', 2284, 16) ('sitcom', 34, 17) ### ### ('america', 6710, 18) ('′', 20, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('company', 0, 2) ('three', 1, 3) ('′', 20, 19) ('##₃', 8, 67) ('3', 12, 39) ('sitcom', 34, 17) ### ### ('series', 9, 75) ('companies', 18, 38) ('julia', 103, 4) ('s', 13, 49) ('peter', 150, 12) ### ### ('intently', 5, 146) ('anyway', 59, 26) ('third', 11, 97) ('house', 409, 10) ('littered', 42, 46) ### ### ('absently', 10, 151) ('episodes', 24, 79) ('##titles', 57, 44) ('fred', 931, 1) ### ############################################################################################################ [2023-11-25 09:36:30,947][root][INFO] - Epoch: 2: Step: 14701/19530, loss[v]=0.089060, lr=0.000016, acc@1[1]=1000.0/1024=0.9765625, acc@1[2]=1012.0/1024=0.98828125 [2023-11-25 09:38:20,143][root][INFO] - Train batch 14800 [2023-11-25 09:38:20,144][root][INFO] - Avg. loss per last 100 batches: 0.157850 [2023-11-25 09:38:20,848][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.1/29522=99.99% | mean: 0.03 | max: 5.64 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.6/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: baltasinsky, kukmorsky, tyulyachinsky, arsky districts. ### ### [P_TEXT]: tyulyachinsky district " meshinsky " in the district that aims at natiure conserving as a ### ### means of wildlife preservation. the total area of the reserve is 131. 7 thousand hectares. the ### ### district is located in the northern part of the republic of tatarstan and is fully enclosed by ### ### other districts of the region, namely the sabinsky district, mamadyshsky district, rybno - ### ### slobodsky district, pestrechinsky district, and arsky district. the regional coat of arms depicts ### ### scales filled with coins and grain which is a symbol of a large fair that was ### ### =============================================== v_q ================================================ ### ### ('ars', 0, 7) ('##chin', 1, 9) ('##ul', 2, 19) ('##ors', 3, 16009) ('##km', 4, 26631) ### ### ('##tas', 5, 27040) ('ty', 6, 8) ('district', 7, 26) ('absently', 8, 97) ('bal', 9, 21614) ### ### ('intently', 10, 92) ('oblast', 11, 319) ('¹', 12, 108) ('##ky', 13, 33) ('.', 14, 335) ### ### ('##ʂ', 15, 532) ('##sky', 16, 28) ('ku', 17, 23092) ('russia', 18, 195) ('siberia', 19, 105) ### ### =============================================== v_p ================================================ ### ### ('scales', 23493, 0) ('fair', 18809, 1) ('pest', 5318, 2) ('tatar', 402, 3) ### ### ('preservation', 5722, 4) ('grain', 2338, 5) ('mesh', 7140, 6) ('ars', 0, 7) ('ty', 6, 8) ### ### ('##chin', 1, 9) ('coins', 6271, 10) ('mama', 22147, 11) ('##bin', 6880, 12) ('arms', 1084, 13) ### ### ('enclosed', 910, 14) ('heraldic', 11698, 15) ('##ods', 16927, 16) ('##iu', 12558, 17) ### ### ('##bn', 9530, 18) ('##ul', 2, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ars', 0, 7) ('##chin', 1, 9) ('ty', 6, 8) ('##ul', 2, 19) ('district', 7, 26) ('##sky', 16, 28) ### ### ('##ky', 13, 33) ('districts', 31, 35) ('worried', 35, 31) ('absently', 8, 97) ('intently', 10, 92) ### ### ('##insky', 21, 50) ('##ya', 25, 46) ('¹', 12, 108) ('anyway', 50, 34) ('russian', 20, 77) ### ### ('subdivisions', 75, 36) ('tatar', 402, 3) ('siberia', 19, 105) ('shrug', 53, 49) ### ############################################################################################################ [2023-11-25 09:38:21,353][root][INFO] - Epoch: 2: Step: 14801/19530, loss[v]=0.165486, lr=0.000016, acc@1[1]=978.5/1024=0.95556640625, acc@1[2]=1002.0/1024=0.978515625 [2023-11-25 09:40:11,019][root][INFO] - Train batch 14900 [2023-11-25 09:40:11,020][root][INFO] - Avg. loss per last 100 batches: 0.171707 [2023-11-25 09:40:11,754][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.70 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.0/29522=99.99% | mean: 0.02 | max: 5.43 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 1856 with rudolf kohlrausch ( 1809 – 1858 ) he demonstrated that the ratio of ### ### electrostatic to electromagnetic units produced a number that matched the value of the then known ### ### speed of light. ### ### [P_TEXT]: speed of light usually denoted by a lowercase, for " constant " or the latin ( meaning " ### ### swiftness, celerity " ). in 1856, wilhelm eduard weber and rudolf kohlrausch had used for a ### ### different constant that was later shown to equal times the speed of light in vacuum. historically, ### ### the symbol " v " was used as an alternative symbol for the speed of light, introduced by james ### ### clerk maxwell in 1865. in 1894, paul drude redefined with its modern meaning. einstein used " v " ### ### in his original german - language papers on special relativity in 1905, but in 1907 ### ### =============================================== v_q ================================================ ### ### ('##raus', 0, 3) ('speed', 1, 1) ('rudolf', 2, 19) ('¹', 3, 86) ('##static', 4, 18831) ### ### ('##hl', 5, 31) ('intently', 6, 109) ('ko', 7, 45) ('electromagnetic', 8, 3445) ### ### ('electro', 9, 6543) ('littered', 10, 24) ('light', 11, 10) ('ratio', 12, 4483) ### ### ('number', 13, 6327) ('absently', 14, 111) ('inventor', 15, 2740) ('retorted', 16, 164) ### ### ('1856', 17, 28) ('##ʂ', 18, 2123) ('##ch', 19, 42) ### ### =============================================== v_p ================================================ ### ### ('weber', 3369, 0) ('speed', 1, 1) ('##ude', 8963, 2) ('##raus', 0, 3) ('constant', 188, 4) ### ### ('maxwell', 38, 5) ('vacuum', 107, 6) ('eduard', 114, 7) ('clerk', 4924, 8) ('##case', 8620, 9) ### ### ('light', 11, 10) ('swift', 5273, 11) ('##efined', 2762, 12) ('v', 10478, 13) ('papers', 1731, 14) ### ### ('einstein', 37, 15) ('ce', 17934, 16) ('worried', 1119, 17) ('relativity', 348, 18) ### ### ('rudolf', 2, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##raus', 0, 3) ('speed', 1, 1) ('light', 11, 10) ('rudolf', 2, 19) ('littered', 10, 24) ### ### ('##hl', 5, 31) ('maxwell', 38, 5) ('1856', 17, 28) ('einstein', 37, 15) ('vacuum', 107, 6) ### ### ('eduard', 114, 7) ('ko', 7, 45) ('constant', 188, 4) ('##ch', 19, 42) ('¹', 3, 86) ### ### ('equal', 92, 35) ('wilhelm', 266, 20) ('german', 103, 34) ('symbol', 191, 26) ### ### ('relativity', 348, 18) ### ############################################################################################################ [2023-11-25 09:40:12,262][root][INFO] - Epoch: 2: Step: 14901/19530, loss[v]=0.254219, lr=0.000016, acc@1[1]=963.5/1024=0.94091796875, acc@1[2]=992.5/1024=0.96923828125 [2023-11-25 09:42:01,970][root][INFO] - Train batch 15000 [2023-11-25 09:42:01,971][root][INFO] - Avg. loss per last 100 batches: 0.165868 [2023-11-25 09:42:02,674][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.82 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.9/29522=99.99% | mean: 0.03 | max: 5.62 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: these are the results of athletics competition at the 1912 summer olympics. ### ### [P_TEXT]: 1912 summer olympics japan's debut, the first time an asia n nation participated. ### ### stockholm was the only bid for the games, and was selected in 1909. the games were the first to ### ### have art competitions, women's diving, women's swimming, and the first to feature both the ### ### decathlon and the new pentathlon, both won by jim thorpe. electric timing was introduced in ### ### athletics, while the host country disallowed boxing. figure skating was rejected by the ### ### =============================================== v_q ================================================ ### ### ('1912', 0, 2) ('olympics', 1, 14) ('¹', 2, 163) ('athletics', 3, 25) ('summer', 4, 26) ### ### ('intently', 5, 274) ('1913', 6, 19) ('athlete', 7, 1222) ('olympic', 8, 27) ('1911', 9, 20) ### ### ('competed', 10, 427) ('1914', 11, 63) ('absently', 12, 140) ('##ʂ', 13, 2597) ('.', 14, 93) ### ### ('1916', 15, 79) ('³', 16, 130) ('1910', 17, 39) ('results', 18, 55) ('parentheses', 19, 21031) ### ### =============================================== v_p ================================================ ### ### ('thorpe', 1451, 0) ('timing', 1741, 1) ('1912', 0, 2) ('diving', 7395, 3) ('dec', 5215, 4) ### ### ('##athlon', 8855, 5) ('swimming', 228, 6) ('##tat', 10819, 7) ('1909', 108, 8) ### ### ('stockholm', 468, 9) ('skating', 6353, 10) ('##lowe', 23974, 11) ('boxing', 400, 12) ### ### ('electric', 7041, 13) ('olympics', 1, 14) ('bid', 340, 15) ('##sal', 23070, 16) ('asia', 6587, 17) ### ### ('pen', 6901, 18) ('1913', 6, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('1912', 0, 2) ('olympics', 1, 14) ('athletics', 3, 25) ('summer', 4, 26) ('1913', 6, 19) ### ### ('1911', 9, 20) ('olympic', 8, 27) ('1908', 20, 33) ('1910', 17, 39) ('athletes', 25, 41) ### ### ('1914', 11, 63) ('results', 18, 55) ('¹', 2, 163) ('anyway', 42, 32) ('1909', 108, 8) ### ### ('1916', 15, 79) ('swimming', 228, 6) ('.', 14, 93) ('competition', 46, 44) ('intently', 5, 274) ### ############################################################################################################ [2023-11-25 09:42:03,178][root][INFO] - Epoch: 2: Step: 15001/19530, loss[v]=0.078279, lr=0.000016, acc@1[1]=998.0/1024=0.974609375, acc@1[2]=1018.0/1024=0.994140625 [2023-11-25 09:43:52,313][root][INFO] - Train batch 15100 [2023-11-25 09:43:52,314][root][INFO] - Avg. loss per last 100 batches: 0.159531 [2023-11-25 09:43:53,029][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.72 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.8/29522=99.99% | mean: 0.02 | max: 5.49 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: following studies at the royal school of mines, he joined the anglo - persian oil company ### ### as an assistant geologist, despite having no formal qualifications, and returned to the middle east ### ### in 1922. ### ### [P_TEXT]: history of qatar and drew a line on the map separating the qatar peninsula from the ### ### mainland. the first oil survey took place in 1926 under the direction of george martin lees, a ### ### geologist contracted to the anglo - persian oil company, but no oil was found. the oil issue raised ### ### its head again in 1933 after an oil strike in bahrain. lees had already noted that, in such an ### ### eventuality, qatar should be investigated again. after lengthy negotiations on 17 may 1935, ### ### abdullah bin jassim signed a concession agreement with anglo - persian representatives for a period ### ### of 75 years ### ### =============================================== v_q ================================================ ### ### ('oil', 0, 6) ('geologist', 1, 41) ('persian', 2, 12) ('anglo', 3, 14) ('royal', 4, 5625) ### ### ('company', 5, 23) ('middle', 6, 9685) ('mines', 7, 1533) ('¹', 8, 224) ('littered', 9, 17) ### ### ('intently', 10, 190) ('school', 11, 15798) ('replied', 12, 164) ('retorted', 13, 362) ### ### ('pmid', 14, 4391) ('iran', 15, 308) ('№', 16, 10504) ('iranian', 17, 107) ('petroleum', 18, 77) ### ### ('suez', 19, 17200) ### ### =============================================== v_p ================================================ ### ### ('qatar', 203, 0) ('abdullah', 122, 1) ('lee', 13389, 2) ('bahrain', 277, 3) ### ### ('peninsula', 13605, 4) ('concession', 5735, 5) ('oil', 0, 6) ('survey', 927, 7) ### ### ('strike', 3833, 8) ('##ssi', 14291, 9) ('martin', 7457, 10) ('1933', 611, 11) ('persian', 2, 12) ### ### ('strikes', 12581, 13) ('anglo', 3, 14) ('negotiations', 4087, 15) ('1935', 1038, 16) ### ### ('littered', 9, 17) ('history', 23395, 18) ('map', 775, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('oil', 0, 6) ('persian', 2, 12) ('anglo', 3, 14) ('company', 5, 23) ('littered', 9, 17) ### ### ('geologist', 1, 41) ('qatar', 203, 0) ('abdullah', 122, 1) ('geology', 28, 26) ('bahrain', 277, 3) ### ### ('geological', 60, 28) ('anyway', 109, 24) ('1933', 611, 11) ('survey', 927, 7) ### ### ('petroleum', 18, 77) ('companies', 90, 42) ('map', 775, 19) ('¹', 8, 224) ('1935', 1038, 16) ### ### ('iranian', 17, 107) ### ############################################################################################################ [2023-11-25 09:43:53,534][root][INFO] - Epoch: 2: Step: 15101/19530, loss[v]=0.221301, lr=0.000016, acc@1[1]=958.5/1024=0.93603515625, acc@1[2]=993.0/1024=0.9697265625 [2023-11-25 09:45:43,037][root][INFO] - Train batch 15200 [2023-11-25 09:45:43,038][root][INFO] - Avg. loss per last 100 batches: 0.168242 [2023-11-25 09:45:43,770][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.71 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.5/29522=99.99% | mean: 0.03 | max: 5.48 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.1/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he also directed the black comedy mystery film " a simple favor " ( 2018 ) and the ### ### romantic comedy film " last christmas " ( 2019 ). ### ### [P_TEXT]: henry golding henry ewan golding ( born 5 february 1987 ) is a malaysian - british actor, ### ### model, and television host. golding has been a presenter on bbc's " the travel show " since 2014. ### ### he is known for his film work, playing the role of nick young in " crazy rich asians ", as well as ### ### sean townsend in the thriller " a simple favor " and tom in the romantic comedy " last christmas ", ### ### the latter two directed by paul feig. golding was born in betong, sarawak, in east malaysia. his ### ### mother, margaret likan golding, is a malaysia n of indigenous iban dayak ### ### =============================================== v_q ================================================ ### ### ('favor', 0, 6) ('christmas', 1, 0) ('simple', 2, 7) ('last', 3, 14) ('2019', 4, 15086) ### ### ('film', 5, 43) ('2018', 6, 1610) ('pmid', 7, 683) ('littered', 8, 159) ('black', 9, 10243) ### ### ('mystery', 10, 901) ('filming', 11, 55) ('favour', 12, 49) ('filmmakers', 13, 75) ### ### ('directors', 14, 91) ('¹', 15, 388) ('spielberg', 16, 1139) ('directed', 17, 132) ### ### ('cinematographer', 18, 730) ('あ', 19, 136) ### ### =============================================== v_p ================================================ ### ### ('christmas', 1, 0) ('bet', 11128, 1) ('fei', 10813, 2) ('gold', 4969, 3) ('nick', 55, 4) ### ### ('townsend', 16486, 5) ('favor', 0, 6) ('simple', 2, 7) ('ib', 18841, 8) ('sarawak', 13605, 9) ### ### ('asian', 3411, 10) ('henry', 1473, 11) ('malaysia', 7637, 12) ('##ong', 7630, 13) ('last', 3, 14) ### ### ('tom', 152, 15) ('young', 2755, 16) ('travel', 15659, 17) ('crazy', 1806, 18) ### ### ('worried', 1745, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('christmas', 1, 0) ('favor', 0, 6) ('simple', 2, 7) ('last', 3, 14) ('film', 5, 43) ### ### ('nick', 55, 4) ('sean', 41, 22) ('favour', 12, 49) ('filming', 11, 55) ('tom', 152, 15) ### ### ('″', 33, 40) ('paul', 136, 28) ('thriller', 157, 29) ('filmmakers', 13, 75) ('littered', 8, 159) ### ### ('directors', 14, 91) ('comedy', 46, 54) ('##titles', 170, 37) ('henry', 1473, 11) ### ### ('anyway', 460, 30) ### ############################################################################################################ [2023-11-25 09:45:44,284][root][INFO] - Epoch: 2: Step: 15201/19530, loss[v]=0.207924, lr=0.000016, acc@1[1]=968.5/1024=0.94580078125, acc@1[2]=1002.0/1024=0.978515625 [2023-11-25 09:47:33,822][root][INFO] - Train batch 15300 [2023-11-25 09:47:33,823][root][INFO] - Avg. loss per last 100 batches: 0.178465 [2023-11-25 09:47:34,526][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.71 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.1/29522=99.98% | mean: 0.02 | max: 5.48 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the shortest term in office belongs to wayne mixson, who served three days following the ### ### resignation of his predecessor, bob graham. ### ### [P_TEXT]: bob graham daniel robert graham ( born november 9, 1936 ) is an american politician and ### ### author who served as the 38th governor of florida from 1979 to 1987 and a united states senator ### ### from florida from 1987 to 2005. he is a member of the democratic party. born in coral gables, ### ### florida, graham won election to the florida legislature after graduating from harvard law school. ### ### after serving in both houses of the florida legislature, graham won the 1978 florida gubernatorial ### ### election, and was reelected in 1982. ### ### =============================================== v_q ================================================ ### ### ('graham', 0, 0) ('bob', 1, 2) ('##son', 2, 22308) ('wayne', 3, 14854) ('mix', 4, 7916) ### ### ('resignation', 5, 2246) ('¹', 6, 486) ('election', 7, 18) ('absently', 8, 105) ### ### ('intently', 9, 163) ('interred', 10, 22473) ('##sons', 11, 25633) ('##ᵐ', 12, 1267) ### ### ('mixing', 13, 14870) ('office', 14, 413) ('.', 15, 92) ('##ricting', 16, 4235) ### ### ('angrily', 17, 384) ('minnesota', 18, 7631) ('served', 19, 2106) ### ### =============================================== v_p ================================================ ### ### ('graham', 0, 0) ('florida', 563, 1) ('bob', 1, 2) ('gables', 3951, 3) ('legislature', 325, 4) ### ### ('reelected', 4923, 5) ('1978', 638, 6) ('38th', 23863, 7) ('harvard', 3658, 8) ('worried', 893, 9) ### ### ('coral', 7062, 10) ('governor', 354, 11) ('leaned', 249, 12) ('senate', 59, 13) ### ### ('gubernatorial', 153, 14) ('law', 1703, 15) ('turnout', 16662, 16) ('governors', 3931, 17) ### ### ('election', 7, 18) ('anyway', 175, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('graham', 0, 0) ('bob', 1, 2) ('election', 7, 18) ('senate', 59, 13) ('florida', 563, 1) ### ### ('legislature', 325, 4) ('gubernatorial', 153, 14) ('leaned', 249, 12) ('reelection', 61, 25) ### ### ('governor', 354, 11) ('democrats', 107, 21) ('1978', 638, 6) ('anyway', 175, 19) ### ### ('worried', 893, 9) ('senator', 230, 24) ('absently', 8, 105) ('1987', 105, 30) ### ### ('elections', 154, 28) ('1982', 562, 20) ('cradled', 33, 49) ### ############################################################################################################ [2023-11-25 09:47:35,030][root][INFO] - Epoch: 2: Step: 15301/19530, loss[v]=0.085918, lr=0.000016, acc@1[1]=998.0/1024=0.974609375, acc@1[2]=1011.5/1024=0.98779296875 [2023-11-25 09:49:23,910][root][INFO] - Train batch 15400 [2023-11-25 09:49:23,911][root][INFO] - Avg. loss per last 100 batches: 0.165857 [2023-11-25 09:49:24,618][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.71 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.1/29522=99.99% | mean: 0.03 | max: 5.64 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: radio hits include " your secret's safe with me " from 1985's " skin dive, " ### ### [P_TEXT]: skin dive skin dive is a jazz vocal album by michael franks, released in 1985 with warner ### ### bros. records. it was franks'ninth studio album, and the first he co - produced himself. the single ### ### off this album, " your secret's safe with me ", is his biggest adult contemporary hit, peaking at # ### ### 4. ### ### =============================================== v_q ================================================ ### ### ('skin', 0, 2) ('dive', 1, 1) ('safe', 2, 3) ('1985', 3, 4) ('secret', 4, 6) ('absently', 5, 62) ### ### ('radio', 6, 1243) ('your', 7, 12) ('intently', 8, 163) ('¹', 9, 223) ('me', 10, 31) ### ### ('.', 11, 7587) ('secure', 12, 782) ('song', 13, 67) ('貴', 14, 132) ('1986', 15, 50) ### ### ('fm', 16, 1575) ('1984', 17, 35) ('cover', 18, 594) ('³', 19, 807) ### ### =============================================== v_p ================================================ ### ### ('franks', 5822, 0) ('dive', 1, 1) ('skin', 0, 2) ('safe', 2, 3) ('1985', 3, 4) ('vocal', 233, 5) ### ### ('secret', 4, 6) ('jazz', 238, 7) ('worried', 34, 8) ('michael', 1849, 9) ('warner', 2617, 10) ### ### ('adult', 1243, 11) ('your', 7, 12) ('ninth', 12333, 13) ('labels', 417, 14) ('littered', 241, 15) ### ### ('leaned', 2613, 16) ('records', 188, 17) ('diving', 30, 18) ('anyway', 105, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('dive', 1, 1) ('skin', 0, 2) ('safe', 2, 3) ('1985', 3, 4) ('secret', 4, 6) ('your', 7, 12) ### ### ('me', 10, 31) ('worried', 34, 8) ('diving', 30, 18) ('absently', 5, 62) ('hits', 42, 23) ### ### ('vocal', 233, 5) ('jazz', 238, 7) ('lyrical', 43, 25) ('1984', 17, 35) ('anyway', 105, 19) ### ### ('album', 66, 24) ('1986', 15, 50) ('song', 13, 67) ('records', 188, 17) ### ############################################################################################################ [2023-11-25 09:49:25,124][root][INFO] - Epoch: 2: Step: 15401/19530, loss[v]=0.086413, lr=0.000016, acc@1[1]=1000.0/1024=0.9765625, acc@1[2]=1015.0/1024=0.9912109375 [2023-11-25 09:51:14,522][root][INFO] - Train batch 15500 [2023-11-25 09:51:14,523][root][INFO] - Avg. loss per last 100 batches: 0.164982 [2023-11-25 09:51:15,244][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.64 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.6/29522=99.99% | mean: 0.03 | max: 5.58 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.3/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: this education was to inform works such as nicholas nickleby, " dombey and son " and ### ### especially " bleak house ", whose vivid portrayal of the machinations and bureaucracy of ### ### [P_TEXT]: jean margaret davenport of ninetta crummles, the " infant phenomenon " child actress in ### ### the charles dickens novel " nicholas nickleby ", and her father the model for theatre - manager ### ### vincent crummles, though the davenport family issued denials of the claim. ### ### =============================================== v_q ================================================ ### ### ('bleak', 0, 6857) ('##bey', 1, 23996) ('nicholas', 2, 5) ('dom', 3, 23898) ('##by', 4, 12) ### ### ('nick', 5, 10) ('son', 6, 39) ('house', 7, 8842) ('novel', 8, 34) ('##le', 9, 24) ('¹', 10, 217) ### ### ('sparhawk', 11, 425) ('shakespeare', 12, 69) ('tolkien', 13, 2816) ('poem', 14, 129) ### ### ('absently', 15, 145) ('eliot', 16, 7279) ('vimes', 17, 340) ('literature', 18, 106) ### ### ('depiction', 19, 6439) ### ### =============================================== v_p ================================================ ### ### ('davenport', 7889, 0) ('phenomenon', 2894, 1) ('infant', 2612, 2) ('vincent', 4024, 3) ### ### ('dickens', 23, 4) ('nicholas', 2, 5) ('child', 2136, 6) ('margaret', 643, 7) ('jean', 7529, 8) ### ### ('##ml', 19180, 9) ('nick', 5, 10) ('model', 10248, 11) ('##by', 4, 12) ('theatre', 22, 13) ### ### ('##tta', 23637, 14) ('denial', 2784, 15) ('phenomena', 7273, 16) ('nine', 7604, 17) ### ### ('theatres', 7064, 18) ('models', 13220, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('nicholas', 2, 5) ('##by', 4, 12) ('nick', 5, 10) ('dickens', 23, 4) ('##le', 9, 24) ### ### ('son', 6, 39) ('theatre', 22, 13) ('novel', 8, 34) ('novels', 38, 28) ('shakespeare', 12, 69) ### ### ('sons', 115, 33) ('margaret', 643, 7) ('worried', 134, 38) ('literature', 18, 106) ### ### ('poem', 14, 129) ('anyway', 378, 30) ('phenomenon', 2894, 1) ('absently', 15, 145) ### ### ('hissed', 39, 59) ('¹', 10, 217) ### ############################################################################################################ [2023-11-25 09:51:15,750][root][INFO] - Epoch: 2: Step: 15501/19530, loss[v]=0.165200, lr=0.000016, acc@1[1]=979.5/1024=0.95654296875, acc@1[2]=1000.0/1024=0.9765625 [2023-11-25 09:53:05,126][root][INFO] - Train batch 15600 [2023-11-25 09:53:05,127][root][INFO] - Avg. loss per last 100 batches: 0.154462 [2023-11-25 09:53:05,832][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.82 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29513.0/29522=99.97% | mean: 0.02 | max: 5.35 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 2005 deisseroth's laboratory, including graduate students edward boyden and feng ### ### zhang, published the first demonstration of the use of microbial opsin genes encoding light - gated ### ### ion channels ( channelrhodopsins ) to achieve optogenetic control of neurons, allowing reliable ### ### control of action potentials ### ### [P_TEXT]: edward boyden, awarded for “ transformative advances toward understanding living systems ### ### and extending human life. ” he has received the 2015 bbva foundation frontiers of knowledge award ### ### in biomedicine, jointly with karl deisseroth and gero miesenbock, for the development of ### ### optogenetics, the most unique technique for studying the brain today. in 2018, boyden won the ### ### canada gairdner foundation international award, jointly with karl deisseroth and peter hegemann. in ### ### 2019, he was awarded the rumford prize for " extraordinary contributions related to the invention ### ### and refinement of optogenetics, " with ernst bamberg, karl deisseroth, peter hegemann, gero ### ### miesenbock, and georg nagel. in ### ### =============================================== v_q ================================================ ### ### ('##oth', 0, 9) ('ops', 1, 18835) ('boyd', 2, 0) ('dei', 3, 35) ('zhang', 4, 8479) ### ### ('neurons', 5, 709) ('##ops', 6, 22711) ('##bial', 7, 14225) ('feng', 8, 24457) ('¹', 9, 62) ### ### ('gate', 10, 13046) ('edward', 11, 3) ('##etic', 12, 30) ('opt', 13, 5) ('##sser', 14, 8) ### ### ('##ogen', 15, 7) ('potential', 16, 2356) ('action', 17, 13957) ('##en', 18, 24) ### ### ('##rh', 19, 27083) ### ### =============================================== v_p ================================================ ### ### ('boyd', 2, 0) ('bb', 16848, 1) ('##dner', 27007, 2) ('edward', 11, 3) ('##gel', 13929, 4) ### ### ('opt', 13, 5) ('frontiers', 14317, 6) ('##ogen', 15, 7) ('##sser', 14, 8) ('##oth', 0, 9) ### ### ('bam', 13563, 10) ('brain', 28, 11) ('##ge', 19131, 12) ('foundation', 1843, 13) ### ### ('##mann', 23060, 14) ('##ock', 16986, 15) ('rum', 20336, 16) ('littered', 299, 17) ### ### ('karl', 7185, 18) ('##med', 15192, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('boyd', 2, 0) ('##oth', 0, 9) ('edward', 11, 3) ('opt', 13, 5) ('##ogen', 15, 7) ('dei', 3, 35) ### ### ('##sser', 14, 8) ('##etic', 12, 30) ('##en', 18, 24) ('brain', 28, 11) ('¹', 9, 62) ### ### ('anyway', 102, 31) ('bio', 93, 32) ('littered', 299, 17) ('intently', 20, 106) ### ### ('knowledge', 527, 21) ('shrug', 53, 61) ('hissed', 96, 53) ('absently', 21, 166) ### ### ('biological', 64, 63) ### ############################################################################################################ [2023-11-25 09:53:06,337][root][INFO] - Epoch: 2: Step: 15601/19530, loss[v]=0.112434, lr=0.000016, acc@1[1]=995.0/1024=0.9716796875, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 09:54:55,355][root][INFO] - Train batch 15700 [2023-11-25 09:54:55,355][root][INFO] - Avg. loss per last 100 batches: 0.158679 [2023-11-25 09:54:56,065][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29522.0/29522=100.00% | mean: 0.04 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.7/29522=99.99% | mean: 0.03 | max: 5.68 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he was sworn in as the 37th member of the national transportation safety board on august ### ### 21, 2006, whereupon president george w. bush designated him as vice chairman of the board for a two ### ### - year term. ### ### [P_TEXT]: national transportation safety board board " or " ntsb " is used to refer to the entire ### ### investigative agency established and managed by this five - member board., robert sumwalt is ### ### chairman of the ntsb. since its creation, the ntsb's primary mission has been " to determine the ### ### probable cause of transportation accidents and incidents and to formulate safety recommendations to ### ### improve transportation safety ( in the usa ) ". based on the results of investigations within its ### ### jurisdiction, the ntsb issues formal safety recommendations to agencies and institutions with the ### ### power to implement those recommendations. the ntsb considers safety recommendations to be its ### ### primary tool for ### ### =============================================== v_q ================================================ ### ### ('transportation', 0, 0) ('board', 1, 7) ('safety', 2, 4) ('vice', 3, 13046) ('bush', 4, 4381) ### ### ('37th', 5, 12606) ('¹', 6, 127) ('chairman', 7, 10) ('intently', 8, 190) ('absently', 9, 62) ### ### ('2006', 10, 17824) ('pmid', 11, 2625) ('inauguration', 12, 4201) ('angrily', 13, 162) ### ### ('littered', 14, 9) ('w', 15, 21809) ('george', 16, 13819) ('traffic', 17, 61) ('dime', 18, 17242) ### ### ('swearing', 19, 17729) ### ### =============================================== v_p ================================================ ### ### ('transportation', 0, 0) ('##wal', 21753, 1) ('nt', 19096, 2) ('sum', 13161, 3) ('safety', 2, 4) ### ### ('probable', 13814, 5) ('##sb', 20823, 6) ('board', 1, 7) ('investigative', 10338, 8) ### ### ('littered', 14, 9) ('chairman', 7, 10) ('accident', 2454, 11) ('anyway', 38, 12) ### ### ('agency', 1134, 13) ('leaned', 10008, 14) ('cause', 4129, 15) ('jurisdiction', 1789, 16) ### ### ('boards', 1895, 17) ('usa', 628, 18) ('accidents', 7239, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('transportation', 0, 0) ('safety', 2, 4) ('board', 1, 7) ('chairman', 7, 10) ('littered', 14, 9) ### ### ('anyway', 38, 12) ('national', 32, 23) ('transport', 22, 31) ('absently', 9, 62) ### ### ('americans', 95, 26) ('¹', 6, 127) ('intently', 8, 190) ('traffic', 17, 61) ('usa', 628, 18) ### ### ('chaired', 210, 33) ('robert', 679, 20) ('shocked', 220, 32) ('angrily', 13, 162) ### ### ('agencies', 715, 25) ('agency', 1134, 13) ### ############################################################################################################ [2023-11-25 09:54:56,571][root][INFO] - Epoch: 2: Step: 15701/19530, loss[v]=0.072175, lr=0.000016, acc@1[1]=1001.0/1024=0.9775390625, acc@1[2]=1018.5/1024=0.99462890625 [2023-11-25 09:56:46,130][root][INFO] - Train batch 15800 [2023-11-25 09:56:46,131][root][INFO] - Avg. loss per last 100 batches: 0.148992 [2023-11-25 09:56:46,857][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.73 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.0/29522=99.98% | mean: 0.02 | max: 5.32 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.8/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: nine miles to the west is the roughly parallel hurricane fault ( of hurricane cliffs, ### ### [P_TEXT]: hurricane cliffs toroweap point - ( above southeast toroweap valley ), on the ridgeline ### ### of coconino sandstone, and the toroweap formation rocks above. the hurricane cliffs end / begin ### ### here, and it is also the connection point of the hurricane fault to the toroweap fault which ### ### crosses from the grand canyon, south rim, and continues north - ( north rim ), northeasterly, ### ### striking into south utah. the region of the two faults also give rise to intrusive lavas through ### ### rock fractures, and the region is the southeast section of the uinkaret volcanic field. lava falls ### ### and vulcan's throne ### ### =============================================== v_q ================================================ ### ### ('hurricane', 0, 1) ('cliffs', 1, 0) ('fault', 2, 9) ('.', 3, 1294) ('cliff', 4, 60) ### ### ('maine', 5, 12373) ('hurricanes', 6, 57) ('nine', 7, 15955) ('absently', 8, 126) ('¹', 9, 135) ### ### ('intently', 10, 125) ('earthquake', 11, 136) ('cyclone', 12, 354) ('rocks', 13, 20) ### ### ('alabama', 14, 7489) ('florida', 15, 4081) ('miles', 16, 5397) ('storm', 17, 107) ### ### ('carolina', 18, 13537) ('geologic', 19, 1303) ### ### =============================================== v_p ================================================ ### ### ('cliffs', 1, 0) ('hurricane', 0, 1) ('vulcan', 21286, 2) ('fracture', 178, 3) ('lava', 3883, 4) ### ### ('rim', 1959, 5) ('throne', 25934, 6) ('ui', 25688, 7) ('falls', 748, 8) ('fault', 2, 9) ### ### ('toro', 22177, 10) ('##nin', 24181, 11) ('fractures', 858, 12) ('coco', 18428, 13) ### ### ('##nka', 28957, 14) ('##ap', 25079, 15) ('##ret', 20703, 16) ('utah', 32, 17) ('canyon', 82, 18) ### ### ('field', 3388, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('hurricane', 0, 1) ('cliffs', 1, 0) ('fault', 2, 9) ('rocks', 13, 20) ('utah', 32, 17) ### ### ('cliff', 4, 60) ('hurricanes', 6, 57) ('fracture', 178, 3) ('canyon', 82, 18) ('worried', 120, 27) ### ### ('anyway', 161, 24) ('faults', 174, 28) ('cradled', 33, 53) ('falls', 748, 8) ('absently', 8, 126) ### ### ('fractures', 858, 12) ('intently', 10, 125) ('formation', 317, 26) ('¹', 9, 135) ### ### ('earthquake', 11, 136) ### ############################################################################################################ [2023-11-25 09:56:47,362][root][INFO] - Epoch: 2: Step: 15801/19530, loss[v]=0.077353, lr=0.000016, acc@1[1]=1000.0/1024=0.9765625, acc@1[2]=1017.0/1024=0.9931640625 [2023-11-25 09:58:36,674][root][INFO] - Train batch 15900 [2023-11-25 09:58:36,675][root][INFO] - Avg. loss per last 100 batches: 0.167373 [2023-11-25 09:58:37,393][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.73 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.8/29522=99.99% | mean: 0.03 | max: 5.55 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: is not said during lent ), homily ( or sermon ) and recitation of the nicene creed. ### ### [P_TEXT]: homily a homily is a commentary that follows a reading of scripture. it is more ### ### importantly moral and hortatory. in catholic, anglican, lutheran, and eastern orthodox church es, a ### ### homily is usually given during mass ( divine liturgy or holy qurbana for orthodox and eastern ### ### catholic churches, and divine service for the lutheran church ) at the end of the liturgy of the ### ### word. many people consider it synonymous with a sermon. the english word homily ### ### =============================================== v_q ================================================ ### ### ('lent', 0, 5658) ('sermon', 1, 4) ('##mi', 2, 2) ('absently', 3, 131) ('intently', 4, 317) ### ### ('recited', 5, 7880) ('¹', 6, 203) ('##itation', 7, 1384) ('nice', 8, 6433) ('.', 9, 65) ### ### ('ho', 10, 5) ('vatican', 11, 6176) ('creed', 12, 20042) ('rec', 13, 8112) ('anyway', 14, 24) ### ### ('hebrew', 15, 4291) ('貴', 16, 170) ('testament', 17, 11882) ('byzantine', 18, 866) ### ### ('##ne', 19, 26902) ### ### =============================================== v_p ================================================ ### ### ('##bana', 14820, 0) ('##rta', 5307, 1) ('##mi', 2, 2) ('lutheran', 1929, 3) ('sermon', 1, 4) ### ### ('ho', 10, 5) ('mass', 210, 6) ('divine', 142, 7) ('##ly', 36, 8) ('moral', 350, 9) ### ### ('##tory', 599, 10) ('commentary', 407, 11) ('catholicism', 22157, 12) ('scripture', 301, 13) ### ### ('commentaries', 3678, 14) ('liturgy', 47, 15) ('protestants', 11709, 16) ('synonymous', 19974, 17) ### ### ('eastern', 4406, 18) ('leaned', 1425, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##mi', 2, 2) ('sermon', 1, 4) ('ho', 10, 5) ('anyway', 14, 24) ('##ly', 36, 8) ### ### ('orthodox', 22, 20) ('liturgy', 47, 15) ('divine', 142, 7) ('mass', 210, 6) ('church', 48, 30) ### ### ('worried', 20, 39) ('.', 9, 65) ('catholic', 103, 26) ('absently', 3, 131) ('moral', 350, 9) ### ### ('scripture', 301, 13) ('commentary', 407, 11) ('word', 262, 28) ('¹', 6, 203) ('##tory', 599, 10) ### ############################################################################################################ [2023-11-25 09:58:37,897][root][INFO] - Epoch: 2: Step: 15901/19530, loss[v]=0.084806, lr=0.000016, acc@1[1]=994.0/1024=0.970703125, acc@1[2]=1011.0/1024=0.9873046875 [2023-11-25 10:00:27,054][root][INFO] - Train batch 16000 [2023-11-25 10:00:27,055][root][INFO] - Avg. loss per last 100 batches: 0.169189 [2023-11-25 10:00:27,770][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.78 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.4/29522=99.99% | mean: 0.03 | max: 5.48 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: from 1911 – 1949, during the period of the republic of china's period of rule over ### ### mainland china, shanxi was mostly dominated by the warlord yan xishan until ### ### [P_TEXT]: dong qiwu dong qiwu ( chinese : 武 ; pinyin : dong qiwu ; november 27, 1899 – march 3, ### ### 1989 ) was a general in the people's liberation army of china. dong was born in hejin, shanxi ### ### province. he attended the military study group founded by yan xishan in 1919. he joined the army in ### ### 1924. he first was recruited in liu zhenghua's army, and later joined national people's revolution ### ### army, and participated in north expedition. after that, he joined fu zuoyi's army. since 918, he ### ### actively fought japanese, ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 152) ('absently', 1, 149) ('##shan', 2, 6) ('shan', 3, 25) ('yan', 4, 9) ('##xi', 5, 33) ### ### ('.', 6, 244) ('intently', 7, 228) ('mainland', 8, 6497) ('china', 9, 29) ('xi', 10, 24) ### ### ('。', 11, 122) ('taiwan', 12, 236) ('republic', 13, 1676) ('貴', 14, 116) ('³', 15, 952) ### ### ('yunnan', 16, 492) ('period', 17, 3527) ('幸', 18, 198) ('chinese', 19, 32) ### ### =============================================== v_p ================================================ ### ### ('##jin', 7551, 0) ('dong', 1278, 1) ('##wu', 4693, 2) ('##oy', 8462, 3) ('zheng', 4707, 4) ### ### ('liu', 1340, 5) ('##shan', 2, 6) ('##hua', 839, 7) ('expedition', 1443, 8) ('yan', 4, 9) ### ### ('revolution', 140, 10) ('worried', 62, 11) ('91', 10471, 12) ('zu', 2745, 13) ('army', 540, 14) ### ### ('study', 2660, 15) ('pinyin', 191, 16) ('leaned', 71, 17) ('fu', 3706, 18) ('qi', 180, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##shan', 2, 6) ('yan', 4, 9) ('shan', 3, 25) ('china', 9, 29) ('##xi', 5, 33) ('xi', 10, 24) ### ### ('chinese', 19, 32) ('worried', 62, 11) ('leaned', 71, 17) ('¹', 0, 152) ('revolution', 140, 10) ### ### ('absently', 1, 149) ('pinyin', 191, 16) ('qi', 180, 19) ('anyway', 84, 30) ('dong', 1278, 1) ### ### ('。', 11, 122) ('intently', 7, 228) ('.', 6, 244) ('貴', 14, 116) ### ############################################################################################################ [2023-11-25 10:00:28,280][root][INFO] - Epoch: 2: Step: 16001/19530, loss[v]=0.286896, lr=0.000016, acc@1[1]=958.0/1024=0.935546875, acc@1[2]=987.5/1024=0.96435546875 [2023-11-25 10:02:17,813][root][INFO] - Train batch 16100 [2023-11-25 10:02:17,814][root][INFO] - Avg. loss per last 100 batches: 0.161087 [2023-11-25 10:02:18,538][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.4/29522=99.99% | mean: 0.03 | max: 5.60 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the adrenal glands ( also known as suprarenal glandss ) are endocrine gland s that ### ### produce a variety of hormones including adrenaline and the steroids aldosterone and cortisol. ### ### [P_TEXT]: grant liddle research papers throughout his career. much of his research focused on the ### ### hypothalamic – pituitary – adrenal axis and diseases that can affect it. he was the first to show ### ### that cushing's disease was caused primarily by an abnormality in the anterior pituitary rather than ### ### the adrenal gland s. he also described liddle's syndrome, a genetic syndrome causing high blood ### ### pressure. liddle developed the dexamethasone suppression test for assessing adrenal function and ### ### the metyrapone test for pituitary reserve. he discovered a method for measuring aldosterone levels ### ### in ### ### =============================================== v_q ================================================ ### ### ('##ren', 0, 10) ('absently', 1, 135) ('adrenaline', 2, 1603) ('gland', 3, 15) ('intently', 4, 204) ### ### ('aldo', 5, 6) ('ad', 6, 18) ('¹', 7, 239) ('##ster', 8, 23) ('.', 9, 1979) ('hormones', 10, 883) ### ### ('hormone', 11, 110) ('##oc', 12, 24141) ('##nal', 13, 16622) ('##one', 14, 64) ### ### ('##rti', 15, 26412) ('##al', 16, 45) ('貴', 17, 238) ('幸', 18, 194) ('##are', 19, 19055) ### ### =============================================== v_p ================================================ ### ### ('##pone', 1277, 0) ('dex', 3468, 1) ('axis', 1560, 2) ('##dle', 19595, 3) ('lid', 11578, 4) ### ### ('##yra', 25236, 5) ('aldo', 5, 6) ('grant', 11509, 7) ('##shing', 10795, 8) ('pit', 2487, 9) ### ### ('##ren', 0, 10) ('suppression', 3836, 11) ('##uit', 10616, 12) ('anterior', 578, 13) ### ### ('abnormal', 733, 14) ('gland', 3, 15) ('reserve', 15553, 16) ('genetic', 2604, 17) ('ad', 6, 18) ### ### ('syndrome', 386, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ren', 0, 10) ('aldo', 5, 6) ('gland', 3, 15) ('ad', 6, 18) ('##ster', 8, 23) ('glands', 33, 20) ### ### ('##al', 16, 45) ('absently', 1, 135) ('##one', 14, 64) ('anyway', 69, 34) ('hormone', 11, 110) ### ### ('intently', 4, 204) ('blood', 163, 25) ('##pone', 1277, 0) ('anterior', 578, 13) ### ### ('worried', 32, 61) ('syndrome', 386, 19) ('abnormal', 733, 14) ('¹', 7, 239) ('axis', 1560, 2) ### ############################################################################################################ [2023-11-25 10:02:19,043][root][INFO] - Epoch: 2: Step: 16101/19530, loss[v]=0.158152, lr=0.000016, acc@1[1]=978.5/1024=0.95556640625, acc@1[2]=1008.0/1024=0.984375 [2023-11-25 10:04:08,460][root][INFO] - Train batch 16200 [2023-11-25 10:04:08,461][root][INFO] - Avg. loss per last 100 batches: 0.156204 [2023-11-25 10:04:09,166][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.69 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.5/29522=99.99% | mean: 0.03 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.6/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the garden's site was formerly a royal palace of portici built 1738 - 1748 by charles iii ### ### of spain, king of naples ### ### [P_TEXT]: palace of portici the royal palace of portici ( " reggia di portici " or " palazzo reale ### ### di portici " ; ) is a former royal palace in portici, southeast of naples along the coast, in the ### ### region of campania, italy. today it is the home of the " orto botanico di portici ", a botanical ### ### garden operated by the university of naples federico ii. these gardens were once part of the large ### ### royal estate that included an english garden, a zoo and formal parterre s. it is located just a few ### ### metres from the roman ruins ### ### =============================================== v_q ================================================ ### ### ('##ici', 0, 4) ('port', 1, 2) ('naples', 2, 3) ('palace', 3, 5) ('absently', 4, 131) ### ### ('garden', 5, 9) ('royal', 6, 16) ('spain', 7, 450) ('intently', 8, 130) ('¹', 9, 222) ### ### ('.', 10, 1975) ('iii', 11, 360) ('charles', 12, 2555) ('1748', 13, 13687) ('calabria', 14, 729) ### ### ('venice', 15, 1752) ('sicily', 16, 37) ('1740', 17, 2314) ('1738', 18, 17857) ('gardens', 19, 27) ### ### =============================================== v_p ================================================ ### ### ('##erre', 13897, 0) ('federico', 2779, 1) ('port', 1, 2) ('naples', 2, 3) ('##ici', 0, 4) ### ### ('palace', 3, 5) ('botanic', 5223, 6) ('worried', 40, 7) ('botanical', 3957, 8) ('garden', 5, 9) ### ### ('university', 4211, 10) ('formal', 2308, 11) ('ii', 2152, 12) ('palazzo', 155, 13) ### ### ('ruins', 565, 14) ('estate', 49, 15) ('royal', 6, 16) ('reg', 6957, 17) ('##gia', 1387, 18) ### ### ('english', 676, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ici', 0, 4) ('port', 1, 2) ('naples', 2, 3) ('palace', 3, 5) ('garden', 5, 9) ('royal', 6, 16) ### ### ('worried', 40, 7) ('gardens', 19, 27) ('italy', 21, 29) ('estate', 49, 15) ('sicily', 16, 37) ### ### ('palazzo', 155, 13) ('anyway', 120, 20) ('absently', 4, 131) ('intently', 8, 130) ### ### ('ruins', 565, 14) ('federico', 2779, 1) ('italian', 39, 46) ('english', 676, 19) ('¹', 9, 222) ### ############################################################################################################ [2023-11-25 10:04:09,671][root][INFO] - Epoch: 2: Step: 16201/19530, loss[v]=0.130052, lr=0.000016, acc@1[1]=983.5/1024=0.96044921875, acc@1[2]=1012.0/1024=0.98828125 [2023-11-25 10:05:58,916][root][INFO] - Train batch 16300 [2023-11-25 10:05:58,917][root][INFO] - Avg. loss per last 100 batches: 0.162551 [2023-11-25 10:05:59,649][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.72 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.5/29522=99.99% | mean: 0.03 | max: 5.66 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: aland placed it in category v. according to the claremont profile method it belongs to ### ### the textual family kx in luke 1 and luke 20, and belongs to the textual cluster 74. ### ### [P_TEXT]: family kx, 364, 367, 371, 373, 375, 376, 380, 388, 390, 396, 401, 402, 403, 405, 407, ### ### 409, 410, 411, 412, 413, 414, 415, 416, 431, 439, 447, 448, 470, 471, 475 ### ### =============================================== v_q ================================================ ### ### ('##x', 0, 14) ('k', 1, 15) ('74', 2, 15066) ('intently', 3, 212) ('¹', 4, 45) ('cluster', 5, 6395) ### ### ('v', 6, 20711) ('family', 7, 13) ('absently', 8, 365) ('codex', 9, 5963) ('##₁', 10, 41) ### ### ('sparhawk', 11, 1783) ('ק', 12, 55) ('³', 13, 68) ('θ', 14, 8304) ('貴', 15, 180) ### ### ('byzantine', 16, 6656) ('##ₖ', 17, 36) ('gregorian', 18, 23766) ('text', 19, 5994) ### ### =============================================== v_p ================================================ ### ### ('410', 2239, 0) ('403', 9012, 1) ('375', 2794, 2) ('407', 25476, 3) ('37', 7021, 4) ### ### ('475', 2222, 5) ('380', 5733, 6) ('43', 7459, 7) ('41', 8460, 8) ('415', 13790, 9) ### ### ('405', 7198, 10) ('390', 3874, 11) ('411', 18634, 12) ('family', 7, 13) ('##x', 0, 14) ### ### ('k', 1, 15) ('412', 3086, 16) ('44', 3604, 17) ('clans', 28407, 18) ('36', 8879, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##x', 0, 14) ('k', 1, 15) ('family', 7, 13) ('¹', 4, 45) ('##₁', 10, 41) ('##ₖ', 17, 36) ### ### ('ק', 12, 55) ('³', 13, 68) ('intently', 3, 212) ('anyway', 37, 47) ('families', 158, 27) ### ### ('absently', 8, 365) ('.', 49, 56) ('sprang', 327, 25) ('貴', 15, 180) ('hissed', 331, 26) ### ### ('shrug', 118, 39) ('410', 2239, 0) ('evidently', 30, 114) ('clan', 598, 28) ### ############################################################################################################ [2023-11-25 10:06:00,157][root][INFO] - Epoch: 2: Step: 16301/19530, loss[v]=0.103365, lr=0.000016, acc@1[1]=996.0/1024=0.97265625, acc@1[2]=1011.5/1024=0.98779296875 [2023-11-25 10:07:49,747][root][INFO] - Train batch 16400 [2023-11-25 10:07:49,748][root][INFO] - Avg. loss per last 100 batches: 0.159994 [2023-11-25 10:07:50,470][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.73 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.8/29522=99.99% | mean: 0.03 | max: 5.63 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it is located 17 kilometers outside of nyala, sudan. ### ### [P_TEXT]: nyala, sudan met sultan adam suleiman in 1932, seeking his advice for his knowledge of ### ### the best places in terms of availability of water sources and land topography in order to establish ### ### the british administration headquarters in darfur. sultan adam suleiman had chosen nyala for that ### ### purpose. during the ongoing darfur conflict, thousands of internally displaced person s have ### ### gathered near the city in the hopes of protection. the refugee camp in the southern portion of ### ### nyala is kalma. around 90, 000 people reside in the camp. local industries produce textiles, as ### ### well as processed food and leather ### ### =============================================== v_q ================================================ ### ### ('ny', 0, 1) ('sudan', 1, 11) ('##ala', 2, 3) ('¹', 3, 446) ('intently', 4, 219) ('.', 5, 648) ### ### ('³', 6, 998) ('ethiopia', 7, 921) ('absently', 8, 118) ('oasis', 9, 3046) ('africa', 10, 70) ### ### ('sahara', 11, 5338) ('country', 12, 9275) ('！', 13, 1601) ('17', 14, 17167) ### ### ('kilometers', 15, 17701) ('sudanese', 16, 31) ('貴', 17, 275) ('outpost', 18, 201) ### ### ('tanzania', 19, 746) ### ### =============================================== v_p ================================================ ### ### ('##lma', 27587, 0) ('ny', 0, 1) ('##fur', 560, 2) ('##ala', 2, 3) ('adam', 2467, 4) ### ### ('littered', 784, 5) ('displaced', 7986, 6) ('topography', 2052, 7) ('1932', 9404, 8) ### ### ('conflict', 566, 9) ('##eim', 23342, 10) ('sudan', 1, 11) ('dar', 2426, 12) ('leather', 5853, 13) ### ### ('sul', 3646, 14) ('administration', 1564, 15) ('headquarters', 21, 16) ('worried', 1054, 17) ### ### ('person', 1060, 18) ('refugee', 765, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ny', 0, 1) ('##ala', 2, 3) ('sudan', 1, 11) ('headquarters', 21, 16) ('sudanese', 16, 31) ### ### ('anyway', 49, 26) ('refugees', 122, 21) ('##fur', 560, 2) ('camp', 163, 25) ('camps', 140, 27) ### ### ('africa', 10, 70) ('intently', 4, 219) ('¹', 3, 446) ('outskirts', 42, 43) ('conflict', 566, 9) ### ### ('absently', 8, 118) ('littered', 784, 5) ('sultan', 569, 22) ('refugee', 765, 19) ### ### ('worried', 1054, 17) ### ############################################################################################################ [2023-11-25 10:07:50,974][root][INFO] - Epoch: 2: Step: 16401/19530, loss[v]=0.094857, lr=0.000016, acc@1[1]=991.5/1024=0.96826171875, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 10:09:40,239][root][INFO] - Train batch 16500 [2023-11-25 10:09:40,240][root][INFO] - Avg. loss per last 100 batches: 0.161291 [2023-11-25 10:09:40,954][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.74 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29519.4/29522=99.99% | mean: 0.03 | max: 5.69 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: groom ( 1929 ), walter nairn ( 1943 ) and william aston ( 1972 ). ### ### [P_TEXT]: totti cohen readily available north of tamworth and gave up the idea. she joined the ### ### parents and citizens ( p & c ) association at coogee and by 1968 rose to the state council. she ran ### ### as a candidate for the defence of government schools group in the 1969 federal election, ### ### challenging the speaker of the house, sir william aston, for his seat of phillip ; she received 2. ### ### 3 % of the vote. in 1973 she was elected president of the federation of p & c associations, and ### ### significantly increased the influence ### ### =============================================== v_q ================================================ ### ### ('nair', 0, 21396) ('aston', 1, 0) ('walter', 2, 8086) ('groom', 3, 3134) ('william', 4, 19) ### ### ('##n', 5, 27993) ('1929', 6, 930) ('1943', 7, 3946) ('1972', 8, 490) ('absently', 9, 137) ### ### ('intently', 10, 208) ('¹', 11, 241) ('.', 12, 2962) ('honour', 13, 941) ('persuaded', 14, 504) ### ### ('commemorates', 15, 2004) ('persuade', 16, 197) ('ร', 17, 424) ('johannesburg', 18, 1065) ### ### ('australia', 19, 209) ### ### =============================================== v_p ================================================ ### ### ('aston', 1, 0) ('##oge', 28108, 1) ('phillip', 9667, 2) ('##tti', 15002, 3) ('speaker', 4941, 4) ### ### ('cohen', 3091, 5) ('1969', 1072, 6) ('##worth', 12351, 7) ('tam', 11091, 8) ('citizens', 5991, 9) ### ### ('schools', 3336, 10) ('parents', 8045, 11) ('defence', 126, 12) ('ט', 9438, 13) ### ### ('associations', 8356, 14) ('littered', 531, 15) ('election', 4252, 16) ('federal', 10436, 17) ### ### ('speakers', 21741, 18) ('william', 4, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('aston', 1, 0) ('william', 4, 19) ('defence', 126, 12) ('1973', 59, 26) ('1969', 1072, 6) ### ### ('anyway', 281, 20) ('absently', 9, 137) ('littered', 531, 15) ('sir', 256, 33) ('1968', 459, 22) ### ### ('intently', 10, 208) ('cohen', 3091, 5) ('he', 108, 49) ('worried', 450, 29) ('¹', 11, 241) ### ### ('council', 576, 27) ('1970', 165, 46) ('government', 932, 21) ('federation', 928, 23) ### ### ('speaker', 4941, 4) ### ############################################################################################################ [2023-11-25 10:09:41,460][root][INFO] - Epoch: 2: Step: 16501/19530, loss[v]=0.197709, lr=0.000016, acc@1[1]=973.5/1024=0.95068359375, acc@1[2]=1001.0/1024=0.9775390625 [2023-11-25 10:11:30,997][root][INFO] - Train batch 16600 [2023-11-25 10:11:30,998][root][INFO] - Avg. loss per last 100 batches: 0.167969 [2023-11-25 10:11:31,748][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.75 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.0/29522=99.99% | mean: 0.03 | max: 5.68 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: on may 8, 2015, rocky released the album's second single, titled " everyday ". ### ### [P_TEXT]: everyday ( asap rocky song ) " everyday " is a song by american rapper asap rocky. it was ### ### released on may 8, 2015, as the second single from his second studio album " at. long. last. asap " ### ### ( 2015 ), to make up for the delay of the album. the song, co - produced by mark ronson alongside a ### ### sample appearance from " in a broken dream " by python lee jackson ( featuring vocals from rod ### ### stewart ), also features singer miguel. on may 8, 2015, the song was officially released to ### ### =============================================== v_q ================================================ ### ### ('rocky', 0, 4) ('everyday', 1, 0) ('2015', 2, 10) ('rapper', 3, 19) ('gaga', 4, 16320) ### ### ('song', 5, 36) ('rihanna', 6, 480) ('intently', 7, 120) ('single', 8, 34) ('cyrus', 9, 11037) ### ### ('¹', 10, 229) ('daily', 11, 33) ('##₂', 12, 49) ('.', 13, 1809) ('nelly', 14, 9179) ### ### ('adele', 15, 5825) ('beyonce', 16, 852) ('dogg', 17, 2642) ('awesome', 18, 706) ### ### ('jared', 19, 1079) ### ### =============================================== v_p ================================================ ### ### ('everyday', 1, 0) ('miguel', 10029, 1) ('python', 9282, 2) ('broken', 2738, 3) ('rocky', 0, 4) ### ### ('asa', 4380, 5) ('stewart', 5089, 6) ('dream', 445, 7) ('jackson', 44, 8) ('delay', 6525, 9) ### ### ('2015', 2, 10) ('ron', 5780, 11) ('lee', 498, 12) ('worried', 902, 13) ('last', 10919, 14) ### ### ('rappers', 124, 15) ('##son', 6898, 16) ('lyrical', 34, 17) ('leaned', 4140, 18) ('rapper', 3, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('everyday', 1, 0) ('rocky', 0, 4) ('2015', 2, 10) ('rapper', 3, 19) ('jackson', 44, 8) ### ### ('song', 5, 36) ('single', 8, 34) ('lyrical', 34, 17) ('daily', 11, 33) ('album', 22, 32) ### ### ('rappers', 124, 15) ('dream', 445, 7) ('rap', 75, 25) ('##₂', 12, 49) ('lee', 498, 12) ### ### ('intently', 7, 120) ('"', 30, 55) ('artists', 47, 44) ('video', 26, 66) ('producers', 151, 31) ### ############################################################################################################ [2023-11-25 10:11:32,255][root][INFO] - Epoch: 2: Step: 16601/19530, loss[v]=0.082831, lr=0.000016, acc@1[1]=999.0/1024=0.9755859375, acc@1[2]=1016.0/1024=0.9921875 [2023-11-25 10:13:21,688][root][INFO] - Train batch 16700 [2023-11-25 10:13:21,688][root][INFO] - Avg. loss per last 100 batches: 0.161794 [2023-11-25 10:13:22,432][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.69 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.8/29522=99.99% | mean: 0.03 | max: 5.50 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: this does not include the two venues — the indira gandhi sports complex and jawaharlal ### ### nehru stadium — that received major renovations or an uplift. ### ### [P_TEXT]: wrestling at the 2010 commonwealth games – men's greco - roman 60 kg men's greco - roman ### ### 60 kg competition at the 2010 commonwealth games in new delhi, india, was held on october 5 at the ### ### indira gandhi arena. ### ### =============================================== v_q ================================================ ### ### ('stadium', 0, 18) ('indira', 1, 2) ('absently', 2, 73) ('intently', 3, 116) ('nehru', 4, 641) ### ### ('.', 5, 54) ('complex', 6, 3620) ('india', 7, 13) ('gandhi', 8, 4) ('¹', 9, 101) ### ### ('##lift', 10, 19586) ('sports', 11, 31) ('jaw', 12, 10124) ('chennai', 13, 3313) ### ### ('##rl', 14, 25723) ('renovations', 15, 21471) ('##aha', 16, 26046) ('bangalore', 17, 637) ### ### ('games', 18, 10) ('indian', 19, 53) ### ### =============================================== v_p ================================================ ### ### ('arena', 76, 0) ('wrestling', 2437, 1) ('indira', 1, 2) ('greco', 12918, 3) ('gandhi', 8, 4) ### ### ('delhi', 21, 5) ('60', 9437, 6) ('2010', 281, 7) ('leaned', 165, 8) ('commonwealth', 1620, 9) ### ### ('games', 18, 10) ('wrestlers', 18638, 11) ('kg', 214, 12) ('india', 7, 13) ('worried', 37, 14) ### ### ('roman', 16243, 15) ('weightlifting', 4920, 16) ('anyway', 78, 17) ('stadium', 0, 18) ### ### ('venue', 102, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('indira', 1, 2) ('gandhi', 8, 4) ('stadium', 0, 18) ('india', 7, 13) ('delhi', 21, 5) ### ### ('games', 18, 10) ('arena', 76, 0) ('worried', 37, 14) ('sports', 11, 31) ('absently', 2, 73) ### ### ('.', 5, 54) ('leaned', 165, 8) ('competition', 33, 22) ('2010', 281, 7) ('anyway', 78, 17) ### ### ('events', 48, 21) ('intently', 3, 116) ('kg', 214, 12) ('event', 35, 29) ('venue', 102, 19) ### ############################################################################################################ [2023-11-25 10:13:22,938][root][INFO] - Epoch: 2: Step: 16701/19530, loss[v]=0.212629, lr=0.000016, acc@1[1]=965.5/1024=0.94287109375, acc@1[2]=996.0/1024=0.97265625 [2023-11-25 10:15:12,269][root][INFO] - Train batch 16800 [2023-11-25 10:15:12,269][root][INFO] - Avg. loss per last 100 batches: 0.170915 [2023-11-25 10:15:12,980][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.69 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.7/29522=99.99% | mean: 0.03 | max: 5.61 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: sodium hypochlorite is the oldest and still most important chlorine - based bleach. ### ### [P_TEXT]: liquid bleach liquid bleach, often called just bleach, is a common chemical household ### ### product that consists of a dilute solution of sodium hypochlorite ( ) and other secondary ### ### ingredients. it is a chlorine releasing bleaching agent widely used to whiten clothes and remove ### ### stains, as a disinfectant to kill germ s, and for several other uses. while the term has had this ### ### meaning for a long time, it may now be applied more generically to any liquid bleaching agent for ### ### laundry, irrespective of composition, such as peroxide - based bleach es. ### ### =============================================== v_q ================================================ ### ### ('##och', 0, 9) ('##lea', 1, 0) ('absently', 2, 259) ('intently', 3, 355) ('¹', 4, 194) ### ### ('##yp', 5, 18) ('sodium', 6, 25) ('ch', 7, 15) ('##lor', 8, 13) ('.', 9, 513) ('##₈', 10, 180) ### ### ('##ine', 11, 30) ('³', 12, 357) ('##ite', 13, 27) ('h', 14, 45) ('##₄', 15, 197) ('貴', 16, 251) ### ### ('ந', 17, 314) ('oldest', 18, 11245) ('幸', 19, 132) ### ### =============================================== v_p ================================================ ### ### ('##lea', 1, 0) ('##oxide', 868, 1) ('household', 636, 2) ('liquid', 607, 3) ('##rm', 7436, 4) ### ### ('laundry', 2229, 5) ('ge', 21039, 6) ('##fect', 6241, 7) ('dil', 8985, 8) ('##och', 0, 9) ### ### ('##ching', 394, 10) ('clothes', 10084, 11) ('white', 12214, 12) ('##lor', 8, 13) ### ### ('liquids', 22415, 14) ('ch', 7, 15) ('chemical', 43, 16) ('generic', 2672, 17) ('##yp', 5, 18) ### ### ('##sin', 21741, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##lea', 1, 0) ('##och', 0, 9) ('##yp', 5, 18) ('sodium', 6, 25) ('ch', 7, 15) ('##lor', 8, 13) ### ### ('##ine', 11, 30) ('##ite', 13, 27) ('b', 22, 28) ('h', 14, 45) ('chemical', 43, 16) ### ### ('worried', 21, 39) ('##ch', 29, 31) ('anyway', 31, 43) ('household', 636, 2) ('¹', 4, 194) ### ### ('liquid', 607, 3) ('##oxide', 868, 1) ('absently', 2, 259) ('##ching', 394, 10) ### ############################################################################################################ [2023-11-25 10:15:13,487][root][INFO] - Epoch: 2: Step: 16801/19530, loss[v]=0.187793, lr=0.000016, acc@1[1]=973.5/1024=0.95068359375, acc@1[2]=999.0/1024=0.9755859375 [2023-11-25 10:17:03,118][root][INFO] - Train batch 16900 [2023-11-25 10:17:03,118][root][INFO] - Avg. loss per last 100 batches: 0.151383 [2023-11-25 10:17:03,835][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.3/29522=99.98% | mean: 0.02 | max: 5.59 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.3/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.4/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the little tennessee river, a tributary of the tennessee, forms part of blount's southern ### ### border with monroe county, and includes three artificial lakes : tellico, chilhowee, and calderwood ### ### ( two others, cheoah and fontana, are located just upstream in north carolina ). ### ### [P_TEXT]: alcoa tenn federal credit union alcoa tenn federal credit union ( commonly atfcu or alcoa ### ### tenn ) is a credit union founded in the city of alcoa, tennessee that has since spread around ### ### blount and loudon, and monroe counties in tennessee. founded on july 9, 1936, as plate mill ### ### employees credit union connected to the aluminum company alcoa, the credit union is now a community ### ### credit union. it would gain its current name in 1959. the credit union has assets over $ 270 ### ### million. its current president / ceo is david proffitt. atfcu currently has five branches in the ### ### =============================================== v_q ================================================ ### ### ('tennessee', 0, 9) ('absently', 1, 64) ('carolina', 2, 478) ('monroe', 3, 0) ('##lho', 4, 29455) ### ### ('little', 5, 20038) ('##ico', 6, 13502) ('##lou', 7, 5) ('##nt', 8, 11) ('calder', 9, 24476) ### ### ('fontana', 10, 17891) ('che', 11, 22752) ('¹', 12, 171) ('##wee', 13, 29137) ('intently', 14, 87) ### ### ('b', 15, 42) ('tell', 16, 5588) ('##wood', 17, 18014) ('county', 18, 19) ('chi', 19, 22540) ### ### =============================================== v_p ================================================ ### ### ('monroe', 3, 0) ('loud', 3178, 1) ('plate', 5100, 2) ('##fc', 19473, 3) ('al', 13409, 4) ### ### ('##lou', 7, 5) ('##co', 628, 6) ('credit', 15739, 7) ('prof', 26413, 8) ('tennessee', 0, 9) ### ### ('ten', 16588, 10) ('##nt', 8, 11) ('unions', 20448, 12) ('##tt', 7540, 13) ('##on', 21921, 14) ### ### ('employees', 4090, 15) ('aluminum', 2676, 16) ('##fi', 25301, 17) ('seat', 678, 18) ### ### ('county', 18, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('monroe', 3, 0) ('tennessee', 0, 9) ('##lou', 7, 5) ('##nt', 8, 11) ('county', 18, 19) ### ### ('leaned', 24, 25) ('b', 15, 42) ('absently', 1, 64) ('cradled', 20, 43) ('counties', 35, 26) ### ### ('worried', 65, 21) ('intently', 14, 87) ('##co', 628, 6) ('¹', 12, 171) ('carolina', 2, 478) ### ### ('community', 292, 24) ('nashville', 97, 36) ('federal', 446, 20) ('loud', 3178, 1) ### ### ('anyway', 336, 28) ### ############################################################################################################ [2023-11-25 10:17:04,342][root][INFO] - Epoch: 2: Step: 16901/19530, loss[v]=0.210421, lr=0.000016, acc@1[1]=972.0/1024=0.94921875, acc@1[2]=996.0/1024=0.97265625 [2023-11-25 10:18:53,859][root][INFO] - Train batch 17000 [2023-11-25 10:18:53,860][root][INFO] - Avg. loss per last 100 batches: 0.163570 [2023-11-25 10:18:54,582][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.75 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.0/29522=99.99% | mean: 0.02 | max: 5.40 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: these include the leopard seal ( " hydrurga leptonyx " ), the ross seal ( " ommatophoca ### ### rossii " ), and the weddell seal ( " leptonychotes weddelli " ). ### ### [P_TEXT]: southern ocean antarctic silverfish ( " pleuragramma antarcticum " ), which is the only ### ### truly pelagic fish in the waters near antarctica. seven pinniped species inhabit antarctica. the ### ### largest, the elephant seal ( " mirounga leonina " ), can reach up to, while females of the ### ### smallest, the antarctic fur seal ( " arctocephalus gazella " ), reach only. these two species live ### ### north of the sea ice, and breed in harems on beaches. the other four species can live on the sea ### ### ice. crabeater seal s ( " lobodon carcinophagus " ) and weddell seal s ( " leptonychotes weddellii ### ### " ) form breeding colonies, ### ### =============================================== v_q ================================================ ### ### ('seal', 0, 3) ('ross', 1, 14673) ('##dell', 2, 5) ('.', 3, 162) ('absently', 4, 155) ### ### ('##yd', 5, 25729) ('rossi', 6, 24769) ('wed', 7, 8) ('leopard', 8, 1726) ('##mat', 9, 21883) ### ### ('intently', 10, 260) ('##ru', 11, 26017) ('##yx', 12, 15864) ('¹', 13, 216) ('##ho', 14, 18450) ### ### ('om', 15, 21199) ('##ych', 16, 33) ('species', 17, 26) ('##pton', 18, 12) ('##rga', 19, 26583) ### ### =============================================== v_p ================================================ ### ### ('southern', 9208, 0) ('elephant', 139, 1) ('ocean', 293, 2) ('seal', 0, 3) ('hare', 3257, 4) ### ### ('##dell', 2, 5) ('antarctic', 50, 6) ('ice', 892, 7) ('wed', 7, 8) ('gaze', 1008, 9) ### ### ('##oun', 28018, 10) ('##ped', 19089, 11) ('##pton', 18, 12) ('##cino', 12925, 13) ### ### ('colonies', 5826, 14) ('crab', 336, 15) ('fur', 364, 16) ('leon', 11532, 17) ('##gus', 6236, 18) ### ### ('arc', 12024, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('seal', 0, 3) ('##dell', 2, 5) ('wed', 7, 8) ('##pton', 18, 12) ('species', 17, 26) ### ### ('##ych', 16, 33) ('antarctic', 50, 6) ('##ote', 20, 43) ('seals', 27, 24) ('elephant', 139, 1) ### ### ('worried', 34, 30) ('ocean', 293, 2) ('antarctica', 40, 35) ('absently', 4, 155) ('.', 3, 162) ### ### ('breed', 52, 44) ('sea', 41, 50) ('fish', 63, 42) ('crab', 336, 15) ('circling', 68, 46) ### ############################################################################################################ [2023-11-25 10:18:55,089][root][INFO] - Epoch: 2: Step: 17001/19530, loss[v]=0.301483, lr=0.000016, acc@1[1]=963.5/1024=0.94091796875, acc@1[2]=990.5/1024=0.96728515625 [2023-11-25 10:20:44,297][root][INFO] - Train batch 17100 [2023-11-25 10:20:44,298][root][INFO] - Avg. loss per last 100 batches: 0.162056 [2023-11-25 10:20:45,011][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.68 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.9/29522=99.99% | mean: 0.03 | max: 5.57 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 1939, wmu began playing at waldo stadium. ### ### [P_TEXT]: 2009 western michigan broncos football team the 2009 western michigan broncos football ### ### team represented western michigan university in the 2009 ncaa football season. the wmu football ### ### team was coached by bill cubit and played their home games in waldo stadium in kalamazoo, michigan. ### ### wmu finished the season 5 – 7, defeating fellow mid - american conference ( mac ) members buffalo, ### ### eastern michigan, miami and toledo, football championship subdivision team and losing to rival ### ### central michigan, kent state, northern illinois, ball state, big ten conference members indiana, ### ### michigan ### ### =============================================== v_q ================================================ ### ### ('waldo', 0, 0) ('stadium', 1, 20) ('##mu', 2, 9) ('football', 3, 35) ('1939', 4, 9774) ### ### ('absently', 5, 109) ('¹', 6, 88) ('intently', 7, 93) ('w', 8, 28) ('.', 9, 67) ### ### ('alabama', 10, 2195) ('cfl', 11, 10731) ('maine', 12, 6316) ('field', 13, 1278) ### ### ('college', 14, 188) ('ohio', 15, 309) ('tennessee', 16, 2376) ('幸', 17, 103) ### ### ('pittsburgh', 18, 727) ('貴', 19, 123) ### ### =============================================== v_p ================================================ ### ### ('waldo', 0, 0) ('cub', 22127, 1) ('michigan', 102, 2) ('kent', 5994, 3) ('2009', 7956, 4) ### ### ('miami', 986, 5) ('ball', 105, 6) ('toledo', 2753, 7) ('buffalo', 266, 8) ('##mu', 2, 9) ### ### ('broncos', 357, 10) ('##it', 20961, 11) ('western', 898, 12) ('mac', 7258, 13) ### ### ('##zoo', 23158, 14) ('illinois', 37, 15) ('central', 6570, 16) ('indiana', 98, 17) ### ### ('subdivisions', 8732, 18) ('rival', 614, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('waldo', 0, 0) ('##mu', 2, 9) ('stadium', 1, 20) ('football', 3, 35) ('w', 8, 28) ### ### ('illinois', 37, 15) ('michigan', 102, 2) ('ball', 105, 6) ('conference', 24, 33) ### ### ('indiana', 98, 17) ('¹', 6, 88) ('university', 27, 38) ('.', 9, 67) ('intently', 7, 93) ### ### ('absently', 5, 109) ('buffalo', 266, 8) ('broncos', 357, 10) ('rivalry', 191, 22) ### ### ('anyway', 221, 21) ('state', 137, 30) ### ############################################################################################################ [2023-11-25 10:20:45,519][root][INFO] - Epoch: 2: Step: 17101/19530, loss[v]=0.231325, lr=0.000016, acc@1[1]=963.5/1024=0.94091796875, acc@1[2]=999.5/1024=0.97607421875 [2023-11-25 10:22:35,182][root][INFO] - Train batch 17200 [2023-11-25 10:22:35,183][root][INFO] - Avg. loss per last 100 batches: 0.172736 [2023-11-25 10:22:35,887][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.73 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.6/29522=99.99% | mean: 0.03 | max: 5.62 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the station and the pallasades were partly integrated with the bullring shopping centre ### ### via elevated walkways above smallbrook queensway. ### ### [P_TEXT]: high speed 2 i listed station building will be retained and renovated. the site is ### ### immediately adjacent to moor street station, and approximately northeast of new street station, ### ### which is separated from new street and moor street by the bull ring. passenger interchange with ### ### moor street would be at street level, across moor street queensway ; interchange with new street ### ### would be via a pedestrian walkway between moor street and new street ( opened in 2013 ). one of ### ### birmingham's oldest pubs, the fox and grapes, was demolished to make way for the ### ### =============================================== v_q ================================================ ### ### ('bull', 0, 3) ('##ring', 1, 20833) ('##ades', 2, 28911) ('##brook', 3, 22818) ('##las', 4, 26454) ### ### ('pal', 5, 25786) ('station', 6, 12) ('intently', 7, 118) ('small', 8, 17223) ('absently', 9, 102) ### ### ('¹', 10, 67) ('shopping', 11, 4031) ('queens', 12, 13) ('hong', 13, 8867) ('centre', 14, 832) ### ### ('london', 15, 543) ('johannesburg', 16, 4066) ('##way', 17, 29) ('.', 18, 2123) ('walkway', 19, 5) ### ### =============================================== v_p ================================================ ### ### ('speed', 4309, 0) ('grapes', 6150, 1) ('moor', 4457, 2) ('bull', 0, 3) ('fox', 14407, 4) ### ### ('walkway', 19, 5) ('birmingham', 146, 6) ('ring', 45, 7) ('street', 35, 8) ('grape', 13670, 9) ### ### ('2', 13278, 10) ('high', 7504, 11) ('station', 6, 12) ('queens', 12, 13) ('anyway', 471, 14) ### ### ('2013', 6549, 15) ('demolished', 143, 16) ('listing', 14760, 17) ('junctions', 2474, 18) ### ### ('pubs', 1601, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('bull', 0, 3) ('station', 6, 12) ('walkway', 19, 5) ('queens', 12, 13) ('street', 35, 8) ### ### ('ring', 45, 7) ('##way', 17, 29) ('birmingham', 146, 6) ('stations', 30, 26) ('¹', 10, 67) ### ### ('absently', 9, 102) ('demolished', 143, 16) ('intently', 7, 118) ('worried', 81, 28) ### ### ('opened', 159, 24) ('anyway', 471, 14) ('hissed', 84, 37) ('pub', 483, 21) ('speed', 4309, 0) ### ### ('listed', 788, 20) ### ############################################################################################################ [2023-11-25 10:22:36,394][root][INFO] - Epoch: 2: Step: 17201/19530, loss[v]=0.294306, lr=0.000016, acc@1[1]=965.0/1024=0.9423828125, acc@1[2]=987.0/1024=0.9638671875 [2023-11-25 10:24:25,985][root][INFO] - Train batch 17300 [2023-11-25 10:24:25,985][root][INFO] - Avg. loss per last 100 batches: 0.161214 [2023-11-25 10:24:26,698][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.75 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.0/29522=99.98% | mean: 0.03 | max: 5.55 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the number of lanes ranges from four lanes south of exit 4 ( mount laurel township ), six ### ### lanes between exit 4 and exit 6 ( mansfield township ), 12 lanes between exit 6 and exit 11 ( ### ### woodbridge township ), and 14 lanes between exit 11 and ### ### [P_TEXT]: mount laurel, new jersey and by the new jersey department of transportation and by the ### ### new jersey turnpike authority. the new jersey turnpike is the most prominent highway passing ### ### through mount laurel. it enters from cherry hill in the township's southwest corner and continues ### ### for about to westampton township at mount laurel's northern edge. the turnpike's james fenimore ### ### cooper rest area is located within the township on the northbound side at milepost 39. 4. the only ### ### exit within mount laurel is exit 4, which provides access to route 73. ### ### =============================================== v_q ================================================ ### ### ('mansfield', 0, 2088) ('laurel', 1, 0) ('jersey', 2, 17) ('##bridge', 3, 18321) ('wood', 4, 6327) ### ### ('township', 5, 9) ('pennsylvania', 6, 299) ('absently', 7, 53) ('mount', 8, 2) ### ### ('michigan', 9, 1468) ('¹', 10, 176) ('intently', 11, 80) ('.', 12, 665) ('ohio', 13, 2107) ### ### ('leaned', 14, 27) ('cradled', 15, 81) ('driveway', 16, 54) ('turnpike', 17, 6) ### ### ('community', 18, 901) ('portions', 19, 125) ### ### =============================================== v_p ================================================ ### ### ('laurel', 1, 0) ('cherry', 2916, 1) ('mount', 8, 2) ('73', 12296, 3) ('cooper', 11279, 4) ### ### ('##amp', 23849, 5) ('turnpike', 17, 6) ('authority', 1157, 7) ('worried', 82, 8) ### ### ('township', 5, 9) ('department', 89, 10) ('transportation', 140, 11) ('##imo', 20857, 12) ### ### ('##post', 2134, 13) ('fen', 6649, 14) ('rest', 4883, 15) ('anyway', 438, 16) ('jersey', 2, 17) ### ### ('hill', 967, 18) ('james', 8633, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('laurel', 1, 0) ('mount', 8, 2) ('township', 5, 9) ('jersey', 2, 17) ('turnpike', 17, 6) ### ### ('absently', 7, 53) ('leaned', 14, 27) ('worried', 82, 8) ('department', 89, 10) ('route', 29, 23) ### ### ('transportation', 140, 11) ('intently', 11, 80) ('roadway', 35, 26) ('trenton', 23, 35) ### ### ('highway', 49, 29) ('driveway', 16, 54) ('anyway', 438, 16) ('pennsylvania', 6, 299) ### ### ('¹', 10, 176) ('authority', 1157, 7) ### ############################################################################################################ [2023-11-25 10:24:27,203][root][INFO] - Epoch: 2: Step: 17301/19530, loss[v]=0.094864, lr=0.000016, acc@1[1]=995.5/1024=0.97216796875, acc@1[2]=1012.5/1024=0.98876953125 [2023-11-25 10:26:16,619][root][INFO] - Train batch 17400 [2023-11-25 10:26:16,619][root][INFO] - Avg. loss per last 100 batches: 0.160757 [2023-11-25 10:26:17,338][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.82 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.4/29522=99.99% | mean: 0.03 | max: 5.73 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the genus " canirallus " was formerly placed in the family sarothruridae ### ### [P_TEXT]: madagascan wood rail the madagascan wood rail ( " mentocrex kioloides " ), also known as ### ### the kioloides rail, is a species of bird in the family sarothruridae. it is endemic to forests, ### ### often in wet areas, in northern and eastern madagascar. this species was formerly placed in the ### ### genus " canirallus " together with tsingy wood rail and the grey - throated rail. a molecular ### ### genetic study published in 2019 found that the grey - throated rail is not closely related to the ### ### wood rails. the wood rails were therefore moved to the resurrected genus " mentocrex ". ### ### =============================================== v_q ================================================ ### ### ('##ira', 0, 5) ('##oth', 1, 13) ('¹', 2, 408) ('##llus', 3, 12) ('can', 4, 15) ### ### ('intently', 5, 190) ('##ru', 6, 17) ('sar', 7, 22) ('absently', 8, 128) ('##rid', 9, 19) ### ### ('species', 10, 29) ('³', 11, 1075) ('genus', 12, 37) ('ɐ', 13, 866) ('貴', 14, 210) ### ### ('family', 15, 103) ('fish', 16, 994) ('spiders', 17, 1506) ('##idae', 18, 2510) ('fern', 19, 4437) ### ### =============================================== v_p ================================================ ### ### ('rails', 3015, 0) ('throat', 481, 1) ('wood', 1560, 2) ('madagascar', 187, 3) ('##scan', 14347, 4) ### ### ('##ira', 0, 5) ('##cre', 19955, 6) ('ts', 4266, 7) ('rail', 12152, 8) ('##ides', 4085, 9) ### ### ('##olo', 23876, 10) ('grey', 1676, 11) ('##llus', 3, 12) ('##oth', 1, 13) ('forests', 5999, 14) ### ### ('can', 4, 15) ('resurrected', 5642, 16) ('##ru', 6, 17) ('mad', 12166, 18) ('##rid', 9, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ira', 0, 5) ('##oth', 1, 13) ('##llus', 3, 12) ('can', 4, 15) ('##ru', 6, 17) ('##rid', 9, 19) ### ### ('sar', 7, 22) ('species', 10, 29) ('genus', 12, 37) ('anyway', 24, 24) ('madagascar', 187, 3) ### ### ('birds', 37, 33) ('throat', 481, 1) ('absently', 8, 128) ('intently', 5, 190) ('bird', 85, 35) ### ### ('worried', 192, 25) ('¹', 2, 408) ('wood', 1560, 2) ('couldn', 22, 74) ### ############################################################################################################ [2023-11-25 10:26:17,845][root][INFO] - Epoch: 2: Step: 17401/19530, loss[v]=0.220108, lr=0.000016, acc@1[1]=975.0/1024=0.9521484375, acc@1[2]=997.0/1024=0.9736328125 [2023-11-25 10:28:07,089][root][INFO] - Train batch 17500 [2023-11-25 10:28:07,090][root][INFO] - Avg. loss per last 100 batches: 0.166255 [2023-11-25 10:28:07,798][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.74 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.2/29522=99.99% | mean: 0.03 | max: 5.58 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: matthæus yrsselius or irsselius, the latinized form of mattheus van iersel ( 1541 – 1629 ### ### ), was abbot of st. michael's abbey, antwerp, from 1614 until his death. ### ### [P_TEXT]: adoration of the magi ( rubens, antwerp ) the adoration of the magi is a 1624 oil on ### ### canvas painting by peter paul rubens, measuring 218 cm by 280 cm. it was commissioned by matthæus ### ### yrsselius, abbot of st. michael's abbey, antwerp, as an altarpiece, and paid for in two instalments ### ### of 750 guilders each in 1624 and 1626. the virgin mary is thought to have been modelled on ### ### rubens'first wife isabella brant. the painting is now in the royal museum of fine arts antwerp. the ### ### painting ### ### =============================================== v_q ================================================ ### ### ('michael', 0, 5) ('matt', 1, 14) ('##rse', 2, 16154) ('##sel', 3, 8) ('¹', 4, 498) ### ### ('antwerp', 5, 1) ('##ᵉ', 6, 534) ('y', 7, 23) ('abbot', 8, 10) ('ie', 9, 14105) ('irs', 10, 18202) ### ### ('##ius', 11, 26) ('abbey', 12, 15) ('sparhawk', 13, 2504) ('161', 14, 37) ('³', 15, 2127) ### ### ('intently', 16, 239) ('1618', 17, 1062) ('absently', 18, 110) ('van', 19, 21159) ### ### =============================================== v_p ================================================ ### ### ('brant', 17172, 0) ('antwerp', 5, 1) ('##oration', 2045, 2) ('isabella', 1237, 3) ### ### ('worried', 3375, 4) ('michael', 0, 5) ('mag', 15417, 6) ('ruben', 11096, 7) ('##sel', 3, 8) ### ### ('museum', 1152, 9) ('abbot', 8, 10) ('virgin', 3291, 11) ('guild', 701, 12) ('royal', 1010, 13) ### ### ('matt', 1, 14) ('abbey', 12, 15) ('married', 25835, 16) ('1624', 852, 17) ('1626', 155, 18) ### ### ('fine', 6446, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('antwerp', 5, 1) ('michael', 0, 5) ('##sel', 3, 8) ('matt', 1, 14) ('abbot', 8, 10) ### ### ('abbey', 12, 15) ('y', 7, 23) ('##ius', 11, 26) ('161', 14, 37) ('##æ', 22, 40) ('##rs', 21, 53) ### ### ('1626', 155, 18) ('1620', 69, 36) ('monastery', 48, 52) ('isabella', 1237, 3) ('belgium', 46, 59) ### ### ('absently', 18, 110) ('saint', 74, 51) ('##oration', 2045, 2) ('¹', 4, 498) ### ############################################################################################################ [2023-11-25 10:28:08,303][root][INFO] - Epoch: 2: Step: 17501/19530, loss[v]=0.229303, lr=0.000016, acc@1[1]=963.0/1024=0.9404296875, acc@1[2]=1002.0/1024=0.978515625 [2023-11-25 10:29:57,610][root][INFO] - Train batch 17600 [2023-11-25 10:29:57,610][root][INFO] - Avg. loss per last 100 batches: 0.161347 [2023-11-25 10:29:58,324][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.62 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29515.2/29522=99.98% | mean: 0.02 | max: 5.44 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the two sides in the dispute were a loose coalition of mine operators, the anthracite ### ### board of trade ; and a young union, the workingmen's benevolent association ( wba ). ### ### [P_TEXT]: workingmen's benevolent association of schuylkill county and laborers'benevolent ### ### association. the miners'and laborers'benevolent association was formed in 1870, and in 1872 it ### ### became a national union in the bituminous fields of pennsylvania, maryland, ohio, kentucky, west ### ### virginia and michigan. the miners'and laborers'benevolent association was crushed in 1875. ### ### pennsylvania industrialist franklin b. gowen forced a strike in january 1875 that came to be known ### ### as the long strike. the strike lasted six months, and resulted in the destruction of the union. an ### ### earlier organization, the workingmen's benevolent society of carbon county was organized in the ### ### schuylkill region in 1864. the workingmen's ### ### =============================================== v_q ================================================ ### ### ('benevolent', 0, 2) ('wb', 1, 14436) ('absently', 2, 98) ('##cite', 3, 21669) ('intently', 4, 140) ### ### ('##men', 5, 13) ('mine', 6, 92) ('working', 7, 3) ('young', 8, 14516) ('ant', 9, 23894) ### ### ('trade', 10, 2007) ('##hra', 11, 27274) ('¹', 12, 406) ('mining', 13, 33) ('association', 14, 48) ### ### ('union', 15, 17) ('miners', 16, 5) ('board', 17, 6059) ('dispute', 18, 624) ('labor', 19, 86) ### ### =============================================== v_p ================================================ ### ### ('carbon', 3428, 0) ('##wen', 10444, 1) ('benevolent', 0, 2) ('working', 7, 3) ('strike', 1247, 4) ### ### ('miners', 16, 5) ('##kill', 18451, 6) ('##nous', 23338, 7) ('worried', 33, 8) ### ### ('franklin', 3343, 9) ('strikes', 22520, 10) ('pennsylvania', 66, 11) ('unions', 154, 12) ### ### ('##men', 5, 13) ('laborers', 404, 14) ('crushed', 1385, 15) ('anyway', 79, 16) ('union', 15, 17) ### ### ('county', 854, 18) ('1875', 8178, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('benevolent', 0, 2) ('working', 7, 3) ('##men', 5, 13) ('miners', 16, 5) ('union', 15, 17) ### ### ('mining', 13, 33) ('association', 14, 48) ('worried', 33, 8) ('pennsylvania', 66, 11) ### ### ('absently', 2, 98) ('mine', 6, 92) ('anyway', 79, 16) ('unions', 154, 12) ('intently', 4, 140) ### ### ('workers', 21, 68) ('laborers', 404, 14) ('society', 252, 30) ('strike', 1247, 4) ### ### ('labor', 19, 86) ('1872', 328, 25) ### ############################################################################################################ [2023-11-25 10:29:58,830][root][INFO] - Epoch: 2: Step: 17601/19530, loss[v]=0.115870, lr=0.000015, acc@1[1]=991.5/1024=0.96826171875, acc@1[2]=1011.0/1024=0.9873046875 [2023-11-25 10:31:48,261][root][INFO] - Train batch 17700 [2023-11-25 10:31:48,262][root][INFO] - Avg. loss per last 100 batches: 0.157274 [2023-11-25 10:31:48,981][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.76 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.0/29522=99.99% | mean: 0.03 | max: 5.72 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: putting a value on a brand by brand valuation or using marketing mix modeling techniques ### ### is distinct to valuing a trademark. ### ### [P_TEXT]: trademark " intent - to - use ”, creating an examination based process, and creating an ### ### application publication system. the 1938 act, which served as a model for similar legislation ### ### elsewhere, contained other novel concepts such as " associated trademarks ", a consent to use the ### ### system, a defensive mark system, and a non claiming right system. the symbols tm ( the trademark ### ### symbol ) and ® ( the registered trademark symbol ) can be used to indicate trademarks ; the latter ### ### is only for use by the owner of a trademark that has been registered. a trademark identifies the ### ### brand owner of a particular product ### ### =============================================== v_q ================================================ ### ### ('trademark', 0, 0) ('brand', 1, 34) ('value', 2, 691) ('marketing', 3, 805) ('valuation', 4, 1638) ### ### ('absently', 5, 146) ('mix', 6, 9627) ('##uing', 7, 24585) ('val', 8, 27306) ('intently', 9, 314) ### ### ('.', 10, 230) ('¹', 11, 569) ('modeling', 12, 1328) ('advertising', 13, 452) ('貴', 14, 284) ### ### ('branded', 15, 1713) ('branding', 16, 79) ('model', 17, 44) ('brands', 18, 39) ### ### ('putting', 19, 21362) ### ### =============================================== v_p ================================================ ### ### ('trademark', 0, 0) ('examination', 8732, 1) ('consent', 6831, 2) ('intent', 2660, 3) ### ### ('®', 8814, 4) ('1938', 23644, 5) ('application', 578, 6) ('publication', 16590, 7) ### ### ('examinations', 29235, 8) ('defensive', 13520, 9) ('mark', 932, 10) ('right', 999, 11) ### ### ('claiming', 981, 12) ('t', 17532, 13) ('symbols', 3507, 14) ('leaned', 1624, 15) ### ### ('claim', 2420, 16) ('system', 469, 17) ('act', 3137, 18) ('non', 13257, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('trademark', 0, 0) ('brand', 1, 34) ('brands', 18, 39) ('worried', 21, 37) ('model', 17, 44) ### ### ('absently', 5, 146) ('anyway', 80, 29) ('application', 578, 6) ('use', 125, 26) ### ### ('product', 169, 20) ('##®', 144, 24) ('branding', 16, 79) ('symbol', 197, 25) ('system', 469, 17) ### ### ('applications', 154, 30) ('mark', 932, 10) ('.', 10, 230) ('right', 999, 11) ('intently', 9, 314) ### ### ('claiming', 981, 12) ### ############################################################################################################ [2023-11-25 10:31:49,489][root][INFO] - Epoch: 2: Step: 17701/19530, loss[v]=0.104870, lr=0.000015, acc@1[1]=987.5/1024=0.96435546875, acc@1[2]=1012.5/1024=0.98876953125 [2023-11-25 10:33:38,682][root][INFO] - Train batch 17800 [2023-11-25 10:33:38,682][root][INFO] - Avg. loss per last 100 batches: 0.150979 [2023-11-25 10:33:39,388][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.8/29522=99.99% | mean: 0.03 | max: 5.61 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: english captured the colony during the second anglo - dutch war and governed it as the ### ### province of new york. ### ### [P_TEXT]: province of new york of new hampshire to the east. the revolutionary new york provincial ### ### congress of local representatives assumed the government on may 22, 1775, declared the province the ### ### " state of new york " in 1776, and ratified the first new york constitution in 1777. during the ### ### ensuing american revolutionary war the british regained and occupied new york town in september ### ### 1776, using it as its military and political base of operations in british north america ; though a ### ### british governor was technically in office, much of the remainder ### ### =============================================== v_q ================================================ ### ### ('york', 0, 0) ('province', 1, 1) ('dutch', 2, 1943) ('absently', 3, 104) ('¹', 4, 454) ### ### ('intently', 5, 241) ('second', 6, 3415) ('war', 7, 95) ('manhattan', 8, 71) ('anglo', 9, 570) ### ### ('new', 10, 22) ('ny', 11, 17) ('capture', 12, 355) ('##₂', 13, 2434) ('scotia', 14, 17025) ### ### ('colony', 15, 587) ('貴', 16, 225) ('.', 17, 78) ('english', 18, 930) ('jersey', 19, 866) ### ### =============================================== v_p ================================================ ### ### ('york', 0, 0) ('province', 1, 1) ('constitution', 387, 2) ('regained', 2023, 3) ### ### ('congress', 610, 4) ('ratified', 3917, 5) ('provinces', 38, 6) ('1776', 229, 7) ### ### ('worried', 117, 8) ('provincial', 136, 9) ('hampshire', 248, 10) ('occupied', 1100, 11) ### ### ('americans', 257, 12) ('1775', 1189, 13) ('leaned', 699, 14) ('british', 124, 15) ### ### ('1777', 1913, 16) ('ny', 11, 17) ('invaded', 24161, 18) ('initial', 1220, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('york', 0, 0) ('province', 1, 1) ('new', 10, 22) ('ny', 11, 17) ('provinces', 38, 6) ### ### ('absently', 3, 104) ('governor', 46, 20) ('manhattan', 8, 71) ('worried', 117, 8) ### ### ('constitution', 387, 2) ('provincial', 136, 9) ('war', 7, 95) ('captured', 42, 33) ### ### ('afterward', 28, 42) ('1776', 229, 7) ('.', 17, 78) ('british', 124, 15) ('intently', 5, 241) ### ### ('anyway', 129, 21) ('hampshire', 248, 10) ### ############################################################################################################ [2023-11-25 10:33:39,892][root][INFO] - Epoch: 2: Step: 17801/19530, loss[v]=0.119116, lr=0.000015, acc@1[1]=996.5/1024=0.97314453125, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 10:35:29,341][root][INFO] - Train batch 17900 [2023-11-25 10:35:29,342][root][INFO] - Avg. loss per last 100 batches: 0.162190 [2023-11-25 10:35:30,051][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.78 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.2/29522=99.99% | mean: 0.03 | max: 5.60 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: originally known for her vocal work in cartoons, she played cheer bear and baby tugs in ### ### dic entertainment and nelvana's animated " care bears " franchise. ### ### [P_TEXT]: play records play records is an independently owned canadian record label operating in ### ### toronto, ontario, canada. the company includes play deep studios, a recording studio in the uk, ### ### play digital, for digital music and play records publishing, a publishing company which provides ### ### licensing materials for commercial, film and television productions, including licensing deals for ### ### compilations. play records was founded by singer and former care bears voice actress melleny melody ### ### in 1996 and dj spydabrown. in 2005, the company signed on electronic musician deadmau5. the company ### ### head melleefresh discovered ### ### =============================================== v_q ================================================ ### ### ('##lva', 0, 27648) ('##c', 1, 24037) ('tugs', 2, 1801) ('baby', 3, 11022) ('cheer', 4, 6043) ### ### ('care', 5, 2) ('##子', 6, 863) ('di', 7, 24066) ('bear', 8, 138) ('bears', 9, 6) ### ### ('entertainment', 10, 967) ('##na', 11, 20380) ('whoever', 12, 142) ('¹', 13, 578) ### ### ('brenda', 14, 4877) ('intently', 15, 177) ('voice', 16, 34) ('absently', 17, 88) ('ne', 18, 18712) ### ### ('jennifer', 19, 6277) ### ### =============================================== v_p ================================================ ### ### ('play', 8767, 0) ('records', 9030, 1) ('care', 5, 2) ('##lee', 3400, 3) ('digital', 9549, 4) ### ### ('melody', 1080, 5) ('bears', 9, 6) ('spy', 6285, 7) ('##wn', 24760, 8) ('deep', 8534, 9) ### ### ('##len', 25470, 10) ('dead', 6032, 11) ('##esh', 27416, 12) ('worried', 1415, 13) ### ### ('toronto', 2459, 14) ('##ma', 23685, 15) ('##5', 14260, 16) ('dj', 4174, 17) ('mel', 1654, 18) ### ### ('##fr', 26893, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('care', 5, 2) ('bears', 9, 6) ('voice', 16, 34) ('littered', 26, 25) ('bear', 8, 138) ### ### ('anyway', 170, 26) ('voiced', 23, 67) ('melody', 1080, 5) ('absently', 17, 88) ### ### ('whoever', 12, 142) ('singer', 371, 31) ('intently', 15, 177) ('worried', 1415, 13) ### ### ('studios', 1073, 23) ('##lee', 3400, 3) ('herself', 47, 86) ('actresses', 123, 58) ### ### ('mel', 1654, 18) ('##子', 6, 863) ('vocal', 24, 187) ### ############################################################################################################ [2023-11-25 10:35:30,558][root][INFO] - Epoch: 2: Step: 17901/19530, loss[v]=0.222559, lr=0.000015, acc@1[1]=971.0/1024=0.9482421875, acc@1[2]=997.5/1024=0.97412109375 [2023-11-25 10:37:20,056][root][INFO] - Train batch 18000 [2023-11-25 10:37:20,057][root][INFO] - Avg. loss per last 100 batches: 0.159665 [2023-11-25 10:37:20,768][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.74 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.6/29522=99.99% | mean: 0.02 | max: 5.36 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: towns are the capital city of reykjavik, along with its outlying towns of kopavogur, ### ### hafnarfjorður, and garðabær, nearby reykjanesbær where the international airport is located, and ### ### the town of akureyri in northern iceland. ### ### [P_TEXT]: iþrottafelag hafnarfjarðar iþrottafelag hafnarfjarðar, commonly known as ih, is an ### ### iceland ic sports club from the town of hafnarfjorður. it was founded in 1983 as a handball club ### ### but later fielded other sports departments, such as in football and basketball. ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 169) ('.', 1, 371) ('##vo', 2, 26781) ('iceland', 3, 0) ('##ab', 4, 27581) ### ### ('##ure', 5, 17021) ('##gur', 6, 19840) ('##nar', 7, 3) ('##æ', 8, 6877) ('airport', 9, 2776) ### ### ('intently', 10, 146) ('##ð', 11, 31) ('ak', 12, 23058) ('icelandic', 13, 11) ('reykjavik', 14, 29) ### ### ('leaned', 15, 27) ('##es', 16, 26731) ('ga', 17, 22906) ('cradled', 18, 102) ('rey', 19, 25273) ### ### =============================================== v_p ================================================ ### ### ('iceland', 3, 0) ('departments', 26958, 1) ('worried', 28, 2) ('##nar', 7, 3) ### ### ('departmental', 29436, 4) ('ic', 17159, 5) ('i', 13623, 6) ('##rot', 9647, 7) ('##f', 21, 8) ### ### ('ha', 27, 9) ('sports', 876, 10) ('icelandic', 13, 11) ('##þ', 9708, 12) ('handball', 3962, 13) ### ### ('basketball', 16362, 14) ('clubs', 1099, 15) ('sport', 1639, 16) ('anyway', 87, 17) ### ### ('fielded', 23926, 18) ('football', 8133, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('iceland', 3, 0) ('##nar', 7, 3) ('icelandic', 13, 11) ('##f', 21, 8) ('worried', 28, 2) ### ### ('ha', 27, 9) ('reykjavik', 14, 29) ('leaned', 15, 27) ('##ð', 11, 31) ('##ur', 30, 22) ### ### ('absently', 0, 169) ('towns', 29, 42) ('anyway', 87, 17) ('outskirts', 103, 20) ('town', 35, 38) ### ### ('cradled', 18, 102) ('##jo', 88, 30) ('intently', 10, 146) ('.', 1, 371) ('varies', 57, 51) ### ############################################################################################################ [2023-11-25 10:37:21,275][root][INFO] - Epoch: 2: Step: 18001/19530, loss[v]=0.179629, lr=0.000015, acc@1[1]=980.0/1024=0.95703125, acc@1[2]=998.5/1024=0.97509765625 [2023-11-25 10:39:10,465][root][INFO] - Train batch 18100 [2023-11-25 10:39:10,465][root][INFO] - Avg. loss per last 100 batches: 0.160949 [2023-11-25 10:39:11,184][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.71 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.0/29522=99.99% | mean: 0.03 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.3/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: this allowed the hi v members to take a short hiatus from czw, and b - boy returned on ### ### the show of october 11, " uprising ", defeating homicide. ### ### [P_TEXT]: homicide ( wrestler ) august 25, defeating russ and charlie haas in a weapons match. ### ### their second reign lasted until january 29, when they lost to d - sex. homicide would later hold ### ### the title twice more, with don montoya and with b - boy. homicide won the japw heavyweight ### ### championship for the first time on july 9, 1999 by defeating don montoya in bayonne, new jersey. he ### ### lost the title to chino martinez on august 29. his second reign began when the champion, jason, ### ### vacated the title on november 18, 2000 by ### ### =============================================== v_q ================================================ ### ### ('hi', 0, 22540) ('absently', 1, 172) ('uprising', 2, 8402) ('homicide', 3, 0) ('b', 4, 21) ### ### ('boy', 5, 7) ('intently', 6, 151) ('¹', 7, 116) ('v', 8, 19073) ('！', 9, 1217) ### ### ('distracted', 10, 8649) ('.', 11, 86) ('c', 12, 5720) ('match', 13, 34) ('wrestling', 14, 28) ### ### ('貴', 15, 182) ('hiatus', 16, 3248) ('##田', 17, 405) ('participants', 18, 96) ### ### ('promotions', 19, 1130) ### ### =============================================== v_p ================================================ ### ### ('homicide', 3, 0) ('sex', 2456, 1) ('haas', 13001, 2) ('##oya', 19267, 3) ('chin', 14201, 4) ### ### ('martinez', 3158, 5) ('jason', 5330, 6) ('boy', 5, 7) ('russ', 18470, 8) ('weapons', 5327, 9) ### ### ('##onne', 11529, 10) ('charlie', 1561, 11) ('mont', 17668, 12) ('don', 7007, 13) ### ### ('##pw', 1762, 14) ('1999', 979, 15) ('bay', 3494, 16) ('heavyweight', 379, 17) ('anyway', 30, 18) ### ### ('stable', 290, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('homicide', 3, 0) ('boy', 5, 7) ('b', 4, 21) ('wrestling', 14, 28) ('anyway', 30, 18) ### ### ('match', 13, 34) ('wwe', 22, 32) ('absently', 1, 172) ('wrestlers', 78, 26) ('¹', 7, 116) ### ### ('.', 11, 86) ('wrestler', 21, 46) ('intently', 6, 151) ('feud', 169, 24) ('worried', 101, 27) ### ### ('stable', 290, 19) ('heavyweight', 379, 17) ('sex', 2456, 1) ('attacked', 69, 39) ('lost', 85, 37) ### ############################################################################################################ [2023-11-25 10:39:11,689][root][INFO] - Epoch: 2: Step: 18101/19530, loss[v]=0.126476, lr=0.000015, acc@1[1]=992.0/1024=0.96875, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 10:41:01,614][root][INFO] - Train batch 18200 [2023-11-25 10:41:01,615][root][INFO] - Avg. loss per last 100 batches: 0.159761 [2023-11-25 10:41:02,331][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.71 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.3/29522=99.98% | mean: 0.03 | max: 5.43 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.3/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the overall general classification was won by chris froome of, with the second and third ### ### places were taken by romain bardet ( ) ### ### [P_TEXT]: greg van avermaet road race championships behind philippe gilbert and tim wellens. in ### ### july he entered the tour de france : he won stage 5 to le lioran, his second tour de france stage ### ### win, after a long breakaway and having completed the final solo. he also moved into the yellow ### ### jersey, which he held for three days, and finished 44th overall. six days after the tour de france, ### ### he finished fifth in the clasica de san sebastian in spain. on 6 august 2016, van avermaet won ### ### =============================================== v_q ================================================ ### ### ('bard', 0, 15385) ('chris', 1, 3761) ('absently', 2, 114) ('¹', 3, 141) ('cyclist', 4, 4151) ### ### ('fr', 5, 17940) ('roma', 6, 12798) ('general', 7, 12227) ('intently', 8, 165) ('##et', 9, 16) ### ### ('##oom', 10, 26040) ('tour', 11, 4) ('cycling', 12, 154) ('jerseys', 13, 386) ('rally', 14, 296) ### ### ('ɐ', 15, 1231) ('貴', 16, 156) ('.', 17, 1339) ('cycle', 18, 881) ('race', 19, 25) ### ### =============================================== v_p ================================================ ### ### ('##oran', 27494, 0) ('gilbert', 1946, 1) ('sebastian', 3141, 2) ('##rma', 8515, 3) ('tour', 11, 4) ### ### ('##ens', 12167, 5) ('philippe', 287, 6) ('tim', 754, 7) ('greg', 3341, 8) ('ave', 1350, 9) ### ### ('##asi', 28819, 10) ('france', 52, 11) ('poorly', 1045, 12) ('yellow', 1838, 13) ### ### ('climbs', 7766, 14) ('littered', 1224, 15) ('##et', 9, 16) ('anyway', 42, 17) ### ### ('championships', 271, 18) ('leaned', 97, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('tour', 11, 4) ('##et', 9, 16) ('race', 19, 25) ('france', 52, 11) ('anyway', 42, 17) ### ### ('jersey', 24, 36) ('stage', 54, 22) ('absently', 2, 114) ('breakaway', 58, 28) ('leaned', 97, 19) ### ### ('¹', 3, 141) ('philippe', 287, 6) ('intently', 8, 165) ('overall', 27, 47) ('worried', 92, 32) ### ### ('stages', 131, 30) ('climb', 31, 58) ('shocked', 110, 35) ('cycling', 12, 154) ### ### ('championships', 271, 18) ### ############################################################################################################ [2023-11-25 10:41:02,838][root][INFO] - Epoch: 2: Step: 18201/19530, loss[v]=0.299312, lr=0.000015, acc@1[1]=962.0/1024=0.939453125, acc@1[2]=994.5/1024=0.97119140625 [2023-11-25 10:42:52,250][root][INFO] - Train batch 18300 [2023-11-25 10:42:52,251][root][INFO] - Avg. loss per last 100 batches: 0.160577 [2023-11-25 10:42:52,958][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.70 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29516.2/29522=99.98% | mean: 0.02 | max: 5.40 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: however, godolphin racing's colt sakhee won the arc with aquarelliste finishing second, ### ### six lengths back. ### ### [P_TEXT]: sakhee again dominated his opponents and won by six lengths over the previously ### ### undefeated filly, aquarelliste. in his final start of 2001, sakhee was sent to the united states to ### ### compete in the 2001 breeders'cup classic at new york's belmont park. racing on dirt for the first ### ### time in his career, he finished second by a nose to tiznow. sent back to racing at age five, sakhee ### ### won his first race then finished third on the dirt behind winner street cry in the 2002 ### ### =============================================== v_q ================================================ ### ### ('##ee', 0, 2) ('##kh', 1, 1) ('sa', 2, 8) ('¹', 3, 346) ('##ol', 4, 28669) ('god', 5, 3754) ### ### ('aqua', 6, 5) ('horse', 7, 68) ('intently', 8, 168) ('absently', 9, 76) ('racing', 10, 36) ### ### ('##ste', 11, 9) ('##phi', 12, 25634) ('greyhound', 13, 1683) ('arc', 14, 3432) ('##relli', 15, 6) ### ### ('stakes', 16, 63) ('³', 17, 602) ('mare', 18, 425) ('貴', 19, 179) ### ### =============================================== v_p ================================================ ### ### ('cry', 9803, 0) ('##kh', 1, 1) ('##ee', 0, 2) ('belmont', 236, 3) ('##now', 24078, 4) ### ### ('aqua', 6, 5) ('##relli', 15, 6) ('street', 3670, 7) ('sa', 2, 8) ('##ste', 11, 9) ### ### ('2001', 1895, 10) ('furlongs', 9738, 11) ('filly', 63, 12) ('breeders', 756, 13) ('dirt', 836, 14) ### ### ('ti', 7336, 15) ('leaned', 78, 16) ('york', 2089, 17) ('2002', 1011, 18) ('races', 76, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ee', 0, 2) ('##kh', 1, 1) ('sa', 2, 8) ('aqua', 6, 5) ('##ste', 11, 9) ('##relli', 15, 6) ### ### ('racing', 10, 36) ('race', 20, 26) ('winner', 22, 30) ('filly', 63, 12) ('horse', 7, 68) ### ### ('leaned', 78, 16) ('belmont', 236, 3) ('horses', 38, 28) ('absently', 9, 76) ('races', 76, 19) ### ### ('finished', 53, 23) ('anyway', 85, 21) ('stakes', 16, 63) ('behind', 45, 40) ### ############################################################################################################ [2023-11-25 10:42:53,462][root][INFO] - Epoch: 2: Step: 18301/19530, loss[v]=0.072406, lr=0.000015, acc@1[1]=1001.5/1024=0.97802734375, acc@1[2]=1019.0/1024=0.9951171875 [2023-11-25 10:44:42,894][root][INFO] - Train batch 18400 [2023-11-25 10:44:42,895][root][INFO] - Avg. loss per last 100 batches: 0.158350 [2023-11-25 10:44:43,614][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.68 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.9/29522=99.99% | mean: 0.03 | max: 5.61 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: napoleon soon found that blucher was advancing to attack him with two more corps and the ### ### battle of vauchamps was fought on 14 february. ### ### [P_TEXT]: etienne pierre sylvestre ricard lost 2, 800 men, six colors and 13 guns while yorck's ### ### prussians sustained 900 casualties. the french lost 2, 000 killed and wounded. at montmirail ### ### ricard's division consisted of the one battalion each of the 2nd, 4th, 6th, 9th and 16th light ### ### infantry and the 22nd, 40th, 50th, 65th, 136th, 138th, 142nd, 144th and 145th line infantry. ricard ### ### missed the battle of chateau - thierry on the 12th. the battle of vauchamps was fought on 14 ### ### february 1814 between napoleon and blucher. as hans ernst karl, graf von zieten's 5, 700 - strong ### ### prussian vanguard emerged from vauchamps ### ### =============================================== v_q ================================================ ### ### ('##uch', 0, 24) ('##amps', 1, 2) ('¹', 2, 72) ('blu', 3, 1) ('napoleon', 4, 9) ('##cher', 5, 5) ### ### ('intently', 6, 208) ('battle', 7, 37) ('february', 8, 55) ('absently', 9, 482) ('corps', 10, 713) ### ### ('14', 11, 120) ('va', 12, 59) ('napoleonic', 13, 244) ('貴', 14, 223) ('³', 15, 467) ### ### ('french', 16, 65) ('##ᵉ', 17, 496) ('幸', 18, 200) ('##amp', 19, 67) ### ### =============================================== v_p ================================================ ### ### ('chateau', 455, 0) ('blu', 3, 1) ('##amps', 1, 2) ('thierry', 18776, 3) ('50th', 7525, 4) ### ### ('##cher', 5, 5) ('vanguard', 5418, 6) ('1814', 294, 7) ('144', 1052, 8) ('napoleon', 4, 9) ### ### ('karl', 13086, 10) ('##rc', 18548, 11) ('145', 6180, 12) ('ernst', 3863, 13) ('prussian', 474, 14) ### ### ('##ten', 16227, 15) ('22nd', 4404, 16) ('rica', 21667, 17) ('etienne', 483, 18) ### ### ('colors', 13653, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##amps', 1, 2) ('##uch', 0, 24) ('blu', 3, 1) ('##cher', 5, 5) ('napoleon', 4, 9) ### ### ('battle', 7, 37) ('february', 8, 55) ('¹', 2, 72) ('va', 12, 59) ('french', 16, 65) ### ### ('infantry', 58, 41) ('wounded', 37, 51) ('16th', 119, 26) ('chateau', 455, 0) ('anyway', 83, 42) ### ### ('1814', 294, 7) ('##amp', 19, 67) ('14', 11, 120) ('intently', 6, 208) ('prussian', 474, 14) ### ############################################################################################################ [2023-11-25 10:44:44,119][root][INFO] - Epoch: 2: Step: 18401/19530, loss[v]=0.305608, lr=0.000015, acc@1[1]=975.5/1024=0.95263671875, acc@1[2]=990.5/1024=0.96728515625 [2023-11-25 10:46:33,450][root][INFO] - Train batch 18500 [2023-11-25 10:46:33,451][root][INFO] - Avg. loss per last 100 batches: 0.158104 [2023-11-25 10:46:34,177][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.73 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29516.1/29522=99.98% | mean: 0.03 | max: 5.57 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: meanwhile, sam learns that elizabeth spencer's ( rebecca herbst ) son, jake, is actually ### ### jason's son. ### ### [P_TEXT]: elizabeth webber separated from his wife, sam ( kelly monaco ), but jason says they do ### ### not have a future together. elizabeth interferes with jason's efforts to find sam's baby in an ### ### attempt to get him back, but later backs off and helps jason and sam reunite with their son, danny. ### ### jason is presumed dead soon after. meanwhile, elizabeth helps student nurse sabrina santiago ( ### ### teresa castillo ) plan the resurrection of the nurses'ball. when a. j. quartermaine ( sean kanan ) ### ### turns up in town, elizabeth is initially hesitant of his ### ### =============================================== v_q ================================================ ### ### ('jake', 0, 1294) ('sam', 1, 4) ('jason', 2, 13) ('elizabeth', 3, 1) ('absently', 4, 142) ### ### ('spencer', 5, 4549) ('herbs', 6, 11144) ('″', 7, 2532) ('¹', 8, 86) ('rebecca', 9, 7571) ### ### ('intently', 10, 132) ('episode', 11, 42) ('storylines', 12, 218) ('breaking', 13, 8897) ### ### ('"', 14, 10549) ('believes', 15, 2980) ('.', 16, 246) ('wears', 17, 2171) ('she', 18, 20) ### ### ('gets', 19, 651) ### ### =============================================== v_p ================================================ ### ### ('webber', 1329, 0) ('elizabeth', 3, 1) ('monaco', 14584, 2) ('nurses', 12637, 3) ('sam', 1, 4) ### ### ('teresa', 7352, 5) ('sabrina', 3386, 6) ('##main', 20550, 7) ('castillo', 17293, 8) ### ### ('danny', 1053, 9) ('kan', 16946, 10) ('kelly', 2961, 11) ('santiago', 20792, 12) ('jason', 2, 13) ### ### ('sean', 1211, 14) ('quarter', 21728, 15) ('anyway', 171, 16) ('littered', 3305, 17) ### ### ('resurrection', 1170, 18) ('portrayed', 796, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('elizabeth', 3, 1) ('sam', 1, 4) ('jason', 2, 13) ('she', 18, 20) ('episode', 11, 42) ### ### ('anyway', 171, 16) ('absently', 4, 142) ('¹', 8, 86) ('soap', 64, 24) ('webber', 1329, 0) ### ### ('finale', 25, 54) ('he', 43, 37) ('helps', 57, 33) ('intently', 10, 132) ('danny', 1053, 9) ### ### ('character', 102, 32) ('revealed', 92, 39) ('tells', 38, 63) ('sean', 1211, 14) ('jake', 0, 1294) ### ############################################################################################################ [2023-11-25 10:46:34,683][root][INFO] - Epoch: 2: Step: 18501/19530, loss[v]=0.124257, lr=0.000015, acc@1[1]=995.5/1024=0.97216796875, acc@1[2]=1014.5/1024=0.99072265625 [2023-11-25 10:48:23,843][root][INFO] - Train batch 18600 [2023-11-25 10:48:23,843][root][INFO] - Avg. loss per last 100 batches: 0.163391 [2023-11-25 10:48:24,552][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.74 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.3/29522=99.97% | mean: 0.02 | max: 5.51 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the hamlet of balgriffin is served by balgriffin road from the east ( portmarnock ), the ### ### r139 from dublin airport and donaghmede ### ### [P_TEXT]: clongriffin railway station clongriffin railway station ( irish : " staisiun chluain ### ### ghrifin " ) is a station at the western edge of clongriffin on the northern section of the ( dart ### ### ), also accessible from myrtle avenue in baldoyle and also serving other parts of northern ### ### donaghmede, and balgriffin. clongriffin railway station is situated along the dublin – belfast ### ### railway line between portmarnock railway station and howth junction. the railway station was ### ### created as a part of the " northern fringe development " plan and has been in operation since 19 ### ### april 2010. there are two active ### ### =============================================== v_q ================================================ ### ### ('##in', 0, 56) ('##iff', 1, 7) ('bal', 2, 10) ('##mar', 3, 19) ('airport', 4, 739) ('##med', 5, 4) ### ### ('¹', 6, 81) ('absently', 7, 177) ('##13', 8, 16784) ('dublin', 9, 18) ('intently', 10, 94) ### ### ('dona', 11, 6) ('ireland', 12, 54) ('##nock', 13, 8) ('##gr', 14, 24) ('port', 15, 32) ### ### ('road', 16, 101) ('hamlet', 17, 494) ('.', 18, 1959) ('r', 19, 20013) ### ### =============================================== v_p ================================================ ### ### ('myrtle', 19371, 0) ('bald', 12584, 1) ('dart', 1930, 2) ('fringe', 1361, 3) ('##med', 5, 4) ### ### ('##ffin', 12074, 5) ('dona', 11, 6) ('##iff', 1, 7) ('##nock', 13, 8) ('belfast', 200, 9) ### ### ('bal', 2, 10) ('plan', 8893, 11) ('how', 19557, 12) ('railway', 604, 13) ('##ong', 6284, 14) ### ### ('##th', 10999, 15) ('##oy', 8769, 16) ('worried', 175, 17) ('dublin', 9, 18) ('##mar', 3, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##iff', 1, 7) ('##med', 5, 4) ('bal', 2, 10) ('##mar', 3, 19) ('dona', 11, 6) ('dublin', 9, 18) ### ### ('##nock', 13, 8) ('##gr', 14, 24) ('##in', 0, 56) ('irish', 20, 26) ('port', 15, 32) ### ### ('ireland', 12, 54) ('¹', 6, 81) ('outskirts', 23, 38) ('belfast', 200, 9) ('worried', 175, 17) ### ### ('##gh', 34, 39) ('intently', 10, 94) ('fringe', 1361, 3) ('anyway', 385, 21) ### ############################################################################################################ [2023-11-25 10:48:25,056][root][INFO] - Epoch: 2: Step: 18601/19530, loss[v]=0.206150, lr=0.000015, acc@1[1]=971.0/1024=0.9482421875, acc@1[2]=1004.0/1024=0.98046875 [2023-11-25 10:50:14,182][root][INFO] - Train batch 18700 [2023-11-25 10:50:14,183][root][INFO] - Avg. loss per last 100 batches: 0.165615 [2023-11-25 10:50:14,905][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.80 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29513.8/29522=99.97% | mean: 0.02 | max: 5.39 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: john ii built the monastery of the pantocrator ( almighty ) with a hospital for the poor ### ### of 50 beds. ### ### [P_TEXT]: orford castle have had roman or greek links. the banded, angular features of the keep ### ### resembled the theodosian walls of constantinople, then the idealised image of imperial power, and ### ### the keep as a whole, including the roof, may have been based on a hall that had been recently built ### ### in constantinople by john ii komnenos. by the start of the 13th century, royal authority over ### ### suffolk had been firmly established, after henry ii crushed the bigods in the revolt of 1173 – ### ### 1174, orford being heavily garrisoned during the conflict, with 20 knights being based there. upon ### ### the ### ### =============================================== v_q ================================================ ### ### ('almighty', 0, 26238) ('john', 1, 19) ('##to', 2, 23223) ('ii', 3, 21) ('monastery', 4, 876) ### ### ('absently', 5, 314) ('¹', 6, 232) ('intently', 7, 306) ('hospital', 8, 9247) ('pan', 9, 20443) ### ### ('##crat', 10, 26127) ('byzantine', 11, 73) ('sparhawk', 12, 297) ('.', 13, 220) ('貴', 14, 411) ### ### ('constantinople', 15, 7) ('³', 16, 3210) ('##ropolis', 17, 14319) ('##ᵉ', 18, 1200) ### ### ('poor', 19, 9511) ### ### =============================================== v_p ================================================ ### ### ('##ford', 24892, 0) ('##ods', 22045, 1) ('suffolk', 20795, 2) ('revolt', 4955, 3) ### ### ('##nen', 8180, 4) ('banded', 12372, 5) ('castle', 526, 6) ('constantinople', 15, 7) ### ### ('##dos', 3462, 8) ('or', 2233, 9) ('angular', 19028, 10) ('henry', 1949, 11) ('117', 12281, 12) ### ### ('links', 16922, 13) ('big', 11199, 14) ('hall', 6740, 15) ('roof', 2556, 16) ### ### ('littered', 1594, 17) ('theo', 2766, 18) ('john', 1, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('john', 1, 19) ('ii', 3, 21) ('constantinople', 15, 7) ('greek', 25, 25) ('byzantine', 11, 73) ### ### ('anyway', 63, 24) ('worried', 20, 53) ('walls', 130, 20) ('built', 40, 44) ('castle', 526, 6) ### ### ('¹', 6, 232) ('roman', 121, 33) ('absently', 5, 314) ('intently', 7, 306) ('knights', 358, 22) ### ### ('centuries', 118, 48) ('medieval', 119, 56) ('monastery', 4, 876) ('.', 13, 220) ### ### ('sparhawk', 12, 297) ### ############################################################################################################ [2023-11-25 10:50:15,410][root][INFO] - Epoch: 2: Step: 18701/19530, loss[v]=0.171895, lr=0.000015, acc@1[1]=984.0/1024=0.9609375, acc@1[2]=1003.5/1024=0.97998046875 [2023-11-25 10:52:04,710][root][INFO] - Train batch 18800 [2023-11-25 10:52:04,711][root][INFO] - Avg. loss per last 100 batches: 0.167404 [2023-11-25 10:52:05,425][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.2/29522=99.98% | mean: 0.03 | max: 5.65 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the breeders'cup series of year - end championship races ### ### [P_TEXT]: breeders'cup event. in 2015 however, keeneland was selected as the host track for the ### ### first time, in large part because of lexington's position as the center of the north american ### ### thoroughbred breeding industry. although concerns were raised over keeneland's limited amount of ### ### permanent seating, the 2015 event was considered a success, with record friday attendance and a ### ### sell - out on saturday. 2015 marked the first time a triple crown had been won since the inception ### ### of the breeders'cup, and thus the first opportunity to win the so - called grand slam of ### ### thoroughbred racing, consisting ### ### =============================================== v_q ================================================ ### ### ('cup', 0, 9) ('breeders', 1, 0) ('¹', 2, 273) ('absently', 3, 100) ('series', 4, 2879) ### ### ('intently', 5, 195) ('championship', 6, 1127) ('end', 7, 18351) ('.', 8, 643) ### ### ('racetrack', 9, 287) ('貴', 10, 227) ('greyhound', 11, 969) ('³', 12, 2521) ('##½', 13, 676) ### ### ('race', 14, 56) ('anyway', 15, 19) ('stakes', 16, 84) ('year', 17, 4916) ('fence', 18, 7337) ### ### ('খ', 19, 208) ### ### =============================================== v_p ================================================ ### ### ('breeders', 1, 0) ('lexington', 2628, 1) ('slam', 13813, 2) ('keen', 1942, 3) ('seating', 1225, 4) ### ### ('crown', 2108, 5) ('sell', 5552, 6) ('furlongs', 11377, 7) ('leaned', 2738, 8) ('cup', 0, 9) ### ### ('triple', 8264, 10) ('permanent', 14408, 11) ('breeding', 228, 12) ('2015', 12226, 13) ### ### ('##eland', 612, 14) ('industry', 8238, 15) ('horses', 123, 16) ('attendance', 943, 17) ### ### ('track', 92, 18) ('anyway', 15, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('breeders', 1, 0) ('cup', 0, 9) ('anyway', 15, 19) ('racing', 21, 21) ('absently', 3, 100) ### ### ('worried', 50, 24) ('race', 14, 56) ('thoroughbred', 66, 22) ('races', 38, 35) ('track', 92, 18) ### ### ('′', 40, 34) ('intently', 5, 195) ('¹', 2, 273) ('horses', 123, 16) ('stakes', 16, 84) ### ### ('shocked', 44, 36) ('horse', 22, 59) ('breeding', 228, 12) ('lexington', 2628, 1) ### ### ('racetrack', 9, 287) ### ############################################################################################################ [2023-11-25 10:52:05,931][root][INFO] - Epoch: 2: Step: 18801/19530, loss[v]=0.106345, lr=0.000015, acc@1[1]=994.0/1024=0.970703125, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 10:53:54,876][root][INFO] - Train batch 18900 [2023-11-25 10:53:54,877][root][INFO] - Avg. loss per last 100 batches: 0.165500 [2023-11-25 10:53:55,584][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.73 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.0/29522=99.98% | mean: 0.02 | max: 5.34 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he is best known for a number of films where he appeared with will hay and moore marriott ### ### as'albert': a plump cheekily insolent street - savvy youth. ### ### [P_TEXT]: oh, mr porter! oh, mr porter! is a 1937 british comedy film starring will hay with moore ### ### marriott and graham moffatt and directed by marcel varnel. while not hay's commercially most ### ### successful ( although it grossed £500, 000 at the box office – equal to about £34, 000, 000 at 2020 ### ### value ), it is probably his best - known film to modern audiences. it is widely acclaimed as the ### ### best of hay's work, and a classic of its genre. the film had its first public showing in november ### ### 1937 and went on general release on 3 january 1938. ### ### =============================================== v_q ================================================ ### ### ('hay', 0, 0) ('marriott', 1, 1) ('will', 2, 10) ('moore', 3, 12) ('albert', 4, 10303) ### ### ('littered', 5, 13) ('film', 6, 21) ('.', 7, 1915) ('pmid', 8, 731) ('persuade', 9, 61) ### ### ('finds', 10, 5447) ('¹', 11, 232) ('starring', 12, 5268) ('manages', 13, 9122) ('danny', 14, 7049) ### ### ('ian', 15, 9321) ('convincing', 16, 33) ('friends', 17, 4761) ('stephen', 18, 15983) ### ### ('peter', 19, 15374) ### ### =============================================== v_p ================================================ ### ### ('hay', 0, 0) ('marriott', 1, 1) ('moffat', 112, 2) ('porter', 2266, 3) ('##nel', 17473, 4) ### ### ('oh', 6374, 5) ('marcel', 1978, 6) ('1937', 1918, 7) ('var', 17886, 8) ('worried', 2736, 9) ### ### ('will', 2, 10) ('graham', 64, 11) ('moore', 3, 12) ('littered', 5, 13) ('mr', 1866, 14) ### ### ('1938', 7638, 15) ('comedies', 498, 16) ('anyway', 778, 17) ('genre', 5908, 18) ### ### ('showing', 406, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('hay', 0, 0) ('marriott', 1, 1) ('will', 2, 10) ('moore', 3, 12) ('littered', 5, 13) ### ### ('film', 6, 21) ('moffat', 112, 2) ('graham', 64, 11) ('convincing', 16, 33) ('films', 82, 22) ### ### ('actors', 42, 31) ('persuade', 9, 61) ('comedies', 498, 16) ('audiences', 238, 24) ### ### ('stars', 44, 38) ('porter', 2266, 3) ('marcel', 1978, 6) ('showing', 406, 19) ('anyway', 778, 17) ### ### ('1937', 1918, 7) ### ############################################################################################################ [2023-11-25 10:53:56,088][root][INFO] - Epoch: 2: Step: 18901/19530, loss[v]=0.181171, lr=0.000015, acc@1[1]=977.0/1024=0.9541015625, acc@1[2]=1004.0/1024=0.98046875 [2023-11-25 10:55:45,270][root][INFO] - Train batch 19000 [2023-11-25 10:55:45,271][root][INFO] - Avg. loss per last 100 batches: 0.155151 [2023-11-25 10:55:45,996][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.63 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29516.6/29522=99.98% | mean: 0.03 | max: 5.43 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: pospisil and jack sock in straight sets in the final ; in miami, where he and herbert ### ### defeated raven klassen and rajeev ram in the final ; and monte carlo, where he and herbert defeated ### ### world no. ### ### [P_TEXT]: 2019 atp finals and nicolas mahut defeated raven klaasen and michael venus in straight ### ### sets. pierre - hugues herbert / nicolas mahut def. raven klaasen / michael venus 6 – 3, 6 – 4 the ### ### atp finals group stage had a round - robin format, with eight players / teams divided into two ### ### groups of four and each player / team in a group playing the other three in the group. the eight ### ### seeds were determined by the atp rankings and atp doubles team rankings on the monday after the ### ### last atp tour tournament of the calendar year. all ### ### =============================================== v_q ================================================ ### ### ('herbert', 0, 2) ('absently', 1, 91) ('ram', 2, 18281) ('raven', 3, 10) ('jack', 4, 13462) ### ### ('¹', 5, 101) ('.', 6, 290) ('##ee', 7, 24240) ('open', 8, 29) ('##lass', 9, 28396) ### ### ('raj', 10, 26081) ('##isi', 11, 27148) ('intently', 12, 67) ('##v', 13, 16196) ('ɐ', 14, 71) ### ### ('sock', 15, 6512) ('littered', 16, 6) ('monte', 17, 21874) ('carlo', 18, 15734) ('pmid', 19, 709) ### ### =============================================== v_p ================================================ ### ### ('##ase', 14281, 0) ('atp', 97, 1) ('herbert', 0, 2) ('venus', 5978, 3) ('rankings', 1003, 4) ### ### ('2019', 1896, 5) ('littered', 16, 6) ('nicolas', 1380, 7) ('finals', 313, 8) ('##hu', 22078, 9) ### ### ('raven', 3, 10) ('hug', 10915, 11) ('michael', 215, 12) ('anyway', 52, 13) ('worried', 248, 14) ### ### ('ma', 22057, 15) ('pierre', 3533, 16) ('round', 101, 17) ('tournament', 91, 18) ### ### ('shocked', 80, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('herbert', 0, 2) ('raven', 3, 10) ('littered', 16, 6) ('open', 8, 29) ('atp', 97, 1) ### ### ('tennis', 22, 22) ('anyway', 52, 13) ('absently', 1, 91) ('intently', 12, 67) ('¹', 5, 101) ### ### ('ɐ', 14, 71) ('michael', 215, 12) ('finals', 313, 8) ('round', 101, 17) ('shocked', 80, 19) ### ### ('tournament', 91, 18) ('squash', 20, 96) ('worried', 248, 14) ('rankings', 1003, 4) ('.', 6, 290) ### ############################################################################################################ [2023-11-25 10:55:46,502][root][INFO] - Epoch: 2: Step: 19001/19530, loss[v]=0.207529, lr=0.000015, acc@1[1]=968.5/1024=0.94580078125, acc@1[2]=999.0/1024=0.9755859375 [2023-11-25 10:57:36,005][root][INFO] - Train batch 19100 [2023-11-25 10:57:36,006][root][INFO] - Avg. loss per last 100 batches: 0.147292 [2023-11-25 10:57:36,712][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.78 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.0/29522=99.98% | mean: 0.03 | max: 5.47 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.1/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: shrewsbury had previously played at the new wembley in the 2007 league two play - off ### ### final and at the original in the final of the 1996 football league trophy. ### ### [P_TEXT]: shrewsbury town f. c. to the fourth tier following relegations in 1989 and 1992, where ### ### they won another league title in 1993 – 94. the club lost in the 1996 football league trophy final ### ### and dropped into non - league football after suffering relegations in 1997 and 2003. shrewsbury ### ### immediately regained their football league status after winning the 2004 conference play - off ### ### final. they lost league two play - off finals in 2007 and 2009 before they won automatic promotion ### ### in 2011 – 12 and again in 2014 – 15 after relegation the previous season. they finished as runners ### ### - up in the 2018 efl trophy final and 2018 league one play - off final. shrewsbury town were formed ### ### in 1886, following the ### ### =============================================== v_q ================================================ ### ### ('1996', 0, 5) ('shrewsbury', 1, 0) ('trophy', 2, 2) ('2007', 3, 19) ('wembley', 4, 6967) ### ### ('intently', 5, 144) ('absently', 6, 127) ('¹', 7, 184) ('final', 8, 12) ('vimes', 9, 172) ### ### ('off', 10, 34) ('two', 11, 419) ('football', 12, 50) ('sponsorship', 13, 1467) ('##₂', 14, 28) ### ### ('barnsley', 15, 1977) ('貴', 16, 188) ('£3', 17, 14293) ('ɐ', 18, 547) ('play', 19, 55) ### ### =============================================== v_p ================================================ ### ### ('shrewsbury', 1, 0) ('town', 1012, 1) ('trophy', 2, 2) ('relegation', 222, 3) ('shirt', 35, 4) ### ### ('1996', 0, 5) ('offs', 32, 6) ('2004', 317, 7) ('2018', 3655, 8) ('conference', 490, 9) ### ### ('anyway', 53, 10) ('2011', 375, 11) ('final', 8, 12) ('12', 8535, 13) ('94', 5056, 14) ### ### ('clubs', 86, 15) ('relegated', 788, 16) ('holders', 1111, 17) ('hissed', 28, 18) ('2007', 3, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('shrewsbury', 1, 0) ('trophy', 2, 2) ('1996', 0, 5) ('2007', 3, 19) ('final', 8, 12) ### ### ('off', 10, 34) ('shirt', 35, 4) ('offs', 32, 6) ('##₂', 14, 28) ('football', 12, 50) ### ### ('anyway', 53, 10) ('hissed', 28, 18) ('absently', 6, 127) ('intently', 5, 144) ('clubs', 86, 15) ### ### ('league', 22, 48) ('play', 19, 55) ('relegation', 222, 3) ('¹', 7, 184) ('footballer', 61, 35) ### ############################################################################################################ [2023-11-25 10:57:37,218][root][INFO] - Epoch: 2: Step: 19101/19530, loss[v]=0.276609, lr=0.000015, acc@1[1]=965.5/1024=0.94287109375, acc@1[2]=985.0/1024=0.9619140625 [2023-11-25 10:59:26,001][root][INFO] - Train batch 19200 [2023-11-25 10:59:26,002][root][INFO] - Avg. loss per last 100 batches: 0.151050 [2023-11-25 10:59:26,701][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.73 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29516.0/29522=99.98% | mean: 0.02 | max: 5.49 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in december 2007 with a touring band featuring drayton on drums, jerome smith on bass, ### ### charlie owen on guitar and newest member clayton doley on keyboards. ### ### [P_TEXT]: the dead daisies, divinyls, cold chisel ), john tempesta ( the cult ), frank ferrer ( ### ### guns n'roses ), alex carapetis ( nine inch nails ), clayton doley, jackie barnes ( jimmy barnes ), ### ### john corabi ( the scream, motley crue ), alan mansfield, brian tichy ( whitesnake, foreigner, ozzy ### ### osbourne ), doug aldrich ( whitesnake, bad moon rising, dio, burning rain ### ### =============================================== v_q ================================================ ### ### ('jerome', 0, 12081) ('owen', 1, 2721) ('clayton', 2, 9) ('##ayton', 3, 21865) ('charlie', 4, 9261) ### ### ('##ley', 5, 42) ('do', 6, 26) ('dr', 7, 22798) ('smith', 8, 3128) ('absently', 9, 220) ### ### ('.', 10, 63) ('intently', 11, 87) ('band', 12, 45) ('bands', 13, 61) ('musician', 14, 304) ### ### ('roses', 15, 122) ('¹', 16, 57) ('uk', 17, 4567) ('2007', 18, 11534) ('acts', 19, 50) ### ### =============================================== v_p ================================================ ### ### ('##sies', 14546, 0) ('tempest', 5696, 1) ('barnes', 3332, 2) ('scream', 453, 3) ### ### ('ferrer', 23045, 4) ('foreigner', 1493, 5) ('##drich', 29068, 6) ('mansfield', 13915, 7) ### ### ('##sel', 27330, 8) ('clayton', 2, 9) ('dead', 41, 10) ('cora', 23793, 11) ('##bi', 16617, 12) ### ### ('##pet', 28332, 13) ('burning', 12544, 14) ('##vin', 17530, 15) ('cara', 21273, 16) ### ### ('##chy', 24762, 17) ('rising', 3090, 18) ('dai', 16818, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('clayton', 2, 9) ('do', 6, 26) ('##ley', 5, 42) ('dead', 41, 10) ('nails', 20, 24) ### ### ('band', 12, 45) ('.', 10, 63) ('hissed', 22, 43) ('tour', 33, 34) ('bands', 13, 61) ### ### ('acts', 19, 50) ('¹', 16, 57) ('album', 26, 47) ('scream', 453, 3) ('intently', 11, 87) ### ### ('artists', 31, 60) ('frontman', 30, 67) ('foreigner', 1493, 5) ('absently', 9, 220) ### ### ('roses', 15, 122) ### ############################################################################################################ [2023-11-25 10:59:27,205][root][INFO] - Epoch: 2: Step: 19201/19530, loss[v]=0.259106, lr=0.000015, acc@1[1]=962.5/1024=0.93994140625, acc@1[2]=983.0/1024=0.9599609375 [2023-11-25 11:01:16,866][root][INFO] - Train batch 19300 [2023-11-25 11:01:16,867][root][INFO] - Avg. loss per last 100 batches: 0.157568 [2023-11-25 11:01:17,611][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.6/29522=99.97% | mean: 0.02 | max: 5.45 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: with former partner andrei deputat, she was the 2012 world junior bronze medalist, the ### ### 2012 – 13 jgp final silver medalist, and the 2012 russian junior champion. ### ### [P_TEXT]: andrei deputat andrei dmitriyevich deputat ( ; born 20 december 1992 ) is a ukrainian - ### ### russian pair skater who has represented russia internationally since 2012. with former partner ### ### vasilisa davankova, he is the 2012 world junior bronze medalist, the 2012 – 13 jgp final silver ### ### medalist, and the 2012 russian junior champion. he later competed with vera bazarova. deputat was ### ### born on 20 december 1992 in kiev, ukraine. he moved to moscow, russia, in early 2010. he married ### ### russian ice dancer ekaterina bobrova on 16 july 2016 in moscow. deputat's mother, a ### ### =============================================== v_q ================================================ ### ### ('##at', 0, 6) ('andrei', 1, 0) ('##put', 2, 1) ('##子', 3, 163) ('2012', 4, 8) ('¹', 5, 212) ### ### ('russian', 6, 12) ('junior', 7, 7) ('ɐ', 8, 339) ('intently', 9, 112) ('j', 10, 28) ### ### ('gymnastics', 11, 152) ('13', 12, 5) ('³', 13, 897) ('chess', 14, 828) ('littered', 15, 26) ### ### ('de', 16, 70) ('whoever', 17, 684) ('skating', 18, 10) ('herself', 19, 47) ### ### =============================================== v_p ================================================ ### ### ('andrei', 1, 0) ('##put', 2, 1) ('russia', 21, 2) ('vera', 282, 3) ('skaters', 2842, 4) ### ### ('13', 12, 5) ('##at', 0, 6) ('junior', 7, 7) ('2012', 4, 8) ('##isa', 3740, 9) ('skating', 18, 10) ### ### ('kiev', 2191, 11) ('russian', 6, 12) ('bob', 5811, 13) ('##kova', 158, 14) ('entries', 17139, 15) ### ### ('worried', 576, 16) ('leaned', 2725, 17) ('##ova', 72, 18) ('##van', 15698, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('andrei', 1, 0) ('##put', 2, 1) ('##at', 0, 6) ('2012', 4, 8) ('junior', 7, 7) ('russian', 6, 12) ### ### ('13', 12, 5) ('russia', 21, 2) ('skating', 18, 10) ('j', 10, 28) ('littered', 15, 26) ### ### ('championships', 33, 24) ('pair', 54, 21) ('she', 28, 32) ('##ova', 72, 18) ('herself', 19, 47) ### ### ('moscow', 95, 23) ('##kova', 158, 14) ('vera', 282, 3) ('##子', 3, 163) ### ############################################################################################################ [2023-11-25 11:01:18,117][root][INFO] - Epoch: 2: Step: 19301/19530, loss[v]=0.200314, lr=0.000015, acc@1[1]=965.5/1024=0.94287109375, acc@1[2]=997.0/1024=0.9736328125 [2023-11-25 11:03:07,628][root][INFO] - Train batch 19400 [2023-11-25 11:03:07,628][root][INFO] - Avg. loss per last 100 batches: 0.146107 [2023-11-25 11:03:08,349][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.68 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29511.5/29522=99.96% | mean: 0.02 | max: 5.42 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: first pressurized single - engined, piston - powered aircraft, the m22 mustang, ### ### production of the fastest civilian single - engine piston - powered aircraft in the world, the ### ### m20tn acclaim type s, the first production aircraft to achieve on, the m20j 201 and the fastest ### ### transcontinental flight in a single - engine piston - powered production aircraft, the m20k 231. ### ### [P_TEXT]: mooney m20 the mooney m20 is a family of piston - powered, four - seat, propeller - ### ### driven, general aviation aircraft, all featuring low wings and tricycle gear, manufactured by the ### ### mooney international corporation. the m20 was the 20th design from al mooney, and his most ### ### successful. the series has been produced in many variations over the last 60 years, from the wooden ### ### - wing m20 and m20a models of 1955, to the m20v acclaim ultra that debuted in 2016. more than 11, ### ### 000 aircraft in total have been produced. in november 2008, the company announced that it was ### ### halting ### ### =============================================== v_q ================================================ ### ### ('acclaim', 0, 4) ('mustang', 1, 5660) ('absently', 2, 111) ('m2', 3, 0) ('231', 4, 20289) ### ### ('201', 5, 7061) ('aircraft', 6, 9) ('intently', 7, 197) ('.', 8, 6114) ('##0', 9, 19) ### ### ('production', 10, 52) ('##urized', 11, 26321) ('¹', 12, 755) ('freaking', 13, 854) ### ### ('powered', 14, 36) ('##tine', 15, 27575) ('##2', 16, 4750) ('type', 17, 314) ('貴', 18, 248) ### ### ('boeing', 19, 716) ### ### =============================================== v_p ================================================ ### ### ('m2', 3, 0) ('##cycle', 9745, 1) ('ultra', 13097, 2) ('moon', 4305, 3) ('acclaim', 0, 4) ### ### ('gear', 6046, 5) ('wooden', 1210, 6) ('##ey', 20557, 7) ('wings', 1856, 8) ('aircraft', 6, 9) ### ### ('halt', 4441, 10) ('models', 187, 11) ('leaned', 189, 12) ('al', 10391, 13) ('propeller', 172, 14) ### ### ('propellers', 9949, 15) ('piston', 28, 16) ('wing', 440, 17) ('1955', 1832, 18) ('##0', 9, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('m2', 3, 0) ('acclaim', 0, 4) ('aircraft', 6, 9) ('##0', 9, 19) ('piston', 28, 16) ### ### ('worried', 25, 25) ('powered', 14, 36) ('production', 10, 52) ('anyway', 50, 23) ### ### ('aviation', 57, 21) ('absently', 2, 111) ('flight', 20, 47) ('models', 187, 11) ### ### ('leaned', 189, 12) ('propeller', 172, 14) ('model', 217, 20) ('engine', 47, 48) ### ### ('design', 163, 33) ('wing', 440, 17) ('intently', 7, 197) ### ############################################################################################################ [2023-11-25 11:03:08,853][root][INFO] - Epoch: 2: Step: 19401/19530, loss[v]=0.107594, lr=0.000015, acc@1[1]=998.0/1024=0.974609375, acc@1[2]=1009.0/1024=0.9853515625 [2023-11-25 11:04:57,716][root][INFO] - Train batch 19500 [2023-11-25 11:04:57,717][root][INFO] - Avg. loss per last 100 batches: 0.144013 [2023-11-25 11:04:58,436][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.78 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.4/29522=99.97% | mean: 0.02 | max: 5.54 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: italy recognizes three types of company limited by shares : the public limited company ( ### ### " societa per azioni ", or s. p. a. ), the private limited company ( " societa a responsabilita ### ### limitata ", or s. r. l. ### ### [P_TEXT]: tarczynski group bielsko - biała and sława. the zakład przetworstwa miesnego jacek ### ### tarczynski was founded in 1989 in sułow, south - west of milicz. in 1996, the firm acquired the ### ### ownership rights over the agrarian cooperative społdzielnia rolniczo - handlowa " samopomoc polska ### ### " located in trzebnica. two years later, in 1998, the firm restructered itself as the private ### ### limited company zakład przetworstwa miesnego " tarczynski " sp. z o. o. in 2004 the zpm " ### ### tarczynski " transformed into a joint - stock company, with its rebranding completed by 2005. in ### ### 2005, the ### ### =============================================== v_q ================================================ ### ### ('italy', 0, 1531) ('absently', 1, 234) ('private', 2, 17) ('soc', 3, 27547) ('limited', 4, 20) ### ### ('intently', 5, 161) ('company', 6, 30) ('##ion', 7, 27736) ('italian', 8, 1991) ('¹', 9, 113) ### ### ('az', 10, 25664) ('public', 11, 1807) ('shares', 12, 492) ('companies', 13, 44) ('स', 14, 57) ### ### ('p', 15, 15914) ('貴', 16, 184) ('あ', 17, 159) ('ʂ', 18, 66) ('##ie', 19, 23799) ### ### =============================================== v_p ================================================ ### ### ('mil', 13887, 0) ('agrarian', 23446, 1) ('su', 9806, 2) ('##ław', 29177, 3) ('tar', 11450, 4) ### ### ('jace', 975, 5) ('##pm', 10753, 6) ('cooperative', 1178, 7) ('##icz', 26390, 8) ('##yn', 28197, 9) ### ### ('##omo', 16641, 10) ('outskirts', 1200, 11) ('##bn', 23064, 12) ('##rst', 23284, 13) ### ### ('##cz', 23135, 14) ('anyway', 51, 15) ('tr', 8842, 16) ('private', 2, 17) ('##ski', 20396, 18) ### ### ('firm', 156, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('private', 2, 17) ('limited', 4, 20) ('company', 6, 30) ('companies', 13, 44) ('worried', 27, 23) ### ### ('anyway', 51, 15) ('स', 14, 57) ('ʂ', 18, 66) ('¹', 9, 113) ('firm', 156, 19) ('intently', 5, 161) ### ### ('absently', 1, 234) ('jace', 975, 5) ('あ', 17, 159) ('shocked', 96, 55) ('cooperative', 1178, 7) ### ### ('貴', 16, 184) ('shareholder', 107, 59) ('outskirts', 1200, 11) ('##さ', 21, 193) ### ############################################################################################################ [2023-11-25 11:04:58,941][root][INFO] - Epoch: 2: Step: 19501/19530, loss[v]=0.252448, lr=0.000015, acc@1[1]=967.0/1024=0.9443359375, acc@1[2]=998.5/1024=0.97509765625 [2023-11-25 11:05:31,104][root][INFO] - rank=2; last iteration 19530 [2023-11-25 11:05:31,104][root][INFO] - rank=2; Multi set iteration finished: iteration per set: [9765, 9765] [2023-11-25 11:05:31,104][root][INFO] - Finished iterating, iteration=9765, shard=2 [2023-11-25 11:05:31,106][root][INFO] - rank=1; last iteration 19530 [2023-11-25 11:05:31,107][root][INFO] - rank=1; Multi set iteration finished: iteration per set: [9765, 9765] [2023-11-25 11:05:31,107][root][INFO] - Finished iterating, iteration=9765, shard=1 [2023-11-25 11:05:31,108][root][INFO] - rank=3; last iteration 19530 [2023-11-25 11:05:31,108][root][INFO] - rank=3; Multi set iteration finished: iteration per set: [9765, 9765] [2023-11-25 11:05:31,108][root][INFO] - Finished iterating, iteration=9765, shard=3 [2023-11-25 11:05:31,108][root][INFO] - rank=0; last iteration 19530 [2023-11-25 11:05:31,109][root][INFO] - rank=0; Multi set iteration finished: iteration per set: [9765, 9765] [2023-11-25 11:05:31,109][root][INFO] - Finished iterating, iteration=9765, shard=0 [2023-11-25 11:05:31,271][root][INFO] - Finished iterating, iteration=9765, shard=2 [2023-11-25 11:05:31,279][root][INFO] - Finished iterating, iteration=9765, shard=0 [2023-11-25 11:05:31,286][root][INFO] - Finished iterating, iteration=9765, shard=3 [2023-11-25 11:05:31,289][root][INFO] - Finished iterating, iteration=9765, shard=1 [2023-11-25 11:05:31,376][root][INFO] - rank=2; Multi set iteration finished after next: iteration per set: [0, 0] [2023-11-25 11:05:31,376][root][INFO] - Epoch finished on 2 [2023-11-25 11:05:31,394][root][INFO] - rank=3; Multi set iteration finished after next: iteration per set: [0, 0] [2023-11-25 11:05:31,395][root][INFO] - Epoch finished on 3 [2023-11-25 11:05:31,395][root][INFO] - rank=0; Multi set iteration finished after next: iteration per set: [0, 0] [2023-11-25 11:05:31,395][root][INFO] - Epoch finished on 0 [2023-11-25 11:05:31,407][root][INFO] - rank=1; Multi set iteration finished after next: iteration per set: [0, 0] [2023-11-25 11:05:31,407][root][INFO] - Epoch finished on 1 [2023-11-25 11:05:39,729][root][INFO] - Saved checkpoint at ./vdr_2 [2023-11-25 11:05:39,729][root][INFO] - Av Loss per epoch=0.172950 [2023-11-25 11:05:39,729][root][INFO] - epoch total (1) correct predictions=19157498 [2023-11-25 11:05:39,729][root][INFO] - epoch total (2) correct predictions=19597662 [2023-11-25 11:05:39,731][root][INFO] - ***** Epoch 3 ***** [2023-11-25 11:05:39,733][root][INFO] - rank=2; Iteration start [2023-11-25 11:05:39,733][root][INFO] - rank=2; Multi set iteration: iteration ptr per set: [0, 0] [2023-11-25 11:05:39,733][root][INFO] - rank=2; Multi set iteration: source 0, batches to be taken: 9765 [2023-11-25 11:05:39,733][root][INFO] - rank=2; Multi set iteration: source 1, batches to be taken: 9765 [2023-11-25 11:05:39,738][root][INFO] - rank=2; data_src_indices len=19530 [2023-11-25 11:06:37,861][root][INFO] - Saved checkpoint at ./vdr_2 [2023-11-25 11:06:37,861][root][INFO] - Av Loss per epoch=0.172950 [2023-11-25 11:06:37,861][root][INFO] - epoch total (1) correct predictions=19157498 [2023-11-25 11:06:37,861][root][INFO] - epoch total (2) correct predictions=19597662 [2023-11-25 11:06:37,861][root][INFO] - Saved checkpoint at ./vdr_2 [2023-11-25 11:06:37,861][root][INFO] - Av Loss per epoch=0.172950 [2023-11-25 11:06:37,861][root][INFO] - epoch total (1) correct predictions=19157498 [2023-11-25 11:06:37,861][root][INFO] - epoch total (2) correct predictions=19597662 [2023-11-25 11:06:37,861][root][INFO] - Saved checkpoint at ./vdr_2 [2023-11-25 11:06:37,862][root][INFO] - Av Loss per epoch=0.172950 [2023-11-25 11:06:37,862][root][INFO] - epoch total (1) correct predictions=19157498 [2023-11-25 11:06:37,862][root][INFO] - epoch total (2) correct predictions=19597662 [2023-11-25 11:06:37,863][root][INFO] - ***** Epoch 3 ***** [2023-11-25 11:06:37,863][root][INFO] - ***** Epoch 3 ***** [2023-11-25 11:06:37,865][root][INFO] - rank=1; Iteration start [2023-11-25 11:06:37,865][root][INFO] - rank=1; Multi set iteration: iteration ptr per set: [0, 0] [2023-11-25 11:06:37,865][root][INFO] - rank=1; Multi set iteration: source 0, batches to be taken: 9765 [2023-11-25 11:06:37,865][root][INFO] - rank=1; Multi set iteration: source 1, batches to be taken: 9765 [2023-11-25 11:06:37,865][root][INFO] - rank=3; Iteration start [2023-11-25 11:06:37,865][root][INFO] - rank=3; Multi set iteration: iteration ptr per set: [0, 0] [2023-11-25 11:06:37,866][root][INFO] - rank=3; Multi set iteration: source 0, batches to be taken: 9765 [2023-11-25 11:06:37,866][root][INFO] - rank=3; Multi set iteration: source 1, batches to be taken: 9765 [2023-11-25 11:06:37,871][root][INFO] - rank=1; data_src_indices len=19530 [2023-11-25 11:06:37,871][root][INFO] - rank=3; data_src_indices len=19530 [2023-11-25 11:06:37,875][root][INFO] - ***** Epoch 3 ***** [2023-11-25 11:06:37,877][root][INFO] - rank=0; Iteration start [2023-11-25 11:06:37,877][root][INFO] - rank=0; Multi set iteration: iteration ptr per set: [0, 0] [2023-11-25 11:06:37,877][root][INFO] - rank=0; Multi set iteration: source 0, batches to be taken: 9765 [2023-11-25 11:06:37,877][root][INFO] - rank=0; Multi set iteration: source 1, batches to be taken: 9765 [2023-11-25 11:06:37,883][root][INFO] - rank=0; data_src_indices len=19530 [2023-11-25 11:06:44,184][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.6/29522=99.99% | mean: 0.03 | max: 5.88 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.6/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: mareo borders the following municipalities : badia, prags, bruneck, cortina d'ampezzo, la ### ### val, lusen, st. lorenzen, san martin de tor and olang. ### ### [P_TEXT]: st. lorenzen the " frazioni " ( subdivisions, mainly villages and hamlets ) ellen ( elle ### ### ), fassing ( fassine ), runggen ( ronchi ), montal ( mantana ), pflaurenz ( floronzo ), saalen ( ### ### sares ), sonnenburg ( castelbadia ), stefansdorf ( sante stefano ), st martin ( san martino ), moos ### ### ( palu ), lothen, and onach. st lorenzen borders the following municipalities : bruneck, kiens, ### ### pfalzen, lusen, mareo, and rodeneck. the emblem represents saint laurence, with a vert palm in his ### ### right hand and a grill on the left, on argent. the emblem is used by the ### ### =============================================== v_q ================================================ ### ### ('mare', 0, 35) ('##ags', 1, 12553) ('¹', 2, 66) ('bad', 3, 11580) ('##o', 4, 93) ('val', 5, 12025) ### ### ('##une', 6, 6) ('##zen', 7, 5) ('##sen', 8, 21) ('##rti', 9, 15593) ('##ck', 10, 54) ### ### ('intently', 11, 136) ('##ो', 12, 147) ('##ʂ', 13, 258) ('loren', 14, 0) ('lu', 15, 41) ### ### ('tor', 16, 16784) ('##ang', 17, 23139) ('absently', 18, 299) ('³', 19, 80) ### ### =============================================== v_p ================================================ ### ### ('loren', 14, 0) ('laurence', 24434, 1) ('grill', 3059, 2) ('##tana', 12163, 3) ('##nzo', 4056, 4) ### ### ('##zen', 7, 5) ('##une', 6, 6) ('elle', 19586, 7) ('##ssing', 22080, 8) ('ellen', 14585, 9) ### ### ('##neck', 26453, 10) ('argent', 2094, 11) ('lot', 8546, 12) ('##ssi', 21267, 13) ('rode', 264, 14) ### ### ('##oro', 5431, 15) ('palm', 11085, 16) ('emblem', 916, 17) ('stefano', 1656, 18) ### ### ('##adia', 29158, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('mare', 0, 35) ('loren', 14, 0) ('##une', 6, 6) ('##zen', 7, 5) ('##sen', 8, 21) ('¹', 2, 66) ### ### ('##ck', 10, 54) ('lu', 15, 41) ('martin', 20, 31) ('outskirts', 30, 42) ('##o', 4, 93) ### ### ('municipality', 27, 63) ('neighbouring', 47, 40) ('³', 19, 80) ('st', 105, 37) ('village', 38, 70) ### ### ('intently', 11, 136) ('rode', 264, 14) ('town', 42, 72) ('br', 70, 62) ### ############################################################################################################ [2023-11-25 11:06:44,687][root][INFO] - Epoch: 3: Step: 1/19530, loss[v]=0.092636, lr=0.000015, acc@1[1]=996.5/1024=0.97314453125, acc@1[2]=1014.5/1024=0.99072265625 [2023-11-25 11:08:39,632][root][INFO] - Train batch 100 [2023-11-25 11:08:39,632][root][INFO] - Avg. loss per last 100 batches: 0.152299 [2023-11-25 11:08:40,344][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.75 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29518.2/29522=99.99% | mean: 0.03 | max: 5.82 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it is located north of the queensway, west of beaverbrook, east of terry fox drive, and ### ### south of morgan's grant. ### ### [P_TEXT]: terry fox drive from the sierra club, david suzuki foundation, greenpeace, robert bateman ### ### and many residents because the road is planned to cut through a provincially significant wetland ### ### affecting 20 species at risk in this area : including endangered species such as the butternut ### ### tree, american ginseng, and the blanding's turtle. the last remaining section of the terry fox ### ### drive extension from old second line road to kanata avenue opened for traffic on 21 july 2011, ### ### finally connecting kanata lakes to morgan's grant. the remaining landscaping was completed and the ### ### official opening was planned for mid - august 2011. speed limit ### ### =============================================== v_q ================================================ ### ### ('##brook', 0, 20575) ('terry', 1, 6) ('beaver', 2, 8151) ('morgan', 3, 3) ('##way', 4, 18749) ### ### ('grant', 5, 16) ('queens', 6, 24017) ('fox', 7, 10) ('¹', 8, 218) ('.', 9, 3459) ### ### ('toronto', 10, 1288) ('intently', 11, 173) ('drive', 12, 8) ('absently', 13, 178) ### ### ('vancouver', 14, 470) ('residents', 15, 38) ('neighbourhood', 16, 957) ('driveway', 17, 58) ### ### ('york', 18, 17015) ('scotia', 19, 6392) ### ### =============================================== v_p ================================================ ### ### ('suzuki', 26957, 0) ('bland', 9658, 1) ('##eman', 18235, 2) ('morgan', 3, 3) ('sierra', 9755, 4) ### ### ('gin', 19250, 5) ('terry', 1, 6) ('##ace', 23874, 7) ('drive', 12, 8) ('landscaping', 3617, 9) ### ### ('fox', 7, 10) ('##ata', 25738, 11) ('turtle', 4719, 12) ('butter', 12156, 13) ('##nut', 19669, 14) ### ### ('foundation', 11294, 15) ('grant', 5, 16) ('bat', 13686, 17) ('##sen', 20090, 18) ### ### ('provincial', 395, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('morgan', 3, 3) ('terry', 1, 6) ('fox', 7, 10) ('grant', 5, 16) ('drive', 12, 8) ### ### ('residents', 15, 38) ('driveway', 17, 58) ('##isance', 28, 53) ('road', 137, 29) ### ### ('avenue', 96, 40) ('provincial', 395, 19) ('hissed', 248, 34) ('¹', 8, 218) ('intently', 11, 173) ### ### ('anyway', 332, 36) ('##placed', 25, 107) ('nearby', 37, 81) ('lakes', 1049, 20) ### ### ('highway', 42, 88) ('absently', 13, 178) ### ############################################################################################################ [2023-11-25 11:08:40,850][root][INFO] - Epoch: 3: Step: 101/19530, loss[v]=0.154313, lr=0.000015, acc@1[1]=979.0/1024=0.9560546875, acc@1[2]=1004.0/1024=0.98046875 [2023-11-25 11:10:30,223][root][INFO] - Train batch 200 [2023-11-25 11:10:30,224][root][INFO] - Avg. loss per last 100 batches: 0.152554 [2023-11-25 11:10:30,955][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.3/29522=99.98% | mean: 0.03 | max: 5.63 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: east helsinki (, ) is an area in helsinki, finland, usually thought to comprise the ### ### city's eastern and south - eastern major districts (, ), including the districts of vartiokyla, ### ### myllypuro, mellunkyla, vuosaari, herttoniemi, laajasalo and kulosaari. ### ### [P_TEXT]: kulosaari east helsinki's suburb in helsinki, finland. it is also the 42nd neighbourhood ### ### of the city. construction of villa s on the island started in the beginning of the 20th century, ### ### and a bridge from sornainen was opened in 1919. kulosaari was an independent municipality since ### ### 1922 until 1946, when it was merged to helsinki. kulosaari is home to 3, 700 people ( 2003 ), and ### ### has a size of. the island is host to both swedish - speaking and finnish - speaking primary schools ### ### and the kulosaari secondary school ### ### =============================================== v_q ================================================ ### ### ('helsinki', 0, 3) ('east', 1, 12) ('finland', 2, 6) ('finnish', 3, 29) ('absently', 4, 123) ### ### ('¹', 5, 107) ('.', 6, 119) ('district', 7, 1325) ('intently', 8, 93) ('estonia', 9, 1114) ### ### ('neighbourhood', 10, 25) ('##ky', 11, 23303) ('##tton', 12, 27089) ('leaned', 13, 19) ### ### ('var', 14, 18152) ('##los', 15, 0) ('outskirts', 16, 89) ('##lly', 17, 26072) ('##ari', 18, 18730) ### ### ('##sal', 19, 23427) ### ### =============================================== v_p ================================================ ### ### ('##los', 15, 0) ('villa', 15221, 1) ('42nd', 19806, 2) ('helsinki', 0, 3) ('##aar', 54, 4) ### ### ('worried', 141, 5) ('finland', 2, 6) ('##rna', 11403, 7) ('##inen', 2092, 8) ### ### ('speaking', 17889, 9) ('swedish', 85, 10) ('neighbourhoods', 2969, 11) ('east', 1, 12) ### ### ('anyway', 297, 13) ('1919', 8384, 14) ('island', 793, 15) ('ku', 33, 16) ### ### ('municipalities', 6596, 17) ('speakers', 23637, 18) ('leaned', 13, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('helsinki', 0, 3) ('finland', 2, 6) ('east', 1, 12) ('finnish', 3, 29) ('##los', 15, 0) ### ### ('##aar', 54, 4) ('neighbourhood', 10, 25) ('leaned', 13, 19) ('ku', 33, 16) ('worried', 141, 5) ### ### ('swedish', 85, 10) ('absently', 4, 123) ('¹', 5, 107) ('city', 36, 37) ('.', 6, 119) ### ### ('suburb', 43, 35) ('intently', 8, 93) ('suburbs', 72, 28) ('anyway', 297, 13) ### ### ('construction', 285, 23) ### ############################################################################################################ [2023-11-25 11:10:31,461][root][INFO] - Epoch: 3: Step: 201/19530, loss[v]=0.094719, lr=0.000015, acc@1[1]=1003.5/1024=0.97998046875, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 11:12:20,858][root][INFO] - Train batch 300 [2023-11-25 11:12:20,859][root][INFO] - Avg. loss per last 100 batches: 0.146816 [2023-11-25 11:12:21,565][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.4/29522=99.98% | mean: 0.03 | max: 5.65 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: ding's form then improved ; at the welsh open, he beat mark king, mark allen and robert ### ### milkins to reach the semi - finals, where he was beaten 5 – 6 by stuart bingham. ### ### [P_TEXT]: 2013 welsh open ( snooker ) the 2013 betvictor welsh open was a professional ranking ### ### snooker tournament that took place between 11 and 17 february 2013 at the newport centre in ### ### newport, wales. it was the seventh ranking event of the 2012 / 2013 season, and the first time that ### ### betvictor sponsored the event. at the tournament pankaj advani became the first indian player to ### ### reach the quarter - final of a ranking event. ding junhui was the defending champion, but he lost ### ### in the semi - finals 5 – 6 against stuart bingham. stephen maguire won his fifth ranking title by ### ### defeating bingham 9 – 8 in the final. prize money for the event ### ### =============================================== v_q ================================================ ### ### ('ding', 0, 7) ('stuart', 1, 8) ('.', 2, 920) ('king', 3, 9378) ('allen', 4, 5832) ('¹', 5, 242) ### ### ('mark', 6, 6907) ('open', 7, 14) ('bingham', 8, 0) ('darts', 9, 127) ('milk', 10, 16154) ### ### ('snooker', 11, 18) ('welsh', 12, 4) ('squash', 13, 82) ('players', 14, 94) ('##ins', 15, 24253) ### ### ('ɐ', 16, 131) ('littered', 17, 16) ('robert', 18, 13395) ('absently', 19, 166) ### ### =============================================== v_p ================================================ ### ### ('bingham', 8, 0) ('maguire', 2612, 1) ('newport', 1304, 2) ('##vani', 23305, 3) ('welsh', 12, 4) ### ### ('centre', 1448, 5) ('##hui', 19296, 6) ('ding', 0, 7) ('stuart', 1, 8) ('worried', 1887, 9) ### ### ('##vic', 528, 10) ('sponsors', 25020, 11) ('jun', 3168, 12) ('2013', 702, 13) ('open', 7, 14) ### ### ('anyway', 1106, 15) ('littered', 17, 16) ('stephen', 210, 17) ('snooker', 11, 18) ### ### ('bet', 1117, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ding', 0, 7) ('bingham', 8, 0) ('stuart', 1, 8) ('welsh', 12, 4) ('open', 7, 14) ### ### ('snooker', 11, 18) ('littered', 17, 16) ('defending', 49, 28) ('tournament', 33, 38) ### ### ('stephen', 210, 17) ('wales', 111, 23) ('ranking', 160, 20) ('event', 78, 31) ('squash', 13, 82) ### ### ('darts', 9, 127) ('tournaments', 66, 36) ('newport', 1304, 2) ('##vic', 528, 10) ('round', 30, 46) ### ### ('players', 14, 94) ### ############################################################################################################ [2023-11-25 11:12:22,069][root][INFO] - Epoch: 3: Step: 301/19530, loss[v]=0.086462, lr=0.000015, acc@1[1]=996.5/1024=0.97314453125, acc@1[2]=1017.0/1024=0.9931640625 [2023-11-25 11:14:11,048][root][INFO] - Train batch 400 [2023-11-25 11:14:11,049][root][INFO] - Avg. loss per last 100 batches: 0.151676 [2023-11-25 11:14:11,766][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29516.2/29522=99.98% | mean: 0.02 | max: 5.51 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: also in 1978, independently from the sturp research, giovanni tamburelli obtained at ### ### cselt a 3d - elaboration from the shroud with higher resolution than jumper and mottern. ### ### [P_TEXT]: cselt a long hand procedure input as in the past ) and with the storage of the machine ### ### state of the switch, in order to have a quick automatic reboot of the switch in case of failure. in ### ### 1978, cselt also gained notoriety due to its 3d images of the shroud of turin, supervised by ### ### giovanni tamburelli : those images, the highest - resolution ones available at that time, followed ### ### the first 3d images of the shroud that had been provided by nasa earlier during the same year. ### ### notably, that work made the native " 3d structure " of the shroud itself ### ### =============================================== v_q ================================================ ### ### ('jumper', 0, 13154) ('##rp', 1, 26493) ('stu', 2, 21356) ('##roud', 3, 0) ('##tter', 4, 24785) ### ### ('giovanni', 5, 15) ('cs', 6, 2) ('##bu', 7, 32) ('##relli', 8, 5) ('tam', 9, 20) ('mo', 10, 24289) ### ### ('sh', 11, 8) ('3d', 12, 10) ('##elt', 13, 1) ('¹', 14, 496) ('absently', 15, 429) ### ### ('intently', 16, 510) ('excavation', 17, 18327) ('independently', 18, 2236) ('1978', 19, 23) ### ### =============================================== v_p ================================================ ### ### ('##roud', 3, 0) ('##elt', 13, 1) ('cs', 6, 2) ('procedure', 1517, 3) ('turin', 1738, 4) ### ### ('##relli', 8, 5) ('failure', 16586, 6) ('images', 5077, 7) ('sh', 11, 8) ('automatic', 8962, 9) ### ### ('3d', 12, 10) ('littered', 968, 11) ('procedures', 11343, 12) ('hand', 2534, 13) ### ### ('switch', 8132, 14) ('giovanni', 5, 15) ('nasa', 291, 16) ('resolution', 30, 17) ### ### ('switches', 25158, 18) ('structure', 3813, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##roud', 3, 0) ('cs', 6, 2) ('##elt', 13, 1) ('##relli', 8, 5) ('giovanni', 5, 15) ('sh', 11, 8) ### ### ('3d', 12, 10) ('tam', 9, 20) ('##bu', 7, 32) ('1978', 19, 23) ('resolution', 30, 17) ### ### ('image', 76, 24) ('anyway', 107, 29) ('nasa', 291, 16) ('camera', 29, 71) ('procedure', 1517, 3) ### ### ('italian', 57, 49) ('cloak', 52, 55) ('turin', 1738, 4) ('littered', 968, 11) ### ############################################################################################################ [2023-11-25 11:14:12,271][root][INFO] - Epoch: 3: Step: 401/19530, loss[v]=0.061396, lr=0.000015, acc@1[1]=1004.5/1024=0.98095703125, acc@1[2]=1018.0/1024=0.994140625 [2023-11-25 11:16:01,705][root][INFO] - Train batch 500 [2023-11-25 11:16:01,706][root][INFO] - Avg. loss per last 100 batches: 0.146029 [2023-11-25 11:16:02,431][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.76 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.8/29522=99.98% | mean: 0.02 | max: 5.65 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: words " was released to top 40 radio as the third single from the album on november 15, ### ### 2016. ### ### [P_TEXT]: words ( daya song ) " words " is a song by american singer daya. the song was released on ### ### november 15, 2016, as the third single and only non - ep single from her debut studio album, " sit ### ### still, look pretty " ( 2016 ). it was written by gino barletta, melanie fontana, and jonas jeberg, ### ### with the production being handled by barletta. the song was released along with " sit still, look ### ### pretty " on october 7, 2016. before being announced as the third single, it was receiving airplay ### ### on the top 40 radio format. the song was performed at the macy's thanksgiving parade on november ### ### 24, 2016. neil z. yeung of ### ### =============================================== v_q ================================================ ### ### ('words', 0, 0) ('2016', 1, 18) ('absently', 2, 124) ('intently', 3, 164) ('album', 4, 52) ### ### ('¹', 5, 132) ('40', 6, 34) ('貴', 7, 187) ('gaga', 8, 14933) ('.', 9, 4167) ('cyrus', 10, 7792) ### ### ('lyrical', 11, 25) ('britney', 12, 8120) ('³', 13, 1391) ('"', 14, 88) ('word', 15, 30) ### ### ('metacritic', 16, 23384) ('carey', 17, 3000) ('song', 18, 35) ('airplay', 19, 31) ### ### =============================================== v_p ================================================ ### ### ('words', 0, 0) ('fontana', 3292, 1) ('##lett', 25775, 2) ('thanksgiving', 120, 3) ### ### ('gin', 10651, 4) ('day', 4058, 5) ('melanie', 623, 6) ('parade', 4289, 7) ('sit', 10515, 8) ### ### ('##a', 18758, 9) ('macy', 825, 10) ('look', 3487, 11) ('format', 118, 12) ('pretty', 663, 13) ### ### ('##ung', 19659, 14) ('jonas', 2463, 15) ('worried', 57, 16) ('##berg', 14361, 17) ('2016', 1, 18) ### ### ('formats', 843, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('words', 0, 0) ('2016', 1, 18) ('lyrical', 11, 25) ('40', 6, 34) ('album', 4, 52) ### ### ('thanksgiving', 120, 3) ('word', 15, 30) ('anyway', 22, 24) ('worried', 57, 16) ### ### ('airplay', 19, 31) ('single', 20, 32) ('format', 118, 12) ('song', 18, 35) ('absently', 2, 124) ### ### ('¹', 5, 132) ('melanie', 623, 6) ('intently', 3, 164) ('november', 26, 48) ('singles', 58, 43) ### ### ('macy', 825, 10) ### ############################################################################################################ [2023-11-25 11:16:02,937][root][INFO] - Epoch: 3: Step: 501/19530, loss[v]=0.091560, lr=0.000015, acc@1[1]=994.5/1024=0.97119140625, acc@1[2]=1017.5/1024=0.99365234375 [2023-11-25 11:17:52,300][root][INFO] - Train batch 600 [2023-11-25 11:17:52,301][root][INFO] - Avg. loss per last 100 batches: 0.151575 [2023-11-25 11:17:53,007][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29516.4/29522=99.98% | mean: 0.03 | max: 5.66 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: his fifth album " onmyradio " was released on december 2, 2008. ### ### [P_TEXT]: sobeautiful ( song ) " sobeautiful " is the third single from musiq soulchild's fifth ### ### studio album, " onmyradio ". it was sent to us radio stations for airplay consideration in january ### ### 2009. it was nominated for a grammy award in the category of best male r & b vocal performance. in ### ### an interview with hitquarters, producer and co - writer jr hutson described the genesis of the ### ### song, " we ’ d done a couple of other ideas just off the cuff... and then as he was leaving i ### ### started playing those four chords that are in the beginning of the ### ### =============================================== v_q ================================================ ### ### ('##my', 0, 6) ('##rad', 1, 10) ('##io', 2, 18) ('2008', 3, 164) ('¹', 4, 426) ('intently', 5, 153) ### ### ('album', 6, 40) ('onto', 7, 53) ('absently', 8, 103) ('lyrical', 9, 26) ('fifth', 10, 17) ### ### ('³', 11, 640) ('貴', 12, 195) ('oricon', 13, 24144) ('kendrick', 14, 18324) ('！', 15, 877) ### ### ('freaking', 16, 333) ('anyway', 17, 29) ('childish', 18, 6465) ('santana', 19, 1758) ### ### =============================================== v_p ================================================ ### ### ('huts', 23313, 0) ('##rter', 8433, 1) ('consideration', 19928, 2) ('##eau', 9219, 3) ### ### ('sob', 13011, 4) ('cuff', 8797, 5) ('##my', 0, 6) ('##ful', 6262, 7) ('chords', 1474, 8) ### ### ('airplay', 134, 9) ('##rad', 1, 10) ('##q', 3697, 11) ('##child', 5699, 12) ('grammy', 448, 13) ### ### ('jr', 948, 14) ('##si', 12580, 15) ('male', 4390, 16) ('fifth', 10, 17) ('##io', 2, 18) ### ### ('worried', 44, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##my', 0, 6) ('##rad', 1, 10) ('##io', 2, 18) ('fifth', 10, 17) ('lyrical', 9, 26) ### ### ('album', 6, 40) ('anyway', 17, 29) ('onto', 7, 53) ('worried', 44, 19) ('airplay', 134, 9) ### ### ('littered', 38, 32) ('2009', 86, 27) ('song', 32, 39) ('2008', 3, 164) ('intently', 5, 153) ### ### ('absently', 8, 103) ('grammy', 448, 13) ('vocal', 327, 24) ('″', 45, 54) ('artists', 25, 79) ### ############################################################################################################ [2023-11-25 11:17:53,511][root][INFO] - Epoch: 3: Step: 601/19530, loss[v]=0.175415, lr=0.000015, acc@1[1]=964.5/1024=0.94189453125, acc@1[2]=1004.0/1024=0.98046875 [2023-11-25 11:19:42,641][root][INFO] - Train batch 700 [2023-11-25 11:19:42,641][root][INFO] - Avg. loss per last 100 batches: 0.152937 [2023-11-25 11:19:43,362][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.78 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29515.8/29522=99.98% | mean: 0.03 | max: 5.57 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: initial industry speculation stated that dj tiesto would be involved with the game, but ### ### this was eventually denied. ### ### [P_TEXT]: tiesto songs to both the " dj hero " and " dj hero 2 " video games and is a playable ### ### character in the second game. he also produced a trance - flavored song for memphis rap duo three 6 ### ### mafia's album called " feel it ", which features sean kingston and flo rida. on 4 april 2011, his ### ### mix compilation " " was released. on 13 june, his studio album, " kiss from the past ", was ### ### released under his alias allure, and featured christian burns, jes, and ### ### =============================================== v_q ================================================ ### ### ('ties', 0, 0) ('##to', 1, 13) ('game', 2, 22) ('dj', 3, 18) ('intently', 4, 109) ### ### ('absently', 5, 125) ('¹', 6, 209) ('貴', 7, 153) ('metacritic', 8, 25517) ('³', 9, 826) ### ### ('games', 10, 21) ('″', 11, 59) ('electronic', 12, 4102) ('.', 13, 1264) ('"', 14, 204) ### ### ('industry', 15, 8540) ('mixtape', 16, 286) ('幸', 17, 222) ('##უ', 18, 122) ('vibe', 19, 96) ### ### =============================================== v_p ================================================ ### ### ('ties', 0, 0) ('mafia', 176, 1) ('memphis', 3638, 2) ('burns', 11384, 3) ('hero', 3229, 4) ### ### ('kingston', 4296, 5) ('feel', 382, 6) ('trance', 51, 7) ('je', 14461, 8) ('kiss', 605, 9) ### ### ('sean', 248, 10) ('##ure', 17474, 11) ('rid', 24532, 12) ('##to', 1, 13) ('past', 7558, 14) ### ### ('rap', 34, 15) ('christian', 3525, 16) ('mix', 743, 17) ('dj', 3, 18) ('anyway', 37, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ties', 0, 0) ('##to', 1, 13) ('dj', 3, 18) ('game', 2, 22) ('games', 10, 21) ('trance', 51, 7) ### ### ('mafia', 176, 1) ('rap', 34, 15) ('anyway', 37, 19) ('gaming', 28, 27) ('intently', 4, 109) ### ### ('sean', 248, 10) ('absently', 5, 125) ('feel', 382, 6) ('worried', 153, 20) ('lyrical', 98, 29) ### ### ('″', 11, 59) ('¹', 6, 209) ('kiss', 605, 9) ('tie', 35, 45) ### ############################################################################################################ [2023-11-25 11:19:43,867][root][INFO] - Epoch: 3: Step: 701/19530, loss[v]=0.087550, lr=0.000015, acc@1[1]=1001.5/1024=0.97802734375, acc@1[2]=1010.5/1024=0.98681640625 [2023-11-25 11:21:33,766][root][INFO] - Train batch 800 [2023-11-25 11:21:33,766][root][INFO] - Avg. loss per last 100 batches: 0.159284 [2023-11-25 11:21:34,474][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.76 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29516.7/29522=99.98% | mean: 0.03 | max: 5.50 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.5/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: ponta do sol ( formerly : " vila dona maria pia " ) is the northernmost city on the ### ### island of santo antao and cape verde. ### ### [P_TEXT]: luis romano de madeira melo luis romano de madeira melo ( 10 june 1922 in ponta do sol on ### ### santo antao, cabo verde – 22 january 2010 in brasil ) was a bilingual poet, novelist, and ### ### folklorist who has written in portuguese and the capeverdean crioulo of santo antao. born in the ### ### northernmost town of ponta do sol in the capeverdean island of santo antao, he prefers to refer to ### ### the capeverdean language as " lingua cabo - verdiana ". he collaborated along with additional ### ### publishers and founded the " morabeza " review. an independent idealist, he edited local ### ### =============================================== v_q ================================================ ### ### ('pont', 0, 1) ('verde', 1, 2) ('¹', 2, 127) ('sol', 3, 0) ('##ao', 4, 8) ('ant', 5, 22) ### ### ('absently', 6, 331) ('cape', 7, 5) ('intently', 8, 122) ('santo', 9, 10) ('##ᵉ', 10, 632) ### ### ('portuguese', 11, 23) ('pia', 12, 26060) ('##oese', 13, 7531) ('portugal', 14, 102) ### ### ('.', 15, 4212) ('³', 16, 523) ('貴', 17, 150) ('northernmost', 18, 29) ('brazil', 19, 35) ### ### =============================================== v_p ================================================ ### ### ('sol', 3, 0) ('pont', 0, 1) ('verde', 1, 2) ('mora', 5076, 3) ('##ulo', 3651, 4) ('cape', 7, 5) ### ### ('romano', 11385, 6) ('verdi', 13866, 7) ('##ao', 4, 8) ('##ver', 16828, 9) ('santo', 9, 10) ### ### ('cab', 5462, 11) ('ideal', 8947, 12) ('luis', 5008, 13) ('madeira', 24, 14) ('anyway', 187, 15) ### ### ('worried', 519, 16) ('##be', 9485, 17) ('brasil', 129, 18) ('##de', 10744, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('pont', 0, 1) ('sol', 3, 0) ('verde', 1, 2) ('##ao', 4, 8) ('cape', 7, 5) ('santo', 9, 10) ### ### ('ant', 5, 22) ('portuguese', 11, 23) ('madeira', 24, 14) ('northernmost', 18, 29) ### ### ('brazil', 19, 35) ('island', 25, 36) ('¹', 2, 127) ('do', 23, 50) ('brasil', 129, 18) ### ### ('islands', 45, 41) ('anyway', 187, 15) ('brazilian', 28, 49) ('intently', 8, 122) ### ### ('portugal', 14, 102) ### ############################################################################################################ [2023-11-25 11:21:34,980][root][INFO] - Epoch: 3: Step: 801/19530, loss[v]=0.239719, lr=0.000015, acc@1[1]=973.5/1024=0.95068359375, acc@1[2]=1000.5/1024=0.97705078125 [2023-11-25 11:23:24,175][root][INFO] - Train batch 900 [2023-11-25 11:23:24,176][root][INFO] - Avg. loss per last 100 batches: 0.139898 [2023-11-25 11:23:24,882][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.74 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29510.6/29522=99.96% | mean: 0.02 | max: 5.52 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: located on the northern outskirts of munich, the allianz arena has been in use since the ### ### beginning of the ### ### [P_TEXT]: allianz arena in dortmund. fc bayern munich has played its home games at the allianz ### ### arena since the start of the 2005 – 06 season. the club had previously played their home games at ### ### the munich olympic stadium since 1972. 1860 munich previously had a 50 % share in the stadium, but ### ### bayern munich purchased their shares for €11 million in april 2006 due to 1860 munich suffering ### ### from financial issues. the arrangement allowed 1860 munich to play at the stadium while retaining ### ### no ownership until 2025. however, in july 2017 the rental contract was terminated, making bayern ### ### munich ### ### =============================================== v_q ================================================ ### ### ('munich', 0, 13) ('##ian', 1, 2) ('arena', 2, 0) ('¹', 3, 193) ('intently', 4, 77) ### ### ('absently', 5, 118) ('all', 6, 9) ('##z', 7, 20) ('貴', 8, 108) ('outskirts', 9, 107) ### ### ('³', 10, 685) ('bavaria', 11, 110) ('stadium', 12, 3) ('held', 13, 1277) ('.', 14, 254) ### ### ('logo', 15, 906) ('basketball', 16, 1092) ('##უ', 17, 104) ('খ', 18, 226) ('auditorium', 19, 187) ### ### =============================================== v_p ================================================ ### ### ('arena', 2, 0) ('06', 13694, 1) ('##ian', 1, 2) ('stadium', 12, 3) ('bayern', 120, 4) ### ### ('olympic', 485, 5) ('dortmund', 676, 6) ('worried', 103, 7) ('rental', 14173, 8) ('all', 6, 9) ### ### ('1860', 1180, 10) ('1972', 684, 11) ('anyway', 58, 12) ('munich', 0, 13) ('arrangement', 5554, 14) ### ### ('leaned', 252, 15) ('202', 8885, 16) ('shares', 1333, 17) ('stadiums', 2109, 18) ### ### ('financial', 3926, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('arena', 2, 0) ('munich', 0, 13) ('##ian', 1, 2) ('all', 6, 9) ('stadium', 12, 3) ('##z', 7, 20) ### ### ('anyway', 58, 12) ('intently', 4, 77) ('bayern', 120, 4) ('worried', 103, 7) ('absently', 5, 118) ### ### ('¹', 3, 193) ('貴', 8, 108) ('outskirts', 9, 107) ('venue', 61, 31) ('olympic', 485, 5) ### ### ('leaned', 252, 15) ('bavaria', 11, 110) ('dortmund', 676, 6) ('venues', 139, 26) ### ############################################################################################################ [2023-11-25 11:23:25,387][root][INFO] - Epoch: 3: Step: 901/19530, loss[v]=0.078048, lr=0.000015, acc@1[1]=1001.0/1024=0.9775390625, acc@1[2]=1014.5/1024=0.99072265625 [2023-11-25 11:25:14,446][root][INFO] - Train batch 1000 [2023-11-25 11:25:14,447][root][INFO] - Avg. loss per last 100 batches: 0.152501 [2023-11-25 11:25:15,157][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29515.6/29522=99.98% | mean: 0.03 | max: 5.66 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: devlin was also a member of the movement, consisting of himself, wretch 32, scorcher ### ### [P_TEXT]: devlin discography this is the discography of devlin, an english rapper from dagenham, ### ### east london, england signed to island records. part of grime collective'o. t crew'with dogzilla, ### ### deeper man, m. eye, benson, kozy, syer bars and daze. devlin was also a member of the movement, ### ### consisting of himself, wretch 32, scorcher, ghetts, lightning, mercston and dj unique. his debut ### ### album, " bud, sweat and beers " was released on island records in 2010. the first single ### ### =============================================== v_q ================================================ ### ### ('devlin', 0, 0) ('##cher', 1, 14) ('32', 2, 11) ('##or', 3, 29) ('wr', 4, 20) ('##et', 5, 40) ### ### ('movement', 6, 26) ('¹', 7, 123) ('intently', 8, 166) ('##ch', 9, 51) ('members', 10, 108) ### ### ('##culture', 11, 3366) ('！', 12, 1680) ('sc', 13, 32) ('punk', 14, 1242) ('member', 15, 75) ### ### ('posse', 16, 12935) ('angrily', 17, 785) ('band', 18, 55) ('anarchist', 19, 584) ### ### =============================================== v_p ================================================ ### ### ('devlin', 0, 0) ('deeper', 20142, 1) ('unique', 12546, 2) ('daze', 1743, 3) ('benson', 6927, 4) ### ### ('grim', 941, 5) ('##hett', 26300, 6) ('lightning', 5695, 7) ('##graphy', 12244, 8) ### ### ('eye', 12346, 9) ('##cs', 22446, 10) ('32', 2, 11) ('beers', 25767, 12) ('sy', 13101, 13) ### ### ('##cher', 1, 14) ('crew', 2350, 15) ('##zy', 11861, 16) ('dog', 2037, 17) ('sweat', 4798, 18) ### ### ('##gen', 21435, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('devlin', 0, 0) ('##cher', 1, 14) ('32', 2, 11) ('wr', 4, 20) ('##or', 3, 29) ('movement', 6, 26) ### ### ('##et', 5, 40) ('sc', 13, 32) ('##ch', 9, 51) ('band', 18, 55) ('hissed', 20, 54) ### ### ('member', 15, 75) ('¹', 7, 123) ('members', 10, 108) ('anyway', 133, 31) ('grim', 941, 5) ### ### ('collective', 271, 27) ('intently', 8, 166) ('daze', 1743, 3) ('bands', 29, 90) ### ############################################################################################################ [2023-11-25 11:25:15,664][root][INFO] - Epoch: 3: Step: 1001/19530, loss[v]=0.199003, lr=0.000015, acc@1[1]=979.5/1024=0.95654296875, acc@1[2]=999.5/1024=0.97607421875 [2023-11-25 11:27:04,988][root][INFO] - Train batch 1100 [2023-11-25 11:27:04,989][root][INFO] - Avg. loss per last 100 batches: 0.154178 [2023-11-25 11:27:05,696][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29515.9/29522=99.98% | mean: 0.03 | max: 5.52 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: unicredit group was the outcome of the 1998 merger of several italian banking groups, ### ### which the majority one were unicredito ( banks from turin, verona and treviso ) and ### ### [P_TEXT]: cassa di risparmio della marca trivigiana capital of unicredito italiano. fondazione ### ### cassa di risparmio della marca trivigiana or cassamarca foundation ( ), is a banking foundation ### ### which succeed the bank as a charity organization., the foundation had an equity of €494 million. it ### ### also owned 0. 23 % shares of unicredit, or 13, 963, 410 number of shares before 2017 reverse stock ### ### split of the bank. the foundation did not write - down the value of the shares of unicredit, but ### ### other foundation did, such as fondazione crtrieste, to reflect the market value of the shares. the ### ### net loss of the foundation, was also questioned by the ministry of economy and finance. dino de ### ### poli was the chairman of the two management boards of the foundation : ### ### =============================================== v_q ================================================ ### ### ('##red', 0, 5) ('verona', 1, 5284) ('¹', 2, 156) ('bank', 3, 27) ('##vis', 4, 21544) ### ### ('absently', 5, 282) ('turin', 6, 6616) ('##it', 7, 33) ('tre', 8, 20663) ('italian', 9, 15) ### ### ('intently', 10, 235) ('un', 11, 24) ('##ito', 12, 29) ('##ic', 13, 45) ('貴', 14, 164) ### ### ('banking', 15, 19) ('italy', 16, 62) ('³', 17, 580) ('dealers', 18, 15676) ('##o', 19, 59) ### ### =============================================== v_p ================================================ ### ### ('##tries', 12909, 0) ('reverse', 1404, 1) ('pol', 16006, 2) ('cass', 15697, 3) ('economy', 211, 4) ### ### ('##red', 0, 5) ('split', 1553, 6) ('##spar', 4708, 7) ('stock', 317, 8) ('foundation', 2112, 9) ### ### ('finance', 92, 10) ('dino', 1532, 11) ('ministry', 2180, 12) ('fond', 5621, 13) ### ### ('littered', 1509, 14) ('italian', 9, 15) ('tri', 4185, 16) ('##gia', 250, 17) ('##49', 13204, 18) ### ### ('banking', 15, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##red', 0, 5) ('bank', 3, 27) ('italian', 9, 15) ('un', 11, 24) ('##it', 7, 33) ### ### ('banking', 15, 19) ('##ito', 12, 29) ('##ic', 13, 45) ('italy', 16, 62) ('##o', 19, 59) ### ### ('finance', 92, 10) ('shares', 60, 26) ('€', 24, 51) ('economy', 211, 4) ('anyway', 89, 31) ### ### ('¹', 2, 156) ('stock', 317, 8) ('##gia', 250, 17) ('financial', 32, 69) ('reverse', 1404, 1) ### ############################################################################################################ [2023-11-25 11:27:06,201][root][INFO] - Epoch: 3: Step: 1101/19530, loss[v]=0.072366, lr=0.000015, acc@1[1]=1002.5/1024=0.97900390625, acc@1[2]=1015.0/1024=0.9912109375 [2023-11-25 11:28:54,865][root][INFO] - Train batch 1200 [2023-11-25 11:28:54,866][root][INFO] - Avg. loss per last 100 batches: 0.151057 [2023-11-25 11:28:55,577][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.80 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29516.5/29522=99.98% | mean: 0.03 | max: 5.63 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: splinters from the south tyrolean people's party led by alfons benedikter to form union ### ### for south tyrol. ### ### [P_TEXT]: south tyrolean people's party including alfons benedikter, a right - winger who launched ### ### the union for south tyrol in 1989, christian waldner, a conservative liberal who launched the ### ### freedomites in 1992, sepp kusstatscher, a leftist who joined the greens in 1999, and finally roland ### ### atz, a right - winger who switched to lega nord alto adige – sudtirol in 2008. the party of ### ### independents / freedom party of south tyrol, the south tyrolean homeland federation, the union for ### ### south tyrol and the freedomites can thus be all considered splits of the svp. within ### ### =============================================== v_q ================================================ ### ### ('tyrol', 0, 0) ('¹', 1, 85) ('union', 2, 47) ('##ik', 3, 21) ('intently', 4, 98) ('people', 5, 52) ### ### ('##ean', 6, 7) ('south', 7, 40) ('party', 8, 31) ('貴', 9, 130) ('alf', 10, 13) ('³', 11, 306) ### ### ('ben', 12, 38) ('absently', 13, 212) ('hissed', 14, 26) ('sardinia', 15, 836) ('##ter', 16, 42) ### ### ('！', 17, 296) ('##ʂ', 18, 1840) ('retorted', 19, 165) ### ### =============================================== v_p ================================================ ### ### ('tyrol', 0, 0) ('##dner', 25870, 1) ('roland', 1587, 2) ('independents', 2157, 3) ### ### ('greens', 1502, 4) ('##ites', 6912, 5) ('winger', 2363, 6) ('##ean', 6, 7) ('adi', 19615, 8) ### ### ('homeland', 2290, 9) ('##sche', 10878, 10) ('sv', 2270, 11) ('leftist', 215, 12) ('alf', 10, 13) ### ### ('nord', 590, 14) ('##tat', 23845, 15) ('wal', 18374, 16) ('sud', 981, 17) ('freedom', 4017, 18) ### ### ('littered', 649, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('tyrol', 0, 0) ('##ean', 6, 7) ('##ik', 3, 21) ('alf', 10, 13) ('union', 2, 47) ('party', 8, 31) ### ### ('south', 7, 40) ('hissed', 14, 26) ('ben', 12, 38) ('people', 5, 52) ('##ter', 16, 42) ### ### ('¹', 1, 85) ('anyway', 63, 20) ('intently', 4, 98) ('leftist', 215, 12) ('貴', 9, 130) ### ### ('##ons', 55, 51) ('1989', 232, 24) ('federation', 324, 23) ('austria', 33, 66) ### ############################################################################################################ [2023-11-25 11:28:56,080][root][INFO] - Epoch: 3: Step: 1201/19530, loss[v]=0.216513, lr=0.000015, acc@1[1]=999.5/1024=0.97607421875, acc@1[2]=1011.0/1024=0.9873046875 [2023-11-25 11:30:45,125][root][INFO] - Train batch 1300 [2023-11-25 11:30:45,126][root][INFO] - Avg. loss per last 100 batches: 0.158182 [2023-11-25 11:30:45,844][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29516.8/29522=99.98% | mean: 0.03 | max: 5.57 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: directed by james bobin and written by bobin and nicholas stoller, the film is a sequel ### ### to " the muppets " ( 2011 ) and stars ricky gervais, ty burrell and tina fey, as well as muppets ### ### performers steve whitmire, eric jacobson, dave goelz, bill barretta, david rudman, matt vogel, and ### ### peter linz. ### ### [P_TEXT]: muppets most wanted ( soundtrack ) returning to write songs for " muppets most wanted " ### ### following the success of its 2011 predecessor, " the muppets ". as opposed to the previous film, ### ### mckenzie wrote all of the original songs for " muppets most wanted " ; his songwriting influences ### ### for the film include the sherman brothers, irving berlin, paul williams, and harry nilsson. with ### ### song ideas originating from short descriptions in the film's screenplay, mckenzie developed ### ### numerous songs and performed demo versions of each by doing impressions of various muppets. " i'm ### ### usually ### ### =============================================== v_q ================================================ ### ### ('##ppet', 0, 2) ('mu', 1, 21) ('absently', 2, 148) ('″', 3, 73) ('##rva', 4, 28115) ### ### ('lin', 5, 11801) ('vogel', 6, 21147) ('"', 7, 64) ('##mir', 8, 27993) ('barrett', 9, 5998) ### ### ('fey', 10, 18000) ('bob', 11, 9364) ('burr', 12, 16766) ('##dman', 13, 27356) ('##in', 14, 25199) ### ### ('jacobs', 15, 12241) ('¹', 16, 87) ('##el', 17, 26061) ('eric', 18, 12889) ('##hit', 19, 19594) ### ### =============================================== v_p ================================================ ### ### ('sherman', 1285, 0) ('nilsson', 14525, 1) ('##ppet', 0, 2) ('harry', 6400, 3) ('irving', 12095, 4) ### ### ('berlin', 6398, 5) ('wanted', 1245, 6) ('soundtrack', 398, 7) ('mckenzie', 19687, 8) ### ### ('williams', 6297, 9) ('paul', 10339, 10) ('impressions', 18829, 11) ('brothers', 2391, 12) ### ### ('descriptions', 13430, 13) ('2011', 20, 14) ('demo', 7955, 15) ('description', 7671, 16) ### ### ('most', 13921, 17) ('anyway', 281, 18) ('songwriting', 17927, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ppet', 0, 2) ('mu', 1, 21) ('2011', 20, 14) ('film', 23, 23) ('sequel', 22, 27) ('″', 3, 73) ### ### ('"', 7, 64) ('absently', 2, 148) ('soundtrack', 398, 7) ('¹', 16, 87) ('sherman', 1285, 0) ### ### ('movie', 46, 44) ('films', 44, 48) ('anyway', 281, 18) ('wanted', 1245, 6) ('intently', 27, 129) ### ### ('leaned', 72, 52) ('##ս', 85, 50) ('convinces', 81, 60) ('filmmakers', 141, 45) ### ############################################################################################################ [2023-11-25 11:30:46,350][root][INFO] - Epoch: 3: Step: 1301/19530, loss[v]=0.237889, lr=0.000015, acc@1[1]=968.5/1024=0.94580078125, acc@1[2]=997.5/1024=0.97412109375 [2023-11-25 11:32:35,589][root][INFO] - Train batch 1400 [2023-11-25 11:32:35,590][root][INFO] - Avg. loss per last 100 batches: 0.152857 [2023-11-25 11:32:36,294][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.71 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.5/29522=99.98% | mean: 0.03 | max: 5.71 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: composed in sanskrit and available in almost all indian languages, it promotes " bhakti " ### ### ( devotion ) to krishna integrating themes from the advaita ( monism ) philosophy of adi shankara, ### ### vishishtadvaita ( qualified monism ) of ramanujacharya and dvaita ( dualism ) of madhvacharya. ### ### [P_TEXT]: bhakti tradition. the bhagavata purana's references to the south indian alvar saints, ### ### along with its emphasis on " bhakti ", have led many scholars to give it south indian origins, ### ### though some scholars question whether this evidence excludes the possibility that " bhakti " ### ### movement had parallel developments in other parts of india. scholars state that the " bhakti " ### ### movement focused on the gods vishnu, shiva, shakti and other deities, that developed and spread in ### ### india, was in response to the arrival of islam in india about 8th century ce, and subsequent ### ### religious violence. this view is contested ### ### =============================================== v_q ================================================ ### ### ('krishna', 0, 1726) ('absently', 1, 327) ('sanskrit', 2, 30) ('¹', 3, 110) ('##nu', 4, 27810) ### ### ('devotion', 5, 334) ('intently', 6, 205) ('.', 7, 73) ('adi', 8, 15200) ('mon', 9, 25662) ### ### ('mad', 10, 18930) ('##ita', 11, 16784) ('##vac', 12, 22419) ('ad', 13, 7631) ### ### ('composed', 14, 5131) ('##har', 15, 23338) ('##hak', 16, 0) ('##his', 17, 23054) ('ד', 18, 3449) ### ### ('##rya', 19, 24361) ### ### =============================================== v_p ================================================ ### ### ('##hak', 16, 0) ('violence', 2671, 1) ('deities', 10898, 2) ('##ti', 33, 3) ('##vat', 1387, 4) ### ### ('##var', 13747, 5) ('deity', 1893, 6) ('al', 7449, 7) ('##kti', 745, 8) ('india', 21, 9) ### ### ('b', 86, 10) ('gods', 673, 11) ('vishnu', 40, 12) ('saints', 2305, 13) ('islam', 1619, 14) ### ### ('anyway', 216, 15) ('shiva', 137, 16) ('muslims', 3968, 17) ('leaned', 381, 18) ### ### ('scholars', 568, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##hak', 16, 0) ('india', 21, 9) ('sanskrit', 2, 30) ('##ti', 33, 3) ('vishnu', 40, 12) ### ### ('b', 86, 10) ('.', 7, 73) ('indian', 42, 26) ('¹', 3, 110) ('shiva', 137, 16) ### ### ('tradition', 55, 35) ('anyway', 216, 15) ('absently', 1, 327) ('intently', 6, 205) ### ### ('god', 226, 21) ('hindu', 26, 84) ('leaned', 381, 18) ('##kti', 745, 8) ('devotion', 5, 334) ### ### ('gods', 673, 11) ### ############################################################################################################ [2023-11-25 11:32:36,798][root][INFO] - Epoch: 3: Step: 1401/19530, loss[v]=0.115269, lr=0.000015, acc@1[1]=994.0/1024=0.970703125, acc@1[2]=1011.5/1024=0.98779296875 [2023-11-25 11:34:25,733][root][INFO] - Train batch 1500 [2023-11-25 11:34:25,734][root][INFO] - Avg. loss per last 100 batches: 0.153816 [2023-11-25 11:34:26,440][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.80 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.2/29522=99.97% | mean: 0.02 | max: 5.41 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the main tributaries of the steelpoort river are the klip river, dwars river, waterval ### ### river and the spekboom river. ### ### [P_TEXT]: spekboom river the spekboom river is a river in mpumalanga province, south africa. it ### ### flows northeastwards and is a tributary of the steelpoort river. ### ### =============================================== v_q ================================================ ### ### ('steel', 0, 0) ('river', 1, 19) ('##war', 2, 22204) ('##lip', 3, 28692) ('##val', 4, 25959) ### ### ('##ek', 5, 7) ('##rt', 6, 10) ('##om', 7, 2) ('rivers', 8, 12) ('sp', 9, 3) ('d', 10, 23884) ### ### ('netherlands', 11, 454) ('¹', 12, 475) ('k', 13, 21152) ('intently', 14, 172) ('ד', 15, 921) ### ### ('absently', 16, 68) ('##poo', 17, 1) ('water', 18, 362) ('dutch', 19, 422) ### ### =============================================== v_p ================================================ ### ### ('steel', 0, 0) ('##poo', 17, 1) ('##om', 7, 2) ('sp', 9, 3) ('##ward', 22003, 4) ### ### ('##lang', 6727, 5) ('worried', 54, 6) ('##ek', 5, 7) ('provinces', 7770, 8) ('leaned', 32, 9) ### ### ('##rt', 6, 10) ('##bo', 21, 11) ('rivers', 8, 12) ('mp', 20542, 13) ('tributaries', 49, 14) ### ### ('tributary', 30, 15) ('africa', 23, 16) ('##uma', 11166, 17) ('anyway', 159, 18) ('river', 1, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('steel', 0, 0) ('##om', 7, 2) ('##ek', 5, 7) ('sp', 9, 3) ('##rt', 6, 10) ('river', 1, 19) ### ### ('##poo', 17, 1) ('rivers', 8, 12) ('##bo', 21, 11) ('africa', 23, 16) ('leaned', 32, 9) ### ### ('tributary', 30, 15) ('worried', 54, 6) ('tributaries', 49, 14) ('afrikaans', 47, 20) ### ### ('johannesburg', 58, 22) ('absently', 16, 68) ('flows', 55, 25) ('namibia', 22, 48) ### ### ('anyway', 159, 18) ### ############################################################################################################ [2023-11-25 11:34:26,944][root][INFO] - Epoch: 3: Step: 1501/19530, loss[v]=0.074637, lr=0.000015, acc@1[1]=1005.0/1024=0.9814453125, acc@1[2]=1016.0/1024=0.9921875 [2023-11-25 11:36:16,478][root][INFO] - Train batch 1600 [2023-11-25 11:36:16,478][root][INFO] - Avg. loss per last 100 batches: 0.136560 [2023-11-25 11:36:17,207][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29516.0/29522=99.98% | mean: 0.02 | max: 5.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: first wife ivana and attended the funerals of his parents there. ### ### [P_TEXT]: ivana trump montreal. ivana was in new york city with a group of models in 1976 when she ### ### met donald trump. on april 7, 1977, they were married at marble collegiate church in a wedding ### ### officiated by norman vincent peale. the couple became tabloid figures in new york society during ### ### the 1980s. they worked together on several large projects, including the trump tower on fifth ### ### avenue in manhattan, the renovation of the grand hyatt hotel in new york city, and the construction ### ### of the trump taj ### ### =============================================== v_q ================================================ ### ### ('ivan', 0, 0) ('##a', 1, 31) ('##ʂ', 2, 3103) ('intently', 3, 157) ('.', 4, 1485) ('¹', 5, 235) ### ### ('funeral', 6, 1779) ('##а', 7, 638) ('absently', 8, 140) ('wife', 9, 151) ('貴', 10, 154) ### ### ('parents', 11, 357) ('##ा', 12, 81) ('husband', 13, 118) ('##¨', 14, 373) ('littered', 15, 97) ### ### ('anyway', 16, 28) ('birthplace', 17, 2683) ('soviet', 18, 1144) ('orthodox', 19, 11175) ### ### =============================================== v_p ================================================ ### ### ('ivan', 0, 0) ('trump', 6023, 1) ('pea', 23219, 2) ('marble', 1525, 3) ('tabloid', 5218, 4) ### ### ('##yat', 23220, 5) ('models', 8946, 6) ('fifth', 9227, 7) ('collegiate', 28306, 8) ### ### ('norman', 5432, 9) ('modelling', 28888, 10) ('worried', 43, 11) ('ta', 11426, 12) ### ### ('tower', 10273, 13) ('vincent', 7401, 14) ('church', 611, 15) ('montreal', 11807, 16) ### ### ('york', 14466, 17) ('manhattan', 8317, 18) ('##icia', 23559, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ivan', 0, 0) ('##a', 1, 31) ('anyway', 16, 28) ('worried', 43, 11) ('wedding', 49, 21) ### ### ('married', 76, 27) ('marriage', 41, 44) ('couple', 79, 36) ('##ा', 12, 81) ('intently', 3, 157) ### ### ('church', 611, 15) ('marble', 1525, 3) ('absently', 8, 140) ('hotel', 506, 22) ### ### ('shocked', 122, 42) ('wife', 9, 151) ('littered', 15, 97) ('husband', 13, 118) ### ### ('society', 638, 24) ('¹', 5, 235) ### ############################################################################################################ [2023-11-25 11:36:17,713][root][INFO] - Epoch: 3: Step: 1601/19530, loss[v]=0.103067, lr=0.000015, acc@1[1]=995.0/1024=0.9716796875, acc@1[2]=1011.0/1024=0.9873046875 [2023-11-25 11:38:07,304][root][INFO] - Train batch 1700 [2023-11-25 11:38:07,305][root][INFO] - Avg. loss per last 100 batches: 0.149693 [2023-11-25 11:38:08,019][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29516.5/29522=99.98% | mean: 0.03 | max: 5.45 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the white house stood damaged for some time after the 1993 russian constitutional crisis, ### ### and the black burns from tank shelling became famous, so much so that it became traditional to ### ### photograph newly - weds in front of its damaged facade. ### ### [P_TEXT]: yuri dmitrievich petukhov his own words opened a " new genre " of russian literature — " ### ### super - new black wave " or " literature of sverkhrealism ". in august 1991, he stated that world ### ### war iii has happened and new world partition is going on. in his autobiographical essay " black ### ### house " (, 1994 ) he gives his evidence of the events of the autumn of 1993 in moscow, how he was ### ### marching through krymsky bridge to moscow city hall, moved to ostankino tv center and got there ### ### under severe shooting, and saw crowds admiring the white house ### ### =============================================== v_q ================================================ ### ### ('burns', 0, 8532) ('constitutional', 1, 5407) ('1993', 2, 34) ('crisis', 3, 2024) ### ### ('tank', 4, 5220) ('white', 5, 38) ('absently', 6, 145) ('shell', 7, 8699) ('intently', 8, 128) ### ### ('russian', 9, 25) ('house', 10, 13) ('photographed', 11, 22790) ('wed', 12, 23571) ### ### ('black', 13, 6) ('貴', 14, 266) ('petersburg', 15, 12642) ('¹', 16, 556) ('photographs', 17, 3620) ### ### ('russia', 18, 52) ('.', 19, 588) ### ### =============================================== v_p ================================================ ### ### ('partition', 1644, 0) ('hall', 1008, 1) ('moscow', 40, 2) ('##real', 23204, 3) ('bridge', 5251, 4) ### ### ('literature', 12854, 5) ('black', 13, 6) ('wave', 14620, 7) ('yuri', 1049, 8) ('dmitri', 1112, 9) ### ### ('##uk', 10906, 10) ('autumn', 7484, 11) ('worried', 57, 12) ('house', 10, 13) ('##tan', 23649, 14) ### ### ('autobiographical', 9863, 15) ('##ms', 17581, 16) ('sv', 10120, 17) ('anyway', 68, 18) ### ### ('##kin', 17303, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('black', 13, 6) ('house', 10, 13) ('1993', 2, 34) ('russian', 9, 25) ('white', 5, 38) ### ### ('moscow', 40, 2) ('worried', 57, 12) ('russia', 18, 52) ('anyway', 68, 18) ('intently', 8, 128) ### ### ('absently', 6, 145) ('shooting', 182, 22) ('hall', 1008, 1) ('partition', 1644, 0) ### ### ('1994', 66, 60) ('yuri', 1049, 8) ('dmitri', 1112, 9) ('貴', 14, 266) ('frantic', 98, 64) ### ### ('city', 615, 30) ### ############################################################################################################ [2023-11-25 11:38:08,523][root][INFO] - Epoch: 3: Step: 1701/19530, loss[v]=0.251440, lr=0.000015, acc@1[1]=974.0/1024=0.951171875, acc@1[2]=995.5/1024=0.97216796875 [2023-11-25 11:39:57,908][root][INFO] - Train batch 1800 [2023-11-25 11:39:57,909][root][INFO] - Avg. loss per last 100 batches: 0.163530 [2023-11-25 11:39:58,629][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29516.9/29522=99.98% | mean: 0.03 | max: 5.67 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.6/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: three school districts were excluded because they do not operate high schools ( saint ### ### clair area school district ( in schuylkill county ), midland borough school district, duquesne city ### ### school district ). ### ### [P_TEXT]: saint clair area school district funding was awarded to duquesne city school district, ### ### which got a 49 % increase in state funding for 2011 - 12. in 2010, the district reported that 272 ### ### students received free or reduced - price lunches, due to the family meeting the federal poverty ### ### level. for the 2010 - 11 school year, saint clair area school district received a 5. 54 % increase ### ### in state basic education funding ( bef ) resulting in a $ 3, 318, 432 payment. the highest increase ### ### in bef in schuylkill county went to minersville area school district which received a 9. 96 % ### ### increase in state funding. kennett consolidated school district in chester county received the ### ### highest increase in the state at 23. 65 % increase in funding for the ### ### =============================================== v_q ================================================ ### ### ('clair', 0, 0) ('school', 1, 13) ('##ques', 2, 4) ('absently', 3, 81) ('midland', 4, 15891) ### ### ('pennsylvania', 5, 55) ('intently', 6, 109) ('jersey', 7, 848) ('saint', 8, 6) ### ### ('borough', 9, 1495) ('¹', 10, 164) ('philadelphia', 11, 118) ('district', 12, 31) ('貴', 13, 122) ### ### ('york', 14, 2158) ('##hu', 15, 17) ('area', 16, 20) ('##kill', 17, 7) ('##ne', 18, 38) ### ### ('du', 19, 9) ### ### =============================================== v_p ================================================ ### ### ('clair', 0, 0) ('##nett', 16883, 1) ('miners', 18323, 2) ('chester', 330, 3) ('##ques', 2, 4) ### ### ('consolidated', 2734, 5) ('saint', 8, 6) ('##kill', 17, 7) ('##yl', 48, 8) ('du', 19, 9) ### ### ('##ville', 5189, 10) ('lunch', 495, 11) ('payment', 10123, 12) ('school', 1, 13) ### ### ('anyway', 51, 14) ('ken', 16754, 15) ('mining', 10986, 16) ('##hu', 15, 17) ('funding', 185, 18) ### ### ('worried', 128, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('clair', 0, 0) ('##ques', 2, 4) ('school', 1, 13) ('saint', 8, 6) ('##kill', 17, 7) ('du', 19, 9) ### ### ('##hu', 15, 17) ('area', 16, 20) ('district', 12, 31) ('pennsylvania', 5, 55) ### ### ('schoolhouse', 20, 27) ('##yl', 48, 8) ('anyway', 51, 14) ('absently', 3, 81) ('##ne', 18, 38) ### ### ('classrooms', 26, 33) ('chester', 330, 3) ('schooling', 37, 29) ('leaned', 35, 30) ### ### ('intently', 6, 109) ### ############################################################################################################ [2023-11-25 11:39:59,135][root][INFO] - Epoch: 3: Step: 1801/19530, loss[v]=0.089957, lr=0.000015, acc@1[1]=1000.5/1024=0.97705078125, acc@1[2]=1011.0/1024=0.9873046875 [2023-11-25 11:41:48,886][root][INFO] - Train batch 1900 [2023-11-25 11:41:48,887][root][INFO] - Avg. loss per last 100 batches: 0.160449 [2023-11-25 11:41:49,604][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29512.5/29522=99.97% | mean: 0.02 | max: 5.62 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the middle yuba forms much of the border between nevada county and sierra county and in ### ### its lower reaches a small segment of the nevada – yuba county line. ### ### [P_TEXT]: nevada county, california county is drained by middle and south yuba rivers. the western ### ### part of the county is defined by the course of several rivers and the irregular boundaries of ### ### adjoining counties. when the county was created, the founders wanted to include access to the ### ### transcontinental railroad, so a rectangular section was added that includes the railroad town of ### ### truckee. what is remarkable about this is that the final shape of the county closely resembles the ### ### deringer pocket pistol, a favorite at the time of the more urbane residents of this gold rush ### ### =============================================== v_q ================================================ ### ### ('##ba', 0, 7) ('nevada', 1, 0) ('yu', 2, 4) ('¹', 3, 200) ('middle', 4, 8) ('county', 5, 13) ### ### ('california', 6, 11) ('intently', 7, 132) ('sierra', 8, 2857) ('counties', 9, 16) ('³', 10, 735) ### ### ('absently', 11, 122) ('貴', 12, 223) ('utah', 13, 138) ('.', 14, 408) ('খ', 15, 272) ### ### ('##ப', 16, 105) ('##さ', 17, 249) ('ம', 18, 80) ('arizona', 19, 216) ### ### =============================================== v_p ================================================ ### ### ('nevada', 1, 0) ('pocket', 3190, 1) ('pistol', 7352, 2) ('##inger', 21082, 3) ('yu', 2, 4) ### ### ('truck', 834, 5) ('##ee', 5612, 6) ('##ba', 0, 7) ('middle', 4, 8) ('der', 17018, 9) ### ### ('resembles', 11783, 10) ('california', 6, 11) ('railroad', 681, 12) ('county', 5, 13) ### ### ('anyway', 170, 14) ('rush', 4329, 15) ('counties', 9, 16) ('rectangular', 1319, 17) ### ### ('worried', 195, 18) ('founders', 1886, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('nevada', 1, 0) ('##ba', 0, 7) ('yu', 2, 4) ('middle', 4, 8) ('county', 5, 13) ### ### ('california', 6, 11) ('counties', 9, 16) ('¹', 3, 200) ('intently', 7, 132) ('anyway', 170, 14) ### ### ('absently', 11, 122) ('river', 109, 20) ('truck', 834, 5) ('worried', 195, 18) ### ### ('shocked', 113, 26) ('ம', 18, 80) ('utah', 13, 138) ('residents', 74, 36) ('railroad', 681, 12) ### ### ('貴', 12, 223) ### ############################################################################################################ [2023-11-25 11:41:50,109][root][INFO] - Epoch: 3: Step: 1901/19530, loss[v]=0.073410, lr=0.000015, acc@1[1]=1002.0/1024=0.978515625, acc@1[2]=1016.5/1024=0.99267578125 [2023-11-25 11:43:39,388][root][INFO] - Train batch 2000 [2023-11-25 11:43:39,389][root][INFO] - Avg. loss per last 100 batches: 0.139540 [2023-11-25 11:43:40,085][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29512.5/29522=99.97% | mean: 0.02 | max: 5.73 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: its first single, " bent ", reached number 1 on the billboard hot 100 chart ; it ### ### [P_TEXT]: bent ( song ) not matter. the video, directed by pedro romhanyi, seems to be a bit of a ### ### play on rob thomas'increased profile following the 1999 mega - hit, " smooth ", since it features ### ### other members of the band mistreating him. it starts with him being hit by a car driven by adam ### ### gaynor. as he lies on the ground, kyle cook shows up and empties thomas of all money in his wallet. ### ### after this, rob recovers from the car hit and walks off, singing the chorus. after the first ### ### chorus, he bumps ### ### =============================================== v_q ================================================ ### ### ('bent', 0, 0) ('.', 1, 432) ('bend', 2, 84) ('intently', 3, 329) ('¹', 4, 327) ('absently', 5, 92) ### ### ('britney', 6, 11920) ('bending', 7, 35) ('gaga', 8, 13358) ('song', 9, 31) ('100', 10, 8722) ### ### ('single', 11, 42) ('"', 12, 51) ('rihanna', 13, 866) ('貴', 14, 395) ('jackson', 15, 5446) ### ### ('hot', 16, 5732) ('bends', 17, 107) ('lyrical', 18, 25) ('！', 19, 510) ### ### =============================================== v_p ================================================ ### ### ('bent', 0, 0) ('smooth', 636, 1) ('kyle', 2149, 2) ('mega', 1165, 3) ('wallet', 8128, 4) ### ### ('adam', 969, 5) ('matter', 5992, 6) ('##yi', 28713, 7) ('thomas', 11843, 8) ('rom', 26778, 9) ### ### ('pedro', 4982, 10) ('##nor', 5458, 11) ('1999', 3022, 12) ('rob', 1178, 13) ('cook', 10154, 14) ### ### ('gay', 597, 15) ('bumps', 1490, 16) ('mist', 1723, 17) ('##han', 23018, 18) ('money', 322, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('bent', 0, 0) ('song', 9, 31) ('bending', 7, 35) ('lyrical', 18, 25) ('video', 21, 22) ### ### ('single', 11, 42) ('worried', 49, 20) ('"', 12, 51) ('smooth', 636, 1) ('videos', 35, 28) ### ### ('bend', 2, 84) ('anyway', 60, 23) ('hits', 59, 26) ('absently', 5, 92) ('chorus', 104, 21) ### ### ('hit', 91, 29) ('shocked', 107, 30) ('singles', 27, 52) ('money', 322, 19) ('adam', 969, 5) ### ############################################################################################################ [2023-11-25 11:43:40,589][root][INFO] - Epoch: 3: Step: 2001/19530, loss[v]=0.089878, lr=0.000014, acc@1[1]=998.0/1024=0.974609375, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 11:45:29,598][root][INFO] - Train batch 2100 [2023-11-25 11:45:29,599][root][INFO] - Avg. loss per last 100 batches: 0.146238 [2023-11-25 11:45:30,311][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.7/29522=99.98% | mean: 0.02 | max: 5.52 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: and 1985 transport acts, gradually deregulated the bus companies. ### ### [P_TEXT]: yorkshire rider yorkshire rider was a bus company in west yorkshire, england. to comply ### ### with the transport act 1985, the west yorkshire passenger transport executive transferred its bus ### ### operations to an arms length company named yorkshire rider on 26 october 1986. on 21 october 1988, ### ### yorkshire rider was privatised, being sold to a management buyout. in august 1989, yorkshire rider ### ### purchased west yorkshire road car company leeds and bradford garages, and associated routes, from ### ### ajs group. in 1989, yorkshire rider was awarded a contract to operate the proposed bradford ### ### trolleybus system, however the project was cancelled. in august 1990, the york businesses of ### ### reynard ### ### =============================================== v_q ================================================ ### ### ('bus', 0, 20) ('1985', 1, 13) ('.', 2, 2273) ('act', 3, 9) ('intently', 4, 166) ### ### ('absently', 5, 83) ('acts', 6, 47) ('transport', 7, 4) ('¹', 8, 310) ('##eg', 9, 27191) ### ### ('貴', 10, 317) ('anyway', 11, 12) ('der', 12, 20291) ('companies', 13, 41) ('eireann', 14, 25728) ### ### ('##ʂ', 15, 3417) ('খ', 16, 283) ('worried', 17, 7) ('cerambycidae', 18, 629) ('stared', 19, 75) ### ### =============================================== v_p ================================================ ### ### ('rider', 2310, 0) ('aj', 17792, 1) ('yorkshire', 327, 2) ('##nard', 25869, 3) ('transport', 7, 4) ### ### ('garage', 2094, 5) ('length', 2136, 6) ('worried', 17, 7) ('trolley', 44, 8) ('act', 3, 9) ### ### ('rey', 10055, 10) ('bradford', 1266, 11) ('anyway', 11, 12) ('1985', 1, 13) ('arms', 2199, 14) ### ### ('executive', 9995, 15) ('lengths', 6999, 16) ('buses', 29, 17) ('businesses', 836, 18) ### ### ('west', 22620, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('transport', 7, 4) ('1985', 1, 13) ('act', 3, 9) ('anyway', 11, 12) ('bus', 0, 20) ### ### ('worried', 17, 7) ('acts', 6, 47) ('trolley', 44, 8) ('yorkshire', 327, 2) ('companies', 13, 41) ### ### ('buses', 29, 17) ('absently', 5, 83) ('1989', 42, 27) ('intently', 4, 166) ('shocked', 134, 25) ### ### ('rider', 2310, 0) ('stared', 19, 75) ('road', 262, 26) ('bradford', 1266, 11) ('company', 149, 34) ### ############################################################################################################ [2023-11-25 11:45:30,818][root][INFO] - Epoch: 3: Step: 2101/19530, loss[v]=0.225726, lr=0.000014, acc@1[1]=956.5/1024=0.93408203125, acc@1[2]=993.5/1024=0.97021484375 [2023-11-25 11:47:20,529][root][INFO] - Train batch 2200 [2023-11-25 11:47:20,529][root][INFO] - Avg. loss per last 100 batches: 0.143131 [2023-11-25 11:47:21,259][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.69 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29516.4/29522=99.98% | mean: 0.03 | max: 5.63 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: among his clients was the watergate complex, which adler named. ### ### [P_TEXT]: watergate complex three accounts of the origin of the name, based on sources inside the ### ### development team : author and playwright warren adler, while working as a publicist for the ### ### developers, came up with the name ; nicolas salgo, a new york financier who suggested the original ### ### site to societa generale immobiliare, acquired the name from marjory hendricks, owner of the water ### ### gate inn ; and three local executives — giuseppe cecchi, an employee of societa generale ### ### immobiliare, nicolas salgo and royce ward — came up with the name, inspired in part by the water ### ### gate inn, and recommended it to executives in the rome office ### ### =============================================== v_q ================================================ ### ### ('adler', 0, 1) ('##gate', 1, 2) ('complex', 2, 5) ('water', 3, 3) ('##¨', 4, 204) ('¹', 5, 106) ### ### ('intently', 6, 296) ('quote', 7, 4058) ('york', 8, 52) ('exclaimed', 9, 161) ('parlor', 10, 668) ### ### ('.', 11, 3338) ('angrily', 12, 476) ('##½', 13, 407) ('！', 14, 2271) ('insists', 15, 127) ### ### ('hiss', 16, 2714) ('lobby', 17, 301) ('yorker', 18, 610) ('harlem', 19, 4567) ### ### =============================================== v_p ================================================ ### ### ('hendricks', 6863, 0) ('adler', 0, 1) ('##gate', 1, 2) ('water', 3, 3) ('nicolas', 15776, 4) ### ### ('complex', 2, 5) ('ward', 3950, 6) ('soc', 17782, 7) ('warren', 2615, 8) ('inn', 3622, 9) ### ### ('rome', 5690, 10) ('royce', 7198, 11) ('##bil', 25093, 12) ('##cchi', 28795, 13) ### ### ('sal', 12174, 14) ('##go', 8591, 15) ('gate', 437, 16) ('littered', 118, 17) ### ### ('giuseppe', 6733, 18) ('development', 2340, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('adler', 0, 1) ('##gate', 1, 2) ('water', 3, 3) ('complex', 2, 5) ('york', 8, 52) ### ### ('littered', 118, 17) ('anyway', 152, 20) ('¹', 5, 106) ('gate', 437, 16) ('##¨', 4, 204) ### ### ('hissed', 142, 35) ('##cier', 273, 27) ('acquainted', 82, 47) ('hotel', 55, 58) ### ### ('exclaimed', 9, 161) ('warren', 2615, 8) ('employees', 735, 28) ('insists', 15, 127) ### ### ('intently', 6, 296) ('compound', 126, 56) ### ############################################################################################################ [2023-11-25 11:47:21,763][root][INFO] - Epoch: 3: Step: 2201/19530, loss[v]=0.102733, lr=0.000014, acc@1[1]=996.5/1024=0.97314453125, acc@1[2]=1011.5/1024=0.98779296875 [2023-11-25 11:49:10,984][root][INFO] - Train batch 2300 [2023-11-25 11:49:10,985][root][INFO] - Avg. loss per last 100 batches: 0.143411 [2023-11-25 11:49:11,700][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.1/29522=99.98% | mean: 0.03 | max: 5.73 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the latter even inflicted a major defeat on jurchen forces and tried to prevent wuzhu ### ### from crossing back to the north bank of the yangtze. ### ### [P_TEXT]: battle of huangtiandang battle of huangtiandang ( ) was an early battle in the jin – song ### ### wars. from 1129, the jin forces led by wuzhu marched southward, successively conquering several ### ### important cities north of the yangtze river. wuzhu sent forces to chase after the song court, but ### ### was ambushed by song forces led by han shizhong and liang hongyu at huangtiandang, northeast of ### ### today's nanjing city. the jin army was trapped there for 48 days, but finally was able to retreat ### ### on the suggestion of a song resident. in july 1129, the ### ### =============================================== v_q ================================================ ### ### ('##zh', 0, 9) ('absently', 1, 128) ('intently', 2, 149) ('¹', 3, 80) ('ju', 4, 14135) ### ### ('##rch', 5, 25665) ('貴', 6, 309) ('wu', 7, 19) ('battle', 8, 16) ('.', 9, 122) ('##u', 10, 69) ### ### ('খ', 11, 234) ('##tze', 12, 20) ('³', 13, 693) ('yang', 14, 48) ('china', 15, 42) ### ### ('battles', 16, 77) ('anyway', 17, 7) ('manchu', 18, 8157) ('chinese', 19, 79) ### ### =============================================== v_p ================================================ ### ### ('liang', 656, 0) ('nanjing', 40, 1) ('112', 5387, 2) ('##tian', 3437, 3) ('##dan', 17996, 4) ### ### ('huang', 152, 5) ('court', 2468, 6) ('anyway', 17, 7) ('jin', 898, 8) ('##zh', 0, 9) ### ### ('wars', 231, 10) ('worried', 87, 11) ('##yu', 3996, 12) ('song', 3222, 13) ('shi', 7436, 14) ### ### ('resident', 2392, 15) ('battle', 8, 16) ('han', 312, 17) ('hong', 217, 18) ('wu', 7, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##zh', 0, 9) ('wu', 7, 19) ('battle', 8, 16) ('anyway', 17, 7) ('##tze', 12, 20) ### ### ('nanjing', 40, 1) ('china', 15, 42) ('¹', 3, 80) ('yang', 14, 48) ('absently', 1, 128) ### ### ('worried', 87, 11) ('huang', 152, 5) ('##u', 10, 69) ('intently', 2, 149) ('river', 47, 29) ### ### ('wars', 231, 10) ('battles', 16, 77) ('.', 9, 122) ('liang', 656, 0) ('hong', 217, 18) ### ############################################################################################################ [2023-11-25 11:49:12,204][root][INFO] - Epoch: 3: Step: 2301/19530, loss[v]=0.095597, lr=0.000014, acc@1[1]=996.5/1024=0.97314453125, acc@1[2]=1011.5/1024=0.98779296875 [2023-11-25 11:51:01,597][root][INFO] - Train batch 2400 [2023-11-25 11:51:01,598][root][INFO] - Avg. loss per last 100 batches: 0.152185 [2023-11-25 11:51:02,318][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.78 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29515.9/29522=99.98% | mean: 0.03 | max: 5.54 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.4/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: ; caves such as, borra caves, indian rock - cut architecture depicting undavalli caves ### ### and the country's second longest caves - the belum caves. ### ### [P_TEXT]: list of tourist attractions in visakhapatnam district visakhapatnam district is a ### ### district in the indian state of andhra pradesh. it has numerous tourist attractions, such as the ### ### simhachalam temple, kailasagiri, and the borra caves. ### ### =============================================== v_q ================================================ ### ### ('caves', 0, 2) ('##rra', 1, 1) ('bel', 2, 25003) ('bo', 3, 8) ('und', 4, 20583) ### ### ('##ava', 5, 22288) ('##um', 6, 17159) ('karnataka', 7, 452) ('.', 8, 720) ('cave', 9, 21) ### ### ('rock', 10, 5648) ('architecture', 11, 534) ('intently', 12, 112) ('absently', 13, 73) ### ### ('indian', 14, 41) ('india', 15, 66) ('cavern', 16, 143) ('cut', 17, 14877) ('mysore', 18, 344) ### ### ('crore', 19, 6521) ### ### =============================================== v_p ================================================ ### ### ('visa', 4125, 0) ('##rra', 1, 1) ('caves', 0, 2) ('sim', 9216, 3) ('##chal', 10995, 4) ### ### ('districts', 3906, 5) ('kai', 13394, 6) ('worried', 23, 7) ('bo', 3, 8) ('##agi', 22045, 9) ### ### ('andhra', 215, 10) ('attractions', 3472, 11) ('##nam', 14798, 12) ('##kha', 3558, 13) ### ### ('tourist', 374, 14) ('tourism', 176, 15) ('leaned', 32, 16) ('district', 434, 17) ### ### ('##las', 11158, 18) ('subdivisions', 10576, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##rra', 1, 1) ('caves', 0, 2) ('bo', 3, 8) ('cave', 9, 21) ('worried', 23, 7) ('leaned', 32, 16) ### ### ('indian', 14, 41) ('india', 15, 66) ('anyway', 79, 23) ('absently', 13, 73) ('andhra', 215, 10) ### ### ('tourism', 176, 15) ('##ि', 35, 33) ('intently', 12, 112) ('tourist', 374, 14) ('cradled', 25, 64) ### ### ('cavern', 16, 143) ('¹', 28, 74) ('temple', 301, 24) ('tourists', 339, 20) ### ############################################################################################################ [2023-11-25 11:51:02,825][root][INFO] - Epoch: 3: Step: 2401/19530, loss[v]=0.184930, lr=0.000014, acc@1[1]=971.0/1024=0.9482421875, acc@1[2]=1002.0/1024=0.978515625 [2023-11-25 11:52:52,263][root][INFO] - Train batch 2500 [2023-11-25 11:52:52,264][root][INFO] - Avg. loss per last 100 batches: 0.143240 [2023-11-25 11:52:53,006][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29516.0/29522=99.98% | mean: 0.03 | max: 5.64 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: originally known as " upper gilmanton ", belmont was chartered in 1727 as a parish of ### ### gilmanton. ### ### [P_TEXT]: gilmanton ironworks, new hampshire gilmanton ironworks ( alternately gilmanton iron works ### ### ) is an unincorporated community in the town of gilmanton in belknap county, new hampshire, united ### ### states. it is located near the eastern boundary of the town, along a stretch of the suncook river ### ### south of the outlet of crystal lake. new hampshire route 140 runs through the village, leading east ### ### to alton and west to the center of gilmanton and then belmont. gilmanton ironworks has a separate ### ### zip code ( 03837 ) from the ### ### =============================================== v_q ================================================ ### ### ('belmont', 0, 1) ('gil', 1, 4) ('##man', 2, 13) ('##ton', 3, 19) ('upper', 4, 3252) ### ### ('shillings', 5, 25147) ('¹', 6, 180) ('.', 7, 3410) ('parish', 8, 810) ('jamaica', 9, 3639) ### ### ('wales', 10, 12454) ('intently', 11, 107) ('barbados', 12, 1661) ('town', 13, 31) ### ### ('cornwall', 14, 3808) ('norfolk', 15, 4395) ('massachusetts', 16, 52) ('england', 17, 3117) ### ### ('somerset', 18, 2883) ('isle', 19, 19231) ### ### =============================================== v_p ================================================ ### ### ('alton', 16995, 0) ('belmont', 0, 1) ('crystal', 20468, 2) ('140', 5420, 3) ('gil', 1, 4) ### ### ('bel', 22480, 5) ('##ok', 22805, 6) ('##works', 221, 7) ('sun', 3247, 8) ('lake', 4798, 9) ### ### ('hampshire', 27, 10) ('worried', 2180, 11) ('iron', 5397, 12) ('##man', 2, 13) ### ### ('outlet', 16832, 14) ('##co', 14330, 15) ('zip', 6659, 16) ('anyway', 2477, 17) ### ### ('route', 1001, 18) ('##ton', 3, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('belmont', 0, 1) ('gil', 1, 4) ('##man', 2, 13) ('##ton', 3, 19) ('hampshire', 27, 10) ### ### ('town', 13, 31) ('##works', 221, 7) ('massachusetts', 16, 52) ('village', 305, 23) ### ### ('towns', 22, 64) ('¹', 6, 180) ('nearby', 86, 38) ('cradled', 276, 28) ('route', 1001, 18) ### ### ('intently', 11, 107) ('residents', 65, 53) ('river', 728, 22) ('140', 5420, 3) ### ### ('worried', 2180, 11) ('sun', 3247, 8) ### ############################################################################################################ [2023-11-25 11:52:53,510][root][INFO] - Epoch: 3: Step: 2501/19530, loss[v]=0.223864, lr=0.000014, acc@1[1]=974.5/1024=0.95166015625, acc@1[2]=998.5/1024=0.97509765625 [2023-11-25 11:54:42,347][root][INFO] - Train batch 2600 [2023-11-25 11:54:42,348][root][INFO] - Avg. loss per last 100 batches: 0.149701 [2023-11-25 11:54:43,047][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29515.2/29522=99.98% | mean: 0.03 | max: 5.61 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: tony canadeo who later played 11 seasons for the green bay packers and was inducted into ### ### the pro football hall of fame. ### ### [P_TEXT]: gonzaga – idaho football rivalry idaho did not have a varsity team the pacific coast ### ### conference ( pcc ) in 1922, while gonzaga remained independent. gonzaga's dominating wins in 1939 ### ### and 1940, both shutouts, were led by halfback tony canadeo, a future member of the pro football ### ### hall of fame. with canadeo in the nfl, idaho rebounded in 1941 to win 21 – 7 in spokane in what was ### ### the final game in during world war ii, gonzaga did not field a team after 1941, while idaho played ### ### in 1942 and went on hiatus in after the war, idaho resumed ### ### =============================================== v_q ================================================ ### ### ('##ade', 0, 0) ('fame', 1, 18) ('absently', 2, 135) ('tony', 3, 6) ('can', 4, 14) ('¹', 5, 153) ### ### ('packers', 6, 1465) ('intently', 7, 112) ('貴', 8, 353) ('nfl', 9, 16) ('pro', 10, 21) ### ### ('##o', 11, 22) ('bay', 12, 12660) ('green', 13, 5495) ('football', 14, 26) ('³', 15, 1537) ### ### ('hall', 16, 54) ('couldn', 17, 57) ('##ப', 18, 168) ('jerseys', 19, 3445) ### ### =============================================== v_p ================================================ ### ### ('##ade', 0, 0) ('gonzaga', 5231, 1) ('rivalry', 740, 2) ('spokane', 1208, 3) ('idaho', 527, 4) ### ### ('1922', 2350, 5) ('tony', 3, 6) ('rebound', 1760, 7) ('1939', 3891, 8) ('1940', 3005, 9) ### ### ('pc', 11513, 10) ('1941', 1635, 11) ('anyway', 40, 12) ('1942', 1355, 13) ('can', 4, 14) ### ### ('pacific', 7924, 15) ('nfl', 9, 16) ('varsity', 2540, 17) ('fame', 1, 18) ('worried', 179, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ade', 0, 0) ('tony', 3, 6) ('can', 4, 14) ('fame', 1, 18) ('nfl', 9, 16) ('pro', 10, 21) ### ### ('##o', 11, 22) ('football', 14, 26) ('anyway', 40, 12) ('absently', 2, 135) ('intently', 7, 112) ### ### ('¹', 5, 153) ('hall', 16, 54) ('couldn', 17, 57) ('field', 35, 31) ('idaho', 527, 4) ### ### ('rivalry', 740, 2) ('worried', 179, 19) ('spokane', 1208, 3) ('game', 58, 39) ### ############################################################################################################ [2023-11-25 11:54:43,552][root][INFO] - Epoch: 3: Step: 2601/19530, loss[v]=0.170584, lr=0.000014, acc@1[1]=974.5/1024=0.95166015625, acc@1[2]=999.0/1024=0.9755859375 [2023-11-25 11:56:32,941][root][INFO] - Train batch 2700 [2023-11-25 11:56:32,942][root][INFO] - Avg. loss per last 100 batches: 0.148657 [2023-11-25 11:56:33,664][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.74 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29515.8/29522=99.98% | mean: 0.03 | max: 5.58 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: emerging from the new wave scene in the late 1970s, the line - up consisted of elliot ### ### easton ( lead guitar ), greg hawkes ( keyboards ), ric ocasek ( rhythm guitar ), benjamin orr ( ### ### bass guitar ), and david robinson ( drums ). ### ### [P_TEXT]: ohio of the black keys, griffin layne, joe dolce, kid cudi, benjamin orr of the cars, ### ### chrissie hynde of the pretenders, william " bootsy " collins, stephanie eulinberg of kid rock's ### ### twisted brown trucker band, and devo. five ohio musicians are rock and roll hall of fame members. ### ### playhouse square in downtown cleveland is the second largest performing arts center in the united ### ### states, home to ten theaters. the ### ### =============================================== v_q ================================================ ### ### ('orr', 0, 2) ('hawke', 1, 27620) ('easton', 2, 17738) ('robinson', 3, 7168) ('##case', 4, 15502) ### ### ('ric', 5, 27676) ('elliot', 6, 4917) ('benjamin', 7, 18) ('absently', 8, 141) ('greg', 9, 15302) ### ### ('wave', 10, 7739) ('david', 11, 15236) ('.', 12, 95) ('scene', 13, 2301) ('band', 14, 39) ### ### ('##k', 15, 27123) ('intently', 16, 246) ('o', 17, 17907) ('musician', 18, 119) ### ### ('reggae', 19, 6263) ### ### =============================================== v_p ================================================ ### ### ('ohio', 2557, 0) ('##lce', 28805, 1) ('orr', 0, 2) ('playhouse', 10376, 3) ('keys', 589, 4) ### ### ('theaters', 25077, 5) ('layne', 9135, 6) ('boots', 2229, 7) ('cleveland', 1631, 8) ### ### ('cars', 5306, 9) ('##sie', 7083, 10) ('dev', 4966, 11) ('twisted', 332, 12) ('griffin', 5078, 13) ### ### ('truck', 12145, 14) ('pretend', 789, 15) ('fame', 465, 16) ('collins', 3472, 17) ### ### ('benjamin', 7, 18) ('downtown', 10712, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('orr', 0, 2) ('benjamin', 7, 18) ('band', 14, 39) ('absently', 8, 141) ('.', 12, 95) ### ### ('bands', 20, 59) ('¹', 23, 61) ('rock', 49, 42) ('twisted', 332, 12) ('hissed', 74, 36) ### ### ('keys', 589, 4) ('frontman', 94, 44) ('fame', 465, 16) ('musicians', 96, 46) ('album', 42, 62) ### ### ('musician', 18, 119) ('ohio', 2557, 0) ('roll', 334, 24) ('pretend', 789, 15) ('drummer', 30, 80) ### ############################################################################################################ [2023-11-25 11:56:34,170][root][INFO] - Epoch: 3: Step: 2701/19530, loss[v]=0.164431, lr=0.000014, acc@1[1]=973.0/1024=0.9501953125, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-25 11:58:23,479][root][INFO] - Train batch 2800 [2023-11-25 11:58:23,480][root][INFO] - Avg. loss per last 100 batches: 0.155087 [2023-11-25 11:58:24,181][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.74 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29517.1/29522=99.98% | mean: 0.03 | max: 5.59 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: bridgnorth is connected to shrewsbury by the a458 road, to telford by the a442 road and ### ### wolverhampton via the a454 road. ### ### [P_TEXT]: eardington eardington is a small village and civil parish in shropshire, england. it is ### ### near the a442 road and is two kilometres south of the town of bridgnorth, along the b4555 road. the ### ### population taken at the 2011 census is shown under bridgnorth. the severn valley railway runs ### ### immediately to the east of the village and there was once a stop on the line, situated about half a ### ### mile to the south, between upper and lower forge, called eardington halt. the half - mile " ### ### eardington forge canal " ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 102) ('wolverhampton', 1, 162) ('##45', 2, 5) ('##rth', 3, 10) ('shrewsbury', 4, 19) ### ### ('##44', 5, 3) ('br', 6, 22) ('##ford', 7, 15129) ('##id', 8, 20) ('##gno', 9, 6) ### ### ('absently', 10, 139) ('intently', 11, 140) ('shropshire', 12, 9) ('³', 13, 1578) ### ### ('tel', 14, 20252) ('.', 15, 2613) ('shillings', 16, 14543) ('##4', 17, 701) ('##2', 18, 17) ### ### ('貴', 19, 164) ### ### =============================================== v_p ================================================ ### ### ('forge', 2902, 0) ('##dington', 2109, 1) ('ear', 7079, 2) ('##44', 5, 3) ('severn', 87, 4) ### ### ('##45', 2, 5) ('##gno', 9, 6) ('##55', 7463, 7) ('worried', 827, 8) ('shropshire', 12, 9) ### ### ('##rth', 3, 10) ('upper', 756, 11) ('lower', 1628, 12) ('anyway', 299, 13) ('leaned', 393, 14) ### ### ('railway', 1978, 15) ('valley', 186, 16) ('##2', 18, 17) ('town', 23, 18) ('shrewsbury', 4, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##44', 5, 3) ('##45', 2, 5) ('##rth', 3, 10) ('##gno', 9, 6) ('shropshire', 12, 9) ### ### ('shrewsbury', 4, 19) ('br', 6, 22) ('##id', 8, 20) ('##2', 18, 17) ('severn', 87, 4) ### ### ('town', 23, 18) ('¹', 0, 102) ('road', 28, 27) ('welsh', 33, 24) ('wolverhampton', 1, 162) ### ### ('valley', 186, 16) ('anyway', 299, 13) ('##dington', 2109, 1) ('absently', 10, 139) ### ### ('worried', 827, 8) ### ############################################################################################################ [2023-11-25 11:58:24,684][root][INFO] - Epoch: 3: Step: 2801/19530, loss[v]=0.205898, lr=0.000014, acc@1[1]=974.0/1024=0.951171875, acc@1[2]=998.0/1024=0.974609375 [2023-11-25 12:00:13,216][root][INFO] - Train batch 2900 [2023-11-25 12:00:13,217][root][INFO] - Avg. loss per last 100 batches: 0.150502 [2023-11-25 12:00:13,927][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29515.9/29522=99.98% | mean: 0.02 | max: 5.48 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 2003, the government arrested and imprisoned a large number of civil activists, a ### ### period known as the " black spring ". ### ### [P_TEXT]: omar pernet hernandez omar pernet hernandez is a cuba n librarian, human rights activist, ### ### and a former political prisoner. he was imprisoned during the black spring in 2003 and released to ### ### exile in spain in 2008. amnesty international considered him a prisoner of conscience. ### ### =============================================== v_q ================================================ ### ### ('spring', 0, 7) ('black', 1, 8) ('intently', 2, 172) ('absently', 3, 116) ('2003', 4, 10) ### ### ('¹', 5, 468) ('anyway', 6, 16) ('.', 7, 143) ('！', 8, 3860) ('activists', 9, 46) ('貴', 10, 513) ### ### ('imprisonment', 11, 409) ('jailed', 12, 788) ('activist', 13, 61) ('civil', 14, 6735) ### ### ('myanmar', 15, 10713) ('bahrain', 16, 2221) ('albania', 17, 1759) ('protests', 18, 40) ### ### ('detained', 19, 347) ### ### =============================================== v_p ================================================ ### ### ('conscience', 1740, 0) ('omar', 5051, 1) ('hernandez', 11739, 2) ('librarian', 25250, 3) ### ### ('amnesty', 470, 4) ('cuba', 1841, 5) ('exile', 89, 6) ('spring', 0, 7) ('black', 1, 8) ### ### ('##net', 13660, 9) ('2003', 4, 10) ('prisoner', 193, 11) ('worried', 23, 12) ('leaned', 1498, 13) ### ### ('human', 601, 14) ('per', 22245, 15) ('anyway', 6, 16) ('prisoners', 302, 17) ('spain', 3841, 18) ### ### ('imprisoned', 28, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('spring', 0, 7) ('black', 1, 8) ('2003', 4, 10) ('anyway', 6, 16) ('worried', 23, 12) ### ### ('exile', 89, 6) ('imprisoned', 28, 19) ('activists', 9, 46) ('amnesty', 470, 4) ### ### ('prisoner', 193, 11) ('activist', 13, 61) ('protests', 18, 40) ('absently', 3, 116) ### ### ('intently', 2, 172) ('prisoners', 302, 17) ('arrested', 21, 44) ('political', 77, 33) ### ### ('conscience', 1740, 0) ('.', 7, 143) ('blacks', 275, 22) ### ############################################################################################################ [2023-11-25 12:00:14,432][root][INFO] - Epoch: 3: Step: 2901/19530, loss[v]=0.148938, lr=0.000014, acc@1[1]=987.5/1024=0.96435546875, acc@1[2]=1008.0/1024=0.984375 [2023-11-25 12:02:03,456][root][INFO] - Train batch 3000 [2023-11-25 12:02:03,457][root][INFO] - Avg. loss per last 100 batches: 0.157401 [2023-11-25 12:02:04,167][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29516.1/29522=99.98% | mean: 0.03 | max: 5.73 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: " i need to be in love " is a song written by richard carpenter, albert hammond and john ### ### bettis. ### ### [P_TEXT]: richard carpenter ( musician ) on the album " now & then " ( 1973 ). the couple had been ### ### dating since the late 1970s. a young mary made a cameo appearance in the carpenters'promotional ### ### video for the song " i need to be in love " ( 1976 ). richard and mary carpenter have five children ### ### : kristi lynn ( the name karen would have given to a daughter ) was born on august 17, 1987 ; traci ### ### tatum on july 25, 1989 ; mindi karen ( named after her late paternal aunt ) on july 7, 1992 ; ### ### collin paul on july 20, 1994 ; and taylor mary on december 5, ### ### =============================================== v_q ================================================ ### ### ('need', 0, 4) ('carpenter', 1, 0) ('″', 2, 18) ('love', 3, 15) ('hammond', 4, 14457) ('¹', 5, 108) ### ### ('##tis', 6, 9811) ('bet', 7, 5500) ('richard', 8, 3) ('и', 9, 49) ('albert', 10, 9432) ### ### ('intently', 11, 81) ('i', 12, 42) ('be', 13, 34) ('³', 14, 1329) ('needed', 15, 70) ('"', 16, 154) ### ### ('貴', 17, 296) ('royalties', 18, 4812) ('absently', 19, 101) ### ### =============================================== v_p ================================================ ### ### ('carpenter', 1, 0) ('collin', 7410, 1) ('tatum', 13407, 2) ('richard', 8, 3) ('need', 0, 4) ### ### ('karen', 2725, 5) ('mind', 4928, 6) ('now', 19011, 7) ('kris', 14550, 8) ('taylor', 1345, 9) ### ### ('mary', 8716, 10) ('aunt', 14757, 11) ('1973', 100, 12) ('1976', 52, 13) ('lynn', 3608, 14) ### ### ('love', 3, 15) ('promotional', 2636, 16) ('##ac', 24003, 17) ('″', 2, 18) ('&', 107, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('carpenter', 1, 0) ('need', 0, 4) ('richard', 8, 3) ('love', 3, 15) ('″', 2, 18) ('и', 9, 49) ### ### ('be', 13, 34) ('i', 12, 42) ('¹', 5, 108) ('intently', 11, 81) ('1976', 52, 13) ('needed', 15, 70) ### ### ('1973', 100, 12) ('&', 107, 19) ('song', 21, 59) ('singles', 30, 47) ('ヘ', 25, 57) ### ### ('lyrical', 104, 28) ('album', 99, 29) ('songs', 22, 85) ### ############################################################################################################ [2023-11-25 12:02:04,672][root][INFO] - Epoch: 3: Step: 3001/19530, loss[v]=0.087304, lr=0.000014, acc@1[1]=998.5/1024=0.97509765625, acc@1[2]=1014.5/1024=0.99072265625 [2023-11-25 12:03:53,086][root][INFO] - Train batch 3100 [2023-11-25 12:03:53,087][root][INFO] - Avg. loss per last 100 batches: 0.147493 [2023-11-25 12:03:53,781][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.4/29522=99.97% | mean: 0.02 | max: 5.74 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: this novel sees the first named appearance of the villain sabbath, who subsequently ### ### appears in many of the following novels. ### ### [P_TEXT]: sabbath ( doctor who ) sabbath is the name of a recurring villain from the eighth doctor ### ### adventures — spin - off novels based on the bbc science fiction television series " doctor who ". ### ### the character was created by lawrence miles and first appeared in " the adventuress of henrietta ### ### street ". originally, miles had intended sabbath to be a one - off character, but bbc books editor ### ### justin richards asked to use the character in a continuing story arc. sabbath was born in 1740. he ### ### was educated at cambridge before being initiated into ### ### =============================================== v_q ================================================ ### ### ('sabbath', 0, 0) ('novel', 1, 27) ('novels', 2, 22) ('.', 3, 370) ('absently', 4, 97) ### ### ('villain', 5, 17) ('evil', 6, 359) ('″', 7, 32) ('¹', 8, 366) ('intently', 9, 130) ### ### ('trilogy', 10, 462) ('tolkien', 11, 8664) ('doctor', 12, 3) ('burroughs', 13, 7990) ### ### ('！', 14, 1874) ('theft', 15, 5714) ('disguised', 16, 17845) ('ronan', 17, 5007) ### ### ('sequels', 18, 4172) ('implied', 19, 2148) ### ### =============================================== v_p ================================================ ### ### ('sabbath', 0, 0) ('henrietta', 12457, 1) ('richards', 3619, 2) ('doctor', 12, 3) ### ### ('justin', 19869, 4) ('miles', 11713, 5) ('adventures', 190, 6) ('lawrence', 9147, 7) ### ### ('worried', 130, 8) ('leaned', 5431, 9) ('editor', 5428, 10) ('eighth', 4567, 11) ### ### ('cambridge', 7689, 12) ('1740', 6808, 13) ('anyway', 261, 14) ('editors', 23868, 15) ### ### ('bbc', 272, 16) ('villain', 5, 17) ('programmes', 2725, 18) ('street', 5230, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('sabbath', 0, 0) ('doctor', 12, 3) ('novels', 2, 22) ('novel', 1, 27) ('villain', 5, 17) ### ### ('″', 7, 32) ('books', 22, 21) ('worried', 130, 8) ('adventures', 190, 6) ('villains', 27, 23) ### ### ('absently', 4, 97) ('characters', 53, 25) ('anyway', 261, 14) ('bbc', 272, 16) ### ### ('character', 85, 31) ('intently', 9, 130) ('richards', 3619, 2) ('.', 3, 370) ### ### ('creatures', 136, 37) ('isbn', 142, 40) ### ############################################################################################################ [2023-11-25 12:03:54,284][root][INFO] - Epoch: 3: Step: 3101/19530, loss[v]=0.124002, lr=0.000014, acc@1[1]=995.5/1024=0.97216796875, acc@1[2]=1012.0/1024=0.98828125 [2023-11-25 12:05:42,644][root][INFO] - Train batch 3200 [2023-11-25 12:05:42,645][root][INFO] - Avg. loss per last 100 batches: 0.151466 [2023-11-25 12:05:43,358][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.75 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29516.6/29522=99.98% | mean: 0.03 | max: 5.63 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: bangladesh, china, papua new guinea, and scotland participated in the women's world ### ### twenty20 qualifier for the first time. ### ### [P_TEXT]: huang zhuo huang zhuo (, born 13 august 1985 ) is a chinese woman cricketer. she ### ### captained the chinese women's cricket team during the 2015 icc women's world twenty20 qualifier. ### ### she also represented china at the 2010 asian games and in the 2014 asian games. she made her ### ### women's twenty20 international ( wt20i ) debut for china on 18 february 2019 against thailand, in ### ### the 2019 icc women's qualifier asia tournament. ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 79) ('cricket', 1, 25) ('bangladesh', 2, 1263) ('twenty20', 3, 22) ### ### ('scotland', 4, 2875) ('women', 5, 5) ('¹', 6, 305) ('papua', 7, 8110) ('china', 8, 2) ### ### ('intently', 9, 92) ('貴', 10, 196) ('qualifier', 11, 6) ('.', 12, 66) ('guinea', 13, 7240) ### ### ('ɐ', 14, 341) ('world', 15, 68) ('খ', 16, 162) ('pakistan', 17, 331) ('india', 18, 126) ### ### ('team', 19, 41) ### ### =============================================== v_p ================================================ ### ### ('thailand', 2034, 0) ('huang', 1570, 1) ('china', 8, 2) ('zhu', 4557, 3) ('qualifiers', 77, 4) ### ### ('women', 5, 5) ('qualifier', 11, 6) ('woman', 270, 7) ('games', 60, 8) ('asia', 205, 9) ### ### ('anyway', 76, 10) ('chinese', 42, 11) ('asian', 40, 12) ('tournament', 58, 13) ('icc', 116, 14) ### ### ('leaned', 85, 15) ('qualification', 97, 16) ('2019', 52, 17) ('worried', 46, 18) ('2010', 130, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('china', 8, 2) ('women', 5, 5) ('qualifier', 11, 6) ('cricket', 1, 25) ('twenty20', 3, 22) ### ### ('asian', 40, 12) ('absently', 0, 79) ('chinese', 42, 11) ('qualifiers', 77, 4) ('games', 60, 8) ### ### ('tournament', 58, 13) ('2019', 52, 17) ('worried', 46, 18) ('anyway', 76, 10) ('.', 12, 66) ### ### ('intently', 9, 92) ('2015', 64, 21) ('leaned', 85, 15) ('team', 19, 41) ('qualification', 97, 16) ### ############################################################################################################ [2023-11-25 12:05:43,864][root][INFO] - Epoch: 3: Step: 3201/19530, loss[v]=0.174358, lr=0.000014, acc@1[1]=977.5/1024=0.95458984375, acc@1[2]=1001.0/1024=0.9775390625 [2023-11-25 12:07:32,779][root][INFO] - Train batch 3300 [2023-11-25 12:07:32,779][root][INFO] - Avg. loss per last 100 batches: 0.150822 [2023-11-25 12:07:33,482][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.70 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29511.0/29522=99.96% | mean: 0.02 | max: 5.31 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 73.4/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: thomas jefferson rusk was one of the most prominent early nacogdoches anglo settlers. ### ### [P_TEXT]: convention of 1836 election. in the nacogdoches district, soldiers under sidney sherman ### ### threatened violence after they were turned away from the polling place. sherman vowed that he " had ### ### come to texas to fight for it and has as soon commence in the town of nacogdoches as elsewhere. " ### ### anxious to avoid an armed fight, election judges asked the public to vote on whether the troops ### ### should be allowed to vote. the soldiers lost by 30 votes, yet refused to back down, insisting that ### ### the voting had been rigged. finally, one of the candidates, thomas jefferson rusk, ### ### =============================================== v_q ================================================ ### ### ('rus', 0, 4) ('¹', 1, 241) ('texas', 2, 11) ('jefferson', 3, 6) ('intently', 4, 163) ### ### ('louisiana', 5, 113) ('absently', 6, 155) ('thomas', 7, 37) ('##k', 8, 39) ('##co', 9, 16) ### ### ('na', 10, 13) ('##ches', 11, 5) ('貴', 12, 461) ('##do', 13, 26) ('³', 14, 3910) ### ### ('proficient', 15, 7077) ('county', 16, 2973) ('খ', 17, 238) ('幸', 18, 299) ('oregon', 19, 161) ### ### =============================================== v_p ================================================ ### ### ('sherman', 503, 0) ('sidney', 3845, 1) ('1836', 1774, 2) ('convention', 3244, 3) ('rus', 0, 4) ### ### ('##ches', 11, 5) ('jefferson', 3, 6) ('soldiers', 342, 7) ('conventions', 28513, 8) ### ### ('rigged', 26624, 9) ('violence', 2338, 10) ('texas', 2, 11) ('anyway', 79, 12) ('na', 10, 13) ### ### ('town', 412, 14) ('littered', 1702, 15) ('##co', 9, 16) ('troops', 2302, 17) ('judges', 16320, 18) ### ### ('voting', 9987, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('rus', 0, 4) ('jefferson', 3, 6) ('texas', 2, 11) ('##ches', 11, 5) ('na', 10, 13) ('##co', 9, 16) ### ### ('##do', 13, 26) ('thomas', 7, 37) ('##k', 8, 39) ('louisiana', 5, 113) ('anyway', 79, 12) ### ### ('¹', 1, 241) ('sherman', 503, 0) ('intently', 4, 163) ('absently', 6, 155) ('residents', 21, 60) ### ### ('soldiers', 342, 7) ('##g', 24, 55) ('worried', 83, 24) ('neighboring', 38, 45) ### ############################################################################################################ [2023-11-25 12:07:33,987][root][INFO] - Epoch: 3: Step: 3301/19530, loss[v]=0.229471, lr=0.000014, acc@1[1]=962.5/1024=0.93994140625, acc@1[2]=995.5/1024=0.97216796875 [2023-11-25 12:09:22,359][root][INFO] - Train batch 3400 [2023-11-25 12:09:22,360][root][INFO] - Avg. loss per last 100 batches: 0.152504 [2023-11-25 12:09:23,052][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29515.9/29522=99.98% | mean: 0.02 | max: 5.43 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: porta cordillera was established in july 2012 by the puerto rico tourism company. ### ### [P_TEXT]: porta caribe porta caribe is a tourism region in southern puerto rico. it was established ### ### in 2003 by the puerto rico tourism company, an agency of the government of puerto rico. when ### ### created in 2003 it consisted of 14 municipalities in the south central zone ( adjuntas, arroyo, ### ### coamo, guayama, guayanilla, jayuya, juana diaz, patillas, penuelas, ponce, salinas, santa isabel, ### ### villaba, yauco ). with the creation of the neighboring porta cordillera zone in ### ### =============================================== v_q ================================================ ### ### ('port', 0, 7) ('¹', 1, 28) ('cord', 2, 47) ('rico', 3, 27) ('puerto', 4, 6) ('tourism', 5, 0) ### ### ('##ille', 6, 31) ('company', 7, 29) ('³', 8, 124) ('intently', 9, 75) ('貴', 10, 130) ### ### ('absently', 11, 163) ('barangay', 12, 5887) ('##ra', 13, 277) ('resort', 14, 823) ('##ப', 15, 114) ### ### ('##a', 16, 92) ('treasure', 17, 1628) ('nautical', 18, 16403) ('খ', 19, 144) ### ### =============================================== v_p ================================================ ### ### ('tourism', 5, 0) ('##ibe', 16959, 1) ('arroyo', 132, 2) ('ponce', 312, 3) ('##jun', 29006, 4) ### ### ('coa', 8813, 5) ('puerto', 4, 6) ('port', 0, 7) ('##illa', 1160, 8) ('##uc', 24692, 9) ### ### ('anyway', 214, 10) ('##uel', 14844, 11) ('worried', 1200, 12) ('government', 6754, 13) ### ### ('zones', 25961, 14) ('##aya', 16140, 15) ('zone', 2051, 16) ('##ayan', 14308, 17) ### ### ('villa', 4920, 18) ('##inas', 6586, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('tourism', 5, 0) ('port', 0, 7) ('puerto', 4, 6) ('¹', 1, 28) ('rico', 3, 27) ('##ille', 6, 31) ### ### ('company', 7, 29) ('cord', 2, 47) ('intently', 9, 75) ('arroyo', 132, 2) ('tourists', 65, 36) ### ### ('³', 8, 124) ('ponce', 312, 3) ('anyway', 214, 10) ('貴', 10, 130) ('tourist', 34, 53) ### ### ('isabel', 167, 24) ('absently', 11, 163) ('companies', 88, 50) ('##a', 16, 92) ### ############################################################################################################ [2023-11-25 12:09:23,556][root][INFO] - Epoch: 3: Step: 3401/19530, loss[v]=0.232294, lr=0.000014, acc@1[1]=980.0/1024=0.95703125, acc@1[2]=1004.0/1024=0.98046875 [2023-11-25 12:11:12,026][root][INFO] - Train batch 3500 [2023-11-25 12:11:12,028][root][INFO] - Avg. loss per last 100 batches: 0.151490 [2023-11-25 12:11:12,744][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29516.1/29522=99.98% | mean: 0.03 | max: 5.60 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: robinson left ireland in 1923, selling his foxrock house, and went to live in ealing as ### ### his life was under threat. ### ### [P_TEXT]: lynch - robinson baronets the robinson, later lynch - robinson baronetcy, of lisnacarrig, ### ### foxrock in the county of dublin, is a title in the baronetage of the united kingdom. it was created ### ### on 30 january 1920 for sir henry robinson, kcb, vice - president of the local government board for ### ### ireland from 1898 to 1920. the second baronet assumed in 1947 by deed poll the additional surname ### ### of lynch. ### ### =============================================== v_q ================================================ ### ### ('robinson', 0, 1) ('ea', 1, 14879) ('##ling', 2, 15605) ('##rock', 3, 0) ('fox', 4, 5) ### ### ('ireland', 5, 18) ('¹', 6, 373) ('house', 7, 6462) ('threat', 8, 8305) ('retorted', 9, 362) ### ### ('realised', 10, 518) ('irish', 11, 43) ('bungalow', 12, 15525) ('1923', 13, 647) ### ### ('mayfair', 14, 8737) ('danger', 15, 11719) ('intently', 16, 202) ('estate', 17, 1243) ### ### ('lived', 18, 1572) ('№', 19, 1354) ### ### =============================================== v_p ================================================ ### ### ('##rock', 3, 0) ('robinson', 0, 1) ('lynch', 1838, 2) ('baronetcy', 9597, 3) ('##nac', 20298, 4) ### ### ('fox', 4, 5) ('dublin', 24, 6) ('baronet', 353, 7) ('kc', 20688, 8) ('worried', 4771, 9) ### ### ('deed', 4898, 10) ('board', 639, 11) ('li', 22096, 12) ('poll', 14943, 13) ### ### ('additional', 17946, 14) ('1920', 251, 15) ('boards', 20181, 16) ('##ri', 17945, 17) ### ### ('ireland', 5, 18) ('anyway', 2517, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('robinson', 0, 1) ('##rock', 3, 0) ('fox', 4, 5) ('ireland', 5, 18) ('dublin', 24, 6) ### ### ('irish', 11, 43) ('baronet', 353, 7) ('1920', 251, 15) ('board', 639, 11) ('uk', 50, 42) ### ### ('lynch', 1838, 2) ('1947', 117, 35) ('local', 811, 25) ('henry', 1176, 21) ('¹', 6, 373) ### ### ('1898', 660, 29) ('1890s', 32, 86) ('british', 40, 90) ('britain', 282, 50) ('sir', 610, 40) ### ############################################################################################################ [2023-11-25 12:11:13,252][root][INFO] - Epoch: 3: Step: 3501/19530, loss[v]=0.187887, lr=0.000014, acc@1[1]=973.5/1024=0.95068359375, acc@1[2]=999.0/1024=0.9755859375 [2023-11-25 12:13:02,068][root][INFO] - Train batch 3600 [2023-11-25 12:13:02,069][root][INFO] - Avg. loss per last 100 batches: 0.156890 [2023-11-25 12:13:02,777][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.76 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.7/29522=99.98% | mean: 0.02 | max: 5.64 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: 1933, include huntsville unit ( 1849 ), wynne unit ( 1883 ), jester i unit ( 1885, brick ### ### building in 1932 ), vance ( harlem / jester ii ) unit ( 1885, brick building in 1933 ), clemens ### ### unit ( 1893 ), ramsey ( i ) unit ( 1908 ), stringfellow ( ramsey ii ) unit ( 1908 ), central unit ( ### ### 1909, rebuilt in 1932 ), goree unit ( 1907 ), darrington unit ( 1917 ), eastham unit ( 1917 ), and ### ### scott ( retrieve ) unit ( 1919 ). ### ### [P_TEXT]: wayne scott unit the wayne scott unit is a texas department of criminal justice ( tdcj ) ### ### prison farm located in unincorporated brazoria county, texas. the unit, southwest of houston, is ### ### along county road 290, south of angleton. scott, which was established in september 1919, has about ### ### of land. the prison was formerly known as the retrieve unit. the prison opened on the grounds of ### ### the former retrieve plantation, which opened in 1839. in 1935 retrieve housed white prisoners. in ### ### 1963, before racial desegregation occurred, the facility ### ### =============================================== v_q ================================================ ### ### ('unit', 0, 6) ('clemens', 1, 14780) ('vance', 2, 2714) ('ramsey', 3, 19444) ('retrieve', 4, 3) ### ### ('harlem', 5, 2469) ('absently', 6, 124) ('##ster', 7, 26872) ('je', 8, 21443) ### ### ('##rington', 9, 26290) ('east', 10, 5221) ('wynn', 11, 11591) ('gore', 12, 8903) ### ### ('brick', 13, 4046) ('##ham', 14, 15333) ('huntsville', 15, 923) ('central', 16, 11990) ### ### ('units', 17, 35) ('scott', 18, 2) ('dar', 19, 16606) ### ### =============================================== v_p ================================================ ### ### ('290', 21823, 0) ('angle', 14148, 1) ('scott', 18, 2) ('retrieve', 4, 3) ('racial', 1104, 4) ### ### ('wayne', 7828, 5) ('unit', 0, 6) ('plantation', 2550, 7) ('bra', 21748, 8) ('td', 19295, 9) ### ### ('prisons', 18743, 10) ('texas', 576, 11) ('houston', 168, 12) ('prison', 112, 13) ### ### ('##gation', 2889, 14) ('plantations', 12244, 15) ('##zo', 26184, 16) ('##ria', 25773, 17) ### ### ('worried', 48, 18) ('farm', 1678, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('unit', 0, 6) ('retrieve', 4, 3) ('scott', 18, 2) ('county', 21, 30) ('units', 17, 35) ### ### ('worried', 48, 18) ('department', 41, 22) ('prison', 112, 13) ('leaned', 31, 41) ### ### ('houston', 168, 12) ('absently', 6, 124) ('driveway', 46, 42) ('texas', 576, 11) ### ### ('racial', 1104, 4) ('1919', 61, 48) ('anyway', 487, 25) ('plantation', 2550, 7) ### ### ('facility', 91, 60) ('intently', 25, 219) ('farm', 1678, 19) ### ############################################################################################################ [2023-11-25 12:13:03,282][root][INFO] - Epoch: 3: Step: 3601/19530, loss[v]=0.069718, lr=0.000014, acc@1[1]=1000.5/1024=0.97705078125, acc@1[2]=1011.5/1024=0.98779296875 [2023-11-25 12:14:51,748][root][INFO] - Train batch 3700 [2023-11-25 12:14:51,749][root][INFO] - Avg. loss per last 100 batches: 0.158298 [2023-11-25 12:14:52,459][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29515.7/29522=99.98% | mean: 0.03 | max: 5.83 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: they captured a convoy of livestock meant for the greek camp and forced a battle, in ### ### which boioannes was defeated and captured ( september 3, 1041 ). ### ### [P_TEXT]: battle of montepeloso lombard - norman revolt against the byzantine empire in southern ### ### italy, with preceding battles at olivento in march, and at montemaggiore in may, both rebel ### ### victories. before the battle, the normans and lombards agreed on choosing atenulf, brother of ### ### pandulf iii of benevento, as their new leader, while the byzantine catepan michael dokeianos was ### ### replaced by exaugustus boioannes. the new overall leader of the revolt, atenulf, was part of the ### ### ruling beneventan family and a rival to the norman sponsor guaimar iv of salerno, but although the ### ### =============================================== v_q ================================================ ### ### ('104', 0, 2063) ('intently', 1, 171) ('absently', 2, 271) ('##anne', 3, 7) ('##io', 4, 56) ### ### ('bo', 5, 50) ('¹', 6, 78) ('greek', 7, 100) ('sparhawk', 8, 562) ('貴', 9, 228) ('106', 10, 8656) ### ### ('battle', 11, 13) ('.', 12, 192) ('³', 13, 1387) ('খ', 14, 216) ('105', 15, 3983) ### ### ('byzantine', 16, 11) ('convoy', 17, 6974) ('forced', 18, 2268) ('anyway', 19, 23) ### ### =============================================== v_p ================================================ ### ### ('##pel', 22067, 0) ('olive', 1780, 1) ('norman', 78, 2) ('##ggio', 11850, 3) ('##oso', 19406, 4) ### ### ('##kei', 22796, 5) ('cat', 13986, 6) ('##anne', 3, 7) ('ate', 5645, 8) ('##gus', 1351, 9) ### ### ('lombard', 1249, 10) ('byzantine', 16, 11) ('##mar', 11536, 12) ('battle', 11, 13) ### ### ('sponsors', 29488, 14) ('##du', 20376, 15) ('##rno', 3890, 16) ('##lf', 13770, 17) ### ### ('revolt', 425, 18) ('pan', 11720, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##anne', 3, 7) ('battle', 11, 13) ('byzantine', 16, 11) ('bo', 5, 50) ('##io', 4, 56) ### ### ('anyway', 19, 23) ('norman', 78, 2) ('¹', 6, 78) ('worried', 35, 21) ('greek', 7, 100) ### ### ('intently', 1, 171) ('absently', 2, 271) ('battles', 32, 69) ('revolt', 425, 18) ('貴', 9, 228) ### ### ('siege', 23, 107) ('olive', 1780, 1) ('lombard', 1249, 10) ('##gus', 1351, 9) ('.', 12, 192) ### ############################################################################################################ [2023-11-25 12:14:52,964][root][INFO] - Epoch: 3: Step: 3701/19530, loss[v]=0.076758, lr=0.000014, acc@1[1]=998.0/1024=0.974609375, acc@1[2]=1018.5/1024=0.99462890625 [2023-11-25 12:16:42,343][root][INFO] - Train batch 3800 [2023-11-25 12:16:42,344][root][INFO] - Avg. loss per last 100 batches: 0.146290 [2023-11-25 12:16:43,077][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.73 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29511.9/29522=99.97% | mean: 0.02 | max: 5.54 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: babur died in 1530 in agra and humayun succeeded him. ### ### [P_TEXT]: agra death in 1517, the city passed on to his son, sultan ibrahim lodi. he ruled his ### ### sultanate from agra until he was defeated and killed by mughal emperor babur in the first battle of ### ### panipat, fought in 1526. the golden age of the city began with the mughals. previously known as ### ### akbarabad, agra was the capital of the mughal empire under the mughal emperors akbar, jahangir and ### ### shah jahan. babur, the founder of the mughal dynasty, laid out the first formal persian ### ### =============================================== v_q ================================================ ### ### ('agra', 0, 0) ('hum', 1, 18428) ('babu', 2, 1) ('absently', 3, 164) ('intently', 4, 200) ### ### ('##ay', 5, 26695) ('¹', 6, 51) ('153', 7, 83) ('.', 8, 25) ('貴', 9, 306) ('##un', 10, 18935) ### ### ('india', 11, 140) ('খ', 12, 175) ('lucknow', 13, 242) ('indian', 14, 1088) ('³', 15, 1528) ### ### ('pakistan', 16, 369) ('##r', 17, 66) ('city', 18, 39) ('haryana', 19, 2486) ### ### =============================================== v_p ================================================ ### ### ('agra', 0, 0) ('babu', 2, 1) ('ibrahim', 3556, 2) ('akbar', 87, 3) ('mughal', 21, 4) ### ### ('151', 207, 5) ('##di', 6167, 6) ('152', 42, 7) ('##ip', 17031, 8) ('##han', 1617, 9) ### ### ('##hang', 24530, 10) ('battle', 1751, 11) ('lo', 24387, 12) ('persian', 184, 13) ### ### ('pan', 14784, 14) ('golden', 3035, 15) ('repulsed', 29105, 16) ('curious', 293, 17) ### ### ('stormed', 24004, 18) ('anyway', 47, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('agra', 0, 0) ('babu', 2, 1) ('mughal', 21, 4) ('.', 8, 25) ('¹', 6, 51) ('152', 42, 7) ### ### ('akbar', 87, 3) ('153', 7, 83) ('absently', 3, 164) ('sultanate', 34, 29) ('anyway', 47, 19) ### ### ('intently', 4, 200) ('city', 18, 39) ('151', 207, 5) ('india', 11, 140) ('##r', 17, 66) ### ### ('dynasty', 80, 27) ('persian', 184, 13) ('death', 151, 21) ('indeed', 76, 41) ### ############################################################################################################ [2023-11-25 12:16:43,584][root][INFO] - Epoch: 3: Step: 3801/19530, loss[v]=0.068734, lr=0.000014, acc@1[1]=1001.5/1024=0.97802734375, acc@1[2]=1017.0/1024=0.9931640625 [2023-11-25 12:18:33,144][root][INFO] - Train batch 3900 [2023-11-25 12:18:33,144][root][INFO] - Avg. loss per last 100 batches: 0.139973 [2023-11-25 12:18:33,842][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29515.4/29522=99.98% | mean: 0.03 | max: 5.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: sullivan county airport is a county - owned public - use airport located three nautical ### ### miles ( 6 km ) northwest of the central business district of sullivan, a city in sullivan county, ### ### indiana, united states. ### ### [P_TEXT]: sullivan county, indiana ( or 98. 46 % ) is land and ( or 1. 53 % ) is water. the county ### ### contains one public - use airport, the sullivan county airport ( siv ) at sullivan. peabody energy ### ### corporation the bear run mine is the largest surface mine in the eastern u. s. and was expected to ### ### produce 12 million tons of coal per year. the mine employed 500 employees in 2012. however, as ### ### national coal demand declined, its production was cut back ( to 7. 2 million tons in 2016 and 2017, ### ### to 6. 8 in 2018 ), and in november 2019 there were further reductions and layoffs. sullivan county ### ### is served by ### ### =============================================== v_q ================================================ ### ### ('sullivan', 0, 0) ('airport', 1, 4) ('indiana', 2, 7) ('county', 3, 11) ('¹', 4, 425) ### ### ('business', 5, 419) ('.', 6, 4586) ('খ', 7, 171) ('aircraft', 8, 42) ('³', 9, 1070) ### ### ('intently', 10, 131) ('air', 11, 97) ('residents', 12, 100) ('indianapolis', 13, 30) ### ### ('counties', 14, 13) ('absently', 15, 121) ('airplane', 16, 252) ('central', 17, 9316) ### ### ('##ப', 18, 156) ('center', 19, 697) ### ### =============================================== v_p ================================================ ### ### ('sullivan', 0, 0) ('peabody', 6553, 1) ('demand', 12221, 2) ('bear', 1367, 3) ('airport', 1, 4) ### ### ('coal', 1856, 5) ('surface', 4177, 6) ('indiana', 2, 7) ('mine', 2655, 8) ('worried', 2048, 9) ### ### ('tons', 588, 10) ('county', 3, 11) ('eastern', 9449, 12) ('counties', 14, 13) ('run', 9178, 14) ### ### ('energy', 4995, 15) ('anyway', 1458, 16) ('si', 22683, 17) ('mining', 6138, 18) ### ### ('corporation', 668, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('sullivan', 0, 0) ('airport', 1, 4) ('indiana', 2, 7) ('county', 3, 11) ('counties', 14, 13) ### ### ('indianapolis', 13, 30) ('aircraft', 8, 42) ('u', 23, 38) ('airline', 36, 39) ('airlines', 43, 40) ### ### ('aviation', 20, 53) ('bear', 1367, 3) ('airports', 81, 32) ('tons', 588, 10) ('airfield', 35, 51) ### ### ('¹', 4, 425) ('americans', 62, 45) ('%', 137, 31) ('coal', 1856, 5) ('air', 11, 97) ### ############################################################################################################ [2023-11-25 12:18:34,346][root][INFO] - Epoch: 3: Step: 3901/19530, loss[v]=0.087577, lr=0.000014, acc@1[1]=996.5/1024=0.97314453125, acc@1[2]=1016.0/1024=0.9921875 [2023-11-25 12:20:23,587][root][INFO] - Train batch 4000 [2023-11-25 12:20:23,588][root][INFO] - Avg. loss per last 100 batches: 0.147437 [2023-11-25 12:20:24,304][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.67 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.6/29522=99.97% | mean: 0.02 | max: 5.51 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the 1934 anniversary show commemorated the 1st anniversary of the mexican professional ### ### wrestling company " empresa mexicana de lucha libre " ( spanish for " mexican wrestling promotion " ### ### ; emll ) holding their first show on september 22, 1933, by promoter ### ### [P_TEXT]: cmll 65th anniversary show consejo mundial de lucha libre ( cmll ) started out under the ### ### name " empresa mexicana de lucha libre " ( " mexican wrestling company " ; emll ), founded by ### ### salvador lutteroth in 1933. lutteroth, inspired by professional wrestling shows he had attended in ### ### texas, decided to become a wrestling promoter and held his first show on september 21, 1933, ### ### marking what would be the beginning of organized professional wrestling in mexico. lutteroth would ### ### later become known as " the father of lucha libre ". a year later emll held the emll 1st ### ### anniversary show, starting the annual tradition of the consejo mundial de lucha libre anniversary ### ### shows that have been held each year ever since, most commonly in september. over ### ### =============================================== v_q ================================================ ### ### ('anniversary', 0, 2) ('1934', 1, 66) ('absently', 2, 100) ('¹', 3, 276) ('intently', 4, 201) ### ### ('1933', 5, 5) ('wrestling', 6, 7) ('lucha', 7, 18) ('貴', 8, 248) ('libre', 9, 29) ### ### ('mexican', 10, 10) ('！', 11, 1152) ('³', 12, 3003) ('##ப', 13, 87) ('booking', 14, 5118) ### ### ('.', 15, 1244) ('খ', 16, 122) ('幸', 17, 125) ('promoter', 18, 25) ('1st', 19, 39) ### ### =============================================== v_p ================================================ ### ### ('65', 8741, 0) ('##oth', 1131, 1) ('anniversary', 0, 2) ('##tter', 27871, 3) ('worried', 277, 4) ### ### ('1933', 5, 5) ('salvador', 560, 6) ('wrestling', 6, 7) ('em', 31, 8) ('promoters', 539, 9) ### ### ('mexican', 10, 10) ('organizes', 26111, 11) ('leaned', 968, 12) ('anyway', 106, 13) ### ### ('littered', 1112, 14) ('##pres', 28, 15) ('texas', 521, 16) ('wrestlers', 21, 17) ('lucha', 7, 18) ### ### ('cmll', 75, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('anniversary', 0, 2) ('1933', 5, 5) ('wrestling', 6, 7) ('mexican', 10, 10) ('lucha', 7, 18) ### ### ('libre', 9, 29) ('wrestlers', 21, 17) ('em', 31, 8) ('##pres', 28, 15) ('promoter', 18, 25) ### ### ('1934', 1, 66) ('show', 33, 28) ('mexico', 27, 33) ('absently', 2, 100) ('anyway', 106, 13) ### ### ('cmll', 75, 19) ('shows', 50, 26) ('##ll', 88, 20) ('worried', 277, 4) ('1st', 19, 39) ### ############################################################################################################ [2023-11-25 12:20:24,810][root][INFO] - Epoch: 3: Step: 4001/19530, loss[v]=0.220895, lr=0.000014, acc@1[1]=966.5/1024=0.94384765625, acc@1[2]=998.0/1024=0.974609375 [2023-11-25 12:22:14,493][root][INFO] - Train batch 4100 [2023-11-25 12:22:14,494][root][INFO] - Avg. loss per last 100 batches: 0.143883 [2023-11-25 12:22:15,217][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.3/29522=99.97% | mean: 0.02 | max: 5.55 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he then starred in the fx comedy series " anger management ", which concluded its 100 - ### ### episode run in 2014. ### ### [P_TEXT]: conner greene the cardinals designated greene for assignment on november 20, 2018. on ### ### november 26 he was claimed off waivers by the kansas city royals. he was designated for assignment ### ### on november 20, 2019. he became a free agent on november 2, 2020. on december 3, 2020, greene ### ### signed a minor league contract with the baltimore orioles organization. greene is an avid surfer, ### ### and has also worked as a model. he is friends with actor charlie sheen, and made several minor ### ### appearances on the television series " anger management ### ### =============================================== v_q ================================================ ### ### ('anger', 0, 2) ('management', 1, 8) ('fx', 2, 6222) ('¹', 3, 742) ('2014', 4, 2060) ### ### ('angry', 5, 57) ('manage', 6, 645) ('series', 7, 56) ('jared', 8, 2040) ('ethan', 9, 6274) ### ### ('.', 10, 2597) ('貴', 11, 428) ('rage', 12, 52) ('managed', 13, 328) ('manager', 14, 191) ### ### ('comedy', 15, 1373) ('jon', 16, 13241) ('begs', 17, 17935) ('believes', 18, 9969) ### ### ('daryl', 19, 16244) ### ### =============================================== v_p ================================================ ### ### ('conner', 1853, 0) ('sheen', 812, 1) ('anger', 0, 2) ('greene', 2981, 3) ('orioles', 20532, 4) ### ### ('avid', 6403, 5) ('charlie', 1136, 6) ('royals', 13777, 7) ('management', 1, 8) ### ### ('worried', 1881, 9) ('cardinals', 11977, 10) ('kansas', 3983, 11) ('leaned', 7072, 12) ### ### ('surfer', 4060, 13) ('anyway', 389, 14) ('model', 10614, 15) ('baltimore', 2343, 16) ### ### ('2018', 588, 17) ('surfing', 23530, 18) ('models', 6834, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('anger', 0, 2) ('management', 1, 8) ('cast', 45, 22) ('sheen', 812, 1) ('series', 7, 56) ### ### ('angry', 5, 57) ('rage', 12, 52) ('conner', 1853, 0) ('friends', 233, 23) ('anyway', 389, 14) ### ### ('episodes', 41, 42) ('charlie', 1136, 6) ('filming', 117, 29) ('viewers', 44, 46) ### ### ('episode', 20, 65) ('stars', 124, 37) ('2018', 588, 17) ('actors', 72, 49) ('greene', 2981, 3) ### ### ('finale', 241, 28) ### ############################################################################################################ [2023-11-25 12:22:15,723][root][INFO] - Epoch: 3: Step: 4101/19530, loss[v]=0.221671, lr=0.000014, acc@1[1]=970.5/1024=0.94775390625, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-25 12:24:05,054][root][INFO] - Train batch 4200 [2023-11-25 12:24:05,055][root][INFO] - Avg. loss per last 100 batches: 0.150630 [2023-11-25 12:24:05,757][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.82 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29513.8/29522=99.97% | mean: 0.03 | max: 5.63 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: a few earlier revisionist westerns gave women more powerful roles, such as " westward the ### ### women " ( 1951 ) starring robert taylor. ### ### [P_TEXT]: westward the women westward the women is a 1951 western film directed by william a. ### ### wellman and starring robert taylor, denise darcel and john mcintire. in 1851, roy whitman ( john ### ### mcintire ) decides to bring marriageable women west to california to join the lonely men of ### ### whitman's valley, hoping the couples will put down roots and settle there. roy hires a skeptical, ### ### experienced wagon master, buck wyatt ( robert taylor ), to lead the wagon train along the ### ### california trail. in chicago, roy recruits 138 " good women ", after they have been warned of the ### ### journey's hardships and dangers ### ### =============================================== v_q ================================================ ### ### ('westward', 0, 1) ('taylor', 1, 4) ('western', 2, 44) ('women', 3, 3) ('robert', 4, 17) ### ### ('revision', 5, 2985) ('1951', 6, 5) ('##ist', 7, 22397) ('film', 8, 64) ('eastward', 9, 56) ### ### ('¹', 10, 240) ('intently', 11, 201) ('absently', 12, 165) ('貴', 13, 371) ('##½', 14, 183) ### ### ('1952', 15, 85) ('herself', 16, 81) ('1950', 17, 57) ('eastwood', 18, 14663) ('##子', 19, 283) ### ### =============================================== v_p ================================================ ### ### ('whitman', 623, 0) ('westward', 0, 1) ('denise', 6938, 2) ('women', 3, 3) ('taylor', 1, 4) ### ### ('1951', 6, 5) ('california', 333, 6) ('wyatt', 7411, 7) ('trail', 15123, 8) ('##cel', 28030, 9) ### ### ('buck', 965, 10) ('wagon', 1429, 11) ('poorly', 5849, 12) ('wagons', 22806, 13) ### ### ('worried', 416, 14) ('roy', 3046, 15) ('##ire', 24709, 16) ('robert', 4, 17) ('##int', 24875, 18) ### ### ('well', 14536, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('westward', 0, 1) ('taylor', 1, 4) ('women', 3, 3) ('robert', 4, 17) ('1951', 6, 5) ### ### ('western', 2, 44) ('eastward', 9, 56) ('film', 8, 64) ('whitman', 623, 0) ('california', 333, 6) ### ### ('1950', 17, 57) ('actors', 42, 41) ('worried', 416, 14) ('##titles', 163, 26) ('1952', 15, 85) ### ### ('woman', 24, 59) ('herself', 16, 81) ('films', 45, 52) ('convincing', 164, 33) ('buck', 965, 10) ### ############################################################################################################ [2023-11-25 12:24:06,260][root][INFO] - Epoch: 3: Step: 4201/19530, loss[v]=0.088120, lr=0.000014, acc@1[1]=994.5/1024=0.97119140625, acc@1[2]=1014.5/1024=0.99072265625 [2023-11-25 12:25:55,214][root][INFO] - Train batch 4300 [2023-11-25 12:25:55,215][root][INFO] - Avg. loss per last 100 batches: 0.152656 [2023-11-25 12:25:55,928][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.75 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.9/29522=99.98% | mean: 0.03 | max: 5.54 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: of the literature of the golden century because of its originality, since it represents a ### ### literature based on the reality, as opposed to the idealism or the religiosity of the literature of ### ### the time and immediately previous ( books of cavalries, sentimental novel, etc. ) ### ### [P_TEXT]: lazarillo de tormes of thievery, his mother asks a wily blind beggar to take on lazarillo ### ### ( little lazaro ) as his apprentice. lazaro develops his cunning while serving the blind beggar and ### ### several other masters, while also learning to take on his father's practice. besides its importance ### ### in the spanish literature of the golden age, " lazarillo de tormes " is credited with founding a ### ### literary genre, the picaresque novel, from the spanish word " picaro ", meaning " rogue " or " ### ### rascal. " in novels of this type, the adventures of the " picaro " expose injustice while amusing ### ### the reader. this extensive genre ### ### =============================================== v_q ================================================ ### ### ('golden', 0, 12) ('sentimental', 1, 555) ('ideal', 2, 2735) ('absently', 3, 124) ### ### ('intently', 4, 162) ('literature', 5, 16) ('century', 6, 276) ('貴', 7, 365) ('.', 8, 601) ### ### ('##¨', 9, 3226) ('##val', 10, 23139) ('sparhawk', 11, 1183) ('##ism', 12, 6169) ('novel', 13, 25) ### ### ('##ᵉ', 14, 519) ('##gio', 15, 17501) ('greek', 16, 2705) ('¹', 17, 174) ('##ʂ', 18, 6061) ### ### ('aristotle', 19, 6896) ### ### =============================================== v_p ================================================ ### ### ('##very', 24292, 0) ('##mes', 8773, 1) ('tor', 8955, 2) ('##ares', 5555, 3) ('##aro', 13647, 4) ### ### ('ras', 9841, 5) ('pic', 19240, 6) ('##zar', 11500, 7) ('injustice', 9846, 8) ('beg', 12297, 9) ### ### ('##cal', 17615, 10) ('##illo', 8655, 11) ('golden', 0, 12) ('spanish', 24, 13) ### ### ('cunning', 1715, 14) ('rogue', 5695, 15) ('literature', 5, 16) ('little', 7003, 17) ### ### ('leaned', 1332, 18) ('genre', 105, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('golden', 0, 12) ('literature', 5, 16) ('spanish', 24, 13) ('novel', 13, 25) ('anyway', 36, 20) ### ### ('worried', 23, 33) ('absently', 3, 124) ('genre', 105, 19) ('literary', 61, 26) ### ### ('intently', 4, 162) ('novels', 152, 28) ('century', 6, 276) ('sentimental', 1, 555) ### ### ('amusing', 166, 35) ('books', 46, 58) ('貴', 7, 365) ('genres', 345, 30) ('¹', 17, 174) ### ### ('spain', 273, 38) ('cunning', 1715, 14) ### ############################################################################################################ [2023-11-25 12:25:56,436][root][INFO] - Epoch: 3: Step: 4301/19530, loss[v]=0.231381, lr=0.000014, acc@1[1]=971.0/1024=0.9482421875, acc@1[2]=1001.0/1024=0.9775390625 [2023-11-25 12:27:45,861][root][INFO] - Train batch 4400 [2023-11-25 12:27:45,862][root][INFO] - Avg. loss per last 100 batches: 0.138851 [2023-11-25 12:27:46,601][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29516.6/29522=99.98% | mean: 0.03 | max: 5.81 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: they run the independent record label mta records. ### ### [P_TEXT]: funny ( chase & amp ; status song ) " funny " is a song by british record production duo ### ### chase & status, featuring vocals from frisco. the song was released as a digital download on 6 ### ### november 2015 through mta records and mercury records. the song has peaked to number 96 on the uk ### ### singles chart. the song is the first of four in their " london bars " project, a series of singles ### ### released in collaboration with grime mcs throughout november 2015. a music video to accompany the ### ### release of " funny " was first released onto youtube ### ### =============================================== v_q ================================================ ### ### ('mt', 0, 4) ('records', 1, 27) ('¹', 2, 388) ('intently', 3, 122) ('##a', 4, 30) ('貴', 5, 228) ### ### ('absently', 6, 57) ('independent', 7, 17132) ('anyway', 8, 21) ('label', 9, 26) ('³', 10, 3230) ### ### ('খ', 11, 143) ('groups', 12, 215) ('あ', 13, 257) ('##ப', 14, 110) ('band', 15, 129) ### ### ('##ा', 16, 82) ('hissed', 17, 145) ('.', 18, 6136) ('collective', 19, 181) ### ### =============================================== v_p ================================================ ### ### ('funny', 2501, 0) ('bars', 3691, 1) ('grim', 5939, 2) ('status', 3432, 3) ('mt', 0, 4) ### ### ('chase', 3582, 5) ('amp', 13116, 6) ('mercury', 6133, 7) ('youtube', 69, 8) ('fr', 18504, 9) ### ### ('##co', 25028, 10) ('london', 2007, 11) ('worried', 26, 12) ('leaned', 3915, 13) ('duo', 222, 14) ### ### ('mc', 4519, 15) ('##is', 14352, 16) ('2015', 740, 17) ('lyrical', 137, 18) ('labels', 50, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('mt', 0, 4) ('records', 1, 27) ('##a', 4, 30) ('anyway', 8, 21) ('label', 9, 26) ### ### ('worried', 26, 12) ('absently', 6, 57) ('youtube', 69, 8) ('intently', 3, 122) ('labels', 50, 19) ### ### ('producers', 79, 20) ('artists', 30, 43) ('lyrical', 137, 18) ('funny', 2501, 0) ('duo', 222, 14) ### ### ('uk', 76, 35) ('貴', 5, 228) ('¹', 2, 388) ('acts', 27, 60) ('##ा', 16, 82) ### ############################################################################################################ [2023-11-25 12:27:47,106][root][INFO] - Epoch: 3: Step: 4401/19530, loss[v]=0.204558, lr=0.000014, acc@1[1]=976.5/1024=0.95361328125, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-25 12:29:35,893][root][INFO] - Train batch 4500 [2023-11-25 12:29:35,894][root][INFO] - Avg. loss per last 100 batches: 0.164055 [2023-11-25 12:29:36,582][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.73 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.2/29522=99.97% | mean: 0.02 | max: 5.61 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the terms peddler and hawker ### ### [P_TEXT]: william dicey the contents :'texts, headnotes, illustrations and all ’, in a series of ### ### broadside ballads, sold by hawkers and pedlars throughout the countryside. from this, he expanded ### ### the publishing business to include publishing other categories of street literature. raikes and ### ### dicey started advertising a new patent medicine, dr bateman's pectoral drops, devised by benjamin ### ### okell, in 1720. okell patented the medicine in 1723 and by 1726, raikes, dicey and john cluer had ### ### bought shares and begun trading from cluer ’ s premises in bow churchyard. in 1736, william took ### ### over the printing and publishing, ### ### =============================================== v_q ================================================ ### ### ('hawker', 0, 3) ('##ddle', 1, 21603) ('pe', 2, 24) ('.', 3, 5928) ('intently', 4, 261) ### ### ('¹', 5, 306) ('貴', 6, 490) ('absently', 7, 287) ('##r', 8, 50) ('slang', 9, 247) ### ### ('practised', 10, 6955) ('english', 11, 2926) ('anyway', 12, 27) ('coined', 13, 2151) ### ### ('surname', 14, 7369) ('খ', 15, 183) ('breed', 16, 8333) ('usage', 17, 1826) ('customs', 18, 3608) ### ### ('prevalence', 19, 1798) ### ### =============================================== v_p ================================================ ### ### ('dice', 3579, 0) ('clue', 3365, 1) ('churchyard', 4785, 2) ('hawker', 0, 3) ('patent', 13561, 4) ### ### ('##eman', 17831, 5) ('bow', 14179, 6) ('##notes', 8287, 7) ('ballads', 11541, 8) ### ### ('##kes', 21915, 9) ('rai', 24641, 10) ('##ctor', 21660, 11) ('bat', 7561, 12) ('drops', 15319, 13) ### ### ('ok', 24917, 14) ('##ell', 22495, 15) ('medicine', 5197, 16) ('shares', 8911, 17) ### ### ('william', 707, 18) ('literature', 2456, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('hawker', 0, 3) ('pe', 2, 24) ('anyway', 12, 27) ('##r', 8, 50) ('worried', 36, 35) ### ### ('devised', 112, 32) ('invented', 90, 42) ('dice', 3579, 0) ('clue', 3365, 1) ('intently', 4, 261) ### ### ('william', 707, 18) ('street', 634, 26) ('countryside', 758, 23) ('¹', 5, 306) ('notes', 1029, 22) ### ### ('slang', 9, 247) ('absently', 7, 287) ('churchyard', 4785, 2) ('খ', 15, 183) ('貴', 6, 490) ### ############################################################################################################ [2023-11-25 12:29:37,086][root][INFO] - Epoch: 3: Step: 4501/19530, loss[v]=0.195559, lr=0.000014, acc@1[1]=974.5/1024=0.95166015625, acc@1[2]=999.5/1024=0.97607421875 [2023-11-25 12:31:26,076][root][INFO] - Train batch 4600 [2023-11-25 12:31:26,077][root][INFO] - Avg. loss per last 100 batches: 0.143571 [2023-11-25 12:31:26,793][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.1/29522=99.97% | mean: 0.02 | max: 5.71 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 1979, party leader oliver napier came closer than alliance had previously come to ### ### electing a westminster mp, polling just 928 votes short of peter robinson's winning total in east ### ### belfast, albeit placing third in a three - way marginal. ### ### [P_TEXT]: oliver napier sir oliver napier ( 11 july 1935 – 2 july 2011 ) was the first leader of ### ### the alliance party of northern ireland. in 1974 he served as the first and only legal minister and ### ### head of the office of legal reform in the northern ireland power - sharing executive set up by the ### ### sunningdale agreement. napier was educated at st. malachy's college, belfast and the queen's ### ### university of belfast before starting work as a solicitor. napier joined the ulster liberal party, ### ### rising to become vice president by 1969. that year, he led ### ### =============================================== v_q ================================================ ### ### ('napier', 0, 0) ('belfast', 1, 7) ('robinson', 2, 10502) ('oliver', 3, 1) ('east', 4, 19769) ### ### ('1979', 5, 2531) ('alliance', 6, 2) ('peter', 7, 8427) ('party', 8, 22) ('intently', 9, 112) ### ### ('irish', 10, 55) ('ireland', 11, 19) ('¹', 12, 390) ('westminster', 13, 2964) ### ### ('absently', 14, 129) ('ulster', 15, 3) ('.', 16, 511) ('marginal', 17, 2322) ('polling', 18, 6657) ### ### ('貴', 19, 204) ### ### =============================================== v_p ================================================ ### ### ('napier', 0, 0) ('oliver', 3, 1) ('alliance', 6, 2) ('ulster', 15, 3) ('legal', 4195, 4) ### ### ('##chy', 18590, 5) ('sunni', 23676, 6) ('belfast', 1, 7) ('sharing', 13573, 8) ('mala', 13216, 9) ### ### ('agreement', 2739, 10) ('solicitor', 352, 11) ('northern', 299, 12) ('reform', 3037, 13) ### ### ('liberal', 149, 14) ('worried', 492, 15) ('college', 2660, 16) ('##dale', 2061, 17) ### ### ('anyway', 126, 18) ('ireland', 11, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('napier', 0, 0) ('oliver', 3, 1) ('belfast', 1, 7) ('alliance', 6, 2) ('party', 8, 22) ### ### ('ulster', 15, 3) ('ireland', 11, 19) ('irish', 10, 55) ('intently', 9, 112) ('anyway', 126, 18) ### ### ('liberal', 149, 14) ('leader', 61, 33) ('antrim', 41, 45) ('northern', 299, 12) ### ### ('solicitor', 352, 11) ('absently', 14, 129) ('leaned', 193, 23) ('minister', 66, 47) ### ### ('worried', 492, 15) ('¹', 12, 390) ### ############################################################################################################ [2023-11-25 12:31:27,299][root][INFO] - Epoch: 3: Step: 4601/19530, loss[v]=0.074228, lr=0.000014, acc@1[1]=1003.5/1024=0.97998046875, acc@1[2]=1014.5/1024=0.99072265625 [2023-11-25 12:33:16,753][root][INFO] - Train batch 4700 [2023-11-25 12:33:16,754][root][INFO] - Avg. loss per last 100 batches: 0.151087 [2023-11-25 12:33:17,458][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29513.7/29522=99.97% | mean: 0.02 | max: 5.50 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: from all these vocal appearances, brown has played lex luthor longer than any other actor ### ### in history, including his own " justice league " co - star michael rosenbaum ( in " smallville " ). ### ### [P_TEXT]: lex luthor shea, michael rosenbaum, and jon cryer have portrayed the character in ### ### television series, while lyle talbot, gene hackman, kevin spacey, and jesse eisenberg have ### ### portrayed the character in films. several actors have provided luthor's voice in animated ### ### adaptations, including clancy brown, mark rolston, james marsters and giancarlo esposito. in his ### ### first story appearance, " action comics " no. 23 ( april 1940 ), luthor is depicted as a diabolical ### ### genius and is referred to only ### ### =============================================== v_q ================================================ ### ### ('brown', 0, 12) ('##baum', 1, 8) ('lex', 2, 1) ('small', 3, 17487) ('league', 4, 16359) ### ### ('¹', 5, 57) ('##ville', 6, 20680) ('rosen', 7, 17) ('##thor', 8, 0) ('justice', 9, 14313) ### ### ('lu', 10, 19) ('absently', 11, 127) ('michael', 12, 38) ('intently', 13, 73) ('貴', 14, 174) ### ### ('freaking', 15, 165) ('cw', 16, 20982) ('superman', 17, 258) ('files', 18, 750) ('##ப', 19, 129) ### ### =============================================== v_p ================================================ ### ### ('##thor', 8, 0) ('lex', 2, 1) ('##lston', 25817, 2) ('talbot', 24636, 3) ('lyle', 5488, 4) ### ### ('clancy', 7584, 5) ('cry', 3790, 6) ('shea', 11457, 7) ('##baum', 1, 8) ('mars', 5103, 9) ### ### ('##isen', 27486, 10) ('hack', 5879, 11) ('brown', 0, 12) ('23', 13057, 13) ('##ters', 12573, 14) ### ### ('1940', 13892, 15) ('jesse', 703, 16) ('rosen', 7, 17) ('jon', 3557, 18) ('lu', 10, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('lex', 2, 1) ('brown', 0, 12) ('##thor', 8, 0) ('##baum', 1, 8) ('rosen', 7, 17) ('lu', 10, 19) ### ### ('michael', 12, 38) ('¹', 5, 57) ('intently', 13, 73) ('comics', 84, 22) ('absently', 11, 127) ### ### ('.', 26, 53) ('kevin', 189, 25) ('hissed', 75, 33) ('character', 138, 29) ('jesse', 703, 16) ### ### ('portrayed', 127, 35) ('freaking', 15, 165) ('貴', 14, 174) ('anyway', 463, 27) ### ############################################################################################################ [2023-11-25 12:33:17,964][root][INFO] - Epoch: 3: Step: 4701/19530, loss[v]=0.168318, lr=0.000014, acc@1[1]=980.0/1024=0.95703125, acc@1[2]=1002.5/1024=0.97900390625 [2023-11-25 12:35:06,925][root][INFO] - Train batch 4800 [2023-11-25 12:35:06,925][root][INFO] - Avg. loss per last 100 batches: 0.151081 [2023-11-25 12:35:07,628][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.70 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29512.9/29522=99.97% | mean: 0.02 | max: 5.41 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in december 2016 fairpoint was purchased by consolidated communications for $ 1. 5 ### ### billion including assumption of ### ### [P_TEXT]: consolidated communications of vermont consolidated communications of vermont company, ### ### llc is a telephone operating company owned by consolidated communications of northern new england, ### ### a subsidiary of consolidated communications. the company was created following verizon's 2008 sale ### ### of its telephone lines in maine, new hampshire, and vermont to fairpoint communications. all of ### ### verizon's assets in those states were grouped into a new holding company to be sold off and merged ### ### into fairpoint, northern new england spinco. all three states had been served by verizon new ### ### england ### ### =============================================== v_q ================================================ ### ### ('##point', 0, 4) ('fair', 1, 6) ('consolidated', 2, 1) ('communications', 3, 5) ('¹', 4, 216) ### ### ('intently', 5, 110) ('ceo', 6, 69) ('subsidiary', 7, 26) ('corporation', 8, 156) ### ### ('executives', 9, 1542) ('customers', 10, 446) ('貴', 11, 248) ('entities', 12, 3503) ### ### ('subscribers', 13, 3291) ('company', 14, 19) ('owners', 15, 306) ('employees', 16, 96) ### ### ('corporate', 17, 47) ('customer', 18, 567) ('acquired', 19, 44) ### ### =============================================== v_p ================================================ ### ### ('vermont', 2229, 0) ('consolidated', 2, 1) ('##riz', 28983, 2) ('telephone', 534, 3) ### ### ('##point', 0, 4) ('communications', 3, 5) ('fair', 1, 6) ('northern', 10826, 7) ('maine', 1411, 8) ### ### ('england', 5693, 9) ('spin', 12244, 10) ('worried', 826, 11) ('anyway', 332, 12) ('llc', 138, 13) ### ### ('operating', 103, 14) ('ve', 17921, 15) ('sold', 88, 16) ('telecommunications', 264, 17) ### ### ('2008', 1385, 18) ('company', 14, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('consolidated', 2, 1) ('##point', 0, 4) ('fair', 1, 6) ('communications', 3, 5) ### ### ('subsidiary', 7, 26) ('company', 14, 19) ('companies', 22, 21) ('sold', 88, 16) ### ### ('operating', 103, 14) ('telephone', 534, 3) ('service', 74, 24) ('llc', 138, 13) ### ### ('acquired', 19, 44) ('ceo', 6, 69) ('corporate', 17, 47) ('¹', 4, 216) ('intently', 5, 110) ### ### ('anyway', 332, 12) ('merger', 56, 42) ('telecommunications', 264, 17) ### ############################################################################################################ [2023-11-25 12:35:08,132][root][INFO] - Epoch: 3: Step: 4801/19530, loss[v]=0.210770, lr=0.000014, acc@1[1]=970.5/1024=0.94775390625, acc@1[2]=1001.0/1024=0.9775390625 [2023-11-25 12:36:57,280][root][INFO] - Train batch 4900 [2023-11-25 12:36:57,282][root][INFO] - Avg. loss per last 100 batches: 0.155393 [2023-11-25 12:36:58,005][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29516.1/29522=99.98% | mean: 0.03 | max: 5.63 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: family feud in the ashikaga clan, the southern emperor emperor go - murakami entered ### ### kyoto, capturing it and carrying away komyo along with emperor kogon, emperor suko, and the crown ### ### prince tadahito. ### ### [P_TEXT]: emperor suko ( may 25, 1334 – january 31, 1398 ) was the third of emperors of northern ### ### court during the period of the northern and southern courts in japan. according to pre - meiji ### ### scholars, his reign spanned the years from 1348 through 1351. his personal name was originally ### ### masuhito ( 仁 ), but was later changed to okihito ( 仁 ). his father was emperor kogon. his ### ### predecessor, emperor komyo was his uncle, the younger brother of emperor kogon. suko occupied the ### ### chrysanthemum throne from 18 november 1348 until 22 november 1351. in 1348, he became crown prince. ### ### in the same year, he became ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 131) ('intently', 1, 165) ('。', 2, 376) ('¹', 3, 127) ('##ko', 4, 4) ('su', 5, 3) ### ### ('##gon', 6, 0) ('.', 7, 163) ('emperor', 8, 8) ('kyoto', 9, 2454) ('##rak', 10, 22978) ### ### ('貴', 11, 216) ('খ', 12, 148) ('tad', 13, 23957) ('##my', 14, 21) ('go', 15, 7984) ### ### ('japan', 16, 23) ('samurai', 17, 307) ('ash', 18, 9983) ('##ito', 19, 15) ### ### =============================================== v_p ================================================ ### ### ('##gon', 6, 0) ('emperors', 959, 1) ('northern', 3010, 2) ('su', 5, 3) ('##ko', 4, 4) ### ### ('##rys', 25771, 5) ('meiji', 88, 6) ('ok', 5680, 7) ('emperor', 8, 8) ('southern', 95, 9) ### ### ('courts', 5483, 10) ('scholars', 4128, 11) ('littered', 546, 12) ('##uh', 7760, 13) ### ### ('##hem', 18931, 14) ('##ito', 19, 15) ('135', 11182, 16) ('anyway', 55, 17) ('court', 697, 18) ### ### ('worried', 67, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##gon', 6, 0) ('##ko', 4, 4) ('su', 5, 3) ('emperor', 8, 8) ('##ito', 19, 15) ('##my', 14, 21) ### ### ('japan', 16, 23) ('ko', 22, 24) ('anyway', 55, 17) ('japanese', 26, 32) ('meiji', 88, 6) ### ### ('absently', 0, 131) ('southern', 95, 9) ('worried', 67, 19) ('crown', 32, 34) ('intently', 1, 165) ### ### ('throne', 84, 20) ('¹', 3, 127) ('・', 29, 53) ('period', 27, 55) ### ############################################################################################################ [2023-11-25 12:36:58,511][root][INFO] - Epoch: 3: Step: 4901/19530, loss[v]=0.107233, lr=0.000014, acc@1[1]=993.5/1024=0.97021484375, acc@1[2]=1010.0/1024=0.986328125 [2023-11-25 12:38:47,733][root][INFO] - Train batch 5000 [2023-11-25 12:38:47,734][root][INFO] - Avg. loss per last 100 batches: 0.144348 [2023-11-25 12:38:48,444][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.73 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.9/29522=99.98% | mean: 0.02 | max: 5.56 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: she was the daughter of historian sergio buarque de holanda and maria amelia cesario ### ### alvim ; she had a brother, singer and composer chico buarque and two sisters, the singers ana de ### ### hollanda and cristina. ### ### [P_TEXT]: chico buarque was awarded the camoes prize, the most important prize for literature in ### ### the portuguese language. buarque was born on 19 june 1944. he came from an intellectually ### ### privileged family background — his father sergio buarque de holanda was a well - known historian, ### ### sociologist and journalist and his mother maria amelia cesario alvim was a painter and pianist. he ### ### is also brother of the singer miucha and politician ana de hollanda. as a child, he was impressed ### ### by the musical style of bossa nova, specifically the work of tom jobim and ### ### =============================================== v_q ================================================ ### ### ('chico', 0, 0) ('##que', 1, 19) ('##ar', 2, 30) ('¹', 3, 628) ('bu', 4, 24) ('##landa', 5, 11) ### ### ('##vi', 6, 13) ('sergio', 7, 6) ('amelia', 8, 12) ('holland', 9, 1) ('cristina', 10, 10474) ### ### ('intently', 11, 204) ('ana', 12, 14) ('##子', 13, 124) ('al', 14, 34) ('singers', 15, 51) ### ### ('貴', 16, 524) ('ho', 17, 158) ('buenos', 18, 23763) ('³', 19, 2985) ### ### =============================================== v_p ================================================ ### ### ('chico', 0, 0) ('holland', 9, 1) ('##oes', 12280, 2) ('nova', 12401, 3) ('job', 12613, 4) ### ### ('##uch', 22207, 5) ('sergio', 7, 6) ('mi', 13138, 7) ('boss', 5408, 8) ('portuguese', 41, 9) ### ### ('cam', 8624, 10) ('##landa', 5, 11) ('amelia', 8, 12) ('##vi', 6, 13) ('ana', 12, 14) ### ### ('tom', 12898, 15) ('cesar', 20, 16) ('prize', 1153, 17) ('worried', 1371, 18) ('##que', 1, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('chico', 0, 0) ('holland', 9, 1) ('sergio', 7, 6) ('##landa', 5, 11) ('##vi', 6, 13) ### ### ('##que', 1, 19) ('amelia', 8, 12) ('bu', 4, 24) ('##ar', 2, 30) ('ana', 12, 14) ('al', 14, 34) ### ### ('cesar', 20, 16) ('portuguese', 41, 9) ('singers', 15, 51) ('brother', 75, 23) ### ### ('portugal', 53, 37) ('singer', 38, 43) ('herself', 64, 35) ('##a', 63, 36) ('##子', 13, 124) ### ############################################################################################################ [2023-11-25 12:38:48,947][root][INFO] - Epoch: 3: Step: 5001/19530, loss[v]=0.182757, lr=0.000014, acc@1[1]=970.5/1024=0.94775390625, acc@1[2]=999.5/1024=0.97607421875 [2023-11-25 12:40:37,683][root][INFO] - Train batch 5100 [2023-11-25 12:40:37,684][root][INFO] - Avg. loss per last 100 batches: 0.141944 [2023-11-25 12:40:38,396][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29515.7/29522=99.98% | mean: 0.02 | max: 5.61 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.3/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.4/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: that year proust also began working on a novel, which was eventually published in 1952 ### ### and titled " jean santeuil " by his posthumous editors. ### ### [P_TEXT]: goodbye to language didier franck's " nietzsche and the shadow of god ", sigmund freud's ### ### " introductory lectures on psycho - analysis ", julien green's " journal ", victor hugo's " ### ### expiation ", emmanuel levinas's " totality and infinity ", cesar pavese's " the house on the hill ### ### ", ezra pound's " works and usury : three essays ", marcel proust's " jean santeuil " and " la ### ### prisonniere ", rainier maria rilke's " duino elegies ", george sand's " elle et lui " and " lettres ### ### a alfred de musset ", louis antoine de saint - just's " rapport a la convention, march 3, 1794 ", ### ### jean - paul sartre's " ### ### =============================================== v_q ================================================ ### ### ('sant', 0, 13) ('##ust', 1, 45) ('##eu', 2, 54) ('##il', 3, 30) ('posthumous', 4, 1714) ### ### ('novel', 5, 99) ('¹', 6, 109) ('pro', 7, 85) ('intently', 8, 122) ('absently', 9, 133) ### ### ('jean', 10, 74) ('貴', 11, 257) ('খ', 12, 165) ('##¨', 13, 575) ('1952', 14, 24395) ('″', 15, 55) ### ### ('french', 16, 161) ('authors', 17, 691) ('³', 18, 1998) ('幸', 19, 179) ### ### =============================================== v_p ================================================ ### ### ('infinity', 3695, 0) ('journal', 577, 1) ('goodbye', 9010, 2) ('##ury', 21731, 3) ### ### ('language', 1736, 4) ('shadow', 3094, 5) ('levin', 22759, 6) ('##pia', 22738, 7) ### ### ('julien', 3489, 8) ('##ves', 5159, 9) ('elle', 15583, 10) ('lectures', 7894, 11) ### ### ('psycho', 1122, 12) ('sant', 0, 13) ('sand', 5596, 14) ('total', 14400, 15) ('##nier', 9556, 16) ### ### ('analysis', 3282, 17) ('hill', 17978, 18) ('freud', 115, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('sant', 0, 13) ('##il', 3, 30) ('##ust', 1, 45) ('##eu', 2, 54) ('jean', 10, 74) ('pro', 7, 85) ### ### ('″', 15, 55) ('novel', 5, 99) ('¹', 6, 109) ('intently', 8, 122) ('"', 28, 60) ### ### ('absently', 9, 133) ('works', 64, 39) ('hugo', 57, 44) ('anyway', 56, 46) ('freud', 115, 19) ### ### ('journal', 577, 1) ('nietzsche', 89, 51) ('antoine', 117, 53) ('pound', 401, 20) ### ############################################################################################################ [2023-11-25 12:40:38,902][root][INFO] - Epoch: 3: Step: 5101/19530, loss[v]=0.190194, lr=0.000014, acc@1[1]=978.5/1024=0.95556640625, acc@1[2]=1007.5/1024=0.98388671875 [2023-11-25 12:42:28,076][root][INFO] - Train batch 5200 [2023-11-25 12:42:28,077][root][INFO] - Avg. loss per last 100 batches: 0.150371 [2023-11-25 12:42:28,831][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.78 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.1/29522=99.97% | mean: 0.02 | max: 5.62 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it was the holding company for the three deep - level " tube " underground railway lines ### ### opened in london during 1906 and 1907 : the baker street and waterloo railway, the charing cross, ### ### euston and hampstead railway and the great northern, piccadilly and brompton railway. ### ### [P_TEXT]: london underground standard stock specification, which ensured that each car would have ### ### 48 seats and two sets of air - operated double doors on each side, providing an opening which was ### ### wide. beyond these basic guidelines, each builder was allowed to build a trailer car to its own ### ### design, although externally, they looked very similar. the control trailer was designed by the ### ### underground electric railways. the stock became known as the 1922 stock or competition stock and ### ### was considered part of the standard stock fleet. the earliest standard stock was built for use on ### ### the hampstead tube, which was extended ### ### =============================================== v_q ================================================ ### ### ('railway', 0, 51) ('absently', 1, 127) ('tube', 2, 10) ('london', 3, 15) ('underground', 4, 5) ### ### ('intently', 5, 152) ('##mpton', 6, 14039) ('1906', 7, 1016) ('¹', 8, 366) ('company', 9, 969) ### ### ('貴', 10, 259) ('baker', 11, 8144) ('deep', 12, 5574) ('hampstead', 13, 0) ('খ', 14, 133) ### ### ('.', 15, 419) ('waterloo', 16, 2731) ('holding', 17, 9754) ('great', 18, 17695) ### ### ('char', 19, 17546) ### ### =============================================== v_p ================================================ ### ### ('hampstead', 13, 0) ('stock', 113, 1) ('trailer', 8144, 2) ('standard', 947, 3) ('1922', 252, 4) ### ### ('underground', 4, 5) ('competition', 2289, 6) ('doors', 2612, 7) ('trailers', 27078, 8) ### ### ('control', 1735, 9) ('tube', 2, 10) ('fleet', 1274, 11) ('electric', 790, 12) ('air', 5452, 13) ### ### ('specifications', 16420, 14) ('london', 3, 15) ('anyway', 150, 16) ('fleets', 27899, 17) ### ### ('specification', 1186, 18) ('earliest', 2195, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('hampstead', 13, 0) ('underground', 4, 5) ('tube', 2, 10) ('london', 3, 15) ('railway', 0, 51) ### ### ('stock', 113, 1) ('1922', 252, 4) ('absently', 1, 127) ('locomotives', 31, 36) ('anyway', 150, 16) ### ### ('railways', 111, 21) ('standard', 947, 3) ('worried', 225, 20) ('intently', 5, 152) ### ### ('trains', 108, 27) ('electric', 790, 12) ('খ', 14, 133) ('fleet', 1274, 11) ### ### ('competition', 2289, 6) ('control', 1735, 9) ### ############################################################################################################ [2023-11-25 12:42:29,339][root][INFO] - Epoch: 3: Step: 5201/19530, loss[v]=0.191145, lr=0.000014, acc@1[1]=977.5/1024=0.95458984375, acc@1[2]=1007.0/1024=0.9833984375 [2023-11-25 12:44:18,442][root][INFO] - Train batch 5300 [2023-11-25 12:44:18,442][root][INFO] - Avg. loss per last 100 batches: 0.149512 [2023-11-25 12:44:19,128][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29515.1/29522=99.98% | mean: 0.02 | max: 5.67 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: those privileges became the basis of the " golden liberty " in the polish – lithuanian ### ### commonwealth. ### ### [P_TEXT]: stanisław staszic improvements of the peasants'fate ( by granting them land and private ### ### rights ). he criticised the szlachta ( polish nobility ) for inefficient governance, and argued ### ### that it showed itself too inept to be allowed to govern alone. he argued for a slight increase in ### ### taxes, which should allow the commonwealth to create an army of 100, 000 that would at least stand ### ### a chance against the still - larger armies of its neighbours. although he preferred republicanism ### ### in theory, in the commonwealth context he agreed that a strengthening of the central ( royal ) ### ### power was the most practical solution ### ### =============================================== v_q ================================================ ### ### ('golden', 0, 10369) ('##ʂ', 1, 1002) ('polish', 2, 10) ('intently', 3, 390) ('liberty', 4, 1530) ### ### ('commonwealth', 5, 6) ('lithuanian', 6, 716) ('absently', 7, 260) ('¹', 8, 472) ('.', 9, 61) ### ### ('lithuania', 10, 508) ('貴', 11, 718) ('poland', 12, 34) ('freedom', 13, 777) ('খ', 14, 131) ### ### ('obliged', 15, 1415) ('decreed', 16, 13934) ('privileges', 17, 103) ('duchy', 18, 1847) ### ### ('vilnius', 19, 405) ### ### =============================================== v_p ================================================ ### ### ('##zic', 2496, 0) ('stanisław', 30, 1) ('nobility', 102, 2) ('peasants', 55, 3) ### ### ('republican', 1572, 4) ('##cht', 10474, 5) ('commonwealth', 5, 6) ('##as', 11395, 7) ### ### ('private', 2296, 8) ('armies', 3582, 9) ('polish', 2, 10) ('governance', 2795, 11) ### ### ('anyway', 65, 12) ('taxes', 13900, 13) ('republicans', 25701, 14) ('royal', 415, 15) ### ### ('100', 8082, 16) ('improvements', 8509, 17) ('monarch', 256, 18) ('##ism', 729, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('polish', 2, 10) ('commonwealth', 5, 6) ('stanisław', 30, 1) ('poland', 12, 34) ### ### ('peasants', 55, 3) ('nobility', 102, 2) ('.', 9, 61) ('anyway', 65, 12) ('absently', 7, 260) ### ### ('intently', 3, 390) ('খ', 14, 131) ('##zic', 2496, 0) ('powers', 26, 69) ('monarch', 256, 18) ### ### ('royal', 415, 15) ('granted', 57, 54) ('privileges', 17, 103) ('privilege', 27, 75) ### ### ('land', 88, 46) ('republican', 1572, 4) ### ############################################################################################################ [2023-11-25 12:44:19,633][root][INFO] - Epoch: 3: Step: 5301/19530, loss[v]=0.216656, lr=0.000014, acc@1[1]=961.0/1024=0.9384765625, acc@1[2]=998.5/1024=0.97509765625 [2023-11-25 12:46:08,350][root][INFO] - Train batch 5400 [2023-11-25 12:46:08,350][root][INFO] - Avg. loss per last 100 batches: 0.155612 [2023-11-25 12:46:09,066][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.73 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29512.9/29522=99.97% | mean: 0.02 | max: 5.38 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: as a result of this, shaun pollock resigned as captain and was replaced by young batsman ### ### graeme smith, although pollock continued to play for the team. ### ### [P_TEXT]: graeme smith graeme craig smith ( born 1 february 1981 ) is a south african cricket ### ### commentator and former cricketer, who played for south africa in all formats. in 2003, he was ### ### appointed captain of the national team, taking over from shaun pollock. he held the position of ### ### test captain until his retirement in 2014. he is regarded as one of the greatest captains of all ### ### time in test cricket. at 22, he was appointed as south africa's youngest ever captain. a tall, left ### ### - handed opening batsman, smith is regarded as one of the best ### ### =============================================== v_q ================================================ ### ### ('shaun', 0, 10) ('graeme', 1, 0) ('pollock', 2, 1) ('smith', 3, 2) ('cricket', 4, 7) ('¹', 5, 377) ### ### ('.', 6, 604) ('absently', 7, 99) ('intently', 8, 171) ('cricketer', 9, 84) ('ɐ', 10, 440) ### ### ('captain', 11, 17) ('貴', 12, 417) ('darts', 13, 2272) ('england', 14, 503) ('খ', 15, 142) ### ### ('test', 16, 11) ('he', 17, 57) ('list', 18, 3597) ('runs', 19, 108) ### ### =============================================== v_p ================================================ ### ### ('graeme', 1, 0) ('pollock', 2, 1) ('smith', 3, 2) ('worried', 651, 3) ('handed', 177, 4) ### ### ('africa', 212, 5) ('batsman', 148, 6) ('cricket', 4, 7) ('formats', 3177, 8) ### ### ('commentator', 363, 9) ('shaun', 0, 10) ('test', 16, 11) ('captains', 261, 12) ('anyway', 483, 13) ### ### ('littered', 1162, 14) ('commentary', 2489, 15) ('wickets', 3488, 16) ('captain', 11, 17) ### ### ('craig', 4772, 18) ('referees', 27142, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('graeme', 1, 0) ('pollock', 2, 1) ('shaun', 0, 10) ('smith', 3, 2) ('cricket', 4, 7) ### ### ('captain', 11, 17) ('test', 16, 11) ('batsman', 148, 6) ('handed', 177, 4) ('africa', 212, 5) ### ### ('captains', 261, 12) ('overs', 32, 36) ('cricketer', 9, 84) ('absently', 7, 99) ### ### ('commentator', 363, 9) ('2003', 199, 20) ('worried', 651, 3) ('match', 33, 43) ('team', 48, 35) ### ### ('he', 17, 57) ### ############################################################################################################ [2023-11-25 12:46:09,571][root][INFO] - Epoch: 3: Step: 5401/19530, loss[v]=0.090856, lr=0.000014, acc@1[1]=997.0/1024=0.9736328125, acc@1[2]=1012.0/1024=0.98828125 [2023-11-25 12:47:58,783][root][INFO] - Train batch 5500 [2023-11-25 12:47:58,784][root][INFO] - Avg. loss per last 100 batches: 0.150084 [2023-11-25 12:47:59,512][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29515.1/29522=99.98% | mean: 0.02 | max: 5.45 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: a separate civic police force, the unarmed dublin metropolitan police, patrolled the ### ### capital, while the cities of derry and belfast, originally with their own police forces, later had ### ### special divisions within the ric. ### ### [P_TEXT]: visit to ireland medal 1911 visit to ireland medals, a medal was awarded to those members ### ### of the royal irish constabulary and the dublin metropolitan police who were on duty at engagements ### ### during king george's visit. additionally, members of the civil service staffs of both police forces ### ### who were on duty at dublin castle during the visit received the medal, as did some members of the ### ### dublin st johns ambulance brigade, the kingstown harbour police and the local parks constabularies. ### ### prior to the irish visit, a delegation of 585 members of the royal irish constabulary had travelled ### ### to london and were on duty at king george v ’ s coronation in june 1911 and received the ### ### =============================================== v_q ================================================ ### ### ('police', 0, 9) ('civic', 1, 10708) ('dublin', 2, 3) ('ric', 3, 19460) ('absently', 4, 115) ### ### ('derry', 5, 4595) ('metropolitan', 6, 20) ('intently', 7, 76) ('.', 8, 416) ('city', 9, 5179) ### ### ('ireland', 10, 18) ('belfast', 11, 377) ('irish', 12, 24) ('¹', 13, 431) ('貴', 14, 241) ### ### ('খ', 15, 178) ('special', 16, 7382) ('constabulary', 17, 8) ('##ப', 18, 294) ('##merie', 19, 1232) ### ### =============================================== v_p ================================================ ### ### ('coronation', 1735, 0) ('##bular', 26021, 1) ('1911', 418, 2) ('dublin', 2, 3) ('johns', 4962, 4) ### ### ('visit', 1861, 5) ('ambulance', 870, 6) ('castle', 213, 7) ('constabulary', 17, 8) ### ### ('police', 0, 9) ('harbour', 364, 10) ('anyway', 179, 11) ('parks', 9670, 12) ('medal', 5234, 13) ### ### ('kings', 13243, 14) ('##town', 1129, 15) ('brigade', 173, 16) ('medals', 6236, 17) ### ### ('ireland', 10, 18) ('brigades', 2549, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('police', 0, 9) ('dublin', 2, 3) ('metropolitan', 6, 20) ('ireland', 10, 18) ### ### ('constabulary', 17, 8) ('irish', 12, 24) ('castle', 213, 7) ('intently', 7, 76) ### ### ('anyway', 179, 11) ('1911', 418, 2) ('absently', 4, 115) ('brigade', 173, 16) ('harbour', 364, 10) ### ### ('ambulance', 870, 6) ('worried', 111, 28) ('coronation', 1735, 0) ('visit', 1861, 5) ### ### ('##town', 1129, 15) ('forces', 28, 55) ('.', 8, 416) ### ############################################################################################################ [2023-11-25 12:48:00,019][root][INFO] - Epoch: 3: Step: 5501/19530, loss[v]=0.163685, lr=0.000014, acc@1[1]=976.0/1024=0.953125, acc@1[2]=999.0/1024=0.9755859375 [2023-11-25 12:49:48,897][root][INFO] - Train batch 5600 [2023-11-25 12:49:48,897][root][INFO] - Avg. loss per last 100 batches: 0.137553 [2023-11-25 12:49:49,596][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.78 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.2/29522=99.97% | mean: 0.02 | max: 5.63 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: during this period, his hearing began to deteriorate, but he continued to conduct, ### ### premiering his third and fifth symphonies in 1804 and 1808, respectively. ### ### [P_TEXT]: electric light orchestra been all but drowned out by the electrified instruments. the ### ### band released their second album " elo 2 " in early 1973, which produced their second uk top 10 and ### ### their first us chart single, an elaborate version of the chuck berry classic " roll over beethoven ### ### " ( which also incorporated the first movement of beethoven's own fifth symphony ). elo also made ### ### their first appearance on " american bandstand ". during the recording of the third album, gibson ### ### was let go after a dispute over money, ### ### =============================================== v_q ================================================ ### ### ('symphony', 0, 1) ('beethoven', 1, 0) ('fifth', 2, 7) ('1808', 3, 6117) ('intently', 4, 105) ### ### ('.', 5, 1507) ('third', 6, 52) ('mozart', 7, 88) ('##¨', 8, 1166) ('1804', 9, 7490) ('¹', 10, 236) ### ### ('absently', 11, 67) ('5th', 12, 79) ('gustav', 13, 12337) ('haydn', 14, 432) ('貴', 15, 217) ### ### ('conductors', 16, 4574) ('premier', 17, 17735) ('1803', 18, 11853) ('##₃', 19, 39) ### ### =============================================== v_p ================================================ ### ### ('beethoven', 1, 0) ('symphony', 0, 1) ('electric', 4684, 2) ('el', 9976, 3) ('symphonies', 195, 4) ### ### ('electrified', 20687, 5) ('gibson', 9514, 6) ('fifth', 2, 7) ('orchestra', 54, 8) ### ### ('light', 2084, 9) ('drowned', 5885, 10) ('berry', 16901, 11) ('roll', 16837, 12) ('2', 17761, 13) ### ### ('##tan', 22826, 14) ('leaned', 6534, 15) ('dispute', 19810, 16) ('1973', 15418, 17) ### ### ('bands', 11394, 18) ('anyway', 63, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('symphony', 0, 1) ('beethoven', 1, 0) ('fifth', 2, 7) ('orchestra', 54, 8) ('third', 6, 52) ### ### ('symphonies', 195, 4) ('anyway', 63, 19) ('##₃', 19, 39) ('intently', 4, 105) ('worried', 127, 20) ### ### ('absently', 11, 67) ('mozart', 7, 88) ('5th', 12, 79) ('″', 259, 21) ('song', 273, 27) ### ### ('¹', 10, 236) ('light', 2084, 9) ('electric', 4684, 2) ('ம', 43, 90) ('貴', 15, 217) ### ############################################################################################################ [2023-11-25 12:49:50,100][root][INFO] - Epoch: 3: Step: 5601/19530, loss[v]=0.205591, lr=0.000014, acc@1[1]=974.0/1024=0.951171875, acc@1[2]=994.5/1024=0.97119140625 [2023-11-25 12:51:39,059][root][INFO] - Train batch 5700 [2023-11-25 12:51:39,060][root][INFO] - Avg. loss per last 100 batches: 0.157513 [2023-11-25 12:51:39,770][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.94 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29515.9/29522=99.98% | mean: 0.03 | max: 5.62 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: clear lake township is located in sangamon county, illinois. ### ### [P_TEXT]: barclay, illinois barclay is an unincorporated community in clear lake and williams ### ### townships, sangamon county, illinois, united states. barclay is located on illinois route 54 and ### ### the canadian national railway, northeast of spaulding. ### ### =============================================== v_q ================================================ ### ### ('illinois', 0, 2) ('clear', 1, 3) ('lake', 2, 12) ('¹', 3, 275) ('sang', 4, 9) ('##amo', 5, 6) ### ### ('township', 6, 22) ('intently', 7, 95) ('county', 8, 25) ('³', 9, 1932) ('貴', 10, 145) ### ### ('খ', 11, 115) ('michigan', 12, 224) ('jersey', 13, 7351) ('.', 14, 3280) ('absently', 15, 65) ### ### ('##ப', 16, 133) ('##n', 17, 97) ('##უ', 18, 226) ('chicago', 19, 32) ### ### =============================================== v_p ================================================ ### ### ('barclay', 13883, 0) ('54', 4872, 1) ('illinois', 0, 2) ('clear', 1, 3) ('spa', 20878, 4) ### ### ('williams', 11501, 5) ('##amo', 5, 6) ('##uld', 26050, 7) ('canadian', 1427, 8) ('sang', 4, 9) ### ### ('worried', 498, 10) ('townships', 368, 11) ('lake', 2, 12) ('railway', 17057, 13) ### ### ('##ing', 9164, 14) ('anyway', 73, 15) ('route', 520, 16) ('routes', 2089, 17) ### ### ('national', 4390, 18) ('counties', 25, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('illinois', 0, 2) ('clear', 1, 3) ('##amo', 5, 6) ('sang', 4, 9) ('lake', 2, 12) ### ### ('township', 6, 22) ('county', 8, 25) ('counties', 25, 19) ('intently', 7, 95) ('anyway', 73, 15) ### ### ('chicago', 19, 32) ('¹', 3, 275) ('townships', 368, 11) ('worried', 498, 10) ('absently', 15, 65) ### ### ('খ', 11, 115) ('unincorporated', 32, 38) ('貴', 10, 145) ('canadian', 1427, 8) ### ### ('population', 155, 26) ### ############################################################################################################ [2023-11-25 12:51:40,276][root][INFO] - Epoch: 3: Step: 5701/19530, loss[v]=0.143698, lr=0.000014, acc@1[1]=985.5/1024=0.96240234375, acc@1[2]=1011.5/1024=0.98779296875 [2023-11-25 12:53:29,960][root][INFO] - Train batch 5800 [2023-11-25 12:53:29,962][root][INFO] - Avg. loss per last 100 batches: 0.151686 [2023-11-25 12:53:30,670][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.82 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29513.6/29522=99.97% | mean: 0.02 | max: 5.57 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 2016, former correspondent samantha bee launched her own late - night talk show " full ### ### frontal with samantha bee ". ### ### [P_TEXT]: jon wurster of guided by voices ( 2006 ). in november, 2006 superchunk played with the ### ### mountain goats at " the daily show " tenth anniversary party at irving plaza in new york city. in ### ### addition to performing a comedy sketch with samantha bee of " daily show ", wurster played three ### ### songs with the mountain goats. as a result, he was invited to drum on the duo's west coast tour in ### ### 2007. wurster played drums and percussion on the mountain goats'2007 album " heretic pride " and ### ### joined the band after the release of the album. ### ### =============================================== v_q ================================================ ### ### ('frontal', 0, 25991) ('bee', 1, 5) ('samantha', 2, 1) ('full', 3, 11865) ### ### ('correspondent', 4, 4418) ('¹', 5, 354) ('reporter', 6, 327) ('absently', 7, 48) ### ### ('journalist', 8, 3749) ('analyst', 9, 9588) ('intently', 10, 162) ('.', 11, 4154) ('貴', 12, 320) ### ### ('！', 13, 2418) ('2016', 14, 8680) ('##子', 15, 504) ('talk', 16, 4781) ('news', 17, 1880) ### ### ('commentator', 18, 2349) ('hillary', 19, 17532) ### ### =============================================== v_p ================================================ ### ### ('irving', 9967, 0) ('samantha', 2, 1) ('goats', 10420, 2) ('guided', 22271, 3) ('##rst', 19598, 4) ### ### ('bee', 1, 5) ('anniversary', 5914, 6) ('daily', 84, 7) ('pride', 4613, 8) ('plaza', 5746, 9) ### ### ('wu', 21789, 10) ('##chu', 28251, 11) ('jon', 923, 12) ('##tic', 18234, 13) ('voices', 3391, 14) ### ### ('mountain', 9560, 15) ('tenth', 15470, 16) ('leaned', 434, 17) ('show', 67, 18) ('tour', 5722, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('samantha', 2, 1) ('bee', 1, 5) ('daily', 84, 7) ('absently', 7, 48) ('show', 67, 18) ### ### ('she', 40, 30) ('herself', 26, 42) ('anyway', 206, 25) ('leaned', 434, 17) ('jon', 923, 12) ### ### ('intently', 10, 162) ('reporter', 6, 327) ('¹', 5, 354) ('worried', 794, 22) ('2007', 1406, 21) ### ### ('hissed', 161, 51) ('venue', 1203, 27) ('hosted', 86, 81) ('貴', 12, 320) ('voices', 3391, 14) ### ############################################################################################################ [2023-11-25 12:53:31,175][root][INFO] - Epoch: 3: Step: 5801/19530, loss[v]=0.190547, lr=0.000014, acc@1[1]=975.0/1024=0.9521484375, acc@1[2]=1005.5/1024=0.98193359375 [2023-11-25 12:55:20,074][root][INFO] - Train batch 5900 [2023-11-25 12:55:20,076][root][INFO] - Avg. loss per last 100 batches: 0.144003 [2023-11-25 12:55:20,775][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29513.6/29522=99.97% | mean: 0.02 | max: 5.61 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: livingston is a village in madison county, illinois, united states. ### ### [P_TEXT]: olive township, madison county, illinois olive township is located in madison county, ### ### illinois, in the united states. as of the 2010 census, its population was 1, 785 and it contained ### ### 842 housing units. the township's administrative offices are located in the village of livingston. ### ### olive township was organized in 1876. abel olive was an early settler. according to the 2010 ### ### census, the township has a total area of, of which ( or 98. 95 % ) is land and ( or 1. 05 % ) is ### ### water. interstate 55, illinois route 4 and the former u. s. route 66 all pass through the township. ### ### there are two villages located in the township : livingston ### ### =============================================== v_q ================================================ ### ### ('livingston', 0, 0) ('illinois', 1, 5) ('madison', 2, 4) ('¹', 3, 198) ('county', 4, 18) ### ### ('.', 5, 317) ('village', 6, 20) ('u', 7, 127) ('neighboring', 8, 201) ('³', 9, 737) ### ### ('sprang', 10, 368) ('prairie', 11, 954) ('portions', 12, 56) ('iowa', 13, 434) ### ### ('indiana', 14, 158) ('chicago', 15, 49) ('counties', 16, 13) ('cabin', 17, 2620) ('খ', 18, 121) ### ### ('jenks', 19, 117) ### ### =============================================== v_p ================================================ ### ### ('livingston', 0, 0) ('olive', 7253, 1) ('55', 3324, 2) ('66', 3078, 3) ('madison', 2, 4) ### ### ('illinois', 1, 5) ('4', 8896, 6) ('abel', 10345, 7) ('township', 76, 8) ('worried', 2399, 9) ### ### ('anyway', 1498, 10) ('interchange', 11949, 11) ('interstate', 584, 12) ('counties', 16, 13) ### ### ('route', 468, 14) ('leaned', 3231, 15) ('seat', 2092, 16) ('passes', 12276, 17) ('county', 4, 18) ### ### ('freeway', 20287, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('livingston', 0, 0) ('illinois', 1, 5) ('madison', 2, 4) ('county', 4, 18) ('village', 6, 20) ### ### ('counties', 16, 13) ('township', 76, 8) ('villages', 26, 23) ('¹', 3, 198) ('settlers', 59, 24) ### ### ('66', 3078, 3) ('.', 5, 317) ('chicago', 15, 49) ('55', 3324, 2) ('portions', 12, 56) ### ### ('route', 468, 14) ('interstate', 584, 12) ('driveway', 274, 28) ('olive', 7253, 1) ### ### ('population', 189, 34) ### ############################################################################################################ [2023-11-25 12:55:21,279][root][INFO] - Epoch: 3: Step: 5901/19530, loss[v]=0.217916, lr=0.000013, acc@1[1]=965.0/1024=0.9423828125, acc@1[2]=994.0/1024=0.970703125 [2023-11-25 12:57:10,424][root][INFO] - Train batch 6000 [2023-11-25 12:57:10,426][root][INFO] - Avg. loss per last 100 batches: 0.146097 [2023-11-25 12:57:11,159][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.04 | max: 4.00 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.4/29522=99.97% | mean: 0.02 | max: 5.73 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: other pupils were thomas simon cool, charles napier hemy, willem linnig the elder, petrus ### ### marius molijn, hendrik albert van trigt and alexis van hamme. ### ### [P_TEXT]: charles napier hemy charles napier hemy ( newcastle - on - tyne 24 may 1841 – 30 ### ### september 1917 falmouth ) was a british painter best known for his marine paintings and his ### ### paintings in the tate collections. he was born to a musical family in newcastle - on - tyne and his ### ### two brothers, thomas and bernard, were also painters. he trained in the government school of ### ### design, newcastle, followed by the antwerp academy and the studio of baron leys. he returned to ### ### london in the 1870s and in 1881 moved to the coastal town of falmouth in cornwall. he produced ### ### painted ### ### =============================================== v_q ================================================ ### ### ('cool', 0, 6387) ('alexis', 1, 12181) ('hem', 2, 4) ('napier', 3, 1) ('ham', 4, 3662) ### ### ('##nig', 5, 23515) ('##gt', 6, 28619) ('tri', 7, 18680) ('lin', 8, 16839) ('##me', 9, 22993) ### ### ('##jn', 10, 28883) ('##rus', 11, 23889) ('charles', 12, 13) ('##li', 13, 23835) ### ### ('lessons', 14, 1160) ('simon', 15, 8122) ('¹', 16, 604) ('thomas', 17, 22) ('hen', 18, 20265) ### ### ('marius', 19, 12685) ### ### =============================================== v_p ================================================ ### ### ('ley', 8942, 0) ('napier', 3, 1) ('antwerp', 75, 2) ('newcastle', 2494, 3) ('hem', 2, 4) ### ### ('bernard', 5955, 5) ('design', 502, 6) ('worried', 1034, 7) ('tate', 3011, 8) ('academy', 393, 9) ### ### ('government', 10220, 10) ('##outh', 20750, 11) ('school', 163, 12) ('charles', 12, 13) ### ### ('##lm', 19321, 14) ('leaned', 73, 15) ('marine', 15225, 16) ('paintings', 40, 17) ### ### ('anyway', 249, 18) ('collections', 595, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('napier', 3, 1) ('hem', 2, 4) ('charles', 12, 13) ('thomas', 17, 22) ('antwerp', 75, 2) ### ### ('paintings', 40, 17) ('painted', 41, 24) ('leaned', 73, 15) ('painting', 39, 29) ### ### ('school', 163, 12) ('painter', 85, 28) ('academy', 393, 9) ('##y', 36, 50) ('design', 502, 6) ### ### ('anyway', 249, 18) ('pupils', 50, 47) ('##ɨ', 124, 39) ('worried', 1034, 7) ### ### ('collection', 248, 26) ('collections', 595, 19) ### ############################################################################################################ [2023-11-25 12:57:11,665][root][INFO] - Epoch: 3: Step: 6001/19530, loss[v]=0.083535, lr=0.000013, acc@1[1]=1004.0/1024=0.98046875, acc@1[2]=1019.0/1024=0.9951171875 [2023-11-25 12:59:01,173][root][INFO] - Train batch 6100 [2023-11-25 12:59:01,174][root][INFO] - Avg. loss per last 100 batches: 0.145088 [2023-11-25 12:59:01,889][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.80 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29512.6/29522=99.97% | mean: 0.02 | max: 5.46 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: king was appointed the manager of the ireland women's senior team in 2000. ### ### [P_TEXT]: susan ronan susan " sue " ronan ( born 21 february 1964 ) is an irish football coach and ### ### former player, who managed the republic of ireland women's national football team from 2010 until ### ### 2016. ronan represented welsox and shelbourne at club level, and made her debut as a player for ### ### ireland in 1988. she won 22 caps and was named fai women's player of the year in 1993. in october ### ### 2010 she succeeded noel king as head coach of the senior republic of ireland women's national ### ### football team. she stood down from the position ### ### =============================================== v_q ================================================ ### ### ('king', 0, 7) ('ireland', 1, 4) ('women', 2, 8) ('¹', 3, 215) ('manager', 4, 58) ### ### ('intently', 5, 134) ('herself', 6, 21) ('³', 7, 1326) ('senior', 8, 24) ('footballer', 9, 63) ### ### ('貴', 10, 172) ('irish', 11, 28) ('ɐ', 12, 360) ('##ท', 13, 192) ('she', 14, 51) ### ### ('gael', 15, 23897) ('team', 16, 29) ('##ப', 17, 140) ('gaa', 18, 13555) ('football', 19, 19) ### ### =============================================== v_p ================================================ ### ### ('susan', 5012, 0) ('ronan', 1698, 1) ('noel', 3724, 2) ('fai', 458, 3) ('ireland', 1, 4) ### ### ('##ox', 28841, 5) ('worried', 1434, 6) ('king', 0, 7) ('women', 2, 8) ('anyway', 81, 9) ### ### ('##ls', 17143, 10) ('##lb', 21970, 11) ('republic', 1821, 12) ('caps', 1067, 13) ### ### ('##our', 22301, 14) ('1988', 7978, 15) ('we', 6974, 16) ('shirt', 27, 17) ('littered', 612, 18) ### ### ('football', 19, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('king', 0, 7) ('ireland', 1, 4) ('women', 2, 8) ('herself', 6, 21) ('senior', 8, 24) ### ### ('irish', 11, 28) ('football', 19, 19) ('manager', 4, 58) ('shirt', 27, 17) ('anyway', 81, 9) ### ### ('team', 16, 29) ('footballer', 9, 63) ('coach', 33, 30) ('fai', 458, 3) ('she', 14, 51) ### ### ('¹', 3, 215) ('intently', 5, 134) ('ronan', 1698, 1) ('players', 55, 35) ('teams', 72, 33) ### ############################################################################################################ [2023-11-25 12:59:02,394][root][INFO] - Epoch: 3: Step: 6101/19530, loss[v]=0.186256, lr=0.000013, acc@1[1]=964.0/1024=0.94140625, acc@1[2]=998.5/1024=0.97509765625 [2023-11-25 13:00:51,504][root][INFO] - Train batch 6200 [2023-11-25 13:00:51,505][root][INFO] - Avg. loss per last 100 batches: 0.142306 [2023-11-25 13:00:52,211][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.74 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29515.5/29522=99.98% | mean: 0.03 | max: 5.69 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: ray burke becoming minister for foreign affairs. ### ### [P_TEXT]: albert reynolds and michael woods in the fianna fail leadership election, and succeeded ### ### haughey as taoiseach on 11 february 1992. the ministers who had been sacked along with reynolds at ### ### the end of 1991 were all appointed to cabinet, while eight members of haughey's cabinet, including ### ### such long - serving haughey loyalists as ray burke, mary o'rourke and gerry collins, were left out. ### ### nine of the twelve junior ministers, many of whom were also haughey supporters, were also ### ### dismissed. reynolds promoted several long - time critics of haughey, like david andrews, seamus ### ### brennan and charlie mccreevy to senior ministerial positions. reynolds also promoted younger ### ### =============================================== v_q ================================================ ### ### ('burke', 0, 4) ('ray', 1, 11) ('foreign', 2, 5696) ('minister', 3, 47) ('intently', 4, 134) ### ### ('absently', 5, 141) ('¹', 6, 52) ('election', 7, 25) ('.', 8, 313) ('affairs', 9, 2821) ### ### ('貴', 10, 130) ('anyway', 11, 23) ('খ', 12, 128) ('zealand', 13, 861) ('australian', 14, 1508) ### ### ('honours', 15, 730) ('！', 16, 881) ('overseas', 17, 5982) ('liberal', 18, 398) ### ### ('australia', 19, 1632) ### ### =============================================== v_p ================================================ ### ### ('reynolds', 5540, 0) ('andrews', 907, 1) ('rourke', 8687, 2) ('woods', 6837, 3) ('burke', 0, 4) ### ### ('albert', 1099, 5) ('brennan', 3000, 6) ('loyalists', 2228, 7) ('gerry', 106, 8) ### ### ('##ugh', 25933, 9) ('seamus', 584, 10) ('ray', 1, 11) ('cabinet', 22, 12) ('##vy', 27526, 13) ### ### ('collins', 1317, 14) ('dismiss', 18553, 15) ('sacked', 2768, 16) ('leadership', 167, 17) ### ### ('tao', 21901, 18) ('junior', 2630, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('burke', 0, 4) ('ray', 1, 11) ('election', 7, 25) ('anyway', 11, 23) ('minister', 3, 47) ### ### ('cabinet', 22, 12) ('¹', 6, 52) ('gerry', 106, 8) ('hissed', 45, 30) ('intently', 4, 134) ### ### ('leadership', 167, 17) ('ireland', 56, 37) ('andrews', 907, 1) ('criticised', 41, 46) ### ### ('absently', 5, 141) ('貴', 10, 130) ('seamus', 584, 10) ('1991', 344, 20) ('albert', 1099, 5) ### ### ('খ', 12, 128) ### ############################################################################################################ [2023-11-25 13:00:52,716][root][INFO] - Epoch: 3: Step: 6201/19530, loss[v]=0.212822, lr=0.000013, acc@1[1]=975.0/1024=0.9521484375, acc@1[2]=1002.0/1024=0.978515625 [2023-11-25 13:02:41,976][root][INFO] - Train batch 6300 [2023-11-25 13:02:41,977][root][INFO] - Avg. loss per last 100 batches: 0.146264 [2023-11-25 13:02:42,706][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.75 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29516.1/29522=99.98% | mean: 0.03 | max: 5.71 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: cindy chang of " the times - picayune " said former live oak elementary school was ### ### performing so poorly that paul vallas, the superintendent of the rsd, gave the school to ### ### [P_TEXT]: renew schools renew schools is a charter school management organization in new orleans, ### ### united states. its headquarters are in the batiste cultural arts academy school facility, in the ### ### former live oak elementary school in the irish channel neighborhood. renew is a charter management ### ### organization that specializes in acquiring low performing schools. around 2009 live oak elementary ### ### school was performing so poorly that paul vallas, the head of the recovery school district ( rsd ), ### ### gave the school to the renew. in 2019, the renew staff announced to their staff via e - mail that ### ### they would ### ### =============================================== v_q ================================================ ### ### ('val', 0, 8) ('oak', 1, 7) ('rs', 2, 6) ('live', 3, 10) ('##las', 4, 12) ('##une', 5, 23900) ### ### ('school', 6, 17) ('intently', 7, 261) ('paul', 8, 32) ('pic', 9, 24353) ('¹', 10, 607) ### ### ('cindy', 11, 11570) ('absently', 12, 73) ('##ay', 13, 28603) ('貴', 14, 309) ### ### ('cerambycidae', 15, 820) ('ד', 16, 168) ('chicago', 17, 426) ('³', 18, 3335) ### ### ('schoolhouse', 19, 36) ### ### =============================================== v_p ================================================ ### ### ('renew', 7118, 0) ('recovery', 5713, 1) ('irish', 1908, 2) ('channel', 5037, 3) ### ### ('orleans', 200, 4) ('charter', 5436, 5) ('rs', 2, 6) ('oak', 1, 7) ('val', 0, 8) ### ### ('##iste', 27973, 9) ('live', 3, 10) ('bat', 15325, 11) ('##las', 4, 12) ('management', 4836, 13) ### ### ('schools', 122, 14) ('littered', 4757, 15) ('email', 4440, 16) ('school', 6, 17) ### ### ('low', 16951, 18) ('facility', 2213, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('val', 0, 8) ('oak', 1, 7) ('rs', 2, 6) ('live', 3, 10) ('##las', 4, 12) ('school', 6, 17) ### ### ('paul', 8, 32) ('schoolhouse', 19, 36) ('elementary', 25, 35) ('neighborhood', 28, 34) ### ### ('orleans', 200, 4) ('classrooms', 39, 33) ('schools', 122, 14) ('absently', 12, 73) ### ### ('neighborhoods', 116, 23) ('anyway', 117, 21) ('louisiana', 245, 24) ('cradled', 110, 42) ### ### ('schooling', 230, 30) ('intently', 7, 261) ### ############################################################################################################ [2023-11-25 13:02:43,213][root][INFO] - Epoch: 3: Step: 6301/19530, loss[v]=0.205570, lr=0.000013, acc@1[1]=972.5/1024=0.94970703125, acc@1[2]=1002.0/1024=0.978515625 [2023-11-25 13:04:32,364][root][INFO] - Train batch 6400 [2023-11-25 13:04:32,365][root][INFO] - Avg. loss per last 100 batches: 0.152539 [2023-11-25 13:04:33,076][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29510.7/29522=99.96% | mean: 0.02 | max: 5.56 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: nothing is known for sure about hermeric before 419, the year in which he is first ### ### mentioned ; namely, he became king of the suebi ( or suevi ) in the city of braga ( bracara augusta ### ### ) according to bishop hydatius ( who wrote his chronicle around the year 470 ). ### ### [P_TEXT]: braga of the newly founded roman province of gallaecia. during the germanic invasions of ### ### the iberian peninsula, the area was conquered by the suebi, a germanic people from central europe. ### ### in 410, the suebi established a kingdom in northwest iberia covering what is present - day's ### ### northern half of portugal, galicia and asturias, which they maintained as " gallaecia ", and had ### ### bracara as their capital. this kingdom was founded by hermeric and lasted for over 150 years. by ### ### about 584, the visigoths took over control of gallaecia ### ### =============================================== v_q ================================================ ### ### ('##meric', 0, 1) ('bra', 1, 0) ('##ati', 2, 19009) ('absently', 3, 86) ('¹', 4, 70) ### ### ('##car', 5, 6) ('intently', 6, 174) ('##ga', 7, 2) ('##bi', 8, 21) ('sue', 9, 7) ('470', 10, 7872) ### ### ('貴', 11, 172) ('her', 12, 38) ('41', 13, 3074) ('##yd', 14, 26415) ('augusta', 15, 6581) ### ### ('³', 16, 1390) ('sparhawk', 17, 1460) ('খ', 18, 127) ('##vi', 19, 18902) ### ### =============================================== v_p ================================================ ### ### ('bra', 1, 0) ('##meric', 0, 1) ('##ga', 7, 2) ('410', 5400, 3) ('invasions', 10850, 4) ### ### ('gall', 501, 5) ('##car', 5, 6) ('sue', 9, 7) ('asturias', 117, 8) ('germanic', 31, 9) ### ### ('galicia', 219, 10) ('conquered', 14129, 11) ('portugal', 391, 12) ('ib', 1568, 13) ### ### ('romans', 292, 14) ('provinces', 2756, 15) ('anyway', 174, 16) ('invaded', 26183, 17) ### ### ('roman', 20, 18) ('##cia', 306, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('bra', 1, 0) ('##meric', 0, 1) ('##ga', 7, 2) ('##car', 5, 6) ('sue', 9, 7) ('##bi', 8, 21) ### ### ('germanic', 31, 9) ('roman', 20, 18) ('her', 12, 38) ('¹', 4, 70) ('probably', 24, 34) ### ### ('iberian', 37, 28) ('asturias', 117, 8) ('absently', 3, 86) ('galicia', 219, 10) ('gall', 501, 5) ### ### ('kingdom', 96, 27) ('.', 41, 48) ('intently', 6, 174) ('anyway', 174, 16) ### ############################################################################################################ [2023-11-25 13:04:33,580][root][INFO] - Epoch: 3: Step: 6401/19530, loss[v]=0.098586, lr=0.000013, acc@1[1]=1004.5/1024=0.98095703125, acc@1[2]=1015.5/1024=0.99169921875 [2023-11-25 13:06:22,303][root][INFO] - Train batch 6500 [2023-11-25 13:06:22,304][root][INFO] - Avg. loss per last 100 batches: 0.145356 [2023-11-25 13:06:23,014][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.9/29522=100.00% | mean: 0.03 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29515.0/29522=99.98% | mean: 0.03 | max: 5.56 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: leather jackets can be designed for many purposes, and specific styles have been ### ### associated with subcultures such as greasers, motorcyclists and bikers, mobsters, military aviators ### ### ( especially during and directly after world war ii ) and ### ### [P_TEXT]: heckler ( character ) of the very few people that's in on his heckler alter - ego. ### ### notable for his bad spelling, ledge provides most of stu's information regarding his opponents, ### ### keeping extensive dossiers on all major crime figures in delta city. francois - stu's slightly anal ### ### - retentive and obsessive but highly creative french cook at " eats. " mr. dude - one of the ### ### patrons of " eats ". a leather jacket - clad overweight and middle - aged greaser who provides stu ### ### with information on underworld activities and is regularly visited by the denizens of delta city, ### ### including generals and monsignor s, for advice on a ### ### =============================================== v_q ================================================ ### ### ('grease', 0, 3) ('leather', 1, 19) ('jacket', 2, 11) ('jackets', 3, 476) ('absently', 4, 166) ### ### ('mob', 5, 9972) ('##culture', 6, 13398) ('.', 7, 154) ('biker', 8, 11769) ('intently', 9, 900) ### ### ('貴', 10, 475) ('aviator', 11, 9738) ('military', 12, 78) ('¹', 13, 845) ('motor', 14, 15079) ### ### ('##cl', 15, 28734) ('খ', 16, 216) ('！', 17, 1650) ('##sters', 18, 24652) ('##cy', 19, 18750) ### ### =============================================== v_p ================================================ ### ### ('delta', 13534, 0) ('dude', 888, 1) ('heck', 3760, 2) ('grease', 0, 3) ('eats', 6635, 4) ### ### ('underworld', 2584, 5) ('##weight', 14507, 6) ('mons', 16442, 7) ('stu', 6373, 8) ### ### ('ledge', 21707, 9) ('generals', 4387, 10) ('jacket', 2, 11) ('spelling', 17090, 12) ### ### ('##ign', 23348, 13) ('##ler', 21875, 14) ('clad', 932, 15) ('alter', 1272, 16) ('anal', 17661, 17) ### ### ('francois', 26715, 18) ('leather', 1, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('grease', 0, 3) ('jacket', 2, 11) ('leather', 1, 19) ('worried', 21, 51) ('military', 12, 78) ### ### ('absently', 4, 166) ('anyway', 58, 37) ('.', 7, 154) ('dude', 888, 1) ('jackets', 3, 476) ### ### ('mafia', 63, 60) ('patrons', 72, 63) ('clad', 932, 15) ('খ', 16, 216) ('貴', 10, 475) ### ### ('alter', 1272, 16) ('underworld', 2584, 5) ('activities', 554, 32) ('ம', 34, 138) ### ### ('heck', 3760, 2) ### ############################################################################################################ [2023-11-25 13:06:23,520][root][INFO] - Epoch: 3: Step: 6501/19530, loss[v]=0.166424, lr=0.000013, acc@1[1]=974.0/1024=0.951171875, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-25 13:08:12,648][root][INFO] - Train batch 6600 [2023-11-25 13:08:12,649][root][INFO] - Avg. loss per last 100 batches: 0.148379 [2023-11-25 13:08:13,367][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.83 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29515.8/29522=99.98% | mean: 0.03 | max: 5.62 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: after a hiatus, talking heads hit their commercial peak in 1983 with the u. s. top 10 hit ### ### " burning down the house " from the album " speaking in tongues " and released the concert film " ### ### stop ### ### [P_TEXT]: kjaq 17, 1999, infinity flipped kycw - fm to all - 80's hits, branded as " 96. 5 the ### ### point ", and changed call letters to kypt. “ the point ” saw some initial success but then declined ### ### in the ratings. in 2001, much of the on - air staff was fired, leading to rumors of a format change ### ### to adult contemporary. however, the station continued with its all - 80's format and brought in a ### ### new air staff. despite the changes, ratings continued to slip. on december 19, 2003, at 5 : 00 p. ### ### m., after playing " burning down the house " by talking heads, kypt flipped ### ### =============================================== v_q ================================================ ### ### ('tongues', 0, 9700) ('stop', 1, 3460) ('talking', 2, 3) ('heads', 3, 8) ('burning', 4, 0) ### ### ('speaking', 5, 183) ('absently', 6, 92) ('house', 7, 4) ('intently', 8, 424) ('¹', 9, 189) ### ### ('1983', 10, 5507) ('貴', 11, 284) ('down', 12, 15) ('খ', 13, 120) ('head', 14, 1132) ### ### ('film', 15, 4846) ('spoke', 16, 5848) ('³', 17, 1057) ('brains', 18, 3489) ('.', 19, 5964) ### ### =============================================== v_p ================================================ ### ### ('burning', 4, 0) ('infinity', 4275, 1) ('##ja', 25204, 2) ('talking', 2, 3) ('house', 7, 4) ### ### ('ratings', 5350, 5) ('##q', 7487, 6) ('point', 2393, 7) ('heads', 3, 8) ('formats', 2950, 9) ### ### ('ky', 14696, 10) ('burned', 3402, 11) ('##pt', 24761, 12) ('format', 1542, 13) ('80', 4232, 14) ### ### ('down', 12, 15) ('lyrical', 91, 16) ('rumors', 3420, 17) ('96', 17275, 18) ('branding', 9357, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('burning', 4, 0) ('talking', 2, 3) ('heads', 3, 8) ('house', 7, 4) ('down', 12, 15) ### ### ('absently', 6, 92) ('speaking', 5, 183) ('lyrical', 91, 16) ('¹', 9, 189) ('খ', 13, 120) ### ### ('song', 39, 40) ('worried', 58, 38) ('anyway', 170, 26) ('貴', 11, 284) ('album', 50, 46) ### ### ('talk', 26, 73) ('ம', 24, 75) ('intently', 8, 424) ('″', 195, 32) ('band', 22, 124) ### ############################################################################################################ [2023-11-25 13:08:13,874][root][INFO] - Epoch: 3: Step: 6601/19530, loss[v]=0.185777, lr=0.000013, acc@1[1]=975.0/1024=0.9521484375, acc@1[2]=1001.0/1024=0.9775390625 [2023-11-25 13:10:02,960][root][INFO] - Train batch 6700 [2023-11-25 13:10:02,962][root][INFO] - Avg. loss per last 100 batches: 0.137087 [2023-11-25 13:10:03,672][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29515.0/29522=99.98% | mean: 0.03 | max: 5.62 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 2009, square enix holdings acquired british game publisher eidos interactive, which ### ### was absorbed into square enix europe to publish titles such as " tomb raider ", " deus ex " and " ### ### just cause ". ### ### [P_TEXT]: warner bros. interactive entertainment warner acquired a total of 10 million shares of ### ### the company, raising its owned amount to 19. 92 %, after an agreement which prevented time warner ### ### from acquiring more shares was scrapped one month earlier. on january 28, 2009, the hollywood ### ### reporter reported the deal also gave warner the rights of the " tomb raider " film series, ### ### previously owned by paramount pictures. on february 12, 2009, sqex ltd, a wholly owned subsidiary ### ### of square enix, announced a takeover offer worth ( 32p per share ) for eidos plc, which was backed ### ### by warner bros. ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 69) ('raider', 1, 1) ('##ix', 2, 11) ('en', 3, 79) ('tomb', 4, 4) ('##ido', 5, 2) ### ### ('ex', 6, 14969) ('cause', 7, 18247) ('interactive', 8, 0) ('¹', 9, 213) ('##us', 10, 23447) ### ### ('intently', 11, 313) ('just', 12, 18219) ('貴', 13, 253) ('game', 14, 791) ('europe', 15, 5673) ### ### ('square', 16, 32) ('de', 17, 26706) ('e', 18, 23) ('খ', 19, 144) ### ### =============================================== v_p ================================================ ### ### ('interactive', 8, 0) ('raider', 1, 1) ('##ido', 5, 2) ('warner', 1619, 3) ('tomb', 4, 4) ### ### ('sq', 1683, 5) ('reporter', 11075, 6) ('paramount', 5307, 7) ('worth', 22319, 8) ('plc', 3713, 9) ### ### ('entertainment', 52, 10) ('##ix', 2, 11) ('##ex', 12242, 12) ('offer', 1061, 13) ### ### ('subsidiaries', 15520, 14) ('″', 53, 15) ('shareholder', 10670, 16) ('2009', 143, 17) ### ### ('film', 853, 18) ('takeover', 411, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('raider', 1, 1) ('interactive', 8, 0) ('##ido', 5, 2) ('tomb', 4, 4) ('##ix', 2, 11) ### ### ('square', 16, 32) ('absently', 0, 69) ('e', 18, 23) ('en', 3, 79) ('entertainment', 52, 10) ### ### ('″', 53, 15) ('raiders', 22, 43) ('series', 40, 33) ('company', 72, 24) ('anyway', 67, 28) ### ### ('¹', 9, 213) ('2009', 143, 17) ('貴', 13, 253) ('ம', 32, 77) ('##ப', 21, 124) ### ############################################################################################################ [2023-11-25 13:10:04,176][root][INFO] - Epoch: 3: Step: 6701/19530, loss[v]=0.209416, lr=0.000013, acc@1[1]=965.0/1024=0.9423828125, acc@1[2]=990.0/1024=0.966796875 [2023-11-25 13:11:53,374][root][INFO] - Train batch 6800 [2023-11-25 13:11:53,375][root][INFO] - Avg. loss per last 100 batches: 0.147036 [2023-11-25 13:11:54,095][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29512.7/29522=99.97% | mean: 0.02 | max: 5.57 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: tomb kv15, located in the valley of the kings in egypt, was used for the burial of ### ### pharaoh seti ii of the nineteenth dynasty. ### ### [P_TEXT]: twosret iii described the late 19th dynasty as a time of chaos. setnakhte usurped the ### ### joint kv14 tomb of seti ii and twosret but reburied seti ii in tomb kv15, while deliberately ### ### replastering and redrawing all images of twosret in tomb kv14 with those of himself. setnakhte's ### ### decisions here may demonstrate his dislike and presumably hatred for twosret since he chose to ### ### reinter seti ii but not twosret. setnakhte's son, ramesses iii, later excluded twosret and even ### ### siptah of the 19th dynasty from his medinet habu list of egyptian kings thereby delegitimizing them ### ### in ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 82) ('##15', 1, 18) ('nineteenth', 2, 38) ('tomb', 3, 6) ('absently', 4, 195) ### ### ('kv', 5, 16) ('egypt', 6, 35) ('intently', 7, 318) ('set', 8, 2) ('valley', 9, 17545) ### ### ('貴', 10, 305) ('kings', 11, 21) ('³', 12, 809) ('ii', 13, 33) ('##i', 14, 29) ('খ', 15, 181) ### ### ('pharaoh', 16, 484) ('burial', 17, 128) ('19th', 18, 8) ('##い', 19, 67) ### ### =============================================== v_p ================================================ ### ### ('##sr', 15510, 0) ('sip', 24845, 1) ('set', 8, 2) ('##tering', 5646, 3) ('##bu', 21874, 4) ### ### ('##et', 9796, 5) ('tomb', 3, 6) ('##14', 73, 7) ('19th', 18, 8) ('med', 13738, 9) ### ### ('chaos', 8718, 10) ('images', 2749, 11) ('dynasty', 32, 12) ('hatred', 14798, 13) ### ### ('egyptian', 84, 14) ('rein', 25510, 15) ('kv', 5, 16) ('two', 14681, 17) ('##15', 1, 18) ### ### ('ha', 19037, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('set', 8, 2) ('tomb', 3, 6) ('##15', 1, 18) ('kv', 5, 16) ('nineteenth', 2, 38) ('egypt', 6, 35) ### ### ('kings', 11, 21) ('19th', 18, 8) ('##i', 14, 29) ('ii', 13, 33) ('¹', 0, 82) ('dynasty', 32, 12) ### ### ('##14', 73, 7) ('egyptian', 84, 14) ('tombs', 115, 25) ('##い', 19, 67) ('absently', 4, 195) ### ### ('anyway', 188, 20) ('ம', 23, 85) ('intently', 7, 318) ### ############################################################################################################ [2023-11-25 13:11:54,602][root][INFO] - Epoch: 3: Step: 6801/19530, loss[v]=0.177829, lr=0.000013, acc@1[1]=972.0/1024=0.94921875, acc@1[2]=998.0/1024=0.974609375 [2023-11-25 13:13:44,241][root][INFO] - Train batch 6900 [2023-11-25 13:13:44,242][root][INFO] - Avg. loss per last 100 batches: 0.140175 [2023-11-25 13:13:44,980][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.0/29522=99.97% | mean: 0.03 | max: 5.77 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the homeland of transkei served in many regards as a " testing ground " for apartheid ### ### policies ; its institutional development started already before the 1959 act, and its attainment of ### ### self - government and independence were therefore implemented earlier than for the other homelands. ### ### [P_TEXT]: transkei within the tbvc - system. chief kaiser daliwonga matanzima was transkei's prime ### ### minister until 1979, when he assumed the office of president, a position he held until 1986. south ### ### african prime minister b. j. vorster justified the declaration of transkei as an independent ### ### republic by referring to " the right of every people to have full control over its own affairs " ### ### and wished " transkei and its leaders god's richest blessings on the road ahead. " a press release ### ### by the african national congress at the time rejected the transkei's independence and condemned it ### ### as " designed ### ### =============================================== v_q ================================================ ### ### ('##kei', 0, 0) ('homeland', 1, 6539) ('apartheid', 2, 215) ('1959', 3, 983) ('trans', 4, 1) ### ### ('absently', 5, 81) ('¹', 6, 192) ('intently', 7, 227) ('貴', 8, 223) ('.', 9, 197) ### ### ('africa', 10, 21) ('খ', 11, 118) ('act', 12, 12478) ('testing', 13, 4311) ('independence', 14, 38) ### ### ('kenya', 15, 662) ('rhodesia', 16, 906) ('pretoria', 17, 379) ('self', 18, 24671) ### ### ('african', 19, 6) ### ### =============================================== v_p ================================================ ### ### ('##kei', 0, 0) ('trans', 4, 1) ('##rst', 12912, 2) ('kaiser', 4925, 3) ('dali', 12810, 4) ### ### ('##vc', 23707, 5) ('african', 19, 6) ('tb', 17979, 7) ('littered', 1420, 8) ### ### ('blessings', 19663, 9) ('##won', 13545, 10) ('congress', 3436, 11) ('##ima', 1574, 12) ### ### ('worried', 88, 13) ('god', 9239, 14) ('##nz', 9692, 15) ('mata', 425, 16) ### ### ('declaration', 1371, 17) ('vo', 7528, 18) ('prime', 1610, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##kei', 0, 0) ('trans', 4, 1) ('african', 19, 6) ('africa', 10, 21) ('independence', 14, 38) ### ### ('absently', 5, 81) ('worried', 88, 13) ('apartheid', 2, 215) ('খ', 11, 118) ('¹', 6, 192) ### ### ('ம', 26, 60) ('.', 9, 197) ('intently', 7, 227) ('貴', 8, 223) ('government', 35, 48) ### ### ('anyway', 225, 20) ('johannesburg', 28, 67) ('mata', 425, 16) ('zimbabwe', 32, 75) ### ### ('republic', 144, 30) ### ############################################################################################################ [2023-11-25 13:13:45,485][root][INFO] - Epoch: 3: Step: 6901/19530, loss[v]=0.090899, lr=0.000013, acc@1[1]=1000.0/1024=0.9765625, acc@1[2]=1015.0/1024=0.9912109375 [2023-11-25 13:15:34,659][root][INFO] - Train batch 7000 [2023-11-25 13:15:34,660][root][INFO] - Avg. loss per last 100 batches: 0.150584 [2023-11-25 13:15:35,374][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.78 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.4/29522=99.97% | mean: 0.03 | max: 5.73 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: out of the furnace is a 2013 american crime drama film directed by scott cooper, from a ### ### screenplay written by cooper and brad ingelsby. ### ### [P_TEXT]: willem dafoe anton yelchin as the titular character that possesses supernatural powers to ### ### see the dead. using motion - capture acting technology, dafoe co - starred alongside elliot page in ### ### david cage's video game " " ( 2013 ) as a paranormal activity researcher who acts as the surrogate ### ### - father - figure to a girl who possesses supernatural powers. the game polarized reviewers, ### ### although dafoe and page's performance were widely praised. in scott cooper's " out of the furnace " ### ### ( 2013 ), starring christian bale, dafoe played the supporting role of a ### ### =============================================== v_q ================================================ ### ### ('furnace', 0, 0) ('cooper', 1, 5) ('out', 2, 18) ('scott', 3, 11) ('″', 4, 49) ('2013', 5, 13) ### ### ('¹', 6, 228) ('brad', 7, 15861) ('film', 8, 73) ('##by', 9, 26301) ('crime', 10, 14885) ### ### ('³', 11, 1845) ('ing', 12, 23975) ('kill', 13, 3098) ('2014', 14, 124) ('prison', 15, 1661) ### ### ('"', 16, 78) ('movie', 17, 154) ('##els', 18, 28028) ('murder', 19, 834) ### ### =============================================== v_p ================================================ ### ### ('furnace', 0, 0) ('page', 7607, 1) ('willem', 18238, 2) ('cage', 1343, 3) ('elliot', 1594, 4) ### ### ('cooper', 1, 5) ('##fo', 26412, 6) ('##lch', 24902, 7) ('bal', 23617, 8) ('##rogate', 676, 9) ### ### ('christian', 7076, 10) ('scott', 3, 11) ('anton', 13614, 12) ('2013', 5, 13) ('motion', 6771, 14) ### ### ('game', 1152, 15) ('capture', 1101, 16) ('paranormal', 110, 17) ('out', 2, 18) ('da', 17995, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('furnace', 0, 0) ('cooper', 1, 5) ('scott', 3, 11) ('out', 2, 18) ('2013', 5, 13) ('″', 4, 49) ### ### ('film', 8, 73) ('paranormal', 110, 17) ('¹', 6, 228) ('dead', 67, 24) ('filming', 72, 31) ### ### ('2012', 22, 63) ('"', 16, 78) ('cage', 1343, 3) ('##rogate', 676, 9) ('revealed', 29, 68) ### ### ('elliot', 1594, 4) ('2014', 14, 124) ('filmed', 98, 44) ('video', 364, 28) ### ############################################################################################################ [2023-11-25 13:15:35,877][root][INFO] - Epoch: 3: Step: 7001/19530, loss[v]=0.180384, lr=0.000013, acc@1[1]=979.5/1024=0.95654296875, acc@1[2]=1003.5/1024=0.97998046875 [2023-11-25 13:17:25,195][root][INFO] - Train batch 7100 [2023-11-25 13:17:25,196][root][INFO] - Avg. loss per last 100 batches: 0.154085 [2023-11-25 13:17:25,920][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.2/29522=99.97% | mean: 0.03 | max: 5.62 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: westerville was once known as " the dry capital of the world " for its strict laws ### ### prohibiting sales of alcohol and for being the home of the anti - saloon league, one of the driving ### ### forces behind prohibition at the beginning of the 20th century. ### ### [P_TEXT]: anti - saloon league young clarke and mary elizabeth tyler of the southern publicity ### ### association. in 1909, the league moved its national headquarters from washington to westerville, ### ### ohio, which had a reputation for supporting temperance. the american issue publishing house, the ### ### publishing arm of the league, was also in westerville. ernest cherrington headed the company. it ### ### printed so many leaflets — over 40 tons of mail per month — that westerville was the smallest town ### ### to have a first class post office. from 1948 until 1950 the group was named ### ### =============================================== v_q ================================================ ### ### ('saloon', 0, 1) ('##ville', 1, 11) ('west', 2, 20) ('##er', 3, 15) ('prohibition', 4, 195) ### ### ('¹', 5, 279) ('anti', 6, 7) ('intently', 7, 222) ('league', 8, 29) ('³', 9, 3255) ('貴', 10, 258) ### ### ('absently', 11, 65) ('খ', 12, 138) ('capital', 13, 862) ('alcohol', 14, 574) ### ### ('notorious', 15, 4540) ('kentucky', 16, 1020) ('dry', 17, 10936) ('liquor', 18, 2514) ### ### ('indiana', 19, 580) ### ### =============================================== v_p ================================================ ### ### ('temperance', 34, 0) ('saloon', 0, 1) ('publicity', 1772, 2) ('tyler', 2287, 3) ### ### ('##rington', 21393, 4) ('leaflets', 12498, 5) ('issue', 3805, 6) ('anti', 6, 7) ### ### ('washington', 2883, 8) ('clarke', 17634, 9) ('ernest', 3139, 10) ('##ville', 1, 11) ### ### ('publishing', 10331, 12) ('cher', 23424, 13) ('elizabeth', 5333, 14) ('##er', 3, 15) ### ### ('worried', 499, 16) ('headquarters', 1934, 17) ('anyway', 102, 18) ('littered', 1858, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('saloon', 0, 1) ('anti', 6, 7) ('##ville', 1, 11) ('##er', 3, 15) ('west', 2, 20) ### ### ('temperance', 34, 0) ('league', 8, 29) ('absently', 11, 65) ('anyway', 102, 18) ### ### ('prohibition', 4, 195) ('ohio', 106, 27) ('¹', 5, 279) ('ம', 23, 74) ('cradled', 76, 36) ### ### ('intently', 7, 222) ('খ', 12, 138) ('publicity', 1772, 2) ('worried', 499, 16) ('tyler', 2287, 3) ### ### ('貴', 10, 258) ### ############################################################################################################ [2023-11-25 13:17:26,426][root][INFO] - Epoch: 3: Step: 7101/19530, loss[v]=0.077400, lr=0.000013, acc@1[1]=998.0/1024=0.974609375, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 13:19:15,583][root][INFO] - Train batch 7200 [2023-11-25 13:19:15,584][root][INFO] - Avg. loss per last 100 batches: 0.152019 [2023-11-25 13:19:16,285][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.78 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29512.0/29522=99.97% | mean: 0.02 | max: 5.62 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: all of sr 533 was re - signed as " old sr 533 " and cr 5 ( with which sr 533 was already ### ### entirely concurrent ) while cr 144 and cr 144a in the southeastern part of the county became state ### ### - maintained as an extension of sr 347. ### ### [P_TEXT]: ohio state route 347 sr 347 was extended about over logan county routes 144a and 144 to ### ### sr 287 as a part of a maintenance swap with logan county. as a part of the swap, the ohio ### ### department of transportation gave up jurisdiction of sr 533 to the county. ### ### =============================================== v_q ================================================ ### ### ('144', 0, 2) ('absently', 1, 38) ('34', 2, 4) ('jersey', 3, 17534) ('intently', 4, 238) ### ### ('5', 5, 12465) ('¹', 6, 146) ('route', 7, 12) ('貴', 8, 239) ('maine', 9, 5742) ('53', 10, 5) ### ### ('##3', 11, 22) ('খ', 12, 109) ('##7', 13, 7) ('county', 14, 13) ('.', 15, 1610) ('##ப', 16, 118) ### ### ('³', 17, 1045) ('幸', 18, 126) ('tennessee', 19, 2084) ### ### =============================================== v_p ================================================ ### ### ('287', 12507, 0) ('logan', 5606, 1) ('144', 0, 2) ('swap', 21279, 3) ('34', 2, 4) ('53', 10, 5) ### ### ('maintenance', 491, 6) ('##7', 13, 7) ('ohio', 22, 8) ('sr', 25, 9) ('department', 185, 10) ### ### ('transportation', 976, 11) ('route', 7, 12) ('county', 14, 13) ('anyway', 65, 14) ### ### ('routes', 47, 15) ('counties', 56, 16) ('146', 80, 17) ('145', 76, 18) ('worried', 39, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('144', 0, 2) ('34', 2, 4) ('53', 10, 5) ('##7', 13, 7) ('route', 7, 12) ('county', 14, 13) ### ### ('ohio', 22, 8) ('##3', 11, 22) ('sr', 25, 9) ('absently', 1, 38) ('routes', 47, 15) ### ### ('worried', 39, 19) ('counties', 56, 16) ('anyway', 65, 14) ('driveway', 43, 21) ('146', 80, 17) ### ### ('145', 76, 18) ('department', 185, 10) ('¹', 6, 146) ('maintenance', 491, 6) ### ############################################################################################################ [2023-11-25 13:19:16,789][root][INFO] - Epoch: 3: Step: 7201/19530, loss[v]=0.075564, lr=0.000013, acc@1[1]=1000.0/1024=0.9765625, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 13:21:05,257][root][INFO] - Train batch 7300 [2023-11-25 13:21:05,258][root][INFO] - Avg. loss per last 100 batches: 0.145136 [2023-11-25 13:21:05,964][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.75 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29515.0/29522=99.98% | mean: 0.03 | max: 5.66 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.5/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: common dreads is the second studio album by english rock band enter shikari, released on ### ### 15 june 2009 and 16 june in the us. ### ### [P_TEXT]: fanfare for the common man the elp version in its opening title. in the 1990s, " fanfare ### ### " began to be used to welcome the winner of the aintree grand national steeplechase from the ### ### racecourse to the winner's enclosure as the timing of the piece roughly matched the time it took ### ### the winner to make the journey. when the enclosure was moved in 2010, " fanfare " was used instead ### ### to announce the procession of competitors from the paddock to the course before the race. the ### ### british band enter shikari included a song titled " fanfare for the conscious man " on their album ### ### " common dreads ". the song is a criticism of british involvement in the iraq war. a version ### ### =============================================== v_q ================================================ ### ### ('common', 0, 10) ('shi', 1, 19) ('¹', 2, 397) ('##kari', 3, 2) ('dread', 4, 3) ('enter', 5, 7) ### ### ('2009', 6, 394) ('³', 7, 3012) ('″', 8, 54) ('album', 9, 34) ('貴', 10, 469) ### ### ('metacritic', 11, 17925) ('খ', 12, 235) ('band', 13, 51) ('intently', 14, 588) ('2008', 15, 3143) ### ### ('lyrical', 16, 32) ('幸', 17, 343) ('2010', 18, 24) ('absently', 19, 104) ### ### =============================================== v_p ================================================ ### ### ('##fare', 19160, 0) ('conscious', 8479, 1) ('##kari', 3, 2) ('dread', 4, 3) ('ain', 6375, 4) ### ### ('iraq', 4051, 5) ('enclosure', 20243, 6) ('enter', 5, 7) ('pad', 14429, 8) ('fan', 2018, 9) ### ### ('common', 0, 10) ('journey', 2043, 11) ('furlongs', 28042, 12) ('timing', 10324, 13) ### ### ('##tree', 4273, 14) ('steep', 8259, 15) ('piece', 723, 16) ('el', 12421, 17) ('##p', 21134, 18) ### ### ('shi', 1, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##kari', 3, 2) ('dread', 4, 3) ('common', 0, 10) ('enter', 5, 7) ('shi', 1, 19) ('album', 9, 34) ### ### ('″', 8, 54) ('2010', 18, 24) ('lyrical', 16, 32) ('band', 13, 51) ('¹', 2, 397) ('bands', 35, 50) ### ### ('2009', 6, 394) ('ம', 21, 97) ('absently', 19, 104) ('albums', 45, 61) ('uk', 44, 62) ### ### ('anyway', 295, 22) ('hissed', 24, 101) ('খ', 12, 235) ### ############################################################################################################ [2023-11-25 13:21:06,469][root][INFO] - Epoch: 3: Step: 7301/19530, loss[v]=0.230952, lr=0.000013, acc@1[1]=974.5/1024=0.95166015625, acc@1[2]=1002.5/1024=0.97900390625 [2023-11-25 13:22:55,348][root][INFO] - Train batch 7400 [2023-11-25 13:22:55,349][root][INFO] - Avg. loss per last 100 batches: 0.148465 [2023-11-25 13:22:56,069][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29513.1/29522=99.97% | mean: 0.03 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: lee falk ( ), born leon harrison gross ( ; april 28, 1911 – march 13, 1999 ), was an ### ### american writer, theater director and producer, best known as the creator of the popular comic ### ### strip s " mandrake the magician " ( 1934 – 2013 ) and " the phantom " ( 1936 – present ). ### ### [P_TEXT]: the phantom the phantom is an american adventure comic strip, first published by lee falk ### ### in february 1936. the main character, the phantom, is a fictional costumed crime - fighter who ### ### operates from the fictional african country of bangalla. the character has been adapted for ### ### television, film and video games. the series began with a daily newspaper strip on february 17, ### ### 1936, followed by a color sunday strip on may 28, 1939 ; both are still running as of. in 1966, ### ### king features stated that " the phantom " ### ### =============================================== v_q ================================================ ### ### ('##lk', 0, 1) ('phantom', 1, 0) ('fa', 2, 14) ('lee', 3, 11) ('##dra', 4, 28750) ('¹', 5, 744) ### ### ('magician', 6, 6395) ('exclaimed', 7, 203) ('strip', 8, 9) ('##½', 9, 930) ('absently', 10, 115) ### ### ('intently', 11, 1044) ('³', 12, 10792) ('খ', 13, 297) ('leon', 14, 20444) ('##ke', 15, 17834) ### ### ('1936', 16, 19) ('silent', 17, 17330) ('貴', 18, 522) ('amusing', 19, 521) ### ### =============================================== v_p ================================================ ### ### ('phantom', 1, 0) ('##lk', 0, 1) ('##alla', 23660, 2) ('bang', 9271, 3) ('color', 2325, 4) ### ### ('strips', 1211, 5) ('sunday', 10691, 6) ('leaned', 4162, 7) ('features', 7530, 8) ('strip', 8, 9) ### ### ('african', 11814, 10) ('lee', 3, 11) ('king', 9918, 12) ('worried', 3550, 13) ('fa', 2, 14) ### ### ('costume', 1193, 15) ('adventure', 2506, 16) ('costumes', 2868, 17) ('newspaper', 5344, 18) ### ### ('1936', 16, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('phantom', 1, 0) ('##lk', 0, 1) ('lee', 3, 11) ('fa', 2, 14) ('strip', 8, 9) ('1936', 16, 19) ### ### ('comic', 26, 26) ('absently', 10, 115) ('1939', 142, 25) ('comics', 97, 27) ('americans', 102, 30) ### ### ('exclaimed', 7, 203) ('¹', 5, 744) ('strips', 1211, 5) ('cartoon', 59, 53) ('cartoons', 76, 51) ### ### ('supposedly', 116, 49) ('1930s', 31, 78) ('color', 2325, 4) ('ம', 28, 99) ### ############################################################################################################ [2023-11-25 13:22:56,576][root][INFO] - Epoch: 3: Step: 7401/19530, loss[v]=0.096345, lr=0.000013, acc@1[1]=993.0/1024=0.9697265625, acc@1[2]=1011.5/1024=0.98779296875 [2023-11-25 13:24:45,703][root][INFO] - Train batch 7500 [2023-11-25 13:24:45,704][root][INFO] - Avg. loss per last 100 batches: 0.145441 [2023-11-25 13:24:46,417][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29513.6/29522=99.97% | mean: 0.03 | max: 5.70 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: panathinaikos used to hold a minor rivalry with aris, mostly during the 1980s, when ### ### panathinaikos and aris were the two biggest dynasties in greek basketball at the time. ### ### [P_TEXT]: aris b. c. first - class partner. the result was total carnage for other teams. the ### ### lethal back court blazed through the greek league for seven consecutive years, with the help of ### ### players such as nikos filippou, michalis romanidis, lefteris subotic, georgios doxakis, vassilis ### ### lipiridis, and others, winning 7 greek league championships, and 5 greek cups ( with one greek cup ### ### being memorably lost to panathinaikos in 1986, with galis performing surprisingly poorly in a ### ### single - elimination game in athens ). especially between 1985 and 1988, the question was not who ### ### =============================================== v_q ================================================ ### ### ('ari', 0, 0) ('absently', 1, 250) ('basketball', 2, 275) ('##athi', 3, 43) ('貴', 4, 237) ### ### ('¹', 5, 103) ('intently', 6, 178) ('.', 7, 163) ('##nai', 8, 26) ('খ', 9, 235) ('greek', 10, 6) ### ### ('##kos', 11, 115) ('pan', 12, 91) ('##s', 13, 51) ('greece', 14, 47) ('athens', 15, 11) ### ### ('³', 16, 1221) ('幸', 17, 173) ('rebounds', 18, 2808) ('##ப', 19, 254) ### ### =============================================== v_p ================================================ ### ### ('ari', 0, 0) ('filippo', 3134, 1) ('carnage', 7980, 2) ('lip', 5811, 3) ('littered', 1231, 4) ### ### ('anyway', 28, 5) ('greek', 10, 6) ('gal', 22895, 7) ('question', 8009, 8) ('lethal', 10585, 9) ### ### ('b', 17221, 10) ('athens', 15, 11) ('memo', 25818, 12) ('##hal', 13313, 13) ('cups', 9196, 14) ### ### ('1986', 259, 15) ('romani', 2358, 16) ('mic', 18841, 17) ('georg', 4063, 18) ('##xa', 9461, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ari', 0, 0) ('greek', 10, 6) ('athens', 15, 11) ('##nai', 8, 26) ('##athi', 3, 43) ### ### ('anyway', 28, 5) ('greece', 14, 47) ('##s', 13, 51) ('¹', 5, 103) ('league', 53, 23) ### ### ('pan', 12, 91) ('##kos', 11, 115) ('intently', 6, 178) ('.', 7, 163) ('absently', 1, 250) ### ### ('worried', 109, 22) ('basketball', 2, 275) ('##ս', 27, 63) ('貴', 4, 237) ('খ', 9, 235) ### ############################################################################################################ [2023-11-25 13:24:46,920][root][INFO] - Epoch: 3: Step: 7501/19530, loss[v]=0.113528, lr=0.000013, acc@1[1]=994.5/1024=0.97119140625, acc@1[2]=1008.5/1024=0.98486328125 [2023-11-25 13:26:35,850][root][INFO] - Train batch 7600 [2023-11-25 13:26:35,851][root][INFO] - Avg. loss per last 100 batches: 0.144365 [2023-11-25 13:26:36,564][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.74 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.7/29522=99.98% | mean: 0.03 | max: 5.78 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he became leader of the official opposition saskatchewan party on july 15, 2004, ### ### replacing elwin hermanson, and ### ### [P_TEXT]: sasktel network branded as infinet ( stylized infinet ), beginning in portions of regina ### ### and saskatoon, and other cities over the next seven years. in january 2013, sasktel announced the ### ### launch of an lte network in the regina and saskatoon areas, with plans to extend coverage into ### ### other major areas of the province by 2014. as of 2013, the company had recorded nearly 616, 000 ### ### wireless subscribers and over 100, 000 max tv subscribers. in july 2015, sasktel acquired six aws - ### ### 1 wireless spectrum licenses from freedom mobile. in 2016, brad wall's saskatchewan party ### ### government proposed ### ### =============================================== v_q ================================================ ### ### ('saskatchewan', 0, 2) ('official', 1, 8645) ('¹', 2, 206) ('littered', 3, 16) ('herman', 4, 23572) ### ### ('##son', 5, 20255) ('leader', 6, 1497) ('party', 7, 11) ('pmid', 8, 3578) ('2004', 9, 7322) ### ### ('angrily', 10, 397) ('##win', 11, 27267) ('saskatoon', 12, 9) ('el', 13, 22724) ### ### ('intently', 14, 259) ('bellowed', 15, 591) ('.', 16, 805) ('whoever', 17, 2407) ('貴', 18, 243) ### ### ('！', 19, 1725) ### ### =============================================== v_p ================================================ ### ### ('sas', 26252, 0) ('##fine', 20987, 1) ('saskatchewan', 0, 2) ('##kt', 17144, 3) ('lt', 2484, 4) ### ### ('brad', 271, 5) ('wall', 6920, 6) ('aw', 7280, 7) ('freedom', 16425, 8) ('saskatoon', 12, 9) ### ### ('mobile', 19784, 10) ('party', 7, 11) ('max', 3785, 12) ('anyway', 164, 13) ('regina', 133, 14) ### ### ('wireless', 11794, 15) ('littered', 3, 16) ('worried', 2054, 17) ('##el', 8960, 18) ### ### ('spectrum', 16893, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('saskatchewan', 0, 2) ('littered', 3, 16) ('party', 7, 11) ('saskatoon', 12, 9) ('brad', 271, 5) ### ### ('regina', 133, 14) ('anyway', 164, 13) ('¹', 2, 206) ('alberta', 29, 86) ('intently', 14, 259) ### ### ('lt', 2484, 4) ('canada', 51, 68) ('opposition', 23, 151) ('provincial', 45, 81) ('ம', 38, 89) ### ### ('politicians', 20, 186) ('angrily', 10, 397) ('agenda', 116, 50) ('貴', 18, 243) ### ### ('government', 318, 27) ### ############################################################################################################ [2023-11-25 13:26:37,070][root][INFO] - Epoch: 3: Step: 7601/19530, loss[v]=0.198267, lr=0.000013, acc@1[1]=974.0/1024=0.951171875, acc@1[2]=1001.5/1024=0.97802734375 [2023-11-25 13:28:26,288][root][INFO] - Train batch 7700 [2023-11-25 13:28:26,289][root][INFO] - Avg. loss per last 100 batches: 0.140607 [2023-11-25 13:28:27,058][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29513.5/29522=99.97% | mean: 0.02 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the new jersey turnpike ( interstate 95 ) passes outside the township in both neighboring ### ### south brunswick and new brunswick, but the closest interchanges are two towns away in edison ( exit ### ### 10 ), east brunswick township ( exit 9 ) and monroe township ( exit 8a ). ### ### [P_TEXT]: north brunswick, new jersey and the netherlands acquired land in 1761 from the lenape ### ### that would become north brunswick. north brunswick was first mentioned in middlesex freeholder ### ### board minutes of february 28, 1779. north brunswick township was incorporated on february 21, 1798, ### ### by the new jersey legislature's township act of 1798 as one of the state's initial group of 104 ### ### townships. portions of the township have since separated to create east brunswick ( february 28, ### ### 1860 ) and milltown ( january 29, 1889 ). territorial exchanges were made with franklin township, ### ### somerset county, new jersey in 1850 and 1858, with ### ### =============================================== v_q ================================================ ### ### ('jersey', 0, 12) ('brunswick', 1, 0) ('monroe', 2, 667) ('township', 3, 2) ('¹', 4, 111) ### ### ('absently', 5, 124) ('edison', 6, 17456) ('95', 7, 17228) ('pennsylvania', 8, 159) ### ### ('south', 9, 3367) ('intently', 10, 172) ('.', 11, 120) ('貴', 12, 168) ('turnpike', 13, 6983) ### ### ('খ', 14, 234) ('cradled', 15, 253) ('rutgers', 16, 226) ('ம', 17, 73) ('幸', 18, 200) ### ### ('neighboring', 19, 52) ### ### =============================================== v_p ================================================ ### ### ('brunswick', 1, 0) ('north', 5024, 1) ('township', 3, 2) ('worried', 763, 3) ('franklin', 914, 4) ### ### ('##holder', 2616, 5) ('middlesex', 950, 6) ('mill', 3477, 7) ('townships', 114, 8) ### ### ('104', 8931, 9) ('1798', 6103, 10) ('somerset', 72, 11) ('jersey', 0, 12) ### ### ('legislature', 6063, 13) ('netherlands', 4352, 14) ('anyway', 676, 15) ('lena', 853, 16) ### ### ('1779', 17419, 17) ('##town', 527, 18) ('act', 6648, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('brunswick', 1, 0) ('jersey', 0, 12) ('township', 3, 2) ('east', 22, 25) ('somerset', 72, 11) ### ### ('townships', 114, 8) ('¹', 4, 111) ('leaned', 28, 30) ('absently', 5, 124) ('neighboring', 19, 52) ### ### ('newark', 35, 33) ('portions', 24, 53) ('.', 11, 120) ('ம', 17, 73) ('worried', 763, 3) ### ### ('pennsylvania', 8, 159) ('intently', 10, 172) ('franklin', 914, 4) ('middlesex', 950, 6) ### ### ('貴', 12, 168) ### ############################################################################################################ [2023-11-25 13:28:27,567][root][INFO] - Epoch: 3: Step: 7701/19530, loss[v]=0.160920, lr=0.000013, acc@1[1]=979.5/1024=0.95654296875, acc@1[2]=1008.0/1024=0.984375 [2023-11-25 13:30:16,694][root][INFO] - Train batch 7800 [2023-11-25 13:30:16,696][root][INFO] - Avg. loss per last 100 batches: 0.154151 [2023-11-25 13:30:17,404][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29513.0/29522=99.97% | mean: 0.02 | max: 5.64 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: among the main reservoirs ( lakes ) in central java are gunung rowo lake ( pati regency ### ### ), gajahmungkur reservoir ( wonogiri regency ), kedungombo reservoir ( boyolali and sragen regency ### ### ), rawa pening lake ( semarang regency ), cacaban reservoir ( tegal regency ), malahayu reservoir ( ### ### brebes regency ), wadaslintang reservoir ( border of kebumen regency and wonosobo regency ), ### ### gembong reservoir ( pati regency ), ### ### [P_TEXT]: tegal regency tegal regency is one of the regencies in the northwest part of central java ### ### province of indonesia, with an area of. the administrative center used to be in tegal city, in the ### ### northwest corner of the regency, but then tegal city was administratively separated from the ### ### regency and formed into its own territory. the city was later replaced as the administrative center ### ### of tegal regency by slawi town, which is a suburb about to the south of the city and within the ### ### district boundary. the name of tegal comes from the word ### ### =============================================== v_q ================================================ ### ### ('java', 0, 6) ('absently', 1, 64) ('regency', 2, 1) ('reservoir', 3, 7656) ('pat', 4, 6325) ### ### ('leaned', 5, 12) ('##og', 6, 28337) ('.', 7, 2116) ('indonesia', 8, 10) ('gun', 9, 11658) ### ### ('貴', 10, 74) ('indonesian', 11, 13) ('central', 12, 21) ('cradled', 13, 89) ('lake', 14, 15694) ### ### ('খ', 15, 131) ('intently', 16, 182) ('##age', 17, 24088) ('endemic', 18, 2635) ('boy', 19, 15701) ### ### =============================================== v_p ================================================ ### ### ('##gal', 20, 0) ('regency', 2, 1) ('##encies', 19448, 2) ('te', 50, 3) ('##aw', 19387, 4) ### ### ('worried', 48, 5) ('java', 0, 6) ('sl', 9725, 7) ('reg', 24987, 8) ('boundary', 6597, 9) ### ### ('indonesia', 8, 10) ('provinces', 2572, 11) ('leaned', 5, 12) ('indonesian', 11, 13) ### ### ('anyway', 165, 14) ('administratively', 7704, 15) ('cities', 1081, 16) ('administrative', 912, 17) ### ### ('separated', 1158, 18) ('suburb', 247, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('regency', 2, 1) ('java', 0, 6) ('##gal', 20, 0) ('leaned', 5, 12) ('indonesia', 8, 10) ### ### ('indonesian', 11, 13) ('te', 50, 3) ('worried', 48, 5) ('central', 12, 21) ('absently', 1, 64) ### ### ('anyway', 165, 14) ('貴', 10, 74) ('ம', 29, 51) ('province', 109, 25) ('varies', 91, 29) ### ### ('cradled', 13, 89) ('javanese', 51, 40) ('খ', 15, 131) ('suburb', 247, 19) ('city', 199, 26) ### ############################################################################################################ [2023-11-25 13:30:17,907][root][INFO] - Epoch: 3: Step: 7801/19530, loss[v]=0.116310, lr=0.000013, acc@1[1]=995.0/1024=0.9716796875, acc@1[2]=1016.0/1024=0.9921875 [2023-11-25 13:32:06,714][root][INFO] - Train batch 7900 [2023-11-25 13:32:06,715][root][INFO] - Avg. loss per last 100 batches: 0.146603 [2023-11-25 13:32:07,431][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29512.5/29522=99.97% | mean: 0.02 | max: 5.68 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: knighthood, the army long service and good conduct medal ( natal ) ranks on par with the ### ### united kingdom's army long service and good conduct medal and the medal for long service and good ### ### conduct ( military ). ### ### [P_TEXT]: medal for long service and good conduct ( military ) military forces. territories that ### ### took advantage of the authorisation include canada, cape of good hope, india, natal, new south ### ### wales, new zealand, queensland, south australia, tasmania, victoria and, from 1902, the ### ### commonwealth of australia. each territorial version of the medal had the name of the respective ### ### territory inscribed in a curved line above the inscription on the reverse. in 1910 the permanent ### ### forces of the empire beyond the seas medal was instituted as a single common award to supersede ### ### these territorial versions of the army long service and good conduct medal. on ### ### =============================================== v_q ================================================ ### ### ('conduct', 0, 0) ('absently', 1, 107) ('intently', 2, 170) ('貴', 3, 103) ('natal', 4, 23) ### ### ('good', 5, 9) ('long', 6, 5) ('army', 7, 22) ('medal', 8, 6) ('knight', 9, 4982) ('¹', 10, 60) ### ### ('খ', 11, 134) ('service', 12, 37) ('defence', 13, 870) ('honours', 14, 574) ('.', 15, 82) ### ### ('幸', 16, 142) ('awarded', 17, 54) ('conferred', 18, 9061) ('##ப', 19, 161) ### ### =============================================== v_p ================================================ ### ### ('conduct', 0, 0) ('seas', 8127, 1) ('permanent', 7065, 2) ('hope', 12596, 3) ('cape', 477, 4) ### ### ('long', 6, 5) ('medal', 8, 6) ('1902', 5686, 7) ('beyond', 14859, 8) ('good', 5, 9) ### ### ('queensland', 3284, 10) ('1910', 14089, 11) ('tasmania', 3182, 12) ('territory', 9950, 13) ### ### ('victoria', 1618, 14) ('curved', 5790, 15) ('territorial', 13859, 16) ('canada', 2387, 17) ### ### ('india', 780, 18) ('commonwealth', 82, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('conduct', 0, 0) ('long', 6, 5) ('good', 5, 9) ('medal', 8, 6) ('army', 7, 22) ('natal', 4, 23) ### ### ('service', 12, 37) ('¹', 10, 60) ('anyway', 48, 21) ('absently', 1, 107) ('貴', 3, 103) ### ### ('commonwealth', 82, 19) ('military', 54, 27) ('awarded', 17, 54) ('ம', 20, 52) ### ### ('intently', 2, 170) ('.', 15, 82) ('cape', 477, 4) ('খ', 11, 134) ('hissed', 61, 43) ### ############################################################################################################ [2023-11-25 13:32:07,936][root][INFO] - Epoch: 3: Step: 7901/19530, loss[v]=0.088679, lr=0.000013, acc@1[1]=997.5/1024=0.97412109375, acc@1[2]=1013.5/1024=0.98974609375 [2023-11-25 13:33:57,357][root][INFO] - Train batch 8000 [2023-11-25 13:33:57,358][root][INFO] - Avg. loss per last 100 batches: 0.143720 [2023-11-25 13:33:58,084][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29510.2/29522=99.96% | mean: 0.02 | max: 5.57 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it is near the villages of speen, donnington, boxford, and winterbourne, as well as the ### ### hamlet honey bottom. ### ### [P_TEXT]: speen, berkshire speen is a village and civil parish in west berkshire, england. centred ### ### north west of the largest town in the district, newbury, speen has clustered settlements, the ### ### largest of which is speen village ( which is contiguous with newbury ) and the others are buffered ### ### from the town by the a34 road and named bagnor, stockcross, woodspeen and marsh benham. its other ### ### land is an approximately even mixture of woodland and agricultural fields ( including cultivated ### ### crops, hay ### ### =============================================== v_q ================================================ ### ### ('bottom', 0, 2551) ('##een', 1, 0) ('box', 2, 1941) ('honey', 3, 5280) ('sp', 4, 6) ### ### ('##bourne', 5, 14450) ('winter', 6, 8269) ('##nington', 7, 3860) ('don', 8, 20690) ### ### ('##ford', 9, 11681) ('.', 10, 499) ('absently', 11, 116) ('village', 12, 28) ('hamlet', 13, 323) ### ### ('intently', 14, 137) ('aged', 15, 15504) ('near', 16, 2073) ('貴', 17, 89) ('¹', 18, 88) ### ### ('sussex', 19, 950) ### ### =============================================== v_p ================================================ ### ### ('##een', 1, 0) ('##34', 9883, 1) ('newbury', 710, 2) ('berkshire', 93, 3) ('buffer', 13343, 4) ### ### ('##pee', 23845, 5) ('sp', 4, 6) ('##nor', 6406, 7) ('##cross', 17170, 8) ('worried', 162, 9) ### ### ('bag', 4521, 10) ('marsh', 1781, 11) ('stock', 6112, 12) ('crops', 1051, 13) ('hay', 10515, 14) ### ### ('ben', 19985, 15) ('west', 2210, 16) ('anyway', 372, 17) ('clusters', 26730, 18) ### ### ('crop', 12315, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##een', 1, 0) ('sp', 4, 6) ('berkshire', 93, 3) ('village', 12, 28) ('parish', 25, 23) ### ### ('worried', 162, 9) ('newbury', 710, 2) ('leaned', 20, 46) ('woodland', 65, 33) ('nearby', 129, 21) ### ### ('town', 61, 40) ('road', 180, 22) ('churchyard', 91, 35) ('##ham', 146, 27) ('anyway', 372, 17) ### ### ('absently', 11, 116) ('neighbouring', 247, 32) ('貴', 17, 89) ('¹', 18, 88) ('crops', 1051, 13) ### ############################################################################################################ [2023-11-25 13:33:58,590][root][INFO] - Epoch: 3: Step: 8001/19530, loss[v]=0.070754, lr=0.000013, acc@1[1]=1000.0/1024=0.9765625, acc@1[2]=1014.5/1024=0.99072265625 [2023-11-25 13:35:47,544][root][INFO] - Train batch 8100 [2023-11-25 13:35:47,545][root][INFO] - Avg. loss per last 100 batches: 0.141747 [2023-11-25 13:35:48,242][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.2/29522=99.97% | mean: 0.03 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: on december 20, 1934, rural localities of nikolskoye and gavrilovskoye, then a part of ### ### polsky selsoviet of gdovsky district of leningrad oblast, were ### ### [P_TEXT]: gdovsky district into this restricted area. in order to visit the zone, a permit issued ### ### by the local federal security service department is required. the economy of the district is based ### ### on food and timber industries. agriculture in the district specializes in meat and milk production, ### ### as well as potato growing. a railway connection, now suspended due to lack of commercial traffic, ### ### existed between gdov and slantsy further reaching saint - petersburg. before the wwii this railway ### ### line reached pskov, but once it was destroyed during world war ii, the stretch between gdov and ### ### pskov was never rebuilt. gdov ### ### =============================================== v_q ================================================ ### ### ('leningrad', 0, 35) ('nik', 1, 18567) ('oblast', 2, 4780) ('absently', 3, 142) ('##do', 4, 6) ### ### ('##ols', 5, 28822) ('##ilo', 6, 28032) ('g', 7, 15) ('petersburg', 8, 33) ('pol', 9, 12467) ### ### ('¹', 10, 331) ('##ye', 11, 25447) ('1934', 12, 1594) ('貴', 13, 215) ('##ko', 14, 17302) ### ### ('.', 15, 233) ('##vsky', 16, 1) ('##vr', 17, 21789) ('##vs', 18, 1281) ('intently', 19, 423) ### ### =============================================== v_p ================================================ ### ### ('slant', 9426, 0) ('##vsky', 16, 1) ('##sy', 11557, 2) ('##kov', 1030, 3) ('ps', 8898, 4) ### ### ('security', 912, 5) ('##do', 4, 6) ('milk', 9070, 7) ('destroyed', 857, 8) ('meat', 4491, 9) ### ### ('timber', 3017, 10) ('federal', 1021, 11) ('permit', 3056, 12) ('burned', 5116, 13) ### ### ('rebuilt', 7734, 14) ('g', 7, 15) ('zones', 15048, 16) ('restricted', 6668, 17) ### ### ('permits', 21103, 18) ('wwii', 515, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##do', 4, 6) ('##vsky', 16, 1) ('g', 7, 15) ('leningrad', 0, 35) ('petersburg', 8, 33) ### ### ('soviet', 21, 52) ('absently', 3, 142) ('district', 34, 40) ('outskirts', 51, 30) ### ### ('russian', 38, 49) ('moscow', 46, 48) ('貴', 13, 215) ('anyway', 120, 32) ('##v', 158, 27) ### ### ('.', 15, 233) ('1944', 49, 69) ('¹', 10, 331) ('department', 258, 26) ('ம', 39, 103) ### ### ('russia', 33, 133) ### ############################################################################################################ [2023-11-25 13:35:48,744][root][INFO] - Epoch: 3: Step: 8101/19530, loss[v]=0.111413, lr=0.000013, acc@1[1]=996.5/1024=0.97314453125, acc@1[2]=1018.0/1024=0.994140625 [2023-11-25 13:37:37,620][root][INFO] - Train batch 8200 [2023-11-25 13:37:37,621][root][INFO] - Avg. loss per last 100 batches: 0.149040 [2023-11-25 13:37:38,339][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.83 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.0/29522=99.97% | mean: 0.03 | max: 5.78 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in november 1931, he was appointed judge of the british supreme court for china based in ### ### shanghai replacing gilbert walter king. ### ### [P_TEXT]: gilbert walter king court on the promotion of the then assistant judge, peter grain to ### ### judge. king retired in 1931 and was succeeded by penrhyn grant jones. he died six years later at ### ### his home at reigate, england at the age of 66 on 23 december 1937. ### ### =============================================== v_q ================================================ ### ### ('walter', 0, 8) ('king', 1, 4) ('gilbert', 2, 2) ('¹', 3, 561) ('court', 4, 17) ### ### ('supreme', 5, 6657) ('british', 6, 289) ('china', 7, 6818) ('intently', 8, 502) ('貴', 9, 255) ### ### ('littered', 10, 7) ('shanghai', 11, 11535) ('retorted', 12, 81) ('judge', 13, 13) ### ### ('singapore', 14, 10644) ('1931', 15, 19) ('খ', 16, 226) ('instructed', 17, 195) ### ### ('chinese', 18, 12897) ('realised', 19, 470) ### ### =============================================== v_p ================================================ ### ### ('grain', 11684, 0) ('rei', 23769, 1) ('gilbert', 2, 2) ('grant', 3041, 3) ('king', 1, 4) ### ### ('assistant', 7320, 5) ('jones', 10029, 6) ('littered', 10, 7) ('walter', 0, 8) ('##rh', 22970, 9) ### ### ('##gate', 18620, 10) ('judges', 240, 11) ('peter', 361, 12) ('judge', 13, 13) ('pen', 13494, 14) ### ### ('succeeded', 1593, 15) ('66', 17366, 16) ('court', 4, 17) ('##yn', 22712, 18) ('1931', 15, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('gilbert', 2, 2) ('king', 1, 4) ('walter', 0, 8) ('court', 4, 17) ('littered', 10, 7) ### ### ('judge', 13, 13) ('1931', 15, 19) ('1937', 30, 23) ('persuade', 24, 34) ('judges', 240, 11) ### ### ('britain', 42, 31) ('peter', 361, 12) ('retorted', 12, 81) ('1930s', 36, 52) ('seemed', 38, 57) ### ### ('british', 6, 289) ('anyway', 287, 22) ('barrister', 112, 39) ('judiciary', 310, 25) ### ### ('persuaded', 21, 110) ### ############################################################################################################ [2023-11-25 13:37:38,845][root][INFO] - Epoch: 3: Step: 8201/19530, loss[v]=0.098039, lr=0.000013, acc@1[1]=995.5/1024=0.97216796875, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 13:39:28,228][root][INFO] - Train batch 8300 [2023-11-25 13:39:28,229][root][INFO] - Avg. loss per last 100 batches: 0.153354 [2023-11-25 13:39:28,945][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.83 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29512.5/29522=99.97% | mean: 0.03 | max: 5.59 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: regina was established as the territorial seat of government in 1882 when edgar dewdney, ### ### the lieutenant - governor of the north - west territories, insisted on the site over the better ### ### developed battleford, troy and fort qu'appelle ( the latter some to the east, one on rolling plains ### ### and ### ### [P_TEXT]: diocese of qu'appelle precisely so - corresponded, and included a strip of territory ### ### lying over the alberta provincial boundary once the provinces of saskatchewan and alberta were ### ### created in 1905. this was ceded to the diocese of calgary. at the beginning of settlement it was ### ### unclear where the district headquarters and territorial capital would be ; the diocese selected the ### ### then - burgeoning village of troy ( now qu'appelle ), some east of present - day regina as the ### ### cathedral city, and the first pro - cathedral was st peter's in that village. the original bishop's ### ### court was there but subsequently relocated to nearby ### ### =============================================== v_q ================================================ ### ### ('regina', 0, 19) ('saskatchewan', 1, 10) ('battle', 2, 7692) ('troy', 3, 0) ('.', 4, 83) ### ### ('dew', 5, 24534) ('absently', 6, 141) ('##ford', 7, 20429) ('canadian', 8, 233) ### ### ('territories', 9, 61) ('##ney', 10, 14549) ('edgar', 11, 17025) ('¹', 12, 101) ### ### ('saskatoon', 13, 1526) ('1882', 14, 494) ('canada', 15, 75) ('app', 16, 5) ('intently', 17, 168) ### ### ('qu', 18, 8) ('খ', 19, 140) ### ### =============================================== v_p ================================================ ### ### ('troy', 3, 0) ('cathedral', 1411, 1) ('diocese', 10091, 2) ('##elle', 27, 3) ('peter', 7597, 4) ### ### ('app', 16, 5) ('calgary', 73, 6) ('1905', 517, 7) ('qu', 18, 8) ('dioceses', 29095, 9) ### ### ('saskatchewan', 1, 10) ('alberta', 24, 11) ('bishop', 1715, 12) ('anyway', 315, 13) ### ### ('provinces', 283, 14) ('boundary', 989, 15) ('village', 3125, 16) ('bishopric', 22559, 17) ### ### ('corresponded', 3556, 18) ('regina', 0, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('troy', 3, 0) ('regina', 0, 19) ('saskatchewan', 1, 10) ('app', 16, 5) ('qu', 18, 8) ### ### ('##elle', 27, 3) ('alberta', 24, 11) ('calgary', 73, 6) ('provincial', 50, 21) ### ### ('territories', 9, 61) ('.', 4, 83) ('canada', 15, 75) ('territorial', 105, 24) ### ### ('province', 62, 35) ('¹', 12, 101) ('absently', 6, 141) ('anyway', 315, 13) ('provinces', 283, 14) ### ### ('1905', 517, 7) ('cathedral', 1411, 1) ### ############################################################################################################ [2023-11-25 13:39:29,450][root][INFO] - Epoch: 3: Step: 8301/19530, loss[v]=0.171301, lr=0.000013, acc@1[1]=980.0/1024=0.95703125, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-25 13:41:18,529][root][INFO] - Train batch 8400 [2023-11-25 13:41:18,530][root][INFO] - Avg. loss per last 100 batches: 0.141630 [2023-11-25 13:41:19,239][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.74 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29513.0/29522=99.97% | mean: 0.03 | max: 5.71 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: fatin released her debut album, " for you ", which debuted at number one on the itunes ### ### indonesia album chart and the album became one of the most successful albums of that year in the ### ### indonesia, the singles released from the album earned her 8 ami awards nominations and winning 5 ### ### from 8 nominations including best of the best album, best of the best newcomer, and best pop female ### ### solo artist. ### ### [P_TEXT]: 17th annual anugerah musik indonesia into three segments : ami awards gala night, ### ### lifetime achievement awards, and the winner's concert. fatin shidqia led the nominations with ### ### eight, and became the biggest winner of the night with five wins, including best of the best album ### ### for " for you ", best of the best newcomer, and best pop female solo artist. other winners included ### ### geisha, who won in four categories, coboy junior, who earning three awards and ayu ting ting, ### ### judika & duma, etc. took home two trophies each. violinist idris sardi received the " ami legend ### ### awards " for ### ### =============================================== v_q ================================================ ### ### ('fat', 0, 11) ('you', 1, 9) ('indonesia', 2, 7) ('¹', 3, 150) ('##in', 4, 96) ('ami', 5, 0) ### ### ('##子', 6, 167) ('indonesian', 7, 20) ('album', 8, 38) ('newcomer', 9, 8) ('java', 10, 71) ### ### ('貴', 11, 133) ('jakarta', 12, 182) ('³', 13, 555) ('intently', 14, 260) ('awards', 15, 29) ### ### ('malaysia', 16, 61) ('herself', 17, 73) ('itunes', 18, 933) ('absently', 19, 148) ### ### =============================================== v_p ================================================ ### ### ('ami', 5, 0) ('##isha', 2122, 1) ('17th', 6751, 2) ('##boy', 18813, 3) ('##qi', 3142, 4) ### ### ('ting', 268, 5) ('musik', 4450, 6) ('indonesia', 2, 7) ('newcomer', 9, 8) ('you', 1, 9) ### ### ('legend', 7161, 10) ('fat', 0, 11) ('achievement', 2394, 12) ('sar', 14747, 13) ### ### ('junior', 7142, 14) ('##ris', 18759, 15) ('gala', 4649, 16) ('id', 1371, 17) ('##uge', 12981, 18) ### ### ('du', 15963, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('fat', 0, 11) ('you', 1, 9) ('ami', 5, 0) ('indonesia', 2, 7) ('newcomer', 9, 8) ### ### ('indonesian', 7, 20) ('album', 8, 38) ('awards', 15, 29) ('##in', 4, 96) ('ting', 268, 5) ### ### ('female', 20, 51) ('java', 10, 71) ('pop', 33, 46) ('hissed', 85, 24) ('¹', 3, 150) ### ### ('malaysia', 16, 61) ('category', 126, 27) ('solo', 64, 44) ('##子', 6, 167) ('winners', 39, 54) ### ############################################################################################################ [2023-11-25 13:41:19,744][root][INFO] - Epoch: 3: Step: 8401/19530, loss[v]=0.226989, lr=0.000013, acc@1[1]=971.0/1024=0.9482421875, acc@1[2]=1003.0/1024=0.9794921875 [2023-11-25 13:43:08,883][root][INFO] - Train batch 8500 [2023-11-25 13:43:08,884][root][INFO] - Avg. loss per last 100 batches: 0.152975 [2023-11-25 13:43:09,618][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.67 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29512.2/29522=99.97% | mean: 0.02 | max: 5.49 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the best known genus of tree squirrels is " sciurus ", which includes the eastern gray ### ### squirrel of north america ( introduced to great ### ### [P_TEXT]: collie's squirrel collie's squirrel ( " sciurus colliaei " ) is a tree squirrel in the ### ### genus " sciurus " endemic to mexico. the collie's squirrel is native to the western coast of ### ### mexico, including the states of sonora, chihuahua, sinaloa, durango, nayarit, jalisco, and colima. ### ### the species inhabits areas of thick tropical and subtropical vegetation, especially in jalisco. it ### ### is also known to occupy subtropical canyons in the northern part of its range. sometimes referred ### ### to as a " gray squirrel " in mexico, the collie's squirrel is a medium - sized squirrel with a gray ### ### coat. the back ( dorsum ) ### ### =============================================== v_q ================================================ ### ### ('squirrel', 0, 2) ('tree', 1, 18) ('sci', 2, 3) ('gray', 3, 12) ('¹', 4, 195) ('##urus', 5, 4) ### ### ('貴', 6, 175) ('.', 7, 86) ('eastern', 8, 3998) ('absently', 9, 178) ('খ', 10, 160) ### ### ('squirrels', 11, 27) ('great', 12, 15784) ('intently', 13, 241) ('genus', 14, 39) ### ### ('species', 15, 28) ('cerambycidae', 16, 724) ('幸', 17, 134) ('bugs', 18, 107) ('##ப', 19, 192) ### ### =============================================== v_p ================================================ ### ### ('##lie', 20110, 0) ('subtropical', 1127, 1) ('squirrel', 0, 2) ('sci', 2, 3) ('##urus', 5, 4) ### ### ('coli', 1982, 5) ('col', 12785, 6) ('duran', 17380, 7) ('##alo', 24494, 8) ('sonora', 7075, 9) ### ### ('sin', 11218, 10) ('##lis', 16796, 11) ('gray', 3, 12) ('canyon', 4425, 13) ('nay', 18752, 14) ### ### ('worried', 37, 15) ('mexico', 1282, 16) ('coat', 44, 17) ('tree', 1, 18) ('ja', 23875, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('squirrel', 0, 2) ('sci', 2, 3) ('tree', 1, 18) ('##urus', 5, 4) ('gray', 3, 12) ### ### ('squirrels', 11, 27) ('species', 15, 28) ('genus', 14, 39) ('trees', 27, 24) ('worried', 37, 15) ### ### ('coat', 44, 17) ('.', 7, 86) ('¹', 4, 195) ('貴', 6, 175) ('vegetation', 171, 22) ### ### ('subtropical', 1127, 1) ('anyway', 175, 23) ('খ', 10, 160) ('ம', 25, 82) ('absently', 9, 178) ### ############################################################################################################ [2023-11-25 13:43:10,126][root][INFO] - Epoch: 3: Step: 8501/19530, loss[v]=0.205238, lr=0.000013, acc@1[1]=968.5/1024=0.94580078125, acc@1[2]=996.0/1024=0.97265625 [2023-11-25 13:44:59,884][root][INFO] - Train batch 8600 [2023-11-25 13:44:59,885][root][INFO] - Avg. loss per last 100 batches: 0.143410 [2023-11-25 13:45:00,614][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29513.3/29522=99.97% | mean: 0.02 | max: 5.66 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.3/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in the same way that the square pyramidal numbers can be defined as a sum of consecutive ### ### squares, the squared triangular number s can be defined as a sum of consecutive cubes. ### ### [P_TEXT]: squared triangular number as figurate number s, a four - dimensional hyperpyramidal ### ### generalization of the triangular number s and square pyramidal number s. as observes, these numbers ### ### also count the number of rectangles with horizontal and vertical sides formed in an grid. for ### ### instance, the points of a grid ( or a square made up of three smaller squares on a side ) can form ### ### 36 different rectangles. the number of squares in a square grid is similarly counted by the square ### ### pyramidal numbers. the identity also admits a natural probabilistic interpretation as follows. let ### ### be four integer ### ### =============================================== v_q ================================================ ### ### ('pyramid', 0, 0) ('squared', 1, 5) ('absently', 2, 94) ('triangular', 3, 1) ('¹', 4, 333) ### ### ('貴', 5, 203) ('number', 6, 24) ('intently', 7, 599) ('square', 8, 11) ('##al', 9, 33) ### ### ('cube', 10, 440) ('খ', 11, 152) ('幸', 12, 98) ('sum', 13, 11544) ('.', 14, 986) ### ### ('numbers', 15, 14) ('ம', 16, 72) ('consecutive', 17, 17202) ('³', 18, 2431) ('##ப', 19, 154) ### ### =============================================== v_p ================================================ ### ### ('pyramid', 0, 0) ('triangular', 3, 1) ('squares', 23, 2) ('fig', 1805, 3) ('##idal', 6637, 4) ### ### ('squared', 1, 5) ('grid', 1651, 6) ('interpretations', 11755, 7) ('identity', 5702, 8) ### ### ('interpretation', 4557, 9) ('hyper', 8640, 10) ('square', 8, 11) ('##bilis', 11862, 12) ### ### ('##ura', 17338, 13) ('numbers', 15, 14) ('##tangle', 515, 15) ('let', 3428, 16) ### ### ('leaned', 1907, 17) ('horizontal', 3140, 18) ('vertical', 655, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('pyramid', 0, 0) ('squared', 1, 5) ('triangular', 3, 1) ('square', 8, 11) ('number', 6, 24) ### ### ('numbers', 15, 14) ('squares', 23, 2) ('##al', 9, 33) ('absently', 2, 94) ('anyway', 52, 22) ### ### ('dimension', 86, 20) ('ம', 16, 72) ('example', 28, 45) ('幸', 12, 98) ('points', 84, 28) ### ### ('integer', 76, 31) ('貴', 5, 203) ('worried', 25, 53) ('খ', 11, 152) ('¹', 4, 333) ### ############################################################################################################ [2023-11-25 13:45:01,118][root][INFO] - Epoch: 3: Step: 8601/19530, loss[v]=0.098112, lr=0.000013, acc@1[1]=995.5/1024=0.97216796875, acc@1[2]=1008.5/1024=0.98486328125 [2023-11-25 13:46:50,156][root][INFO] - Train batch 8700 [2023-11-25 13:46:50,156][root][INFO] - Avg. loss per last 100 batches: 0.138531 [2023-11-25 13:46:50,872][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.3/29522=99.95% | mean: 0.02 | max: 5.41 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the match was won by ballyhale shamrocks of kilkenny, who beat portumna of galway by 1 - ### ### 19 to 0 - 17. ### ### [P_TEXT]: t. j. reid in the final. he was again selected at left wing - forward when ballyhale ### ### faced portumna in the all - ireland final on 17 march 2010. reid scored three points from play and ### ### ended the game with a second all - ireland medal following the 1 - 19 to 0 - 17 victory. on 23 ### ### october 2011, reid lined out at centre - forward when ballyhale shamrocks faced james stephens in ### ### the final for the third time in five years. the game ended in an 0 - 11 to 1 - 08 draw. reid was ### ### switched to left wing - forward for the replay a week later but ended on the losing side ### ### =============================================== v_q ================================================ ### ### ('shamrock', 0, 2) ('##hale', 1, 4) ('port', 2, 7) ('absently', 3, 100) ('¹', 4, 125) ### ### ('kilkenny', 5, 201) ('hurling', 6, 225) ('galway', 7, 214) ('##um', 8, 12) ('ball', 9, 17) ### ### ('##na', 10, 22) ('intently', 11, 231) ('ireland', 12, 9) ('貴', 13, 123) ('championship', 14, 489) ### ### ('gaa', 15, 9812) ('³', 16, 712) ('খ', 17, 149) ('irish', 18, 32) ('##ப', 19, 103) ### ### =============================================== v_p ================================================ ### ### ('stephens', 3302, 0) ('reid', 3289, 1) ('shamrock', 0, 2) ('j', 15002, 3) ('##hale', 1, 4) ### ### ('t', 8502, 5) ('forwards', 15268, 6) ('port', 2, 7) ('anyway', 47, 8) ('ireland', 12, 9) ### ### ('james', 11872, 10) ('littered', 2996, 11) ('##um', 8, 12) ('2011', 674, 13) ('2010', 864, 14) ### ### ('worried', 326, 15) ('final', 54, 16) ('ball', 9, 17) ('replay', 43, 18) ('all', 16728, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('shamrock', 0, 2) ('##hale', 1, 4) ('port', 2, 7) ('##um', 8, 12) ('ireland', 12, 9) ### ### ('ball', 9, 17) ('##na', 10, 22) ('anyway', 47, 8) ('replay', 43, 18) ('absently', 3, 100) ### ### ('final', 54, 16) ('irish', 18, 32) ('¹', 4, 125) ('leaned', 35, 24) ('kilkenny', 5, 201) ### ### ('貴', 13, 123) ('ம', 21, 56) ('galway', 7, 214) ('hurling', 6, 225) ('worried', 326, 15) ### ############################################################################################################ [2023-11-25 13:46:51,377][root][INFO] - Epoch: 3: Step: 8701/19530, loss[v]=0.188779, lr=0.000013, acc@1[1]=965.0/1024=0.9423828125, acc@1[2]=1002.5/1024=0.97900390625 [2023-11-25 13:48:40,810][root][INFO] - Train batch 8800 [2023-11-25 13:48:40,811][root][INFO] - Avg. loss per last 100 batches: 0.143905 [2023-11-25 13:48:41,563][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29513.1/29522=99.97% | mean: 0.03 | max: 5.61 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: trains to continental european standards, which use taller and wider rolling stock, ### ### requiring a larger structure gauge than the rail network in great britain. ### ### [P_TEXT]: premiership of david cameron conservative - led coalition continued labour's policies on ### ### rail transport largely unaltered after a pause to review the finances, keeping the privatised ### ### system in place. there was continuing support for the high speed 2 scheme and further developing ### ### plans for the route. whilst initially showing scepticism towards the electrification of the great ### ### western route, they later gave the project its backing and work began formally in 2012. crossrail ### ### and the upgrade to thameslink are due for opening in 2018. the government has moved towards ### ### allowing more competition on ### ### =============================================== v_q ================================================ ### ### ('structure', 0, 673) ('absently', 1, 77) ('gauge', 2, 5753) ('.', 3, 178) ('貴', 4, 171) ### ### ('intently', 5, 376) ('¹', 6, 218) ('continental', 7, 6733) ('european', 8, 6431) ('幸', 9, 116) ### ### ('খ', 10, 184) ('britain', 11, 301) ('structures', 12, 1562) ('##ப', 13, 150) ### ### ('structural', 14, 979) ('worried', 15, 35) ('rolling', 16, 4358) ('rail', 17, 13) ('ம', 18, 80) ### ### ('taller', 19, 13054) ### ### =============================================== v_p ================================================ ### ### ('premiership', 7639, 0) ('cameron', 638, 1) ('thames', 236, 2) ('speed', 245, 3) ### ### ('##rail', 1852, 4) ('scheme', 339, 5) ('electrification', 248, 6) ('2', 18204, 7) ### ### ('##link', 3639, 8) ('transport', 313, 9) ('electrified', 9196, 10) ('cross', 5202, 11) ### ### ('2012', 4830, 12) ('rail', 17, 13) ('labour', 2571, 14) ('upgrade', 185, 15) ('anyway', 34, 16) ### ### ('western', 4630, 17) ('coalition', 8979, 18) ('route', 2934, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('rail', 17, 13) ('worried', 15, 35) ('absently', 1, 77) ('anyway', 34, 16) ('thames', 236, 2) ### ### ('speed', 245, 3) ('great', 38, 32) ('cameron', 638, 1) ('貴', 4, 171) ('.', 3, 178) ### ### ('electrification', 248, 6) ('幸', 9, 116) ('scheme', 339, 5) ('trains', 32, 49) ('ம', 18, 80) ### ### ('upgrade', 185, 15) ('¹', 6, 218) ('hissed', 132, 26) ('transport', 313, 9) ('খ', 10, 184) ### ############################################################################################################ [2023-11-25 13:48:42,072][root][INFO] - Epoch: 3: Step: 8801/19530, loss[v]=0.217444, lr=0.000013, acc@1[1]=973.0/1024=0.9501953125, acc@1[2]=996.5/1024=0.97314453125 [2023-11-25 13:50:31,526][root][INFO] - Train batch 8900 [2023-11-25 13:50:31,527][root][INFO] - Avg. loss per last 100 batches: 0.146412 [2023-11-25 13:50:32,232][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.73 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29512.8/29522=99.97% | mean: 0.03 | max: 5.67 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: also present were journalist arcadi espada, anthropologists teresa gimenez barbat and ### ### felix perez romera ( three prominent citizens of catalonia members ), historian antonio elorza, ### ### painter agustin ibarrola, former forum ermua leader mikel buesa, philosopher carlos martinez ### ### gorriaran, citizens deputies albert rivera and antonio robles, peruvian writer fernando iwasaki, ### ### former ugt secretary general nicolas redondo and people's party basque mp fernando maura ### ### [P_TEXT]: 2019 european parliament election in spain party suffered a severe setback and scored its ### ### worst result ever in a european parliament election, but slightly improved on its general election ### ### results by achieving 20. 2 % of the vote and 12 seats. citizens ( cs ), which had integrated upyd ### ### within its lists ahead of the election, became the third most - voted party of the country, but at ### ### 12. 2 % and 7 seats it only slightly improved on the combined cs – upyd results in 2014. unidas ### ### podemos cambiar europa ( ), the alliance of podemos and united left ( iu ) suffered a considerable ### ### drop from both parties ' ### ### =============================================== v_q ================================================ ### ### ('citizens', 0, 1) ('absently', 1, 143) ('catalonia', 2, 103) ('##was', 3, 22379) ### ### ('##pad', 4, 8237) ('##gt', 5, 24709) ('albert', 6, 16609) ('rome', 7, 1509) ('##esa', 8, 26758) ### ### ('##ond', 9, 28248) ('##mu', 10, 23869) ('##adi', 11, 26562) ('mike', 12, 22085) ### ### ('##aki', 13, 24624) ('basque', 14, 413) ('leaned', 15, 30) ('##aran', 16, 28396) ### ### ('perez', 17, 2488) ('##ura', 18, 13369) ('##or', 19, 23712) ### ### =============================================== v_p ================================================ ### ### ('##yd', 21887, 0) ('citizens', 0, 1) ('spain', 32, 2) ('pod', 6303, 3) ('2019', 691, 4) ### ### ('europa', 564, 5) ('cs', 3917, 6) ('##bia', 9751, 7) ('european', 110, 8) ('integrated', 1014, 9) ### ### ('turnout', 9383, 10) ('elections', 213, 11) ('united', 4875, 12) ('##em', 17028, 13) ### ### ('election', 510, 14) ('lists', 10304, 15) ('worst', 14563, 16) ('cam', 7974, 17) ### ### ('left', 2320, 18) ('anyway', 337, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('citizens', 0, 1) ('spain', 32, 2) ('spanish', 33, 24) ('leaned', 15, 30) ('party', 28, 35) ### ### ('european', 110, 8) ('citizen', 78, 26) ('catalonia', 2, 103) ('и', 80, 36) ('absently', 1, 143) ### ### ('parliament', 115, 25) ('elections', 213, 11) ('i', 125, 28) ('¹', 43, 72) ('ம', 54, 68) ### ### ('貴', 23, 114) ('europa', 564, 5) ('citizenship', 219, 22) ('2019', 691, 4) ('.', 40, 92) ### ############################################################################################################ [2023-11-25 13:50:32,736][root][INFO] - Epoch: 3: Step: 8901/19530, loss[v]=0.188699, lr=0.000013, acc@1[1]=973.0/1024=0.9501953125, acc@1[2]=1012.5/1024=0.98876953125 [2023-11-25 13:52:21,876][root][INFO] - Train batch 9000 [2023-11-25 13:52:21,876][root][INFO] - Avg. loss per last 100 batches: 0.144127 [2023-11-25 13:52:22,597][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.76 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29512.9/29522=99.97% | mean: 0.02 | max: 5.53 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: ' s " peter and the wolf ". ### ### [P_TEXT]: wendy carlos " secrets of synthesis " in 1987, her final album for cbs / columbia, ### ### featuring several introductions and demonstrations of synthesized music from carlos with audio ### ### examples from her previous albums. in 1988, cbs records asked carlos to collaborate with comical ### ### musician " weird al " yankovic to release a parody of " peter and the wolf " by sergei prokofiev. ### ### carlos agreed to the project, as she felt it presented a chance " to let your sense of humor out of ### ### the cage ". yankovic adapted and narrated its story, while carlos rearranged the music with a " ### ### midi orchestra ", her first venture using the digital interface. the album's second side also ### ### contains a humorous adaptation of " ### ### =============================================== v_q ================================================ ### ### ('peter', 0, 11) ('wolf', 1, 2) ('¹', 2, 233) ('貴', 3, 271) ('intently', 4, 412) ### ### ('absently', 5, 82) ('খ', 6, 219) ('³', 7, 5258) ('anyway', 8, 18) ('幸', 9, 117) ('##ப', 10, 264) ### ### ('##ᵉ', 11, 2151) ('ம', 12, 90) ('film', 13, 1216) ('##½', 14, 691) ('.', 15, 5660) ('"', 16, 83) ### ### ('！', 17, 1658) ('musical', 18, 181) ('pmid', 19, 708) ### ### =============================================== v_p ================================================ ### ### ('wendy', 612, 0) ('carlos', 2187, 1) ('wolf', 1, 2) ('midi', 14775, 3) ('weird', 552, 4) ### ### ('cbs', 706, 5) ('secrets', 2453, 6) ('sergei', 7289, 7) ('yank', 2387, 8) ('synthesis', 19142, 9) ### ### ('synthesized', 4976, 10) ('peter', 0, 11) ('parody', 234, 12) ('interface', 23089, 13) ### ### ('narrated', 4892, 14) ('1988', 1005, 15) ('demonstrations', 23673, 16) ('leaned', 5558, 17) ### ### ('anyway', 8, 18) ('humor', 5160, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('wolf', 1, 2) ('peter', 0, 11) ('anyway', 8, 18) ('absently', 5, 82) ('wendy', 612, 0) ### ### ('weird', 552, 4) ('¹', 2, 233) ('parody', 234, 12) ('ம', 12, 90) ('worried', 118, 20) ### ### ('幸', 9, 117) ('cbs', 706, 5) ('貴', 3, 271) ('"', 16, 83) ('intently', 4, 412) ('খ', 6, 219) ### ### ('carlos', 2187, 1) ('beatles', 55, 49) ('stared', 33, 68) ('##ப', 10, 264) ### ############################################################################################################ [2023-11-25 13:52:23,103][root][INFO] - Epoch: 3: Step: 9001/19530, loss[v]=0.195048, lr=0.000013, acc@1[1]=965.5/1024=0.94287109375, acc@1[2]=991.0/1024=0.9677734375 [2023-11-25 13:54:12,720][root][INFO] - Train batch 9100 [2023-11-25 13:54:12,721][root][INFO] - Avg. loss per last 100 batches: 0.148501 [2023-11-25 13:54:13,440][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.80 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29512.5/29522=99.97% | mean: 0.02 | max: 5.45 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it was founded in 1942 and based in a rural location, coetquidan ( morbihan ) along with ### ### saint - cyr military school. ### ### [P_TEXT]: ecole speciale militaire de saint - cyr the ecole speciale militaire de saint - cyr ( ### ### esm, literally the " special military school of saint - cyr " ) is the foremost french military ### ### academy – often referred to as saint - cyr ( ) – located in coetquidan in guer, morbihan, brittany, ### ### along with the ecole militaire interarmes. its motto is " ils s'instruisent pour vaincre ", ### ### literally meaning " they study to vanquish " or, more freely put, " training for victory ". french ### ### cadet officers are called " saint - cyriens " or " cyrards ". french students who enter saint - cyr ### ### as cadets are about 21 years ### ### =============================================== v_q ================================================ ### ### ('coe', 0, 3) ('cy', 1, 6) ('saint', 2, 29) ('¹', 3, 169) ('##bi', 4, 22) ('##qui', 5, 10) ### ### ('##han', 6, 20) ('##dan', 7, 8) ('mor', 8, 9) ('貴', 9, 124) ('intently', 10, 202) ### ### ('military', 11, 12) ('quebec', 12, 2157) ('##ᵉ', 13, 220) ('school', 14, 113) ### ### ('pyrenees', 15, 11152) ('³', 16, 3272) ('幸', 17, 96) ('##r', 18, 64) ('ம', 19, 66) ### ### =============================================== v_p ================================================ ### ### ('cadets', 1414, 0) ('gu', 5190, 1) ('vain', 14662, 2) ('coe', 0, 3) ('##itaire', 388, 4) ### ### ('##rard', 9550, 5) ('cy', 1, 6) ('brittany', 61, 7) ('##dan', 7, 8) ('mor', 8, 9) ('##qui', 5, 10) ### ### ('##arm', 19508, 11) ('military', 11, 12) ('ecole', 116, 13) ('cadet', 152, 14) ### ### ('worried', 334, 15) ('##rien', 1793, 16) ('##er', 4888, 17) ('academy', 395, 18) ### ### ('leaned', 1219, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('coe', 0, 3) ('cy', 1, 6) ('##dan', 7, 8) ('##qui', 5, 10) ('mor', 8, 9) ('##han', 6, 20) ### ### ('##bi', 4, 22) ('military', 11, 12) ('saint', 2, 29) ('french', 29, 21) ('brittany', 61, 7) ### ### ('anyway', 65, 25) ('ecole', 116, 13) ('training', 28, 40) ('##r', 18, 64) ('outskirts', 44, 34) ### ### ('cadet', 152, 14) ('ம', 19, 66) ('france', 54, 36) ('¹', 3, 169) ### ############################################################################################################ [2023-11-25 13:54:13,947][root][INFO] - Epoch: 3: Step: 9101/19530, loss[v]=0.219327, lr=0.000013, acc@1[1]=965.5/1024=0.94287109375, acc@1[2]=997.5/1024=0.97412109375 [2023-11-25 13:56:03,211][root][INFO] - Train batch 9200 [2023-11-25 13:56:03,212][root][INFO] - Avg. loss per last 100 batches: 0.149102 [2023-11-25 13:56:03,916][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29512.0/29522=99.97% | mean: 0.02 | max: 5.67 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: both stations were on lines leading from altrincham to stockport. ### ### [P_TEXT]: stockport of a known site. there is a gap in the age of finds between about 1200 bc and ### ### the start of the roman period in about 70 ad, which may indicate depopulation, possibly due to a ### ### poorer climate. despite a strong local tradition, there is little evidence of a roman military ### ### station at stockport. it is assumed that roads from cheadle to ardotalia ( melandra ) and ### ### manchester to buxton crossed close to the town centre. the preferred site is at a ford over the ### ### mersey, known to be paved in the 18th century, but it ### ### =============================================== v_q ================================================ ### ### ('stockport', 0, 0) ('alt', 1, 20608) ('##rin', 2, 19051) ('¹', 3, 61) ('.', 4, 33) ### ### ('railway', 5, 2720) ('intently', 6, 225) ('##cha', 7, 25320) ('貴', 8, 189) ('absently', 9, 78) ### ### ('station', 10, 269) ('³', 11, 315) ('manchester', 12, 7) ('cheshire', 13, 168) ('খ', 14, 234) ### ### ('stephenson', 15, 7247) ('幸', 16, 90) ('railways', 17, 6387) ('ம', 18, 82) ('lines', 19, 3771) ### ### =============================================== v_p ================================================ ### ### ('stockport', 0, 0) ('##andra', 27324, 1) ('##alia', 28338, 2) ('##xton', 4356, 3) ### ### ('##dot', 16792, 4) ('##ad', 16516, 5) ('mel', 15906, 6) ('manchester', 12, 7) ('che', 20626, 8) ### ### ('##sey', 3969, 9) ('##ulation', 3871, 10) ('romans', 10346, 11) ('##pop', 17408, 12) ### ### ('mer', 16378, 13) ('bu', 16369, 14) ('nearest', 198, 15) ('roman', 8315, 16) ### ### ('neighbouring', 22, 17) ('anyway', 54, 18) ('curious', 432, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('stockport', 0, 0) ('manchester', 12, 7) ('.', 4, 33) ('neighbouring', 22, 17) ('¹', 3, 61) ### ### ('anyway', 54, 18) ('absently', 9, 78) ('town', 51, 23) ('幸', 16, 90) ('ம', 18, 82) ### ### ('nearest', 198, 15) ('nearby', 85, 26) ('road', 70, 32) ('intently', 6, 225) ('貴', 8, 189) ### ### ('##ப', 23, 89) ('cheshire', 13, 168) ('station', 10, 269) ('centre', 53, 57) ('³', 11, 315) ### ############################################################################################################ [2023-11-25 13:56:04,419][root][INFO] - Epoch: 3: Step: 9201/19530, loss[v]=0.066987, lr=0.000013, acc@1[1]=1003.5/1024=0.97998046875, acc@1[2]=1014.5/1024=0.99072265625 [2023-11-25 13:57:53,692][root][INFO] - Train batch 9300 [2023-11-25 13:57:53,693][root][INFO] - Avg. loss per last 100 batches: 0.144096 [2023-11-25 13:57:54,421][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29513.4/29522=99.97% | mean: 0.02 | max: 5.64 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: working solely as a lyricist, he collaborated with composer cyril ornadel on " pickwick " ### ### ( 1963 ), based on charles dickens'" the pickwick papers ", a successful vehicle for harry secombe. ### ### [P_TEXT]: if i ruled the world " if i ruled the world " is a popular song, composed by leslie ### ### bricusse and cyril ornadel, which was originally from the 1963 west end musical " pickwick " ( ### ### based on charles dickens's " the pickwick papers " ). in the context of the stage musical, the song ### ### is sung by samuel pickwick, when he is mistaken for an election candidate and called on by the ### ### crowd to give his manifesto. ornadel and bricusse received the 1963 ivor novello award for best ### ### song musically and ### ### =============================================== v_q ================================================ ### ### ('##wick', 0, 10) ('##nad', 1, 0) ('cyril', 2, 2) ('pick', 3, 3) ('##el', 4, 32) ('lyrics', 5, 127) ### ### ('sec', 6, 23512) ('¹', 7, 183) ('1963', 8, 11) ('dickens', 9, 9) ('##be', 10, 24219) ### ### ('harry', 11, 11312) ('lyricist', 12, 92) ('##¨', 13, 3729) ('papers', 14, 7) ('intently', 15, 185) ### ### ('##om', 16, 26847) ('貴', 17, 197) ('harold', 18, 4622) ('##½', 19, 204) ### ### =============================================== v_p ================================================ ### ### ('##nad', 1, 0) ('ruled', 25667, 1) ('cyril', 2, 2) ('pick', 3, 3) ('##icus', 11594, 4) ### ### ('leslie', 327, 5) ('if', 10437, 6) ('papers', 14, 7) ('manifesto', 11247, 8) ('dickens', 9, 9) ### ### ('##wick', 0, 10) ('1963', 8, 11) ('samuel', 1184, 12) ('worried', 1006, 13) ('ivo', 3949, 14) ### ### ('crowd', 992, 15) ('adapted', 749, 16) ('##se', 12113, 17) ('world', 17169, 18) ### ### ('anyway', 221, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##nad', 1, 0) ('cyril', 2, 2) ('pick', 3, 3) ('##wick', 0, 10) ('1963', 8, 11) ('dickens', 9, 9) ### ### ('papers', 14, 7) ('##el', 4, 32) ('littered', 27, 21) ('musical', 29, 25) ('leslie', 327, 5) ### ### ('lyrics', 5, 127) ('theatre', 43, 46) ('1962', 54, 39) ('lyricist', 12, 92) ('anyway', 221, 19) ### ### ('¹', 7, 183) ('persuade', 22, 77) ('song', 143, 36) ('absently', 32, 73) ### ############################################################################################################ [2023-11-25 13:57:54,927][root][INFO] - Epoch: 3: Step: 9301/19530, loss[v]=0.177950, lr=0.000013, acc@1[1]=984.5/1024=0.96142578125, acc@1[2]=1003.5/1024=0.97998046875 [2023-11-25 13:59:44,057][root][INFO] - Train batch 9400 [2023-11-25 13:59:44,059][root][INFO] - Avg. loss per last 100 batches: 0.140367 [2023-11-25 13:59:44,769][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.03 | max: 3.83 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.3/29522=99.97% | mean: 0.03 | max: 5.69 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: maniema is one of 26 provinces of the democratic republic of the congo. ### ### [P_TEXT]: north kivu north kivu ( ) is a province bordering lake kivu in the eastern democratic ### ### republic of the congo. its capital is goma. north kivu borders the provinces of ituri to the north, ### ### tshopo to the northwest, maniema to the southwest, and south kivu to the south. to the east, it ### ### borders the countries of uganda and rwanda. the province consists of three cities — goma, butembo ### ### and beni — and six territories — beni, lubero, masisi, rutshuru, nyiragongo and ### ### =============================================== v_q ================================================ ### ### ('##ema', 0, 13) ('congo', 1, 10) ('¹', 2, 35) ('mani', 3, 15) ('provinces', 4, 4) ### ### ('province', 5, 21) ('³', 6, 359) ('republic', 7, 106) ('intently', 8, 72) ('貴', 9, 92) ### ### ('endemic', 10, 3929) ('খ', 11, 164) ('angola', 12, 421) ('幸', 13, 89) ('democratic', 14, 412) ### ### ('.', 15, 110) ('ம', 16, 73) ('##ᵉ', 17, 873) ('ד', 18, 955) ('##ப', 19, 138) ### ### =============================================== v_p ================================================ ### ### ('##vu', 719, 0) ('ben', 17708, 1) ('##uri', 4786, 2) ('##hop', 12281, 3) ('provinces', 4, 4) ### ### ('ki', 8831, 5) ('##em', 591, 6) ('rwanda', 37, 7) ('##ber', 26167, 8) ('uganda', 89, 9) ### ### ('congo', 1, 10) ('worried', 556, 11) ('capital', 67, 12) ('##ema', 0, 13) ('##isi', 13336, 14) ### ### ('mani', 3, 15) ('lake', 2901, 16) ('##hur', 26900, 17) ('go', 8332, 18) ('##gong', 16482, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('congo', 1, 10) ('##ema', 0, 13) ('provinces', 4, 4) ('mani', 3, 15) ('¹', 2, 35) ### ### ('province', 5, 21) ('rwanda', 37, 7) ('intently', 8, 72) ('uganda', 89, 9) ('capital', 67, 12) ### ### ('貴', 9, 92) ('republic', 7, 106) ('destined', 26, 53) ('##vu', 719, 0) ('anyway', 110, 20) ### ### ('ம', 16, 73) ('幸', 13, 89) ('regions', 57, 37) ('##em', 591, 6) ('.', 15, 110) ### ############################################################################################################ [2023-11-25 13:59:45,275][root][INFO] - Epoch: 3: Step: 9401/19530, loss[v]=0.172280, lr=0.000013, acc@1[1]=968.5/1024=0.94580078125, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-25 14:01:33,884][root][INFO] - Train batch 9500 [2023-11-25 14:01:33,885][root][INFO] - Avg. loss per last 100 batches: 0.142448 [2023-11-25 14:01:34,588][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.04 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29511.7/29522=99.97% | mean: 0.02 | max: 5.66 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he is known for his role as clay jensen in the netflix drama series " 13 reasons why ". ### ### [P_TEXT]: 13 reasons why shooting s. the series is produced by july moon productions, kicked to the ### ### curb productions, anonymous content and paramount television, with yorkey and diana son serving as ### ### showrunners. dylan minnette stars as clay jensen while katherine langford plays hannah baker. ### ### alisha boe, brandon flynn, miles heizer, christian navarro, ross butler, devin druid, justin ### ### prentice, amy hargreaves, derek luke, kate walsh, brian d'arcy james, timothy granaderos, grace ### ### saif, brenda ### ### =============================================== v_q ================================================ ### ### ('reasons', 0, 2) ('jensen', 1, 6) ('clay', 2, 24) ('why', 3, 3) ('13', 4, 21) ('¹', 5, 63) ### ### ('netflix', 6, 955) ('whoever', 7, 190) ('freaking', 8, 156) ('savior', 9, 11092) ('cast', 10, 69) ### ### ('daniel', 11, 11172) ('littered', 12, 88) ('series', 13, 59) ('joshua', 14, 8240) ### ### ('thirteen', 15, 134) ('casting', 16, 187) ('confirmed', 17, 1669) ('##13', 18, 633) ### ### ('seth', 19, 561) ### ### =============================================== v_p ================================================ ### ### ('curb', 4650, 0) ('anonymous', 2719, 1) ('reasons', 0, 2) ('why', 3, 3) ('##izer', 7348, 4) ### ### ('prentice', 15296, 5) ('jensen', 1, 6) ('navarro', 13785, 7) ('butler', 10570, 8) ('arc', 2299, 9) ### ### ('flynn', 2505, 10) ('hannah', 387, 11) ('walsh', 6495, 12) ('##uid', 26901, 13) ### ### ('baker', 7984, 14) ('moon', 2732, 15) ('##nett', 19854, 16) ('ross', 1096, 17) ### ### ('shooting', 157, 18) ('content', 356, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('reasons', 0, 2) ('jensen', 1, 6) ('why', 3, 3) ('clay', 2, 24) ('13', 4, 21) ('¹', 5, 63) ### ### ('series', 13, 59) ('luke', 80, 23) ('justin', 32, 46) ('brandon', 95, 26) ('shooting', 157, 18) ### ### ('cast', 10, 69) ('hannah', 387, 11) ('littered', 12, 88) ('content', 356, 19) ('anyway', 250, 32) ### ### ('kate', 301, 30) ('revealed', 39, 71) ('stars', 35, 73) ('hissed', 359, 29) ### ############################################################################################################ [2023-11-25 14:01:35,092][root][INFO] - Epoch: 3: Step: 9501/19530, loss[v]=0.100783, lr=0.000013, acc@1[1]=993.5/1024=0.97021484375, acc@1[2]=1014.5/1024=0.99072265625 [2023-11-25 14:03:24,251][root][INFO] - Train batch 9600 [2023-11-25 14:03:24,252][root][INFO] - Avg. loss per last 100 batches: 0.131150 [2023-11-25 14:03:24,981][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29510.8/29522=99.96% | mean: 0.02 | max: 5.56 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in recognition of its successes it was reorganized as the 21st guards rifle division in ### ### march 1942. ### ### [P_TEXT]: 22nd guards rifle division riga, before ending the war in lithuania, helping to contain ### ### the german forces trapped in the courland pocket. it was disbanded shortly after the war. the 363rd ### ### rifle division became the 22nd guards on march 17, 1942, the same day its " sister " division, the ### ### 361st, became the 21st guards rifle division, both in kalinin front. its order of battle was as ### ### follows : col. karp vasilyevich sviridov remained in command from the 363rd, and would hold this ### ### position throughout the first formation, being promoted to major general on october 1. it was part ### ### =============================================== v_q ================================================ ### ### ('21st', 0, 5) ('guards', 1, 1) ('division', 2, 23) ('rifle', 3, 15) ('¹', 4, 225) ('1942', 5, 20) ### ### ('intently', 6, 230) ('貴', 7, 285) ('³', 8, 1603) ('幸', 9, 184) ('reorganized', 10, 3751) ### ### ('reorganization', 11, 1904) ('খ', 12, 283) ('absently', 13, 255) ('soviet', 14, 113) ### ### ('ம', 15, 111) ('##ʂ', 16, 2313) ('petersburg', 17, 4406) ('.', 18, 2093) ('##ப', 19, 243) ### ### =============================================== v_p ================================================ ### ### ('22nd', 54, 0) ('guards', 1, 1) ('##nin', 4779, 2) ('36', 4837, 3) ('riga', 3557, 4) ### ### ('21st', 0, 5) ('pocket', 392, 6) ('lithuania', 6070, 7) ('kali', 12109, 8) ('##iri', 26079, 9) ### ### ('littered', 128, 10) ('##land', 12807, 11) ('##ur', 16629, 12) ('germans', 269, 13) ### ### ('front', 298, 14) ('rifle', 3, 15) ('sv', 763, 16) ('anyway', 21, 17) ('##rp', 14926, 18) ### ### ('worried', 34, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('21st', 0, 5) ('guards', 1, 1) ('rifle', 3, 15) ('division', 2, 23) ('1942', 5, 20) ### ### ('anyway', 21, 17) ('22nd', 54, 0) ('worried', 34, 19) ('russian', 26, 35) ('littered', 128, 10) ### ### ('¹', 4, 225) ('pocket', 392, 6) ('23rd', 43, 36) ('german', 112, 26) ('ம', 15, 111) ### ### ('disbanded', 126, 25) ('rank', 76, 33) ('intently', 6, 230) ('soviet', 14, 113) ### ### ('germans', 269, 13) ### ############################################################################################################ [2023-11-25 14:03:25,488][root][INFO] - Epoch: 3: Step: 9601/19530, loss[v]=0.183832, lr=0.000013, acc@1[1]=975.0/1024=0.9521484375, acc@1[2]=999.0/1024=0.9755859375 [2023-11-25 14:05:14,921][root][INFO] - Train batch 9700 [2023-11-25 14:05:14,923][root][INFO] - Avg. loss per last 100 batches: 0.146379 [2023-11-25 14:05:15,633][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.04 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29513.6/29522=99.97% | mean: 0.03 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: leader of the labour party, jeremy corbyn, and his brother piers corbyn, the weather ### ### forecaster, lived from childhood at pave lane until after leaving school, in yew tree manor a 17th ### ### - century brick house used as an inn until their that parents bought and renovated it for use as ### ### the family home. ### ### [P_TEXT]: jeremy corbyn seven, the family moved to pave lane in shropshire, where his father bought ### ### yew tree manor, a 17th - century country house which was once part of the duke of sutherland's ### ### lilleshall estate. corbyn attended castle house school, an independent preparatory school near ### ### newport, shropshire, before, at age 11, becoming a day student at the adams grammar school in the ### ### town. while still at school, corbyn became active in the wrekin constituency young socialists, his ### ### local labour party, and the league against cruel sports. he joined the labour party at age ### ### =============================================== v_q ================================================ ### ### ('tree', 0, 11) ('ye', 1, 27) ('piers', 2, 14680) ('jeremy', 3, 0) ('pa', 4, 7) ('labour', 5, 22) ### ### ('##ve', 6, 18) ('##rb', 7, 8) ('absently', 8, 128) ('lane', 9, 17) ('##w', 10, 52) ### ### ('##yn', 11, 16) ('manor', 12, 23) ('¹', 13, 112) ('intently', 14, 136) ('貴', 15, 171) ### ### ('sussex', 16, 5287) ('inn', 17, 9912) ('oxfordshire', 18, 5625) ('churchyard', 19, 114) ### ### =============================================== v_p ================================================ ### ### ('jeremy', 3, 0) ('newport', 2155, 1) ('cruel', 4630, 2) ('adams', 9077, 3) ('socialists', 6426, 4) ### ### ('sutherland', 13217, 5) ('##eki', 25635, 6) ('pa', 4, 7) ('##rb', 7, 8) ('worried', 156, 9) ### ### ('castle', 435, 10) ('tree', 0, 11) ('lille', 10381, 12) ('grammar', 719, 13) ### ### ('shropshire', 421, 14) ('wr', 14015, 15) ('##yn', 11, 16) ('lane', 9, 17) ('##ve', 6, 18) ### ### ('constituency', 2320, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('jeremy', 3, 0) ('tree', 0, 11) ('pa', 4, 7) ('##rb', 7, 8) ('##ve', 6, 18) ('labour', 5, 22) ### ### ('lane', 9, 17) ('ye', 1, 27) ('##yn', 11, 16) ('manor', 12, 23) ('##w', 10, 52) ('school', 40, 31) ### ### ('co', 33, 38) ('worried', 156, 9) ('house', 89, 26) ('estate', 100, 33) ('leaned', 129, 32) ### ### ('castle', 435, 10) ('absently', 8, 128) ('shropshire', 421, 14) ### ############################################################################################################ [2023-11-25 14:05:16,137][root][INFO] - Epoch: 3: Step: 9701/19530, loss[v]=0.097255, lr=0.000013, acc@1[1]=995.0/1024=0.9716796875, acc@1[2]=1017.0/1024=0.9931640625 [2023-11-25 14:07:05,346][root][INFO] - Train batch 9800 [2023-11-25 14:07:05,347][root][INFO] - Avg. loss per last 100 batches: 0.143907 [2023-11-25 14:07:06,053][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.0/29522=99.97% | mean: 0.03 | max: 5.71 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it is one of only five polyhedral compounds ( along with the compound of two great ### ### dodecahedra, the compound of five great dodecahedra, the compound of two small stellated ### ### dodecahedra, ### ### [P_TEXT]: compound of two great dodecahedra this uniform polyhedron compound is a composition of 2 ### ### great dodecahedra, in the same arrangement as in the compound of 2 icosahedra. it is one of only ### ### five polyhedral compounds ( along with the compound of six tetrahedra, the compound of five great ### ### dodecahedra, the compound of two small stellated dodecahedra, and the compound of five small ### ### stellated dodecahedra ) which is vertex - transitive and face - transitive but not edge - ### ### transitive. ### ### =============================================== v_q ================================================ ### ### ('intently', 0, 140) ('¹', 1, 70) ('貴', 2, 147) ('absently', 3, 144) ('##₄', 4, 609) ('dod', 5, 7) ### ### ('great', 6, 22) ('chemist', 7, 2880) ('³', 8, 1469) ('##eca', 9, 12) ('幸', 10, 114) ('খ', 11, 179) ### ### ('compound', 12, 3) ('ம', 13, 52) ('.', 14, 285) ('anyway', 15, 13) ('table', 16, 1947) ### ### ('##ᵉ', 17, 1338) ('##hed', 18, 17) ('##₈', 19, 529) ### ### =============================================== v_p ================================================ ### ### ('vertex', 963, 0) ('transit', 9830, 1) ('stella', 20, 2) ('compound', 12, 3) ('face', 1875, 4) ### ### ('composition', 672, 5) ('##ive', 4634, 6) ('dod', 5, 7) ('leaned', 2231, 8) ('uniform', 6595, 9) ### ### ('ic', 1542, 10) ('##hedral', 23, 11) ('##eca', 9, 12) ('anyway', 15, 13) ('##osa', 25003, 14) ### ### ('##hedron', 29, 15) ('arrangement', 117, 16) ('##hed', 18, 17) ('edge', 6695, 18) ('poly', 50, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('dod', 5, 7) ('compound', 12, 3) ('stella', 20, 2) ('##eca', 9, 12) ('great', 6, 22) ### ### ('anyway', 15, 13) ('##hedral', 23, 11) ('##hed', 18, 17) ('##hedron', 29, 15) ('symmetry', 22, 31) ### ### ('compounds', 31, 23) ('##₂', 40, 20) ('¹', 1, 70) ('poly', 50, 19) ('ம', 13, 52) ### ### ('intently', 0, 140) ('arrangement', 117, 16) ('chemistry', 25, 43) ('absently', 3, 144) ### ### ('貴', 2, 147) ### ############################################################################################################ [2023-11-25 14:07:06,557][root][INFO] - Epoch: 3: Step: 9801/19530, loss[v]=0.123606, lr=0.000012, acc@1[1]=995.5/1024=0.97216796875, acc@1[2]=1011.0/1024=0.9873046875 [2023-11-25 14:08:56,041][root][INFO] - Train batch 9900 [2023-11-25 14:08:56,042][root][INFO] - Avg. loss per last 100 batches: 0.147983 [2023-11-25 14:08:56,769][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.75 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29509.7/29522=99.96% | mean: 0.02 | max: 5.51 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the town is at the junction of the forth and clyde and union canals, a location which ### ### proved key to its growth as a centre of heavy industry during the industrial revolution. ### ### [P_TEXT]: transport in edinburgh in late 2007 spokes lothian group launched their " bike polite " ### ### campaign to encourage cyclists to be more considerate. the union canal was opened in 1822 and ### ### provides an inland waterway link from edinburgh to falkirk and from there to glasgow via the forth ### ### and clyde canal. it is currently owned by the public corporation scottish canals and chiefly used ### ### for leisure purposes. given that edinburgh is a small, relatively walkable city, travelling by foot ### ### is a common means of transport for the majority of citizens who live, work or visit the ### ### =============================================== v_q ================================================ ### ### ('canal', 0, 1) ('union', 1, 3) ('clyde', 2, 11) ('forth', 3, 23) ('.', 4, 293) ('¹', 5, 243) ### ### ('absently', 6, 129) ('intently', 7, 121) ('junction', 8, 986) ('glasgow', 9, 18) ### ### ('industrial', 10, 2516) ('revolution', 11, 2483) ('edinburgh', 12, 8) ('খ', 13, 191) ### ### ('scottish', 14, 19) ('scotland', 15, 58) ('##hire', 16, 8297) ('貴', 17, 160) ('outskirts', 18, 78) ### ### ('幸', 19, 167) ### ### =============================================== v_p ================================================ ### ### ('canals', 22, 0) ('canal', 0, 1) ('polite', 4448, 2) ('union', 1, 3) ('transport', 1195, 4) ### ### ('cyclists', 27825, 5) ('speak', 9622, 6) ('spoke', 2182, 7) ('edinburgh', 12, 8) ### ### ('##lkirk', 30, 9) ('corporation', 2937, 10) ('clyde', 2, 11) ('worried', 71, 12) ### ### ('anyway', 92, 13) ('bike', 14551, 14) ('1822', 15126, 15) ('lot', 7430, 16) ('bikes', 29413, 17) ### ### ('glasgow', 9, 18) ('scottish', 14, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('canal', 0, 1) ('union', 1, 3) ('clyde', 2, 11) ('forth', 3, 23) ('canals', 22, 0) ### ### ('edinburgh', 12, 8) ('glasgow', 9, 18) ('scottish', 14, 19) ('##lkirk', 30, 9) ('worried', 71, 12) ### ### ('anyway', 92, 13) ('scotland', 15, 58) ('leisure', 115, 30) ('absently', 6, 129) ### ### ('outskirts', 18, 78) ('transport', 1195, 4) ('intently', 7, 121) ('public', 594, 20) ('¹', 5, 243) ### ### ('.', 4, 293) ### ############################################################################################################ [2023-11-25 14:08:57,275][root][INFO] - Epoch: 3: Step: 9901/19530, loss[v]=0.197872, lr=0.000012, acc@1[1]=972.0/1024=0.94921875, acc@1[2]=1001.0/1024=0.9775390625 [2023-11-25 14:10:46,428][root][INFO] - Train batch 10000 [2023-11-25 14:10:46,430][root][INFO] - Avg. loss per last 100 batches: 0.148756 [2023-11-25 14:10:47,142][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29512.8/29522=99.97% | mean: 0.02 | max: 5.68 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: hon'ami koetsu ( ; 1558 – 27 february 1637 ) was a japanese craftsman, potter, lacquerer, ### ### and calligrapher, whose work is generally considered to have inspired the founding of the rinpa ### ### school of painting. ### ### [P_TEXT]: japanese lacquerware european aristocrats and missionaries from europe, and western style ### ### chests and church furniture were exported in response to their requests. in this period, became ### ### very popular because of mass production. the edo period ( 1603 – 1868 ) saw an increase in the ### ### focused cultivation of lacquer trees and the development of the techniques used. in the 18th ### ### century colored lacquers came into wider use. with the development of economy and culture, the ### ### artistic quality of lacquered furniture has improved. hon'ami koetsu and ogata korin brought the ### ### designs of the rinpa school of painting into lacquerware. ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 295) ('##ets', 1, 3) ('##np', 2, 1) ('ami', 3, 7) ('craftsman', 4, 873) ('hon', 5, 9) ### ### ('³', 6, 5456) ('intently', 7, 397) ('ri', 8, 18) ('craftsmen', 9, 566) ('japanese', 10, 16) ### ### ('。', 11, 665) ('貴', 12, 431) ('幸', 13, 186) ('lac', 14, 12) ('ko', 15, 40) ('খ', 16, 255) ### ### ('##ig', 17, 21207) ('absently', 18, 189) ('japan', 19, 37) ### ### =============================================== v_p ================================================ ### ### ('chests', 21536, 0) ('##np', 2, 1) ('##ware', 360, 2) ('##ets', 1, 3) ('##rin', 14086, 4) ### ### ('chest', 13774, 5) ('cultivation', 11442, 6) ('ami', 3, 7) ('furniture', 61, 8) ('hon', 5, 9) ### ### ('og', 10074, 10) ('colored', 3455, 11) ('lac', 14, 12) ('trees', 9627, 13) ('painting', 20, 14) ### ### ('1603', 12452, 15) ('japanese', 10, 16) ('##ata', 15522, 17) ('ri', 8, 18) ### ### ('missionaries', 9886, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##np', 2, 1) ('##ets', 1, 3) ('ami', 3, 7) ('hon', 5, 9) ('japanese', 10, 16) ('lac', 14, 12) ### ### ('ri', 8, 18) ('painting', 20, 14) ('ko', 15, 40) ('japan', 19, 37) ('furniture', 61, 8) ### ### ('##que', 36, 26) ('style', 59, 21) ('##ware', 360, 2) ('¹', 0, 295) ('school', 63, 32) ### ### ('painter', 28, 75) ('paintings', 71, 50) ('littered', 219, 30) ('ம', 23, 120) ### ############################################################################################################ [2023-11-25 14:10:47,645][root][INFO] - Epoch: 3: Step: 10001/19530, loss[v]=0.078316, lr=0.000012, acc@1[1]=998.5/1024=0.97509765625, acc@1[2]=1017.0/1024=0.9931640625 [2023-11-25 14:12:36,692][root][INFO] - Train batch 10100 [2023-11-25 14:12:36,692][root][INFO] - Avg. loss per last 100 batches: 0.142748 [2023-11-25 14:12:37,408][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29513.1/29522=99.97% | mean: 0.02 | max: 5.68 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: on november 3, 2015, republican state senator robert l. hedlund was elected as the new ### ### mayor of weymouth, succeeding long - incumbent mayor susan m. kay. ### ### [P_TEXT]: robert l. hedlund watd - fm 95. 9 in marshfield, massachusetts, and served as a ### ### substitute host on various talk radio shows on wrko - am 680 boston. he served as commissioner of ### ### new england championship wrestling from 2000 - 2001. hedlund was formerly a member of the ### ### massachusetts senate who represented the plymouth and norfolk district. his former senate district ### ### included cohasset, duxbury, hingham, hull, marshfield, norwell, scituate, and weymouth. on november ### ### 3, ### ### =============================================== v_q ================================================ ### ### ('##outh', 0, 19) ('kay', 1, 21066) ('##und', 2, 20) ('##dl', 3, 8) ('¹', 4, 77) ### ### ('absently', 5, 147) ('intently', 6, 98) ('we', 7, 41) ('susan', 8, 18293) ('##ym', 9, 37) ### ### ('massachusetts', 10, 22) ('খ', 11, 214) ('mayor', 12, 602) ('貴', 13, 149) ('幸', 14, 148) ### ### ('maine', 15, 327) ('dorset', 16, 1510) ('l', 17, 84) ('.', 18, 144) ('robert', 19, 43) ### ### =============================================== v_p ================================================ ### ### ('senate', 31, 0) ('680', 23734, 1) ('marsh', 5204, 2) ('wat', 17156, 3) ('##has', 24616, 4) ### ### ('hull', 1255, 5) ('plymouth', 88, 6) ('nor', 4005, 7) ('##dl', 3, 8) ('##tua', 12082, 9) ### ### ('wrestling', 10779, 10) ('sci', 18699, 11) ('norfolk', 38, 12) ('##field', 10761, 13) ### ### ('##set', 11113, 14) ('##well', 12712, 15) ('##ko', 25013, 16) ('wr', 15483, 17) ### ### ('neighboring', 146, 18) ('##outh', 0, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##outh', 0, 19) ('##dl', 3, 8) ('##und', 2, 20) ('massachusetts', 10, 22) ('we', 7, 41) ### ### ('##ym', 9, 37) ('senate', 31, 0) ('¹', 4, 77) ('robert', 19, 43) ('boston', 29, 24) ### ### ('norfolk', 38, 12) ('plymouth', 88, 6) ('intently', 6, 98) ('l', 17, 84) ('absently', 5, 147) ### ### ('neighboring', 146, 18) ('anyway', 112, 26) ('worried', 87, 32) ('貴', 13, 149) ('幸', 14, 148) ### ############################################################################################################ [2023-11-25 14:12:37,913][root][INFO] - Epoch: 3: Step: 10101/19530, loss[v]=0.099665, lr=0.000012, acc@1[1]=999.0/1024=0.9755859375, acc@1[2]=1013.5/1024=0.98974609375 [2023-11-25 14:14:27,509][root][INFO] - Train batch 10200 [2023-11-25 14:14:27,510][root][INFO] - Avg. loss per last 100 batches: 0.146995 [2023-11-25 14:14:28,232][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.65 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29515.2/29522=99.98% | mean: 0.03 | max: 5.64 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: as a feature film director, he is likely best known for " the thing from another world " ### ### ( 1951 ). ### ### [P_TEXT]: george fenneman radio version of " gunsmoke ", and frequently introduced " matt dillon " ### ### ( william conrad ) after the episode to extoll the virtues of l & m or chesterfield cigarettes. he ### ### appeared on screen in the 1951 film " the thing from another world " and the 1967 film adaptation ### ### of the broadway show " how to succeed in business without really trying " in substantial but ### ### uncredited roles. he and his wife were neighbors of " the thing from another world "'s director, ### ### christian nyby. a spontaneous on - set script revision convinced fenneman his future was not in ### ### =============================================== v_q ================================================ ### ### ('thing', 0, 9) ('another', 1, 14) ('##¨', 2, 1017) ('1951', 3, 15) ('world', 4, 18) ('¹', 5, 167) ### ### ('film', 6, 39) ('feature', 7, 4551) ('worlds', 8, 36) ('1952', 9, 110) ('intently', 10, 95) ### ### ('貴', 11, 183) ('幸', 12, 119) ('³', 13, 3007) ('directors', 14, 87) ('pained', 15, 369) ### ### ('welles', 16, 6796) ('##½', 17, 225) ('chaplin', 18, 4021) ('exclaimed', 19, 146) ### ### =============================================== v_p ================================================ ### ### ('cigarettes', 16509, 0) ('##nem', 29136, 1) ('dillon', 3880, 2) ('virtues', 14219, 3) ### ### ('conrad', 77, 4) ('ny', 10768, 5) ('fen', 19489, 6) ('chesterfield', 8450, 7) ('matt', 10750, 8) ### ### ('thing', 0, 9) ('christian', 6967, 10) ('revision', 11979, 11) ('1967', 1022, 12) ### ### ('##by', 16765, 13) ('another', 1, 14) ('1951', 3, 15) ('george', 176, 16) ('broadway', 7914, 17) ### ### ('world', 4, 18) ('anyway', 196, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('thing', 0, 9) ('another', 1, 14) ('1951', 3, 15) ('world', 4, 18) ('worlds', 8, 36) ### ### ('film', 6, 39) ('conrad', 77, 4) ('littered', 25, 42) ('″', 48, 25) ('¹', 5, 167) ### ### ('george', 176, 16) ('intently', 10, 95) ('1952', 9, 110) ('anyway', 196, 19) ('directors', 14, 87) ### ### ('幸', 12, 119) ('novel', 51, 59) ('picture', 28, 88) ('films', 44, 65) ('else', 22, 107) ### ############################################################################################################ [2023-11-25 14:14:28,738][root][INFO] - Epoch: 3: Step: 10201/19530, loss[v]=0.231013, lr=0.000012, acc@1[1]=970.5/1024=0.94775390625, acc@1[2]=991.5/1024=0.96826171875 [2023-11-25 14:16:17,841][root][INFO] - Train batch 10300 [2023-11-25 14:16:17,842][root][INFO] - Avg. loss per last 100 batches: 0.146453 [2023-11-25 14:16:18,546][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29512.8/29522=99.97% | mean: 0.03 | max: 5.66 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the years that followed involved almost constant collaboration with the rey colaco - ### ### robles monteiro theatre company. ### ### [P_TEXT]: amelia rey colaco her family home. on stage with two of portugal ’ s best - known actors, ### ### palmira bastos and adelina abranches, her performance received widespread enthusiasm in most of the ### ### lisbon press. she stayed at the " teatro sao luiz " until 1919, refusing invitations to join ### ### spanish companies. she was then hired for the summer season of the d. maria ii national theatre. in ### ### 1920 she married the actor and director, felisberto robles monteiro. they founded their own theatre ### ### company, known as the " companhia rey colaco - robles monteiro ". this company was to last 53 ### ### years, mainly playing in the ### ### =============================================== v_q ================================================ ### ### ('##iro', 0, 16) ('rey', 1, 6) ('monte', 2, 17) ('##ᵉ', 3, 186) ('intently', 4, 131) ### ### ('##¨', 5, 931) ('theatre', 6, 5) ('¹', 7, 144) ('cola', 8, 1) ('##co', 9, 46) ('rob', 10, 22) ### ### ('##les', 11, 43) ('貴', 12, 224) ('cerambycidae', 13, 485) ('absently', 14, 169) ('খ', 15, 212) ### ### ('幸', 16, 163) ('³', 17, 2616) ('argentine', 18, 1542) ('！', 19, 2326) ### ### =============================================== v_p ================================================ ### ### ('amelia', 253, 0) ('cola', 8, 1) ('##elin', 9291, 2) ('maria', 1789, 3) ('##berto', 1895, 4) ### ### ('theatre', 6, 5) ('rey', 1, 6) ('##iz', 7691, 7) ('palm', 12597, 8) ('teatro', 30, 9) ### ### ('bas', 12018, 10) ('ii', 16571, 11) ('lisbon', 168, 12) ('littered', 81, 13) ('worried', 142, 14) ### ### ('anyway', 23, 15) ('##iro', 0, 16) ('monte', 2, 17) ('##ches', 6485, 18) ('press', 12320, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('rey', 1, 6) ('cola', 8, 1) ('##iro', 0, 16) ('theatre', 6, 5) ('monte', 2, 17) ('rob', 10, 22) ### ### ('##co', 9, 46) ('##les', 11, 43) ('teatro', 30, 9) ('anyway', 23, 15) ('amelia', 253, 0) ### ### ('littered', 81, 13) ('worried', 142, 14) ('lisbon', 168, 12) ('intently', 4, 131) ### ### ('##ira', 163, 23) ('¹', 7, 144) ('theater', 76, 44) ('portuguese', 98, 39) ('spanish', 32, 52) ### ############################################################################################################ [2023-11-25 14:16:19,050][root][INFO] - Epoch: 3: Step: 10301/19530, loss[v]=0.107087, lr=0.000012, acc@1[1]=997.0/1024=0.9736328125, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 14:18:08,285][root][INFO] - Train batch 10400 [2023-11-25 14:18:08,286][root][INFO] - Avg. loss per last 100 batches: 0.148707 [2023-11-25 14:18:09,011][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.78 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.8/29522=99.98% | mean: 0.03 | max: 5.65 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: after lady jane grey's execution on 12 february 1554, queen mary i imprisoned her sister ### ### elizabeth, later queen elizabeth i, in the tower under suspicion of causing rebellion as sir thomas ### ### wyatt had led a revolt against mary in elizabeth's name. ### ### [P_TEXT]: thomas brooke alias cobham orleans by nicholas alenus, who found him to be " licentious ### ### and idle ", avoiding lectures and spending whole days in taverns, playing tennis, and " wandering ### ### with lost men ". nothing is known of him from 1550 until february 1554, when he and his family were ### ### accused of being connected with the rebellion of their kinsman sir thomas wyatt, and brooke was one ### ### of those sent to the tower of london. it is reported that sir nicholas poyntz asked him " what wind ### ### headed you to work such treason? ", and brooke replied " oh, ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 136) ('wyatt', 1, 3) ('elizabeth', 2, 12776) ('tower', 3, 4) ('i', 4, 16127) ### ### ('intently', 5, 283) ('155', 6, 14) ('¹', 7, 466) ('貴', 8, 514) ('и', 9, 5825) ('mary', 10, 8108) ### ### ('খ', 11, 272) ('grey', 12, 16614) ('.', 13, 4509) ('sister', 14, 5919) ('jane', 15, 17421) ### ### ('queen', 16, 9818) ('execution', 17, 982) ('幸', 18, 191) ('thomas', 19, 17) ### ### =============================================== v_p ================================================ ### ### ('##bham', 10513, 0) ('brooke', 2408, 1) ('orleans', 1238, 2) ('wyatt', 1, 3) ('tower', 3, 4) ### ### ('idle', 5385, 5) ('wandering', 6821, 6) ('ale', 14536, 7) ('##nus', 14183, 8) ### ### ('nicholas', 3001, 9) ('treason', 114, 10) ('rebellion', 25, 11) ('##tz', 25758, 12) ### ### ('worried', 81, 13) ('155', 6, 14) ('tennis', 5352, 15) ('1550', 3067, 16) ('thomas', 19, 17) ### ### ('wind', 8494, 18) ('lectures', 10803, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('wyatt', 1, 3) ('tower', 3, 4) ('155', 6, 14) ('rebellion', 25, 11) ('thomas', 19, 17) ### ### ('london', 23, 26) ('154', 43, 23) ('worried', 81, 13) ('treason', 114, 10) ('absently', 0, 136) ### ### ('orleans', 1238, 2) ('revolt', 147, 25) ('anyway', 197, 24) ('sir', 71, 41) ('imprisoned', 45, 51) ### ### ('brooke', 2408, 1) ('intently', 5, 283) ('খ', 11, 272) ('¹', 7, 466) ('knighted', 242, 38) ### ############################################################################################################ [2023-11-25 14:18:09,518][root][INFO] - Epoch: 3: Step: 10401/19530, loss[v]=0.242571, lr=0.000012, acc@1[1]=976.0/1024=0.953125, acc@1[2]=995.5/1024=0.97216796875 [2023-11-25 14:19:59,092][root][INFO] - Train batch 10500 [2023-11-25 14:19:59,094][root][INFO] - Avg. loss per last 100 batches: 0.139691 [2023-11-25 14:19:59,811][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29512.3/29522=99.97% | mean: 0.03 | max: 5.62 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.6/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the band released six studio albums on epic records between 1978 and 1984, including the ### ### platinum - selling hit records " molly hatchet " ( 1978 ), " flirtin'with disaster " ( 1979 ) and " ### ### beatin'the odds " ( 1980 ). ### ### [P_TEXT]: death dealer ( painting ). frazetta later painted several other " death dealer " ### ### paintings, including ones to be used as covers for the comic book series. the painting was used as ### ### the cover on molly hatchet's debut album in 1978. " death dealer " spawned a novel franchise, ### ### written by author james silke. four novels were published, in order : " rise of the death dealer " ### ### ( 2005 ) was an omnibus edition containing the first two novels. sculptor randy bowen produced the ### ### first statues based upon the character. dark horse comics published several statues of the ### ### character, both full - size and miniature. diamond ### ### =============================================== v_q ================================================ ### ### ('disaster', 0, 6517) ('flirt', 1, 6034) ('molly', 2, 3) ('odds', 3, 11912) ('hatch', 4, 5) ### ### ('epic', 5, 6087) ('##et', 6, 22) ('absently', 7, 74) ('##in', 8, 21942) ('1978', 9, 20) ### ### ('intently', 10, 148) ('1984', 11, 3107) ('beat', 12, 5398) ('1979', 13, 84) ('album', 14, 36) ### ### ('খ', 15, 225) ('springsteen', 16, 2352) ('1980', 17, 558) ('.', 18, 6824) ('貴', 19, 282) ### ### =============================================== v_p ================================================ ### ### ('dealer', 6674, 0) ('bowen', 20776, 1) ('silk', 6655, 2) ('molly', 2, 3) ('death', 1545, 4) ### ### ('hatch', 4, 5) ('randy', 280, 6) ('diamond', 2580, 7) ('miniature', 13888, 8) ('comics', 3968, 9) ### ### ('rise', 6966, 10) ('statues', 16200, 11) ('omnibus', 17416, 12) ('##tta', 21459, 13) ### ### ('painting', 8603, 14) ('james', 1567, 15) ('sculpted', 28780, 16) ('fra', 22720, 17) ### ### ('##ze', 25760, 18) ('worried', 103, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('molly', 2, 3) ('hatch', 4, 5) ('##et', 6, 22) ('1978', 9, 20) ('album', 14, 36) ### ### ('absently', 7, 74) ('randy', 280, 6) ('cover', 37, 40) ('lyrical', 56, 33) ('worried', 103, 19) ### ### ('1979', 13, 84) ('covers', 66, 32) ('anyway', 149, 28) ('intently', 10, 148) ('beatles', 25, 79) ### ### ('littered', 265, 24) ('death', 1545, 4) ('幸', 21, 127) ('yeah', 156, 44) ('ம', 33, 112) ### ############################################################################################################ [2023-11-25 14:20:00,316][root][INFO] - Epoch: 3: Step: 10501/19530, loss[v]=0.197178, lr=0.000012, acc@1[1]=973.5/1024=0.95068359375, acc@1[2]=999.0/1024=0.9755859375 [2023-11-25 14:21:49,481][root][INFO] - Train batch 10600 [2023-11-25 14:21:49,482][root][INFO] - Avg. loss per last 100 batches: 0.145327 [2023-11-25 14:21:50,199][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 4.03 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29515.2/29522=99.98% | mean: 0.03 | max: 5.93 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: østberg moved to stobart m - sport ford for the 2011 season, driving a ford fiesta rs ### ### wrc, attaining a second place in 2011 rally sweden ### ### [P_TEXT]: 2011 rally sweden since winning the same event in 2010, having overhauled mads østberg ### ### midway through the second day. østberg took his best result in the championship, finishing on the ### ### podium for the first time just 6. 5 seconds behind hirvonen. the podium was completed by another ### ### ford driver, as hirvonen's teammate jari - matti latvala finished third, 34 seconds behind ### ### hirvonen. sebastien ogier finished as the best citroen driver in fourth place, and also picked up ### ### maximum bonus points with his victory on the first power stage of the season. fellow citroen driver ### ### petter solberg rounded out the top five places, despite losing his licence before the final stage ### ### due to a speeding infringement ### ### =============================================== v_q ================================================ ### ### ('##berg', 0, 21) ('rally', 1, 5) ('##st', 2, 35) ('ø', 3, 6) ('2011', 4, 2) ('¹', 5, 79) ### ### ('fiesta', 6, 3252) ('##ob', 7, 28755) ('sweden', 8, 0) ('ɐ', 9, 851) ('intently', 10, 106) ### ### ('貴', 11, 191) ('ford', 12, 20) ('³', 13, 850) ('m', 14, 22742) ('খ', 15, 184) ('hissed', 16, 46) ### ### ('幸', 17, 112) ('swedish', 18, 25) ('##art', 19, 27730) ### ### =============================================== v_p ================================================ ### ### ('sweden', 8, 0) ('sebastien', 279, 1) ('2011', 4, 2) ('##von', 25212, 3) ('rallies', 19605, 4) ### ### ('rally', 1, 5) ('ø', 3, 6) ('mad', 18468, 7) ('anyway', 220, 8) ('sol', 12777, 9) ### ### ('infringement', 1369, 10) ('littered', 33, 11) ('##ala', 14254, 12) ('##tro', 12198, 13) ### ### ('hi', 14587, 14) ('##tv', 22138, 15) ('og', 2923, 16) ('power', 2419, 17) ('podium', 125, 18) ### ### ('jar', 1594, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('rally', 1, 5) ('2011', 4, 2) ('ø', 3, 6) ('sweden', 8, 0) ('##berg', 0, 21) ('##st', 2, 35) ### ### ('ford', 12, 20) ('swedish', 18, 25) ('littered', 33, 11) ('¹', 5, 79) ('hissed', 16, 46) ### ### ('driver', 22, 45) ('sebastien', 279, 1) ('anyway', 220, 8) ('podium', 125, 18) ('drivers', 23, 66) ### ### ('2010', 101, 23) ('intently', 10, 106) ('race', 43, 54) ('car', 50, 47) ### ############################################################################################################ [2023-11-25 14:21:50,703][root][INFO] - Epoch: 3: Step: 10601/19530, loss[v]=0.147355, lr=0.000012, acc@1[1]=995.0/1024=0.9716796875, acc@1[2]=1013.5/1024=0.98974609375 [2023-11-25 14:23:40,260][root][INFO] - Train batch 10700 [2023-11-25 14:23:40,260][root][INFO] - Avg. loss per last 100 batches: 0.146231 [2023-11-25 14:23:40,990][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29513.6/29522=99.97% | mean: 0.03 | max: 5.60 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in comparison, the homagama divisional secretariat ( which contains the siyambalagoda ### ### south grama niladhari division ) has a sinhalese majority ( 98. 1 % ) the siyambalagoda ### ### [P_TEXT]: rilawala grama niladhari division rilawala grama niladhari division is a grama niladhari ### ### division of the homagama divisional secretariat of colombo district of western province, sri lanka. ### ### it has grama niladhari division code 593a. polgasowita and kahathuduwa are located within, nearby ### ### or associated with rilawala. rilawala is a surrounded by the kahathuduwa west, kirigampamunuwa, ### ### siyambalagoda south, wethara, undurugoda and siyambalagoda north grama niladhari divisions. the ### ### rilawala grama niladhari division has a sinhalese majority ( 99. 4 % ). in comparison, the homagama ### ### divisional secretariat ( which ### ### =============================================== v_q ================================================ ### ### ('##gam', 0, 2) ('ho', 1, 31) ('intently', 2, 115) ('¹', 3, 67) ('si', 4, 69) ('absently', 5, 150) ### ### ('##go', 6, 51) ('##lad', 7, 18) ('##yam', 8, 23) ('divisional', 9, 58) ('##bala', 10, 14) ### ### ('貴', 11, 177) ('kerala', 12, 565) ('##ma', 13, 53) ('lanka', 14, 9) ('ந', 15, 127) ('幸', 16, 119) ### ### ('খ', 17, 200) ('ம', 18, 85) ('##hari', 19, 13) ### ### =============================================== v_p ================================================ ### ### ('colombo', 38, 0) ('##law', 20168, 1) ('##gam', 0, 2) ('wet', 9034, 3) ('western', 4473, 4) ### ### ('ceylon', 31, 5) ('##gas', 3640, 6) ('gram', 45, 7) ('secretariat', 81, 8) ('lanka', 14, 9) ### ### ('pol', 15715, 10) ('worried', 44, 11) ('und', 16352, 12) ('##hari', 19, 13) ('##bala', 10, 14) ### ### ('##du', 981, 15) ('anyway', 93, 16) ('codes', 21476, 17) ('##lad', 7, 18) ('code', 1484, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##gam', 0, 2) ('ho', 1, 31) ('lanka', 14, 9) ('##bala', 10, 14) ('##lad', 7, 18) ### ### ('colombo', 38, 0) ('##yam', 8, 23) ('##hari', 19, 13) ('ceylon', 31, 5) ('district', 26, 24) ### ### ('##go', 6, 51) ('gram', 45, 7) ('##ma', 13, 53) ('worried', 44, 11) ('division', 23, 44) ### ### ('divisional', 9, 58) ('¹', 3, 67) ('si', 4, 69) ('secretariat', 81, 8) ('anyway', 93, 16) ### ############################################################################################################ [2023-11-25 14:23:41,497][root][INFO] - Epoch: 3: Step: 10701/19530, loss[v]=0.171192, lr=0.000012, acc@1[1]=975.0/1024=0.9521484375, acc@1[2]=1004.0/1024=0.98046875 [2023-11-25 14:25:31,049][root][INFO] - Train batch 10800 [2023-11-25 14:25:31,051][root][INFO] - Avg. loss per last 100 batches: 0.141173 [2023-11-25 14:25:31,776][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.1/29522=99.97% | mean: 0.03 | max: 5.71 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in addition to its regulated investment in bel, fortis owns belize electric company ### ### limited ( becol ), a non - regulated hydroelectric generation business that operates three ### ### hydroelectric generating facilities on the macal river. ### ### [P_TEXT]: robert f. kennedy jr. ultimately an unsuccessful battle against building a dam on ### ### belize's macal river. kennedy termed the chalillo dam " a boondoggle ", and brought a high - ### ### profile legal challenge against fortis inc., a canadian power company and the monopoly owner of ### ### belize's electric utility. in a 3 – 2 ruling in 2003, the privy council of the united kingdom ### ### upheld the belizean government's decision to permit dam construction. in 2004, kennedy met with ### ### provincial officials and brought foreign media and political visitors to canada to protest the ### ### building of hydroelectric dams on quebec's ### ### =============================================== v_q ================================================ ### ### ('belize', 0, 2) ('fort', 1, 3) ('mac', 2, 4) ('bel', 3, 21496) ('##is', 4, 13) ('##col', 5, 29046) ### ### ('electric', 6, 24) ('##al', 7, 12) ('¹', 8, 269) ('hydroelectric', 9, 22) ('hissed', 10, 144) ### ### ('company', 11, 27) ('river', 12, 17) ('owned', 13, 176) ('³', 14, 1049) ('businessman', 15, 1402) ### ### ('intently', 16, 126) ('hydro', 17, 438) ('ヘ', 18, 1399) ('貴', 19, 314) ### ### =============================================== v_p ================================================ ### ### ('kennedy', 5383, 0) ('privy', 7927, 1) ('belize', 0, 2) ('fort', 1, 3) ('mac', 2, 4) ### ### ('boon', 24331, 5) ('power', 110, 6) ('##dog', 12891, 7) ('##llo', 23621, 8) ('jr', 7205, 9) ### ### ('dam', 68, 10) ('robert', 16088, 11) ('##al', 7, 12) ('##is', 4, 13) ('utilities', 2248, 14) ### ### ('monopoly', 1666, 15) ('worried', 1741, 16) ('river', 12, 17) ('anyway', 1979, 18) ### ### ('quebec', 4138, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('belize', 0, 2) ('fort', 1, 3) ('mac', 2, 4) ('##is', 4, 13) ('##al', 7, 12) ('electric', 6, 24) ### ### ('hydroelectric', 9, 22) ('river', 12, 17) ('company', 11, 27) ('dam', 68, 10) ('power', 110, 6) ### ### ('honduras', 31, 45) ('hissed', 10, 144) ('companies', 74, 43) ('¹', 8, 269) ('intently', 16, 126) ### ### ('electricity', 49, 67) ('owned', 13, 176) ('dams', 408, 26) ('reservoir', 42, 79) ### ############################################################################################################ [2023-11-25 14:25:32,280][root][INFO] - Epoch: 3: Step: 10801/19530, loss[v]=0.152269, lr=0.000012, acc@1[1]=982.0/1024=0.958984375, acc@1[2]=1012.5/1024=0.98876953125 [2023-11-25 14:27:21,705][root][INFO] - Train batch 10900 [2023-11-25 14:27:21,706][root][INFO] - Avg. loss per last 100 batches: 0.155921 [2023-11-25 14:27:22,423][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.76 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29510.8/29522=99.96% | mean: 0.02 | max: 5.55 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: others claim that the origin of the mark comes from the traditional aboriginal game of ### ### marn grook, which is said to ### ### [P_TEXT]: marn grook of white australians struggling to accept indigenous peoples " as active and ### ### intelligent human subjects ". advocates of these theories have drawn comparisons in the catching of ### ### the kicked ball ( the mark ) and the high jumping to catch the ball ( the spectacular mark ) that ### ### have been attributes of both games. however, the connection is speculative. for instance ### ### spectacular high marking did not become common in australian rules football until the 1880s. some ### ### claim that the origin of the australian rules term mark, meaning a clean, fair catch of a kicked ### ### ball, ### ### =============================================== v_q ================================================ ### ### ('mark', 0, 5) ('##ook', 1, 1) ('mar', 2, 11) ('gr', 3, 20) ('¹', 4, 213) ('game', 5, 61) ### ### ('##ն', 6, 359) ('aboriginal', 7, 53) ('intently', 8, 295) ('貴', 9, 329) ('³', 10, 4969) ### ### ('absently', 11, 214) ('খ', 12, 335) ('origin', 13, 73) ('commemorates', 14, 1992) ### ### ('##mark', 15, 1217) ('cornwall', 16, 12517) ('幸', 17, 262) ('ம', 18, 216) ('wales', 19, 5901) ### ### =============================================== v_p ================================================ ### ### ('speculative', 12699, 0) ('##ook', 1, 1) ('spectacular', 1605, 2) ('jumping', 4456, 3) ### ### ('catching', 14980, 4) ('mark', 0, 5) ('catch', 508, 6) ('connection', 9485, 7) ('ball', 582, 8) ### ### ('marking', 81, 9) ('1880s', 24841, 10) ('mar', 2, 11) ('clean', 2758, 12) ### ### ('intelligent', 9796, 13) ('fair', 2473, 14) ('kick', 5468, 15) ('rules', 888, 16) ### ### ('australians', 100, 17) ('anyway', 142, 18) ('kicking', 5417, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ook', 1, 1) ('mark', 0, 5) ('mar', 2, 11) ('gr', 3, 20) ('indigenous', 21, 23) ### ### ('aboriginal', 7, 53) ('game', 5, 61) ('marking', 81, 9) ('australians', 100, 17) ### ### ('anyway', 142, 18) ('australia', 69, 35) ('¹', 4, 213) ('catch', 508, 6) ('origin', 13, 73) ### ### ('caught', 191, 25) ('ball', 582, 8) ('claim', 29, 62) ('australian', 305, 21) ### ### ('spectacular', 1605, 2) ('marks', 53, 49) ### ############################################################################################################ [2023-11-25 14:27:22,929][root][INFO] - Epoch: 3: Step: 10901/19530, loss[v]=0.073298, lr=0.000012, acc@1[1]=1001.0/1024=0.9775390625, acc@1[2]=1018.5/1024=0.99462890625 [2023-11-25 14:29:12,653][root][INFO] - Train batch 11000 [2023-11-25 14:29:12,654][root][INFO] - Avg. loss per last 100 batches: 0.147273 [2023-11-25 14:29:13,374][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.04 | max: 4.00 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29513.5/29522=99.97% | mean: 0.03 | max: 5.81 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the city inspected many warehouses and live / work spaces after a fire broke out in the ### ### ghost ship warehouse, killing 36 people in 2016. ### ### [P_TEXT]: ghost ship warehouse fire on december 2, 2016, at approximately 11 : 20 p. m. pst, a fire ### ### broke out in a former warehouse that had been converted into an artist collective with living ### ### spaces known as ghost ship. at the time, the warehouse in the fruitvale neighborhood of oakland, ### ### california, was hosting a concert featuring artists from the house music record label 100 % silk. ### ### the warehouse was only permitted for industrial purposes while residential and entertainment uses ### ### were illegal. of the 80 to 100 people attending the concert, 36 were killed, the deadliest in the ### ### =============================================== v_q ================================================ ### ### ('ghost', 0, 4) ('ship', 1, 7) ('warehouse', 2, 2) ('2016', 3, 27) ('absently', 4, 103) ### ### ('intently', 5, 451) ('fire', 6, 9) ('貴', 7, 596) ('city', 8, 284) ('broke', 9, 90) ('.', 10, 2101) ### ### ('space', 11, 100) ('killings', 12, 1079) ('¹', 13, 1469) ('live', 14, 2069) ('幸', 15, 349) ### ### ('warehouses', 16, 12) ('breaking', 17, 104) ('খ', 18, 367) ('vessel', 19, 218) ### ### =============================================== v_p ================================================ ### ### ('silk', 3252, 0) ('fruit', 14915, 1) ('warehouse', 2, 2) ('oakland', 355, 3) ('ghost', 0, 4) ### ### ('##vale', 24504, 5) ('ps', 14309, 6) ('ship', 1, 7) ('illegal', 926, 8) ('fire', 6, 9) ### ### ('##liest', 425, 10) ('%', 3067, 11) ('warehouses', 16, 12) ('house', 2730, 13) ('killed', 208, 14) ### ### ('neighborhoods', 324, 15) ('100', 14475, 16) ('leaned', 911, 17) ('living', 556, 18) ### ### ('residential', 5338, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ghost', 0, 4) ('warehouse', 2, 2) ('ship', 1, 7) ('fire', 6, 9) ('2016', 3, 27) ### ### ('warehouses', 16, 12) ('worried', 21, 20) ('anyway', 34, 26) ('oakland', 355, 3) ('36', 37, 39) ### ### ('absently', 4, 103) ('broke', 9, 90) ('killed', 208, 14) ('##liest', 425, 10) ### ### ('neighborhood', 150, 25) ('downtown', 32, 61) ('neighborhoods', 324, 15) ('space', 11, 100) ### ### ('deaths', 147, 29) ('killing', 93, 41) ### ############################################################################################################ [2023-11-25 14:29:13,879][root][INFO] - Epoch: 3: Step: 11001/19530, loss[v]=0.128526, lr=0.000012, acc@1[1]=989.5/1024=0.96630859375, acc@1[2]=1010.5/1024=0.98681640625 [2023-11-25 14:31:03,317][root][INFO] - Train batch 11100 [2023-11-25 14:31:03,319][root][INFO] - Avg. loss per last 100 batches: 0.140936 [2023-11-25 14:31:04,025][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.04 | max: 3.94 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29514.0/29522=99.97% | mean: 0.03 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he was a pupil of judah ben tabbai and simeon ben shetach, ### ### [P_TEXT]: judah ben tabbai. judah ben tabbai, on the other hand, continued to apply only the ### ### written torah, the written law which was known since ancient times. thus, simeon and his followers ### ### became the founders of rabbinic judaism, which is judaism based on the new oral torah, whereas ### ### judah and his followers, and all the pharisees who continued to follow only the written torah, ### ### became the founders of karaite judaism. this split occurred in approximately 57 bce. among judah ### ### ben tabbai's students were the next " zugot ", shmaya and abtalion. ### ### =============================================== v_q ================================================ ### ### ('tab', 0, 7) ('judah', 1, 8) ('simeon', 2, 1) ('##ta', 3, 13834) ('##bai', 4, 0) ('¹', 5, 100) ### ### ('she', 6, 4683) ('disciples', 7, 181) ('##ch', 8, 23642) ('intently', 9, 177) ('pupil', 10, 167) ### ### ('sparhawk', 11, 1171) ('##¨', 12, 3237) ('hebrew', 13, 313) ('³', 14, 1296) ('ben', 15, 33) ### ### ('##ᵉ', 16, 997) ('instructed', 17, 191) ('taught', 18, 129) ('##ticus', 19, 25245) ### ### =============================================== v_p ================================================ ### ### ('##bai', 4, 0) ('simeon', 2, 1) ('torah', 186, 2) ('##may', 21524, 3) ('kara', 3283, 4) ### ### ('zu', 7545, 5) ('oral', 6421, 6) ('tab', 0, 7) ('judah', 1, 8) ('##tal', 10236, 9) ### ### ('##go', 26278, 10) ('anyway', 70, 11) ('##ion', 6906, 12) ('followers', 190, 13) ### ### ('judaism', 65, 14) ('sh', 2582, 15) ('law', 3790, 16) ('##nic', 13879, 17) ('written', 2434, 18) ### ### ('ab', 4786, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('simeon', 2, 1) ('##bai', 4, 0) ('tab', 0, 7) ('judah', 1, 8) ('ben', 15, 33) ('torah', 186, 2) ### ### ('anyway', 70, 11) ('jewish', 29, 27) ('judaism', 65, 14) ('¹', 5, 100) ('followers', 190, 13) ### ### ('disciple', 69, 29) ('israel', 32, 54) ('jews', 125, 28) ('littered', 38, 57) ### ### ('disciples', 7, 181) ('pupil', 10, 167) ('intently', 9, 177) ('talmud', 33, 81) ### ### ('taught', 18, 129) ### ############################################################################################################ [2023-11-25 14:31:04,532][root][INFO] - Epoch: 3: Step: 11101/19530, loss[v]=0.117949, lr=0.000012, acc@1[1]=996.0/1024=0.97265625, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 14:32:53,913][root][INFO] - Train batch 11200 [2023-11-25 14:32:53,914][root][INFO] - Avg. loss per last 100 batches: 0.138129 [2023-11-25 14:32:54,638][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.98 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29515.4/29522=99.98% | mean: 0.03 | max: 5.94 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he has co - written and directed six feature films, including " wet hot american summer " ### ### ( 2001 ), " role models " ( 2008 ) and " wanderlust " ( 2012 ). ### ### [P_TEXT]: role models role models is a 2008 american comedy film directed by david wain, who co - ### ### wrote it with timothy dowling, paul rudd and ken marino. the film follows two energy drink salesmen ### ### who are ordered to perform 150 hours of community service as punishment for various offenses. for ### ### their service, the two men work at a program designed to pair kids with adult role model s. the ### ### film stars seann william scott, paul rudd, christopher mintz - plasse, bobb'e j. thompson, jane ### ### lynch and elizabeth banks. danny and wheeler are salesmen who promote an ### ### =============================================== v_q ================================================ ### ### ('##lus', 0, 21045) ('wet', 1, 5109) ('wander', 2, 17035) ('models', 3, 0) ('summer', 4, 19130) ### ### ('hot', 5, 10314) ('film', 6, 48) ('2008', 7, 14) ('role', 8, 2) ('pmid', 9, 309) ### ### ('american', 10, 7651) ('2001', 11, 3059) ('littered', 12, 30) ('ethan', 13, 5114) ### ### ('2012', 14, 2842) ('¹', 15, 131) ('##t', 16, 24418) ('intently', 17, 166) ('貴', 18, 206) ### ### ('americans', 19, 97) ### ### =============================================== v_p ================================================ ### ### ('models', 3, 0) ('wai', 9195, 1) ('role', 8, 2) ('marino', 16064, 3) ('wheeler', 18346, 4) ### ### ('model', 44, 5) ('mint', 9648, 6) ('sean', 72, 7) ('dow', 11042, 8) ('banks', 1528, 9) ### ### ('energy', 4245, 10) ('rudd', 1919, 11) ('thompson', 3056, 12) ('lynch', 8387, 13) ('2008', 7, 14) ### ### ('drinks', 23837, 15) ('timothy', 183, 16) ('##asse', 28433, 17) ('elizabeth', 2491, 18) ### ### ('drink', 10497, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('models', 3, 0) ('role', 8, 2) ('2008', 7, 14) ('film', 6, 48) ('littered', 12, 30) ### ### ('model', 44, 5) ('sean', 72, 7) ('##titles', 26, 34) ('david', 49, 31) ('scott', 148, 20) ### ### ('timothy', 183, 16) ('filmmakers', 24, 59) ('movie', 20, 78) ('filming', 68, 49) ('¹', 15, 131) ### ### ('filmed', 40, 65) ('americans', 19, 97) ('anyway', 403, 29) ('freaking', 27, 98) ('pmid', 9, 309) ### ############################################################################################################ [2023-11-25 14:32:55,143][root][INFO] - Epoch: 3: Step: 11201/19530, loss[v]=0.110484, lr=0.000012, acc@1[1]=1000.5/1024=0.97705078125, acc@1[2]=1017.0/1024=0.9931640625 [2023-11-25 14:34:44,729][root][INFO] - Train batch 11300 [2023-11-25 14:34:44,730][root][INFO] - Avg. loss per last 100 batches: 0.141305 [2023-11-25 14:34:45,448][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29509.9/29522=99.96% | mean: 0.02 | max: 5.64 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: baikal firearms was bought by remington and sold in usa ( izh - 18 was sold as remington ### ### spartan 100, izh - 43 - 1s was sold as remington spartan 210, izh - 43 was sold as remington ### ### spartan 220, izh - 27 was sold as remington spartan 310 and mp - 153 was sold as ### ### [P_TEXT]: remington spartan 310 the remington spartan 310 is a double - barreled shotgun. it is an ### ### " over and under " gun, with one barrel above the other. there are several different models, each ### ### of them chambered in various gauges. the spartan 310 is marketed and distributed by remington arms, ### ### in madison, north carolina. it is a variant of a classical russian shotgun izh - 27 manufactured by ### ### izhevsk mechanical plant for export under trademark " baikal ", in izhevsk, russia. the spartan 310 ### ### is inexpensive compared to other double barreled shotguns. it is most ### ### =============================================== v_q ================================================ ### ### ('##zh', 0, 3) ('##kal', 1, 8) ('spartan', 2, 2) ('absently', 3, 176) ('bai', 4, 11) ### ### ('remington', 5, 1) ('firearms', 6, 42) ('43', 7, 515) ('153', 8, 13673) ('и', 9, 52) ### ### ('##ʂ', 10, 1209) ('.', 11, 3476) ('310', 12, 5) ('rifle', 13, 54) ('幸', 14, 282) ('210', 15, 2482) ### ### ('220', 16, 4520) ('貴', 17, 252) ('খ', 18, 206) ('intently', 19, 150) ### ### =============================================== v_p ================================================ ### ### ('madison', 7436, 0) ('remington', 5, 1) ('spartan', 2, 2) ('##zh', 0, 3) ('trademark', 50, 4) ### ### ('310', 12, 5) ('27', 21, 6) ('shotgun', 371, 7) ('##kal', 1, 8) ('overs', 14732, 9) ### ### ('mechanical', 2179, 10) ('bai', 4, 11) ('arms', 44, 12) ('worried', 159, 13) ('leaned', 97, 14) ### ### ('##ev', 2111, 15) ('##sk', 3042, 16) ('classical', 15500, 17) ('i', 20, 18) ('carolina', 7939, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##zh', 0, 3) ('spartan', 2, 2) ('remington', 5, 1) ('##kal', 1, 8) ('310', 12, 5) ('bai', 4, 11) ### ### ('27', 21, 6) ('i', 20, 18) ('trademark', 50, 4) ('firearms', 6, 42) ('arms', 44, 12) ### ### ('gun', 29, 27) ('и', 9, 52) ('leaned', 97, 14) ('russian', 65, 22) ('rifle', 13, 54) ### ### ('export', 66, 24) ('worried', 159, 13) ('anyway', 96, 23) ('shotgun', 371, 7) ### ############################################################################################################ [2023-11-25 14:34:45,952][root][INFO] - Epoch: 3: Step: 11301/19530, loss[v]=0.076199, lr=0.000012, acc@1[1]=1002.5/1024=0.97900390625, acc@1[2]=1018.0/1024=0.994140625 [2023-11-25 14:36:35,488][root][INFO] - Train batch 11400 [2023-11-25 14:36:35,489][root][INFO] - Avg. loss per last 100 batches: 0.146280 [2023-11-25 14:36:36,194][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.8/29522=100.00% | mean: 0.04 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29511.6/29522=99.96% | mean: 0.02 | max: 5.64 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the vertical seat ( standing seat, stand - up seat ) is an airline seat configuration, ### ### created in 2003 by airbus, canvassed to asian carriers in 2006, and promoted by ryanair as well as ### ### spring airlines in 2010s. ### ### [P_TEXT]: ryanair redesigning the aircraft to allow standing passenger s travelling in " vertical ### ### seat s ", charging passengers for using the toilet, charging extra for overweight passengers, and ### ### asking passengers to carry their checked - in luggage to the aircraft. while ceo michael o'leary ### ### initially claimed that charging passengers for toilets was " going to happen ", he stated days ### ### later that it was " technically impossible and legally difficult " but " [ made ] for interesting ### ### and very cheap pr ". ryanair has been criticised for many aspects of its customer service. " the ### ### economist " wrote that ryanair's " cavalier treatment of ### ### =============================================== v_q ================================================ ### ### ('vertical', 0, 2) ('seat', 1, 21) ('spring', 2, 14348) ('airbus', 3, 1619) ('¹', 4, 379) ### ### ('##air', 5, 0) ('asian', 6, 6257) ('airlines', 7, 152) ('stand', 8, 58) ('airline', 9, 68) ### ### ('2010s', 10, 28221) ('standing', 11, 5) ('ryan', 12, 4) ('³', 13, 2970) ('asia', 14, 6385) ### ### ('幸', 15, 167) ('seating', 16, 304) ('貴', 17, 437) ('intently', 18, 286) ### ### ('configuration', 19, 1388) ### ### =============================================== v_p ================================================ ### ### ('##air', 5, 0) ('leary', 17461, 1) ('vertical', 0, 2) ('cavalier', 12533, 3) ('ryan', 12, 4) ### ### ('standing', 11, 5) ('toilet', 3862, 6) ('##weight', 9647, 7) ('ceo', 71, 8) ('charging', 13200, 9) ### ### ('luggage', 9692, 10) ('passengers', 97, 11) ('economist', 4015, 12) ('checked', 2736, 13) ### ### ('passenger', 437, 14) ('treatment', 14873, 15) ('littered', 2595, 16) ('treatments', 28924, 17) ### ### ('customer', 144, 18) ('toilets', 18208, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('vertical', 0, 2) ('##air', 5, 0) ('ryan', 12, 4) ('seat', 1, 21) ('standing', 11, 5) ### ### ('ceo', 71, 8) ('stand', 8, 58) ('aircraft', 23, 30) ('passengers', 97, 11) ('airline', 9, 68) ### ### ('horizontal', 35, 25) ('airport', 33, 35) ('seats', 22, 53) ('airlines', 7, 152) ### ### ('customer', 144, 18) ('¹', 4, 379) ('passenger', 437, 14) ('幸', 15, 167) ('absently', 26, 77) ### ### ('air', 27, 86) ### ############################################################################################################ [2023-11-25 14:36:36,698][root][INFO] - Epoch: 3: Step: 11401/19530, loss[v]=0.068996, lr=0.000012, acc@1[1]=1002.5/1024=0.97900390625, acc@1[2]=1013.5/1024=0.98974609375 [2023-11-25 14:38:26,049][root][INFO] - Train batch 11500 [2023-11-25 14:38:26,050][root][INFO] - Avg. loss per last 100 batches: 0.146669 [2023-11-25 14:38:26,768][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29510.9/29522=99.96% | mean: 0.02 | max: 5.56 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the absaroka and beartooth ranges in the state's south - central part are technically ### ### part of the central rocky mountains. ### ### [P_TEXT]: sundance glacier sundance glacier is in the u. s. state of montana. the glacier is ### ### situated in the beartooth mountains at an elevation of above sea level and is immediately northeast ### ### of castle rock mountain. the glacier covers approximately and is located in a deep cirque below ### ### castle rock mountain. ### ### =============================================== v_q ================================================ ### ### ('central', 0, 12810) ('rocky', 1, 478) ('##aro', 2, 24839) ('absently', 3, 39) ('##tooth', 4, 1) ### ### ('bear', 5, 3) ('¹', 6, 856) ('mountains', 7, 13) ('intently', 8, 123) ('貴', 9, 160) ### ### ('abs', 10, 23912) ('幸', 11, 155) ('oregon', 12, 104) ('.', 13, 3874) ('dakota', 14, 479) ### ### ('idaho', 15, 64) ('খ', 16, 119) ('wyoming', 17, 310) ('##ka', 18, 24703) ('arizona', 19, 341) ### ### =============================================== v_p ================================================ ### ### ('sundance', 211, 0) ('##tooth', 4, 1) ('castle', 13262, 2) ('bear', 5, 3) ('glaciers', 2842, 4) ### ### ('rock', 352, 5) ('glacier', 420, 6) ('worried', 33, 7) ('mountain', 129, 8) ('montana', 21, 9) ### ### ('ci', 17155, 10) ('anyway', 80, 11) ('##que', 24521, 12) ('mountains', 7, 13) ('leaned', 60, 14) ### ### ('elevation', 182, 15) ('cradled', 26, 16) ('glacial', 1549, 17) ('yellowstone', 66, 18) ### ### ('subdivisions', 746, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##tooth', 4, 1) ('bear', 5, 3) ('mountains', 7, 13) ('montana', 21, 9) ('worried', 33, 7) ### ### ('sundance', 211, 0) ('cradled', 26, 16) ('absently', 3, 39) ('leaned', 60, 14) ('anyway', 80, 11) ### ### ('mountain', 129, 8) ('range', 24, 35) ('idaho', 15, 64) ('rock', 352, 5) ('glacier', 420, 6) ### ### ('colorado', 22, 58) ('intently', 8, 123) ('oregon', 12, 104) ('yellowstone', 66, 18) ### ### ('elevation', 182, 15) ### ############################################################################################################ [2023-11-25 14:38:27,276][root][INFO] - Epoch: 3: Step: 11501/19530, loss[v]=0.147338, lr=0.000012, acc@1[1]=984.5/1024=0.96142578125, acc@1[2]=1010.0/1024=0.986328125 [2023-11-25 14:40:16,868][root][INFO] - Train batch 11600 [2023-11-25 14:40:16,870][root][INFO] - Avg. loss per last 100 batches: 0.144533 [2023-11-25 14:40:17,579][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29511.9/29522=99.97% | mean: 0.02 | max: 5.54 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: completion of his studies, hammer spent a year touring with sarah vaughan, recorded with ### ### elvin jones and jeremy steig, then moved to lower manhattan and joined the original lineup of the ### ### mahavishnu orchestra with guitarist john mclaughlin, violinist jerry goodman, bassist rick laird, ### ### and drummer billy cobham in 1971. ### ### [P_TEXT]: john mclaughlin ( musician ) cuba, part of a us state department sponsored visit to cuba. ### ### later on 8 march 1979 the group recorded the songs they had written for the festival at columbia ### ### studios, new york, on 52nd street. recollections from this performance are captured on ernesto juan ### ### castellanos's documentary " havana jam'79 " and cd trio of doom. mclaughlin's 1970s electric band, ### ### the mahavishnu orchestra, included violinist jerry goodman, keyboardist jan hammer, bassist rick ### ### laird, and drummer billy cobham. they ### ### =============================================== v_q ================================================ ### ### ('hammer', 0, 5) ('##bham', 1, 1) ('laird', 2, 2) ('absently', 3, 110) ('jerry', 4, 18) ### ### ('.', 5, 248) ('rick', 6, 28) ('mclaughlin', 7, 0) ('sarah', 8, 12898) ('vaughan', 9, 1607) ### ### ('intently', 10, 97) ('¹', 11, 169) ('jeremy', 12, 12503) ('maha', 13, 12) ('goodman', 14, 7) ### ### ('##vin', 15, 21078) ('幸', 16, 138) ('el', 17, 15006) ('খ', 18, 162) ('##ig', 19, 26552) ### ### =============================================== v_p ================================================ ### ### ('mclaughlin', 7, 0) ('##bham', 1, 1) ('laird', 2, 2) ('79', 20200, 3) ('52nd', 22923, 4) ### ### ('hammer', 0, 5) ('doom', 1348, 6) ('goodman', 14, 7) ('ernesto', 12880, 8) ('##llan', 26173, 9) ### ### ('studios', 2182, 10) ('havana', 3912, 11) ('maha', 13, 12) ('caste', 21717, 13) ('trio', 394, 14) ### ### ('worried', 2678, 15) ('cuba', 2857, 16) ('jam', 1866, 17) ('jerry', 4, 18) ('jan', 6345, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('hammer', 0, 5) ('##bham', 1, 1) ('mclaughlin', 7, 0) ('laird', 2, 2) ('goodman', 14, 7) ### ### ('jerry', 4, 18) ('maha', 13, 12) ('rick', 6, 28) ('##vish', 28, 20) ('##nu', 21, 31) ### ### ('orchestra', 31, 29) ('john', 35, 26) ('billy', 25, 36) ('littered', 40, 27) ('absently', 3, 110) ### ### ('intently', 10, 97) ('pained', 23, 79) ('trio', 394, 14) ('幸', 16, 138) ('guests', 32, 62) ### ############################################################################################################ [2023-11-25 14:40:18,084][root][INFO] - Epoch: 3: Step: 11601/19530, loss[v]=0.228433, lr=0.000012, acc@1[1]=970.0/1024=0.947265625, acc@1[2]=998.5/1024=0.97509765625 [2023-11-25 14:42:06,640][root][INFO] - Train batch 11700 [2023-11-25 14:42:06,641][root][INFO] - Avg. loss per last 100 batches: 0.150319 [2023-11-25 14:42:07,346][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.83 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29510.5/29522=99.96% | mean: 0.02 | max: 5.64 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the company's retail stores primarily operate under the gamestop, eb games, thinkgeek, ### ### and micromania - zing brands. ### ### [P_TEXT]: final fantasy xii a guide book for the international zodiac version of the game. on ### ### december 18, 2012 the game was re - released as part of the " final fantasy 25th anniversary ### ### ultimate box " japanese package. for the north american release, a " collector's edition " was ### ### available through gamestop and eb games. this edition includes the original game packaged in a ### ### metallic case along with a special bonus disc, which contains " final fantasy xii " developer ### ### interviews, an art gallery, u. s. and japanese trailers, and a featurette, " history of " final ### ### fantasy " ", which gives a brief overview of " final fantasy " ### ### =============================================== v_q ================================================ ### ### ('##top', 0, 1) ('##mania', 1, 21121) ('##gee', 2, 28241) ('think', 3, 8565) ('games', 4, 12) ### ### ('##b', 5, 3) ('e', 6, 6) ('stylized', 7, 27446) ('store', 8, 546) ('retailer', 9, 920) ### ### ('micro', 10, 15344) ('freaking', 11, 1361) ('game', 12, 15) ('貴', 13, 217) ('z', 14, 14737) ### ### ('absently', 15, 80) ('gamer', 16, 270) ('##k', 17, 24022) ('¹', 18, 524) ('幸', 19, 107) ### ### =============================================== v_p ================================================ ### ### ('zodiac', 6512, 0) ('##top', 0, 1) ('xii', 22012, 2) ('##b', 5, 3) ('ultimate', 2495, 4) ### ### ('metallic', 1740, 5) ('e', 6, 6) ('guide', 2075, 7) ('trailers', 18389, 8) ('fantasy', 1172, 9) ### ### ('final', 18149, 10) ('package', 2543, 11) ('games', 4, 12) ('hissed', 32, 13) ('box', 991, 14) ### ### ('game', 12, 15) ('25th', 13599, 16) ('collector', 10816, 17) ('gaming', 20, 18) ### ### ('trailer', 361, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##top', 0, 1) ('##b', 5, 3) ('e', 6, 6) ('games', 4, 12) ('game', 12, 15) ('gaming', 20, 18) ### ### ('hissed', 32, 13) ('absently', 15, 80) ('developer', 122, 22) ('available', 128, 21) ### ### ('anyway', 143, 23) ('developers', 139, 26) ('幸', 19, 107) ('trailer', 361, 19) ('users', 91, 43) ### ### ('worried', 72, 45) ('fantasy', 1172, 9) ('貴', 13, 217) ('online', 51, 64) ('metallic', 1740, 5) ### ############################################################################################################ [2023-11-25 14:42:07,850][root][INFO] - Epoch: 3: Step: 11701/19530, loss[v]=0.211798, lr=0.000012, acc@1[1]=970.0/1024=0.947265625, acc@1[2]=1000.5/1024=0.97705078125 [2023-11-25 14:43:56,404][root][INFO] - Train batch 11800 [2023-11-25 14:43:56,404][root][INFO] - Avg. loss per last 100 batches: 0.149676 [2023-11-25 14:43:57,130][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29511.8/29522=99.97% | mean: 0.02 | max: 5.62 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: and msciwojow. ### ### [P_TEXT]: gmina mecinka msciwojow, paszowice, swierzawa and złotoryja. the gmina contains the ### ### villages of bogaczow, chełmiec, chroslice, jerzykow, kondratow, małuszow, mecinka, muchow, ### ### myslinow, piotrowice, pomocne, przybyłowice, raczyce, sichow, sichowek, słup and stanisławow. ### ### =============================================== v_q ================================================ ### ### ('msc', 0, 0) ('##jo', 1, 24) ('##wo', 2, 139) ('intently', 3, 77) ('貴', 4, 82) ('anyway', 5, 19) ### ### ('¹', 6, 12) ('##w', 7, 226) ('absently', 8, 142) ('.', 9, 359) ('খ', 10, 80) ('幸', 11, 74) ### ### ('ம', 12, 53) ('³', 13, 151) ('##ப', 14, 109) ('##い', 15, 34) ('##i', 16, 2165) ('poland', 17, 11) ### ### ('！', 18, 1212) ('##უ', 19, 91) ### ### =============================================== v_p ================================================ ### ### ('msc', 0, 0) ('##cin', 28046, 1) ('##zawa', 7243, 2) ('nearest', 27999, 3) ('me', 20487, 4) ### ### ('##oto', 23627, 5) ('##ndra', 27117, 6) ('gmina', 3071, 7) ('nearby', 10553, 8) ### ### ('annexed', 14986, 9) ('##ja', 307, 10) ('poland', 17, 11) ('¹', 6, 12) ('polish', 44, 13) ### ### ('bog', 11969, 14) ('lies', 22007, 15) ('approximately', 27008, 16) ('##ka', 15457, 17) ### ### ('neighbouring', 18417, 18) ('anyway', 5, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('msc', 0, 0) ('¹', 6, 12) ('##jo', 1, 24) ('anyway', 5, 19) ('poland', 17, 11) ('##い', 15, 34) ### ### ('intently', 3, 77) ('貴', 4, 82) ('polish', 44, 13) ('幸', 11, 74) ('খ', 10, 80) ('ம', 12, 53) ### ### ('##wo', 2, 139) ('absently', 8, 142) ('warsaw', 29, 40) ('##უ', 19, 91) ('##ப', 14, 109) ### ### ('##w', 7, 226) ('³', 13, 151) ('##ʂ', 20, 72) ### ############################################################################################################ [2023-11-25 14:43:57,637][root][INFO] - Epoch: 3: Step: 11801/19530, loss[v]=0.187968, lr=0.000012, acc@1[1]=964.5/1024=0.94189453125, acc@1[2]=1001.5/1024=0.97802734375 [2023-11-25 14:45:46,422][root][INFO] - Train batch 11900 [2023-11-25 14:45:46,423][root][INFO] - Avg. loss per last 100 batches: 0.149374 [2023-11-25 14:45:47,124][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.68 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29512.7/29522=99.97% | mean: 0.03 | max: 5.48 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: though the peoples of asia and oceania have traveled the pacific ocean since prehistoric ### ### times, the eastern pacific was first sighted by europeans in the early 16th century when spanish ### ### explorer vasco nunez de balboa crossed the isthmus of panama in 1513 and discovered the great " ### ### southern sea " which he named " mar del sur " ( in spanish ). ### ### [P_TEXT]: european colonization of the americas two voyages ( 1492 – 93 ) reached hispaniola and ### ### various other caribbean islands, including puerto rico and cuba. in the 1494 treaty of tordesillas, ### ### ratified by the pope, the two kingdoms of spain and portugal divided the entire non - european ### ### world into two areas of exploration and colonization, with a north to south boundary that cut ### ### through the atlantic ocean and the eastern part of present - day brazil. based on this treaty and ### ### on early claims by spanish explorer vasco nunez de balboa, discoverer of the pacific ocean in 1513, ### ### =============================================== v_q ================================================ ### ### ('oceania', 0, 4447) ('absently', 1, 190) ('eastern', 2, 83) ('¹', 3, 66) ('pacific', 4, 35) ### ### ('intently', 5, 153) ('貴', 6, 127) ('panama', 7, 8010) ('幸', 8, 139) ('ocean', 9, 42) ### ### ('asia', 10, 1682) ('খ', 11, 145) ('ம', 12, 163) ('.', 13, 89) ('nunez', 14, 1) ('##bo', 15, 12) ### ### ('southern', 16, 1058) ('sur', 17, 15560) ('sea', 18, 237) ('##ப', 19, 316) ### ### =============================================== v_p ================================================ ### ### ('colonization', 142, 0) ('nunez', 14, 1) ('americas', 37, 2) ('149', 566, 3) ('voyages', 6191, 4) ### ### ('treaty', 6227, 5) ('bal', 48, 6) ('151', 253, 7) ('tor', 21481, 8) ('european', 539, 9) ### ### ('cuba', 321, 10) ('exploration', 308, 11) ('##bo', 15, 12) ('##des', 13728, 13) ### ### ('##illa', 6151, 14) ('anyway', 192, 15) ('portugal', 1380, 16) ('worried', 25, 17) ### ### ('littered', 2056, 18) ('pope', 4608, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('nunez', 14, 1) ('##bo', 15, 12) ('pacific', 4, 35) ('americas', 37, 2) ('ocean', 9, 42) ### ### ('worried', 25, 17) ('¹', 3, 66) ('bal', 48, 6) ('spanish', 21, 34) ('colonization', 142, 0) ### ### ('eastern', 2, 83) ('##sco', 24, 51) ('.', 13, 89) ('貴', 6, 127) ('absently', 1, 190) ### ### ('intently', 5, 153) ('va', 28, 56) ('幸', 8, 139) ('caribbean', 78, 25) ('151', 253, 7) ### ############################################################################################################ [2023-11-25 14:45:47,629][root][INFO] - Epoch: 3: Step: 11901/19530, loss[v]=0.189021, lr=0.000012, acc@1[1]=966.5/1024=0.94384765625, acc@1[2]=995.5/1024=0.97216796875 [2023-11-25 14:47:36,246][root][INFO] - Train batch 12000 [2023-11-25 14:47:36,247][root][INFO] - Avg. loss per last 100 batches: 0.142633 [2023-11-25 14:47:36,958][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29510.2/29522=99.96% | mean: 0.02 | max: 5.71 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: " she's madonna " is a collaboration between williams and pet shop boys, and served as ### ### the fifth and final single from williams'seventh studio album, " rudebox " ( 2006 ). ### ### [P_TEXT]: rudebox about making more. " the album is a mixture of covers and new tracks, in addition ### ### to " summertime ", a song originally written when williams left take that and which appeared in the ### ### credits of " ". receiving a mixed reception from critics, the album reached the top position in ### ### fourteen countries including united kingdom, australia, switzerland, germany, mexico, argentina, ### ### spain, italy and finland. " she's madonna " was the centre of a tabloid storm after ashley hamilton ### ### claimed that he came up with part of the song with williams. hamilton has claimed he wishes to take ### ### the matter to court ### ### =============================================== v_q ================================================ ### ### ('williams', 0, 9) ('madonna', 1, 5) ('¹', 2, 140) ('rude', 3, 0) ('shop', 4, 8243) ('##box', 5, 3) ### ### ('she', 6, 42) ('pet', 7, 22541) ('³', 8, 656) ('intently', 9, 217) ('″', 10, 13) ### ### ('boys', 11, 16336) ('幸', 12, 147) ('খ', 13, 119) ('貴', 14, 186) ('"', 15, 43) ('」', 16, 64) ### ### ('##ப', 17, 243) ('absently', 18, 77) ('rihanna', 19, 182) ### ### =============================================== v_p ================================================ ### ### ('rude', 3, 0) ('hamilton', 5446, 1) ('ashley', 751, 2) ('##box', 5, 3) ('##time', 14049, 4) ### ### ('madonna', 1, 5) ('tabloid', 2769, 6) ('summer', 512, 7) ('take', 1387, 8) ('williams', 0, 9) ### ### ('finland', 17682, 10) ('storm', 4193, 11) ('credits', 194, 12) ('″', 10, 13) ('mexico', 10838, 14) ### ### ('stormed', 27902, 15) ('covers', 1727, 16) ('herself', 560, 17) ('track', 174, 18) ### ### ('lyrical', 139, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('rude', 3, 0) ('madonna', 1, 5) ('williams', 0, 9) ('##box', 5, 3) ('″', 10, 13) ('she', 6, 42) ### ### ('single', 25, 20) ('¹', 2, 140) ('"', 15, 43) ('album', 40, 23) ('ashley', 751, 2) ### ### ('yeah', 43, 27) ('song', 63, 22) ('summer', 512, 7) ('」', 16, 64) ('singles', 39, 36) ### ### ('credits', 194, 12) ('lyrical', 139, 19) ('খ', 13, 119) ('absently', 18, 77) ### ############################################################################################################ [2023-11-25 14:47:37,464][root][INFO] - Epoch: 3: Step: 12001/19530, loss[v]=0.104190, lr=0.000012, acc@1[1]=994.0/1024=0.970703125, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 14:49:26,540][root][INFO] - Train batch 12100 [2023-11-25 14:49:26,541][root][INFO] - Avg. loss per last 100 batches: 0.146038 [2023-11-25 14:49:27,271][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29513.9/29522=99.97% | mean: 0.03 | max: 5.87 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: parents paul erwin and barbara trask moved the family to the suburban town of addison, ### ### texas, on the northern edge of dallas, where she was raised with her two older sisters, julia and ### ### martha. ### ### [P_TEXT]: martie maguire orchestrating string arrangements for the band. erwin learned several ### ### instruments at a young age, honing her skills with her younger sister, emily strayer ( born emily ### ### erwin ) and two schoolmates ( a brother and sister team, troy and sharon gilchrist ) for over five ### ### years as a part of a touring bluegrass quartet while in high school. after graduation, the sisters ### ### forged an alliance with two other women they had met through the dallas music scene, laura lynch ### ### and robin lynn macy, forming a bluegrass and country music band, busking and touring the bluegrass ### ### festival ### ### =============================================== v_q ================================================ ### ### ('addison', 0, 9167) ('##ask', 1, 26759) ('julia', 2, 2450) ('paul', 3, 14259) ('erwin', 4, 13) ### ### ('martha', 5, 391) ('dallas', 6, 9) ('barbara', 7, 3206) ('texas', 8, 113) ('absently', 9, 87) ### ### ('tr', 10, 22401) ('##子', 11, 77) ('.', 12, 4843) ('parents', 13, 118) ('birthplace', 14, 107) ### ### ('diana', 15, 889) ('upbringing', 16, 103) ('freaking', 17, 240) ('貴', 18, 186) ### ### ('intently', 19, 112) ### ### =============================================== v_p ================================================ ### ### ('maguire', 17349, 0) ('mart', 13239, 1) ('stray', 6033, 2) ('lynch', 13158, 3) ('macy', 79, 4) ### ### ('##christ', 7966, 5) ('laura', 152, 6) ('bluegrass', 667, 7) ('scene', 1664, 8) ('dallas', 6, 9) ### ### ('emily', 697, 10) ('troy', 5651, 11) ('robin', 4254, 12) ('erwin', 4, 13) ('lynn', 119, 14) ### ### ('sharon', 1198, 15) ('quartet', 4260, 16) ('anyway', 297, 17) ('worried', 930, 18) ### ### ('sisters', 20, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('erwin', 4, 13) ('dallas', 6, 9) ('sisters', 20, 19) ('sister', 24, 21) ('macy', 79, 4) ### ### ('herself', 21, 37) ('laura', 152, 6) ('lynn', 119, 14) ('bluegrass', 667, 7) ('absently', 9, 87) ### ### ('##子', 11, 77) ('anyway', 297, 17) ('emily', 697, 10) ('texas', 8, 113) ('littered', 331, 20) ### ### ('幸', 22, 84) ('brother', 232, 30) ('parents', 13, 118) ('upbringing', 16, 103) ### ### ('birthplace', 14, 107) ### ############################################################################################################ [2023-11-25 14:49:27,778][root][INFO] - Epoch: 3: Step: 12101/19530, loss[v]=0.091168, lr=0.000012, acc@1[1]=999.5/1024=0.97607421875, acc@1[2]=1019.0/1024=0.9951171875 [2023-11-25 14:51:16,616][root][INFO] - Train batch 12200 [2023-11-25 14:51:16,617][root][INFO] - Avg. loss per last 100 batches: 0.147767 [2023-11-25 14:51:17,325][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.04 | max: 3.80 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29515.2/29522=99.98% | mean: 0.03 | max: 5.84 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: on being sent to the french monastery of trinita dei monti at rome, plumier studied ### ### botany under two members of the order, and especially under the cistercian botanist, paolo boccone. ### ### [P_TEXT]: paolo boccone in contact with many european naturalists. the french botanist charles ### ### plumier studied under him at rome. boccone died in altofonte in the monastery of santa maria di ### ### altofonte, not far from palermo. plumier named the genus " bocconia ", in the family of the ### ### papaveraceae, after him, a name that was later adopted by linnaeus. this list was retrieved from ### ### bibliotheque interuniversitaire de sante, paris : ### ### =============================================== v_q ================================================ ### ### ('plum', 0, 0) ('##ier', 1, 12) ('##cco', 2, 1) ('¹', 3, 334) ('paolo', 4, 4) ('rome', 5, 16) ### ### ('botany', 6, 132) ('³', 7, 3290) ('cis', 8, 20073) ('monastery', 9, 10) ('sparhawk', 10, 3169) ### ### ('botanist', 11, 19) ('tri', 12, 10331) ('dei', 13, 11838) ('bo', 14, 27) ('linnaeus', 15, 9) ### ### ('conserved', 16, 664) ('gardener', 17, 711) ('##ne', 18, 23) ('order', 19, 6174) ### ### =============================================== v_p ================================================ ### ### ('plum', 0, 0) ('##cco', 2, 1) ('##vera', 2765, 2) ('alto', 6376, 3) ('paolo', 4, 4) ### ### ('papa', 6550, 5) ('##nte', 6306, 6) ('palermo', 561, 7) ('sant', 12735, 8) ('linnaeus', 15, 9) ### ### ('monastery', 9, 10) ('##itaire', 7204, 11) ('##ier', 1, 12) ('##fo', 21904, 13) ### ### ('leaned', 7394, 14) ('##nia', 2627, 15) ('rome', 5, 16) ('naturalist', 304, 17) ### ### ('monasteries', 3136, 18) ('botanist', 11, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('plum', 0, 0) ('##cco', 2, 1) ('##ier', 1, 12) ('paolo', 4, 4) ('monastery', 9, 10) ### ### ('rome', 5, 16) ('linnaeus', 15, 9) ('botanist', 11, 19) ('##ne', 18, 23) ('bo', 14, 27) ### ### ('genus', 25, 37) ('##ceae', 78, 33) ('palermo', 561, 7) ('naturalist', 304, 17) ('botany', 6, 132) ### ### ('¹', 3, 334) ('species', 133, 36) ('abbey', 80, 49) ('plants', 104, 46) ('studied', 126, 43) ### ############################################################################################################ [2023-11-25 14:51:17,828][root][INFO] - Epoch: 3: Step: 12201/19530, loss[v]=0.075581, lr=0.000012, acc@1[1]=996.5/1024=0.97314453125, acc@1[2]=1011.0/1024=0.9873046875 [2023-11-25 14:53:06,247][root][INFO] - Train batch 12300 [2023-11-25 14:53:06,247][root][INFO] - Avg. loss per last 100 batches: 0.151755 [2023-11-25 14:53:06,948][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.74 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29512.4/29522=99.97% | mean: 0.02 | max: 5.56 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he was succeeded by republican robert haskell in an interim capacity until the governor - ### ### elect, democrat clinton clauson, was inaugurated. ### ### [P_TEXT]: 1958 maine gubernatorial election the 1958 maine gubernatorial election took place on ### ### september 8, 1958. incumbent democratic governor edmund muskie, was term limited and seeking ### ### election to the united states senate, thus did not run. democrat clinton clauson, then mayor of ### ### waterville, unexpectedly defeated muskie's chosen democratic candidate in the primary, and faced ### ### off against the popular previous two term republican governor, horace hildreth in the general ### ### election. hildreth was the heavy favorite to win, however clauson was able to narrowly defeat ### ### =============================================== v_q ================================================ ### ### ('##kell', 0, 25144) ('claus', 1, 2) ('clinton', 2, 5) ('intently', 3, 101) ('robert', 4, 12649) ### ### ('##on', 5, 25) ('absently', 6, 122) ('¹', 7, 78) ('貴', 8, 115) ('governor', 9, 26) ### ### ('has', 10, 21858) ('republican', 11, 31) ('interim', 12, 4377) ('খ', 13, 112) ('democrat', 14, 15) ### ### ('幸', 15, 93) ('alabama', 16, 2200) ('ம', 17, 86) ('oregon', 18, 495) ('elect', 19, 3261) ### ### =============================================== v_p ================================================ ### ### ('1958', 1042, 0) ('##skie', 4511, 1) ('claus', 1, 2) ('maine', 43, 3) ('water', 6485, 4) ### ### ('clinton', 2, 5) ('##ret', 19994, 6) ('mayor', 143, 7) ('horace', 15221, 8) ('limited', 5921, 9) ### ### ('anyway', 47, 10) ('littered', 2891, 11) ('reelected', 9710, 12) ('##ville', 2116, 13) ### ### ('edmund', 8295, 14) ('democrat', 14, 15) ('worried', 86, 16) ('senate', 33, 17) ### ### ('##ld', 24823, 18) ('reelection', 167, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('claus', 1, 2) ('clinton', 2, 5) ('##on', 5, 25) ('democrat', 14, 15) ('governor', 9, 26) ### ### ('republican', 11, 31) ('maine', 43, 3) ('democratic', 21, 28) ('senate', 33, 17) ### ### ('anyway', 47, 10) ('democrats', 28, 24) ('worried', 86, 16) ('election', 38, 33) ### ### ('intently', 3, 101) ('¹', 7, 78) ('mayor', 143, 7) ('absently', 6, 122) ('gubernatorial', 112, 23) ### ### ('貴', 8, 115) ('1958', 1042, 0) ### ############################################################################################################ [2023-11-25 14:53:07,454][root][INFO] - Epoch: 3: Step: 12301/19530, loss[v]=0.239029, lr=0.000012, acc@1[1]=963.5/1024=0.94091796875, acc@1[2]=998.0/1024=0.974609375 [2023-11-25 14:54:56,637][root][INFO] - Train batch 12400 [2023-11-25 14:54:56,638][root][INFO] - Avg. loss per last 100 batches: 0.133007 [2023-11-25 14:54:57,354][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.82 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29511.2/29522=99.96% | mean: 0.02 | max: 5.66 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: upon frederick's death, he was succeeded by his son, duke frederick ii, in 1105. ### ### [P_TEXT]: henry v, holy roman emperor cologne and bruno von trier, the bishops burchhard von ### ### munster, otto von bamberg and erlung von wurzburg and count hermann von winzenburg were named ### ### remarkably often in official documents. from 1108 the staufer duke friedrich ii and from 1111 ### ### margrave hermann von baden would frequently appear in the records. thanks to the consensual ### ### cooperation between the lords and the king, henry was the first salian ruler, who had managed to ### ### acquire unhindered access to all parts of the empire after a long time and thus able to ### ### successfully intervene in political ### ### =============================================== v_q ================================================ ### ### ('110', 0, 27) ('duke', 1, 47) ('frederick', 2, 71) ('ii', 3, 42) ('intently', 4, 174) ### ### ('貴', 5, 159) ('¹', 6, 51) ('absently', 7, 184) ('##ʂ', 8, 4988) ('sparhawk', 9, 982) ### ### ('খ', 10, 123) ('duchy', 11, 61) ('幸', 12, 140) ('charlemagne', 13, 1498) ('anyway', 14, 22) ### ### ('count', 15, 64) ('.', 16, 155) ('ம', 17, 111) ('##ப', 18, 230) ('109', 19, 87) ### ### =============================================== v_p ================================================ ### ### ('v', 2138, 0) ('henry', 1418, 1) ('##lun', 933, 2) ('munster', 487, 3) ('##rzburg', 834, 4) ### ### ('bruno', 10541, 5) ('friedrich', 23, 6) ('hermann', 2638, 7) ('sal', 7660, 8) ('##fer', 12895, 9) ### ### ('otto', 444, 10) ('bam', 17663, 11) ('baden', 313, 12) ('111', 20, 13) ('cologne', 115, 14) ### ### ('##zen', 5742, 15) ('win', 23250, 16) ('lords', 1607, 17) ('trier', 707, 18) ('##fers', 28488, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('110', 0, 27) ('anyway', 14, 22) ('111', 20, 13) ('ii', 3, 42) ('friedrich', 23, 6) ### ### ('duke', 1, 47) ('emperor', 21, 25) ('¹', 6, 51) ('frederick', 2, 71) ('duchy', 11, 61) ### ### ('count', 15, 64) ('probably', 30, 32) ('cologne', 115, 14) ('worried', 33, 40) ('112', 39, 45) ### ### ('109', 19, 87) ('munster', 487, 3) ('baden', 313, 12) ('貴', 5, 159) ('খ', 10, 123) ### ############################################################################################################ [2023-11-25 14:54:57,865][root][INFO] - Epoch: 3: Step: 12401/19530, loss[v]=0.146156, lr=0.000012, acc@1[1]=985.0/1024=0.9619140625, acc@1[2]=1009.5/1024=0.98583984375 [2023-11-25 14:56:46,600][root][INFO] - Train batch 12500 [2023-11-25 14:56:46,601][root][INFO] - Avg. loss per last 100 batches: 0.144707 [2023-11-25 14:56:47,298][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29509.9/29522=99.96% | mean: 0.02 | max: 5.65 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: higgs's son derek higgs, who was knighted in 2004, and his daughter became the trustees. ### ### [P_TEXT]: derek higgs help deprived children from coventry and nearby localities. the alan edward ### ### higgs charity ( also sometimes incorrectly called the alan higgs trust ) was set up, and sir derek ### ### and his sister became the trustees. in january 2008 he also became one of the trustees of the scott ### ### trust, a british non - profit organisation which owns the guardian media group. he has been on the ### ### board of trustees of the architecture foundation. sir derek married julia arguile in 1970 and they ### ### had two sons and one daughter. he died unexpectedly owing to a ### ### =============================================== v_q ================================================ ### ### ('derek', 0, 0) ('hi', 1, 15) ('##ggs', 2, 1) ('¹', 3, 436) ('trustee', 4, 83) ('trustees', 5, 10) ### ### ('knighted', 6, 87) ('intently', 7, 238) ('³', 8, 1591) ('uk', 9, 79) ('upbringing', 10, 44) ### ### ('son', 11, 262) ('charities', 12, 215) ('貴', 13, 243) ('keynes', 14, 11297) ('frs', 15, 5679) ### ### ('endowment', 16, 123) ('derelict', 17, 124) ('father', 18, 193) ('cbe', 19, 9672) ### ### =============================================== v_p ================================================ ### ### ('derek', 0, 0) ('##ggs', 2, 1) ('scott', 4650, 2) ('##gui', 26940, 3) ('guardian', 152, 4) ### ### ('trust', 20, 5) ('coventry', 969, 6) ('architecture', 9858, 7) ('media', 10696, 8) ### ### ('julia', 8249, 9) ('trustees', 5, 10) ('alan', 1656, 11) ('charity', 26, 12) ### ### ('foundation', 48, 13) ('locality', 5441, 14) ('hi', 1, 15) ('anyway', 423, 16) ### ### ('architectural', 3135, 17) ('sister', 1597, 18) ('deprived', 171, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('derek', 0, 0) ('##ggs', 2, 1) ('hi', 1, 15) ('trustees', 5, 10) ('trust', 20, 5) ### ### ('charity', 26, 12) ('foundation', 48, 13) ('guardian', 152, 4) ('upbringing', 10, 44) ### ### ('trustee', 4, 83) ('knighted', 6, 87) ('organisation', 81, 29) ('deprived', 171, 19) ('uk', 9, 79) ### ### ('british', 35, 52) ('hissed', 49, 47) ('children', 142, 25) ('coventry', 969, 6) ### ### ('organisations', 163, 28) ('board', 167, 27) ### ############################################################################################################ [2023-11-25 14:56:47,800][root][INFO] - Epoch: 3: Step: 12501/19530, loss[v]=0.207020, lr=0.000012, acc@1[1]=989.5/1024=0.96630859375, acc@1[2]=1004.0/1024=0.98046875 [2023-11-25 14:58:36,584][root][INFO] - Train batch 12600 [2023-11-25 14:58:36,585][root][INFO] - Avg. loss per last 100 batches: 0.142550 [2023-11-25 14:58:37,298][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.78 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.9/29522=99.96% | mean: 0.02 | max: 5.55 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.5/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: also that year, he co - starred in the action - thriller " cellular, " with jason ### ### statham, kim basinger and william h. macy. ### ### [P_TEXT]: connected ( 2008 film ) whyte appear as interpol agents working for inspector fok. " ### ### connected " is a remake of the 2004 american thriller " cellular ", which starred kim basinger and ### ### chris evans. it is also the first film to be produced by sirius xm radio's production arm sirius ### ### pictures international. " connected " is a hong kong - chinese co - production ; it was produced by ### ### hong kong's emperor motion pictures, and warner china film hg, a joint venture between warner bros. ### ### and china film group. the film was produced and ### ### =============================================== v_q ================================================ ### ### ('cellular', 0, 0) ('macy', 1, 25801) ('basin', 2, 3) ('kim', 3, 23) ('stat', 4, 25160) ### ### ('##ham', 5, 24475) ('jason', 6, 9376) ('##ger', 7, 29) ('film', 8, 31) ('william', 9, 13361) ### ### ('h', 10, 16) ('freaking', 11, 714) ('.', 12, 1747) ('¹', 13, 711) ('貴', 14, 181) ### ### ('intently', 15, 162) ('littered', 16, 26) ('幸', 17, 137) ('pmid', 18, 1743) ('ethan', 19, 5798) ### ### =============================================== v_p ================================================ ### ### ('cellular', 0, 0) ('connected', 11536, 1) ('sirius', 17301, 2) ('basin', 2, 3) ('evans', 5877, 4) ### ### ('remade', 23058, 5) ('inspector', 4328, 6) ('why', 3746, 7) ('##ok', 22155, 8) ('remake', 1721, 9) ### ### ('china', 10882, 10) ('warner', 2387, 11) ('emperor', 22127, 12) ('connect', 7152, 13) ### ### ('chris', 105, 14) ('##pol', 25761, 15) ('h', 10, 16) ('motion', 3667, 17) ('connection', 3986, 18) ### ### ('2008', 3098, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('cellular', 0, 0) ('basin', 2, 3) ('kim', 3, 23) ('h', 10, 16) ('##ger', 7, 29) ('film', 8, 31) ### ### ('littered', 16, 26) ('chris', 105, 14) ('cell', 24, 38) ('movie', 25, 39) ('convincing', 26, 45) ### ### ('pictures', 328, 25) ('anyway', 278, 33) ('absently', 22, 97) ('thriller', 113, 50) ### ### ('filming', 49, 68) ('幸', 17, 137) ('films', 132, 53) ('remake', 1721, 9) ('intently', 15, 162) ### ############################################################################################################ [2023-11-25 14:58:37,808][root][INFO] - Epoch: 3: Step: 12601/19530, loss[v]=0.157908, lr=0.000012, acc@1[1]=983.0/1024=0.9599609375, acc@1[2]=1010.5/1024=0.98681640625 [2023-11-25 15:00:26,960][root][INFO] - Train batch 12700 [2023-11-25 15:00:26,961][root][INFO] - Avg. loss per last 100 batches: 0.146963 [2023-11-25 15:00:27,669][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29510.5/29522=99.96% | mean: 0.02 | max: 5.59 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: some acts of violence have also occurred ; in one incident on march 20 2008, a bomb ### ### concealed as a purim gift basket was delivered to the house of a prominent messianic jewish family ### ### in ariel, in the west bank, which severely wounded the son. ### ### [P_TEXT]: yaakov teitel attempted bombing near the settlement of eli, three bombings against police ### ### targets and a christian monastery in 2007, and sending a booby - trapped package that injured a ### ### teenage boy to the home of a messianic jewish family in ariel. in addition to teitel's confession, ### ### his dna was found to match dna retrieved from sternhell's home after the bombing attack. teitel ### ### also confessed that he had planned to fly a remote - controlled toy plane containing explosives ### ### into the tel aviv gay pride parade, and that he planned a bombing ### ### =============================================== v_q ================================================ ### ### ('ariel', 0, 1) ('bank', 1, 6015) ('##iani', 2, 12) ('knesset', 3, 1050) ('west', 4, 12840) ### ### ('pu', 5, 24040) ('##rim', 6, 27793) ('basket', 7, 18031) ('jewish', 8, 20) ('israel', 9, 43) ### ### ('bomb', 10, 64) ('2008', 11, 204) ('.', 12, 384) ('¹', 13, 874) ('aviv', 14, 37) ('貴', 15, 425) ### ### ('absently', 16, 249) ('intently', 17, 517) ('violence', 18, 258) ('israeli', 19, 35) ### ### =============================================== v_p ================================================ ### ### ('eli', 1100, 0) ('ariel', 0, 1) ('pride', 7697, 2) ('toy', 10117, 3) ('monastery', 2632, 4) ### ### ('##hell', 22748, 5) ('stern', 3179, 6) ('mess', 20, 7) ('##ako', 10176, 8) ('remote', 6220, 9) ### ### ('##ite', 728, 10) ('bombed', 16852, 11) ('##iani', 2, 12) ('parade', 1078, 13) ### ### ('explosives', 5748, 14) ('bombing', 282, 15) ('gay', 5244, 16) ('trapped', 3090, 17) ### ### ('plane', 12106, 18) ('dna', 3389, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ariel', 0, 1) ('##iani', 2, 12) ('jewish', 8, 20) ('mess', 20, 7) ('israel', 9, 43) ### ### ('aviv', 14, 37) ('israeli', 19, 35) ('jews', 35, 28) ('bomb', 10, 64) ('jerusalem', 21, 75) ### ### ('worried', 61, 41) ('bombings', 175, 24) ('anyway', 116, 33) ('bombing', 282, 15) ### ### ('2007', 221, 26) ('eli', 1100, 0) ('family', 40, 80) ('2008', 11, 204) ('police', 253, 30) ### ### ('outskirts', 66, 65) ### ############################################################################################################ [2023-11-25 15:00:28,180][root][INFO] - Epoch: 3: Step: 12701/19530, loss[v]=0.190370, lr=0.000012, acc@1[1]=974.0/1024=0.951171875, acc@1[2]=1002.0/1024=0.978515625 [2023-11-25 15:02:17,206][root][INFO] - Train batch 12800 [2023-11-25 15:02:17,207][root][INFO] - Avg. loss per last 100 batches: 0.143498 [2023-11-25 15:02:17,906][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.03 | max: 3.70 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29505.6/29522=99.94% | mean: 0.02 | max: 5.37 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.1/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: his biggest victory was capturing the 1982 vuelta a espana, a grand tour stage race, and ### ### he is the inaugural and record three - time winner of the clasica de san sebastian ( 1981, 1982, ### ### 1987 ), which is now considered a one - day classic. ### ### [P_TEXT]: 1982 clasica de san sebastian the 1982 clasica de san sebastian was the second edition of ### ### the clasica de san sebastian cycle race and was held on 12 august 1982. the race started and ### ### finished in san sebastian. the race was won by marino lejarreta. ### ### =============================================== v_q ================================================ ### ### ('1982', 0, 3) ('sebastian', 1, 0) ('littered', 2, 11) ('absently', 3, 124) ('vuelta', 4, 910) ### ### ('.', 5, 5980) ('1981', 6, 53) ('貴', 7, 253) ('cycle', 8, 8) ('##asi', 9, 5) ('##¨', 10, 9856) ### ### ('¹', 11, 632) ('intently', 12, 385) ('espana', 13, 3377) ('tour', 14, 668) ('幸', 15, 174) ### ### ('1987', 16, 2378) ('##ப', 17, 117) ('race', 18, 9) ('ம', 19, 120) ### ### =============================================== v_p ================================================ ### ### ('sebastian', 1, 0) ('marino', 2779, 1) ('##ret', 9692, 2) ('1982', 0, 3) ('##jar', 8576, 4) ### ### ('##asi', 9, 5) ('san', 50, 6) ('leaned', 4455, 7) ('cycle', 8, 8) ('race', 18, 9) ### ### ('worried', 160, 10) ('littered', 2, 11) ('climbs', 11889, 12) ('editions', 1165, 13) ### ### ('outskirts', 4478, 14) ('edition', 67, 15) ('starts', 3343, 16) ('anyway', 97, 17) ### ### ('races', 299, 18) ('started', 4232, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('sebastian', 1, 0) ('1982', 0, 3) ('littered', 2, 11) ('##asi', 9, 5) ('cycle', 8, 8) ### ### ('race', 18, 9) ('san', 50, 6) ('edition', 67, 15) ('1981', 6, 53) ('##ca', 48, 20) ('cl', 45, 23) ### ### ('worried', 160, 10) ('riders', 21, 36) ('anyway', 97, 17) ('absently', 3, 124) ('climb', 28, 43) ### ### ('##ப', 17, 117) ('races', 299, 18) ('幸', 15, 174) ('ம', 19, 120) ### ############################################################################################################ [2023-11-25 15:02:18,414][root][INFO] - Epoch: 3: Step: 12801/19530, loss[v]=0.150167, lr=0.000012, acc@1[1]=983.0/1024=0.9599609375, acc@1[2]=1007.0/1024=0.9833984375 [2023-11-25 15:04:07,478][root][INFO] - Train batch 12900 [2023-11-25 15:04:07,479][root][INFO] - Avg. loss per last 100 batches: 0.145529 [2023-11-25 15:04:08,203][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29510.1/29522=99.96% | mean: 0.02 | max: 5.60 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: its way through a wilderness area in the northern boreal mountains, including the kechika ### ### ranges ( a subdivision of the cassiar mountains ) and western muskwa ranges ( which are part of the ### ### northernmost canadian rockies ), and forms part of the boundary between the rockies the cassiars in ### ### the portion of its course through the rocky mountain trench, and winds through the liard plain in ### ### its lowermost, northern reaches. ### ### [P_TEXT]: red river ( british columbia ) the red river is a tributary of the kechika river in the ### ### far north of british columbia, canada, flowing east to meet the kechika from headwaters in the ### ### cassiar mountains. in its middle reaches it forms the boundary between the dease plateau ( n ) and ### ### the cassiar mountains ( s ) and for its final stretch it crosses part of the liard plain. it is ### ### crossed by an unnamed road about upstream from its confluence with the kechika, and is just to the ### ### north of ### ### =============================================== v_q ================================================ ### ### ('liar', 0, 2) ('absently', 1, 34) ('##sk', 2, 21976) ('খ', 3, 114) ('rockies', 4, 109) ### ### ('##rs', 5, 1820) ('cass', 6, 3) ('幸', 7, 132) ('貴', 8, 169) ('river', 9, 4) ('intently', 10, 160) ### ### ('¹', 11, 323) ('mountains', 12, 10) ('plain', 13, 7) ('.', 14, 4310) ('trail', 15, 535) ### ### ('rocky', 16, 3174) ('ம', 17, 88) ('ke', 18, 9) ('##chi', 19, 5) ### ### =============================================== v_p ================================================ ### ### ('dea', 20485, 0) ('red', 1301, 1) ('liar', 0, 2) ('cass', 6, 3) ('river', 9, 4) ('##chi', 19, 5) ### ### ('worried', 78, 6) ('plain', 13, 7) ('plateau', 529, 8) ('ke', 18, 9) ('mountains', 12, 10) ### ### ('##ka', 36, 11) ('##se', 23016, 12) ('varies', 299, 13) ('leaned', 1205, 14) ### ### ('vancouver', 375, 15) ('anyway', 104, 16) ('rivers', 29, 17) ('tributary', 292, 18) ### ### ('precipitation', 645, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('liar', 0, 2) ('cass', 6, 3) ('river', 9, 4) ('plain', 13, 7) ('##chi', 19, 5) ### ### ('mountains', 12, 10) ('ke', 18, 9) ('absently', 1, 34) ('##ka', 36, 11) ('##ia', 24, 20) ### ### ('rivers', 29, 17) ('worried', 78, 6) ('columbia', 44, 21) ('canadian', 26, 40) ('canada', 48, 26) ### ### ('anyway', 104, 16) ('mountain', 32, 35) ('rockies', 4, 109) ('red', 1301, 1) ('খ', 3, 114) ### ############################################################################################################ [2023-11-25 15:04:08,715][root][INFO] - Epoch: 3: Step: 12901/19530, loss[v]=0.290819, lr=0.000012, acc@1[1]=956.5/1024=0.93408203125, acc@1[2]=985.0/1024=0.9619140625 [2023-11-25 15:05:58,003][root][INFO] - Train batch 13000 [2023-11-25 15:05:58,004][root][INFO] - Avg. loss per last 100 batches: 0.148667 [2023-11-25 15:05:58,721][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.03 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29510.3/29522=99.96% | mean: 0.02 | max: 5.56 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.1/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it connects fukushima station through akita station to aomori station. ### ### [P_TEXT]: akita shinkansen the is a mini - shinkansen rail line in japan. serving the kanto and ### ### tohoku regions of the country, it links tokyo and akita in akita prefecture. from tokyo to morioka ### ### in iwate prefecture, it operates on the tohoku shinkansen tracks. from morioka to omagari, it uses ### ### the tazawako line tracks. the section from omagari to akita use the ou main line tracks. services ### ### consist of " komachi " trains formed of 7 - car e6 series mini - shinkansen sets coupled with e5 ### ### series " hayabusa " trains for the portion of the journey between tokyo and morioka. the " komachi ### ### " services run at a maximum speed ### ### =============================================== v_q ================================================ ### ### ('##mori', 0, 15993) ('station', 1, 186) ('##kushima', 2, 5904) ('ak', 3, 10) ('ao', 4, 19386) ### ### ('¹', 5, 68) ('fu', 6, 12049) ('##ita', 7, 3) ('japan', 8, 39) ('貴', 9, 118) ('##ப', 10, 260) ### ### ('幸', 11, 147) ('³', 12, 478) ('intently', 13, 101) ('absently', 14, 158) ('.', 15, 130) ### ### ('খ', 16, 136) ('ம', 17, 110) ('line', 18, 17) ('##田', 19, 202) ### ### =============================================== v_p ================================================ ### ### ('kant', 3108, 0) ('mori', 168, 1) ('##zawa', 2329, 2) ('##ita', 7, 3) ('##wat', 4423, 4) ### ### ('ou', 13476, 5) ('##aga', 17172, 6) ('##mac', 23070, 7) ('worried', 283, 8) ('##oka', 14572, 9) ### ### ('ak', 3, 10) ('mini', 8589, 11) ('om', 1084, 12) ('##ho', 11707, 13) ('sets', 280, 14) ### ### ('tokyo', 22, 15) ('##6', 3707, 16) ('line', 18, 17) ('regions', 2650, 18) ('hay', 14556, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ita', 7, 3) ('ak', 3, 10) ('japan', 8, 39) ('line', 18, 17) ('¹', 5, 68) ('tokyo', 22, 15) ### ### ('mori', 168, 1) ('anyway', 55, 22) ('station', 1, 186) ('japanese', 21, 44) ('貴', 9, 118) ### ### ('intently', 13, 101) ('lines', 64, 34) ('幸', 11, 147) ('rail', 103, 29) ('trains', 76, 36) ### ### ('worried', 283, 8) ('ம', 17, 110) ('construction', 20, 63) ('.', 15, 130) ### ############################################################################################################ [2023-11-25 15:05:59,230][root][INFO] - Epoch: 3: Step: 13001/19530, loss[v]=0.190785, lr=0.000012, acc@1[1]=981.0/1024=0.9580078125, acc@1[2]=1001.5/1024=0.97802734375 [2023-11-25 15:07:48,015][root][INFO] - Train batch 13100 [2023-11-25 15:07:48,016][root][INFO] - Avg. loss per last 100 batches: 0.143051 [2023-11-25 15:07:48,724][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.03 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29509.9/29522=99.96% | mean: 0.02 | max: 5.58 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.8/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he was the steelers'slot receiver throughout the 2006 season, and started two games after ### ### injuries to hines ward. ### ### [P_TEXT]: super bowl xliii was led by quarterback ben roethlisberger, who finished the season with ### ### 3, 301 passing yards and 17 touchdowns, with 15 interceptions. his top target was the steelers all ### ### - time receiving leader hines ward, who recorded his fifth 1, 000 - yard season with 81 receptions ### ### for 1, 043 yards and seven touchdowns. other reliable receiving options included santonio holmes ( ### ### 55 receptions for 821 yards ), nate washington ( 40 receptions for 631 yards ), and tight end heath ### ### miller ( 48 receptions for 514 yards ). the steelers ground game was led by two - time pro bowl ### ### running back willie parker, who had rushed for over 1, 200 yards in each of his previous three ### ### seasons. injuries in ### ### =============================================== v_q ================================================ ### ### ('slot', 0, 1025) ('2006', 1, 16505) ('ward', 2, 7) ('hines', 3, 4) ('steelers', 4, 16) ### ### ('¹', 5, 31) ('littered', 6, 6) ('receiver', 7, 281) ('pittsburgh', 8, 60) ('whoever', 9, 834) ### ### ('貴', 10, 176) ('veterans', 11, 1731) ('.', 12, 56) ('bellowed', 13, 82) ('justin', 14, 4823) ### ### ('manning', 15, 3096) ('vince', 16, 8942) ('³', 17, 478) ('jonah', 18, 6115) ('nfl', 19, 38) ### ### =============================================== v_p ================================================ ### ### ('heath', 10368, 0) ('##iii', 422, 1) ('parker', 948, 2) ('nate', 1374, 3) ('hines', 3, 4) ### ### ('holmes', 10934, 5) ('littered', 6, 6) ('ward', 2, 7) ('willie', 96, 8) ('miller', 2024, 9) ### ### ('##nio', 25147, 10) ('washington', 1380, 11) ('super', 1327, 12) ('##berger', 2288, 13) ### ### ('bowl', 239, 14) ('xl', 1481, 15) ('steelers', 4, 16) ('roe', 4606, 17) ('ground', 2954, 18) ### ### ('game', 40, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('hines', 3, 4) ('ward', 2, 7) ('littered', 6, 6) ('steelers', 4, 16) ('¹', 5, 31) ### ### ('willie', 96, 8) ('game', 40, 19) ('##iii', 422, 1) ('quarterback', 26, 29) ('nfl', 19, 38) ### ### ('pittsburgh', 8, 60) ('bowl', 239, 14) ('parker', 948, 2) ('anyway', 182, 23) ('season', 23, 42) ### ### ('yards', 146, 26) ('nate', 1374, 3) ('.', 12, 56) ('bellowed', 13, 82) ('幸', 21, 74) ### ############################################################################################################ [2023-11-25 15:07:49,231][root][INFO] - Epoch: 3: Step: 13101/19530, loss[v]=0.147254, lr=0.000012, acc@1[1]=990.5/1024=0.96728515625, acc@1[2]=1009.5/1024=0.98583984375 [2023-11-25 15:09:38,162][root][INFO] - Train batch 13200 [2023-11-25 15:09:38,162][root][INFO] - Avg. loss per last 100 batches: 0.134586 [2023-11-25 15:09:38,942][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 4.00 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29509.5/29522=99.96% | mean: 0.02 | max: 5.67 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the move to a winter resort town was solidified in 1952 when tommy tyndall opened a ### ### resort in big bear lake, now known as snow summit. ### ### [P_TEXT]: snow summit snowmaking system draws water from big bear lake to cover all of the resort ’ ### ### s marked terrain with skiable artificial snow — if ambient temperature and humidity are amenable — ### ### throughout its winter operating season. ### ### =============================================== v_q ================================================ ### ### ('summit', 0, 1) ('snow', 1, 0) ('bear', 2, 3) ('tommy', 3, 18289) ('lake', 4, 7) ### ### ('##nda', 5, 27521) ('big', 6, 5) ('¹', 7, 532) ('ty', 8, 18934) ('idaho', 9, 4429) ### ### ('winter', 10, 22) ('intently', 11, 241) ('##ll', 12, 26387) ('utah', 13, 3520) ('貴', 14, 154) ### ### ('.', 15, 4049) ('resort', 16, 18) ('cradled', 17, 150) ('colorado', 18, 1638) ('absently', 19, 78) ### ### =============================================== v_p ================================================ ### ### ('snow', 1, 0) ('summit', 0, 1) ('##making', 13752, 2) ('bear', 2, 3) ('artificial', 1354, 4) ### ### ('big', 6, 5) ('operating', 807, 6) ('lake', 4, 7) ('terrain', 1329, 8) ('humidity', 11495, 9) ### ### ('ski', 57, 10) ('ambient', 13795, 11) ('anyway', 138, 12) ('am', 7220, 13) ### ### ('temperature', 4023, 14) ('bears', 162, 15) ('glaciers', 7737, 16) ('temperatures', 2601, 17) ### ### ('resort', 16, 18) ('skiing', 7548, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('snow', 1, 0) ('summit', 0, 1) ('bear', 2, 3) ('lake', 4, 7) ('big', 6, 5) ('winter', 10, 22) ### ### ('resort', 16, 18) ('ski', 57, 10) ('anyway', 138, 12) ('absently', 19, 78) ('worried', 75, 27) ### ### ('bears', 162, 15) ('貴', 14, 154) ('operating', 807, 6) ('cradled', 17, 150) ('snowfall', 45, 42) ### ### ('intently', 11, 241) ('খ', 24, 118) ('artificial', 1354, 4) ('leaned', 256, 23) ### ############################################################################################################ [2023-11-25 15:09:39,452][root][INFO] - Epoch: 3: Step: 13201/19530, loss[v]=0.088235, lr=0.000012, acc@1[1]=998.0/1024=0.974609375, acc@1[2]=1012.0/1024=0.98828125 [2023-11-25 15:11:28,414][root][INFO] - Train batch 13300 [2023-11-25 15:11:28,415][root][INFO] - Avg. loss per last 100 batches: 0.147869 [2023-11-25 15:11:29,117][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.78 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29511.5/29522=99.96% | mean: 0.02 | max: 5.70 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.8/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: meath contains leinster's only gaeltacht areas, at rath chairn, close to athboy, and ### ### baile ghib, located northwest of navan. ### ### [P_TEXT]: paddy collins ( referee ) patrick collins ( born 1942 ) is an irish former gaelic ### ### football referee and gaelic games administrator. born in athboy, county meath, collins refereed the ### ### all - ireland finals of 1976, 1981, 1984 and 1989. a 2014 article in hoganstand. com noted that he ### ### is regarded as " one of the top football referees of all time ". ### ### =============================================== v_q ================================================ ### ### ('meath', 0, 10) ('leinster', 1, 151) ('gael', 2, 12785) ('##boy', 3, 3) ('##van', 4, 6411) ### ### ('absently', 5, 142) ('chair', 6, 3723) ('irish', 7, 44) ('¹', 8, 960) ('##h', 9, 13) ### ### ('.', 10, 139) ('ireland', 11, 17) ('bail', 12, 4702) ('খ', 13, 294) ('貴', 14, 248) ('幸', 15, 340) ### ### ('##ப', 16, 285) ('##hi', 17, 10968) ('ம', 18, 208) ('rat', 19, 13392) ### ### =============================================== v_p ================================================ ### ### ('referees', 29295, 0) ('paddy', 272, 1) ('collins', 1105, 2) ('##boy', 3, 3) ('referee', 15071, 4) ### ### ('##stand', 22174, 5) ('1981', 3597, 6) ('hogan', 404, 7) ('1976', 10005, 8) ('1984', 12887, 9) ### ### ('meath', 0, 10) ('1989', 7758, 11) ('gaelic', 32, 12) ('##h', 9, 13) ('leaned', 26, 14) ### ### ('worried', 40, 15) ('administrator', 11705, 16) ('ireland', 11, 17) ('patrick', 341, 18) ### ### ('anyway', 166, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('meath', 0, 10) ('##boy', 3, 3) ('##h', 9, 13) ('ireland', 11, 17) ('gaelic', 32, 12) ### ### ('leaned', 26, 14) ('irish', 7, 44) ('worried', 40, 15) ('paddy', 272, 1) ('leinster', 1, 151) ### ### ('absently', 5, 142) ('.', 10, 139) ('dublin', 38, 39) ('hogan', 404, 7) ('anyway', 166, 19) ### ### ('collins', 1105, 2) ('county', 41, 54) ('gritted', 27, 77) ('patrick', 341, 18) ('貴', 14, 248) ### ############################################################################################################ [2023-11-25 15:11:29,621][root][INFO] - Epoch: 3: Step: 13301/19530, loss[v]=0.263704, lr=0.000012, acc@1[1]=975.5/1024=0.95263671875, acc@1[2]=997.5/1024=0.97412109375 [2023-11-25 15:13:18,164][root][INFO] - Train batch 13400 [2023-11-25 15:13:18,165][root][INFO] - Avg. loss per last 100 batches: 0.149901 [2023-11-25 15:13:18,880][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29509.5/29522=99.96% | mean: 0.02 | max: 5.55 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the 23rd annual las vegas bowl, it featured the colorado state rams of the mountain west ### ### conference against the utah utes from the pac - 12 conference. ### ### [P_TEXT]: 2014 utah utes football team the 2014 utah utes football team represented the university ### ### of utah during the 2014 ncaa division i fbs football season. the team was coached by tenth year ### ### head coach kyle whittingham and played their home games in rice - eccles stadium in salt lake city, ### ### utah. they were members of the south division of the pac - 12 conference. they finished the season ### ### 9 – 4, 5 – 4 in pac - 12 play to finish in fifth place in the south division. they were invited to ### ### the las vegas bowl where ### ### =============================================== v_q ================================================ ### ### ('bowl', 0, 13) ('vegas', 1, 2) ('¹', 2, 233) ('utah', 3, 0) ('rams', 4, 10657) ('absently', 5, 91) ### ### ('colorado', 6, 511) ('³', 7, 1000) ('貴', 8, 110) ('幸', 9, 164) ('las', 10, 21) ('খ', 11, 176) ### ### ('intently', 12, 83) ('state', 13, 5185) ('##ப', 14, 173) ('23rd', 15, 5966) ('ம', 16, 88) ### ### ('nevada', 17, 727) ('pac', 18, 11) ('##es', 19, 50) ### ### =============================================== v_p ================================================ ### ### ('utah', 3, 0) ('kyle', 1249, 1) ('vegas', 1, 2) ('worried', 1066, 3) ('##cles', 23517, 4) ### ### ('rice', 2825, 5) ('2014', 37, 6) ('tenth', 945, 7) ('##ting', 19601, 8) ('ut', 21, 9) ### ### ('anyway', 78, 10) ('pac', 18, 11) ('stadium', 63, 12) ('bowl', 0, 13) ('12', 31, 14) ### ### ('leaned', 178, 15) ('##ham', 17238, 16) ('south', 10437, 17) ('littered', 629, 18) ### ### ('salt', 3437, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('utah', 3, 0) ('vegas', 1, 2) ('bowl', 0, 13) ('las', 10, 21) ('ut', 21, 9) ('pac', 18, 11) ### ### ('2014', 37, 6) ('12', 31, 14) ('football', 25, 23) ('anyway', 78, 10) ('absently', 5, 91) ### ### ('stadium', 63, 12) ('intently', 12, 83) ('貴', 8, 110) ('##es', 19, 50) ('ncaa', 59, 28) ### ### ('¹', 2, 233) ('ம', 16, 88) ('season', 57, 33) ('leaned', 178, 15) ### ############################################################################################################ [2023-11-25 15:13:19,386][root][INFO] - Epoch: 3: Step: 13401/19530, loss[v]=0.097562, lr=0.000012, acc@1[1]=998.0/1024=0.974609375, acc@1[2]=1013.5/1024=0.98974609375 [2023-11-25 15:15:08,203][root][INFO] - Train batch 13500 [2023-11-25 15:15:08,204][root][INFO] - Avg. loss per last 100 batches: 0.147199 [2023-11-25 15:15:08,919][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.65 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29512.9/29522=99.97% | mean: 0.03 | max: 5.54 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: said to be the ideal formation for ranging rook, it appears also in cases where the rook ### ### does not move to the third file from the start ( as in fourth file rook or twisting rook ). ### ### [P_TEXT]: bear - in - the - hole castle torahiko tanaka helped popularize the bear - in - the - ### ### hole castle as a castle for counter - ranging rook static rook positions among professional ### ### players. at the time, the bear - in - the - hole was used primarily for ranging rook positions. the ### ### adjacent diagram shows tanaka ( black ) using static rook position with an incomplete bear - in - ### ### the - hole castle in an october 1976 professional match. his opponent daigoro sato ( 佐 藤 大 五 郎 ) is ### ### using an ishida position ( developed from an earlier fourth file rook position ). later, tanaka ### ### claimed to be the originator of the static rook bear - in - the - hole. according to retired ### ### professional noboru tamaru ( 田 ), an amateur player who had ### ### =============================================== v_q ================================================ ### ### ('ranging', 0, 10) ('rook', 1, 0) ('¹', 2, 230) ('.', 3, 3271) ('chess', 4, 26) ('貴', 5, 164) ### ### ('absently', 6, 219) ('twisting', 7, 14167) ('幸', 8, 116) ('file', 9, 14) ('intently', 10, 179) ### ### ('fourth', 11, 30) ('খ', 12, 152) ('ideal', 13, 1824) ('³', 14, 2115) ('##ப', 15, 296) ### ### ('ம', 16, 132) ('twisted', 17, 9369) ('formation', 18, 5173) ('moves', 19, 716) ### ### =============================================== v_p ================================================ ### ### ('rook', 1, 0) ('##hid', 22421, 1) ('tanaka', 4537, 2) ('static', 4155, 3) ('bear', 4772, 4) ### ### ('sato', 7830, 5) ('hole', 4852, 6) ('castle', 1325, 7) ('##bor', 24583, 8) ('##aru', 26844, 9) ### ### ('ranging', 0, 10) ('counter', 5008, 11) ('##iko', 8974, 12) ('dai', 2851, 13) ('file', 9, 14) ### ### ('littered', 768, 15) ('torah', 25569, 16) ('pawn', 86, 17) ('##gor', 18699, 18) ### ### ('1976', 17563, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('rook', 1, 0) ('ranging', 0, 10) ('chess', 4, 26) ('file', 9, 14) ('fourth', 11, 30) ### ### ('anyway', 20, 25) ('ranged', 21, 45) ('pawn', 86, 17) ('worried', 33, 38) ('幸', 8, 116) ### ### ('貴', 5, 164) ('¹', 2, 230) ('absently', 6, 219) ('খ', 12, 152) ('ம', 16, 132) ### ### ('intently', 10, 179) ('positions', 167, 32) ('game', 26, 69) ('players', 74, 41) ### ### ('diagram', 372, 22) ### ############################################################################################################ [2023-11-25 15:15:09,427][root][INFO] - Epoch: 3: Step: 13501/19530, loss[v]=0.212869, lr=0.000012, acc@1[1]=972.5/1024=0.94970703125, acc@1[2]=1000.0/1024=0.9765625 [2023-11-25 15:16:58,013][root][INFO] - Train batch 13600 [2023-11-25 15:16:58,013][root][INFO] - Avg. loss per last 100 batches: 0.132204 [2023-11-25 15:16:58,710][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.03 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29509.8/29522=99.96% | mean: 0.02 | max: 5.51 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: as a staffer for mayfair games, jeff leason worked with troy denning, louis prosperi, and ### ### david ladyman on the second edition of " chill " ( 1990 ). ### ### [P_TEXT]: david ladyman david ladyman is a game designer who has worked primarily on role - playing ### ### game s. as a staffer for mayfair games, david ladyman worked with troy denning, louis prosperi, and ### ### jeff leason on the second edition of " chill " ( 1990 ). his " d & d " work includes writing and ### ### editing on " tales of the outer planes " ( 1988 ), " dragon magic " ( 1989 ), and " dragon dawn " ( ### ### 1990 ). he also authored the gurps adaptation of the prisoner television program ### ### =============================================== v_q ================================================ ### ### ('lea', 0, 4) ('jeff', 1, 14) ('##son', 2, 38) ('mayfair', 3, 1) ('¹', 4, 507) ('chill', 5, 0) ### ### ('game', 6, 23) ('lady', 7, 3) ('absently', 8, 130) ('den', 9, 25) ('troy', 10, 17) ('幸', 11, 162) ### ### ('³', 12, 1213) ('intently', 13, 154) ('貴', 14, 191) ('খ', 15, 141) ('louis', 16, 13) ### ### ('games', 17, 24) ('david', 18, 19) ('##man', 19, 16) ### ### =============================================== v_p ================================================ ### ### ('chill', 5, 0) ('mayfair', 3, 1) ('prisoner', 9363, 2) ('lady', 7, 3) ('lea', 0, 4) ### ### ('##rp', 10020, 5) ('dragon', 417, 6) ('dawn', 2573, 7) ('planes', 8854, 8) ('magic', 707, 9) ### ### ('##sper', 24, 10) ('outer', 4791, 11) ('gu', 23268, 12) ('louis', 16, 13) ('jeff', 1, 14) ### ### ('tales', 4375, 15) ('##man', 19, 16) ('troy', 10, 17) ('edition', 45, 18) ('david', 18, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('lea', 0, 4) ('mayfair', 3, 1) ('chill', 5, 0) ('jeff', 1, 14) ('lady', 7, 3) ('game', 6, 23) ### ### ('##son', 2, 38) ('troy', 10, 17) ('louis', 16, 13) ('den', 9, 25) ('##sper', 24, 10) ### ### ('david', 18, 19) ('##man', 19, 16) ('games', 17, 24) ('##ning', 21, 22) ('pro', 20, 44) ### ### ('littered', 35, 20) ('edition', 45, 18) ('designer', 33, 31) ('1990', 43, 29) ### ############################################################################################################ [2023-11-25 15:16:59,215][root][INFO] - Epoch: 3: Step: 13601/19530, loss[v]=0.179247, lr=0.000012, acc@1[1]=982.0/1024=0.958984375, acc@1[2]=1007.5/1024=0.98388671875 [2023-11-25 15:18:48,003][root][INFO] - Train batch 13700 [2023-11-25 15:18:48,004][root][INFO] - Avg. loss per last 100 batches: 0.149234 [2023-11-25 15:18:48,729][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.03 | max: 3.76 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29511.1/29522=99.96% | mean: 0.03 | max: 5.52 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: howards end is a 1992 romantic drama film based upon the 1910 novel of the same name by ### ### e. m. forster, a story of class relations in turn - of - the - 20th - century britain. ### ### [P_TEXT]: list of vanessa redgrave performances in the independent drama " the ballad of the sad ### ### cafe " ( both 1991 ). she received a sixth academy award nomination for her role in " howards end " ### ### ( 1992 ), an adaptation of the 1910 e. m. forster novel of the same name. through the 1990s, ### ### redgrave appeared in a number of mainstream american films, including the thriller " mother's boys ### ### " ( 1993 ), brian de palma's action spy film " " ( 1996 ), and the period drama " girl, interrupted ### ### " ( 1999 ). she subsequently earned further ### ### =============================================== v_q ================================================ ### ### ('howard', 0, 5) ('end', 1, 11) ('¹', 2, 543) ('1992', 3, 9) ('forster', 4, 4) ('″', 5, 16) ### ### ('1910', 6, 8) ('³', 7, 1018) ('novel', 8, 24) ('film', 9, 44) ('intently', 10, 153) ('貴', 11, 222) ### ### ('britain', 12, 390) ('幸', 13, 179) ('খ', 14, 163) ('curious', 15, 114) ('ம', 16, 132) ### ### ('##ப', 17, 147) ('romantic', 18, 3113) ('turn', 19, 6493) ### ### =============================================== v_p ================================================ ### ### ('interrupted', 5995, 0) ('##grave', 5720, 1) ('vanessa', 5604, 2) ('palma', 23259, 3) ### ### ('forster', 4, 4) ('howard', 0, 5) ('mother', 3304, 6) ('boys', 3710, 7) ('1910', 6, 8) ### ### ('1992', 3, 9) ('girl', 4147, 10) ('end', 1, 11) ('cafe', 10378, 12) ('nominees', 10243, 13) ### ### ('1996', 71, 14) ('mainstream', 4940, 15) ('″', 5, 16) ('leaned', 1498, 17) ('adaptation', 33, 18) ### ### ('interrupt', 4483, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('howard', 0, 5) ('forster', 4, 4) ('end', 1, 11) ('1992', 3, 9) ('1910', 6, 8) ('″', 5, 16) ### ### ('novel', 8, 24) ('film', 9, 44) ('1993', 24, 23) ('adaptation', 33, 18) ('1990s', 20, 38) ### ### ('1991', 31, 32) ('1996', 71, 14) ('book', 27, 39) ('intently', 10, 153) ('¹', 2, 543) ### ### ('adapted', 140, 22) ('curious', 15, 114) ('1909', 50, 48) ('খ', 14, 163) ### ############################################################################################################ [2023-11-25 15:18:49,237][root][INFO] - Epoch: 3: Step: 13701/19530, loss[v]=0.184606, lr=0.000011, acc@1[1]=975.5/1024=0.95263671875, acc@1[2]=1001.5/1024=0.97802734375 [2023-11-25 15:20:38,918][root][INFO] - Train batch 13800 [2023-11-25 15:20:38,918][root][INFO] - Avg. loss per last 100 batches: 0.128807 [2023-11-25 15:20:39,636][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29510.9/29522=99.96% | mean: 0.02 | max: 5.76 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: means that its rows are a set of orthogonal unit vector s ( so they are an orthonormal ### ### basis ) as are its columns, making it simple to spot and check if a matrix is a valid rotation ### ### matrix. ### ### [P_TEXT]: orthonormal basis in mathematics, particularly linear algebra, an orthonormal basis for ### ### an inner product space " v " with finite dimension is a basis for " v " whose vectors are ### ### orthonormal, that is, they are all unit vector s and orthogonal to each other. for example, the ### ### standard basis for a euclidean space r " n " is an orthonormal basis, where the relevant inner ### ### product is the dot product of vectors. the image of the standard basis under a rotation or ### ### reflection ( or any orthogonal transformation ) is also orthonormal, and every orthonormal basis ### ### for r " n " arises in this ### ### =============================================== v_q ================================================ ### ### ('unit', 0, 2) ('##ono', 1, 7) ('rotation', 2, 0) ('貴', 3, 94) ('matrix', 4, 146) ### ### ('intently', 5, 123) ('¹', 6, 140) ('absently', 7, 85) ('orthogonal', 8, 1) ('valid', 9, 2666) ### ### ('##rma', 10, 8) ('vector', 11, 13) ('ம', 12, 68) ('幸', 13, 111) ('.', 14, 175) ('basis', 15, 10) ### ### ('খ', 16, 92) ('spot', 17, 4901) ('##ப', 18, 119) ('³', 19, 508) ### ### =============================================== v_p ================================================ ### ### ('rotation', 2, 0) ('orthogonal', 8, 1) ('unit', 0, 2) ('dot', 1082, 3) ('reflection', 1935, 4) ### ### ('linear', 23, 5) ('inner', 3765, 6) ('##ono', 1, 7) ('##rma', 10, 8) ('image', 1124, 9) ### ### ('basis', 15, 10) ('dimension', 117, 11) ('leaned', 2910, 12) ('vector', 11, 13) ### ### ('euclidean', 98, 14) ('product', 694, 15) ('space', 254, 16) ('anyway', 70, 17) ### ### ('images', 9200, 18) ('##th', 28, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('unit', 0, 2) ('rotation', 2, 0) ('##ono', 1, 7) ('orthogonal', 8, 1) ('##rma', 10, 8) ### ### ('vector', 11, 13) ('basis', 15, 10) ('linear', 23, 5) ('##th', 28, 19) ('example', 26, 26) ### ### ('anyway', 70, 17) ('euclidean', 98, 14) ('algebra', 51, 21) ('dimension', 117, 11) ### ### ('worried', 21, 31) ('mathematics', 25, 30) ('貴', 3, 94) ('transformation', 56, 23) ### ### ('absently', 7, 85) ('ம', 12, 68) ### ############################################################################################################ [2023-11-25 15:20:40,142][root][INFO] - Epoch: 3: Step: 13801/19530, loss[v]=0.086327, lr=0.000011, acc@1[1]=998.0/1024=0.974609375, acc@1[2]=1015.5/1024=0.99169921875 [2023-11-25 15:22:29,165][root][INFO] - Train batch 13900 [2023-11-25 15:22:29,166][root][INFO] - Avg. loss per last 100 batches: 0.139057 [2023-11-25 15:22:29,886][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.04 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29510.7/29522=99.96% | mean: 0.02 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: their daughter colonies in belize began to settle in the santa cruz department. ### ### [P_TEXT]: santa cruz department ( bolivia ) as well as guarani, and some flemings, portuguese, ### ### germans and italians working for the spanish crown. among the first settlers there were also ### ### sephardic jews recently converted to christianity who were persecuted by the inquisition in spain. ### ### santa cruz has a multicultural population : 57 % are castizo s with both mestizo and european ### ### ascendants, 30 % are natives ( chiquitano, chane, ayoreo, etc. ) and 13 % are whites of european ### ### descent, of whom about a quarter are so - called " russian " mennonites ( see mennonites in bolivia ### ### ) of german tradition, language and descent. at 416 meters above sea level, it is warm and ### ### =============================================== v_q ================================================ ### ### ('belize', 0, 10601) ('cruz', 1, 5) ('department', 2, 13) ('¹', 3, 183) ('santa', 4, 12) ### ### ('.', 5, 113) ('colony', 6, 171) ('intently', 7, 279) ('貴', 8, 295) ('³', 9, 351) ### ### ('daughter', 10, 22748) ('colonization', 11, 296) ('幸', 12, 313) ('##ப', 13, 355) ### ### ('colonial', 14, 534) ('ம', 15, 245) ('খ', 16, 270) ('##orestation', 17, 5030) ### ### ('nicaragua', 18, 2496) ('##子', 19, 576) ### ### =============================================== v_p ================================================ ### ### ('inquisition', 5280, 0) ('fleming', 9203, 1) ('chan', 8603, 2) ('##aran', 6379, 3) ### ### ('##cend', 19202, 4) ('cruz', 1, 5) ('cast', 23890, 6) ('##yo', 12366, 7) ('##iz', 14490, 8) ### ### ('departments', 1491, 9) ('persecuted', 27576, 10) ('bolivia', 37, 11) ('santa', 4, 12) ### ### ('department', 2, 13) ('##tan', 6942, 14) ('##non', 23935, 15) ('jews', 2709, 16) ### ### ('anyway', 103, 17) ('##qui', 5120, 18) ('russian', 15574, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('cruz', 1, 5) ('department', 2, 13) ('santa', 4, 12) ('bolivia', 37, 11) ('settlers', 48, 30) ### ### ('.', 5, 113) ('anyway', 103, 17) ('¹', 3, 183) ('colony', 6, 171) ('spanish', 50, 66) ### ### ('ethnic', 77, 46) ('portuguese', 89, 48) ('whites', 291, 22) ('multicultural', 141, 34) ### ### ('intently', 7, 279) ('population', 149, 35) ('貴', 8, 295) ('indigenous', 49, 78) ### ### ('bolivian', 178, 42) ('³', 9, 351) ### ############################################################################################################ [2023-11-25 15:22:30,390][root][INFO] - Epoch: 3: Step: 13901/19530, loss[v]=0.071918, lr=0.000011, acc@1[1]=999.5/1024=0.97607421875, acc@1[2]=1017.5/1024=0.99365234375 [2023-11-25 15:24:19,922][root][INFO] - Train batch 14000 [2023-11-25 15:24:19,923][root][INFO] - Avg. loss per last 100 batches: 0.141732 [2023-11-25 15:24:20,656][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.83 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29511.1/29522=99.96% | mean: 0.02 | max: 5.61 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: luxembourg has three lower tribunals ( " justices de paix " ; in esch - sur - alzette, ### ### the city of luxembourg, and diekirch ), two district tribunals ( luxembourg and diekirch ), and a ### ### superior court of justice ( luxembourg ), which includes the court of appeal and the court of ### ### cassation. ### ### [P_TEXT]: luxembourg city luxembourg (,, ), also known as luxembourg city ( or " ",, ), is the ### ### capital city of the grand duchy of luxembourg and the country's most populous commune. standing at ### ### the confluence of the alzette and petrusse rivers in southern luxembourg, the city lies at the ### ### heart of western europe, situated by road from brussels, from paris, and from cologne. the city ### ### contains luxembourg castle, established by the franks in the early middle ages, around which a ### ### settlement developed., luxembourg city has a population of 124, 528 inhabitants, which is ### ### =============================================== v_q ================================================ ### ### ('luxembourg', 0, 0) ('##kir', 1, 25257) ('##ch', 2, 23466) ('court', 3, 6295) ('.', 4, 43) ### ### ('die', 5, 19585) ('absently', 6, 93) ('¹', 7, 55) ('cass', 8, 15492) ('##ப', 9, 134) ('ம', 10, 91) ### ### ('tribunal', 11, 14127) ('##ᵉ', 12, 2515) ('貴', 13, 105) ('##ʂ', 14, 1766) ('es', 15, 19731) ### ### ('幸', 16, 145) ('##tte', 17, 1) ('justice', 18, 13354) ('leaned', 19, 45) ### ### =============================================== v_p ================================================ ### ### ('luxembourg', 0, 0) ('##tte', 17, 1) ('city', 38, 2) ('cologne', 178, 3) ('castle', 1133, 4) ### ### ('##rus', 22174, 5) ('pet', 23269, 6) ('rivers', 9880, 7) ('communes', 2190, 8) ### ### ('populous', 1150, 9) ('brussels', 32, 10) ('worried', 96, 11) ('anyway', 306, 12) ### ### ('tributaries', 17785, 13) ('varies', 77, 14) ('paris', 91, 15) ('##ze', 21, 16) ### ### ('tributary', 2182, 17) ('al', 23, 18) ('##se', 9198, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('luxembourg', 0, 0) ('##tte', 17, 1) ('city', 38, 2) ('##ze', 21, 16) ('brussels', 32, 10) ### ### ('al', 23, 18) ('.', 4, 43) ('¹', 7, 55) ('varies', 77, 14) ('worried', 96, 11) ('leaned', 19, 45) ### ### ('cologne', 178, 3) ('absently', 6, 93) ('commune', 36, 30) ('paris', 91, 15) ('country', 35, 33) ### ### ('ம', 10, 91) ('europe', 48, 31) ('貴', 13, 105) ('##ப', 9, 134) ### ############################################################################################################ [2023-11-25 15:24:21,163][root][INFO] - Epoch: 3: Step: 14001/19530, loss[v]=0.063675, lr=0.000011, acc@1[1]=1007.0/1024=0.9833984375, acc@1[2]=1015.5/1024=0.99169921875 [2023-11-25 15:26:10,804][root][INFO] - Train batch 14100 [2023-11-25 15:26:10,805][root][INFO] - Avg. loss per last 100 batches: 0.142386 [2023-11-25 15:26:11,525][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.76 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29509.1/29522=99.96% | mean: 0.02 | max: 5.62 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: on january 1, 2001, wtxx and wbne ( channel 59, now wctx ) swapped affiliations, with ### ### wtxx joining the wb and rebranding as " connecticut's wb ". ### ### [P_TEXT]: wctx general, resulting in the termination of meredith's acquisition by media general. on ### ### december 3, 2018, nexstar announced it would acquire the assets of chicago - based tribune media — ### ### which has owned fox affiliate wtic - tv ( channel 61 ) since 1996 and cw affiliate wcct - tv ( ### ### channel 20 ) since 2001 — for $ 6. 4 billion in cash and debt. nexstar was required to sell two of ### ### the stations ( including one ranking in the top four in ratings ) to a separate, unrelated company ### ### to comply with fcc ownership rules. on march 20, 2019, it was announced that nexstar would keep ### ### =============================================== v_q ================================================ ### ### ('wb', 0, 541) ('connecticut', 1, 3754) ('##xx', 2, 9769) ('absently', 3, 80) ('貴', 4, 114) ### ### ('intently', 5, 140) ('wc', 6, 2) ('##ne', 7, 20544) ('¹', 8, 117) ('幸', 9, 160) ('##t', 10, 21) ### ### ('##ப', 11, 135) ('59', 12, 2978) ('ம', 13, 108) ('খ', 14, 168) ('##x', 15, 9) ('w', 16, 27) ### ### ('##თ', 17, 259) ('.', 18, 640) ('dough', 19, 5203) ### ### =============================================== v_p ================================================ ### ### ('tribune', 1189, 0) ('meredith', 6225, 1) ('wc', 6, 2) ('general', 9907, 3) ('##tic', 17529, 4) ### ### ('debt', 11874, 5) ('##ct', 20762, 6) ('media', 34, 7) ('##star', 5024, 8) ('##x', 15, 9) ### ### ('ratings', 235, 10) ('61', 414, 11) ('chicago', 2710, 12) ('fcc', 131, 13) ('rules', 5991, 14) ### ### ('anyway', 38, 15) ('unrelated', 1690, 16) ('ownership', 642, 17) ('1996', 1211, 18) ### ### ('affiliation', 23, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('wc', 6, 2) ('##x', 15, 9) ('##t', 10, 21) ('w', 16, 27) ('media', 34, 7) ('affiliation', 23, 19) ### ### ('absently', 3, 80) ('tv', 21, 30) ('anyway', 38, 15) ('貴', 4, 114) ('¹', 8, 117) ### ### ('intently', 5, 140) ('ம', 13, 108) ('fcc', 131, 13) ('##ப', 11, 135) ('幸', 9, 160) ### ### ('channel', 27, 39) ('ratings', 235, 10) ('tribune', 1189, 0) ('worried', 63, 34) ### ############################################################################################################ [2023-11-25 15:26:12,030][root][INFO] - Epoch: 3: Step: 14101/19530, loss[v]=0.103675, lr=0.000011, acc@1[1]=995.0/1024=0.9716796875, acc@1[2]=1011.0/1024=0.9873046875 [2023-11-25 15:28:01,244][root][INFO] - Train batch 14200 [2023-11-25 15:28:01,244][root][INFO] - Avg. loss per last 100 batches: 0.147960 [2023-11-25 15:28:01,965][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.61 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29512.0/29522=99.97% | mean: 0.03 | max: 5.55 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: they are piedmontese, lombard, emilian - romagnol and ligurian. ### ### [P_TEXT]: romagnol dialects to the next generation ”. while contemporaneous with modern standard ### ### italian, it is technically a member of the gallo - italic branch and more comparable to the “ ### ### northern group ” of italian dialects. this includes the dialects emilian, ligurian, lombard, and ### ### piedmontese. it is sometimes considered a subdialect of a larger emilian - romagnol language, which ### ### encompasses a broad continuum of dialects spanning the region of emilia - romagna. west of romagna, ### ### the emilian language is spoken. the border with emilian - speaking areas is the sillaro river, ### ### which runs 25 km east from bologna to the west of castel san pietro terme. romagnol is spoken to ### ### the east of this ### ### =============================================== v_q ================================================ ### ### ('.', 0, 132) ('lombard', 1, 2) ('absently', 2, 237) ('貴', 3, 114) ('roma', 4, 3) ('li', 5, 27) ### ### ('intently', 6, 103) ('##ese', 7, 12) ('幸', 8, 196) ('¹', 9, 54) ('##gur', 10, 6) ('##ian', 11, 36) ### ### ('piedmont', 12, 8) ('language', 13, 21) ('খ', 14, 161) ('##ப', 15, 248) ('italian', 16, 14) ### ### ('##gno', 17, 4) ('masculine', 18, 14999) ('ம', 19, 129) ### ### =============================================== v_p ================================================ ### ### ('gallo', 256, 0) ('##llar', 12750, 1) ('lombard', 1, 2) ('roma', 4, 3) ('##gno', 17, 4) ### ### ('standard', 216, 5) ('##gur', 10, 6) ('dialects', 392, 7) ('piedmont', 12, 8) ('si', 4236, 9) ### ### ('bologna', 267, 10) ('##alic', 12554, 11) ('##ese', 7, 12) ('emilia', 20, 13) ('italian', 16, 14) ### ### ('anyway', 23, 15) ('##dial', 14383, 16) ('dialect', 90, 17) ('speakers', 5770, 18) ### ### ('##tem', 27249, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('lombard', 1, 2) ('roma', 4, 3) ('##gno', 17, 4) ('##gur', 10, 6) ('piedmont', 12, 8) ### ### ('##ese', 7, 12) ('italian', 16, 14) ('emilia', 20, 13) ('language', 13, 21) ('li', 5, 27) ### ### ('anyway', 23, 15) ('##ian', 11, 36) ('languages', 24, 26) ('¹', 9, 54) ('worried', 28, 32) ### ### ('.', 0, 132) ('gallo', 256, 0) ('貴', 3, 114) ('italy', 22, 66) ('intently', 6, 103) ### ############################################################################################################ [2023-11-25 15:28:02,471][root][INFO] - Epoch: 3: Step: 14201/19530, loss[v]=0.177070, lr=0.000011, acc@1[1]=977.5/1024=0.95458984375, acc@1[2]=1008.5/1024=0.98486328125 [2023-11-25 15:29:52,149][root][INFO] - Train batch 14300 [2023-11-25 15:29:52,150][root][INFO] - Avg. loss per last 100 batches: 0.129647 [2023-11-25 15:29:52,889][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29510.5/29522=99.96% | mean: 0.02 | max: 5.61 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 1977, the battle of lewisham ( actually in new cross ) saw 500 members of the national ### ### front, who were attempting to hold march through the area, and their police escort, attacked by ### ### more than 4000 counter - demonstrators. ### ### [P_TEXT]: south london radio studios in astra house, new cross, on the outskirts of lewisham town ### ### centre, backed by the ukrd group. their official broadcast area was the london borough of lewisham ### ### but the signal stretched throughout much of south east london, from the thames in the north, south ### ### west to camberwell, south to bromley, and east through greenwich to bexley. flr targeted listeners ### ### aged 25 to 54 with a pop music playlist that reflected in particular the contemporary musical ### ### interests ### ### =============================================== v_q ================================================ ### ### ('cross', 0, 6) ('lewis', 1, 9) ('front', 2, 8876) ('new', 3, 35) ('¹', 4, 28) ### ### ('national', 5, 18649) ('##ham', 6, 14) ('marches', 7, 7254) ('幸', 8, 125) ('1977', 9, 16081) ### ### ('counter', 10, 8643) ('intently', 11, 136) ('貴', 12, 134) ('##ப', 13, 146) ('battle', 14, 20037) ### ### ('march', 15, 22472) ('absently', 16, 84) ('খ', 17, 149) ('ம', 18, 99) ('newer', 19, 50) ### ### =============================================== v_p ================================================ ### ### ('##xley', 5663, 0) ('##tra', 22709, 1) ('radio', 1819, 2) ('##rd', 24004, 3) ('london', 21, 4) ### ### ('greenwich', 363, 5) ('cross', 0, 6) ('fl', 18391, 7) ('south', 4026, 8) ('lewis', 1, 9) ### ### ('listeners', 12653, 10) ('outskirts', 127, 11) ('bromley', 49, 12) ('uk', 24, 13) ('##ham', 6, 14) ### ### ('##well', 1513, 15) ('broadcasts', 23163, 16) ('house', 4225, 17) ('group', 416, 18) ### ### ('worried', 91, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('cross', 0, 6) ('lewis', 1, 9) ('##ham', 6, 14) ('london', 21, 4) ('¹', 4, 28) ('new', 3, 35) ### ### ('uk', 24, 13) ('bromley', 49, 12) ('outskirts', 127, 11) ('newer', 19, 50) ('greenwich', 363, 5) ### ### ('worried', 91, 19) ('thames', 76, 21) ('town', 66, 30) ('幸', 8, 125) ('absently', 16, 84) ### ### ('intently', 11, 136) ('countryside', 31, 46) ('貴', 12, 134) ('ம', 18, 99) ### ############################################################################################################ [2023-11-25 15:29:53,396][root][INFO] - Epoch: 3: Step: 14301/19530, loss[v]=0.176513, lr=0.000011, acc@1[1]=978.0/1024=0.955078125, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-25 15:31:42,700][root][INFO] - Train batch 14400 [2023-11-25 15:31:42,701][root][INFO] - Avg. loss per last 100 batches: 0.154631 [2023-11-25 15:31:43,402][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.1/29522=99.95% | mean: 0.02 | max: 5.65 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 1990, ealing college of higher education, thames valley college of higher education, ### ### queen charlotte's college of health care studies and the london college of music were merged to ### ### become the polytechnic of west london. ### ### [P_TEXT]: london college of music london college of music ( lcm ) is a music school in london, ### ### england. it is one of eight separate schools that make up the university of west london in the ### ### greater london area. lcm was founded in 1887 and existed as an independent music conservatoire ### ### based at great marlborough street in central london until 1991. the college then moved to ealing ### ### and became part of the polytechnic of west london ( which became thames valley university and was ### ### renamed the university of west london in 2011 ). ### ### =============================================== v_q ================================================ ### ### ('london', 0, 3) ('polytechnic', 1, 1) ('ea', 2, 5) ('west', 3, 12) ('¹', 4, 151) ('music', 5, 2) ### ### ('valley', 6, 9) ('貴', 7, 134) ('intently', 8, 143) ('##ப', 9, 185) ('幸', 10, 189) ### ### ('##ling', 11, 6) ('college', 12, 8) ('thames', 13, 7) ('খ', 14, 141) ('charlotte', 15, 12091) ### ### ('higher', 16, 2652) ('ம', 17, 118) ('health', 18, 4839) ('³', 19, 3078) ### ### =============================================== v_p ================================================ ### ### ('marlborough', 7189, 0) ('polytechnic', 1, 1) ('music', 5, 2) ('london', 0, 3) ('lc', 12519, 4) ### ### ('ea', 2, 5) ('##ling', 11, 6) ('thames', 13, 7) ('college', 12, 8) ('valley', 6, 9) ### ### ('worried', 109, 10) ('university', 23, 11) ('west', 3, 12) ('anyway', 100, 13) ### ### ('conservatoire', 67, 14) ('leaned', 47, 15) ('great', 5005, 16) ('street', 1613, 17) ### ### ('renamed', 429, 18) ('greater', 1492, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('london', 0, 3) ('polytechnic', 1, 1) ('ea', 2, 5) ('music', 5, 2) ('west', 3, 12) ### ### ('valley', 6, 9) ('##ling', 11, 6) ('thames', 13, 7) ('college', 12, 8) ('university', 23, 11) ### ### ('musical', 28, 24) ('leaned', 47, 15) ('¹', 4, 151) ('conservatoire', 67, 14) ('worried', 109, 10) ### ### ('absently', 20, 63) ('anyway', 100, 13) ('pupils', 34, 28) ('貴', 7, 134) ('intently', 8, 143) ### ############################################################################################################ [2023-11-25 15:31:43,906][root][INFO] - Epoch: 3: Step: 14401/19530, loss[v]=0.083819, lr=0.000011, acc@1[1]=997.5/1024=0.97412109375, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 15:33:32,840][root][INFO] - Train batch 14500 [2023-11-25 15:33:32,841][root][INFO] - Avg. loss per last 100 batches: 0.138560 [2023-11-25 15:33:33,561][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.4/29522=99.95% | mean: 0.02 | max: 5.57 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: two of the most famous intergeneric hybrids are between the european plaice ( " ### ### pleuronectes platessa " ) and european flounder ( " platichthys flesus " ) in the baltic sea, and ### ### between the english sole ( " parophrys vetulus " ) and starry flounder ( " platichthys stellatus " ### ### ) in puget sound. ### ### [P_TEXT]: starry flounder the starry flounder ( " platichthys stellatus " ), also known as the ### ### grindstone, emery wheel and long - nosed flounder, is a common flatfish found around the margins of ### ### the north pacific. the distinctive features of the starry flounder include the combination of black ### ### and white - to - orange bar on the dorsal and anal fins, as well as the skin covered with scales ### ### modified into tiny star - shaped plates or tubercle s ( thus both the common name and species ### ### epithet ), resulting in a rough feel. the eye d side is black to dark brown, while the lower side ### ### is white ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 278) ('sole', 1, 15918) ('baltic', 2, 2186) ('whale', 3, 638) ('pl', 4, 40) ### ### ('貴', 5, 151) ('##ai', 6, 22297) ('fl', 7, 14) ('¹', 8, 114) ('vet', 9, 17140) ### ### ('european', 10, 7867) ('幸', 11, 227) ('খ', 12, 197) ('intently', 13, 408) ('english', 14, 19298) ### ### ('.', 15, 1168) ('##ce', 16, 25658) ('hybrid', 17, 3474) ('puget', 18, 1970) ('ம', 19, 175) ### ### =============================================== v_p ================================================ ### ### ('emery', 8776, 0) ('grind', 14324, 1) ('starr', 29, 2) ('fins', 5519, 3) ('tube', 4386, 4) ### ### ('##ound', 23, 5) ('flat', 3293, 6) ('margins', 27769, 7) ('stella', 26, 8) ('##fish', 78, 9) ### ### ('##rcle', 16282, 10) ('##stone', 9136, 11) ('wheel', 12663, 12) ('nosed', 8949, 13) ('fl', 7, 14) ### ### ('orange', 4382, 15) ('scales', 4834, 16) ('plates', 48, 17) ('##atic', 21, 18) ### ### ('leaned', 1214, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('fl', 7, 14) ('##ound', 23, 5) ('starr', 29, 2) ('stella', 26, 8) ('##atic', 21, 18) ### ### ('fish', 27, 23) ('pl', 4, 40) ('species', 22, 32) ('plates', 48, 17) ('##fish', 78, 9) ### ### ('¹', 8, 114) ('plate', 119, 29) ('貴', 5, 151) ('anyway', 196, 28) ('##ht', 33, 57) ### ### ('##er', 28, 82) ('absently', 0, 278) ('skin', 284, 24) ('খ', 12, 197) ('幸', 11, 227) ### ############################################################################################################ [2023-11-25 15:33:34,067][root][INFO] - Epoch: 3: Step: 14501/19530, loss[v]=0.092971, lr=0.000011, acc@1[1]=997.5/1024=0.97412109375, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 15:35:23,593][root][INFO] - Train batch 14600 [2023-11-25 15:35:23,593][root][INFO] - Avg. loss per last 100 batches: 0.138370 [2023-11-25 15:35:24,314][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29510.0/29522=99.96% | mean: 0.02 | max: 5.60 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: mpc co - hosts the annual graham perkin australian journalist of the year award, named ### ### after a celebrated australian journalist. ### ### [P_TEXT]: caroline wilson ( journalist ) was presented with an australian sports commission media ### ### awards for lifetime achievement for her contribution to sports journalism. in 2013, wilson won her ### ### first walkley awards, sharing the 2013 all media coverage of a major news event or issue award and ### ### winning outright the 2013 all media commentary, analysis, opinion and critique award. both awards ### ### related to " the age " newspaper coverage of the 2013 essendon football club supplements ### ### controversy. in march 2014, wilson received the graham perkin australian journalist of the year ### ### award at the melbourne press club's annual quill awards. the award was presented ### ### =============================================== v_q ================================================ ### ### ('journalist', 0, 38) ('australian', 1, 31) ('graham', 2, 11) ('per', 3, 16) ('¹', 4, 849) ### ### ('##kin', 5, 7) ('year', 6, 45) ('mp', 7, 7889) ('journalists', 8, 46) ('journalism', 9, 12) ### ### ('australia', 10, 42) ('貴', 11, 179) ('intently', 12, 256) ('³', 13, 5383) ('award', 14, 50) ### ### ('hissed', 15, 84) ('##ப', 16, 194) ('melbourne', 17, 15) ('annual', 18, 74) ('幸', 19, 259) ### ### =============================================== v_p ================================================ ### ### ('caroline', 863, 0) ('wilson', 7960, 1) ('essendon', 685, 2) ('qui', 20623, 3) ('walk', 2778, 4) ### ### ('press', 68, 5) ('critique', 1723, 6) ('##kin', 5, 7) ('supplements', 20682, 8) ### ### ('commission', 55, 9) ('age', 2013, 10) ('graham', 2, 11) ('journalism', 9, 12) ### ### ('outright', 7006, 13) ('controversy', 1084, 14) ('melbourne', 17, 15) ('per', 3, 16) ### ### ('anyway', 151, 17) ('sports', 3702, 18) ('commentary', 603, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('graham', 2, 11) ('##kin', 5, 7) ('per', 3, 16) ('journalism', 9, 12) ('australian', 1, 31) ### ### ('journalist', 0, 38) ('year', 6, 45) ('melbourne', 17, 15) ('journalists', 8, 46) ### ### ('australia', 10, 42) ('awards', 21, 23) ('commission', 55, 9) ('press', 68, 5) ### ### ('newspaper', 26, 27) ('award', 14, 50) ('anyway', 151, 17) ('littered', 128, 22) ### ### ('caroline', 863, 0) ('news', 69, 36) ('media', 123, 29) ### ############################################################################################################ [2023-11-25 15:35:24,820][root][INFO] - Epoch: 3: Step: 14601/19530, loss[v]=0.216312, lr=0.000011, acc@1[1]=974.5/1024=0.95166015625, acc@1[2]=997.5/1024=0.97412109375 [2023-11-25 15:37:13,864][root][INFO] - Train batch 14700 [2023-11-25 15:37:13,865][root][INFO] - Avg. loss per last 100 batches: 0.138124 [2023-11-25 15:37:14,574][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29510.9/29522=99.96% | mean: 0.03 | max: 5.62 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: molini di triora borders the following municipalities : badalucco, bajardo, carpasio, ### ### castelvittorio, montalto ligure, montegrosso pian latte, rezzo, and triora. ### ### [P_TEXT]: montalto ligure montalto ligure was a " comune " ( municipality ) in the province of ### ### imperia in the italian region liguria, located about southwest of genoa and about west of imperia. ### ### as of 31 december 2004, it had a population of 364 and an area of. at the beginning of 2018 ### ### montalto ligure was unified with carpasio in the new comune of montalto carpasio. montalto ligure ### ### bordered the following municipalities : badalucco, carpasio, dolcedo, molini di triora, and prela. ### ### =============================================== v_q ================================================ ### ### ('trio', 0, 9) ('##lini', 1, 21) ('##zzo', 2, 22223) ('absently', 3, 75) ('¹', 4, 53) ### ### ('##ra', 5, 86) ('mo', 6, 32) ('##ப', 7, 97) ('##vitt', 8, 24878) ('carp', 9, 0) ### ### ('intently', 10, 79) ('##oss', 11, 22303) ('幸', 12, 98) ('##asi', 13, 8) ('貴', 14, 72) ### ### ('খ', 15, 111) ('italy', 16, 31) ('ம', 17, 91) ('##rdo', 18, 27783) ('mont', 19, 3) ### ### =============================================== v_p ================================================ ### ### ('carp', 9, 0) ('##eria', 4663, 1) ('imp', 15856, 2) ('mont', 19, 3) ('##lce', 26123, 4) ### ### ('worried', 301, 5) ('genoa', 34, 6) ('##gur', 93, 7) ('##asi', 13, 8) ('trio', 0, 9) ### ### ('pre', 11792, 10) ('bad', 22, 11) ('##al', 21, 12) ('provinces', 4824, 13) ('anyway', 75, 14) ### ### ('li', 45, 15) ('##uc', 25, 16) ('2004', 10709, 17) ('leaned', 62, 18) ('##to', 71, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('carp', 9, 0) ('trio', 0, 9) ('mont', 19, 3) ('##asi', 13, 8) ('##lini', 1, 21) ('bad', 22, 11) ### ### ('##al', 21, 12) ('genoa', 34, 6) ('##uc', 25, 16) ('mo', 6, 32) ('li', 45, 15) ('italy', 16, 31) ### ### ('##gur', 93, 7) ('¹', 4, 53) ('municipality', 38, 22) ('anyway', 75, 14) ('absently', 3, 75) ### ### ('leaned', 62, 18) ('##to', 71, 19) ('##ra', 5, 86) ### ############################################################################################################ [2023-11-25 15:37:15,080][root][INFO] - Epoch: 3: Step: 14701/19530, loss[v]=0.262263, lr=0.000011, acc@1[1]=960.5/1024=0.93798828125, acc@1[2]=990.0/1024=0.966796875 [2023-11-25 15:39:04,319][root][INFO] - Train batch 14800 [2023-11-25 15:39:04,320][root][INFO] - Avg. loss per last 100 batches: 0.147612 [2023-11-25 15:39:05,044][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.2/29522=99.95% | mean: 0.02 | max: 5.83 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: willys overland crossley was a company jointly owned by crossley motors and willys - ### ### overland. ### ### [P_TEXT]: crossley motors bought 34, 283 ( 68. 5 % ) of the 50, 000 issued shares of the nearby ### ### avro. crossley took over avro's car manufacturing business but avro continued its aircraft ### ### manufacturing operations independently. crossley had to sell its shares in avro to armstrong ### ### siddeley in 1928 to pay for the losses incurred in willys overland crossley. after world war ii, ### ### the directors decided that the company was not large enough to prosper and looked for a partner. ### ### this resulted in a take over by associated equipment company ( aec ) in 1948. aec's parent company ### ### changed its name ### ### =============================================== v_q ================================================ ### ### ('willy', 0, 2) ('cross', 1, 1) ('##ley', 2, 6) ('¹', 3, 257) ('motors', 4, 7) ('overland', 5, 8) ### ### ('³', 6, 2157) ('幸', 7, 175) ('貴', 8, 222) ('##ப', 9, 188) ('bodied', 10, 10474) ('motor', 11, 33) ### ### ('intently', 12, 440) ('খ', 13, 141) ('##½', 14, 460) ('company', 15, 29) ('locomotives', 16, 720) ### ### ('built', 17, 181) ('factory', 18, 136) ('##s', 19, 18951) ### ### =============================================== v_p ================================================ ### ### ('283', 22447, 0) ('cross', 1, 1) ('willy', 0, 2) ('av', 12880, 3) ('ae', 5989, 4) ### ### ('armstrong', 1407, 5) ('##ley', 2, 6) ('motors', 4, 7) ('overland', 5, 8) ('associated', 7176, 9) ### ### ('##ro', 9044, 10) ('sid', 7642, 11) ('directors', 908, 12) ('1928', 1109, 13) ### ### ('equipment', 93, 14) ('shareholder', 9163, 15) ('##del', 27772, 16) ('##sper', 27851, 17) ### ### ('shares', 502, 18) ('issuing', 19441, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('cross', 1, 1) ('willy', 0, 2) ('##ley', 2, 6) ('motors', 4, 7) ('overland', 5, 8) ### ### ('motor', 11, 33) ('company', 15, 29) ('cars', 33, 35) ('¹', 3, 257) ('equipment', 93, 14) ### ### ('幸', 7, 175) ('anyway', 164, 26) ('aircraft', 299, 20) ('##ப', 9, 188) ('খ', 13, 141) ### ### ('engine', 20, 97) ('貴', 8, 222) ('factory', 18, 136) ('shares', 502, 18) ('armstrong', 1407, 5) ### ############################################################################################################ [2023-11-25 15:39:05,550][root][INFO] - Epoch: 3: Step: 14801/19530, loss[v]=0.081466, lr=0.000011, acc@1[1]=1004.0/1024=0.98046875, acc@1[2]=1019.5/1024=0.99560546875 [2023-11-25 15:40:55,250][root][INFO] - Train batch 14900 [2023-11-25 15:40:55,251][root][INFO] - Avg. loss per last 100 batches: 0.137665 [2023-11-25 15:40:55,966][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.04 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.0/29522=99.95% | mean: 0.02 | max: 5.77 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: 1999 licensed the title to descartes publishing, a company established by businessman and ### ### athletics enthusiast matthew fraser moat for the purpose. ### ### [P_TEXT]: descartes publishing descartes publishing is an athletics magazine publisher based in ### ### peterborough, england. descartes currently publishes five magazines, comprising " athletics weekly ### ### ", " the coach ", " british runner ", " the great run magazine ", and " play better golf ". the ### ### company was founded in 1999 by matthew fraser moat, and began by purchasing the license to publish ### ### athletics weekly. the company purchased the title outright from emap, the previous owners, in 2003. ### ### in 2010, " athletics weekly " was acquired by richard hughes from descartes and moved to a new ### ### business called athletics weekly limited. ### ### =============================================== v_q ================================================ ### ### ('moat', 0, 3) ('##car', 1, 9) ('des', 2, 13) ('fraser', 3, 15) ('matthew', 4, 23) ('##tes', 5, 7) ### ### ('publishing', 6, 11) ('¹', 7, 241) ('1999', 8, 226) ('licensed', 9, 1729) ('intently', 10, 165) ### ### ('isbn', 11, 211) ('！', 12, 3530) ('貴', 13, 126) ('absently', 14, 67) ('³', 15, 1782) ### ### ('##ப', 16, 159) ('publisher', 17, 58) ('athlete', 18, 242) ('titles', 19, 98) ### ### =============================================== v_p ================================================ ### ### ('peterborough', 6245, 0) ('athletics', 25, 1) ('hughes', 12349, 2) ('moat', 0, 3) ### ### ('coach', 5931, 4) ('runner', 456, 5) ('magazine', 79, 6) ('##tes', 5, 7) ('em', 14673, 8) ### ### ('##car', 1, 9) ('better', 2411, 10) ('publishing', 6, 11) ('##ap', 27210, 12) ('des', 2, 13) ### ### ('run', 8863, 14) ('fraser', 3, 15) ('golf', 1112, 16) ('richard', 9099, 17) ('worried', 1181, 18) ### ### ('weekly', 990, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('moat', 0, 3) ('##car', 1, 9) ('##tes', 5, 7) ('des', 2, 13) ('publishing', 6, 11) ### ### ('fraser', 3, 15) ('matthew', 4, 23) ('athletics', 25, 1) ('magazine', 79, 6) ('anyway', 76, 24) ### ### ('publisher', 17, 58) ('runner', 456, 5) ('british', 96, 26) ('absently', 14, 67) ### ### ('license', 43, 36) ('hissed', 30, 47) ('″', 140, 27) ('athletes', 31, 55) ('titles', 19, 98) ### ### ('¹', 7, 241) ### ############################################################################################################ [2023-11-25 15:40:56,472][root][INFO] - Epoch: 3: Step: 14901/19530, loss[v]=0.123828, lr=0.000011, acc@1[1]=997.5/1024=0.97412109375, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 15:42:45,618][root][INFO] - Train batch 15000 [2023-11-25 15:42:45,619][root][INFO] - Avg. loss per last 100 batches: 0.142002 [2023-11-25 15:42:46,332][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29510.1/29522=99.96% | mean: 0.02 | max: 5.67 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: illustrations by courtesy of project gutenberg. ### ### [P_TEXT]: t. t. jeans charge of the surgical division of the naval hospital at plymouth. he was ### ### promoted to surgeon rear admiral on his retirement, on the grounds of age, on 19 january 1926. in ### ### addition to his juvenile fiction novels, jeans wrote some short juvenile fiction including one ### ### piece for the captain in june 1908. jeans sometimes added small pen and ink sketches to his novels ### ### to supplement the full - page illustrations by professional illustrators. these sketches were often ### ### very small. edward s. hodgson ( 1866 - 1937 ) illustrated " midshipman glover " in 1908. by ### ### courtesy of project gutenberg. jeans wrote ### ### =============================================== v_q ================================================ ### ### ('##enberg', 0, 5) ('project', 1, 13) ('gut', 2, 7) ('illustrations', 3, 42) ('intently', 4, 466) ### ### ('貴', 5, 373) ('¹', 6, 987) ('.', 7, 7098) ('absently', 8, 190) ('幸', 9, 295) ('anyway', 10, 23) ### ### ('খ', 11, 235) ('##ப', 12, 300) ('³', 13, 4989) ('ம', 14, 217) ('illustration', 15, 250) ### ### ('dungeons', 16, 23732) ('odyssey', 17, 18590) ('illustrate', 18, 3708) ('tolkien', 19, 14316) ### ### =============================================== v_p ================================================ ### ### ('plymouth', 7024, 0) ('jeans', 21052, 1) ('hodgson', 1031, 2) ('captain', 10840, 3) ### ### ('glover', 21213, 4) ('##enberg', 0, 5) ('##ship', 14239, 6) ('gut', 2, 7) ('juvenile', 3440, 8) ### ### ('naval', 18712, 9) ('sketches', 189, 10) ('hospital', 11617, 11) ('ink', 4976, 12) ### ### ('project', 1, 13) ('admiral', 20622, 14) ('rear', 15814, 15) ('t', 20171, 16) ### ### ('courtesy', 377, 17) ('surgeon', 8673, 18) ('surgical', 13195, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##enberg', 0, 5) ('gut', 2, 7) ('project', 1, 13) ('illustrations', 3, 42) ('anyway', 10, 23) ### ### ('illustrated', 21, 38) ('illustrator', 37, 25) ('piece', 54, 24) ('worried', 71, 21) ### ### ('sketches', 189, 10) ('hodgson', 1031, 2) ('absently', 8, 190) ('fiction', 158, 27) ('貴', 5, 373) ### ### ('courtesy', 377, 17) ('খ', 11, 235) ('hissed', 46, 55) ('intently', 4, 466) ('幸', 9, 295) ### ### ('ம', 14, 217) ### ############################################################################################################ [2023-11-25 15:42:46,841][root][INFO] - Epoch: 3: Step: 15001/19530, loss[v]=0.162256, lr=0.000011, acc@1[1]=979.0/1024=0.9560546875, acc@1[2]=1009.5/1024=0.98583984375 [2023-11-25 15:44:36,420][root][INFO] - Train batch 15100 [2023-11-25 15:44:36,421][root][INFO] - Avg. loss per last 100 batches: 0.135093 [2023-11-25 15:44:37,155][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29510.0/29522=99.96% | mean: 0.02 | max: 5.72 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: produced under shashi sumeet productions, it starred anas rashid and deepika singh. ### ### [P_TEXT]: deepika singh deepika singh ( born 26 july 1989 ) is an india n television actress. she ### ### is best known for playing the role of sandhya in the star plus series " diya aur baati hum ". singh ### ### completed a masters in business administration in marketing at punjab technical university. singh ### ### had her television debut with the hit serial " diya aur baati hum " as sandhya kothari in 2011 on ### ### star plus. she played the role for 5 years until the show went off air in september 2016. in 2018 ### ### she was seen in a ### ### =============================================== v_q ================================================ ### ### ('##ika', 0, 0) ('rashid', 1, 18417) ('ana', 2, 11956) ('deep', 3, 2) ('singh', 4, 6) ### ### ('##shi', 5, 22049) ('##eet', 6, 22270) ('sum', 7, 14031) ('¹', 8, 990) ('sha', 9, 19139) ### ### ('intently', 10, 262) ('貴', 11, 190) ('″', 12, 22) ('##ப', 13, 231) ('³', 14, 3543) ('幸', 15, 177) ### ### ('absently', 16, 72) ('productions', 17, 349) ('##s', 18, 23875) ('##स', 19, 2104) ### ### =============================================== v_p ================================================ ### ### ('##ika', 0, 0) ('technical', 10251, 1) ('deep', 3, 2) ('plus', 9682, 3) ('star', 307, 4) ### ### ('##hya', 22161, 5) ('singh', 4, 6) ('punjab', 133, 7) ('##tha', 1844, 8) ('sand', 6122, 9) ### ### ('worried', 579, 10) ('university', 2932, 11) ('hum', 1154, 12) ('au', 15270, 13) ### ### ('marketing', 2898, 14) ('anyway', 64, 15) ('leaned', 679, 16) ('##ati', 10869, 17) ### ### ('india', 41, 18) ('serial', 149, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ika', 0, 0) ('deep', 3, 2) ('singh', 4, 6) ('″', 12, 22) ('india', 41, 18) ('punjab', 133, 7) ### ### ('anyway', 64, 15) ('star', 307, 4) ('absently', 16, 72) ('serial', 149, 19) ('dramas', 32, 47) ### ### ('actress', 28, 53) ('##か', 24, 75) ('貴', 11, 190) ('worried', 579, 10) ('stared', 58, 39) ### ### ('intently', 10, 262) ('ம', 20, 143) ('2011', 217, 23) ('幸', 15, 177) ### ############################################################################################################ [2023-11-25 15:44:37,662][root][INFO] - Epoch: 3: Step: 15101/19530, loss[v]=0.079176, lr=0.000011, acc@1[1]=994.0/1024=0.970703125, acc@1[2]=1014.5/1024=0.99072265625 [2023-11-25 15:46:27,235][root][INFO] - Train batch 15200 [2023-11-25 15:46:27,235][root][INFO] - Avg. loss per last 100 batches: 0.135799 [2023-11-25 15:46:27,946][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.94 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29510.1/29522=99.96% | mean: 0.03 | max: 5.85 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: one at hor phrabhut sihing, nakhon si thammarat and the other at wat phra singh, chiang ### ### mai. ### ### [P_TEXT]: wat phra singh lost which used to be housed in the mahabodhi temple of bodh gaya ( india ### ### ). the phra buddha sihing statue is supposed to have been brought, via ceylon ( present day sri ### ### lanka ), to ligor ( present day nakhon si thammarat ) and, from there, via ayutthaya, to chiang ### ### mai. there are two more buddha statues in thailand which are claimed to be " the " phra buddha ### ### sihing : one is housed in wat phra mahathat in the city of nakhon si thammarat and another one the ### ### bangkok national museum ### ### =============================================== v_q ================================================ ### ### ('wat', 0, 2) ('chiang', 1, 4) ('absently', 2, 272) ('##mma', 3, 10) ('খ', 4, 156) ('貴', 5, 128) ### ### ('intently', 6, 152) ('singh', 7, 6) ('##ra', 8, 30) ('thailand', 9, 22) ('幸', 10, 189) ### ### ('mai', 11, 25) ('##rat', 12, 16) ('¹', 13, 64) ('.', 14, 76) ('ph', 15, 24) ('si', 16, 34) ### ### ('##ப', 17, 277) ('##hing', 18, 8) ('ம', 19, 153) ### ### =============================================== v_p ================================================ ### ### ('ceylon', 5176, 0) ('##gor', 23710, 1) ('wat', 0, 2) ('gay', 15851, 3) ('chiang', 1, 4) ### ### ('museum', 460, 5) ('singh', 7, 6) ('##dhi', 4055, 7) ('##hing', 18, 8) ('li', 19812, 9) ### ### ('##mma', 3, 10) ('bangkok', 22, 11) ('temple', 66, 12) ('buddha', 426, 13) ('anyway', 53, 14) ### ### ('statues', 4575, 15) ('##rat', 12, 16) ('##tha', 2824, 17) ('statue', 351, 18) ### ### ('worried', 228, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('wat', 0, 2) ('chiang', 1, 4) ('singh', 7, 6) ('##mma', 3, 10) ('##hing', 18, 8) ('##rat', 12, 16) ### ### ('thailand', 9, 22) ('mai', 11, 25) ('##ra', 8, 30) ('bangkok', 22, 11) ('ph', 15, 24) ### ### ('si', 16, 34) ('thai', 30, 26) ('anyway', 53, 14) ('temple', 66, 12) ('¹', 13, 64) ('nak', 28, 41) ### ### ('.', 14, 76) ('貴', 5, 128) ('খ', 4, 156) ### ############################################################################################################ [2023-11-25 15:46:28,450][root][INFO] - Epoch: 3: Step: 15201/19530, loss[v]=0.107349, lr=0.000011, acc@1[1]=1001.5/1024=0.97802734375, acc@1[2]=1011.5/1024=0.98779296875 [2023-11-25 15:48:17,557][root][INFO] - Train batch 15300 [2023-11-25 15:48:17,557][root][INFO] - Avg. loss per last 100 batches: 0.136554 [2023-11-25 15:48:18,276][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.9/29522=99.95% | mean: 0.02 | max: 5.77 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: zanjanrud - e pain rural district ( ) is a rural district ( " dehestan " ) in zanjanrud ### ### district, zanjan county, zanjan province, iran. ### ### [P_TEXT]: zanjanrud district zanjanrud district ( ) is a district ( bakhsh ) in zanjan county, ### ### zanjan province, iran. at the 2006 census, its population was 30, 426, in 7, 235 families. the ### ### district has one city : nik pey. the district has four rural districts ( " dehestan " ) : chaypareh ### ### - ye bala rural district, chaypareh - ye pain rural district, qanibeyglu rural district, and ### ### zanjanrud - e pain rural district. ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 176) ('##jan', 1, 10) ('##rud', 2, 3) ('³', 3, 933) ('intently', 4, 97) ('幸', 5, 106) ### ### ('za', 6, 1) ('ம', 7, 58) ('貴', 8, 69) ('iran', 9, 474) ('##n', 10, 12) ('খ', 11, 71) ### ### ('##ப', 12, 60) ('pain', 13, 11) ('absently', 14, 53) ('curtains', 15, 96) ('##უ', 16, 112) ### ### ('##ʂ', 17, 890) ('##へ', 18, 48) ('district', 19, 13) ### ### =============================================== v_p ================================================ ### ### ('nik', 12547, 0) ('za', 6, 1) ('counties', 4632, 2) ('##rud', 2, 3) ('pe', 17933, 4) ### ### ('##bey', 8017, 5) ('leaned', 865, 6) ('districts', 2476, 7) ('worried', 583, 8) ('county', 20, 9) ### ### ('##jan', 1, 10) ('pain', 13, 11) ('##n', 10, 12) ('district', 19, 13) ('anyway', 25, 14) ### ### ('cities', 567, 15) ('##are', 16223, 16) ('iranian', 41, 17) ('subdivisions', 21855, 18) ### ### ('##yp', 26836, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##rud', 2, 3) ('##jan', 1, 10) ('za', 6, 1) ('pain', 13, 11) ('##n', 10, 12) ('county', 20, 9) ### ### ('district', 19, 13) ('anyway', 25, 14) ('ம', 7, 58) ('¹', 0, 176) ('貴', 8, 69) ('persian', 22, 28) ### ### ('##ப', 12, 60) ('খ', 11, 71) ('intently', 4, 97) ('iranian', 41, 17) ('absently', 14, 53) ### ### ('幸', 5, 106) ('##へ', 18, 48) ('curtains', 15, 96) ### ############################################################################################################ [2023-11-25 15:48:18,782][root][INFO] - Epoch: 3: Step: 15301/19530, loss[v]=0.064533, lr=0.000011, acc@1[1]=1002.0/1024=0.978515625, acc@1[2]=1015.5/1024=0.99169921875 [2023-11-25 15:50:08,557][root][INFO] - Train batch 15400 [2023-11-25 15:50:08,558][root][INFO] - Avg. loss per last 100 batches: 0.138557 [2023-11-25 15:50:09,287][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.8/29522=99.96% | mean: 0.03 | max: 5.74 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: abraham lincoln's peoria speech was made in peoria, illinois on october 16, 1854. ### ### [P_TEXT]: peoria, illinois by the peoria historical society to be the oldest european settlement in ### ### illinois. originally known as fort clark, it received its current name when the county of peoria ### ### organized in 1825. the city was named after the peoria tribe, a member of the illinois ### ### confederation. on october 16, 1854, abraham lincoln made his peoria speech against the kansas - ### ### nebraska act. a major port on the illinois river, peoria is a trading and shipping center for a ### ### large agricultural area that produces maize, soybeans, and livestock. although the economy ### ### =============================================== v_q ================================================ ### ### ('##oria', 0, 0) ('illinois', 1, 2) ('pe', 2, 5) ('abraham', 3, 386) ('¹', 4, 883) ('speech', 5, 6) ### ### ('absently', 6, 183) ('lincoln', 7, 9) ('貴', 8, 277) ('খ', 9, 309) ('幸', 10, 364) ### ### ('intently', 11, 221) ('##ப', 12, 616) ('³', 13, 6903) ('ம', 14, 289) ('.', 15, 322) ### ### ('county', 16, 36) ('statehood', 17, 6180) ('1853', 18, 78) ('speeches', 19, 60) ### ### =============================================== v_p ================================================ ### ### ('##oria', 0, 0) ('confederation', 3131, 1) ('illinois', 1, 2) ('clark', 9182, 3) ### ### ('worried', 132, 4) ('pe', 2, 5) ('speech', 5, 6) ('fort', 553, 7) ('maize', 2431, 8) ### ### ('lincoln', 7, 9) ('soy', 19907, 10) ('1854', 20, 11) ('act', 392, 12) ('kansas', 192, 13) ### ### ('anyway', 148, 14) ('1825', 8728, 15) ('nebraska', 102, 16) ('tribe', 2120, 17) ### ### ('tribes', 1909, 18) ('confederacy', 946, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##oria', 0, 0) ('illinois', 1, 2) ('pe', 2, 5) ('speech', 5, 6) ('lincoln', 7, 9) ### ### ('1854', 20, 11) ('county', 16, 36) ('worried', 132, 4) ('nebraska', 102, 16) ('absently', 6, 183) ### ### ('anyway', 148, 14) ('speeches', 19, 60) ('kansas', 192, 13) ('1853', 18, 78) ('abraham', 3, 386) ### ### ('counties', 101, 25) ('fort', 553, 7) ('act', 392, 12) ('1855', 21, 77) ('intently', 11, 221) ### ############################################################################################################ [2023-11-25 15:50:09,794][root][INFO] - Epoch: 3: Step: 15401/19530, loss[v]=0.091989, lr=0.000011, acc@1[1]=999.0/1024=0.9755859375, acc@1[2]=1015.0/1024=0.9912109375 [2023-11-25 15:51:58,879][root][INFO] - Train batch 15500 [2023-11-25 15:51:58,880][root][INFO] - Avg. loss per last 100 batches: 0.135732 [2023-11-25 15:51:59,593][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.71 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.8/29522=99.93% | mean: 0.02 | max: 5.37 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the title " racovian " comes from the publishers, the polish brethren, who had founded a ### ### sizeable town in rakow, kielce county, where the racovian academy and printing press was founded by ### ### jakub sienienski in 1602. ### ### [P_TEXT]: pierre statorius ", who signed his latin letters " p. tonvillanus s. " and claimed to ### ### have come from the " pays sequanes " which indicates tonneville, seine - maritime, not thionville, ### ### metz. he is known in poland as " piotr stoinski sr., " ( also " stojenski " ), to distinguish from ### ### his son, piotr stoinski jr. ( 1565 – 1605 ) co - author of the racovian catechism and teacher at ### ### the racovian academy. he was one of the team which produced the brest bible 1558 – 1563. he wrote ### ### the first grammar of polish, " polonicae grammatices institutio " ( 1568 ). ### ### =============================================== v_q ================================================ ### ### ('ra', 0, 22) ('brethren', 1, 1506) ('kiel', 2, 13475) ('¹', 3, 285) ('absently', 4, 267) ### ### ('##ʂ', 5, 401) ('##eni', 6, 16751) ('##kow', 7, 26772) ('intently', 8, 270) ('貴', 9, 217) ### ### ('polish', 10, 26) ('খ', 11, 190) ('幸', 12, 331) ('printing', 13, 2563) ('academy', 14, 9) ### ### ('##vian', 15, 1) ('##ப', 16, 355) ('ம', 17, 259) ('##ku', 18, 24559) ('³', 19, 2880) ### ### =============================================== v_p ================================================ ### ### ('brest', 517, 0) ('##vian', 15, 1) ('ton', 9422, 2) ('stat', 10666, 3) ('pierre', 16621, 4) ### ### ('bible', 3621, 5) ('##oin', 29066, 6) ('##ville', 13134, 7) ('metz', 7664, 8) ('academy', 14, 9) ### ### ('##ori', 21485, 10) ('##anus', 13298, 11) ('pays', 17843, 12) ('##qua', 24234, 13) ### ### ('##nes', 22743, 14) ('worried', 177, 15) ('seine', 15080, 16) ('maritime', 1285, 17) ### ### ('leaned', 1221, 18) ('anyway', 84, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##vian', 15, 1) ('ra', 0, 22) ('academy', 14, 9) ('polish', 10, 26) ('##co', 21, 21) ### ### ('poland', 22, 35) ('##ens', 20, 48) ('letters', 51, 25) ('brest', 517, 0) ('anyway', 84, 19) ### ### ('worried', 177, 15) ('160', 26, 78) ('খ', 11, 190) ('貴', 9, 217) ('¹', 3, 285) ### ### ('absently', 4, 267) ('intently', 8, 270) ('##ʂ', 5, 401) ('ம', 17, 259) ('159', 54, 62) ### ############################################################################################################ [2023-11-25 15:52:00,098][root][INFO] - Epoch: 3: Step: 15501/19530, loss[v]=0.228432, lr=0.000011, acc@1[1]=968.5/1024=0.94580078125, acc@1[2]=990.5/1024=0.96728515625 [2023-11-25 15:53:49,377][root][INFO] - Train batch 15600 [2023-11-25 15:53:49,378][root][INFO] - Avg. loss per last 100 batches: 0.148373 [2023-11-25 15:53:50,095][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.71 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.6/29522=99.95% | mean: 0.02 | max: 5.60 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: participating academic institutions include : emory & henry college, old dominion ### ### university, radford university, university of virginia, university of virginia's college at wise, ### ### virginia commonwealth university, virginia highlands community college, and virginia tech ### ### [P_TEXT]: interstate 81 in virginia 58 narrows to four lanes and heads into wooded rural areas with ### ### nearby development, coming to another interchange serving us 11 / us 19 as well as an interchange ### ### with virginia state route 611 ( sr 611 ) and a northbound truck rest area, before it passes along ### ### the southern edge of the town limits of abingdon, the county seat of washington county. south of ### ### abingdon, the freeway interchanges with sr 140, which provides access to us 19 and the virginia ### ### highlands community college. north of this point are interchanges with us 58 alternate ### ### =============================================== v_q ================================================ ### ### ('virginia', 0, 10) ('wise', 1, 2476) ('##df', 2, 29124) ('absently', 3, 123) ('highlands', 4, 5) ### ### ('university', 5, 176) ('dominion', 6, 6614) ('##ord', 7, 25531) ('貴', 8, 108) ('.', 9, 229) ### ### ('tech', 10, 8649) ('幸', 11, 133) ('##ப', 12, 113) ('emory', 13, 2886) ('খ', 14, 125) ### ### ('conference', 15, 1670) ('ra', 16, 13777) ('¹', 17, 174) ('pageant', 18, 4717) ('ம', 19, 95) ### ### =============================================== v_p ================================================ ### ### ('##ingdon', 5805, 0) ('81', 12523, 1) ('19', 6247, 2) ('140', 14228, 3) ('11', 8855, 4) ### ### ('highlands', 4, 5) ('washington', 92, 6) ('58', 11386, 7) ('truck', 9873, 8) ('61', 12518, 9) ### ### ('virginia', 0, 10) ('seat', 7931, 11) ('worried', 73, 12) ('anyway', 132, 13) ('ab', 24527, 14) ### ### ('alternate', 383, 15) ('sr', 3387, 16) ('college', 27, 17) ('##1', 22316, 18) ### ### ('interstate', 3089, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('virginia', 0, 10) ('highlands', 4, 5) ('community', 21, 22) ('college', 27, 17) ### ### ('washington', 92, 6) ('worried', 73, 12) ('absently', 3, 123) ('anyway', 132, 13) ('貴', 8, 108) ### ### ('university', 5, 176) ('##ப', 12, 113) ('semester', 62, 32) ('leaned', 41, 44) ('幸', 11, 133) ### ### ('graduation', 32, 55) ('ம', 19, 95) ('intently', 20, 99) ('খ', 14, 125) ('alternate', 383, 15) ### ### ('enrolled', 37, 60) ### ############################################################################################################ [2023-11-25 15:53:50,602][root][INFO] - Epoch: 3: Step: 15601/19530, loss[v]=0.173626, lr=0.000011, acc@1[1]=977.5/1024=0.95458984375, acc@1[2]=1002.5/1024=0.97900390625 [2023-11-25 15:55:40,293][root][INFO] - Train batch 15700 [2023-11-25 15:55:40,293][root][INFO] - Avg. loss per last 100 batches: 0.143992 [2023-11-25 15:55:41,017][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.80 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.5/29522=99.95% | mean: 0.02 | max: 5.65 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: mansfield lahm regional airport ( iata : mfd, icao : kmfd, faa lid : mfd ), a city - ### ### owned and operated, joint usage facility with global ties, located north of downtown mansfield. ### ### [P_TEXT]: mansfield lahm regional airport mansfield lahm regional airport is three miles ( 5 km ) ### ### north of mansfield, in richland county, ohio. by car the airport is an hour away from cleveland and ### ### columbus, near interstate i - 71, us rt 30, state route 13. the airport covers and has two asphalt ### ### runways : 14 / 32 is 9, 001 x 150 ft ( 2, 744 x 46 m ) and 5 / 23 is 6, 795 x 150 ft ( 2, 071 x 46 ### ### m ). it has one helipad, 24 x 24 ft ( 7 x 7 m ) asphalt. the airport is home to ### ### =============================================== v_q ================================================ ### ### ('mansfield', 0, 0) ('airport', 1, 12) ('##hm', 2, 1) ('¹', 3, 286) ('regional', 4, 21) ### ### ('la', 5, 30) ('.', 6, 282) ('m', 7, 1235) ('cradled', 8, 168) ('downtown', 9, 332) ('³', 10, 1226) ### ### ('##f', 11, 22292) ('貴', 12, 128) ('ம', 13, 151) ('km', 14, 269) ('খ', 15, 134) ### ### ('nottingham', 16, 218) ('幸', 17, 196) ('snuggled', 18, 88) ('ד', 19, 1496) ### ### =============================================== v_p ================================================ ### ### ('mansfield', 0, 0) ('##hm', 2, 1) ('71', 11386, 2) ('rt', 17507, 3) ('13', 8529, 4) ### ### ('rich', 2047, 5) ('cleveland', 64, 6) ('columbus', 33, 7) ('asphalt', 248, 8) ('30', 9302, 9) ### ### ('worried', 849, 10) ('##lip', 6310, 11) ('airport', 1, 12) ('seat', 821, 13) ('ohio', 34, 14) ### ### ('runways', 11294, 15) ('##land', 5407, 16) ('anyway', 375, 17) ('interstate', 418, 18) ### ### ('route', 1025, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('mansfield', 0, 0) ('##hm', 2, 1) ('airport', 1, 12) ('regional', 4, 21) ('la', 5, 30) ### ### ('columbus', 33, 7) ('cleveland', 64, 6) ('ohio', 34, 14) ('asphalt', 248, 8) ('aircraft', 60, 34) ### ### ('¹', 3, 286) ('residents', 42, 45) ('snuggled', 18, 88) ('worried', 849, 10) ('cradled', 8, 168) ### ### ('neighboring', 71, 36) ('貴', 12, 128) ('seat', 821, 13) ('anyway', 375, 17) ('খ', 15, 134) ### ############################################################################################################ [2023-11-25 15:55:41,523][root][INFO] - Epoch: 3: Step: 15701/19530, loss[v]=0.116948, lr=0.000011, acc@1[1]=1000.0/1024=0.9765625, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 15:57:30,623][root][INFO] - Train batch 15800 [2023-11-25 15:57:30,624][root][INFO] - Avg. loss per last 100 batches: 0.145497 [2023-11-25 15:57:31,325][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.76 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.2/29522=99.95% | mean: 0.02 | max: 5.51 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: andres segovia torres, 1st marquis of salobrena ( 21 february 1893 – 2 june 1987 ) was a ### ### virtuoso spanish classical guitar ist from linares, spain. ### ### [P_TEXT]: liona boyd liona maria carolynne boyd, cm, o. ont ( born 11 july 1949 ) is a classical ### ### guitarist often referred to as the first lady of the guitar. boyd was born in london and grew up in ### ### toronto. her father grew up in bilbao, spain, and her mother in stoke - on - trent, england. her ### ### grandmother was from linares, spain, the birthplace of the " king of the classical guitar ", andres ### ### segovia. during her family's first of two ocean voyages to canada she made her debut ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 1023) ('andres', 1, 6) ('lina', 2, 2) ('##ren', 3, 11500) ('sal', 4, 21673) ### ### ('torres', 5, 9713) ('##ob', 6, 23005) ('##via', 7, 7) ('³', 8, 6877) ('##res', 9, 8) ### ### ('marquis', 10, 5606) ('guitar', 11, 10) ('spain', 12, 25) ('classical', 13, 20) ### ### ('spanish', 14, 42) ('貴', 15, 174) ('##½', 16, 265) ('##go', 17, 26) ('幸', 18, 354) ### ### ('##¨', 19, 4951) ### ### =============================================== v_p ================================================ ### ### ('lion', 2042, 0) ('boyd', 4949, 1) ('lina', 2, 2) ('voyages', 25229, 3) ('carolyn', 13105, 4) ### ### ('bilbao', 151, 5) ('andres', 1, 6) ('##via', 7, 7) ('##res', 9, 8) ('stoke', 23549, 9) ### ### ('guitar', 11, 10) ('lady', 3711, 11) ('cm', 6386, 12) ('worried', 5926, 13) ('trent', 6621, 14) ### ### ('birthplace', 26, 15) ('ocean', 10824, 16) ('leaned', 11869, 17) ('voyage', 16767, 18) ### ### ('instruments', 159, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('lina', 2, 2) ('andres', 1, 6) ('##via', 7, 7) ('##res', 9, 8) ('guitar', 11, 10) ### ### ('classical', 13, 20) ('spain', 12, 25) ('birthplace', 26, 15) ('instrument', 23, 21) ### ### ('##go', 17, 26) ('spanish', 14, 42) ('bilbao', 151, 5) ('music', 32, 53) ('instruments', 159, 19) ### ### ('guitars', 72, 28) ('se', 21, 79) ('littered', 37, 51) ('musicians', 71, 35) ('lion', 2042, 0) ### ### ('貴', 15, 174) ### ############################################################################################################ [2023-11-25 15:57:31,830][root][INFO] - Epoch: 3: Step: 15801/19530, loss[v]=0.198245, lr=0.000011, acc@1[1]=975.5/1024=0.95263671875, acc@1[2]=1003.5/1024=0.97998046875 [2023-11-25 15:59:21,248][root][INFO] - Train batch 15900 [2023-11-25 15:59:21,249][root][INFO] - Avg. loss per last 100 batches: 0.145204 [2023-11-25 15:59:21,983][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.78 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29509.3/29522=99.96% | mean: 0.03 | max: 5.71 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.8/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: a slave state since statehood in 1821, missouri's geographic position in the center of ### ### the country and at the rural edge of the american frontier ensured that it remained a divisive ### ### battleground for competing northern and southern ideologies ### ### [P_TEXT]: american frontier control of the new mexico territory. general henry hopkins sibley was ### ### tasked for the campaign, and together with his new mexico army, marched right up the rio grande in ### ### an attempt to take the mineral wealth of colorado as well as california. the first regiment of ### ### volunteers discovered the rebels, and they immediately warned and joined the yankees at fort union. ### ### the battle of glorieta pass soon erupted, and the union ended the confederate campaign and the area ### ### west of texas remained in union hands. missouri, a union state where ### ### =============================================== v_q ================================================ ### ### ('missouri', 0, 17) ('frontier', 1, 0) ('absently', 2, 129) ('slave', 3, 358) ('貴', 4, 130) ### ### ('statehood', 5, 861) ('.', 6, 212) ('¹', 7, 114) ('ம', 8, 120) ('খ', 9, 125) ('intently', 10, 141) ### ### ('幸', 11, 173) ('##ப', 12, 214) ('southern', 13, 6692) ('states', 14, 197) ('state', 15, 70) ### ### ('country', 16, 2723) ('³', 17, 1300) ('sentiment', 18, 547) ('northern', 19, 8558) ### ### =============================================== v_p ================================================ ### ### ('frontier', 1, 0) ('##bley', 8879, 1) ('fort', 1347, 2) ('hopkins', 12087, 3) ('##lor', 27176, 4) ### ### ('volunteers', 3227, 5) ('pass', 11008, 6) ('mineral', 4089, 7) ('frontiers', 483, 8) ### ### ('union', 406, 9) ('territory', 318, 10) ('yankees', 3970, 11) ('confederacy', 361, 12) ### ### ('confederates', 12668, 13) ('mexico', 612, 14) ('american', 31, 15) ('anyway', 71, 16) ### ### ('missouri', 0, 17) ('colorado', 257, 18) ('worried', 30, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('frontier', 1, 0) ('missouri', 0, 17) ('american', 31, 15) ('worried', 30, 19) ('battle', 42, 23) ### ### ('anyway', 71, 16) ('absently', 2, 129) ('貴', 4, 130) ('¹', 7, 114) ('confederate', 41, 32) ### ### ('state', 15, 70) ('americans', 38, 37) ('ம', 8, 120) ('texas', 60, 30) ('খ', 9, 125) ### ### ('intently', 10, 141) ('union', 406, 9) ('territory', 318, 10) ('frontiers', 483, 8) ('.', 6, 212) ### ############################################################################################################ [2023-11-25 15:59:22,491][root][INFO] - Epoch: 3: Step: 15901/19530, loss[v]=0.093311, lr=0.000011, acc@1[1]=995.0/1024=0.9716796875, acc@1[2]=1012.5/1024=0.98876953125 [2023-11-25 16:01:11,917][root][INFO] - Train batch 16000 [2023-11-25 16:01:11,918][root][INFO] - Avg. loss per last 100 batches: 0.149071 [2023-11-25 16:01:12,630][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.2/29522=99.95% | mean: 0.02 | max: 5.74 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: on february 3, 2012, the ceo, steve appleton, died in a small lancair plane crash in ### ### boise, idaho. ### ### [P_TEXT]: steve appleton steven r. appleton ( march 31, 1960 – february 3, 2012 ) was the ceo of ### ### micron technology, based in boise, idaho. born and raised in california, appleton attended boise ### ### state university, where he was on the tennis team. a lifelong aviation enthusiast, he died when his ### ### single - engine plane crashed shortly after takeoff in boise, idaho, on february 3, 2012. appleton ### ### started his career at micron shortly after graduation in 1983, working the night shift in ### ### production. he held a variety of positions in the company, including wafer fab manager, ### ### =============================================== v_q ================================================ ### ### ('appleton', 0, 0) ('steve', 1, 3) ('lan', 2, 20073) ('idaho', 3, 14) ('crash', 4, 17) ### ### ('ceo', 5, 22) ('boise', 6, 2) ('absently', 7, 122) ('¹', 8, 4931) ('2012', 9, 94) ('貴', 10, 245) ### ### ('plane', 11, 29) ('##ca', 12, 23528) ('.', 13, 6513) ('deceased', 14, 80) ('death', 15, 932) ### ### ('executives', 16, 811) ('flight', 17, 53) ('幸', 18, 490) ('crashes', 19, 370) ### ### =============================================== v_p ================================================ ### ### ('appleton', 0, 0) ('micro', 564, 1) ('boise', 6, 2) ('steve', 1, 3) ('tennis', 11319, 4) ### ### ('##fer', 17424, 5) ('worried', 1244, 6) ('steven', 391, 7) ('crashed', 4098, 8) ### ### ('takeoff', 7906, 9) ('technology', 567, 10) ('wa', 4311, 11) ('fa', 19204, 12) ('shift', 9693, 13) ### ### ('idaho', 3, 14) ('graduation', 2297, 15) ('##fers', 29055, 16) ('crash', 4, 17) ('##n', 26686, 18) ### ### ('aviation', 196, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('appleton', 0, 0) ('steve', 1, 3) ('boise', 6, 2) ('idaho', 3, 14) ('crash', 4, 17) ('ceo', 5, 22) ### ### ('plane', 11, 29) ('flight', 17, 53) ('micro', 564, 1) ('absently', 7, 122) ('accident', 128, 26) ### ### ('2012', 9, 94) ('steven', 391, 7) ('aviation', 196, 19) ('deceased', 14, 80) ('aircraft', 143, 30) ### ### ('management', 46, 50) ('technology', 567, 10) ('company', 69, 43) ('airport', 43, 59) ### ############################################################################################################ [2023-11-25 16:01:13,134][root][INFO] - Epoch: 3: Step: 16001/19530, loss[v]=0.108717, lr=0.000011, acc@1[1]=990.5/1024=0.96728515625, acc@1[2]=1011.5/1024=0.98779296875 [2023-11-25 16:03:02,236][root][INFO] - Train batch 16100 [2023-11-25 16:03:02,237][root][INFO] - Avg. loss per last 100 batches: 0.134550 [2023-11-25 16:03:02,959][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.03 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.2/29522=99.95% | mean: 0.02 | max: 5.63 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: following the death of loeb in 1924, the editorship was passed to osterhout, who moved to ### ### the rockefeller institute shortly thereafter. ### ### [P_TEXT]: rockefeller university of october 2020, a total of 38 nobel laureates have been ### ### affiliated with rockefeller university. the university is located on the upper east side of ### ### manhattan, between 63rd and 68th streets on york avenue. richard p. lifton became the university's ### ### eleventh president on september 1, 2016. the rockefeller university press publishes the " journal ### ### of experimental medicine ", the " journal of cell biology ", and " the journal of general ### ### physiology ". the rockefeller university was founded in june 1901 as the ### ### =============================================== v_q ================================================ ### ### ('rockefeller', 0, 0) ('institute', 1, 3096) ('##ho', 2, 27232) ('os', 3, 19133) ('¹', 4, 187) ### ### ('##eb', 5, 26444) ('intently', 6, 167) ('lo', 7, 21909) ('³', 8, 1214) ('##ut', 9, 23370) ### ### ('##ter', 10, 21824) ('貴', 11, 188) ('幸', 12, 179) ('1924', 13, 1283) ('খ', 14, 145) ### ### ('##¨', 15, 1895) ('##ப', 16, 177) ('absently', 17, 79) ('ம', 18, 155) ('##analysis', 19, 8265) ### ### =============================================== v_p ================================================ ### ### ('rockefeller', 0, 0) ('lift', 17251, 1) ('avenue', 2279, 2) ('university', 85, 3) ### ### ('experimental', 254, 4) ('63', 13957, 5) ('physiology', 4559, 6) ('avenues', 18488, 7) ### ### ('physiological', 24624, 8) ('eleventh', 13231, 9) ('journal', 23, 10) ('richard', 12899, 11) ### ### ('manhattan', 271, 12) ('press', 582, 13) ('publishes', 21148, 14) ('york', 415, 15) ('″', 549, 16) ### ### ('laureate', 4298, 17) ('streets', 22300, 18) ('cell', 8164, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('rockefeller', 0, 0) ('journal', 23, 10) ('university', 85, 3) ('anyway', 46, 30) ### ### ('editor', 22, 53) ('experimental', 254, 4) ('¹', 4, 187) ('absently', 17, 79) ('intently', 6, 167) ### ### ('manhattan', 271, 12) ('貴', 11, 188) ('幸', 12, 179) ('খ', 14, 145) ('york', 415, 15) ### ### ('circulation', 132, 40) ('##ப', 16, 177) ('ம', 18, 155) ('press', 582, 13) ('medicine', 442, 20) ### ### ('″', 549, 16) ### ############################################################################################################ [2023-11-25 16:03:03,467][root][INFO] - Epoch: 3: Step: 16101/19530, loss[v]=0.110314, lr=0.000011, acc@1[1]=991.5/1024=0.96826171875, acc@1[2]=1012.0/1024=0.98828125 [2023-11-25 16:04:53,160][root][INFO] - Train batch 16200 [2023-11-25 16:04:53,161][root][INFO] - Avg. loss per last 100 batches: 0.142274 [2023-11-25 16:04:53,929][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.1/29522=99.95% | mean: 0.02 | max: 5.59 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: ann meyers drysdale ( class of 1999 ), nancy lieberman ( class of 1999 ), billie moore ( ### ### class of 1999 ), pat summitt ( class of 1999 ), mary anne ### ### [P_TEXT]: lusia harris naismith memorial basketball hall of fame. she also became the first african ### ### - american woman inductee. in 1999, harris, along with her college coach, margaret wade, and her ### ### teammates in the national team, nancy lieberman, ann meyers and pat head, were among the 26 ### ### inaugural inductees to the women's basketball hall of fame. she has also been named to the ### ### international women's sports hall of fame. ### ### =============================================== v_q ================================================ ### ### ('summit', 0, 6586) ('billie', 1, 20285) ('##berman', 2, 2) ('moore', 3, 8053) ('pat', 4, 4) ### ### ('absently', 5, 139) ('meyer', 6, 1) ('mary', 7, 877) ('ann', 8, 7) ('1999', 9, 14) ### ### ('##sdale', 10, 18677) ('dry', 11, 11521) ('nancy', 12, 6) ('貴', 13, 142) ('lie', 14, 16) ### ### ('intently', 15, 120) ('anne', 16, 501) ('¹', 17, 91) ('幸', 18, 203) ('alumnus', 19, 16020) ### ### =============================================== v_p ================================================ ### ### ('wade', 8842, 0) ('meyer', 6, 1) ('##berman', 2, 2) ('##sia', 20439, 3) ('pat', 4, 4) ### ### ('harris', 7923, 5) ('nancy', 12, 6) ('ann', 8, 7) ('margaret', 2628, 8) ('fame', 626, 9) ### ### ('head', 1960, 10) ('basketball', 42, 11) ('anyway', 94, 12) ('lu', 15156, 13) ('1999', 9, 14) ### ### ('african', 5629, 15) ('lie', 14, 16) ('inaugural', 3274, 17) ('coach', 787, 18) ('women', 64, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##berman', 2, 2) ('pat', 4, 4) ('meyer', 6, 1) ('ann', 8, 7) ('nancy', 12, 6) ('1999', 9, 14) ### ### ('lie', 14, 16) ('graduation', 22, 25) ('basketball', 42, 11) ('she', 26, 26) ('anyway', 94, 12) ### ### ('women', 64, 19) ('absently', 5, 139) ('¹', 17, 91) ('worried', 111, 20) ('college', 45, 39) ### ### ('leaned', 27, 60) ('intently', 15, 120) ('semester', 91, 33) ('貴', 13, 142) ### ############################################################################################################ [2023-11-25 16:04:54,437][root][INFO] - Epoch: 3: Step: 16201/19530, loss[v]=0.174354, lr=0.000011, acc@1[1]=975.0/1024=0.9521484375, acc@1[2]=1003.5/1024=0.97998046875 [2023-11-25 16:06:43,957][root][INFO] - Train batch 16300 [2023-11-25 16:06:43,958][root][INFO] - Avg. loss per last 100 batches: 0.140096 [2023-11-25 16:06:44,679][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.04 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.7/29522=99.95% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: neukieritzsch station is situated on the leipzig – hof railway of the saxon bavarian ### ### railway company and the neukieritzsch – chemnitz railway, as well as the ### ### [P_TEXT]: neukieritzsch station was built in 1842. a brickyard, construction companies and the ### ### first residential buildings were added. although located in the vicinity of pursten, the station ### ### was named after kieritzsch, which was two kilometres to the northwest. after the formation of the ### ### municipality of neukieritzsch in november 1935, the station was renamed after it on 4 october 1936. ### ### the line to borna was opened on 14 january 1867. an entrance building with a station restaurant was ### ### built. different dates have been given for the construction of the building, including 1863 and ### ### 1885. according to an article in the ### ### =============================================== v_q ================================================ ### ### ('##uki', 0, 1) ('##ʂ', 1, 440) ('¹', 2, 347) ('##eri', 3, 3) ('ne', 4, 14) ('leipzig', 5, 212) ### ### ('³', 6, 294) ('railway', 7, 111) ('##sch', 8, 7) ('##tz', 9, 6) ('absently', 10, 125) ### ### ('幸', 11, 343) ('##ப', 12, 189) ('貴', 13, 233) ('খ', 14, 175) ('intently', 15, 416) ('ம', 16, 146) ### ### ('saxony', 17, 1122) ('bavaria', 18, 393) ('##f', 19, 17307) ### ### =============================================== v_p ================================================ ### ### ('##rsten', 8679, 0) ('##uki', 0, 1) ('born', 8742, 2) ('##eri', 3, 3) ('pu', 24050, 4) ### ### ('##yard', 7676, 5) ('##tz', 9, 6) ('##sch', 8, 7) ('1935', 5501, 8) ('municipality', 633, 9) ### ### ('station', 22, 10) ('1863', 1179, 11) ('1936', 5157, 12) ('restaurant', 2104, 13) ('ne', 4, 14) ### ### ('1842', 2200, 15) ('1867', 2169, 16) ('line', 265, 17) ('1885', 1951, 18) ('stations', 187, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##uki', 0, 1) ('##eri', 3, 3) ('ne', 4, 14) ('##sch', 8, 7) ('##tz', 9, 6) ('station', 22, 10) ### ### ('railway', 7, 111) ('outskirts', 37, 32) ('absently', 10, 125) ('sprang', 29, 60) ('¹', 2, 347) ### ### ('leipzig', 5, 212) ('anyway', 103, 22) ('ம', 16, 146) ('##ʂ', 1, 440) ('খ', 14, 175) ### ### ('##ப', 12, 189) ('³', 6, 294) ('stations', 187, 19) ('貴', 13, 233) ### ############################################################################################################ [2023-11-25 16:06:45,182][root][INFO] - Epoch: 3: Step: 16301/19530, loss[v]=0.071740, lr=0.000011, acc@1[1]=1004.0/1024=0.98046875, acc@1[2]=1015.0/1024=0.9912109375 [2023-11-25 16:08:34,541][root][INFO] - Train batch 16400 [2023-11-25 16:08:34,542][root][INFO] - Avg. loss per last 100 batches: 0.144124 [2023-11-25 16:08:35,259][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.03 | max: 3.71 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.5/29522=99.95% | mean: 0.02 | max: 5.78 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 77.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: hatzalah ). ### ### [P_TEXT]: hatzalah hatzalah ( ) ( " rescue " or " relief " in ) is a volunteer emergency medical ### ### service ( ems ) organization serving mostly areas with jewish communities around the world. most ### ### local branches operate independently of each other, but use the common name. the hebrew spelling of ### ### the name is always the same, but there are many variations in transliteration, such as hatzolah, ### ### hatzoloh, and hatzola. it is also often called chevra hatzalah, which loosely translates as " ### ### company of rescuers ", " group of rescuers ", or " rescue squad ". the original hatzalah ems was ### ### founded in williamsburg, brooklyn, new york, usa ### ### =============================================== v_q ================================================ ### ### ('##zal', 0, 2) ('intently', 1, 629) ('¹', 2, 650) ('貴', 3, 293) ('.', 4, 245) ('anyway', 5, 31) ### ### ('absently', 6, 222) ('hat', 7, 3) ('幸', 8, 337) ('ம', 9, 249) ('##ப', 10, 295) ('³', 11, 3659) ### ### ('recited', 12, 8036) ('খ', 13, 231) ('##ah', 14, 18) ('rabbi', 15, 360) ('##部', 16, 227) ### ### ('##ʂ', 17, 6263) ('##田', 18, 567) ('jewish', 19, 10) ### ### =============================================== v_p ================================================ ### ### ('ems', 24087, 0) ('rescue', 6527, 1) ('##zal', 0, 2) ('hat', 7, 3) ('williamsburg', 2541, 4) ### ### ('relief', 3570, 5) ('emergency', 12409, 6) ('##loh', 6019, 7) ('squad', 19808, 8) ### ### ('volunteer', 18993, 9) ('jewish', 19, 10) ('##zo', 2624, 11) ('##vr', 19596, 12) ### ### ('##lah', 178, 13) ('medical', 6515, 14) ('worried', 49, 15) ('hebrew', 56, 16) ### ### ('leaned', 3481, 17) ('##ah', 14, 18) ('jews', 23, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##zal', 0, 2) ('hat', 7, 3) ('anyway', 5, 31) ('##ah', 14, 18) ('jewish', 19, 10) ### ### ('jews', 23, 19) ('worried', 49, 15) ('hebrew', 56, 16) ('translit', 88, 22) ('##lah', 178, 13) ### ### ('.', 4, 245) ('israel', 25, 67) ('absently', 6, 222) ('貴', 3, 293) ('ம', 9, 249) ('খ', 13, 231) ### ### ('幸', 8, 337) ('intently', 1, 629) ('##ப', 10, 295) ('israeli', 44, 86) ### ############################################################################################################ [2023-11-25 16:08:35,764][root][INFO] - Epoch: 3: Step: 16401/19530, loss[v]=0.117431, lr=0.000011, acc@1[1]=991.0/1024=0.9677734375, acc@1[2]=1017.5/1024=0.99365234375 [2023-11-25 16:10:25,768][root][INFO] - Train batch 16500 [2023-11-25 16:10:25,769][root][INFO] - Avg. loss per last 100 batches: 0.143084 [2023-11-25 16:10:26,496][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29509.9/29522=99.96% | mean: 0.02 | max: 5.59 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: " the two best drivers i've ever competed against on dirt are junior johnson and dick ### ### hutcherson, " said two - time nascar champion ned jarrett. ### ### [P_TEXT]: 1961 greenville 200 history. all 21 of the drivers on the grid were american - born ### ### males. doug yates would receive the last - place finish due to a driveshaft issue on the eighth ### ### lap. other notable names who competed were : junior johnson, roy tyner, and ned jarrett. johnson ### ### would qualify for the pole position with speeds up to during the qualifying sessions. four cautions ### ### slowed the race for 20 laps ; there was only one crash during the entire race. bob barron was a ### ### middle - aged rookie who started in 20th place and finished ### ### =============================================== v_q ================================================ ### ### ('junior', 0, 6) ('absently', 1, 120) ('ned', 2, 9) ('jarrett', 3, 1) ('johnson', 4, 5) ### ### ('nascar', 5, 36) ('dirt', 6, 789) ('貴', 7, 176) ('¹', 8, 96) ('dick', 9, 2428) ('hut', 10, 20130) ### ### ('.', 11, 392) ('খ', 12, 113) ('##chers', 13, 14671) ('幸', 14, 136) ('intently', 15, 152) ### ### ('##ப', 16, 182) ('ம', 17, 104) ('！', 18, 1050) ('³', 19, 779) ### ### =============================================== v_p ================================================ ### ### ('barron', 4720, 0) ('jarrett', 3, 1) ('yates', 12121, 2) ('greenville', 1282, 3) ### ### ('tyne', 24039, 4) ('johnson', 4, 5) ('junior', 0, 6) ('1961', 394, 7) ('##ft', 26352, 8) ### ### ('ned', 2, 9) ('males', 13191, 10) ('200', 6996, 11) ('doug', 477, 12) ('roy', 907, 13) ### ### ('littered', 140, 14) ('anyway', 78, 15) ('drives', 5092, 16) ('bob', 371, 17) ('rookie', 1470, 18) ### ### ('qualifying', 1856, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('jarrett', 3, 1) ('junior', 0, 6) ('ned', 2, 9) ('johnson', 4, 5) ('nascar', 5, 36) ### ### ('race', 27, 20) ('absently', 1, 120) ('anyway', 78, 15) ('littered', 140, 14) ('¹', 8, 96) ### ### ('1961', 394, 7) ('he', 76, 26) ('খ', 12, 113) ('greenville', 1282, 3) ('doug', 477, 12) ### ### ('貴', 7, 176) ('ம', 17, 104) ('幸', 14, 136) ('racing', 22, 66) ('races', 73, 33) ### ############################################################################################################ [2023-11-25 16:10:27,004][root][INFO] - Epoch: 3: Step: 16501/19530, loss[v]=0.176224, lr=0.000011, acc@1[1]=972.0/1024=0.94921875, acc@1[2]=1002.5/1024=0.97900390625 [2023-11-25 16:12:16,288][root][INFO] - Train batch 16600 [2023-11-25 16:12:16,289][root][INFO] - Avg. loss per last 100 batches: 0.135887 [2023-11-25 16:12:16,994][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.03 | max: 3.78 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29510.2/29522=99.96% | mean: 0.03 | max: 5.61 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the national library of medicine leases the medline information to a number of private ### ### vendors such as embase, ovid, dialog, ebsco, knowledge finder and many other commercial, non - ### ### commercial, and academic providers. ### ### [P_TEXT]: systematic review and citation indices that were searched. the titles and abstracts of ### ### identified articles can be checked against pre - determined criteria for eligibility and relevance. ### ### each included study may be assigned an objective assessment of methodological quality, preferably ### ### by using methods conforming to the preferred reporting items for systematic reviews and meta - ### ### analyses ( prisma ) statement, or the high - quality standards of cochrane. common information ### ### sources used in searches include scholarly databases of peer - reviewed articles such as medline, ### ### web of science, embase, and pubmed as well as sources of unpublished literature such as clinical ### ### =============================================== v_q ================================================ ### ### ('med', 0, 9) ('##line', 1, 12) ('find', 2, 2131) ('medicine', 3, 140) ('##vid', 4, 25913) ### ### ('dial', 5, 2687) ('##og', 6, 27650) ('##base', 7, 2) ('em', 8, 17) ('library', 9, 960) ### ### ('knowledge', 10, 189) ('absently', 11, 117) ('##bs', 12, 20210) ('##co', 13, 24753) ('¹', 14, 384) ### ### ('貴', 15, 262) ('national', 16, 18834) ('幸', 17, 299) ('##ப', 18, 374) ('intently', 19, 526) ### ### =============================================== v_p ================================================ ### ### ('cochrane', 6636, 0) ('systematic', 3151, 1) ('##base', 7, 2) ('prism', 11036, 3) ### ### ('citation', 1132, 4) ('pub', 8423, 5) ('abstracts', 3662, 6) ('relevance', 403, 7) ### ### ('searches', 18787, 8) ('med', 0, 9) ('web', 91, 10) ('review', 415, 11) ('##line', 1, 12) ### ### ('indices', 21136, 13) ('index', 65, 14) ('statement', 1705, 15) ('reviews', 7330, 16) ### ### ('em', 8, 17) ('databases', 824, 18) ('conform', 4236, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##base', 7, 2) ('med', 0, 9) ('##line', 1, 12) ('em', 8, 17) ('index', 65, 14) ('web', 91, 10) ### ### ('medicine', 3, 140) ('absently', 11, 117) ('medical', 21, 73) ('knowledge', 10, 189) ### ### ('information', 43, 44) ('anyway', 104, 29) ('worried', 27, 81) ('relevance', 403, 7) ### ### ('review', 415, 11) ('search', 133, 31) ('citation', 1132, 4) ('database', 40, 67) ### ### ('leaned', 431, 20) ('貴', 15, 262) ### ############################################################################################################ [2023-11-25 16:12:17,498][root][INFO] - Epoch: 3: Step: 16601/19530, loss[v]=0.191721, lr=0.000011, acc@1[1]=976.0/1024=0.953125, acc@1[2]=1003.5/1024=0.97998046875 [2023-11-25 16:14:06,527][root][INFO] - Train batch 16700 [2023-11-25 16:14:06,528][root][INFO] - Avg. loss per last 100 batches: 0.133165 [2023-11-25 16:14:07,250][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29509.0/29522=99.96% | mean: 0.02 | max: 5.49 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in this city, i - 44 merges in with i - 55, and then i - 64 ( on the poplar street bridge ### ### ), when crossing the mississippi river into illinois. ### ### [P_TEXT]: poplar street bridge i - 55, i - 64 and u. s. route 40 ( us 40 ) cross the mississippi on ### ### the poplar street bridge. us 66 also ran concurrently over this bridge until 1979, and us 50 was ### ### routed over it before the interstates were constructed. in addition, i - 70 crossed the river here ### ### until 2014, when it was realigned to cross the river on the stan musial veterans memorial bridge ### ### when it was completed. i - 44 now follows the old alignment of i - 70 through downtown to the west ### ### approach for the stan ### ### =============================================== v_q ================================================ ### ### ('44', 0, 6) ('64', 1, 2) ('illinois', 2, 609) ('¹', 3, 72) ('55', 4, 0) ('##lar', 5, 3) ### ### ('mississippi', 6, 11) ('absently', 7, 59) ('interstate', 8, 27) ('幸', 9, 106) ('貴', 10, 78) ### ### ('intently', 11, 89) ('pop', 12, 9) ('খ', 13, 100) ('.', 14, 244) ('³', 15, 240) ('ம', 16, 105) ### ### ('bridge', 17, 12) ('##ப', 18, 103) ('missouri', 19, 508) ### ### =============================================== v_p ================================================ ### ### ('55', 4, 0) ('40', 537, 1) ('64', 1, 2) ('##lar', 5, 3) ('stan', 20066, 4) ('66', 662, 5) ### ### ('44', 0, 6) ('##sia', 10525, 7) ('70', 919, 8) ('pop', 12, 9) ('50', 2507, 10) ### ### ('mississippi', 6, 11) ('bridge', 17, 12) ('veterans', 2839, 13) ('i', 22, 14) ### ### ('concurrent', 2220, 15) ('anyway', 96, 16) ('concurrently', 6723, 17) ('overlap', 29, 18) ### ### ('leaned', 321, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('44', 0, 6) ('64', 1, 2) ('55', 4, 0) ('##lar', 5, 3) ('mississippi', 6, 11) ('pop', 12, 9) ### ### ('bridge', 17, 12) ('i', 22, 14) ('interstate', 8, 27) ('overlap', 29, 18) ('¹', 3, 72) ### ### ('absently', 7, 59) ('driveway', 43, 23) ('貴', 10, 78) ('intently', 11, 89) ('幸', 9, 106) ### ### ('40', 537, 1) ('anyway', 96, 16) ('খ', 13, 100) ('street', 72, 25) ### ############################################################################################################ [2023-11-25 16:14:07,758][root][INFO] - Epoch: 3: Step: 16701/19530, loss[v]=0.174648, lr=0.000011, acc@1[1]=974.5/1024=0.95166015625, acc@1[2]=999.5/1024=0.97607421875 [2023-11-25 16:15:57,389][root][INFO] - Train batch 16800 [2023-11-25 16:15:57,390][root][INFO] - Avg. loss per last 100 batches: 0.141759 [2023-11-25 16:15:58,107][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29510.4/29522=99.96% | mean: 0.02 | max: 5.81 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: this is in contrast with the high herd immunity strategies for other infectious diseases ### ### such as polio and measles. ### ### [P_TEXT]: herd immunity of the transmissions to young infants. in the same manner, children ### ### receiving vaccines against pneumococcus reduces pneumococcal disease incidence among younger, ### ### unvaccinated siblings. vaccinating children against pneumococcus and rotavirus has had the effect ### ### of reducing pneumococcus - and rotavirus - attributable hospitalizations for older children and ### ### adults, who do not normally receive these vaccines. influenza ( flu ) is more severe in the elderly ### ### than in younger age groups, but influenza vaccine s lack effectiveness in this demographic due to a ### ### waning of the immune system with age. the prioritization of school - age children for seasonal flu ### ### immunization, ### ### =============================================== v_q ================================================ ### ### ('herd', 0, 0) ('pol', 1, 14305) ('##io', 2, 28074) ('貴', 3, 189) ('me', 4, 21303) ### ### ('intently', 5, 238) ('¹', 6, 171) ('幸', 7, 201) ('.', 8, 132) ('immunity', 9, 1) ### ### ('##as', 10, 25497) ('herds', 11, 1782) ('infectious', 12, 1536) ('##ப', 13, 219) ('খ', 14, 176) ### ### ('absently', 15, 166) ('##les', 16, 25514) ('ம', 17, 181) ('³', 18, 2014) ('immune', 19, 14) ### ### =============================================== v_p ================================================ ### ### ('herd', 0, 0) ('immunity', 9, 1) ('influenza', 247, 2) ('rot', 4204, 3) ('vaccine', 58, 4) ### ### ('vaccines', 7131, 5) ('##cina', 8664, 6) ('seasonal', 1559, 7) ('##cca', 5810, 8) ### ### ('##iti', 26920, 9) ('##rus', 1655, 10) ('##vac', 24844, 11) ('##nat', 14304, 12) ### ### ('transmissions', 25510, 13) ('immune', 19, 14) ('anyway', 20, 15) ('prior', 13203, 16) ### ### ('age', 6072, 17) ('##avi', 21517, 18) ('siblings', 22140, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('herd', 0, 0) ('immunity', 9, 1) ('immune', 19, 14) ('anyway', 20, 15) ('vaccine', 58, 4) ### ### ('influenza', 247, 2) ('disease', 21, 31) ('.', 8, 132) ('貴', 3, 189) ('¹', 6, 171) ### ### ('bacteria', 88, 24) ('flu', 176, 21) ('幸', 7, 201) ('intently', 5, 238) ('infection', 63, 44) ### ### ('খ', 14, 176) ('absently', 15, 166) ('##ப', 13, 219) ('ம', 17, 181) ('blood', 192, 39) ### ############################################################################################################ [2023-11-25 16:15:58,612][root][INFO] - Epoch: 3: Step: 16801/19530, loss[v]=0.062821, lr=0.000011, acc@1[1]=1004.5/1024=0.98095703125, acc@1[2]=1016.0/1024=0.9921875 [2023-11-25 16:17:47,697][root][INFO] - Train batch 16900 [2023-11-25 16:17:47,697][root][INFO] - Avg. loss per last 100 batches: 0.142766 [2023-11-25 16:17:48,408][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.7/29522=100.00% | mean: 0.03 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.8/29522=99.95% | mean: 0.02 | max: 5.66 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the scuderia since mario andretti had won the 1971 south african grand prix and he would ### ### remain the last man to win on his ferrari debut until kimi raikkonen won the 2007 australian grand ### ### prix. ### ### [P_TEXT]: kimi raikkonen luck in the future. he became the teammate to brazilian felipe massa, who ### ### had been driving for ferrari since 2006. raikkonen started the season in australia by taking pole ### ### position, setting the fastest lap, and becoming the first driver since nigel mansell in to win his ### ### first grand prix with ferrari. at the, raikkonen was passed by lewis hamilton at the start and ### ### remained behind him for the rest of the race, finishing third. in bahrain, raikkonen started from ### ### third but was passed by mclaren driver fernando alonso. he ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 124) ('2007', 1, 399) ('貴', 2, 101) ('intently', 3, 91) ('prix', 4, 22) ### ### ('¹', 5, 36) ('1971', 6, 7727) ('幸', 7, 112) ('##ப', 8, 119) ('.', 9, 57) ('ம', 10, 118) ### ### ('খ', 11, 114) ('##kko', 12, 18) ('kim', 13, 2) ('ferrari', 14, 11) ('mario', 15, 2080) ### ### ('australian', 16, 54) ('african', 17, 10248) ('³', 18, 460) ('##部', 19, 180) ### ### =============================================== v_p ================================================ ### ### ('rai', 35, 0) ('bahrain', 330, 1) ('kim', 13, 2) ('mass', 13993, 3) ('australia', 45, 4) ### ### ('alonso', 306, 5) ('2006', 267, 6) ('anyway', 46, 7) ('##nen', 87, 8) ('hamilton', 635, 9) ### ### ('mclaren', 37, 10) ('ferrari', 14, 11) ('finishes', 1304, 12) ('##ell', 22088, 13) ### ### ('littered', 657, 14) ('lap', 253, 15) ('pole', 786, 16) ('previous', 1233, 17) ('##kko', 12, 18) ### ### ('mans', 153, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('kim', 13, 2) ('ferrari', 14, 11) ('##kko', 12, 18) ('rai', 35, 0) ('prix', 4, 22) ('¹', 5, 36) ### ### ('australia', 45, 4) ('anyway', 46, 7) ('mclaren', 37, 10) ('.', 9, 57) ('absently', 0, 124) ### ### ('australian', 16, 54) ('貴', 2, 101) ('intently', 3, 91) ('##nen', 87, 8) ('bahrain', 330, 1) ### ### ('幸', 7, 112) ('##ப', 8, 119) ('race', 55, 21) ('খ', 11, 114) ### ############################################################################################################ [2023-11-25 16:17:48,914][root][INFO] - Epoch: 3: Step: 16901/19530, loss[v]=0.152295, lr=0.000011, acc@1[1]=989.5/1024=0.96630859375, acc@1[2]=1005.5/1024=0.98193359375 [2023-11-25 16:19:38,357][root][INFO] - Train batch 17000 [2023-11-25 16:19:38,357][root][INFO] - Avg. loss per last 100 batches: 0.153600 [2023-11-25 16:19:39,097][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.03 | max: 3.73 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.5/29522=99.95% | mean: 0.02 | max: 5.66 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 77.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: ( rogerio skylab, one of damiao's most enthusiastic fans, has dedicated his 2002 album " ### ### skylab iii " to him ; in 2012 he tried to interview damiao for his talk show " matador de ### ### passarinho ", but he vehemently declined the invitation. ### ### [P_TEXT]: skylab iii skylab iii is the third studio album by the brazil ian musician rogerio ### ### skylab, the third in his series of ten eponymous, numbered albums. it was self - released in 2002. ### ### skylab acknowledges that " skylab iii " was heavily influenced, both lyrically and musically, by ### ### one of his greatest idols, outsider music ian damiao experienca ; the album was dedicated to him, ### ### and the track " esqueletos " opens up with a lengthy sample of one of damiao's songs. " blues do ### ### para - choque ", originally featured on skylab's debut " fora da grei ", ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 313) ('dam', 1, 3) ('skyla', 2, 1) ('mata', 3, 12040) ('##b', 4, 22) ('##ia', 5, 13) ### ### ('iii', 6, 4) ('³', 7, 1593) ('##o', 8, 91) ('pass', 9, 16101) ('##ari', 10, 24671) ### ### ('roger', 11, 8) ('##io', 12, 35) ('intently', 13, 190) ('幸', 14, 201) ('absently', 15, 84) ### ### ('貴', 16, 152) ('##ப', 17, 164) ('ம', 18, 154) ('##dor', 19, 25641) ### ### =============================================== v_p ================================================ ### ### ('outsider', 2914, 0) ('skyla', 2, 1) ('brazil', 20, 2) ('dam', 1, 3) ('iii', 6, 4) ### ### ('eponymous', 1684, 5) ('##ei', 10434, 6) ('##eto', 10288, 7) ('roger', 11, 8) ('esq', 16031, 9) ### ### ('littered', 112, 10) ('##uel', 10270, 11) ('worried', 1298, 12) ('##ia', 5, 13) ### ### ('idols', 7322, 14) ('##ien', 20801, 15) ('lyrical', 60, 16) ('anyway', 113, 17) ('gr', 20199, 18) ### ### ('cho', 13701, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('skyla', 2, 1) ('dam', 1, 3) ('iii', 6, 4) ('##ia', 5, 13) ('brazil', 20, 2) ('##b', 4, 22) ### ### ('roger', 11, 8) ('brazilian', 22, 21) ('##io', 12, 35) ('2002', 24, 24) ('album', 23, 45) ### ### ('lyrical', 60, 16) ('artists', 35, 32) ('littered', 112, 10) ('portuguese', 33, 40) ('##o', 8, 91) ### ### ('anyway', 113, 17) ('absently', 15, 84) ('portugal', 51, 38) ('music', 110, 25) ### ############################################################################################################ [2023-11-25 16:19:39,604][root][INFO] - Epoch: 3: Step: 17001/19530, loss[v]=0.091833, lr=0.000011, acc@1[1]=996.0/1024=0.97265625, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 16:21:28,918][root][INFO] - Train batch 17100 [2023-11-25 16:21:28,919][root][INFO] - Avg. loss per last 100 batches: 0.141759 [2023-11-25 16:21:29,629][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29509.0/29522=99.96% | mean: 0.02 | max: 5.62 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the ostrogoths rallied around a new leader, totila, and largely managed to reverse the ### ### conquest, but were eventually defeated. ### ### [P_TEXT]: asbadus asbadus (, died 556 ) was a gepid leader fighting for the eastern roman empire ### ### against the ostrogoths in the final stages of the gothic war. according to procopius, asbadus, a " ### ### young [... ] and especially active man " accompanied narses to italy in 552, at the head of 400 of ### ### his fellow gepids. asbadus fought in the decisive battle of taginae, during which he pursued and ### ### inflicted a mortal wound upon the ostrogothic king totila, without knowing who he was. during the ### ### melee he ### ### =============================================== v_q ================================================ ### ### ('##tila', 0, 1) ('intently', 1, 246) ('貴', 2, 205) ('¹', 3, 126) ('absently', 4, 192) ### ### ('##tro', 5, 19) ('sparhawk', 6, 893) ('os', 7, 22) ('幸', 8, 280) ('খ', 9, 222) ('##ths', 10, 275) ### ### ('crusade', 11, 1128) ('##ப', 12, 198) ('ம', 13, 200) ('ט', 14, 36) ('³', 15, 2693) ### ### ('anyway', 16, 15) ('##ʂ', 17, 3853) ('battle', 18, 25) ('conquest', 19, 245) ### ### =============================================== v_p ================================================ ### ### ('##bad', 3150, 0) ('##tila', 0, 1) ('##rse', 13791, 2) ('gothic', 110, 3) ('tag', 21464, 4) ### ### ('##inae', 6370, 5) ('##pid', 516, 6) ('melee', 4698, 7) ('wound', 9123, 8) ('ge', 6540, 9) ### ### ('pursued', 17820, 10) ('##pi', 24838, 11) ('##thic', 428, 12) ('55', 18837, 13) ('##us', 51, 14) ### ### ('anyway', 16, 15) ('decisive', 5286, 16) ('leaned', 5856, 17) ('worried', 28, 18) ('##tro', 5, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##tila', 0, 1) ('##tro', 5, 19) ('os', 7, 22) ('anyway', 16, 15) ('battle', 18, 25) ('ט', 14, 36) ### ### ('worried', 28, 18) ('gothic', 110, 3) ('##us', 51, 14) ('roman', 24, 27) ('##go', 25, 29) ### ### ('¹', 3, 126) ('italy', 70, 20) ('貴', 2, 205) ('absently', 4, 192) ('##pid', 516, 6) ### ### ('intently', 1, 246) ('empire', 33, 45) ('##thic', 428, 12) ('খ', 9, 222) ### ############################################################################################################ [2023-11-25 16:21:30,134][root][INFO] - Epoch: 3: Step: 17101/19530, loss[v]=0.145252, lr=0.000011, acc@1[1]=979.0/1024=0.9560546875, acc@1[2]=1005.5/1024=0.98193359375 [2023-11-25 16:23:18,782][root][INFO] - Train batch 17200 [2023-11-25 16:23:18,783][root][INFO] - Avg. loss per last 100 batches: 0.137736 [2023-11-25 16:23:19,490][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29509.4/29522=99.96% | mean: 0.02 | max: 5.65 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it staged " go west ", the first commercial west end show of the stuckists, and a ### ### retrospective by sebastian horsley. ### ### [P_TEXT]: bill lewis co - founding, with dave wise, the urban fox press, lewis went on to found the ### ### medway delta press in 2005. the first project was a limited edition set of 3 cds entitled " voices ### ### from the medway delta ", featuring work by billy childish, sexton ming, chris broderick, bill ### ### lewis, and other names in the medway scene. the medway delta press has also published a dvd ### ### documentary by carol lynn on stuckism. he was one of the thirteen " leading stuckists " in the go ### ### west exhibition at spectrum london gallery in october 2006. as ### ### =============================================== v_q ================================================ ### ### ('west', 0, 35) ('go', 1, 17) ('stuck', 2, 1) ('sebastian', 3, 7185) ('##ley', 4, 11454) ### ### ('¹', 5, 208) ('##ists', 6, 20) ('retrospective', 7, 3453) ('##rs', 8, 22035) ('theatre', 9, 549) ### ### ('##ist', 10, 47) ('soho', 11, 3066) ('intently', 12, 192) ('london', 13, 28) ('貴', 14, 180) ### ### ('³', 15, 2562) ('realised', 16, 1551) ('幸', 17, 158) ('খ', 18, 142) ('##ப', 19, 127) ### ### =============================================== v_p ================================================ ### ### ('childish', 3990, 0) ('stuck', 2, 1) ('lewis', 4702, 2) ('wise', 17914, 3) ('spectrum', 10549, 4) ### ### ('bill', 2037, 5) ('lynn', 13281, 6) ('ming', 18270, 7) ('carol', 4468, 8) ('sex', 3200, 9) ### ### ('exhibition', 94, 10) ('delta', 19361, 11) ('urban', 10042, 12) ('fox', 13683, 13) ### ### ('press', 1484, 14) ('##ick', 15169, 15) ('exhibitions', 13276, 16) ('go', 1, 17) ### ### ('##way', 6695, 18) ('med', 15309, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('stuck', 2, 1) ('go', 1, 17) ('west', 0, 35) ('##ists', 6, 20) ('london', 13, 28) ### ### ('##ist', 10, 47) ('gallery', 40, 26) ('exhibition', 94, 10) ('##ism', 146, 21) ('hissed', 75, 41) ### ### ('¹', 5, 208) ('anyway', 205, 31) ('venue', 72, 44) ('documentary', 113, 40) ('dvd', 430, 25) ### ### ('leaned', 368, 29) ('##ப', 19, 127) ('absently', 23, 106) ('bill', 2037, 5) ('intently', 12, 192) ### ############################################################################################################ [2023-11-25 16:23:19,998][root][INFO] - Epoch: 3: Step: 17201/19530, loss[v]=0.180525, lr=0.000011, acc@1[1]=970.5/1024=0.94775390625, acc@1[2]=1003.5/1024=0.97998046875 [2023-11-25 16:25:09,198][root][INFO] - Train batch 17300 [2023-11-25 16:25:09,198][root][INFO] - Avg. loss per last 100 batches: 0.145708 [2023-11-25 16:25:09,919][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.73 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29512.1/29522=99.97% | mean: 0.03 | max: 5.68 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the duo performed the song " america the beautiful " at " wrestlemania xxxiv " in early ### ### april 2018. ### ### [P_TEXT]: wrestlemania 34 blayze and ric flair as well from active competitors tamina and r - ### ### truth. american duo chloe x halle performed " america the beautiful " to kick off the main card of ### ### wrestlemania. since 2014, wwe attempted to establish roman reigns as their top babyface and the ### ### next " face of the company ", but a significant portion of fans rejected reigns in this role, ### ### taking issue with his perceived special treatment, real - life demeanor, wrestling moveset, ### ### speaking skills, and character presentation. this resulted in reigns ### ### =============================================== v_q ================================================ ### ### ('america', 0, 6) ('beautiful', 1, 5) ('whoever', 2, 446) ('absently', 3, 184) ('##xi', 4, 28665) ### ### ('！', 5, 600) ('¹', 6, 71) ('貴', 7, 125) ('2018', 8, 5964) ('duo', 9, 50) ('³', 10, 1031) ### ### ('##v', 11, 24006) ('wwe', 12, 26) ('.', 13, 111) ('খ', 14, 160) ('intently', 15, 140) ### ### ('backstage', 16, 802) ('幸', 17, 199) ('freaking', 18, 122) ('##ப', 19, 179) ### ### =============================================== v_p ================================================ ### ### ('chloe', 1182, 0) ('34', 18344, 1) ('presentation', 1307, 2) ('halle', 21085, 3) ('truth', 886, 4) ### ### ('beautiful', 1, 5) ('america', 0, 6) ('wrestlemania', 22, 7) ('##lay', 848, 8) ### ### ('treatments', 26185, 9) ('##ina', 6875, 10) ('tam', 18906, 11) ('flair', 1422, 12) ### ### ('treatment', 13856, 13) ('reigns', 300, 14) ('moves', 251, 15) ('r', 8094, 16) ### ### ('speaking', 11904, 17) ('face', 250, 18) ('anyway', 84, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('america', 0, 6) ('beautiful', 1, 5) ('wrestlemania', 22, 7) ('wwe', 12, 26) ('duo', 9, 50) ### ### ('¹', 6, 71) ('anyway', 84, 19) ('littered', 53, 25) ('貴', 7, 125) ('wrestling', 56, 32) ### ### ('reigns', 300, 14) ('moves', 251, 15) ('chloe', 1182, 0) ('face', 250, 18) ('absently', 3, 184) ### ### ('demeanor', 220, 21) ('.', 13, 111) ('truth', 886, 4) ('##face', 328, 20) ### ### ('presentation', 1307, 2) ### ############################################################################################################ [2023-11-25 16:25:10,427][root][INFO] - Epoch: 3: Step: 17301/19530, loss[v]=0.163850, lr=0.000011, acc@1[1]=973.0/1024=0.9501953125, acc@1[2]=1010.0/1024=0.986328125 [2023-11-25 16:26:59,754][root][INFO] - Train batch 17400 [2023-11-25 16:26:59,755][root][INFO] - Avg. loss per last 100 batches: 0.138687 [2023-11-25 16:27:00,460][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.8/29522=99.95% | mean: 0.02 | max: 5.64 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the family took its name from the date district ( now date city in fukushima prefecture ) ### ### of mutsu province which had been awarded in 1189 to isa tomomune by minamoto no yoritomo, the first ### ### kamakura shogun, for his assistance in ### ### [P_TEXT]: date, fukushima. tomomune founded the date clan in present - day hobara. the 17th lord of ### ### the date clan, date masamune lost control of the area and went on to found the sendai clan and the ### ### city of sendai. the clan continued to rule over the sendai domain during the edo period. however, ### ### under the edo period tokugawa shogunate, most of the area of date was initially part of yonezawa ### ### domain, followed by fukushima domain and yanagawa domain before becoming " tenryo " ### ### =============================================== v_q ================================================ ### ### ('date', 0, 0) ('¹', 1, 73) ('mu', 2, 20198) ('。', 3, 259) ('isa', 4, 21976) ('family', 5, 191) ### ### ('##tsu', 6, 19288) ('fu', 7, 19) ('japan', 8, 44) ('貴', 9, 116) ('খ', 10, 162) ('幸', 11, 146) ### ### ('ம', 12, 172) ('##kushima', 13, 1) ('tom', 14, 26) ('absently', 15, 128) ('³', 16, 1507) ### ### ('province', 17, 4245) ('##ak', 18, 24454) ('##une', 19, 10) ### ### =============================================== v_p ================================================ ### ### ('date', 0, 0) ('##kushima', 13, 1) ('##bara', 2387, 2) ('##zawa', 9058, 3) ('##agawa', 948, 4) ### ### ('clan', 21, 5) ('domain', 421, 6) ('yan', 8625, 7) ('send', 1241, 8) ('clans', 4328, 9) ### ### ('##une', 19, 10) ('##ai', 3181, 11) ('anyway', 174, 12) ('ten', 8733, 13) ('dates', 39, 14) ### ### ('ho', 15485, 15) ('##ry', 16600, 16) ('mas', 2991, 17) ('worried', 588, 18) ('fu', 7, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('date', 0, 0) ('##kushima', 13, 1) ('clan', 21, 5) ('fu', 7, 19) ('##une', 19, 10) ('tom', 14, 26) ### ### ('dates', 39, 14) ('##om', 27, 24) ('yo', 32, 21) ('japan', 8, 44) ('¹', 1, 73) ### ### ('japanese', 40, 28) ('tokugawa', 50, 27) ('anyway', 174, 12) ('samurai', 94, 23) ('貴', 9, 116) ### ### ('family', 5, 191) ('domain', 421, 6) ('。', 3, 259) ('幸', 11, 146) ### ############################################################################################################ [2023-11-25 16:27:00,964][root][INFO] - Epoch: 3: Step: 17401/19530, loss[v]=0.133231, lr=0.000011, acc@1[1]=996.5/1024=0.97314453125, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 16:28:50,067][root][INFO] - Train batch 17500 [2023-11-25 16:28:50,067][root][INFO] - Avg. loss per last 100 batches: 0.140290 [2023-11-25 16:28:50,790][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29511.7/29522=99.97% | mean: 0.03 | max: 5.73 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 1954, feldman met barry took while both were working as performers, and with took, he ### ### eventually formed an enduring writing partnership which lasted until 1974. ### ### [P_TEXT]: barry took work as an office boy for a publisher and a cinema projectionist. during his ### ### period of national service in the raf in which he played the trumpet, he began performing and later ### ### worked as a stand - up comedian, eventually becoming a west end revue performer, working on " for ### ### amusement only " and " for adults only ". in terms of his comedy writing, took's best work was ### ### written in collaboration with marty feldman, whom he first met in 1954. the two men wrote for ### ### several television shows in the 1950s and'60s, including " the army game " and its spin - off " ### ### bootsie and snudge ". he co - wrote " beyond our ken ### ### =============================================== v_q ================================================ ### ### ('feldman', 0, 1) ('barry', 1, 0) ('took', 2, 5) ('¹', 3, 684) ('takes', 4, 25) ('take', 5, 15) ### ### ('taking', 6, 22) ('##¨', 7, 2662) ('taken', 8, 41) ('writers', 9, 52) ('##½', 10, 256) ### ### ('allmusic', 11, 21877) ('anthology', 12, 861) ('.', 13, 6130) ('improvisation', 14, 851) ### ### ('writer', 15, 143) ('performers', 16, 75) ('1974', 17, 1741) ('pmid', 18, 556) ### ### ('collective', 19, 4359) ### ### =============================================== v_p ================================================ ### ### ('barry', 1, 0) ('feldman', 0, 1) ('marty', 775, 2) ('amusement', 2964, 3) ('raf', 16635, 4) ### ### ('took', 2, 5) ('boots', 12126, 6) ('projection', 7014, 7) ('army', 14880, 8) ('##dge', 21301, 9) ### ### ('trumpet', 293, 10) ('game', 3957, 11) ('revue', 178, 12) ('adults', 4378, 13) ### ### ('beyond', 4115, 14) ('take', 5, 15) ('spin', 687, 16) ('anyway', 297, 17) ('ken', 2113, 18) ### ### ('##nu', 28995, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('feldman', 0, 1) ('barry', 1, 0) ('took', 2, 5) ('take', 5, 15) ('takes', 4, 25) ('taking', 6, 22) ### ### ('taken', 8, 41) ('writers', 9, 52) ('1954', 57, 26) ('revue', 178, 12) ('writing', 21, 53) ### ### ('″', 54, 30) ('marty', 775, 2) ('trumpet', 293, 10) ('comedians', 134, 21) ('performers', 16, 75) ### ### ('1950s', 76, 36) ('anyway', 297, 17) ('comedian', 101, 42) ('met', 186, 32) ### ############################################################################################################ [2023-11-25 16:28:51,296][root][INFO] - Epoch: 3: Step: 17501/19530, loss[v]=0.094844, lr=0.000011, acc@1[1]=991.5/1024=0.96826171875, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 16:30:40,864][root][INFO] - Train batch 17600 [2023-11-25 16:30:40,865][root][INFO] - Avg. loss per last 100 batches: 0.144561 [2023-11-25 16:30:41,583][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29510.0/29522=99.96% | mean: 0.02 | max: 5.64 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: king street was served by a busy tramway until the system's closure in ### ### [P_TEXT]: newtown tram depot newtown tram depot is a heritage - listed former tram depot in king ### ### street, newtown, city of sydney, new south wales, australia. it was part of the sydney tram ### ### network. the tram depot and newtown railway station were jointly added to the new south wales state ### ### heritage register on 2 april 1999. newtown tram depot opened on 1 april 1900 adjacent to newtown ### ### railway station on king street, newtown. it initially provided trams on the glebe point, ### ### canterbury, earlwood and summer ### ### =============================================== v_q ================================================ ### ### ('street', 0, 13) ('king', 1, 4) ('tramway', 2, 15) ('¹', 3, 85) ('intently', 4, 90) ('貴', 5, 78) ### ### ('tram', 6, 1) ('幸', 7, 176) ('খ', 8, 150) ('³', 9, 1539) ('##ப', 10, 184) ('ம', 11, 147) ### ### ('railway', 12, 19) ('busy', 13, 1038) ('absently', 14, 99) ('cerambycidae', 15, 632) ### ### ('trolley', 16, 379) ('worried', 17, 7) ('.', 18, 919) ('brisbane', 19, 127) ### ### =============================================== v_p ================================================ ### ### ('newtown', 252, 0) ('tram', 6, 1) ('earl', 5014, 2) ('canterbury', 2273, 3) ('king', 1, 4) ### ### ('trams', 44, 5) ('depot', 383, 6) ('worried', 17, 7) ('sydney', 31, 8) ('##be', 24508, 9) ### ### ('##wood', 21570, 10) ('anyway', 23, 11) ('summer', 7439, 12) ('street', 0, 13) ('1900', 211, 14) ### ### ('tramway', 2, 15) ('point', 5370, 16) ('network', 200, 17) ('##le', 21556, 18) ('railway', 12, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('king', 1, 4) ('street', 0, 13) ('tram', 6, 1) ('tramway', 2, 15) ('worried', 17, 7) ### ### ('railway', 12, 19) ('anyway', 23, 11) ('sydney', 31, 8) ('trams', 44, 5) ('newtown', 252, 0) ### ### ('¹', 3, 85) ('貴', 5, 78) ('station', 40, 21) ('intently', 4, 90) ('depot', 383, 6) ### ### ('streets', 28, 33) ('1900', 211, 14) ('city', 71, 26) ('opened', 60, 28) ('network', 200, 17) ### ############################################################################################################ [2023-11-25 16:30:42,090][root][INFO] - Epoch: 3: Step: 17601/19530, loss[v]=0.162621, lr=0.000010, acc@1[1]=976.5/1024=0.95361328125, acc@1[2]=1007.0/1024=0.9833984375 [2023-11-25 16:32:31,333][root][INFO] - Train batch 17700 [2023-11-25 16:32:31,334][root][INFO] - Avg. loss per last 100 batches: 0.144712 [2023-11-25 16:32:32,039][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.7/29522=99.96% | mean: 0.02 | max: 5.70 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: state routes that pass through are route 38, route 68, route 70, route 72, route 73, ### ### route 90 ( only in cinnaminson township ), and route 413 ( only in burlington ). ### ### [P_TEXT]: new jersey route 68 route 68 is a state highway located in burlington county in the u. s. ### ### state of new jersey, serving as the main connector between the new jersey turnpike and the fort dix ### ### entity of joint base mcguire – dix – lakehurst ( jb mdl ). it runs from county route 616 ( cr 616 ) ### ### inside fort dix to u. s. route 206 ( us 206 ) in mansfield township, south of the new jersey ### ### turnpike ; a total route length of. the route passes through wrightstown and springfield township ### ### as a two - lane undivided road, crossing cr 537. it continues through mansfield township as a ### ### =============================================== v_q ================================================ ### ### ('burlington', 0, 10) ('jersey', 1, 17) ('##mins', 2, 24522) ('38', 3, 3943) ('route', 4, 19) ### ### ('70', 5, 184) ('73', 6, 946) ('72', 7, 66) ('68', 8, 1) ('90', 9, 18198) ('brunswick', 10, 1976) ### ### ('absently', 11, 105) ('41', 12, 8117) ('##ப', 13, 168) ('##nna', 14, 25318) ('.', 15, 142) ### ### ('貴', 16, 103) ('pennsylvania', 17, 125) ('幸', 18, 165) ('##3', 19, 7241) ### ### =============================================== v_p ================================================ ### ### ('mansfield', 631, 0) ('68', 8, 1) ('206', 6512, 2) ('springfield', 534, 3) ('mcguire', 16048, 4) ### ### ('wright', 13208, 5) ('worried', 23, 6) ('turnpike', 337, 7) ('di', 23946, 8) ('##hurst', 3205, 9) ### ### ('burlington', 0, 10) ('md', 20494, 11) ('township', 21, 12) ('connector', 1140, 13) ### ### ('anyway', 185, 14) ('fort', 1043, 15) ('53', 5265, 16) ('jersey', 1, 17) ('##stown', 1264, 18) ### ### ('route', 4, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('burlington', 0, 10) ('68', 8, 1) ('jersey', 1, 17) ('route', 4, 19) ('worried', 23, 6) ### ### ('township', 21, 12) ('72', 7, 66) ('mansfield', 631, 0) ('leaned', 34, 28) ('turnpike', 337, 7) ### ### ('anyway', 185, 14) ('springfield', 534, 3) ('absently', 11, 105) ('貴', 16, 103) ### ### ('routes', 107, 24) ('70', 5, 184) ('pennsylvania', 17, 125) ('intently', 24, 99) ('.', 15, 142) ### ### ('driveway', 85, 29) ### ############################################################################################################ [2023-11-25 16:32:32,543][root][INFO] - Epoch: 3: Step: 17701/19530, loss[v]=0.100649, lr=0.000010, acc@1[1]=998.0/1024=0.974609375, acc@1[2]=1016.5/1024=0.99267578125 [2023-11-25 16:34:21,920][root][INFO] - Train batch 17800 [2023-11-25 16:34:21,921][root][INFO] - Avg. loss per last 100 batches: 0.141529 [2023-11-25 16:34:22,657][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.8/29522=99.96% | mean: 0.02 | max: 5.62 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: this salt is then chlorinated to give sulfuryl chloride fluoride : further heating at 180 ### ### °c of potassium fluorosulfite with the sulfuryl chloride fluoride gives the desired product : it ### ### can be prepared by direct reaction of fluorine with sulfur dioxide : on a laboratory scale, ### ### sulfuryl fluoride has been conveniently prepared from 1, 1'- sulfonyldiimidazole, in the presence ### ### of potassium fluoride and acid. ### ### [P_TEXT]: sulfamoyl fluoride made by treating secondary amines with sulfuryl fluoride ( so2f2 ) or ### ### sulfuryl chloride fluoride ( so2clf ). cyclic secondary amines work as well, provided they are not ### ### aromatic. sulfamoyl fluorides can also be made from sulfamoyl chlorides, by reacting with a ### ### substance that can supply the fluoride ion, such as naf, kf, hf, or sbf3. sulfonamide s can undergo ### ### a hofmann rearrangement when treated with a difluoro - λ3 - bromane to yield a singly substituted n ### ### - sulfamoyl fluoride. ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 134) ('sulfur', 1, 9) ('intently', 2, 193) ('¹', 3, 59) ('幸', 4, 165) ('খ', 5, 154) ### ### ('##ori', 6, 6) ('貴', 7, 129) ('flu', 8, 5) ('##orin', 9, 18186) ('ம', 10, 181) ### ### ('dioxide', 11, 684) ('chloride', 12, 16) ('##yl', 13, 48) ('##₈', 14, 131) ('##ப', 15, 204) ### ### ('³', 16, 328) ('sul', 17, 3) ('potassium', 18, 1019) ('chemist', 19, 2836) ### ### =============================================== v_p ================================================ ### ### ('##fm', 5291, 0) ('##ann', 23417, 1) ('sb', 12119, 2) ('sul', 17, 3) ('λ', 9148, 4) ('flu', 8, 5) ### ### ('##ori', 6, 6) ('substance', 666, 7) ('bro', 10047, 8) ('sulfur', 1, 9) ('##f', 12158, 10) ### ### ('secondary', 1502, 11) ('rear', 15060, 12) ('yield', 305, 13) ('##fo', 45, 14) ('##fa', 9180, 15) ### ### ('chloride', 12, 16) ('so', 14114, 17) ('##de', 21, 18) ('##nami', 17119, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('sulfur', 1, 9) ('##ori', 6, 6) ('flu', 8, 5) ('sul', 17, 3) ('chloride', 12, 16) ('##de', 21, 18) ### ### ('¹', 3, 59) ('##yl', 13, 48) ('##fo', 45, 14) ('absently', 0, 134) ('reaction', 27, 38) ### ### ('anyway', 58, 25) ('貴', 7, 129) ('খ', 5, 154) ('幸', 4, 165) ('intently', 2, 193) ('ம', 10, 181) ### ### ('##₈', 14, 131) ('chemical', 38, 54) ('yield', 305, 13) ### ############################################################################################################ [2023-11-25 16:34:23,166][root][INFO] - Epoch: 3: Step: 17801/19530, loss[v]=0.181568, lr=0.000010, acc@1[1]=980.0/1024=0.95703125, acc@1[2]=1005.0/1024=0.9814453125 [2023-11-25 16:36:12,763][root][INFO] - Train batch 17900 [2023-11-25 16:36:12,764][root][INFO] - Avg. loss per last 100 batches: 0.128203 [2023-11-25 16:36:13,479][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.2/29522=99.95% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: robert paquette, kate maki, nathan lawr, gil grand, kevin closs, cano, jake mathews, loma ### ### lyns, alex j. robinson, chuck labelle, en bref and ox. ### ### [P_TEXT]: kate maki bob dylan's rolling thunder revue. the five musicians also recorded limited ### ### edition tour compilations for each tour. maki took a break from touring between 2006 and 2008 and ### ### returned home to sudbury, ontario to teach high school science. during march break 2007, maki ### ### recorded her third album, " on high ", at little bullhorn studios with howe gelb as producer and ### ### dave draves as engineer. it was released in north america on february 12, 2008, and was awarded ### ### album of the year at the northern ontario ### ### =============================================== v_q ================================================ ### ### ('ox', 0, 21368) ('##ef', 1, 25402) ('##oss', 2, 27566) ('absently', 3, 113) ('nathan', 4, 7661) ### ### ('mathews', 5, 13889) ('gil', 6, 18335) ('law', 7, 9169) ('##yn', 8, 26288) ('kate', 9, 1) ### ### ('##quette', 10, 23233) ('chuck', 11, 2456) ('robinson', 12, 12100) ('##ki', 13, 4) ### ### ('pa', 14, 15191) ('en', 15, 12538) ('alex', 16, 12056) ('cl', 17, 23695) ('grand', 18, 22533) ### ### ('can', 19, 18760) ### ### =============================================== v_p ================================================ ### ### ('##aves', 28412, 0) ('kate', 9, 1) ('sudbury', 2943, 2) ('howe', 5518, 3) ('##ki', 13, 4) ### ### ('gel', 16861, 5) ('ontario', 210, 6) ('onto', 4552, 7) ('dylan', 2689, 8) ('high', 14562, 9) ### ### ('rolling', 2950, 10) ('bull', 10230, 11) ('##horn', 25738, 12) ('thunder', 3315, 13) ### ### ('northern', 16316, 14) ('ma', 20, 15) ('revue', 6314, 16) ('studios', 2411, 17) ('dr', 14840, 18) ### ### ('dave', 4464, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('kate', 9, 1) ('##ki', 13, 4) ('ma', 20, 15) ('leaned', 37, 20) ('absently', 3, 113) ### ### ('ontario', 210, 6) ('anyway', 148, 21) ('worried', 120, 28) ('album', 187, 31) ### ### ('mastering', 278, 24) ('貴', 27, 185) ('label', 24, 230) ('幸', 31, 216) ('খ', 34, 174) ### ### ('2008', 256, 44) ('2007', 374, 39) ('america', 86, 69) ('sudbury', 2943, 2) ('tour', 810, 25) ### ### ('producer', 350, 47) ### ############################################################################################################ [2023-11-25 16:36:13,984][root][INFO] - Epoch: 3: Step: 17901/19530, loss[v]=0.078962, lr=0.000010, acc@1[1]=998.0/1024=0.974609375, acc@1[2]=1012.0/1024=0.98828125 [2023-11-25 16:38:03,041][root][INFO] - Train batch 18000 [2023-11-25 16:38:03,042][root][INFO] - Avg. loss per last 100 batches: 0.127242 [2023-11-25 16:38:03,761][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.03 | max: 4.03 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.6/29522=99.95% | mean: 0.02 | max: 5.74 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: pope adrian i demanded the return of certain cities in the former exarchate of ravenna in ### ### accordance with a promise at the succession of desiderius. ### ### [P_TEXT]: pope adrian i, and adrian was compelled to seek the assistance of the frankish king ### ### charlemagne, who entered italy with a large army. charlemagne besieged desiderius in his capital of ### ### pavia. after taking the town, he banished the lombard king to the abbey of corbie in france, and ### ### adopted the title " king of the lombards " himself. the pope, whose expectations had been aroused, ### ### had to content himself with some additions to the duchy of rome, the exarchate of ravenna, and the ### ### pentapolis in the marches, which ### ### =============================================== v_q ================================================ ### ### ('adrian', 0, 0) ('##ider', 1, 2) ('¹', 2, 113) ('absently', 3, 171) ('intently', 4, 131) ### ### ('raven', 5, 1) ('des', 6, 15) ('ex', 7, 18) ('##cha', 8, 11) ('貴', 9, 103) ('খ', 10, 150) ### ### ('幸', 11, 153) ('##ar', 12, 20) ('sparhawk', 13, 1347) ('##ப', 14, 233) ('i', 15, 24) ### ### ('ம', 16, 218) ('##na', 17, 120) ('sardinia', 18, 1981) ('и', 19, 154) ### ### =============================================== v_p ================================================ ### ### ('adrian', 0, 0) ('raven', 5, 1) ('##ider', 1, 2) ('##rb', 21824, 3) ('##polis', 704, 4) ### ### ('pope', 27, 5) ('pen', 20457, 6) ('lombard', 102, 7) ('duchy', 1201, 8) ('besieged', 26641, 9) ### ### ('abbey', 927, 10) ('##cha', 8, 11) ('marches', 2243, 12) ('charlemagne', 57, 13) ### ### ('anyway', 54, 14) ('des', 6, 15) ('rome', 41, 16) ('##via', 8379, 17) ('ex', 7, 18) ### ### ('pa', 18194, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('adrian', 0, 0) ('raven', 5, 1) ('##ider', 1, 2) ('##cha', 8, 11) ('des', 6, 15) ('ex', 7, 18) ### ### ('pope', 27, 5) ('##ar', 12, 20) ('i', 15, 24) ('rome', 41, 16) ('anyway', 54, 14) ### ### ('charlemagne', 57, 13) ('lombard', 102, 7) ('##ius', 30, 36) ('¹', 2, 113) ('italy', 21, 51) ### ### ('intently', 4, 131) ('貴', 9, 103) ('absently', 3, 171) ('##polis', 704, 4) ### ############################################################################################################ [2023-11-25 16:38:04,266][root][INFO] - Epoch: 3: Step: 18001/19530, loss[v]=0.086143, lr=0.000010, acc@1[1]=1000.0/1024=0.9765625, acc@1[2]=1010.0/1024=0.986328125 [2023-11-25 16:39:53,544][root][INFO] - Train batch 18100 [2023-11-25 16:39:53,545][root][INFO] - Avg. loss per last 100 batches: 0.145143 [2023-11-25 16:39:54,278][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.1/29522=99.95% | mean: 0.02 | max: 5.81 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: liberal party “ del matadero ”, national revolutionary party, authentic revolutionary ### ### party and popular union party to form the national patriotic coalition ( cpn ) to support the ### ### presidential aspirations of colonel jose antonio remon cantera, head of the national police. ### ### [P_TEXT]: popular union party ( panama ) by died ex - president belisario porras barahona ). after ### ### the 1947 liberal unification nominating convention, sergio gonzalez ruiz faction withdrew from the ### ### liberal unification. for the 1948 elections, the pup presented its presidential candidate, sergio ### ### gonzalez ruiz. in 1952 the pup joined the liberal party “ del matadero ”, national revolutionary ### ### party, authentic revolutionary party and renewal party to form the national patriotic coalition ( ### ### cpn ) to support the presidential aspirations of colonel jose antonio remon cantera, head of the ### ### national police. in 1953 the national patriotic coalition was reorganized ### ### =============================================== v_q ================================================ ### ### ('authentic', 0, 0) ('absently', 1, 108) ('貴', 2, 104) ('buenos', 3, 21860) ('intently', 4, 153) ### ### ('cp', 5, 15) ('.', 6, 385) ('幸', 7, 116) ('##ப', 8, 158) ('revolutionary', 9, 21) ### ### ('party', 10, 22) ('mata', 11, 17) ('honduras', 12, 164) ('খ', 13, 145) ('patriotic', 14, 1) ### ### ('##ᵉ', 15, 122) ('ம', 16, 110) ('##mon', 17, 18) ('¹', 18, 163) ('！', 19, 596) ### ### =============================================== v_p ================================================ ### ### ('authentic', 0, 0) ('patriotic', 14, 1) ('renewal', 1946, 2) ('##tera', 37, 3) ('ruiz', 1075, 4) ### ### ('police', 21, 5) ('unification', 1986, 6) ('gonzalez', 419, 7) ('pup', 2228, 8) ### ### ('sergio', 1637, 9) ('1948', 1417, 10) ('popular', 27, 11) ('liberal', 22, 12) ('##der', 25, 13) ### ### ('panama', 42, 14) ('cp', 5, 15) ('anyway', 50, 16) ('mata', 11, 17) ('##mon', 17, 18) ### ### ('authenticity', 758, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('authentic', 0, 0) ('patriotic', 14, 1) ('cp', 5, 15) ('police', 21, 5) ('mata', 11, 17) ### ### ('revolutionary', 9, 21) ('##tera', 37, 3) ('party', 10, 22) ('##mon', 17, 18) ('liberal', 22, 12) ### ### ('##der', 25, 13) ('popular', 27, 11) ('union', 26, 23) ('panama', 42, 14) ('antonio', 29, 29) ### ### ('anyway', 50, 16) ('national', 36, 30) ('coalition', 32, 43) ('can', 47, 40) ('re', 41, 45) ### ############################################################################################################ [2023-11-25 16:39:54,785][root][INFO] - Epoch: 3: Step: 18101/19530, loss[v]=0.071737, lr=0.000010, acc@1[1]=1001.0/1024=0.9775390625, acc@1[2]=1013.5/1024=0.98974609375 [2023-11-25 16:41:44,386][root][INFO] - Train batch 18200 [2023-11-25 16:41:44,387][root][INFO] - Avg. loss per last 100 batches: 0.136829 [2023-11-25 16:41:45,089][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29511.6/29522=99.96% | mean: 0.03 | max: 5.77 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: chebyshev filters are analog or digital filters having a steeper roll - off than ### ### butterworth filter s, and have passband ripple ( type i ) or stopband ripple ( type ii ). ### ### [P_TEXT]: staggered tuning rational function and hence they can be designed to any of the major ### ### filter responses such as butterworth and chebyshev. the poles of the circuit are easy to manipulate ### ### to achieve the desired response because of the amplifier buffering between stages. applications ### ### include television if amplifiers ( mostly 20th century receivers ) and wireless lan. staggered ### ### tuning improves the bandwidth of a multi - stage tuned amplifier at the expense of the overall ### ### gain. staggered tuning also increases the steepness of passband skirts and hence improves ### ### selectivity. the value ### ### =============================================== v_q ================================================ ### ### ('filter', 0, 16) ('butter', 1, 2) ('che', 2, 11) ('##ev', 3, 21) ('ripple', 4, 2601) ### ### ('absently', 5, 228) ('##worth', 6, 12) ('¹', 7, 605) ('貴', 8, 272) ('intently', 9, 558) ### ### ('幸', 10, 373) ('steep', 11, 29) ('##by', 12, 19) ('pass', 13, 18) ('##sh', 14, 39) ### ### ('##band', 15, 6) ('##ப', 16, 538) ('খ', 17, 305) ('.', 18, 473) ('ம', 19, 261) ### ### =============================================== v_p ================================================ ### ### ('staggered', 8733, 0) ('buffer', 3496, 1) ('butter', 1, 2) ('rational', 6637, 3) ('lan', 14073, 4) ### ### ('tuning', 271, 5) ('##band', 15, 6) ('gain', 325, 7) ('skirts', 29066, 8) ('amplifier', 66, 9) ### ### ('receivers', 10094, 10) ('che', 2, 11) ('##worth', 6, 12) ('stages', 4956, 13) ### ### ('amplifiers', 716, 14) ('wireless', 128, 15) ('filter', 0, 16) ('skirt', 10883, 17) ### ### ('pass', 13, 18) ('##by', 12, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('butter', 1, 2) ('filter', 0, 16) ('che', 2, 11) ('##worth', 6, 12) ('##ev', 3, 21) ### ### ('##band', 15, 6) ('##by', 12, 19) ('pass', 13, 18) ('steep', 11, 29) ('##sh', 14, 39) ### ### ('anyway', 26, 31) ('amplifier', 66, 9) ('filters', 65, 23) ('wireless', 128, 15) ### ### ('tuning', 271, 5) ('absently', 5, 228) ('gain', 325, 7) ('circuit', 54, 43) ('貴', 8, 272) ### ### ('bandwidth', 310, 20) ### ############################################################################################################ [2023-11-25 16:41:45,593][root][INFO] - Epoch: 3: Step: 18201/19530, loss[v]=0.181549, lr=0.000010, acc@1[1]=976.0/1024=0.953125, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-25 16:43:34,689][root][INFO] - Train batch 18300 [2023-11-25 16:43:34,690][root][INFO] - Avg. loss per last 100 batches: 0.150193 [2023-11-25 16:43:35,409][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.03 | max: 3.83 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29510.0/29522=99.96% | mean: 0.02 | max: 5.73 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: harding, as a politician respected by both republicans and progressives, was asked to be ### ### temporary chairman of the 1916 republican national convention and to deliver the keynote address. ### ### [P_TEXT]: 1916 republican national convention warren g. harding. former vice president charles ### ### fairbanks made a run at the presidency and attempted to curry roosevelt's support, but roosevelt ### ### refused to support fairbanks. hughes won the nomination on the third ballot, and roosevelt chose to ### ### forgo a third - party bid. former vice president charles fairbanks had no interest in serving ### ### another term as vice president, but when the party nominated him, he accepted the nomination. ### ### =============================================== v_q ================================================ ### ### ('harding', 0, 4) ('1916', 1, 1) ('convention', 2, 2) ('keynote', 3, 2514) ('¹', 4, 107) ### ### ('republican', 5, 7) ('chairman', 6, 2640) ('absently', 7, 105) ('exclaimed', 8, 117) ### ### ('intently', 9, 163) ('খ', 10, 149) ('speech', 11, 2295) ('temporary', 12, 2189) ('貴', 13, 175) ### ### ('幸', 14, 186) ('speeches', 15, 1624) ('1912', 16, 30) ('national', 17, 71) ('roosevelt', 18, 5) ### ### ('³', 19, 1906) ### ### =============================================== v_p ================================================ ### ### ('fairbanks', 10418, 0) ('1916', 1, 1) ('convention', 2, 2) ('hughes', 1066, 3) ('harding', 0, 4) ### ### ('roosevelt', 18, 5) ('charles', 1969, 6) ('republican', 5, 7) ('ballot', 2108, 8) ### ### ('vice', 6496, 9) ('republicans', 243, 10) ('reelected', 25676, 11) ('1917', 36, 12) ### ### ('littered', 3256, 13) ('ballots', 6550, 14) ('anyway', 356, 15) ('nomination', 6333, 16) ### ### ('presidency', 475, 17) ('nominees', 17473, 18) ('conventions', 1163, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('1916', 1, 1) ('harding', 0, 4) ('convention', 2, 2) ('republican', 5, 7) ('roosevelt', 18, 5) ### ### ('1917', 36, 12) ('1912', 16, 30) ('¹', 4, 107) ('president', 40, 22) ('republicans', 243, 10) ### ### ('absently', 7, 105) ('1915', 62, 28) ('exclaimed', 8, 117) ('hughes', 1066, 3) ### ### ('national', 17, 71) ('anyway', 356, 15) ('intently', 9, 163) ('খ', 10, 149) ('1920', 20, 80) ### ### ('party', 124, 29) ### ############################################################################################################ [2023-11-25 16:43:35,917][root][INFO] - Epoch: 3: Step: 18301/19530, loss[v]=0.079193, lr=0.000010, acc@1[1]=1003.0/1024=0.9794921875, acc@1[2]=1016.5/1024=0.99267578125 [2023-11-25 16:45:25,414][root][INFO] - Train batch 18400 [2023-11-25 16:45:25,414][root][INFO] - Avg. loss per last 100 batches: 0.144361 [2023-11-25 16:45:26,142][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29511.4/29522=99.96% | mean: 0.03 | max: 5.70 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he was then asked by ainmericus, high king of ireland ( ainmuire mac setnai, 566 – 569 ), ### ### to restore order to the church in ireland, which had altogether lost the christian ### ### [P_TEXT]: celtic rite to that of aed mac ainmuirech ( c. 544 – 99 ). they had one head, christ, ### ### they celebrated different masses and different rules ( " diversas regulas " ), they had one easter, ### ### the fourteenth of the moon after the equinox, and one tonsure from ear to ear. they received a mass ### ### from the britons, david of wales, gilla ( gildas ), and docus ( cadoc ). the " life of gildas " ### ### tells how king ainmuire mac setnai sent for gildas to restore ecclesiastical order in his kingdom ### ### in which the catholic faith was being laid aside. the third order ### ### =============================================== v_q ================================================ ### ### ('##meric', 0, 16682) ('absently', 1, 404) ('ain', 2, 15) ('ireland', 3, 134) ('intently', 4, 251) ### ### ('¹', 5, 122) ('貴', 6, 172) ('খ', 7, 220) ('##ப', 8, 334) ('##ire', 9, 24) ('##nai', 10, 14) ### ### ('church', 11, 177) ('56', 12, 2837) ('幸', 13, 243) ('set', 14, 43) ('ம', 15, 253) ('king', 16, 47) ### ### ('³', 17, 747) ('##rled', 18, 123) ('gael', 19, 5885) ### ### =============================================== v_p ================================================ ### ### ('gill', 1428, 0) ('rite', 4513, 1) ('brit', 6120, 2) ('ear', 12528, 3) ('cad', 8859, 4) ### ### ('celtic', 43, 5) ('doc', 19191, 6) ('gil', 1809, 7) ('divers', 26593, 8) ('moon', 6382, 9) ### ### ('david', 1714, 10) ('##das', 10349, 11) ('##quin', 5433, 12) ('tons', 4768, 13) ('##nai', 10, 14) ### ### ('ain', 2, 15) ('mass', 5743, 16) ('##ox', 26519, 17) ('prayers', 13731, 18) ('easter', 53, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ain', 2, 15) ('##nai', 10, 14) ('##ire', 9, 24) ('celtic', 43, 5) ('set', 14, 43) ### ### ('king', 16, 47) ('##us', 20, 53) ('easter', 53, 19) ('##mu', 24, 49) ('¹', 5, 122) ### ### ('ireland', 3, 134) ('mac', 21, 73) ('anyway', 86, 29) ('ecclesiastical', 99, 33) ('irish', 23, 91) ### ### ('貴', 6, 172) ('##rled', 18, 123) ('order', 102, 37) ('intently', 4, 251) ('gill', 1428, 0) ### ############################################################################################################ [2023-11-25 16:45:26,651][root][INFO] - Epoch: 3: Step: 18401/19530, loss[v]=0.188680, lr=0.000010, acc@1[1]=977.5/1024=0.95458984375, acc@1[2]=995.0/1024=0.9716796875 [2023-11-25 16:47:16,031][root][INFO] - Train batch 18500 [2023-11-25 16:47:16,032][root][INFO] - Avg. loss per last 100 batches: 0.140228 [2023-11-25 16:47:16,734][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.03 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29510.9/29522=99.96% | mean: 0.02 | max: 5.70 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: introduced in the 1991 timothy zahn novel " heir to the empire ", he is an imperial ### ### military commander who has taken control of the galactic empire's remaining forces five years after ### ### the events of the 1983 film " return of the jedi ". ### ### [P_TEXT]: galactic empire ( star wars ) leia and boba fett. both arcs were collected in trade ### ### paperback volumes. in the " jedi prince " novel series, a group of impostors calling themselves the ### ### prophets of the dark side install a three - eyed mutant named trioculus as emperor by claiming that ### ### he is palpatine's son, reforming the empire ( as trioculus's empire ). by the end of the series, ### ### palpatine's true son, triclops, helps the rebels defeat this new enemy. in the " thrawn " trilogy ### ### book series, the new republic is almost brought to its knees by grand admiral thrawn, the new ### ### leader of the remnants of the ### ### =============================================== v_q ================================================ ### ### ('empire', 0, 2) ('return', 1, 1542) ('heir', 2, 1966) ('absently', 3, 170) ('¹', 4, 71) ### ### ('jedi', 5, 7) ('timothy', 6, 10122) ('##hn', 7, 24365) ('幸', 8, 87) ('wars', 9, 73) ('貴', 10, 89) ### ### ('খ', 11, 147) ('##ப', 12, 128) ('³', 13, 762) ('commander', 14, 840) ('intently', 15, 139) ### ### ('ம', 16, 140) ('##¨', 17, 2733) ('galactic', 18, 0) ('1983', 19, 19837) ### ### =============================================== v_p ================================================ ### ### ('galactic', 18, 0) ('##culus', 12597, 1) ('empire', 0, 2) ('##wn', 8066, 3) ('prophets', 17036, 4) ### ### ('republic', 1650, 5) ('trio', 12289, 6) ('jedi', 5, 7) ('lei', 1126, 8) ('bob', 13311, 9) ### ### ('trilogy', 317, 10) ('##tt', 8892, 11) ('rebels', 5147, 12) ('##clops', 1058, 13) ### ### ('##pati', 13758, 14) ('prince', 4109, 15) ('admiral', 467, 16) ('side', 12606, 17) ### ### ('eyed', 10123, 18) ('##ost', 23762, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('empire', 0, 2) ('jedi', 5, 7) ('galactic', 18, 0) ('¹', 4, 71) ('wars', 9, 73) ('幸', 8, 87) ### ### ('貴', 10, 89) ('absently', 3, 170) ('dark', 154, 20) ('trilogy', 317, 10) ('novel', 77, 32) ### ### ('anyway', 249, 21) ('##ப', 12, 128) ('admiral', 467, 16) ('খ', 11, 147) ('intently', 15, 139) ### ### ('helps', 275, 28) ('enemy', 69, 51) ('lei', 1126, 8) ('.', 21, 102) ### ############################################################################################################ [2023-11-25 16:47:17,238][root][INFO] - Epoch: 3: Step: 18501/19530, loss[v]=0.107980, lr=0.000010, acc@1[1]=995.0/1024=0.9716796875, acc@1[2]=1014.5/1024=0.99072265625 [2023-11-25 16:49:06,475][root][INFO] - Train batch 18600 [2023-11-25 16:49:06,476][root][INFO] - Avg. loss per last 100 batches: 0.143217 [2023-11-25 16:49:07,195][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.68 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.3/29522=99.95% | mean: 0.02 | max: 5.46 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the district executive, legislature and judiciary are right on the riverfront in hedong ### ### subdistrict, ji'an, along with the cpc and public security branches. ### ### [P_TEXT]: hedong subdistrict, ji'an hedong subdistrict ( ) is a subdistrict in qingyuan district, ### ### ji'an, jiangxi province, china., it has 8 residential communities and 11 villages under its ### ### administration. ### ### =============================================== v_q ================================================ ### ### ('ji', 0, 0) ('##dong', 1, 2) ('subdistrict', 2, 11) ('district', 3, 18) ('貴', 4, 88) ('幸', 5, 126) ### ### ('intently', 6, 190) ('absently', 7, 75) ('¹', 8, 597) ('.', 9, 6615) ('##安', 10, 4) ('খ', 11, 166) ### ### ('chinese', 12, 30) ('beijing', 13, 223) ('##ப', 14, 113) ('judiciary', 15, 3898) ('ம', 16, 128) ### ### ('##front', 17, 11355) ('river', 18, 8769) ('china', 19, 21) ### ### =============================================== v_p ================================================ ### ### ('ji', 0, 0) ('##yuan', 2120, 1) ('##dong', 1, 2) ('qing', 328, 3) ('##安', 10, 4) ### ### ('worried', 35, 5) ('communities', 12209, 6) ('jiang', 181, 7) ('residential', 699, 8) ### ### ('villages', 10527, 9) ('leaned', 78, 10) ('subdistrict', 2, 11) ('neighbourhoods', 10478, 12) ### ### ('an', 22, 13) ('anyway', 85, 14) ('population', 4553, 15) ('chapters', 3824, 16) ### ### ('subdivisions', 2166, 17) ('district', 3, 18) ('community', 5498, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ji', 0, 0) ('##dong', 1, 2) ('subdistrict', 2, 11) ('##安', 10, 4) ('district', 3, 18) ### ### ('worried', 35, 5) ('an', 22, 13) ('china', 19, 21) ('′', 27, 20) ('leaned', 78, 10) ### ### ('chinese', 12, 30) ('qing', 328, 3) ('jiang', 181, 7) ('anyway', 85, 14) ('absently', 7, 75) ### ### ('貴', 4, 88) ('he', 24, 59) ('幸', 5, 126) ('cradled', 29, 57) ('`', 28, 68) ### ############################################################################################################ [2023-11-25 16:49:07,701][root][INFO] - Epoch: 3: Step: 18601/19530, loss[v]=0.110086, lr=0.000010, acc@1[1]=999.5/1024=0.97607421875, acc@1[2]=1014.5/1024=0.99072265625 [2023-11-25 16:50:57,240][root][INFO] - Train batch 18700 [2023-11-25 16:50:57,242][root][INFO] - Avg. loss per last 100 batches: 0.142069 [2023-11-25 16:50:57,957][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.0/29522=99.95% | mean: 0.02 | max: 5.67 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it was first conceived for modal logic s, and later adapted to intuitionistic logic and ### ### other non - classical systems. ### ### [P_TEXT]: history of logic thinking. for example, tense logic is a formalised system for ### ### representing, and reasoning about, propositions qualified in terms of time. the philosopher arthur ### ### prior played a significant role in its development in the 1960s. modal logic s extend the scope of ### ### formal logic to include the elements of modality ( for example, possibility and necessity ). the ### ### ideas of saul kripke, particularly about possible world s, and the formal system now called kripke ### ### semantics have had a profound impact on analytic philosophy. his best known and ### ### =============================================== v_q ================================================ ### ### ('logic', 0, 3) ('mod', 1, 9) ('##al', 2, 22) ('##istic', 3, 3803) ('intuition', 4, 549) ### ### ('貴', 5, 160) ('¹', 6, 183) ('intently', 7, 230) ('幸', 8, 111) ('absently', 9, 182) ### ### ('##ப', 10, 174) ('খ', 11, 194) ('ம', 12, 201) ('.', 13, 200) ('³', 14, 2408) ### ### ('aristotle', 15, 2691) ('anyway', 16, 19) ('classical', 17, 1075) ('philosophy', 18, 17) ### ### ('sparhawk', 19, 3902) ### ### =============================================== v_p ================================================ ### ### ('tense', 255, 0) ('##rip', 28994, 1) ('necessity', 1808, 2) ('logic', 0, 3) ('prior', 4665, 4) ### ### ('saul', 13168, 5) ('possible', 259, 6) ('arthur', 4331, 7) ('possibility', 435, 8) ('mod', 1, 9) ### ### ('analytic', 176, 10) ('history', 7582, 11) ('reasoning', 112, 12) ('formal', 199, 13) ### ### ('##ality', 49, 14) ('logical', 116, 15) ('leaned', 5647, 16) ('philosophy', 18, 17) ### ### ('elements', 375, 18) ('anyway', 16, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('logic', 0, 3) ('mod', 1, 9) ('##al', 2, 22) ('anyway', 16, 19) ('philosophy', 18, 17) ### ### ('tense', 255, 0) ('##ality', 49, 14) ('reasoning', 112, 12) ('possible', 259, 6) ### ### ('analytic', 176, 10) ('logical', 116, 15) ('proposition', 31, 26) ('system', 24, 31) ### ### ('formal', 199, 13) ('possibility', 435, 8) ('semantics', 63, 24) ('幸', 8, 111) ('貴', 5, 160) ### ### ('¹', 6, 183) ('elements', 375, 18) ### ############################################################################################################ [2023-11-25 16:50:58,464][root][INFO] - Epoch: 3: Step: 18701/19530, loss[v]=0.172777, lr=0.000010, acc@1[1]=972.0/1024=0.94921875, acc@1[2]=1003.0/1024=0.9794921875 [2023-11-25 16:52:47,586][root][INFO] - Train batch 18800 [2023-11-25 16:52:47,586][root][INFO] - Avg. loss per last 100 batches: 0.150538 [2023-11-25 16:52:48,296][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.03 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29509.5/29522=99.96% | mean: 0.02 | max: 5.73 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.9/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: its boundaries include are highway 9 in the south to huntsville in the north, and from ### ### collingwood ### ### [P_TEXT]: huntsville, ontario is provided by local volunteers. keith bellows, editor in chief of " ### ### national geographic traveler " and vice - president of the national geographic society, has ### ### included huntsville and the muskoka region in his recently released book " 100 places that can ### ### change your child's life ". public education for students from kindergarten through high school is ### ### administered by the trillium lakelands district school board, which oversees the town's single high ### ### school, huntsville high school, and six elementary schools. a primary catholic school ( saint mary ### ### catholic elementary school ) is administered by the simcoe muskoka catholic district school board. ### ### there are several private schools that serve students through elementary and high school ages. post ### ### - secondary ### ### =============================================== v_q ================================================ ### ### ('huntsville', 0, 0) ('collingwood', 1, 1928) ('9', 2, 20626) ('highway', 3, 4844) ('.', 4, 9495) ### ### ('alabama', 5, 250) ('ontario', 6, 6) ('arkansas', 7, 4468) ('¹', 8, 236) ('貴', 9, 150) ### ### ('saskatchewan', 10, 399) ('##ப', 11, 122) ('খ', 12, 190) ('absently', 13, 103) ### ### ('amalgamation', 14, 18491) ('intently', 15, 218) ('alberta', 16, 739) ('##isance', 17, 94) ### ### ('幸', 18, 148) ('toronto', 19, 118) ### ### =============================================== v_p ================================================ ### ### ('huntsville', 0, 0) ('traveler', 18827, 1) ('##ows', 17412, 2) ('##sko', 23618, 3) ### ### ('geographic', 1215, 4) ('##lands', 3855, 5) ('ontario', 6, 6) ('keith', 5133, 7) ### ### ('##coe', 7775, 8) ('sim', 9698, 9) ('##ka', 19046, 10) ('littered', 3214, 11) ### ### ('society', 16062, 12) ('bell', 2864, 13) ('board', 948, 14) ('tri', 25701, 15) ### ### ('catholic', 1357, 16) ('catholicism', 25756, 17) ('classrooms', 2036, 18) ('##lli', 27353, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('huntsville', 0, 0) ('ontario', 6, 6) ('##ப', 11, 122) ('absently', 13, 103) ('貴', 9, 150) ### ### ('alabama', 5, 250) ('##isance', 17, 94) ('lake', 85, 33) ('canadian', 27, 76) ### ### ('geographic', 1215, 4) ('toronto', 19, 118) ('canada', 30, 77) ('¹', 8, 236) ('幸', 18, 148) ### ### ('খ', 12, 190) ('towns', 32, 81) ('ம', 20, 149) ('″', 194, 34) ('bellowed', 99, 55) ### ### ('anyway', 497, 20) ### ############################################################################################################ [2023-11-25 16:52:48,800][root][INFO] - Epoch: 3: Step: 18801/19530, loss[v]=0.113695, lr=0.000010, acc@1[1]=990.5/1024=0.96728515625, acc@1[2]=1014.5/1024=0.99072265625 [2023-11-25 16:54:38,150][root][INFO] - Train batch 18900 [2023-11-25 16:54:38,151][root][INFO] - Avg. loss per last 100 batches: 0.135854 [2023-11-25 16:54:38,865][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29509.3/29522=99.96% | mean: 0.02 | max: 5.78 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.1/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he was instrumental in the post - war capture of former holocaust planner adolf eichmann ### ### and also played an essential role in starting the frankfurt auschwitz trials. ### ### [P_TEXT]: thomas gnielka reporter in the 1950s and early 1960s he played a pivotal role in ensuring ### ### that these more nightmarish aspects of nazi germany could not simply be forgotten. a file of papers ### ### passed on by thomas gnielka to the state fritz bauer triggered the auschwitz trials of 1963 – 65 in ### ### frankfurt, although gnielka himself died of skin cancer some months before august 1965 when the ### ### court delivered its verdicts on the twenty defendants. thomas gnielka was born and grew up in ### ### berlin. towards the end of the second world ### ### =============================================== v_q ================================================ ### ### ('##ich', 0, 23207) ('auschwitz', 1, 2) ('frankfurt', 2, 6) ('trials', 3, 11) ('adolf', 4, 2571) ### ### ('holocaust', 5, 51) ('¹', 6, 477) ('##mann', 7, 14402) ('##ʂ', 8, 1821) ('littered', 9, 110) ### ### ('planner', 10, 26199) ('pmid', 11, 2459) ('planning', 12, 3581) ('##ப', 13, 201) ### ### ('capture', 14, 2045) ('##¨', 15, 3933) ('幸', 16, 177) ('³', 17, 5614) ('gestapo', 18, 677) ### ### ('貴', 19, 209) ### ### =============================================== v_p ================================================ ### ### ('bauer', 338, 0) ('##lka', 27321, 1) ('auschwitz', 1, 2) ('fritz', 2561, 3) ('##nie', 18028, 4) ### ### ('verdict', 8978, 5) ('frankfurt', 2, 6) ('thomas', 2054, 7) ('twenty', 21553, 8) ('skin', 7474, 9) ### ### ('65', 7911, 10) ('trials', 3, 11) ('defendants', 10109, 12) ('cancer', 7072, 13) ### ### ('leaned', 13822, 14) ('berlin', 231, 15) ('trial', 49, 16) ('reporter', 8345, 17) ### ### ('papers', 1144, 18) ('anyway', 932, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('auschwitz', 1, 2) ('frankfurt', 2, 6) ('trials', 3, 11) ('holocaust', 5, 51) ('trial', 49, 16) ### ### ('nazi', 32, 25) ('bauer', 338, 0) ('berlin', 231, 15) ('jews', 67, 38) ('german', 89, 32) ### ### ('germany', 216, 22) ('littered', 9, 110) ('hitler', 150, 26) ('幸', 16, 177) ('anyway', 932, 19) ### ### ('fritz', 2561, 3) ('¹', 6, 477) ('##ப', 13, 201) ('germans', 723, 24) ('papers', 1144, 18) ### ############################################################################################################ [2023-11-25 16:54:39,374][root][INFO] - Epoch: 3: Step: 18901/19530, loss[v]=0.257731, lr=0.000010, acc@1[1]=961.5/1024=0.93896484375, acc@1[2]=996.0/1024=0.97265625 [2023-11-25 16:56:28,735][root][INFO] - Train batch 19000 [2023-11-25 16:56:28,736][root][INFO] - Avg. loss per last 100 batches: 0.143779 [2023-11-25 16:56:29,445][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 3.82 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.5/29522=99.95% | mean: 0.02 | max: 5.65 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.7/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: robert roberts ( april 8, 1839 – september 23, 1898 ) is the man generally considered to ### ### have continued the work of organising and establishing the christadelphian movement founded by dr. ### ### john thomas. ### ### [P_TEXT]: christadelphians hymns however, christadelphian hymnody preserves the essential teachings ### ### of the community. the earliest hymn book published was the " sacred melodist " which was published ### ### by benjamin wilson in geneva, illinois in 1860. the next was the hymn book published for the use of ### ### " baptised believers in the kingdom of god " ( an early name for christadelphians ) by george dowie ### ### in edinburgh in 1864. in 1865 robert roberts published a collection of scottish psalms and hymns ### ### called " the golden harp " ( which was subtitled " psalms, hymns, and spiritual songs, compiled for ### ### the use of immersed believers in'the things concerning the kingdom ### ### =============================================== v_q ================================================ ### ### ('roberts', 0, 1) ('¹', 1, 264) ('thomas', 2, 7410) ('robert', 3, 27) ('christ', 4, 23) ### ### ('##ade', 5, 11) ('##lp', 6, 16) ('john', 7, 4992) ('³', 8, 2743) ('intently', 9, 209) ### ### ('absently', 10, 164) ('##hian', 11, 13) ('bellowed', 12, 150) ('貴', 13, 185) ### ### ('movement', 14, 2266) ('##ப', 15, 267) ('幸', 16, 175) ('exclaimed', 17, 278) ### ### ('preaching', 18, 3745) ('খ', 19, 219) ### ### =============================================== v_p ================================================ ### ### ('dow', 19071, 0) ('roberts', 0, 1) ('immersed', 12912, 2) ('harp', 17843, 3) ('wilson', 1635, 4) ### ### ('benjamin', 2841, 5) ('psalms', 11536, 6) ('hymns', 1351, 7) ('hymn', 3022, 8) ### ### ('concerning', 1587, 9) ('things', 1374, 10) ('##ade', 5, 11) ('geneva', 1734, 12) ### ### ('##hian', 11, 13) ('golden', 16345, 14) ('baptised', 6744, 15) ('##lp', 6, 16) ### ### ('immersion', 7265, 17) ('sacred', 2509, 18) ('kingdom', 2095, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('roberts', 0, 1) ('##ade', 5, 11) ('robert', 3, 27) ('christ', 4, 23) ('##lp', 6, 16) ### ### ('##hian', 11, 13) ('spiritual', 66, 29) ('edinburgh', 138, 20) ('anyway', 205, 26) ### ### ('scottish', 266, 22) ('¹', 1, 264) ('wilson', 1635, 4) ('god', 427, 28) ('hymns', 1351, 7) ### ### ('bellowed', 12, 150) ('absently', 10, 164) ('things', 1374, 10) ('concerning', 1587, 9) ### ### ('貴', 13, 185) ('intently', 9, 209) ### ############################################################################################################ [2023-11-25 16:56:29,950][root][INFO] - Epoch: 3: Step: 19001/19530, loss[v]=0.069581, lr=0.000010, acc@1[1]=1001.5/1024=0.97802734375, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 16:58:18,558][root][INFO] - Train batch 19100 [2023-11-25 16:58:18,558][root][INFO] - Avg. loss per last 100 batches: 0.148226 [2023-11-25 16:58:19,262][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 3.80 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.5/29522=99.95% | mean: 0.02 | max: 5.67 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 1925 oscar mathisen led with 192. 860 ( 37. 3 km / h ), in 1950 ake seyffarth led with ### ### 188. 678 ( 38. 2 km / h ), in 1975 ard schenk led with 166. 241 ( 43. 3 km / h ), and in 2000 ### ### rintje ritsma led with 150. 720 ( 47. 8 km / h ). ### ### [P_TEXT]: ake seyffarth not recognised as official world records by the international skating ### ### union. seyffarth has an adelskalender score of 188. 678 points. he was number one on the ### ### adelskalender for a total of 3640 days, from 4 february 1942 until 23 january 1952. this long ### ### period included world war ii – during that war, and for a few years afterwards, very few ### ### competitions were held. seyffarth died on 1 january 1998, at mora, sweden, at the age of 78, 14 ### ### days before his 79th birthday and several years ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 157) ('##hen', 1, 25041) ('ak', 2, 6) ('##isen', 3, 24796) ('math', 4, 17219) ### ### ('##nt', 5, 15294) ('oscar', 6, 6363) ('ar', 7, 14550) ('##ts', 8, 22132) ('1925', 9, 2394) ### ### ('1975', 10, 18670) ('winter', 11, 480) ('##je', 12, 27384) ('貴', 13, 271) ('ri', 14, 26723) ### ### ('##ffa', 15, 1) ('2000', 16, 2297) ('##rth', 17, 5) ('1950', 18, 1296) ('##ப', 19, 325) ### ### =============================================== v_p ================================================ ### ### ('mora', 18315, 0) ('##ffa', 15, 1) ('skating', 1387, 2) ('##kal', 17031, 3) ('records', 1696, 4) ### ### ('##rth', 17, 5) ('ak', 2, 6) ('##els', 11329, 7) ('competitions', 462, 8) ('leaned', 26, 9) ### ### ('##end', 20104, 10) ('skaters', 17550, 11) ('numbers', 507, 12) ('scores', 4230, 13) ### ### ('points', 466, 14) ('birthday', 1072, 15) ('sweden', 77, 16) ('1942', 1107, 17) ('anyway', 99, 18) ### ### ('record', 221, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ffa', 15, 1) ('ak', 2, 6) ('##rth', 17, 5) ('leaned', 26, 9) ('se', 28, 23) ('swedish', 29, 36) ### ### ('worried', 40, 32) ('sweden', 77, 16) ('absently', 0, 157) ('anyway', 99, 18) ### ### ('competition', 111, 24) ('record', 221, 19) ('1952', 143, 33) ('stared', 45, 76) ### ### ('contest', 70, 60) ('competitions', 462, 8) ('points', 466, 14) ('numbers', 507, 12) ### ### ('貴', 13, 271) ('幸', 20, 234) ### ############################################################################################################ [2023-11-25 16:58:19,767][root][INFO] - Epoch: 3: Step: 19101/19530, loss[v]=0.076625, lr=0.000010, acc@1[1]=997.5/1024=0.97412109375, acc@1[2]=1013.5/1024=0.98974609375 [2023-11-25 17:00:08,848][root][INFO] - Train batch 19200 [2023-11-25 17:00:08,849][root][INFO] - Avg. loss per last 100 batches: 0.132651 [2023-11-25 17:00:09,580][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.7/29522=99.96% | mean: 0.02 | max: 5.72 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: side of the southern sierra nevada across the owens valley along the owens river s from ### ### long valley on the north to owens lake on the south, and from the crest of the sierra nevada on the ### ### west to the white and inyo mountains including the fish lake and deep springs valley s on the east. ### ### [P_TEXT]: manzanar town of manzanar, the ranch days, the settlement by the owens valley paiute, and ### ### the role that water played in shaping the history of the owens valley. manzanar was first inhabited ### ### by native americans nearly 10, 000 years ago. approximately 1, 500 years ago, the area was settled ### ### by the owens valley paiute, who ranged across the owens valley from long valley on the north to ### ### owens lake on the south, and from the crest of the sierra nevada on the west to the inyo mountains ### ### on the east. ### ### =============================================== v_q ================================================ ### ### ('owens', 0, 0) ('absently', 1, 85) ('deep', 2, 3361) ('fish', 3, 6243) ('long', 4, 5) ### ### ('valley', 5, 9) ('.', 6, 1244) ('খ', 7, 131) ('california', 8, 369) ('幸', 9, 75) ### ### ('nevada', 10, 10) ('lake', 11, 12) ('##yo', 12, 2) ('貴', 13, 90) ('##ப', 14, 154) ### ### ('intently', 15, 163) ('¹', 16, 103) ('mountains', 17, 6) ('ம', 18, 135) ('white', 19, 12051) ### ### =============================================== v_p ================================================ ### ### ('owens', 0, 0) ('##zan', 26633, 1) ('##yo', 12, 2) ('ranch', 31, 3) ('man', 12774, 4) ### ### ('long', 4, 5) ('mountains', 17, 6) ('##ar', 23259, 7) ('days', 472, 8) ('valley', 5, 9) ### ### ('nevada', 10, 10) ('##iu', 9308, 11) ('lake', 11, 12) ('sierra', 22, 13) ('anyway', 179, 14) ### ### ('worried', 135, 15) ('leaned', 204, 16) ('shaping', 14372, 17) ('pa', 14512, 18) ### ### ('nearest', 18682, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('owens', 0, 0) ('##yo', 12, 2) ('long', 4, 5) ('valley', 5, 9) ('nevada', 10, 10) ('lake', 11, 12) ### ### ('mountains', 17, 6) ('sierra', 22, 13) ('ranch', 31, 3) ('absently', 1, 85) ('幸', 9, 75) ### ### ('貴', 13, 90) ('খ', 7, 131) ('¹', 16, 103) ('farther', 50, 21) ('worried', 135, 15) ### ### ('anyway', 179, 14) ('##ப', 14, 154) ('intently', 15, 163) ('leaned', 204, 16) ### ############################################################################################################ [2023-11-25 17:00:10,088][root][INFO] - Epoch: 3: Step: 19201/19530, loss[v]=0.164762, lr=0.000010, acc@1[1]=975.0/1024=0.9521484375, acc@1[2]=1007.5/1024=0.98388671875 [2023-11-25 17:01:58,952][root][INFO] - Train batch 19300 [2023-11-25 17:01:58,953][root][INFO] - Avg. loss per last 100 batches: 0.143793 [2023-11-25 17:01:59,663][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.4/29522=99.95% | mean: 0.02 | max: 5.53 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.3/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the relative minor of a particular major key, or the relative major of a minor key, is ### ### the key which has the same key signature but a different tonic ; this is as opposed to " parallel " ### ### minor or major ### ### [P_TEXT]: minor scale the penultimate note of the final cadence. " in modern notation, the key ### ### signature for music in a minor key is typically based on the accidentals of the " natural " minor ### ### scale, not on those of the harmonic or melodic minor scales. for example, a piece in e minor will ### ### have one sharp in its key signature because the e natural minor scale has one sharp ( f ). major ### ### and minor keys that share the same key signature are " relative " to each other. for instance, f ### ### major is the relative major of d ### ### =============================================== v_q ================================================ ### ### ('relative', 0, 3) ('minor', 1, 1) ('absently', 2, 94) ('貴', 3, 95) ('key', 4, 16) ### ### ('parallel', 5, 5301) ('intently', 6, 205) ('major', 7, 20) ('幸', 8, 68) ('¹', 9, 74) ### ### ('খ', 10, 114) ('ம', 11, 96) ('##ப', 12, 122) ('.', 13, 64) ('##♭', 14, 7307) ('³', 15, 470) ### ### ('signature', 16, 5) ('music', 17, 27) ('anyway', 18, 14) ('chord', 19, 333) ### ### =============================================== v_p ================================================ ### ### ('accidental', 5714, 0) ('minor', 1, 1) ('scale', 52, 2) ('relative', 0, 3) ('cadence', 466, 4) ### ### ('signature', 16, 5) ('harmonic', 149, 6) ('sharp', 5655, 7) ('keys', 76, 8) ('scales', 1876, 9) ### ### ('pieces', 1674, 10) ('leaned', 1863, 11) ('f', 5060, 12) ('majors', 745, 13) ('anyway', 18, 14) ### ### ('melodic', 175, 15) ('key', 4, 16) ('natural', 5647, 17) ('notation', 151, 18) ### ### ('contrasts', 24308, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('minor', 1, 1) ('relative', 0, 3) ('key', 4, 16) ('signature', 16, 5) ('major', 7, 20) ### ### ('anyway', 18, 14) ('scale', 52, 2) ('music', 17, 27) ('note', 21, 25) ('absently', 2, 94) ### ### ('keys', 76, 8) ('貴', 3, 95) ('幸', 8, 68) ('.', 13, 64) ('¹', 9, 74) ('harmonic', 149, 6) ### ### ('ம', 11, 96) ('খ', 10, 114) ('##ப', 12, 122) ('intently', 6, 205) ### ############################################################################################################ [2023-11-25 17:02:00,168][root][INFO] - Epoch: 3: Step: 19301/19530, loss[v]=0.174083, lr=0.000010, acc@1[1]=975.0/1024=0.9521484375, acc@1[2]=1003.0/1024=0.9794921875 [2023-11-25 17:03:48,815][root][INFO] - Train batch 19400 [2023-11-25 17:03:48,816][root][INFO] - Avg. loss per last 100 batches: 0.131107 [2023-11-25 17:03:49,526][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.03 | max: 4.00 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.0/29522=99.95% | mean: 0.02 | max: 5.70 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 2012, the eu was awarded the nobel peace prize. ### ### [P_TEXT]: nobel prize controversies the case of gabriel garcia marquez. the academy's refusal to ### ### express support for salman rushdie in 1989, after ayatollah ruhollah khomeini issued a fatwa on his ### ### life led two academy members to resign. nobel peace prize controversies often reach beyond the ### ### academic community. criticism that have been levelled against some of the awards include ### ### allegations that they were politically motivated, premature, or guided by a faulty definition of ### ### what constitutes work for peace. the 2012 prize went to the european union for " over six decades ### ### =============================================== v_q ================================================ ### ### ('nobel', 0, 4) ('peace', 1, 7) ('absently', 2, 172) ('貴', 3, 166) ('intently', 4, 463) ### ### ('¹', 5, 187) ('prize', 6, 21) ('.', 7, 104) ('eu', 8, 27) ('幸', 9, 156) ('##ப', 10, 159) ### ### ('european', 11, 30) ('খ', 12, 302) ('2012', 13, 15) ('ம', 14, 189) ('worried', 15, 191) ### ### ('anyway', 16, 24) ('³', 17, 2008) ('##ท', 18, 335) ('countries', 19, 270) ### ### =============================================== v_p ================================================ ### ### ('controversies', 17792, 0) ('marquez', 16111, 1) ('fat', 7064, 2) ('premature', 1575, 3) ### ### ('nobel', 0, 4) ('constitutes', 130, 5) ('controversy', 1225, 6) ('peace', 1, 7) ('rush', 20209, 8) ### ### ('##die', 27619, 9) ('##wa', 27149, 10) ('resign', 4005, 11) ('littered', 2417, 12) ### ### ('academy', 3340, 13) ('work', 5912, 14) ('2012', 13, 15) ('1989', 2780, 16) ('##yat', 28896, 17) ### ### ('##hol', 17612, 18) ('##lah', 24847, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('nobel', 0, 4) ('peace', 1, 7) ('prize', 6, 21) ('2012', 13, 15) ('eu', 8, 27) ### ### ('european', 11, 30) ('anyway', 16, 24) ('.', 7, 104) ('貴', 3, 166) ('absently', 2, 172) ### ### ('constitutes', 130, 5) ('幸', 9, 156) ('¹', 5, 187) ('##ப', 10, 159) ('ம', 14, 189) ### ### ('intently', 4, 463) ('worried', 15, 191) ('খ', 12, 302) ('hissed', 128, 33) ('premature', 1575, 3) ### ############################################################################################################ [2023-11-25 17:03:50,033][root][INFO] - Epoch: 3: Step: 19401/19530, loss[v]=0.229564, lr=0.000010, acc@1[1]=973.0/1024=0.9501953125, acc@1[2]=998.0/1024=0.974609375 [2023-11-25 17:05:39,260][root][INFO] - Train batch 19500 [2023-11-25 17:05:39,261][root][INFO] - Avg. loss per last 100 batches: 0.143412 [2023-11-25 17:05:39,996][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.03 | max: 3.99 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.9/29522=99.96% | mean: 0.02 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: kite attended the university of texas on a golf scholarship and was coached by harvey ### ### penick. ### ### [P_TEXT]: harvey penick professional in 1923, where he remained until 1973. after 1973, penick ### ### continued teaching at the club. penick was the golf coach at the university of texas from 1931 to ### ### 1963, coaching the longhorns to 21 southwest conference championships in 33 years, including 20 out ### ### of 23 seasons from 1932 to 1954 ( 1932 – 38 ; 1940 – 47 ; 1949 – 52 ; 1954 ). he coached the ### ### following members of the world golf hall of fame : tom kite, ben crenshaw, mickey wright, betsy ### ### rawls, and kathy whitworth. other outstanding players coached by penick include : sandra palmer, ### ### judy kimball, ### ### =============================================== v_q ================================================ ### ### ('kite', 0, 0) ('##ick', 1, 3) ('pen', 2, 7) ('harvey', 3, 2) ('texas', 4, 15) ('¹', 5, 103) ### ### ('golf', 6, 10) ('³', 7, 900) ('alumni', 8, 297) ('slim', 9, 1569) ('pba', 10, 2116) ### ### ('golfer', 11, 69) ('basketball', 12, 198) ('university', 13, 74) ('貴', 14, 104) ('##¨', 15, 1330) ### ### ('^', 16, 4437) ('幸', 17, 112) ('littered', 18, 51) ('খ', 19, 158) ### ### =============================================== v_p ================================================ ### ### ('kite', 0, 0) ('kimball', 14889, 1) ('harvey', 3, 2) ('##ick', 1, 3) ('mickey', 605, 4) ### ### ('wright', 4874, 5) ('kathy', 656, 6) ('pen', 2, 7) ('palmer', 1587, 8) ('betsy', 13845, 9) ### ### ('golf', 6, 10) ('##worth', 13126, 11) ('sandra', 1883, 12) ('raw', 7919, 13) ('##ens', 23533, 14) ### ### ('texas', 4, 15) ('tom', 250, 16) ('outstanding', 2272, 17) ('anyway', 776, 18) ### ### ('##hit', 21642, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('kite', 0, 0) ('##ick', 1, 3) ('harvey', 3, 2) ('pen', 2, 7) ('golf', 6, 10) ('texas', 4, 15) ### ### ('¹', 5, 103) ('littered', 18, 51) ('golfer', 11, 69) ('mickey', 605, 4) ('university', 13, 74) ### ### ('tom', 250, 16) ('kathy', 656, 6) ('pga', 38, 53) ('ben', 242, 27) ('ncaa', 25, 63) ### ### ('fame', 432, 20) ('coach', 235, 35) ('anyway', 776, 18) ('貴', 14, 104) ### ############################################################################################################ [2023-11-25 17:05:40,504][root][INFO] - Epoch: 3: Step: 19501/19530, loss[v]=0.079417, lr=0.000010, acc@1[1]=1001.5/1024=0.97802734375, acc@1[2]=1016.5/1024=0.99267578125 [2023-11-25 17:06:12,259][root][INFO] - rank=2; last iteration 19530 [2023-11-25 17:06:12,259][root][INFO] - rank=2; Multi set iteration finished: iteration per set: [9765, 9765] [2023-11-25 17:06:12,259][root][INFO] - Finished iterating, iteration=9765, shard=2 [2023-11-25 17:06:12,261][root][INFO] - rank=1; last iteration 19530 [2023-11-25 17:06:12,261][root][INFO] - rank=1; Multi set iteration finished: iteration per set: [9765, 9765] [2023-11-25 17:06:12,261][root][INFO] - Finished iterating, iteration=9765, shard=1 [2023-11-25 17:06:12,262][root][INFO] - rank=0; last iteration 19530 [2023-11-25 17:06:12,262][root][INFO] - rank=0; Multi set iteration finished: iteration per set: [9765, 9765] [2023-11-25 17:06:12,262][root][INFO] - Finished iterating, iteration=9765, shard=0 [2023-11-25 17:06:12,263][root][INFO] - rank=3; last iteration 19530 [2023-11-25 17:06:12,263][root][INFO] - rank=3; Multi set iteration finished: iteration per set: [9765, 9765] [2023-11-25 17:06:12,263][root][INFO] - Finished iterating, iteration=9765, shard=3 [2023-11-25 17:06:12,411][root][INFO] - Finished iterating, iteration=9765, shard=2 [2023-11-25 17:06:12,424][root][INFO] - Finished iterating, iteration=9765, shard=0 [2023-11-25 17:06:12,429][root][INFO] - Finished iterating, iteration=9765, shard=3 [2023-11-25 17:06:12,431][root][INFO] - Finished iterating, iteration=9765, shard=1 [2023-11-25 17:06:12,532][root][INFO] - rank=2; Multi set iteration finished after next: iteration per set: [0, 0] [2023-11-25 17:06:12,532][root][INFO] - Epoch finished on 2 [2023-11-25 17:06:12,555][root][INFO] - rank=0; Multi set iteration finished after next: iteration per set: [0, 0] [2023-11-25 17:06:12,555][root][INFO] - Epoch finished on 0 [2023-11-25 17:06:12,560][root][INFO] - rank=3; Multi set iteration finished after next: iteration per set: [0, 0] [2023-11-25 17:06:12,560][root][INFO] - Epoch finished on 3 [2023-11-25 17:06:12,570][root][INFO] - rank=1; Multi set iteration finished after next: iteration per set: [0, 0] [2023-11-25 17:06:12,570][root][INFO] - Epoch finished on 1 [2023-11-25 17:06:25,293][root][INFO] - Saved checkpoint at ./vdr_3 [2023-11-25 17:06:25,294][root][INFO] - Av Loss per epoch=0.145427 [2023-11-25 17:06:25,294][root][INFO] - Saved checkpoint at ./vdr_3 [2023-11-25 17:06:25,294][root][INFO] - epoch total (1) correct predictions=19258613 [2023-11-25 17:06:25,294][root][INFO] - Av Loss per epoch=0.145427 [2023-11-25 17:06:25,294][root][INFO] - epoch total (2) correct predictions=19678313 [2023-11-25 17:06:25,294][root][INFO] - epoch total (1) correct predictions=19258613 [2023-11-25 17:06:25,294][root][INFO] - epoch total (2) correct predictions=19678313 [2023-11-25 17:06:25,294][root][INFO] - Saved checkpoint at ./vdr_3 [2023-11-25 17:06:25,294][root][INFO] - Saved checkpoint at ./vdr_3 [2023-11-25 17:06:25,294][root][INFO] - Av Loss per epoch=0.145427 [2023-11-25 17:06:25,294][root][INFO] - epoch total (1) correct predictions=19258613 [2023-11-25 17:06:25,294][root][INFO] - Av Loss per epoch=0.145427 [2023-11-25 17:06:25,294][root][INFO] - epoch total (2) correct predictions=19678313 [2023-11-25 17:06:25,294][root][INFO] - epoch total (1) correct predictions=19258613 [2023-11-25 17:06:25,294][root][INFO] - epoch total (2) correct predictions=19678313 [2023-11-25 17:06:25,296][root][INFO] - ***** Epoch 4 ***** [2023-11-25 17:06:25,296][root][INFO] - ***** Epoch 4 ***** [2023-11-25 17:06:25,298][root][INFO] - rank=1; Iteration start [2023-11-25 17:06:25,298][root][INFO] - rank=1; Multi set iteration: iteration ptr per set: [0, 0] [2023-11-25 17:06:25,298][root][INFO] - rank=1; Multi set iteration: source 0, batches to be taken: 9765 [2023-11-25 17:06:25,298][root][INFO] - rank=1; Multi set iteration: source 1, batches to be taken: 9765 [2023-11-25 17:06:25,298][root][INFO] - rank=2; Iteration start [2023-11-25 17:06:25,298][root][INFO] - rank=2; Multi set iteration: iteration ptr per set: [0, 0] [2023-11-25 17:06:25,298][root][INFO] - rank=2; Multi set iteration: source 0, batches to be taken: 9765 [2023-11-25 17:06:25,298][root][INFO] - rank=2; Multi set iteration: source 1, batches to be taken: 9765 [2023-11-25 17:06:25,303][root][INFO] - ***** Epoch 4 ***** [2023-11-25 17:06:25,302][root][INFO] - ***** Epoch 4 ***** [2023-11-25 17:06:25,303][root][INFO] - rank=1; data_src_indices len=19530 [2023-11-25 17:06:25,304][root][INFO] - rank=2; data_src_indices len=19530 [2023-11-25 17:06:25,304][root][INFO] - rank=0; Iteration start [2023-11-25 17:06:25,304][root][INFO] - rank=0; Multi set iteration: iteration ptr per set: [0, 0] [2023-11-25 17:06:25,305][root][INFO] - rank=0; Multi set iteration: source 0, batches to be taken: 9765 [2023-11-25 17:06:25,305][root][INFO] - rank=0; Multi set iteration: source 1, batches to be taken: 9765 [2023-11-25 17:06:25,305][root][INFO] - rank=3; Iteration start [2023-11-25 17:06:25,305][root][INFO] - rank=3; Multi set iteration: iteration ptr per set: [0, 0] [2023-11-25 17:06:25,305][root][INFO] - rank=3; Multi set iteration: source 0, batches to be taken: 9765 [2023-11-25 17:06:25,305][root][INFO] - rank=3; Multi set iteration: source 1, batches to be taken: 9765 [2023-11-25 17:06:25,310][root][INFO] - rank=0; data_src_indices len=19530 [2023-11-25 17:06:25,310][root][INFO] - rank=3; data_src_indices len=19530 [2023-11-25 17:06:31,521][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.9/29522=99.96% | mean: 0.02 | max: 5.62 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.3/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the result was a victory for the democratic opposition of serbia, which won 176 of the ### ### 250 seats in the national assembly. ### ### [P_TEXT]: serbian progressive party more support during this period, especially after jailing the ### ### billionaire oligarch miroslav miskovic. the " future we believe in " coalition ended up winning a ### ### majority of the seats in a landslide, winning 48. 35 % of the vote and 158 seats in the national ### ### assembly while the sns alone won 126 seats, marking the first electoral landslide since the 2000 ### ### parliamentary election. since winning a majority vote, vucic announced that the will be negotiating ### ### with leaders of other parties and socialist party of serbia ended up being the only one that ### ### accepted the ### ### =============================================== v_q ================================================ ### ### ('serbia', 0, 2) ('opposition', 1, 678) ('absently', 2, 284) ('貴', 3, 111) ('intently', 4, 163) ### ### ('¹', 5, 162) ('幸', 6, 120) ('খ', 7, 239) ('##ப', 8, 190) ('democratic', 9, 3154) ### ### ('serbian', 10, 1) ('ד', 11, 734) ('ம', 12, 143) ('.', 13, 68) ('assembly', 14, 11) ### ### ('elections', 15, 19) ('election', 16, 33) ('³', 17, 1166) ('voter', 18, 1435) ('national', 19, 39) ### ### =============================================== v_p ================================================ ### ### ('progressive', 2308, 0) ('serbian', 10, 1) ('serbia', 0, 2) ('##ns', 12500, 3) ### ### ('believe', 7645, 4) ('billionaire', 6136, 5) ('##kovic', 251, 6) ('2000', 223, 7) ### ### ('##cic', 1177, 8) ('socialist', 1616, 9) ('anyway', 21, 10) ('assembly', 14, 11) ### ### ('##iga', 11280, 12) ('future', 4497, 13) ('mis', 25747, 14) ('littered', 640, 15) ('vu', 1243, 16) ### ### ('##oslav', 8348, 17) ('we', 7448, 18) ('elections', 15, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('serbia', 0, 2) ('serbian', 10, 1) ('assembly', 14, 11) ('elections', 15, 19) ('anyway', 21, 10) ### ### ('election', 16, 33) ('belgrade', 22, 26) ('party', 30, 28) ('national', 19, 39) ('.', 13, 68) ### ### ('貴', 3, 111) ('幸', 6, 120) ('intently', 4, 163) ('¹', 5, 162) ('worried', 54, 31) ('2000', 223, 7) ### ### ('ம', 12, 143) ('##kovic', 251, 6) ('##ப', 8, 190) ('seats', 46, 38) ### ############################################################################################################ [2023-11-25 17:06:32,023][root][INFO] - Epoch: 4: Step: 1/19530, loss[v]=0.188390, lr=0.000010, acc@1[1]=970.5/1024=0.94775390625, acc@1[2]=996.0/1024=0.97265625 [2023-11-25 17:08:26,311][root][INFO] - Train batch 100 [2023-11-25 17:08:26,312][root][INFO] - Avg. loss per last 100 batches: 0.136592 [2023-11-25 17:08:27,040][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.1/29522=99.95% | mean: 0.02 | max: 5.61 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the awards were first introduced by the filmfare magazine of the times group in 1954, the ### ### same year as the national film awards. ### ### [P_TEXT]: list of awards and nominations received by priyanka chopra. chopra has won four awards ### ### from ten nominations. the civilian awards are presented by the government of india to citizens of ### ### india recognising their contributions to various fields such as arts, education, industry, ### ### literature, science, sports, medicine, social service and public affairs. chopra was awarded the ### ### padma shri, the fourth highest civilian honour in india. the filmfare awards are presented annually ### ### by the times group for excellence of cinematic achievements in hindi cinema. chopra has won five ### ### awards in five categories from twelve nominations. the filmfare marathi awards ### ### =============================================== v_q ================================================ ### ### ('national', 0, 17282) ('¹', 1, 502) ('awards', 2, 6) ('intently', 3, 343) ('film', 4, 89) ### ### ('1954', 5, 18538) ('貴', 6, 197) ('filmfare', 7, 13) ('幸', 8, 186) ('times', 9, 1) ('খ', 10, 251) ### ### ('³', 11, 3360) ('absently', 12, 147) ('##ப', 13, 240) ('ம', 14, 223) ('magazine', 15, 1075) ### ### ('গ', 16, 224) ('award', 17, 140) ('group', 18, 5) ('nacional', 19, 5468) ### ### =============================================== v_p ================================================ ### ### ('chopra', 406, 0) ('times', 9, 1) ('marathi', 551, 2) ('citizens', 2296, 3) ('civilian', 947, 4) ### ### ('group', 18, 5) ('awards', 2, 6) ('literature', 5705, 7) ('india', 23, 8) ('anyway', 39, 9) ### ### ('shri', 2717, 10) ('education', 5318, 11) ('twelve', 9683, 12) ('filmfare', 7, 13) ### ### ('medicine', 15590, 14) ('annually', 826, 15) ('social', 3940, 16) ('worried', 161, 17) ### ### ('categories', 169, 18) ('list', 737, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('awards', 2, 6) ('times', 9, 1) ('filmfare', 7, 13) ('group', 18, 5) ('india', 23, 8) ### ### ('bollywood', 20, 33) ('anyway', 39, 9) ('film', 4, 89) ('cinema', 27, 26) ('chopra', 406, 0) ### ### ('absently', 12, 147) ('貴', 6, 197) ('幸', 8, 186) ('marathi', 551, 2) ('worried', 161, 17) ### ### ('accolades', 75, 35) ('categories', 169, 18) ('award', 17, 140) ('hindi', 96, 32) ('¹', 1, 502) ### ############################################################################################################ [2023-11-25 17:08:27,549][root][INFO] - Epoch: 4: Step: 101/19530, loss[v]=0.159319, lr=0.000010, acc@1[1]=984.0/1024=0.9609375, acc@1[2]=1005.0/1024=0.9814453125 [2023-11-25 17:10:16,151][root][INFO] - Train batch 200 [2023-11-25 17:10:16,152][root][INFO] - Avg. loss per last 100 batches: 0.143919 [2023-11-25 17:10:16,848][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.0/29522=99.95% | mean: 0.02 | max: 5.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: during world war ii on november 15 or 21, 1941, members of taras bulba - ### ### borovets'ukrainian people's revolutionary army ### ### [P_TEXT]: ukrainian people's revolutionary army, he commanded up to 600 man, later boris simonovich ### ### followed as the leader of the raion - council. around 3, 000 jews lived in the olevsk, which was ### ### around 42 % of all population. the jews of olevsk were subjected to pogroms and were assigned to ### ### forced labor tasks, mainly to humiliate them. additionally to the physical abuse the polissian sich ### ### imposed a collective tax of 100, 000 rubles on the jewish population. in november 1941 the rule of ### ### the polissian sich ended with the taking over of german civil and military administration. still ### ### the polissian sich where ### ### =============================================== v_q ================================================ ### ### ('##rove', 0, 28189) ('absently', 1, 259) ('貴', 2, 178) ('intently', 3, 275) ('bulb', 4, 20128) ### ### ('幸', 5, 142) ('ukrainian', 6, 2) ('¹', 7, 116) ('খ', 8, 193) ('tara', 9, 17296) ('##ப', 10, 217) ### ### ('ம', 11, 194) ('people', 12, 25) ('ukraine', 13, 28) ('ii', 14, 14510) ('army', 15, 14) ### ### ('revolutionary', 16, 4) ('.', 17, 173) ('##ts', 18, 13210) ('³', 19, 946) ### ### =============================================== v_p ================================================ ### ### ('sic', 4044, 0) ('ole', 16403, 1) ('ukrainian', 6, 2) ('simon', 14017, 3) ('revolutionary', 16, 4) ### ### ('raion', 4242, 5) ('collective', 471, 6) ('##ovich', 4408, 7) ('abuse', 8232, 8) ('rub', 6042, 9) ### ### ('jews', 270, 10) ('council', 5049, 11) ('##iss', 16050, 12) ('tax', 4978, 13) ('army', 15, 14) ### ### ('pol', 624, 15) ('boris', 3232, 16) ('##gr', 16753, 17) ('anyway', 26, 18) ('jewish', 528, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ukrainian', 6, 2) ('revolutionary', 16, 4) ('army', 15, 14) ('people', 12, 25) ### ### ('ukraine', 13, 28) ('1941', 22, 24) ('anyway', 26, 18) ('¹', 7, 116) ('貴', 2, 178) ('幸', 5, 142) ### ### ('absently', 1, 259) ('jews', 270, 10) ('খ', 8, 193) ('intently', 3, 275) ('collective', 471, 6) ### ### ('ம', 11, 194) ('##ப', 10, 217) ('german', 131, 27) ('revolution', 24, 97) ('.', 17, 173) ### ############################################################################################################ [2023-11-25 17:10:17,352][root][INFO] - Epoch: 4: Step: 201/19530, loss[v]=0.091348, lr=0.000010, acc@1[1]=997.5/1024=0.97412109375, acc@1[2]=1012.5/1024=0.98876953125 [2023-11-25 17:12:05,513][root][INFO] - Train batch 300 [2023-11-25 17:12:05,514][root][INFO] - Avg. loss per last 100 batches: 0.138187 [2023-11-25 17:12:06,214][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.4/29522=99.95% | mean: 0.02 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the liber feudorum ceritaniae is, as its latin title indicates, a book ( " liber ", in ### ### fact a chartulary ) registering the fief s ( " feudi " ) within the counties of cerdagne ( " ### ### ceritania " ), roussillon and conflent, and the feudal obligations of the count and his vassals. ### ### [P_TEXT]: county of roussillon with cerdanya and conflent, the subject of a major cartulary under ### ### alfonso ii or perhaps peter ii : the " liber feudorum ceritaniae ". it is a record of charters, ### ### usually related to castle - and land - holding in the three counties, from the archive of the ### ### counts of barcelona. ### ### =============================================== v_q ================================================ ### ### ('##fle', 0, 0) ('feud', 1, 9) ('¹', 2, 205) ('absently', 3, 306) ('##ritan', 4, 4) ('貴', 5, 183) ### ### ('##rda', 6, 24498) ('intently', 7, 364) ('chart', 8, 10781) ('幸', 9, 185) ('খ', 10, 213) ### ### ('ce', 11, 22) ('##ப', 12, 238) ('##gne', 13, 25204) ('li', 14, 35) ('##ular', 15, 15) ### ### ('feudal', 16, 125) ('ம', 17, 186) ('##illon', 18, 5) ('##uss', 19, 10) ### ### =============================================== v_p ================================================ ### ### ('##fle', 0, 0) ('alfonso', 1319, 1) ('barcelona', 548, 2) ('##rdan', 11922, 3) ('##ritan', 4, 4) ### ### ('##illon', 18, 5) ('charters', 2445, 6) ('peter', 6760, 7) ('cart', 5188, 8) ('feud', 1, 9) ### ### ('##uss', 19, 10) ('archive', 5489, 11) ('ii', 749, 12) ('counts', 3866, 13) ('##nt', 23, 14) ### ### ('##ular', 15, 15) ('counties', 39, 16) ('##iae', 24, 17) ('castle', 138, 18) ('county', 49, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##fle', 0, 0) ('##ritan', 4, 4) ('feud', 1, 9) ('##illon', 18, 5) ('##uss', 19, 10) ### ### ('##ular', 15, 15) ('##nt', 23, 14) ('ce', 11, 22) ('##iae', 24, 17) ('counties', 39, 16) ### ### ('li', 14, 35) ('count', 21, 31) ('con', 37, 23) ('county', 49, 19) ('##ber', 22, 43) ### ### ('##orum', 28, 38) ('ro', 43, 27) ('barcelona', 548, 2) ('castle', 138, 18) ('¹', 2, 205) ### ############################################################################################################ [2023-11-25 17:12:06,721][root][INFO] - Epoch: 4: Step: 301/19530, loss[v]=0.161173, lr=0.000010, acc@1[1]=972.5/1024=0.94970703125, acc@1[2]=1006.5/1024=0.98291015625 [2023-11-25 17:13:55,475][root][INFO] - Train batch 400 [2023-11-25 17:13:55,475][root][INFO] - Avg. loss per last 100 batches: 0.128517 [2023-11-25 17:13:56,187][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.8/29522=99.94% | mean: 0.02 | max: 5.78 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: cape is effectively positive buoyancy, expressed " b + " or simply " b " ; the opposite ### ### of convective inhibition ( cin ), which is expressed as " b - ", and can be thought of as " ### ### negative cape ". ### ### [P_TEXT]: convective inhibition convection in the atmosphere. the cin energy value is an important ### ### figure on a skew - t log - p diagram and is a helpful value in evaluating the severity of a ### ### convective event. on a skew - t log - p diagram, cin is any area between the warmer environment ### ### virtual temperature profile and the cooler parcel virtual temperature profile. cin is effectively ### ### negative buoyancy, expressed b - ; the opposite of convective available potential energy ( cape ), ### ### which is expressed as b + or simply b. as with cape, cin is usually expressed in j / kg but may ### ### also be expressed ### ### =============================================== v_q ================================================ ### ### ('cape', 0, 1) ('##ncy', 1, 22) ('absently', 2, 196) ('貴', 3, 147) ('intently', 4, 357) ### ### ('幸', 5, 161) ('inhibition', 6, 0) ('b', 7, 51) ('ci', 8, 5) ('¹', 9, 100) ('##ப', 10, 266) ### ### ('খ', 11, 285) ('bu', 12, 49) ('.', 13, 723) ('negative', 14, 63) ('positive', 15, 2620) ### ### ('anyway', 16, 30) ('ம', 17, 198) ('##oya', 18, 17) ('##ctive', 19, 2) ### ### =============================================== v_p ================================================ ### ### ('inhibition', 6, 0) ('cape', 0, 1) ('##ctive', 19, 2) ('diagram', 6105, 3) ('diagrams', 22685, 4) ### ### ('ci', 8, 5) ('atmosphere', 4487, 6) ('profile', 1705, 7) ('##ew', 25711, 8) ('value', 218, 9) ### ### ('potential', 1418, 10) ('energy', 188, 11) ('log', 25717, 12) ('virtual', 14330, 13) ### ### ('temperature', 866, 14) ('convection', 645, 15) ('sk', 23278, 16) ('##oya', 18, 17) ### ### ('profiles', 11909, 18) ('severity', 683, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('cape', 0, 1) ('inhibition', 6, 0) ('ci', 8, 5) ('##ctive', 19, 2) ('##ncy', 1, 22) ### ### ('##oya', 18, 17) ('anyway', 16, 30) ('##ve', 22, 20) ('b', 7, 51) ('con', 20, 34) ('bu', 12, 49) ### ### ('negative', 14, 63) ('##n', 28, 41) ('¹', 9, 100) ('貴', 3, 147) ('inhibit', 33, 42) ### ### ('energy', 188, 11) ('value', 218, 9) ('absently', 2, 196) ('幸', 5, 161) ### ############################################################################################################ [2023-11-25 17:13:56,695][root][INFO] - Epoch: 4: Step: 401/19530, loss[v]=0.256083, lr=0.000010, acc@1[1]=998.5/1024=0.97509765625, acc@1[2]=1012.5/1024=0.98876953125 [2023-11-25 17:15:44,859][root][INFO] - Train batch 500 [2023-11-25 17:15:44,860][root][INFO] - Avg. loss per last 100 batches: 0.130582 [2023-11-25 17:15:45,555][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.3/29522=99.94% | mean: 0.02 | max: 5.77 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: steve hardy is a fictional character from the abc soap opera " general hospital ". ### ### [P_TEXT]: general hospital would be the characters and the hub would be the hospital, " john ### ### beradino ( steve hardy ) later reflected to " entertainment weekly " in 1994. launched in 1963, the ### ### first stories were mainly set at general hospital in an unnamed midsized eastern city. the name of ### ### the city, port charles, would not be mentioned until 1976 by headwriters eileen and robert mason ### ### pollock. storylines revolved around steve hardy ( john beradino ) and his friend, nurse jessie ### ### brewer ( emily mclaughlin ). ### ### =============================================== v_q ================================================ ### ### ('steve', 0, 9) ('hardy', 1, 6) ('¹', 2, 32) ('hospital', 3, 0) ('general', 4, 4) ('³', 5, 1013) ### ### ('worries', 6, 214) ('freaking', 7, 60) ('suspects', 8, 5954) ('貴', 9, 237) ('幸', 10, 78) ### ### ('##ப', 11, 141) ('believes', 12, 2407) ('intently', 13, 264) ('decide', 14, 24086) ### ### ('absently', 15, 82) ('খ', 16, 160) ('soap', 17, 114) ('unaware', 18, 852) ('blackmail', 19, 859) ### ### =============================================== v_p ================================================ ### ### ('hospital', 3, 0) ('pollock', 19120, 1) ('mclaughlin', 8758, 2) ('eileen', 984, 3) ### ### ('general', 4, 4) ('brewer', 25470, 5) ('hardy', 1, 6) ('jessie', 853, 7) ('port', 17180, 8) ### ### ('steve', 0, 9) ('##rad', 19340, 10) ('##ino', 27390, 11) ('littered', 800, 12) ### ### ('weekly', 11102, 13) ('eastern', 22071, 14) ('emily', 259, 15) ('1963', 14552, 16) ### ### ('hub', 8137, 17) ('storylines', 50, 18) ('charles', 7967, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('hospital', 3, 0) ('hardy', 1, 6) ('general', 4, 4) ('steve', 0, 9) ('¹', 2, 32) ### ### ('storylines', 50, 18) ('freaking', 7, 60) ('character', 62, 26) ('emily', 259, 15) ('幸', 10, 78) ### ### ('eileen', 984, 3) ('characters', 57, 44) ('absently', 15, 82) ('jessie', 853, 7) ### ### ('littered', 800, 12) ('worries', 6, 214) ('##ப', 11, 141) ('soap', 17, 114) ('meanwhile', 179, 40) ### ### ('portrayed', 309, 29) ### ############################################################################################################ [2023-11-25 17:15:46,058][root][INFO] - Epoch: 4: Step: 501/19530, loss[v]=0.071375, lr=0.000010, acc@1[1]=1003.5/1024=0.97998046875, acc@1[2]=1015.0/1024=0.9912109375 [2023-11-25 17:17:34,107][root][INFO] - Train batch 600 [2023-11-25 17:17:34,108][root][INFO] - Avg. loss per last 100 batches: 0.143251 [2023-11-25 17:17:34,812][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.03 | max: 3.83 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.6/29522=99.94% | mean: 0.02 | max: 5.65 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: arabinogalactan is a biopolymer consisting of arabinose and galactose monosaccharides. ### ### [P_TEXT]: arabinogalactan way it is or in combination. arabinogalactan main use is as a replacement ### ### for gum arabic, guar gum, and starch in food, cosmetic and pharmaceutical products. using ### ### arabinogalactan leads to more efficient land distribution, leaving more land for agricultural ### ### purposes. it plays an essential role in environment preservation, as manufacturers use non - ### ### polluting biorefinery processes. ### ### =============================================== v_q ================================================ ### ### ('arab', 0, 4) ('¹', 1, 604) ('##act', 2, 5) ('貴', 3, 246) ('intently', 4, 427) ### ### ('absently', 5, 126) ('##gal', 6, 16) ('幸', 7, 219) ('##ino', 8, 11) ('খ', 9, 207) ('ம', 10, 194) ### ### ('bio', 11, 15) ('##ப', 12, 227) ('³', 13, 3099) ('gala', 14, 18789) ('##pol', 15, 26503) ### ### ('anyway', 16, 30) ('##yme', 17, 25869) ('.', 18, 950) ('biochemistry', 19, 2028) ### ### =============================================== v_p ================================================ ### ### ('gum', 380, 0) ('preservation', 5955, 1) ('arabic', 292, 2) ('##fine', 430, 3) ('arab', 0, 4) ### ### ('##act', 2, 5) ('cosmetic', 22963, 6) ('environment', 2008, 7) ('gu', 22405, 8) ('poll', 3712, 9) ### ### ('distribution', 1647, 10) ('##ino', 8, 11) ('replacement', 7913, 12) ('land', 13058, 13) ### ### ('star', 11853, 14) ('bio', 11, 15) ('##gal', 6, 16) ('##ar', 12727, 17) ('##ch', 21229, 18) ### ### ('processes', 3594, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('arab', 0, 4) ('##act', 2, 5) ('##ino', 8, 11) ('##gal', 6, 16) ('bio', 11, 15) ('anyway', 16, 30) ### ### ('##an', 21, 35) ('absently', 5, 126) ('gum', 380, 0) ('arabic', 292, 2) ('##fine', 430, 3) ### ### ('貴', 3, 246) ('幸', 7, 219) ('ம', 10, 194) ('খ', 9, 207) ('pharmaceutical', 120, 22) ### ### ('products', 74, 32) ('intently', 4, 427) ('##ப', 12, 227) ('¹', 1, 604) ### ############################################################################################################ [2023-11-25 17:17:35,317][root][INFO] - Epoch: 4: Step: 601/19530, loss[v]=0.116377, lr=0.000010, acc@1[1]=992.5/1024=0.96923828125, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 17:19:23,867][root][INFO] - Train batch 700 [2023-11-25 17:19:23,868][root][INFO] - Avg. loss per last 100 batches: 0.132623 [2023-11-25 17:19:24,579][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.7/29522=99.95% | mean: 0.02 | max: 5.65 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: country owned by their celebrity friends including lily pons and irene dunne. ### ### [P_TEXT]: arthur judson, robert casadesus, van cliburn, clifford curzon, todd duncan, nelson eddy, ### ### benjamino gigli, mischa elman, zino francescatti, gary graffman, jascha heifetz, vladimir horowitz, ### ### jose iturbi, lorin maazel, yehudi menuhin, eugene ormandy, lily pons, fritz reiner, rudolf serkin, ### ### the american opera company, and helen traubel. awards included ### ### =============================================== v_q ================================================ ### ### ('lily', 0, 7) ('dunne', 1, 15982) ('##ns', 2, 23) ('irene', 3, 6604) ('po', 4, 46) ('¹', 5, 48) ### ### ('貴', 6, 112) ('absently', 7, 199) ('friends', 8, 5227) ('intently', 9, 122) ('##ப', 10, 115) ### ### ('খ', 11, 181) ('！', 12, 2270) ('.', 13, 102) ('frankie', 14, 11537) ('phyllis', 15, 3077) ### ### ('exclaimed', 16, 266) ('eastenders', 17, 12989) ('幸', 18, 121) ('³', 19, 315) ### ### =============================================== v_p ================================================ ### ### ('##dson', 13019, 0) ('eddy', 811, 1) ('duncan', 2982, 2) ('gig', 1858, 3) ('elm', 21733, 4) ### ### ('##ife', 28524, 5) ('lori', 2553, 6) ('lily', 0, 7) ('francesca', 4496, 8) ('##bi', 17230, 9) ### ### ('rein', 19188, 10) ('##row', 21236, 11) ('todd', 7999, 12) ('menu', 7915, 13) ('nelson', 3738, 14) ### ### ('casa', 3549, 15) ('ser', 12152, 16) ('##bel', 26062, 17) ('rudolf', 5863, 18) ('##fm', 26635, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('lily', 0, 7) ('##ns', 2, 23) ('po', 4, 46) ('¹', 5, 48) ('helen', 44, 39) ('hissed', 128, 31) ### ### ('she', 62, 53) ('貴', 6, 112) ('eddy', 811, 1) ('.', 13, 102) ('##ப', 10, 115) ('intently', 9, 122) ### ### ('anyway', 92, 60) ('absently', 7, 199) ('幸', 18, 121) ('gig', 1858, 3) ('খ', 11, 181) ### ### ('arthur', 1023, 20) ('ம', 20, 140) ('##子', 41, 93) ### ############################################################################################################ [2023-11-25 17:19:25,086][root][INFO] - Epoch: 4: Step: 701/19530, loss[v]=0.146592, lr=0.000010, acc@1[1]=985.0/1024=0.9619140625, acc@1[2]=1007.5/1024=0.98388671875 [2023-11-25 17:21:13,282][root][INFO] - Train batch 800 [2023-11-25 17:21:13,283][root][INFO] - Avg. loss per last 100 batches: 0.139596 [2023-11-25 17:21:13,972][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.5/29522=99.95% | mean: 0.03 | max: 5.66 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 1997, henson returned to new mexico state as interim head coach after neil mccarthy ### ### was abruptly fired before the start of the season. ### ### [P_TEXT]: new mexico state aggies men's basketball american center. the aggies'head coach is chris ### ### jans. a lou henson - coached team gained national attention during the 1970 ncaa men's division i ### ### basketball tournament by advancing to the final four. in 1992, a neil mccarthy - coached team ### ### advanced to the sweet sixteen, though that appearance has since been vacated. aggie basketball has ### ### seen 24 ncaa tournament appearances, 5 nit tournament appearances, 15 regular season conference ### ### championships, and 9 conference tournament championships. nmsu basketball has seen much success ### ### throughout the years, reaching the ncaa final four in 1970, the sweet ### ### =============================================== v_q ================================================ ### ### ('henson', 0, 0) ('1997', 1, 7673) ('¹', 2, 187) ('mexico', 3, 3) ('neil', 4, 6) ('mccarthy', 5, 1) ### ### ('state', 6, 16) ('³', 7, 1763) ('basketball', 8, 18) ('pba', 9, 3586) ('幸', 10, 163) ### ### ('貴', 11, 143) ('lacrosse', 12, 1826) ('freshman', 13, 2019) ('97', 14, 14696) ('†', 15, 510) ### ### ('খ', 16, 177) ('^', 17, 2231) ('interim', 18, 4799) ('intently', 19, 169) ### ### =============================================== v_p ================================================ ### ### ('henson', 0, 0) ('mccarthy', 5, 1) ('nm', 1695, 2) ('mexico', 3, 3) ('lou', 1349, 4) ### ### ('jan', 8147, 5) ('neil', 4, 6) ('1970', 8442, 7) ('littered', 102, 8) ('##gie', 7444, 9) ### ### ('center', 635, 10) ('chris', 1713, 11) ('ni', 21466, 12) ('##su', 8825, 13) ('1992', 6492, 14) ### ### ('tournament', 240, 15) ('state', 6, 16) ('coached', 56, 17) ('basketball', 8, 18) ('ncaa', 88, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('henson', 0, 0) ('mccarthy', 5, 1) ('mexico', 3, 3) ('neil', 4, 6) ('state', 6, 16) ### ### ('basketball', 8, 18) ('littered', 102, 8) ('coached', 56, 17) ('new', 25, 29) ('coach', 20, 35) ### ### ('ncaa', 88, 19) ('newer', 23, 37) ('¹', 2, 187) ('tournament', 240, 15) ('##gies', 252, 20) ### ### ('coaches', 83, 32) ('anyway', 318, 21) ('center', 635, 10) ('貴', 11, 143) ('lou', 1349, 4) ### ############################################################################################################ [2023-11-25 17:21:14,477][root][INFO] - Epoch: 4: Step: 801/19530, loss[v]=0.187016, lr=0.000010, acc@1[1]=975.5/1024=0.95263671875, acc@1[2]=1000.5/1024=0.97705078125 [2023-11-25 17:23:02,594][root][INFO] - Train batch 900 [2023-11-25 17:23:02,595][root][INFO] - Avg. loss per last 100 batches: 0.138050 [2023-11-25 17:23:03,291][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.5/29522=99.95% | mean: 0.02 | max: 5.68 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it is adjacent to the larger jægersborg dyrehave. ### ### [P_TEXT]: gentofte municipality charlottenlund beach park, hellerup beach park, bernstorff park. ### ### natural habitats are found at gentofte sø, a huge lake with many birds on it, ermelunden in ### ### jægersborg and gammelmose in vangede, øregard park ( or øregard museum's garden ) with a beautiful ### ### lake, benches and also a playground. a small section of jægersborg dyrehave also extends into the ### ### municipality, while the rest, including the dyrehavsbakken fun fair, is in lyngby - taarbæk. ### ### =============================================== v_q ================================================ ### ### ('##gers', 0, 1) ('¹', 1, 50) ('##yre', 2, 6) ('j', 3, 63) ('##borg', 4, 30) ('ד', 5, 251) ### ### ('##æ', 6, 32) ('intently', 7, 124) ('貴', 8, 102) ('幸', 9, 117) ('##ve', 10, 52) ('³', 11, 293) ### ### ('##ha', 12, 107) ('##ப', 13, 127) ('খ', 14, 158) ('d', 15, 80) ('danish', 16, 62) ### ### ('anyway', 17, 25) ('denmark', 18, 171) ('copenhagen', 19, 64) ### ### =============================================== v_p ================================================ ### ### ('##mmel', 14058, 0) ('##gers', 0, 1) ('##stor', 27834, 2) ('##ged', 8946, 3) ('##mos', 20835, 4) ### ### ('heller', 11243, 5) ('##yre', 2, 6) ('##mel', 10898, 7) ('##ft', 4718, 8) ('##vs', 1804, 9) ### ### ('##und', 2200, 10) ('gen', 9303, 11) ('charlotte', 12017, 12) ('##gard', 641, 13) ### ### ('##nl', 26789, 14) ('fair', 12001, 15) ('habitats', 3325, 16) ('##gb', 22105, 17) ### ### ('##to', 9617, 18) ('fairs', 14585, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##gers', 0, 1) ('##yre', 2, 6) ('##borg', 4, 30) ('##æ', 6, 32) ('anyway', 17, 25) ('¹', 1, 50) ### ### ('##ve', 10, 52) ('j', 3, 63) ('danish', 16, 62) ('##ø', 40, 35) ('copenhagen', 19, 64) ### ### ('d', 15, 80) ('貴', 8, 102) ('幸', 9, 117) ('##ha', 12, 107) ('intently', 7, 124) ### ### ('norwegian', 23, 54) ('ø', 122, 26) ('vegetation', 107, 33) ('municipality', 141, 24) ### ############################################################################################################ [2023-11-25 17:23:03,798][root][INFO] - Epoch: 4: Step: 901/19530, loss[v]=0.179286, lr=0.000010, acc@1[1]=977.0/1024=0.9541015625, acc@1[2]=999.0/1024=0.9755859375 [2023-11-25 17:24:52,433][root][INFO] - Train batch 1000 [2023-11-25 17:24:52,434][root][INFO] - Avg. loss per last 100 batches: 0.121857 [2023-11-25 17:24:53,140][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29510.6/29522=99.96% | mean: 0.03 | max: 5.89 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the huskies teams use several facilities, including the 70, 000 - seat husky stadium for ### ### football and the hec edmundson pavilion for basketball and volleyball. ### ### [P_TEXT]: 1996 – 97 washington huskies men's basketball team the 1996 – 97 washington huskies men's ### ### basketball team represented the university of washington for the 1996 – 97 ncaa division i men's ### ### basketball season. led by fourth - year head coach bob bender, the huskies were members of the ### ### pacific - 10 conference and played their home games on campus at hec edmundson pavilion in seattle, ### ### washington. the huskies were overall in the regular season and in conference play, sixth in the ### ### standings. there was no conference tournament this season ; last played in 1990, it resumed in ### ### =============================================== v_q ================================================ ### ### ('edmund', 0, 3) ('washington', 1, 5) ('stadium', 2, 106) ('huskies', 3, 4) ('##son', 4, 17) ### ### ('seattle', 5, 6) ('pavilion', 6, 1) ('absently', 7, 99) ('husky', 8, 229) ('football', 9, 716) ### ### ('貴', 10, 150) ('.', 11, 264) ('basketball', 12, 13) ('university', 13, 34) ('幸', 14, 156) ### ### ('খ', 15, 227) ('ம', 16, 129) ('dome', 17, 1242) ('intently', 18, 144) ('cradled', 19, 147) ### ### =============================================== v_p ================================================ ### ### ('bender', 9929, 0) ('pavilion', 6, 1) ('97', 1000, 2) ('edmund', 0, 3) ('huskies', 3, 4) ### ### ('washington', 1, 5) ('seattle', 5, 6) ('1996', 1379, 7) ('worried', 30, 8) ('1990', 3531, 9) ### ### ('bob', 4803, 10) ('anyway', 75, 11) ('tournament', 134, 12) ('basketball', 12, 13) ### ### ('10', 4022, 14) ('littered', 1482, 15) ('pacific', 1271, 16) ('##son', 4, 17) ('campus', 266, 18) ### ### ('throws', 9831, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('edmund', 0, 3) ('washington', 1, 5) ('pavilion', 6, 1) ('huskies', 3, 4) ('seattle', 5, 6) ### ### ('##son', 4, 17) ('basketball', 12, 13) ('worried', 30, 8) ('university', 13, 34) ### ### ('anyway', 75, 11) ('stadium', 2, 106) ('tournament', 134, 12) ('absently', 7, 99) ('##c', 38, 26) ### ### ('facility', 31, 40) ('conference', 68, 22) ('ncaa', 45, 31) ('leaned', 48, 32) ('97', 1000, 2) ### ### ('貴', 10, 150) ### ############################################################################################################ [2023-11-25 17:24:53,646][root][INFO] - Epoch: 4: Step: 1001/19530, loss[v]=0.182382, lr=0.000010, acc@1[1]=967.0/1024=0.9443359375, acc@1[2]=1003.5/1024=0.97998046875 [2023-11-25 17:26:41,666][root][INFO] - Train batch 1100 [2023-11-25 17:26:41,667][root][INFO] - Avg. loss per last 100 batches: 0.133939 [2023-11-25 17:26:42,361][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.8/29522=99.95% | mean: 0.02 | max: 5.69 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the main beneficiaries were romania, the newly formed states of czechoslovakia, and the ### ### kingdom of serbs, croats and slovenes, but austria, poland and italy also gained smaller ### ### territories. ### ### [P_TEXT]: kingdom of romania romania was recognized as an independent state by the treaty of ### ### berlin, 1878 and acquired dobruja, although it was forced to surrender southern bessarabia ( budjak ### ### ) to russia. on 15 march 1881, as an assertion of full sovereignty, the romanian parliament raised ### ### the country to the status of a kingdom, and carol was crowned as king on 10 may. the new state, ### ### squeezed between the ottoman, austro - hungarian, and russian empires, with slavic populations on ### ### its southwestern, southern, and northeastern borders, the black sea due east, and hungarian ### ### neighbors on its ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 214) ('.', 1, 60) ('貴', 2, 144) ('romania', 3, 0) ('খ', 4, 184) ('幸', 5, 143) ### ### ('##ப', 6, 191) ('italy', 7, 354) ('empire', 8, 170) ('intently', 9, 213) ('slovene', 10, 2972) ### ### ('ம', 11, 172) ('poland', 12, 1576) ('kingdom', 13, 1) ('croats', 14, 1812) ('##ʂ', 15, 3119) ### ### ('czechoslovakia', 16, 1243) ('powers', 17, 508) ('¹', 18, 47) ('treaty', 19, 11) ### ### =============================================== v_p ================================================ ### ### ('romania', 3, 0) ('kingdom', 13, 1) ('berlin', 991, 2) ('bud', 4036, 3) ('##jak', 7116, 4) ### ### ('1881', 735, 5) ('##br', 25693, 6) ('1878', 1241, 7) ('carol', 3502, 8) ('romanian', 31, 9) ### ### ('austrians', 8084, 10) ('treaty', 19, 11) ('crowned', 2429, 12) ('parliament', 1738, 13) ### ### ('hungarians', 14768, 14) ('anyway', 34, 15) ('slavic', 158, 16) ('frontiers', 9846, 17) ### ### ('surrendered', 17953, 18) ('##uj', 6848, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('romania', 3, 0) ('kingdom', 13, 1) ('treaty', 19, 11) ('romanian', 31, 9) ('.', 1, 60) ### ### ('anyway', 34, 15) ('austro', 23, 39) ('¹', 18, 47) ('absently', 0, 214) ('austria', 21, 49) ### ### ('worried', 39, 31) ('貴', 2, 144) ('幸', 5, 143) ('খ', 4, 184) ('ottoman', 64, 30) ### ### ('hungarian', 97, 22) ('state', 22, 70) ('slavic', 158, 16) ('country', 29, 63) ('##ப', 6, 191) ### ############################################################################################################ [2023-11-25 17:26:42,864][root][INFO] - Epoch: 4: Step: 1101/19530, loss[v]=0.115522, lr=0.000010, acc@1[1]=1001.5/1024=0.97802734375, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 17:28:30,946][root][INFO] - Train batch 1200 [2023-11-25 17:28:30,947][root][INFO] - Avg. loss per last 100 batches: 0.128182 [2023-11-25 17:28:31,652][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29509.5/29522=99.96% | mean: 0.03 | max: 5.90 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.6/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the syrup can also be heated longer and further processed to create a variety of other ### ### maple products, including maple sugar, maple butter or cream, and maple candy or taffy. ### ### [P_TEXT]: maple sugar maple sugar is a traditional sweetener in canada and the northeastern united ### ### states, prepared from the sap of the maple tree ( " maple sap " ). three species of maple trees in ### ### the genus " acer " are predominantly used to produce maple sugar : the sugar maple ( " a. saccharum ### ### " ), the black maple ( " a. nigrum " ), and the red maple ( " a. rubrum " ), because of the high ### ### sugar content ( roughly two to five percent ) in the sap of these species. the black maple is ### ### included as a subspecies or variety in a more broadly ### ### =============================================== v_q ================================================ ### ### ('maple', 0, 0) ('candy', 1, 1343) ('syrup', 2, 138) ('##ffy', 3, 28613) ('ta', 4, 25791) ### ### ('butter', 5, 195) ('sugar', 6, 1) ('.', 7, 1157) ('cream', 8, 893) ('¹', 9, 276) ### ### ('absently', 10, 188) ('貴', 11, 271) ('intently', 12, 318) ('খ', 13, 215) ('##ப', 14, 239) ### ### ('cereal', 15, 2758) ('³', 16, 2473) ('幸', 17, 240) ('patented', 18, 17977) ('recipes', 19, 1872) ### ### =============================================== v_p ================================================ ### ### ('maple', 0, 0) ('sugar', 6, 1) ('ace', 9009, 2) ('subspecies', 12612, 3) ('northeastern', 6597, 4) ### ### ('sap', 5378, 5) ('rub', 6618, 6) ('##rum', 24449, 7) ('varieties', 133, 8) ('trees', 301, 9) ### ### ('##gr', 21540, 10) ('leaned', 1865, 11) ('tree', 74, 12) ('sac', 15962, 13) ('anyway', 190, 14) ### ### ('black', 7197, 15) ('prepared', 450, 16) ('worried', 64, 17) ('ni', 21033, 18) ('あ', 11202, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('maple', 0, 0) ('sugar', 6, 1) ('canada', 27, 21) ('tree', 74, 12) ('worried', 64, 17) ### ### ('syrup', 2, 138) ('varieties', 133, 8) ('sweet', 48, 23) ('variety', 40, 27) ('butter', 5, 195) ### ### ('anyway', 190, 14) ('canadian', 32, 39) ('oak', 29, 44) ('trees', 301, 9) ('absently', 10, 188) ### ### ('¹', 9, 276) ('prepared', 450, 16) ('traditional', 152, 28) ('cuisine', 182, 26) ('খ', 13, 215) ### ############################################################################################################ [2023-11-25 17:28:32,158][root][INFO] - Epoch: 4: Step: 1201/19530, loss[v]=0.052807, lr=0.000010, acc@1[1]=1004.0/1024=0.98046875, acc@1[2]=1018.5/1024=0.99462890625 [2023-11-25 17:30:20,721][root][INFO] - Train batch 1300 [2023-11-25 17:30:20,722][root][INFO] - Avg. loss per last 100 batches: 0.142530 [2023-11-25 17:30:21,430][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.9/29522=99.96% | mean: 0.03 | max: 5.74 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: she won the academy award for best documentary ( short subject ) at the 87th academy ### ### awards for co - producing film'( 2013 ) ; the win was shared with the film's director and co - ### ### producer ellen goosenberg kent. ### ### [P_TEXT]: ellen goosenberg kent ellen goosenberg kent is an american film producer and director. ### ### she is best known for directing and co - producing documentary film'( 2013 ), which won the academy ### ### award for best documentary ( short subject ) at the 87th academy awards ; the win was shared with ### ### producer dana perry. throughout her career, she has worked on numerous films, mostly on television ### ### documentaries, including " i have tourette's but tourette's doesn't have me " ( 2005 ) and'( 2007 ### ### ). she has won four emmy awards out of six nominations for her work on hbo ### ### =============================================== v_q ================================================ ### ### ('kent', 0, 4) ('87', 1, 9) ('ellen', 2, 2) ('academy', 3, 16) ('documentary', 4, 7) ### ### ('goose', 5, 5) ('##nberg', 6, 6) ('##子', 7, 693) ('¹', 8, 539) ('2013', 9, 17) ### ### ('freaking', 10, 645) ('intently', 11, 151) ('貴', 12, 116) ('film', 13, 37) ('short', 14, 28) ### ### ('幸', 15, 108) ('##th', 16, 57) ('##ப', 17, 117) ('³', 18, 1628) ('subject', 19, 19) ### ### =============================================== v_p ================================================ ### ### ('dana', 571, 0) ('perry', 1897, 1) ('ellen', 2, 2) ('documentaries', 143, 3) ('kent', 0, 4) ### ### ('goose', 5, 5) ('##nberg', 6, 6) ('documentary', 4, 7) ('doesn', 3554, 8) ('87', 1, 9) ### ### ('##ette', 4471, 10) ('tour', 20813, 11) ('ceremony', 343, 12) ('hbo', 1021, 13) ('awards', 46, 14) ### ### ('leaned', 6553, 15) ('academy', 3, 16) ('2013', 9, 17) ('worried', 800, 18) ('subject', 19, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('kent', 0, 4) ('ellen', 2, 2) ('documentary', 4, 7) ('87', 1, 9) ('goose', 5, 5) ('##nberg', 6, 6) ### ### ('academy', 3, 16) ('2013', 9, 17) ('subject', 19, 19) ('short', 14, 28) ('documentaries', 143, 3) ### ### ('awards', 46, 14) ('film', 13, 37) ('littered', 42, 24) ('filmmakers', 21, 40) ('dana', 571, 0) ### ### ('##th', 16, 57) ('ceremony', 343, 12) ('absently', 22, 61) ('producers', 101, 27) ### ############################################################################################################ [2023-11-25 17:30:21,934][root][INFO] - Epoch: 4: Step: 1301/19530, loss[v]=0.086394, lr=0.000010, acc@1[1]=1000.5/1024=0.97705078125, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 17:32:09,931][root][INFO] - Train batch 1400 [2023-11-25 17:32:09,932][root][INFO] - Avg. loss per last 100 batches: 0.131569 [2023-11-25 17:32:10,623][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.5/29522=99.95% | mean: 0.02 | max: 5.71 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.4/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the u. s. women's team smashed a new olympic record to recapture their freestyle relay ### ### title with the help of a sterling anchor leg from allison schmitt. ### ### [P_TEXT]: dana vollmer vollmer also competed in the 4×200 - meter freestyle relay. she swam the ### ### second leg with a time of 1 : 56. 02, as the u. s. team won gold with a time of 7 : 42. 92. in her ### ### final event, the 4×100 - meter medley relay, vollmer won another gold with missy franklin, rebecca ### ### soni and allison schmitt. swimming the butterfly leg, vollmer recorded a split time of 55. 48, as ### ### the u. s. team set a new world record with a time of 3 : 52. 05, bettering the previous record of 3 ### ### : 52. 19 set by china in 2009. at the 2016 united states olympic trials, the u. s. qualifying meet ### ### for the rio olympics, vollmer qualified for the u. s. olympic ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 119) ('##hmi', 1, 2) ('貴', 2, 127) ('allison', 3, 7) ('intently', 4, 153) ### ### ('freestyle', 5, 15) ('anchor', 6, 12147) ('##tt', 7, 11) ('খ', 8, 116) ('幸', 9, 144) ('ம', 10, 96) ### ### ('##ப', 11, 109) ('¹', 12, 133) ('women', 13, 160) ('u', 14, 30) ('relay', 15, 27) ### ### ('sterling', 16, 15799) ('olympics', 17, 12) ('.', 18, 1161) ('³', 19, 426) ### ### =============================================== v_p ================================================ ### ### ('##lmer', 25755, 0) ('dana', 15260, 1) ('##hmi', 1, 2) ('franklin', 3852, 3) ('rebecca', 1684, 4) ### ### ('trials', 7423, 5) ('missy', 317, 6) ('allison', 3, 7) ('vol', 15170, 8) ('son', 23291, 9) ### ### ('##200', 6346, 10) ('##tt', 7, 11) ('olympics', 17, 12) ('anyway', 32, 13) ('china', 4128, 14) ### ### ('freestyle', 5, 15) ('split', 7554, 16) ('butterfly', 993, 17) ('throws', 8721, 18) ### ### ('swimming', 67, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##hmi', 1, 2) ('allison', 3, 7) ('freestyle', 5, 15) ('##tt', 7, 11) ('olympics', 17, 12) ### ### ('relay', 15, 27) ('u', 14, 30) ('states', 21, 25) ('record', 23, 20) ('anyway', 32, 13) ### ### ('olympic', 39, 23) ('usa', 24, 41) ('absently', 0, 119) ('sc', 20, 54) ('swimming', 67, 19) ### ### ('貴', 2, 127) ('ம', 10, 96) ('leg', 42, 38) ('intently', 4, 153) ('খ', 8, 116) ### ############################################################################################################ [2023-11-25 17:32:11,127][root][INFO] - Epoch: 4: Step: 1401/19530, loss[v]=0.128592, lr=0.000010, acc@1[1]=983.5/1024=0.96044921875, acc@1[2]=1011.0/1024=0.9873046875 [2023-11-25 17:33:59,258][root][INFO] - Train batch 1500 [2023-11-25 17:33:59,259][root][INFO] - Avg. loss per last 100 batches: 0.137295 [2023-11-25 17:33:59,967][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.76 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29505.7/29522=99.94% | mean: 0.02 | max: 5.65 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: poundley would have been working on the much larger farm complex which he had designed ### ### for john naylor at leighton hall near welshpool. ### ### [P_TEXT]: leighton hall, powys also highly innovative. the leighton buildings can be compared with ### ### the smaller farm steading illustrated in poundley's " cottage architecture " of 1857, which fairly ### ### closely resembles a smaller farm at upper pengelli on naylor's kerry estate. leighton home farm is ### ### also imitated by home farm, courtleaze, coleshill, oxfordshire, which was designed and laid out ### ### between 1852 and 1854 by e. moore, land agent to the 2nd earl radnor. these are the two most ### ### remarkable buildings in the home farm complex and are listed grade ii *. they are at the north - ### ### west side of the farm and were ### ### =============================================== v_q ================================================ ### ### ('nay', 0, 11) ('leighton', 1, 0) ('##pool', 2, 15061) ('pound', 3, 2) ('##lor', 4, 15) ### ### ('¹', 5, 321) ('welsh', 6, 137) ('##ley', 7, 23) ('hall', 8, 20) ('intently', 9, 238) ### ### ('farm', 10, 5) ('貴', 11, 179) ('খ', 12, 235) ('john', 13, 18446) ('absently', 14, 224) ### ### ('幸', 15, 227) ('³', 16, 3292) ('cambridge', 17, 1857) ('##ப', 18, 266) ('rebuilding', 19, 2195) ### ### =============================================== v_p ================================================ ### ### ('leighton', 1, 0) ('kerry', 10096, 1) ('pound', 3, 2) ('moore', 10781, 3) ('cottage', 1709, 4) ### ### ('farm', 10, 5) ('cole', 9143, 6) ('##elli', 26200, 7) ('peng', 26855, 8) ('pow', 8524, 9) ### ### ('##ading', 13231, 10) ('nay', 0, 11) ('##lea', 20559, 12) ('##nor', 21669, 13) ### ### ('##shi', 25682, 14) ('##lor', 4, 15) ('anyway', 321, 16) ('court', 2037, 17) ### ### ('architecture', 45, 18) ('illustrated', 2565, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('leighton', 1, 0) ('nay', 0, 11) ('pound', 3, 2) ('##lor', 4, 15) ('farm', 10, 5) ('hall', 8, 20) ### ### ('##ley', 7, 23) ('architecture', 45, 18) ('architect', 22, 44) ('oxfordshire', 97, 26) ### ### ('welsh', 6, 137) ('designed', 32, 53) ('anyway', 321, 16) ('buildings', 62, 40) ('oxford', 46, 47) ### ### ('estate', 211, 27) ('architectural', 84, 42) ('1857', 313, 28) ('design', 25, 77) ('¹', 5, 321) ### ############################################################################################################ [2023-11-25 17:34:00,473][root][INFO] - Epoch: 4: Step: 1501/19530, loss[v]=0.105098, lr=0.000010, acc@1[1]=994.0/1024=0.970703125, acc@1[2]=1012.5/1024=0.98876953125 [2023-11-25 17:35:49,193][root][INFO] - Train batch 1600 [2023-11-25 17:35:49,193][root][INFO] - Avg. loss per last 100 batches: 0.132381 [2023-11-25 17:35:49,892][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.7/29522=99.95% | mean: 0.03 | max: 5.89 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.6/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: after retiring from the chicago bulls in early 1999, michael jordan became the washington ### ### wizards'president of basketball operations as well as a minority owner in january 2000. ### ### [P_TEXT]: michael jordan also holds the highest television rating of any game in nba history. with ### ### phil jackson's contract expiring, the pending departures of scottie pippen and dennis rodman ### ### looming, and being in the latter stages of an owner - induced lockout of nba players, jordan ### ### retired for the second time on january 13, 1999. on january 19, 2000, jordan returned to the nba ### ### not as a player, but as part owner and president of basketball operations for the washington ### ### wizards. jordan's responsibilities with the wizards were comprehensive. he controlled all aspects ### ### of ### ### =============================================== v_q ================================================ ### ### ('wizards', 0, 7) ('absently', 1, 53) ('jordan', 2, 0) ('washington', 3, 18) ('michael', 4, 1) ### ### ('¹', 5, 75) ('幸', 6, 121) ('貴', 7, 234) ('bulls', 8, 2627) ('খ', 9, 106) ('chicago', 10, 4178) ### ### ('nba', 11, 6) ('intently', 12, 192) ('.', 13, 116) ('³', 14, 756) ('owner', 15, 25) ('ம', 16, 73) ### ### ('##ப', 17, 119) ('basketball', 18, 30) ('executives', 19, 1960) ### ### =============================================== v_p ================================================ ### ### ('jordan', 2, 0) ('michael', 4, 1) ('rod', 14378, 2) ('phil', 4420, 3) ('pip', 14884, 4) ### ### ('jackson', 382, 5) ('nba', 11, 6) ('wizards', 0, 7) ('littered', 476, 8) ('scott', 3236, 9) ### ### ('ratings', 5864, 10) ('dennis', 4708, 11) ('anyway', 167, 12) ('1999', 23, 13) ### ### ('worried', 223, 14) ('rating', 1713, 15) ('##pen', 23449, 16) ('2000', 25, 17) ### ### ('washington', 3, 18) ('lock', 9996, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('jordan', 2, 0) ('michael', 4, 1) ('wizards', 0, 7) ('nba', 11, 6) ('washington', 3, 18) ### ### ('owner', 15, 25) ('absently', 1, 53) ('1999', 23, 13) ('2000', 25, 17) ('basketball', 18, 30) ### ### ('¹', 5, 75) ('president', 20, 33) ('খ', 9, 106) ('幸', 6, 121) ('ம', 16, 73) ('.', 13, 116) ### ### ('jackson', 382, 5) ('anyway', 167, 12) ('stared', 31, 43) ('owners', 27, 48) ### ############################################################################################################ [2023-11-25 17:35:50,397][root][INFO] - Epoch: 4: Step: 1601/19530, loss[v]=0.076703, lr=0.000010, acc@1[1]=998.0/1024=0.974609375, acc@1[2]=1013.5/1024=0.98974609375 [2023-11-25 17:37:38,436][root][INFO] - Train batch 1700 [2023-11-25 17:37:38,437][root][INFO] - Avg. loss per last 100 batches: 0.130763 [2023-11-25 17:37:39,128][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.5/29522=99.95% | mean: 0.02 | max: 5.60 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: maple ridge - pitt meadows and mission - kent were created from dewdney during the 1989 ### ### redistribution. ### ### [P_TEXT]: maple ridge - pitt meadows maple ridge - pitt meadows is a provincial electoral district ### ### for the legislative assembly of british columbia, canada. the 2008 re - distribution is an ### ### amalgamation of the old maple ridge - pitt meadows riding with an addition of a small portion of ### ### the old maple ridge - mission. the riding was created for the 1991 election from part of the dual - ### ### member dewdney riding. ### ### =============================================== v_q ================================================ ### ### ('mission', 0, 1) ('dew', 1, 3) ('kent', 2, 8492) ('##ney', 3, 5) ('maple', 4, 2) ('meadows', 5, 4) ### ### ('ridge', 6, 11) ('pitt', 7, 10) ('¹', 8, 594) ('absently', 9, 93) ('intently', 10, 108) ### ### ('幸', 11, 202) ('貴', 12, 137) ('##ப', 13, 148) ('খ', 14, 146) ('1989', 15, 500) ('.', 16, 3110) ### ### ('district', 17, 154) ('ம', 18, 96) ('ontario', 19, 318) ### ### =============================================== v_p ================================================ ### ### ('1991', 348, 0) ('mission', 0, 1) ('maple', 4, 2) ('dew', 1, 3) ('meadows', 5, 4) ('##ney', 3, 5) ### ### ('dual', 14399, 6) ('riding', 46, 7) ('election', 28, 8) ('2008', 5692, 9) ('pitt', 7, 10) ### ### ('ridge', 6, 11) ('distribution', 2024, 12) ('anyway', 201, 13) ('worried', 92, 14) ### ### ('columbia', 94, 15) ('legislative', 27, 16) ('leaned', 126, 17) ('subdivisions', 168, 18) ### ### ('vancouver', 66, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('mission', 0, 1) ('dew', 1, 3) ('maple', 4, 2) ('##ney', 3, 5) ('meadows', 5, 4) ('ridge', 6, 11) ### ### ('pitt', 7, 10) ('election', 28, 8) ('legislative', 27, 16) ('riding', 46, 7) ('##d', 21, 30) ### ### ('provincial', 37, 26) ('assembly', 30, 34) ('worried', 92, 14) ('absently', 9, 93) ### ### ('vancouver', 66, 19) ('columbia', 94, 15) ('1991', 348, 0) ('intently', 10, 108) ### ### ('anyway', 201, 13) ### ############################################################################################################ [2023-11-25 17:37:39,632][root][INFO] - Epoch: 4: Step: 1701/19530, loss[v]=0.148608, lr=0.000010, acc@1[1]=983.5/1024=0.96044921875, acc@1[2]=1007.0/1024=0.9833984375 [2023-11-25 17:39:27,996][root][INFO] - Train batch 1800 [2023-11-25 17:39:27,997][root][INFO] - Avg. loss per last 100 batches: 0.129545 [2023-11-25 17:39:28,715][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.1/29522=99.95% | mean: 0.03 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: however, there are vulnerabilities to tcp including denial of service, connection ### ### hijacking, tcp veto, and reset attack. ### ### [P_TEXT]: denial - of - service attack sent to tcp port 139 of the victim's machine, causing it to ### ### lock up and display a blue screen of death. attackers have found a way to exploit a number of bugs ### ### in peer - to - peer servers to initiate ddos attacks. the most aggressive of these peer - to - peer ### ### - ddos attacks exploits dc + +. with peer - to - peer there is no botnet and the attacker does not ### ### have to communicate with the clients it subverts. instead, the attacker acts as a " puppet master, ### ### " instructing clients of large peer - to - peer file sharing hubs to disconnect from their peer - ### ### to - peer network and to ### ### =============================================== v_q ================================================ ### ### ('reset', 0, 3154) ('absently', 1, 179) ('tc', 2, 8) ('denial', 3, 2) ('貴', 4, 223) ### ### ('service', 5, 31) ('.', 6, 164) ('幸', 7, 277) ('intently', 8, 275) ('vulnerability', 9, 78) ### ### ('##ப', 10, 250) ('veto', 11, 1987) ('##jack', 12, 15771) ('##p', 13, 62) ('খ', 14, 227) ### ### ('attack', 15, 10) ('ம', 16, 185) ('hi', 17, 23630) ('vu', 18, 26879) ('¹', 19, 325) ### ### =============================================== v_p ================================================ ### ### ('puppet', 4771, 0) ('dd', 4269, 1) ('denial', 3, 2) ('##net', 1040, 3) ('##vert', 13101, 4) ### ### ('bot', 4933, 5) ('peer', 3508, 6) ('139', 9611, 7) ('tc', 2, 8) ('dc', 1977, 9) ('attack', 15, 10) ### ### ('master', 8649, 11) ('port', 4186, 12) ('attackers', 680, 13) ('servers', 1395, 14) ### ### ('communicate', 6659, 15) ('screen', 1279, 16) ('exploit', 28, 17) ('victim', 185, 18) ### ### ('##nne', 18582, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('denial', 3, 2) ('tc', 2, 8) ('attack', 15, 10) ('service', 5, 31) ('exploit', 28, 17) ### ### ('attacks', 25, 29) ('##p', 13, 62) ('anyway', 30, 35) ('vulnerability', 9, 78) ### ### ('absently', 1, 179) ('network', 46, 43) ('victim', 185, 18) ('.', 6, 164) ('貴', 4, 223) ### ### ('##net', 1040, 3) ('intently', 8, 275) ('幸', 7, 277) ('ம', 16, 185) ('##ப', 10, 250) ### ### ('clients', 235, 33) ### ############################################################################################################ [2023-11-25 17:39:29,222][root][INFO] - Epoch: 4: Step: 1801/19530, loss[v]=0.067393, lr=0.000010, acc@1[1]=1001.0/1024=0.9775390625, acc@1[2]=1015.5/1024=0.99169921875 [2023-11-25 17:41:17,647][root][INFO] - Train batch 1900 [2023-11-25 17:41:17,648][root][INFO] - Avg. loss per last 100 batches: 0.127086 [2023-11-25 17:41:18,355][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29509.7/29522=99.96% | mean: 0.03 | max: 5.87 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: a railway collision ( also known as " a railroad wreck " ) is a 1900 british short silent ### ### drama film, directed by walter r. booth and produced by robert w. paul. ### ### [P_TEXT]: walter r. booth ( both 1899 ). many of their early collaborations, such as " hindoo ### ### jugglers " and " chinese magic " ( both 1900 ) were based on conjuring tricks, whilst " a railway ### ### collision " ( 1900 ) pioneered the use of scale models. they reached the height of their ### ### collaboration in 1901 ; with simple trick films, such as " undressing extraordinary ", " the waif ### ### and the wizard " and " an over - incubated baby " which relied on jump - cuts, " the devil in the ### ### studio " and " artistic creation " which integrated hand drawn elements, and " cheese mites ; or, ### ### lilliputians ### ### =============================================== v_q ================================================ ### ### ('booth', 0, 0) ('paul', 1, 22952) ('wreck', 2, 1320) ('collision', 3, 1) ('¹', 4, 234) ### ### ('railroad', 5, 503) ('silent', 6, 8894) ('1900', 7, 32) ('walter', 8, 2) ('″', 9, 33) ### ### ('w', 10, 14813) ('robert', 11, 11104) ('³', 12, 911) ('railway', 13, 38) ('r', 14, 24) ### ### ('film', 15, 43) ('##½', 16, 145) ('chaplin', 17, 2474) ('খ', 18, 159) ('collided', 19, 1161) ### ### =============================================== v_p ================================================ ### ### ('booth', 0, 0) ('collision', 3, 1) ('walter', 8, 2) ('cheese', 6832, 3) ('wai', 19257, 4) ### ### ('hind', 16440, 5) ('wizard', 7435, 6) ('magic', 8391, 7) ('devil', 5503, 8) ('trick', 2053, 9) ### ### ('scale', 7571, 10) ('mit', 16728, 11) ('chinese', 10782, 12) ('drawn', 12438, 13) ### ### ('##ress', 22620, 14) ('lil', 14830, 15) ('studio', 7376, 16) ('jug', 14667, 17) ### ### ('##ring', 16281, 18) ('##uba', 29270, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('booth', 0, 0) ('collision', 3, 1) ('walter', 8, 2) ('1900', 7, 32) ('″', 9, 33) ('r', 14, 24) ### ### ('railway', 13, 38) ('film', 15, 43) ('1901', 22, 31) ('¹', 4, 234) ('films', 24, 58) ('"', 20, 74) ### ### ('1899', 66, 37) ('railroad', 5, 503) ('##½', 16, 145) ('cinema', 60, 56) ('wreck', 2, 1320) ### ### ('খ', 18, 159) ('train', 25, 127) ('幸', 21, 210) ### ############################################################################################################ [2023-11-25 17:41:18,860][root][INFO] - Epoch: 4: Step: 1901/19530, loss[v]=0.069424, lr=0.000010, acc@1[1]=999.5/1024=0.97607421875, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 17:43:06,830][root][INFO] - Train batch 2000 [2023-11-25 17:43:06,831][root][INFO] - Avg. loss per last 100 batches: 0.131689 [2023-11-25 17:43:07,528][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.04 | max: 4.01 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29511.2/29522=99.96% | mean: 0.03 | max: 5.96 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.7/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: renaissance capital was established in 1995 by a group of founding partners, including ### ### stephen jennings, boris jordan, leonid rozhetskin, anton kudryashov and richard deitz. ### ### [P_TEXT]: leonid rozhetskin and then as the head of his own law firm, representing clients such as ### ### the international finance corporation ( a division of the world bank ), credit suisse, morgan ### ### grenfell and " the moscow times ". in 1995, rozhetskin's focus shifted from the law to financial ### ### ventures. rozhetskin was part of a group that founded renaissance capital, russia's first ### ### investment bank, in partnership with boris jordan, an american of russian origin, and new zealand ### ### er stephen jennings. in 1998, rozhetskin left renaissance capital to ### ### =============================================== v_q ================================================ ### ### ('renaissance', 0, 1) ('jennings', 1, 0) ('##ets', 2, 7) ('boris', 3, 5) ('capital', 4, 11) ### ### ('jordan', 5, 3) ('stephen', 6, 12) ('¹', 7, 155) ('leonid', 8, 2) ('##tz', 9, 12571) ### ### ('dei', 10, 25198) ('##dry', 11, 29357) ('##ash', 12, 25119) ('absently', 13, 116) ### ### ('anton', 14, 17105) ('³', 15, 2005) ('ku', 16, 23401) ('richard', 17, 12952) ('##kin', 18, 18) ### ### ('founder', 19, 494) ### ### =============================================== v_p ================================================ ### ### ('jennings', 1, 0) ('renaissance', 0, 1) ('leonid', 8, 2) ('jordan', 5, 3) ('morgan', 6178, 4) ### ### ('boris', 3, 5) ('##fell', 28631, 6) ('##ets', 2, 7) ('finance', 185, 8) ('moscow', 43, 9) ### ### ('times', 4025, 10) ('capital', 4, 11) ('stephen', 6, 12) ('zealand', 12662, 13) ('bank', 95, 14) ### ### ('##zh', 27, 15) ('sui', 24251, 16) ('russian', 36, 17) ('##kin', 18, 18) ('leaned', 381, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('renaissance', 0, 1) ('jennings', 1, 0) ('boris', 3, 5) ('jordan', 5, 3) ('##ets', 2, 7) ### ### ('leonid', 8, 2) ('capital', 4, 11) ('stephen', 6, 12) ('##zh', 27, 15) ('##kin', 18, 18) ### ### ('moscow', 43, 9) ('ro', 23, 30) ('russian', 36, 17) ('bank', 95, 14) ('russia', 39, 37) ### ### ('finance', 185, 8) ('clients', 57, 34) ('investment', 73, 23) ('partnership', 93, 20) ### ### ('1995', 51, 40) ### ############################################################################################################ [2023-11-25 17:43:08,032][root][INFO] - Epoch: 4: Step: 2001/19530, loss[v]=0.077050, lr=0.000009, acc@1[1]=1002.5/1024=0.97900390625, acc@1[2]=1018.5/1024=0.99462890625 [2023-11-25 17:44:56,405][root][INFO] - Train batch 2100 [2023-11-25 17:44:56,406][root][INFO] - Avg. loss per last 100 batches: 0.139112 [2023-11-25 17:44:57,126][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.80 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.5/29522=99.95% | mean: 0.02 | max: 5.70 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 1999, he played robin goodfellow / puck in the michael hoffman adaptation of " a ### ### midsummer night's dream ". ### ### [P_TEXT]: a midsummer night's dream ( 1999 film ) a midsummer night's dream is a 1999 romantic ### ### comedy fantasy film based on the play " a midsummer night's dream " by william shakespeare. it was ### ### directed by michael hoffman. the ensemble cast features kevin kline as bottom, michelle pfeiffer ### ### and rupert everett as titania and oberon, stanley tucci as puck, and calista flockhart, anna friel, ### ### christian bale, and dominic west as the four lovers. in 19th century monte athena, in the ### ### =============================================== v_q ================================================ ### ### ('hoffman', 0, 2) ('##fell', 1, 28835) ('robin', 2, 3743) ('puck', 3, 3) ('1999', 4, 7) ### ### ('midsummer', 5, 4) ('¹', 6, 58) ('貴', 7, 161) ('.', 8, 165) ('michael', 9, 34) ### ### ('littered', 10, 49) ('##¨', 11, 1726) ('night', 12, 29) ('dream', 13, 8) ('absently', 14, 82) ### ### ('幸', 15, 153) ('খ', 16, 184) ('pmid', 17, 541) ('##ow', 18, 28447) ('intently', 19, 172) ### ### =============================================== v_p ================================================ ### ### ('athena', 4580, 0) ('bottom', 2464, 1) ('hoffman', 0, 2) ('puck', 3, 3) ('midsummer', 5, 4) ### ### ('titan', 6525, 5) ('everett', 14558, 6) ('1999', 4, 7) ('dream', 13, 8) ('bal', 6283, 9) ### ### ('obe', 11045, 10) ('##line', 8746, 11) ('rupert', 207, 12) ('flock', 2455, 13) ### ### ('dominic', 230, 14) ('worried', 367, 15) ('##cci', 20622, 16) ('##iel', 3808, 17) ### ### ('##iff', 17465, 18) ('monte', 10600, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('hoffman', 0, 2) ('puck', 3, 3) ('1999', 4, 7) ('midsummer', 5, 4) ('dream', 13, 8) ### ### ('michael', 9, 34) ('night', 12, 29) ('littered', 10, 49) ('¹', 6, 58) ('anyway', 32, 21) ### ### ('convincing', 23, 47) ('absently', 14, 82) ('rupert', 207, 12) ('film', 37, 48) ### ### ('dominic', 230, 14) ('貴', 7, 161) ('##子', 56, 46) ('.', 8, 165) ('worried', 367, 15) ### ### ('play', 68, 44) ### ############################################################################################################ [2023-11-25 17:44:57,633][root][INFO] - Epoch: 4: Step: 2101/19530, loss[v]=0.067938, lr=0.000009, acc@1[1]=1002.0/1024=0.978515625, acc@1[2]=1016.0/1024=0.9921875 [2023-11-25 17:46:46,068][root][INFO] - Train batch 2200 [2023-11-25 17:46:46,069][root][INFO] - Avg. loss per last 100 batches: 0.132120 [2023-11-25 17:46:46,761][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.83 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.9/29522=99.95% | mean: 0.02 | max: 5.62 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: mary and max is a 2009 australian stop - motion adult - animated comedy - drama film ### ### written and directed by adam elliot and was his first animated feature film. ### ### [P_TEXT]: elena kats - chernin york theme in adam elliot's feature film, " mary and max ". kats - ### ### chernin is a represented composer of the australian music centre. a portrait of kats - chernin by ### ### australian portrait artist wendy sharpe was acquired by the national portrait gallery ( australia ) ### ### in 2019. elena kats - chernin was appointed an officer of the order of australia ( ao ) in january ### ### 2019 " for distinguished service to the performing arts, particularly to music, as an orchestral, ### ### operatic and chamber music composer ". ### ### =============================================== v_q ================================================ ### ### ('max', 0, 4) ('mary', 1, 7) ('elliot', 2, 0) ('″', 3, 26) ('¹', 4, 850) ('adam', 5, 5) ### ### ('2009', 6, 9272) ('adult', 7, 18085) ('animated', 8, 3107) ('³', 9, 3339) ('film', 10, 30) ### ### ('stop', 11, 9574) ('"', 12, 271) ('hissed', 13, 74) ('films', 14, 39) ('খ', 15, 166) ### ### ('幸', 16, 198) ('##ப', 17, 164) ('australian', 18, 15) ('animation', 19, 1437) ### ### =============================================== v_p ================================================ ### ### ('elliot', 2, 0) ('elena', 14995, 1) ('sharpe', 9121, 2) ('theme', 2257, 3) ('max', 0, 4) ### ### ('adam', 5, 5) ('wendy', 2741, 6) ('mary', 1, 7) ('kat', 10264, 8) ('##nin', 20742, 9) ### ### ('portrait', 12838, 10) ('centre', 10182, 11) ('cher', 15621, 12) ('york', 9392, 13) ### ### ('themes', 2184, 14) ('australian', 18, 15) ('australia', 29, 16) ('anyway', 1520, 17) ### ### ('music', 10176, 18) ('ao', 23239, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('elliot', 2, 0) ('max', 0, 4) ('mary', 1, 7) ('adam', 5, 5) ('″', 3, 26) ('film', 10, 30) ### ### ('australian', 18, 15) ('australia', 29, 16) ('films', 14, 39) ('hissed', 13, 74) ('movie', 31, 48) ### ### ('feature', 35, 56) ('leaned', 320, 22) ('¹', 4, 850) ('sydney', 67, 41) ('filming', 155, 34) ### ### ('movies', 24, 78) ('filmmakers', 44, 57) ('theme', 2257, 3) ('absently', 26, 84) ### ############################################################################################################ [2023-11-25 17:46:47,265][root][INFO] - Epoch: 4: Step: 2201/19530, loss[v]=0.178625, lr=0.000009, acc@1[1]=982.5/1024=0.95947265625, acc@1[2]=1008.5/1024=0.98486328125 [2023-11-25 17:48:35,085][root][INFO] - Train batch 2300 [2023-11-25 17:48:35,086][root][INFO] - Avg. loss per last 100 batches: 0.125182 [2023-11-25 17:48:35,781][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29509.5/29522=99.96% | mean: 0.03 | max: 5.81 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: his final film appearance was in " the legend of bagger vance " ( 2000 ). ### ### [P_TEXT]: grantland rice russell. for many years, a portion of one floor of the columbia university ### ### graduate school of journalism was designated the " grantland rice suite ". grantland avenue in his ### ### hometown of murfreesboro, tennessee, was named in his honor. rice was mentioned in an " i love lucy ### ### " episode entitled " the camping trip ", and was portrayed by actor lane smith, also a native of ### ### tennessee, in " the legend of bagger vance ". on june 8, 2011, espn's bill simmons launched a ### ### sports and popular culture website titled ### ### =============================================== v_q ================================================ ### ### ('vance', 0, 1) ('bag', 1, 11) ('legend', 2, 17) ('##ger', 3, 36) ('##¨', 4, 1129) ('¹', 5, 340) ### ### ('film', 6, 315) ('2000', 7, 12186) ('intently', 8, 189) ('貴', 9, 156) ('vaughn', 10, 2009) ### ### ('freaking', 11, 391) ('³', 12, 5206) ('খ', 13, 177) ('幸', 14, 194) ('##ப', 15, 178) ### ### ('subtly', 16, 1140) ('pmid', 17, 2366) ('convinces', 18, 83) ('convincing', 19, 89) ### ### =============================================== v_p ================================================ ### ### ('camping', 9786, 0) ('vance', 0, 1) ('lane', 8464, 2) ('simmons', 2208, 3) ('rice', 4938, 4) ### ### ('journalism', 9460, 5) ('espn', 16926, 6) ('grant', 1028, 7) ('smith', 5850, 8) ('lucy', 2648, 9) ### ### ('##land', 16532, 10) ('bag', 1, 11) ('trip', 1357, 12) ('##sboro', 20707, 13) ### ### ('graduate', 2557, 14) ('tennessee', 4708, 15) ('bill', 162, 16) ('legend', 2, 17) ### ### ('worried', 1194, 18) ('columbia', 4616, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('vance', 0, 1) ('bag', 1, 11) ('legend', 2, 17) ('##ger', 3, 36) ('littered', 27, 22) ### ### ('anyway', 61, 24) ('bill', 162, 16) ('convinces', 18, 83) ('grant', 1028, 7) ### ### ('convincing', 19, 89) ('貴', 9, 156) ('intently', 8, 189) ('simmons', 2208, 3) ### ### ('episodes', 247, 32) ('portrayed', 64, 64) ('¹', 5, 340) ('legends', 140, 42) ('trip', 1357, 12) ### ### ('film', 6, 315) ('খ', 13, 177) ### ############################################################################################################ [2023-11-25 17:48:36,288][root][INFO] - Epoch: 4: Step: 2301/19530, loss[v]=0.178346, lr=0.000009, acc@1[1]=984.5/1024=0.96142578125, acc@1[2]=1005.0/1024=0.9814453125 [2023-11-25 17:50:24,749][root][INFO] - Train batch 2400 [2023-11-25 17:50:24,750][root][INFO] - Avg. loss per last 100 batches: 0.126347 [2023-11-25 17:50:25,476][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.04 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29509.0/29522=99.96% | mean: 0.03 | max: 5.71 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.2/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: she encouraged catherine nicholson and harriet desmoines to bring the lesbian - feminist ### ### journal " sinister wisdom " to lincoln. ### ### [P_TEXT]: feminist separatism " lesbian nation : the feminist solution " ( simon & schuster, 1973 ) ### ### is a collection of essays written by jill johnston, that were originally printed in " the village ### ### voice ", where johnston discusses elements of breaking off from the male - dominated institutions. ### ### " for lesbians only : a separatist anthology " ( onlywomen press, 1988 ), edited by julia penelope ### ### and sarah lucia hoagland, is a collection of writings on lesbian separatism. notable us lesbian ### ### separatist periodicals include " common lives / lesbian lives " ( iowa, 1980 – 1996 ), " lesbian ### ### connection " ( michigan, 1974 – present ), " sinister wisdom ### ### =============================================== v_q ================================================ ### ### ('lincoln', 0, 11992) ('catherine', 1, 2612) ('nicholson', 2, 8221) ('wisdom', 3, 7) ### ### ('harriet', 4, 3670) ('des', 5, 20771) ('##ines', 6, 20173) ('sinister', 7, 4) ('lesbian', 8, 5) ### ### ('##mo', 9, 26585) ('¹', 10, 273) ('lincolnshire', 11, 15549) ('pained', 12, 555) ### ### ('feminist', 13, 8) ('herself', 14, 51) ('##子', 15, 68) ('helen', 16, 702) ('anarchist', 17, 104) ### ### ('townspeople', 18, 8742) ('martha', 19, 2879) ### ### =============================================== v_p ================================================ ### ### ('johnston', 9272, 0) ('penelope', 596, 1) ('jill', 2347, 2) ('##arat', 26401, 3) ### ### ('sinister', 7, 4) ('lesbian', 8, 5) ('lucia', 8506, 6) ('wisdom', 3, 7) ('feminist', 13, 8) ### ### ('sep', 22522, 9) ('breaking', 13162, 10) ('julia', 241, 11) ('voice', 9538, 12) ### ### ('connection', 7109, 13) ('lives', 1055, 14) ('solution', 8468, 15) ('##wo', 14563, 16) ### ### ('##ag', 22394, 17) ('village', 4810, 18) ('schuster', 4355, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('wisdom', 3, 7) ('sinister', 7, 4) ('lesbian', 8, 5) ('feminist', 13, 8) ('herself', 14, 51) ### ### ('magazine', 46, 34) ('##子', 15, 68) ('julia', 241, 11) ('penelope', 596, 1) ('anthology', 122, 21) ### ### ('sarah', 157, 23) ('press', 105, 32) ('feminism', 126, 31) ('she', 144, 42) ('writings', 147, 43) ### ### ('publication', 520, 20) ('anarchist', 17, 104) ('jill', 2347, 2) ('lives', 1055, 14) ### ### ('littered', 299, 41) ### ############################################################################################################ [2023-11-25 17:50:25,985][root][INFO] - Epoch: 4: Step: 2401/19530, loss[v]=0.173582, lr=0.000009, acc@1[1]=974.0/1024=0.951171875, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-25 17:52:14,193][root][INFO] - Train batch 2500 [2023-11-25 17:52:14,194][root][INFO] - Avg. loss per last 100 batches: 0.130880 [2023-11-25 17:52:14,881][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.82 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.0/29522=99.95% | mean: 0.02 | max: 5.89 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: meanwhile, qing attention became temporarily focused on the khalka prince chingunjav, a ### ### descendant of genghis khan, who between the summer of 1756 and january 1757 mounted the most ### ### serious khalka mongol rebellion against the qing until its demise in 1911. ### ### [P_TEXT]: ten great campaigns summer of 1756, amursana started a dzungar revolt against the chinese ### ### with the help of prince chingunjav. the qing empire reacted at the start of 1757 and sent general ### ### zhaohui with support from burhan al - din and khwaja - i jahan. among several battles, the most ### ### important ones were illustrated in qianlong's paintings. the dzungar leader ayushi defected to the ### ### qing side and attacked the dzungar camp at gadan - ola ( battle of gadan - ola ). general zhaohui ### ### defeated the dzungars in two battles : the battle of oroi - jalatu ( 1758 ) and the battle of ### ### khurungui ( 1758 ). in the first battle, zhaohui attacked amursana's ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 386) ('খ', 1, 193) ('¹', 2, 87) ('貴', 3, 243) ('intently', 4, 293) ('幸', 5, 242) ### ### ('ching', 6, 6) ('ம', 7, 196) ('##v', 8, 42) ('mongol', 9, 475) ('.', 10, 478) ('##ப', 11, 210) ### ### ('##un', 12, 24) ('rebellion', 13, 57) ('##hal', 14, 25325) ('gen', 15, 377) ('##ka', 16, 22915) ### ### ('##ghi', 17, 25111) ('qing', 18, 18) ('##ja', 19, 51) ### ### =============================================== v_p ================================================ ### ### ('campaigns', 4017, 0) ('##hui', 18245, 1) ('ten', 13426, 2) ('oro', 13950, 3) ('##lat', 16337, 4) ### ### ('##nga', 8253, 5) ('ching', 6, 6) ('##hwa', 10003, 7) ('ol', 17046, 8) ('##urs', 3538, 9) ### ### ('##shi', 6343, 10) ('##rh', 13691, 11) ('##yu', 3821, 12) ('revolt', 394, 13) ('zhao', 2923, 14) ### ### ('qi', 1411, 15) ('##dan', 15087, 16) ('paintings', 8368, 17) ('qing', 18, 18) ('great', 272, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ching', 6, 6) ('##un', 12, 24) ('qing', 18, 18) ('##v', 8, 42) ('1757', 23, 23) ### ### ('rebellion', 13, 57) ('##ja', 19, 51) ('¹', 2, 87) ('1756', 29, 35) ('prince', 31, 36) ### ### ('worried', 22, 69) ('anyway', 52, 29) ('খ', 1, 193) ('1758', 95, 33) ('absently', 0, 386) ### ### ('貴', 3, 243) ('幸', 5, 242) ('ம', 7, 196) ('intently', 4, 293) ('great', 272, 19) ### ############################################################################################################ [2023-11-25 17:52:15,386][root][INFO] - Epoch: 4: Step: 2501/19530, loss[v]=0.197812, lr=0.000009, acc@1[1]=972.0/1024=0.94921875, acc@1[2]=1001.5/1024=0.97802734375 [2023-11-25 17:54:03,316][root][INFO] - Train batch 2600 [2023-11-25 17:54:03,317][root][INFO] - Avg. loss per last 100 batches: 0.138789 [2023-11-25 17:54:04,020][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.1/29522=99.95% | mean: 0.02 | max: 5.70 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: , overlooking upper new york bay near the border of today's st. george and tompkinsville. ### ### [P_TEXT]: st. george, staten island led to the population of new brighton doubling between 1840 and ### ### 1860. in 1858, angry staten island residents burned down the quarantine in what became known as the ### ### staten island quarantine war. three years afterward, the onset of the american civil war resulted ### ### in large changes to the neighborhood's land use. initially, the local economy suffered due to ### ### cessation of trade with the southern united states, but because of the union army's demand for ### ### material, many entrepreneurs and workers moved to new york city, including to staten island's north ### ### =============================================== v_q ================================================ ### ### ('##pkins', 0, 3500) ('george', 1, 1) ('york', 2, 37) ('upper', 3, 6729) ('absently', 4, 246) ### ### ('貴', 5, 368) ('intently', 6, 405) ('খ', 7, 281) ('幸', 8, 358) ('¹', 9, 795) ('##ப', 10, 344) ### ### ('bay', 11, 8033) ('ம', 12, 266) ('overlooking', 13, 14915) ('st', 14, 3) ('tom', 15, 14812) ### ### ('##ville', 16, 16510) ('.', 17, 373) ('³', 18, 3990) ('staten', 19, 5) ### ### =============================================== v_p ================================================ ### ### ('brighton', 1813, 0) ('george', 1, 1) ('doubling', 13159, 2) ('st', 14, 3) ('burned', 7221, 4) ### ### ('staten', 19, 5) ('1858', 12593, 6) ('burning', 6521, 7) ('worried', 41, 8) ('1840', 1411, 9) ### ### ('anyway', 53, 10) ('southern', 8769, 11) ('confederacy', 11789, 12) ('##aran', 27118, 13) ### ### ('##tine', 18036, 14) ('island', 79, 15) ('confederates', 24035, 16) ('americans', 201, 17) ### ### ('trade', 610, 18) ('war', 265, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('george', 1, 1) ('st', 14, 3) ('staten', 19, 5) ('york', 2, 37) ('worried', 41, 8) ### ### ('anyway', 53, 10) ('island', 79, 15) ('settlers', 49, 28) ('saint', 39, 48) ('absently', 4, 246) ### ### ('ny', 29, 63) ('খ', 7, 281) ('afterward', 90, 26) ('residents', 143, 20) ('newer', 97, 30) ### ### ('貴', 5, 368) ('brighton', 1813, 0) ('americans', 201, 17) ('幸', 8, 358) ('ம', 12, 266) ### ############################################################################################################ [2023-11-25 17:54:04,526][root][INFO] - Epoch: 4: Step: 2601/19530, loss[v]=0.082296, lr=0.000009, acc@1[1]=997.5/1024=0.97412109375, acc@1[2]=1012.0/1024=0.98828125 [2023-11-25 17:55:52,912][root][INFO] - Train batch 2700 [2023-11-25 17:55:52,912][root][INFO] - Avg. loss per last 100 batches: 0.132627 [2023-11-25 17:55:53,685][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.8/29522=99.95% | mean: 0.02 | max: 5.74 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the large department store salling is located in søndergade and there is a sculpture and ### ### [P_TEXT]: aarhus lighting of the christmas lights on the salling department store in søndergade has ### ### also become an attraction in recent times, packing the pedestrianised city centre with thousands of ### ### revellers. significant dates such as saint lucy's day, " sankt hans " ( saint john's eve ) and ### ### fastelavn are traditionally celebrated with numerous events across the city. the beech forests of ### ### riis skov and marselisborg occupy the hills along the coast to the north and south, and apart from ### ### the city centre, sandy beaches form the coastline of the entire ### ### =============================================== v_q ================================================ ### ### ('sal', 0, 14) ('##nder', 1, 3) ('##ling', 2, 4) ('##ga', 3, 35) ('¹', 4, 95) ('貴', 5, 257) ### ### ('copenhagen', 6, 58) ('খ', 7, 203) ('intently', 8, 269) ('sculpture', 9, 9534) ('norway', 10, 112) ### ### ('oslo', 11, 219) ('##de', 12, 52) ('##ப', 13, 302) ('幸', 14, 230) ('norwegian', 15, 82) ### ### ('³', 16, 365) ('ம', 17, 187) ('denmark', 18, 122) ('absently', 19, 183) ### ### =============================================== v_p ================================================ ### ### ('aarhus', 60, 0) ('lucy', 8489, 1) ('fast', 9235, 2) ('##nder', 1, 3) ('##ling', 2, 4) ### ### ('##eller', 1204, 5) ('##elis', 14455, 6) ('lights', 3240, 7) ('saint', 4031, 8) ('sk', 987, 9) ### ### ('hans', 283, 10) ('beech', 11696, 11) ('sank', 2299, 12) ('store', 20, 13) ('sal', 0, 14) ### ### ('christmas', 303, 15) ('ri', 11885, 16) ('forest', 650, 17) ('forests', 5853, 18) ### ### ('##borg', 93, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##nder', 1, 3) ('sal', 0, 14) ('##ling', 2, 4) ('##ga', 3, 35) ('aarhus', 60, 0) ### ### ('store', 20, 13) ('##de', 12, 52) ('copenhagen', 6, 58) ('##ø', 21, 36) ('¹', 4, 95) ### ### ('anyway', 40, 23) ('danish', 23, 50) ('norwegian', 15, 82) ('norway', 10, 112) ('##borg', 93, 19) ### ### ('worried', 25, 63) ('denmark', 18, 122) ('খ', 7, 203) ('hans', 283, 10) ('貴', 5, 257) ### ############################################################################################################ [2023-11-25 17:55:54,193][root][INFO] - Epoch: 4: Step: 2701/19530, loss[v]=0.157445, lr=0.000009, acc@1[1]=981.5/1024=0.95849609375, acc@1[2]=1000.5/1024=0.97705078125 [2023-11-25 17:57:42,498][root][INFO] - Train batch 2800 [2023-11-25 17:57:42,499][root][INFO] - Avg. loss per last 100 batches: 0.126162 [2023-11-25 17:57:43,192][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.98 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.0/29522=99.95% | mean: 0.02 | max: 5.70 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: western australia, south australia, tasmania, queensland and victoria first met as the ### ### state librarians council in march 1973, with the aim of forming a peak body to represent and lobby ### ### for their organisations in dealings with the commonwealth government. ### ### [P_TEXT]: state library of south australia of the state library's heritage collections and is of ### ### international importance. the library manages, in collaboration with the history trust of south ### ### australia, the centre of democracy on the corner of north terrace and kintore avenue. the centre's ### ### gallery exhibits treasures from history trust and state library collections, as well as items on ### ### loan from state records of south australia, the art gallery of south australia, the courts ### ### authority, parliament house, government house and private lenders. as a member library of national ### ### and state libraries australia, the organisation collaborated on the creation of the national ### ### edeposit ( ned ) system, which ### ### =============================================== v_q ================================================ ### ### ('librarian', 0, 53) ('library', 1, 1) ('¹', 2, 120) ('commonwealth', 3, 5387) ('australia', 4, 10) ### ### ('tasmania', 5, 156) ('peak', 6, 8365) ('貴', 7, 91) ('council', 8, 2182) ('queensland', 9, 279) ### ### ('absently', 10, 118) ('intently', 11, 114) ('幸', 12, 135) ('খ', 13, 124) ('state', 14, 8) ### ### ('##ப', 15, 180) ('ம', 16, 106) ('³', 17, 958) ('australian', 18, 28) ('victoria', 19, 1236) ### ### =============================================== v_p ================================================ ### ### ('terrace', 12964, 0) ('library', 1, 1) ('##tore', 24199, 2) ('courts', 145, 3) ### ### ('libraries', 28, 4) ('trust', 855, 5) ('democracy', 2950, 6) ('parliament', 719, 7) ### ### ('state', 14, 8) ('authority', 116, 9) ('australia', 4, 10) ('anyway', 180, 11) ('hissed', 30, 12) ### ### ('kin', 21600, 13) ('ned', 9091, 14) ('history', 2779, 15) ('terraces', 28699, 16) ### ### ('government', 22, 17) ('gallery', 2421, 18) ('##osi', 27207, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('library', 1, 1) ('australia', 4, 10) ('state', 14, 8) ('libraries', 28, 4) ('librarian', 0, 53) ### ### ('australian', 18, 28) ('government', 22, 17) ('hissed', 30, 12) ('adelaide', 24, 24) ### ### ('courts', 145, 3) ('authority', 116, 9) ('south', 63, 27) ('¹', 2, 120) ('貴', 7, 91) ### ### ('anyway', 180, 11) ('organisations', 48, 32) ('organisation', 140, 22) ('absently', 10, 118) ### ### ('intently', 11, 114) ('tasmania', 5, 156) ### ############################################################################################################ [2023-11-25 17:57:43,695][root][INFO] - Epoch: 4: Step: 2801/19530, loss[v]=0.071600, lr=0.000009, acc@1[1]=999.0/1024=0.9755859375, acc@1[2]=1016.0/1024=0.9921875 [2023-11-25 17:59:31,799][root][INFO] - Train batch 2900 [2023-11-25 17:59:31,800][root][INFO] - Avg. loss per last 100 batches: 0.132905 [2023-11-25 17:59:32,503][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29509.2/29522=99.96% | mean: 0.03 | max: 5.82 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the closure of dartford's major employers : seagers'engineering works, j & e hall ### ### international, vickers, the reduction and subsequent closure of burroughs wellcome ( now ### ### glaxosmithkline ), and the redevelopment of nearby bexleyheath as a shopping town in the 1970s ( ### ### and the more recent development of the bluewater shopping centre ), have had a negative effect on ### ### the economy of dartford, but the town is still home to ### ### [P_TEXT]: ruxley along north cray road to bexley, bexleyheath and eventually dartford then ### ### bluewater shopping centre. the 321 bus, which has been a 24 - hour bus since 2009, terminates at ### ### the tesco superstore in ruxley, from there it takes passengers west through foots cray, sidcup, new ### ### eltham, eltham, lee, lewisham and new cross. the 321 bus has been running since 1997, before this ### ### time the 21 bus, which now only travels from central london to lewisham, had a longer route which ### ### took it from central london through ruxley, and it terminated at either ### ### =============================================== v_q ================================================ ### ### ('dart', 0, 4) ('##water', 1, 7) ('burroughs', 2, 20710) ('##ford', 3, 46) ('##hea', 4, 27) ### ### ('vickers', 5, 12937) ('##xley', 6, 0) ('##mith', 7, 28588) ('##gers', 8, 25156) ('¹', 9, 34) ### ### ('absently', 10, 192) ('##th', 11, 128) ('blue', 12, 18) ('cradled', 13, 619) ('.', 14, 97) ### ### ('leaned', 15, 179) ('##com', 16, 23905) ('kent', 17, 5279) ('খ', 18, 113) ('##la', 19, 28275) ### ### =============================================== v_p ================================================ ### ### ('##xley', 6, 0) ('321', 22578, 1) ('##stor', 17227, 2) ('##tham', 6375, 3) ('dart', 0, 4) ### ### ('ru', 21193, 5) ('lee', 8118, 6) ('##water', 1, 7) ('##sco', 13822, 8) ('##cup', 19029, 9) ### ### ('sid', 2620, 10) ('##ay', 9450, 11) ('lewis', 6013, 12) ('21', 13203, 13) ('cross', 8436, 14) ### ### ('foot', 1840, 15) ('bus', 713, 16) ('parish', 309, 17) ('blue', 12, 18) ('el', 15891, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##xley', 6, 0) ('dart', 0, 4) ('##water', 1, 7) ('##hea', 4, 27) ('blue', 12, 18) ('¹', 9, 34) ### ### ('##ford', 3, 46) ('centre', 20, 37) ('london', 28, 32) ('shopping', 52, 29) ('worried', 80, 23) ### ### ('intently', 27, 55) ('.', 14, 97) ('ம', 24, 82) ('##th', 11, 128) ('খ', 18, 113) ### ### ('parish', 309, 17) ('貴', 31, 72) ('幸', 23, 112) ('absently', 10, 192) ### ############################################################################################################ [2023-11-25 17:59:33,009][root][INFO] - Epoch: 4: Step: 2901/19530, loss[v]=0.227137, lr=0.000009, acc@1[1]=975.5/1024=0.95263671875, acc@1[2]=1002.0/1024=0.978515625 [2023-11-25 18:01:21,752][root][INFO] - Train batch 3000 [2023-11-25 18:01:21,753][root][INFO] - Avg. loss per last 100 batches: 0.137302 [2023-11-25 18:01:22,461][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.6/29522=100.00% | mean: 0.04 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29509.6/29522=99.96% | mean: 0.03 | max: 5.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in the same year he aquatinted and published a view of the serpentine river, hyde park, ### ### etched by jacob schnebbelie in 1787. ### ### [P_TEXT]: jacob schnebbelie his name appearing as the publisher on some of the plates. a view of ### ### the serpentine river, hyde park, etched by schnebbelie in 1787, was aquatinted by francis jukes and ### ### published in 1796. schnebbelie died of rheumatic fever at his residence in poland street, london, ### ### on 21 february 1792. he left a widow and three children, for whom provision was made by the society ### ### of antiquaries. one of his sons, robert blemmell schnebbelie ( 1781 – 1847 ), also practised as a ### ### topographical artist. ### ### =============================================== v_q ================================================ ### ### ('aqua', 0, 1) ('hyde', 1, 2) ('etched', 2, 139) ('serpent', 3, 0) ('##hn', 4, 21) ('park', 5, 19) ### ### ('jacob', 6, 5) ('##bel', 7, 8) ('##eb', 8, 30) ('¹', 9, 915) ('engraved', 10, 388) ### ### ('##ine', 11, 25) ('river', 12, 29) ('engraver', 13, 825) ('##ie', 14, 54) ('view', 15, 36) ### ### ('貴', 16, 443) ('illustrate', 17, 12672) ('##tin', 18, 17) ('sc', 19, 97) ### ### =============================================== v_p ================================================ ### ### ('serpent', 3, 0) ('aqua', 0, 1) ('hyde', 1, 2) ('topographic', 1364, 3) ('poland', 8768, 4) ### ### ('jacob', 6, 5) ('##kes', 22342, 6) ('ju', 25250, 7) ('##bel', 7, 8) ('##umatic', 8360, 9) ### ### ('plates', 2114, 10) ('##lem', 16162, 11) ('fever', 15870, 12) ('##mel', 9617, 13) ### ### ('1796', 542, 14) ('##qua', 21084, 15) ('francis', 387, 16) ('##tin', 18, 17) ### ### ('engraving', 458, 18) ('park', 5, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('aqua', 0, 1) ('hyde', 1, 2) ('serpent', 3, 0) ('jacob', 6, 5) ('##bel', 7, 8) ('##hn', 4, 21) ### ### ('park', 5, 19) ('##eb', 8, 30) ('##ine', 11, 25) ('river', 12, 29) ('##tin', 18, 17) ### ### ('view', 15, 36) ('1787', 21, 35) ('##ie', 14, 54) ('published', 34, 39) ('etched', 2, 139) ### ### ('littered', 33, 57) ('london', 65, 37) ('sc', 19, 97) ('francis', 387, 16) ### ############################################################################################################ [2023-11-25 18:01:22,967][root][INFO] - Epoch: 4: Step: 3001/19530, loss[v]=0.091931, lr=0.000009, acc@1[1]=996.5/1024=0.97314453125, acc@1[2]=1017.5/1024=0.99365234375 [2023-11-25 18:03:11,393][root][INFO] - Train batch 3100 [2023-11-25 18:03:11,393][root][INFO] - Avg. loss per last 100 batches: 0.134899 [2023-11-25 18:03:12,090][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.7/29522=99.95% | mean: 0.02 | max: 5.67 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: their song was called " waterloo ", which was written by bjorn and benny. ### ### [P_TEXT]: sweden in the eurovision song contest 2015, and. prior to the 2015 contest, sweden had ### ### participated in the eurovision song contest fifty - four times since its first entry in 1958. ### ### sweden had won the contest on five occasions : in 1974 with the song " waterloo " performed by ### ### abba, in 1984 with the song " diggi - loo diggi - ley " performed by herreys, in 1991 with the song ### ### " fangad av en stormvind " performed by carola, in 1999 with the song " take me to your heaven " ### ### performed by charlotte nilsson, and in 2012 with the ### ### =============================================== v_q ================================================ ### ### ('waterloo', 0, 0) ('bjorn', 1, 8262) ('benny', 2, 5101) ('.', 3, 284) ('eurovision', 4, 89) ### ### ('¹', 5, 125) ('খ', 6, 129) ('intently', 7, 113) ('andersson', 8, 2050) ('song', 9, 21) ### ### ('貴', 10, 153) ('absently', 11, 164) ('##ப', 12, 159) ('″', 13, 38) ('"', 14, 83) ('ம', 15, 136) ### ### ('幸', 16, 134) ('...', 17, 91) ('³', 18, 298) ('iceland', 19, 1266) ### ### =============================================== v_p ================================================ ### ### ('waterloo', 0, 0) ('carol', 2191, 1) ('herr', 9523, 2) ('ley', 6973, 3) ('sweden', 42, 4) ### ### ('dig', 12765, 5) ('nilsson', 179, 6) ('1958', 229, 7) ('charlotte', 15516, 8) ('contest', 77, 9) ### ### ('heaven', 7012, 10) ('fang', 6243, 11) ('contests', 16511, 12) ('1974', 268, 13) ### ### ('anyway', 73, 14) ('2015', 21305, 15) ('your', 3748, 16) ('storm', 10489, 17) ('ט', 7122, 18) ### ### ('##ey', 12669, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('waterloo', 0, 0) ('song', 9, 21) ('sweden', 42, 4) ('contest', 77, 9) ('swedish', 21, 29) ### ### ('″', 13, 38) ('anyway', 73, 14) ('nilsson', 179, 6) ('1958', 229, 7) ('eurovision', 4, 89) ### ### ('1974', 268, 13) ('songs', 29, 42) ('intently', 7, 113) ('¹', 5, 125) ('খ', 6, 129) ('.', 3, 284) ### ### ('"', 14, 83) ('worried', 171, 30) ('貴', 10, 153) ('...', 17, 91) ### ############################################################################################################ [2023-11-25 18:03:12,594][root][INFO] - Epoch: 4: Step: 3101/19530, loss[v]=0.195572, lr=0.000009, acc@1[1]=977.5/1024=0.95458984375, acc@1[2]=1007.0/1024=0.9833984375 [2023-11-25 18:05:00,934][root][INFO] - Train batch 3200 [2023-11-25 18:05:00,934][root][INFO] - Avg. loss per last 100 batches: 0.135728 [2023-11-25 18:05:01,643][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.03 | max: 3.80 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.1/29522=99.95% | mean: 0.02 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: also been seen on videos such as " live in paris " playing a blond 1958 gibson byrdland. ### ### [P_TEXT]: gibson byrdland using a byrdland, an unusual choice considering nugent's high - volume ### ### style of music. the hollow - bodied design of the guitar caused feedback at higher levels of gain ### ### and volume, which would normally make it impractical for hard rock and similar styles, but nugent ### ### controlled this feedback and incorporated it into his playing. british guitar player john ### ### mclaughlin used a sunburst byrdland with a scalloped fretboard. other famous byrdland players are ### ### anthony wilson, louie shelton, david t. walker and james blood ulmer. the guitar is currently ### ### available as part of gibson's custom series ### ### =============================================== v_q ================================================ ### ### ('paris', 0, 6493) ('byrd', 1, 6) ('live', 2, 7650) ('blond', 3, 4033) ('gibson', 4, 2) ### ### ('##land', 5, 13) ('¹', 6, 146) ('intently', 7, 132) ('貴', 8, 157) ('fender', 9, 81) ### ### ('1958', 10, 18427) ('খ', 11, 134) ('³', 12, 1745) ('guitar', 13, 29) ('幸', 14, 161) ### ### ('##ப', 15, 176) ('video', 16, 3304) ('fleetwood', 17, 2933) ('videos', 18, 7507) ('.', 19, 785) ### ### =============================================== v_p ================================================ ### ### ('shelton', 3535, 0) ('anthony', 6853, 1) ('gibson', 4, 2) ('gain', 4677, 3) ('walker', 6305, 4) ### ### ('louie', 2043, 5) ('byrd', 1, 6) ('feedback', 8788, 7) ('mclaughlin', 4143, 8) ('wilson', 2119, 9) ### ### ('##boards', 27578, 10) ('blood', 13203, 11) ('bodied', 2886, 12) ('##land', 5, 13) ### ### ('##rst', 26256, 14) ('custom', 261, 15) ('##gent', 27951, 16) ('ul', 23278, 17) ### ### ('##mer', 7359, 18) ('hollow', 5750, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('byrd', 1, 6) ('gibson', 4, 2) ('##land', 5, 13) ('guitar', 13, 29) ('guitars', 20, 44) ### ### ('anyway', 64, 25) ('fender', 9, 81) ('¹', 6, 146) ('custom', 261, 15) ('rocker', 26, 62) ### ### ('intently', 7, 132) ('貴', 8, 157) ('খ', 11, 134) ('instrument', 102, 38) ('幸', 14, 161) ### ### ('louie', 2043, 5) ('littered', 563, 26) ('##ப', 15, 176) ('rock', 207, 37) ('wilson', 2119, 9) ### ############################################################################################################ [2023-11-25 18:05:02,148][root][INFO] - Epoch: 4: Step: 3201/19530, loss[v]=0.095335, lr=0.000009, acc@1[1]=993.5/1024=0.97021484375, acc@1[2]=1015.0/1024=0.9912109375 [2023-11-25 18:06:51,072][root][INFO] - Train batch 3300 [2023-11-25 18:06:51,073][root][INFO] - Avg. loss per last 100 batches: 0.130206 [2023-11-25 18:06:51,782][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.2/29522=99.95% | mean: 0.02 | max: 5.74 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: el fassi, a member of the istiqlal party, replaced independent driss jettou. ### ### [P_TEXT]: istiqlal party the liberation of all the territories that once were part of morocco. in ### ### january 2006, istiqlal criticized spanish prime minister jose luis rodriguez zapatero's visit to ### ### the spanish cities of ceuta and melilla on the north african coast, reflecting its nationalist ### ### heritage. istiqlal won 52 out of 325 seats in the parliamentary election held on 7 september 2007, ### ### more than any other party, and subsequently the party's leader, abbas el fassi, was named prime ### ### minister by king mohammed vi on 19 september 2007. the ### ### =============================================== v_q ================================================ ### ### ('##ssi', 0, 1) ('ist', 1, 0) ('##iss', 2, 28677) ('fa', 3, 23) ('##iq', 4, 4) ('##lal', 5, 8) ### ### ('¹', 6, 142) ('el', 7, 28) ('intently', 8, 186) ('jett', 9, 2824) ('貴', 10, 208) ### ### ('absently', 11, 234) ('幸', 12, 199) ('খ', 13, 195) ('tunisia', 14, 349) ('##ப', 15, 217) ### ### ('dr', 16, 17464) ('ம', 17, 156) ('algerian', 18, 1197) ('³', 19, 1741) ### ### =============================================== v_p ================================================ ### ### ('ist', 1, 0) ('##ssi', 0, 1) ('morocco', 47, 2) ('##uta', 16885, 3) ('##iq', 4, 4) ### ### ('abbas', 1354, 5) ('##pate', 23909, 6) ('##illa', 10915, 7) ('##lal', 5, 8) ('moroccan', 60, 9) ### ### ('325', 20384, 10) ('mel', 16443, 11) ('rodriguez', 4796, 12) ('mohammed', 237, 13) ### ### ('littered', 1559, 14) ('vi', 20135, 15) ('african', 223, 16) ('2007', 339, 17) ('anyway', 34, 18) ### ### ('spanish', 2947, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ssi', 0, 1) ('ist', 1, 0) ('##iq', 4, 4) ('##lal', 5, 8) ('fa', 3, 23) ('el', 7, 28) ### ### ('morocco', 47, 2) ('elections', 21, 32) ('anyway', 34, 18) ('moroccan', 60, 9) ('party', 28, 50) ### ### ('¹', 6, 142) ('intently', 8, 186) ('貴', 10, 208) ('幸', 12, 199) ('mohammed', 237, 13) ### ### ('খ', 13, 195) ('election', 53, 55) ('absently', 11, 234) ('ம', 17, 156) ### ############################################################################################################ [2023-11-25 18:06:52,288][root][INFO] - Epoch: 4: Step: 3301/19530, loss[v]=0.106674, lr=0.000009, acc@1[1]=996.0/1024=0.97265625, acc@1[2]=1012.5/1024=0.98876953125 [2023-11-25 18:08:40,673][root][INFO] - Train batch 3400 [2023-11-25 18:08:40,674][root][INFO] - Avg. loss per last 100 batches: 0.128098 [2023-11-25 18:08:41,374][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 3.98 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.9/29522=99.95% | mean: 0.03 | max: 5.95 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.1/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the sassanian empire was established in estakhr by ardashir i. ### ### [P_TEXT]: ardashir i of power in iran since the first days of the sasanians'uprising. in the ### ### sasanian dawn, iran included a union of kingdoms and noble landlords ( liegemen ), each of which ### ### possessed a various degree of independence from the central government and were economically ### ### connected to it by different channels. in other words, a type of feudal society under the rule of ### ### large owners stood in the iranian plateau while in the mesopotamia n deserts, the urban culture and ### ### pathway cities were the face of society more often. the first sasanian shahanshahs ### ### =============================================== v_q ================================================ ### ### ('est', 0, 17173) ('##hir', 1, 12) ('##das', 2, 1) ('intently', 3, 394) ('absently', 4, 213) ### ### ('貴', 5, 384) ('##akh', 6, 26024) ('¹', 7, 327) ('খ', 8, 224) ('sas', 9, 8) ('幸', 10, 344) ### ### ('ar', 11, 9) ('ம', 12, 272) ('##ப', 13, 293) ('.', 14, 107) ('sparhawk', 15, 703) ### ### ('##r', 16, 27977) ('i', 17, 31) ('caspian', 18, 4372) ('empire', 19, 331) ### ### =============================================== v_p ================================================ ### ### ('pathway', 26364, 0) ('##das', 2, 1) ('pathways', 25446, 2) ('plateau', 463, 3) ### ### ('liege', 10690, 4) ('mesopotamia', 78, 5) ('uprising', 2499, 6) ('##ans', 275, 7) ('sas', 9, 8) ### ### ('ar', 11, 9) ('iranian', 58, 10) ('feudal', 529, 11) ('##hir', 1, 12) ('path', 4535, 13) ### ### ('deserts', 6331, 14) ('union', 887, 15) ('kingdoms', 206, 16) ('dawn', 1227, 17) ### ### ('cities', 69, 18) ('iran', 25, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##das', 2, 1) ('##hir', 1, 12) ('sas', 9, 8) ('ar', 11, 9) ('i', 17, 31) ('iran', 25, 19) ### ### ('anyway', 31, 22) ('worried', 32, 26) ('iranian', 58, 10) ('mesopotamia', 78, 5) ### ### ('cities', 69, 18) ('persian', 57, 30) ('leaned', 76, 29) ('.', 14, 107) ('absently', 4, 213) ### ### ('##ans', 275, 7) ('খ', 8, 224) ('kingdoms', 206, 16) ('city', 60, 53) ('plateau', 463, 3) ### ############################################################################################################ [2023-11-25 18:08:41,878][root][INFO] - Epoch: 4: Step: 3401/19530, loss[v]=0.067849, lr=0.000009, acc@1[1]=1000.5/1024=0.97705078125, acc@1[2]=1021.0/1024=0.9970703125 [2023-11-25 18:10:30,384][root][INFO] - Train batch 3500 [2023-11-25 18:10:30,385][root][INFO] - Avg. loss per last 100 batches: 0.131898 [2023-11-25 18:10:31,104][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.6/29522=99.95% | mean: 0.02 | max: 5.76 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the series editor was asif siddiqi, the author of " challenge to apollo : the soviet ### ### union and the space race, 1945 - 1974 ". ### ### [P_TEXT]: asif azam siddiqi and recommend options for the future of nasa's human spaceflight ### ### program. siddiqi is also the editor of the series " rockets and people " which are the four volume ### ### english - language translation of the memoirs of boris chertok, a leading designer who worked under ### ### sergei korolev. these volumes are being published by the nasa history division. he was featured in ### ### the nova wgbh - tv special " astrospies " broadcast in 2008, and has been featured in various media ### ### as a specialist in the history of spaceflight. he has received the american historical ### ### association's fellowship in aerospace history, the eugene m. emme astronautical literature award, ### ### the ### ### =============================================== v_q ================================================ ### ### ('##if', 0, 7) ('challenge', 1, 15692) ('##qi', 2, 18) ('sid', 3, 16) ('¹', 4, 185) ### ### ('race', 5, 8220) ('apollo', 6, 1108) ('##di', 7, 29) ('intently', 8, 277) ('absently', 9, 139) ### ### ('幸', 10, 170) ('soviet', 11, 452) ('space', 12, 37) ('editor', 13, 42) ('貴', 14, 259) ### ### ('³', 15, 2214) ('খ', 16, 178) ('reviewer', 17, 1926) ('##ப', 18, 208) ('documentary', 19, 1574) ### ### =============================================== v_p ================================================ ### ### ('nova', 16212, 0) ('astros', 7427, 1) ('##gb', 26350, 2) ('##tok', 27908, 3) ('##pies', 9868, 4) ### ### ('fellowship', 2762, 5) ('##rol', 17727, 6) ('##if', 0, 7) ('boris', 4670, 8) ('rockets', 2303, 9) ### ### ('nasa', 34, 10) ('sergei', 4463, 11) ('memoirs', 590, 12) ('cher', 14730, 13) ### ### ('history', 2618, 14) ('historical', 3964, 15) ('sid', 3, 16) ('association', 2312, 17) ### ### ('##qi', 2, 18) ('aerospace', 281, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##if', 0, 7) ('sid', 3, 16) ('##qi', 2, 18) ('##di', 7, 29) ('space', 12, 37) ('nasa', 34, 10) ### ### ('editor', 13, 42) ('astronaut', 53, 25) ('″', 33, 58) ('rocket', 42, 47) ('series', 21, 94) ### ### ('¹', 4, 185) ('absently', 9, 139) ('aerospace', 281, 19) ('spacecraft', 44, 71) ### ### ('memoirs', 590, 12) ('memoir', 71, 59) ('幸', 10, 170) ('খ', 16, 178) ('volume', 94, 66) ### ############################################################################################################ [2023-11-25 18:10:31,611][root][INFO] - Epoch: 4: Step: 3501/19530, loss[v]=0.141746, lr=0.000009, acc@1[1]=991.5/1024=0.96826171875, acc@1[2]=1009.5/1024=0.98583984375 [2023-11-25 18:12:20,283][root][INFO] - Train batch 3600 [2023-11-25 18:12:20,284][root][INFO] - Avg. loss per last 100 batches: 0.133799 [2023-11-25 18:12:20,990][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.99 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.3/29522=99.95% | mean: 0.03 | max: 5.85 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.3/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the liturgy of the sacrament of the eucharist, called the mass in the west and divine ### ### liturgy or other names in the east, is the principal liturgy of the catholic church. ### ### [P_TEXT]: eucharist corbie, whose most well - known and influential work is an exposition on the ### ### nature of the eucharist written around 831, entitled " de corpore et sanguine domini ". he was ### ### canonized in 1073 by pope gregory vii. his works are edited in " patrologia latina " vol. 120 ( ### ### 1852 ). most christians, even those who deny that there is any real change in the elements used, ### ### recognize a special presence of christ in this rite. but christians differ about exactly how, where ### ### and how long christ is present in it. catholicism, eastern orthodoxy, ### ### =============================================== v_q ================================================ ### ### ('mass', 0, 1207) ('liturgy', 1, 471) ('absently', 2, 137) ('catholic', 3, 156) ('貴', 4, 446) ### ### ('sacrament', 5, 24) ('intently', 6, 535) ('.', 7, 164) ('幸', 8, 296) ('¹', 9, 154) ('ம', 10, 198) ### ### ('##ப', 11, 237) ('খ', 12, 230) ('church', 13, 98) ('worried', 14, 58) ('rite', 15, 20) ### ### ('##rist', 16, 2) ('divine', 17, 157) ('anyway', 18, 36) ('##cha', 19, 6) ### ### =============================================== v_p ================================================ ### ### ('patrol', 17296, 0) ('##rb', 24163, 1) ('##rist', 16, 2) ('eu', 25, 3) ('gregory', 3357, 4) ### ### ('latin', 57, 5) ('##cha', 19, 6) ('1852', 13577, 7) ('107', 12270, 8) ('catholicism', 354, 9) ### ### ('christ', 257, 10) ('leaned', 1481, 11) ('exposition', 6383, 12) ('83', 5492, 13) ### ### ('elements', 857, 14) ('dom', 449, 15) ('corp', 16252, 16) ('120', 6130, 17) ### ### ('christians', 503, 18) ('vii', 7564, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##rist', 16, 2) ('sacrament', 5, 24) ('rite', 15, 20) ('##cha', 19, 6) ('eu', 25, 3) ### ### ('worried', 14, 58) ('anyway', 18, 36) ('latin', 57, 5) ('absently', 2, 137) ('catholic', 3, 156) ### ### ('.', 7, 164) ('church', 13, 98) ('canon', 48, 26) ('¹', 9, 154) ('liturgical', 32, 55) ### ### ('ம', 10, 198) ('liturgy', 1, 471) ('eastern', 50, 40) ('##ப', 11, 237) ('christ', 257, 10) ### ############################################################################################################ [2023-11-25 18:12:21,496][root][INFO] - Epoch: 4: Step: 3601/19530, loss[v]=0.167039, lr=0.000009, acc@1[1]=991.0/1024=0.9677734375, acc@1[2]=1011.5/1024=0.98779296875 [2023-11-25 18:14:09,944][root][INFO] - Train batch 3700 [2023-11-25 18:14:09,944][root][INFO] - Avg. loss per last 100 batches: 0.125784 [2023-11-25 18:14:10,652][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 3.94 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.7/29522=99.95% | mean: 0.02 | max: 5.89 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.2/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: theron martin of anime news network labelled the " hestia factor " as one of the reasons ### ### for " danmachi " s success, citing inori minase's voice acting, character development, as well ### ### [P_TEXT]: inori minase rabbit? ", noel in " celestial method ", hestia in " is it wrong to try to ### ### pick up girls in a dungeon? ", rem in " ", chito in " girls'last tour ", shoko makinohara in " ### ### rascal does not dream of bunny girl senpai " and " rascal does not dream of a dreaming girl ", ### ### alongside tamaki " kimari " mari in " a place further than the universe ". minase was born in tokyo ### ### on december 2, 1995. she had begun watching anime when she was in ### ### =============================================== v_q ================================================ ### ### ('##sti', 0, 2) ('minas', 1, 0) ('¹', 2, 87) ('dan', 3, 18445) ('##ori', 4, 5) ('##mac', 5, 27345) ### ### ('sailor', 6, 14800) ('³', 7, 467) ('anime', 8, 27) ('factor', 9, 18454) ('##hi', 10, 14644) ### ### ('貴', 11, 171) ('″', 12, 23) ('幸', 13, 111) ('voice', 14, 1900) ('##ron', 15, 24041) ### ### ('absently', 16, 89) ('##ப', 17, 114) ('intently', 18, 157) ('খ', 19, 134) ### ### =============================================== v_p ================================================ ### ### ('minas', 1, 0) ('noel', 5253, 1) ('##sti', 0, 2) ('celestial', 2226, 3) ('dungeon', 1170, 4) ### ### ('##ori', 4, 5) ('dreaming', 20221, 6) ('mari', 3206, 7) ('bunny', 13327, 8) ('##ari', 1657, 9) ### ### ('method', 4341, 10) ('wrong', 4771, 11) ('tour', 9431, 12) ('##oh', 1852, 13) ('rabbit', 799, 14) ### ### ('chi', 3811, 15) ('##aki', 1230, 16) ('girls', 426, 17) ('universe', 86, 18) ('tam', 6318, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('minas', 1, 0) ('##sti', 0, 2) ('##ori', 4, 5) ('″', 12, 23) ('anime', 8, 27) ('¹', 2, 87) ### ### ('japanese', 31, 41) ('universe', 86, 18) ('absently', 16, 89) ('"', 26, 60) ('幸', 13, 111) ### ### ('sequel', 102, 37) ('characters', 43, 51) ('japan', 41, 55) ('##ப', 17, 114) ('hissed', 67, 48) ### ### ('girls', 426, 17) ('girl', 372, 20) ('貴', 11, 171) ('ம', 22, 104) ### ############################################################################################################ [2023-11-25 18:14:11,157][root][INFO] - Epoch: 4: Step: 3701/19530, loss[v]=0.061663, lr=0.000009, acc@1[1]=1005.0/1024=0.9814453125, acc@1[2]=1017.0/1024=0.9931640625 [2023-11-25 18:16:00,107][root][INFO] - Train batch 3800 [2023-11-25 18:16:00,108][root][INFO] - Avg. loss per last 100 batches: 0.133381 [2023-11-25 18:16:00,830][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.2/29522=99.95% | mean: 0.02 | max: 5.77 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it was completed between 1908 - 1910, after the french bombardment and occupation of ### ### casablanca but before the treaty of fes in 1912, which officially established the french ### ### protectorate. ### ### [P_TEXT]: united nations square ( casablanca ) united nations square (, ) is a public square in the ### ### center of casablanca, morocco. it has been central in the history of casablanca. the area outside ### ### the walls of the old medina that is now united nations square, used to be the location of the " ### ### souq kbir " ( ), also referred to as, before french colonization. in 1908, after the french ### ### bombardment and invasion of casablanca, the french commander charles martial joseph dessigny ### ### ordered the construction of a clock tower in ### ### =============================================== v_q ================================================ ### ### ('casablanca', 0, 1) ('fe', 1, 14494) ('protectorate', 2, 2346) ('treaty', 3, 237) ('¹', 4, 927) ### ### ('bombardment', 5, 14) ('french', 6, 18) ('sahara', 7, 1421) ('খ', 8, 299) ('morocco', 9, 11) ### ### ('貴', 10, 319) ('.', 11, 233) ('intently', 12, 264) ('occupation', 13, 94) ('ம', 14, 187) ### ### ('1912', 15, 64) ('absently', 16, 259) ('³', 17, 5573) ('幸', 18, 268) ('##ப', 19, 438) ### ### =============================================== v_p ================================================ ### ### ('medina', 1955, 0) ('casablanca', 0, 1) ('##gny', 10831, 2) ('martial', 5152, 3) ### ### ('square', 998, 4) ('kb', 8967, 5) ('clock', 3816, 6) ('worried', 387, 7) ('nations', 292, 8) ### ### ('colonization', 786, 9) ('walls', 284, 10) ('morocco', 9, 11) ('anyway', 50, 12) ('1908', 27, 13) ### ### ('bombardment', 5, 14) ('invasion', 399, 15) ('invaded', 28167, 16) ('moroccan', 70, 17) ### ### ('french', 6, 18) ('##si', 8895, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('casablanca', 0, 1) ('morocco', 9, 11) ('bombardment', 5, 14) ('french', 6, 18) ('1908', 27, 13) ### ### ('anyway', 50, 12) ('1912', 15, 64) ('moroccan', 70, 17) ('france', 28, 43) ('occupation', 13, 94) ### ### ('nations', 292, 8) ('walls', 284, 10) ('worried', 387, 7) ('treaty', 3, 237) ('invasion', 399, 15) ### ### ('medina', 1955, 0) ('ம', 14, 187) ('square', 998, 4) ('colonization', 786, 9) ('1910', 20, 172) ### ############################################################################################################ [2023-11-25 18:16:01,337][root][INFO] - Epoch: 4: Step: 3801/19530, loss[v]=0.160239, lr=0.000009, acc@1[1]=976.0/1024=0.953125, acc@1[2]=1006.5/1024=0.98291015625 [2023-11-25 18:17:50,384][root][INFO] - Train batch 3900 [2023-11-25 18:17:50,385][root][INFO] - Avg. loss per last 100 batches: 0.131120 [2023-11-25 18:17:51,090][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.4/29522=99.95% | mean: 0.02 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: on may 12, 2008, he exchanged wedding vows with rapper remy ma over the phone while she ### ### was incarcerated. ### ### [P_TEXT]: remy ma she was joining'for its sixth season, along with her husband papoose. she has ### ### continued to appear on the show as a series regular in seasons seven, eight, nine and ten, as well ### ### as starring in the spin - off show ', and the specials " remy & papoose : a merry mackie holiday ", ### ### " dirty little secrets ", " the love edition ", " love & hip hop awards : most certified " and " 40 ### ### greatest love & hip hop moments : the reboot ". in 2016, fat joe revealed that he and ma would be ### ### releasing a joint album. on march 2, 2016, the joint ### ### =============================================== v_q ================================================ ### ### ('remy', 0, 0) ('ma', 1, 2) ('wedding', 2, 1234) ('¹', 3, 238) ('incarcerated', 4, 25953) ### ### ('ம', 5, 106) ('³', 6, 1217) ('freaking', 7, 983) ('celebrities', 8, 372) ('貴', 9, 192) ### ### ('kendrick', 10, 14576) ('prison', 11, 4435) ('rapper', 12, 84) ('kanye', 13, 19935) ### ### ('inmate', 14, 27905) ('2008', 15, 5435) ('！', 16, 993) ('vow', 17, 15445) ('absently', 18, 89) ### ### ('bubba', 19, 7828) ### ### =============================================== v_p ================================================ ### ### ('remy', 0, 0) ('secrets', 10193, 1) ('ma', 1, 2) ('edition', 2285, 3) ('##poo', 15837, 4) ### ### ('fat', 169, 5) ('joe', 1112, 6) ('mack', 8099, 7) ('dirty', 266, 8) ('specials', 4493, 9) ### ### ('##boot', 2372, 10) ('merry', 19619, 11) ('love', 1788, 12) ('awards', 1421, 13) ### ### ('certification', 4945, 14) ('sixth', 7378, 15) ('holiday', 3566, 16) ('show', 985, 17) ### ### ('nine', 16986, 18) ('certified', 3809, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('remy', 0, 0) ('ma', 1, 2) ('ceremony', 47, 23) ('fat', 169, 5) ('revealed', 94, 22) ### ### ('dirty', 266, 8) ('rapper', 12, 84) ('ம', 5, 106) ('absently', 18, 89) ('¹', 3, 238) ### ### ('hip', 110, 39) ('貴', 9, 192) ('joe', 1112, 6) ('rappers', 71, 65) ('hissed', 366, 32) ### ### ('##ப', 21, 128) ('2016', 609, 20) ('hop', 281, 38) ('herself', 95, 66) ('edition', 2285, 3) ### ############################################################################################################ [2023-11-25 18:17:51,594][root][INFO] - Epoch: 4: Step: 3901/19530, loss[v]=0.074823, lr=0.000009, acc@1[1]=996.5/1024=0.97314453125, acc@1[2]=1015.0/1024=0.9912109375 [2023-11-25 18:19:40,212][root][INFO] - Train batch 4000 [2023-11-25 18:19:40,212][root][INFO] - Avg. loss per last 100 batches: 0.129643 [2023-11-25 18:19:40,919][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.6/29522=99.95% | mean: 0.02 | max: 5.84 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: this core is further embellished with goalpariya and kamrupi idioms and forms. ### ### [P_TEXT]: miya people na - asamiyas were honest in their intent to learn the assamese language and ### ### send their children to assamese - medium schools. the migrant muslims thus came to be known as the ### ### na - asamiyas. to this day, there is not a single bengali - medium school in the areas dominated by ### ### the na - asamiyas. with the official acceptance of assamese language and close proximity to ### ### goalpariya speakers, the miya have gradually also developed a creole miya home language. the miya ### ### officially adopted the assamese language during the language movement. in formal and official ### ### meetings, they ### ### =============================================== v_q ================================================ ### ### ('goal', 0, 5) ('##par', 1, 8) ('kam', 2, 18253) ('##rup', 3, 26025) ('##iya', 4, 14) ('貴', 5, 288) ### ### ('intently', 6, 280) ('¹', 7, 851) ('nepal', 8, 62) ('幸', 9, 252) ('খ', 10, 275) ('id', 11, 18498) ### ### ('ம', 12, 168) ('##ப', 13, 259) ('.', 14, 744) ('proficient', 15, 1976) ('absently', 16, 132) ### ### ('³', 17, 2238) ('bengali', 18, 15) ('medium', 19, 22) ### ### =============================================== v_p ================================================ ### ### ('mi', 16643, 0) ('assam', 69, 1) ('##miya', 4531, 2) ('##ya', 341, 3) ('creole', 510, 4) ### ### ('goal', 0, 5) ('muslims', 10242, 6) ('##ese', 500, 7) ('##par', 1, 8) ('migrants', 27897, 9) ### ### ('language', 26, 10) ('asa', 3842, 11) ('movement', 630, 12) ('anyway', 20, 13) ('##iya', 4, 14) ### ### ('bengali', 18, 15) ('muslim', 4276, 16) ('speakers', 7371, 17) ('schooling', 2487, 18) ### ### ('leaned', 3517, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('goal', 0, 5) ('##par', 1, 8) ('##iya', 4, 14) ('bengali', 18, 15) ('anyway', 20, 13) ### ### ('language', 26, 10) ('assam', 69, 1) ('medium', 19, 22) ('worried', 30, 25) ('nepal', 8, 62) ### ### ('languages', 34, 24) ('##ya', 341, 3) ('creole', 510, 4) ('##ese', 500, 7) ('bangladesh', 23, 79) ### ### ('absently', 16, 132) ('ம', 12, 168) ('intently', 6, 280) ('貴', 5, 288) ('幸', 9, 252) ### ############################################################################################################ [2023-11-25 18:19:41,424][root][INFO] - Epoch: 4: Step: 4001/19530, loss[v]=0.168404, lr=0.000009, acc@1[1]=984.5/1024=0.96142578125, acc@1[2]=1006.5/1024=0.98291015625 [2023-11-25 18:21:30,508][root][INFO] - Train batch 4100 [2023-11-25 18:21:30,509][root][INFO] - Avg. loss per last 100 batches: 0.133646 [2023-11-25 18:21:31,270][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29508.3/29522=99.95% | mean: 0.02 | max: 5.81 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: crofton locks are a flight of locks on the kennet and avon canal, near the village of ### ### great bedwyn, wiltshire, england. ### ### [P_TEXT]: kennet and avon canal centre of reading. wilton water by crofton locks and the kennet ### ### valley gravel pit s provide habitats for breeding and wintering waterfowl. several species of ### ### odonata ( dragonflies and damselflies ) and other invertebrates have also been recorded. common ### ### reed ( " phragmites australis " ) is among the plant species growing along the edges of the canal. ### ### measures to preserve and create water vole ( " arvicola amphibius or a. terrestrisis " ) habitat ### ### have had considerable impact on the restoration of the canal, and new " vole - friendly " ### ### techniques of bank protection ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 565) ('canal', 1, 9) ('croft', 2, 2) ('locks', 3, 24) ('bed', 4, 4585) ### ### ('wiltshire', 5, 3346) ('³', 6, 2140) ('##on', 7, 69) ('curious', 8, 323) ('##wyn', 9, 24920) ### ### ('##net', 10, 4) ('lock', 11, 83) ('খ', 12, 195) ('great', 13, 16285) ('ம', 14, 147) ('幸', 15, 182) ### ### ('ken', 16, 12) ('luckily', 17, 74) ('avon', 18, 6) ('derelict', 19, 161) ### ### =============================================== v_p ================================================ ### ### ('wilton', 687, 0) ('reed', 3717, 1) ('croft', 2, 2) ('##flies', 9908, 3) ('##net', 10, 4) ### ### ('reading', 3437, 5) ('avon', 18, 6) ('gravel', 2819, 7) ('vol', 10567, 8) ('canal', 1, 9) ### ### ('terre', 18863, 10) ('##vic', 27277, 11) ('ken', 16, 12) ('pit', 8131, 13) ('##ris', 21067, 14) ### ### ('dams', 1851, 15) ('##rag', 22594, 16) ('protection', 2072, 17) ('##don', 6487, 18) ### ### ('water', 59, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('croft', 2, 2) ('canal', 1, 9) ('##net', 10, 4) ('locks', 3, 24) ('avon', 18, 6) ('ken', 16, 12) ### ### ('##on', 7, 69) ('water', 59, 19) ('lock', 11, 83) ('wilton', 687, 0) ('luckily', 17, 74) ### ### ('common', 68, 38) ('hissed', 100, 37) ('¹', 0, 565) ('ம', 14, 147) ('খ', 12, 195) ('幸', 15, 182) ### ### ('curious', 8, 323) ('derelict', 19, 161) ('churchyard', 58, 78) ### ############################################################################################################ [2023-11-25 18:21:31,779][root][INFO] - Epoch: 4: Step: 4101/19530, loss[v]=0.137210, lr=0.000009, acc@1[1]=992.0/1024=0.96875, acc@1[2]=1011.5/1024=0.98779296875 [2023-11-25 18:23:20,686][root][INFO] - Train batch 4200 [2023-11-25 18:23:20,687][root][INFO] - Avg. loss per last 100 batches: 0.137320 [2023-11-25 18:23:21,390][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.82 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.6/29522=99.95% | mean: 0.02 | max: 5.68 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the mortlake tapestry works outside london were founded ### ### [P_TEXT]: lyme park stone, simulating marble. the entrance hall, which is in the east range, was ### ### remodelled by leoni. it is asymmetrical and contains giant pilasters and a screen of three fluted ### ### ionic columns. the doorway to the courtyard has an open pediment. a hinged picture can be swung out ### ### from the wall to reveal a squint looking into the entrance hall. also in the entrance hall are ### ### tapestries which were woven at the mortlake tapestry works between 1623 and 1636. they were ### ### originally in the leghs'london home in belgrave square ### ### =============================================== v_q ================================================ ### ### ('##lake', 0, 2) ('mort', 1, 7) ('tapestry', 2, 6) ('¹', 3, 596) ('works', 4, 31) ('london', 5, 35) ### ### ('intently', 6, 293) ('³', 7, 2411) ('幸', 8, 244) ('貴', 9, 296) ('খ', 10, 204) ### ### ('fireplace', 11, 1161) ('depict', 12, 1231) ('##ப', 13, 301) ('factory', 14, 655) ### ### ('relic', 15, 386) ('ம', 16, 153) ('workshop', 17, 584) ('modelled', 18, 3666) ('cloth', 19, 65) ### ### =============================================== v_p ================================================ ### ### ('##yme', 13045, 0) ('ionic', 20659, 1) ('##lake', 0, 2) ('leon', 14663, 3) ('##grave', 1696, 4) ### ### ('tapes', 20499, 5) ('tapestry', 2, 6) ('mort', 1, 7) ('##hs', 20433, 8) ('1623', 23701, 9) ### ### ('leg', 8686, 10) ('bel', 18008, 11) ('##tries', 143, 12) ('##uin', 13232, 13) ('sq', 2831, 14) ### ### ('park', 2303, 15) ('163', 13727, 16) ('marble', 176, 17) ('screen', 404, 18) ('square', 526, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##lake', 0, 2) ('mort', 1, 7) ('tapestry', 2, 6) ('works', 4, 31) ('london', 5, 35) ### ### ('##tries', 143, 12) ('marble', 176, 17) ('cloth', 19, 65) ('anyway', 138, 36) ('screen', 404, 18) ### ### ('flute', 232, 30) ('square', 526, 19) ('weaving', 22, 83) ('##grave', 1696, 4) ('¹', 3, 596) ### ### ('幸', 8, 244) ('intently', 6, 293) ('খ', 10, 204) ('ம', 16, 153) ('woven', 839, 25) ### ############################################################################################################ [2023-11-25 18:23:21,894][root][INFO] - Epoch: 4: Step: 4201/19530, loss[v]=0.144311, lr=0.000009, acc@1[1]=978.5/1024=0.95556640625, acc@1[2]=1012.0/1024=0.98828125 [2023-11-25 18:25:10,604][root][INFO] - Train batch 4300 [2023-11-25 18:25:10,604][root][INFO] - Avg. loss per last 100 batches: 0.131669 [2023-11-25 18:25:11,318][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.3/29522=99.94% | mean: 0.02 | max: 5.58 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: churchill, valentine, and " cromwell " models as the vickers tank periscope mk. iv. ### ### [P_TEXT]: t - 70 with two 85 - hp gaz - 203 engines, a mark 4 commander's periscope replacing a ### ### vision slit, and other improvements. the one - man turret of the soviet light tanks made co - ### ### ordinating a tank platoon nearly impossible, because the commanders were kept busy acquiring ### ### targets, loading and firing the main gun and machine gun, and commanding their drivers. the ### ### infantry tank role was already considered obsolete. the su - 76 self - propelled gun was better ### ### suited for infantry support, its 76. 2 - mm gun capable of firing a larger high explosive shell. ### ### industrial resources could be redirected from light tanks to ### ### =============================================== v_q ================================================ ### ### ('cromwell', 0, 3285) ('##cope', 1, 5) ('churchill', 2, 1356) ('valentine', 3, 13830) ### ### ('vickers', 4, 1730) ('.', 5, 133) ('tank', 6, 10) ('iv', 7, 9685) ('##is', 8, 63) ('খ', 9, 181) ### ### ('absently', 10, 202) ('貴', 11, 297) ('幸', 12, 175) ('##ப', 13, 213) ('ம', 14, 159) ### ### ('turret', 15, 29) ('intently', 16, 355) ('¹', 17, 203) ('per', 18, 18) ('british', 19, 7413) ### ### =============================================== v_p ================================================ ### ### ('70', 15274, 0) ('slit', 3054, 1) ('203', 9496, 2) ('vision', 5631, 3) ('platoon', 1416, 4) ### ### ('##cope', 1, 5) ('t', 12675, 6) ('su', 24726, 7) ('76', 15861, 8) ('##dina', 12855, 9) ### ### ('tank', 6, 10) ('turrets', 1825, 11) ('explosives', 11231, 12) ('commander', 307, 13) ### ### ('gun', 108, 14) ('explosive', 3486, 15) ('engines', 565, 16) ('engine', 115, 17) ('per', 18, 18) ### ### ('tanks', 41, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##cope', 1, 5) ('tank', 6, 10) ('per', 18, 18) ('turret', 15, 29) ('##is', 8, 63) ### ### ('tanks', 41, 19) ('anyway', 29, 32) ('gun', 108, 14) ('.', 5, 133) ('engine', 115, 17) ### ### ('mm', 192, 22) ('commander', 307, 13) ('খ', 9, 181) ('shell', 230, 25) ('soviet', 276, 21) ### ### ('absently', 10, 202) ('幸', 12, 175) ('guns', 204, 34) ('loading', 73, 46) ('ம', 14, 159) ### ############################################################################################################ [2023-11-25 18:25:11,826][root][INFO] - Epoch: 4: Step: 4301/19530, loss[v]=0.155124, lr=0.000009, acc@1[1]=978.0/1024=0.955078125, acc@1[2]=1007.5/1024=0.98388671875 [2023-11-25 18:27:00,730][root][INFO] - Train batch 4400 [2023-11-25 18:27:00,731][root][INFO] - Avg. loss per last 100 batches: 0.140369 [2023-11-25 18:27:01,440][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.4/29522=99.95% | mean: 0.02 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: at the 1989 world rowing championships at bled, yugoslavia, he won a bronze in the men's ### ### four with ian wright, alastair mackintosh, and bill coventry. ### ### [P_TEXT]: bill coventry william michael coventry ( born 26 january 1967 ) is a former new zealand ### ### rower. coventry was born in hamilton in 1967. he represented new zealand at the 1988 summer ### ### olympics in the coxless four in a team with campbell clayton - greene, geoff cotter, and neil ### ### gibson, where they came seventh. at the 1989 world rowing championships at bled, yugoslavia, he won ### ### a bronze in the men's four with ian wright, alastair mackintosh, and ### ### =============================================== v_q ================================================ ### ### ('coventry', 0, 0) ('ian', 1, 14) ('wright', 2, 1) ('littered', 3, 26) ('rowing', 4, 6) ### ### ('ɐ', 5, 236) ('boat', 6, 48) ('rower', 7, 42) ('##int', 8, 22) ('##sta', 9, 36) ('1989', 10, 4) ### ### ('bill', 11, 5) ('pmid', 12, 607) ('¹', 13, 150) ('.', 14, 641) ('貴', 15, 125) ('খ', 16, 135) ### ### ('absently', 17, 136) ('rowe', 18, 372) ('##ப', 19, 127) ### ### =============================================== v_p ================================================ ### ### ('coventry', 0, 0) ('wright', 2, 1) ('bled', 34, 2) ('gibson', 6117, 3) ('1989', 10, 4) ### ### ('bill', 11, 5) ('rowing', 4, 6) ('neil', 74, 7) ('cot', 15166, 8) ('geoff', 68, 9) ### ### ('yugoslavia', 87, 10) ('leaned', 1116, 11) ('1988', 44, 12) ('hamilton', 1476, 13) ('ian', 1, 14) ### ### ('zealand', 299, 15) ('##osh', 21, 16) ('campbell', 587, 17) ('worried', 1017, 18) ### ### ('##less', 12374, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('coventry', 0, 0) ('wright', 2, 1) ('rowing', 4, 6) ('1989', 10, 4) ('ian', 1, 14) ('bill', 11, 5) ### ### ('littered', 3, 26) ('##int', 8, 22) ('bled', 34, 2) ('##osh', 21, 16) ('##sta', 9, 36) ### ### ('championships', 27, 24) ('1988', 44, 12) ('ala', 36, 23) ('rower', 7, 42) ('fours', 29, 35) ### ### ('neil', 74, 7) ('geoff', 68, 9) ('boat', 6, 48) ('men', 35, 37) ### ############################################################################################################ [2023-11-25 18:27:01,947][root][INFO] - Epoch: 4: Step: 4401/19530, loss[v]=0.194525, lr=0.000009, acc@1[1]=972.5/1024=0.94970703125, acc@1[2]=1000.5/1024=0.97705078125 [2023-11-25 18:28:50,350][root][INFO] - Train batch 4500 [2023-11-25 18:28:50,350][root][INFO] - Avg. loss per last 100 batches: 0.135938 [2023-11-25 18:28:51,039][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.5/29522=100.00% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.2/29522=99.95% | mean: 0.02 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.6/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: on both sides were eager to avoid an open battle, so members of the clergy brokered a ### ### truce, to the annoyance of both henry and stephen. ### ### [P_TEXT]: treaty of wallingford the treaty of wallingford, also known as the treaty of winchester ### ### or the treaty of westminster, was an agreement reached in england in the summer of 1153. it ### ### effectively ended a civil war known as " the anarchy " ( 1135 – 54 ), caused by a dispute over the ### ### english crown between king stephen and his cousin empress matilda. the treaty of wallingford ### ### allowed stephen to keep the throne until his death ( which was to come in october 1154 ), but ### ### ensured that he would be succeeded by matilda's son henry ii. in 1153, ### ### =============================================== v_q ================================================ ### ### ('stephen', 0, 7) ('battle', 1, 750) ('henry', 2, 14) ('intently', 3, 192) ('貴', 4, 139) ### ### ('truce', 5, 25) ('sparhawk', 6, 2056) ('absently', 7, 174) ('খ', 8, 132) ('.', 9, 216) ### ### ('##ப', 10, 134) ('clergy', 11, 1429) ('幸', 12, 156) ('ம', 13, 141) ('anyway', 14, 15) ### ### ('¹', 15, 178) ('open', 16, 6796) ('³', 17, 2846) ('canterbury', 18, 1347) ('peace', 19, 119) ### ### =============================================== v_p ================================================ ### ### ('matilda', 439, 0) ('westminster', 1263, 1) ('treaty', 114, 2) ('winchester', 126, 3) ### ### ('anarchy', 1642, 4) ('wall', 743, 5) ('115', 2131, 6) ('stephen', 0, 7) ('worried', 35, 8) ### ### ('##ford', 2797, 9) ('113', 4303, 10) ('treaties', 18746, 11) ('cousin', 7123, 12) ### ### ('succeeded', 3502, 13) ('henry', 2, 14) ('anyway', 14, 15) ('civil', 1369, 16) ### ### ('leaned', 7043, 17) ('54', 18622, 18) ('##ing', 10637, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('stephen', 0, 7) ('henry', 2, 14) ('anyway', 14, 15) ('truce', 5, 25) ('worried', 35, 8) ### ### ('treaty', 114, 2) ('winchester', 126, 3) ('wars', 28, 23) ('matilda', 439, 0) ('england', 24, 26) ### ### ('medieval', 23, 30) ('貴', 4, 139) ('intently', 3, 192) ('wall', 743, 5) ('খ', 8, 132) ### ### ('##ப', 10, 134) ('westminster', 1263, 1) ('english', 57, 29) ('absently', 7, 174) ('ம', 13, 141) ### ############################################################################################################ [2023-11-25 18:28:51,542][root][INFO] - Epoch: 4: Step: 4501/19530, loss[v]=0.101893, lr=0.000009, acc@1[1]=1003.0/1024=0.9794921875, acc@1[2]=1009.0/1024=0.9853515625 [2023-11-25 18:30:39,907][root][INFO] - Train batch 4600 [2023-11-25 18:30:39,908][root][INFO] - Avg. loss per last 100 batches: 0.133246 [2023-11-25 18:30:40,616][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.2/29522=99.93% | mean: 0.02 | max: 5.61 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it was established in august 1915 when the northwestern front was split into the northern ### ### and western front following the great retreat, and existed until the demobilization of the russian ### ### army in 1918 ### ### [P_TEXT]: brusilov offensive the emperor approved brusilov's plan. the offensive aimed to capture ### ### the cities of kovel and lviv ( in present - day western ukraine ) ; the central powers had ### ### recovered both these cities in 1915. although the stavka had approved brusilov's plan, his request ### ### for supporting offensives by the neighboring front s ( the western under evert and northern under ### ### aleksey kuropatkin ) was denied. mounting pressure from the western allies caused the russians to ### ### hurry their preparations. brusilov amassed four armies totaling 40 infantry divisions and 15 ### ### cavalry divisions. he faced 39 austrian infantry divisions ### ### =============================================== v_q ================================================ ### ### ('northwestern', 0, 1419) ('western', 1, 20) ('retreat', 2, 1362) ('front', 3, 5) ('¹', 4, 205) ### ### ('northern', 5, 13) ('russian', 6, 33) ('খ', 7, 174) ('貴', 8, 192) ('intently', 9, 159) ### ### ('幸', 10, 173) ('1915', 11, 7) ('great', 12, 11349) ('1918', 13, 157) ('ம', 14, 150) ### ### ('³', 15, 1798) ('##ப', 16, 188) ('absently', 17, 265) ('army', 18, 78) ('petersburg', 19, 3764) ### ### =============================================== v_p ================================================ ### ### ('##vel', 20316, 0) ('##ilo', 25538, 1) ('ever', 18279, 2) ('offensive', 294, 3) ('lviv', 2819, 4) ### ### ('front', 3, 5) ('littered', 1352, 6) ('1915', 11, 7) ('##av', 2960, 8) ('russians', 192, 9) ### ### ('worried', 60, 10) ('##rop', 28752, 11) ('##kin', 10519, 12) ('northern', 5, 13) ### ### ('ukraine', 747, 14) ('austrians', 20538, 15) ('ko', 10807, 16) ('anyway', 20, 17) ### ### ('##t', 17457, 18) ('fronts', 857, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('front', 3, 5) ('1915', 11, 7) ('western', 1, 20) ('northern', 5, 13) ('russian', 6, 33) ### ### ('anyway', 20, 17) ('1914', 37, 27) ('worried', 60, 10) ('army', 18, 78) ('intently', 9, 159) ### ### ('offensive', 294, 3) ('¹', 4, 205) ('খ', 7, 174) ('ம', 14, 150) ('1918', 13, 157) ('幸', 10, 173) ### ### ('armies', 58, 32) ('russians', 192, 9) ('貴', 8, 192) ('##ப', 16, 188) ### ############################################################################################################ [2023-11-25 18:30:41,123][root][INFO] - Epoch: 4: Step: 4601/19530, loss[v]=0.194337, lr=0.000009, acc@1[1]=979.0/1024=0.9560546875, acc@1[2]=1001.5/1024=0.97802734375 [2023-11-25 18:32:30,327][root][INFO] - Train batch 4700 [2023-11-25 18:32:30,327][root][INFO] - Avg. loss per last 100 batches: 0.123546 [2023-11-25 18:32:31,048][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.78 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.6/29522=99.94% | mean: 0.02 | max: 5.70 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the dukedom takes its name from the town where windsor castle, a residence of english ### ### monarchs since the time of henry i, following the norman conquest, is situated. ### ### [P_TEXT]: windsor castle did not spend much of his reign at windsor castle. he continued to spend ### ### most of his time at fort belvedere in the great park, where he had lived whilst prince of wales. ### ### edward created a small aerodrome at the castle on smith's lawn, now used as a golf - course. ### ### edward's reign was short - lived and he broadcast his abdication speech to the british empire from ### ### the castle in december 1936, adopting the title of duke of windsor. his successor, george vi also ### ### preferred his own original ### ### =============================================== v_q ================================================ ### ### ('windsor', 0, 0) ('castle', 1, 1) ('duke', 2, 18) ('.', 3, 124) ('¹', 4, 86) ('absently', 5, 119) ### ### ('##dom', 6, 27152) ('খ', 7, 161) ('貴', 8, 234) ('intently', 9, 226) ('henry', 10, 7675) ### ### ('ம', 11, 121) ('england', 12, 1721) ('##ப', 13, 196) ('幸', 14, 148) ('³', 15, 427) ### ### ('i', 16, 20652) ('и', 17, 4043) ('##部', 18, 219) ('english', 19, 3922) ### ### =============================================== v_p ================================================ ### ### ('windsor', 0, 0) ('castle', 1, 1) ('lawn', 4177, 2) ('aerodrome', 16872, 3) ('##ved', 14099, 4) ### ### ('park', 1283, 5) ('smith', 13189, 6) ('speech', 2977, 7) ('great', 801, 8) ('fort', 377, 9) ### ### ('whilst', 109, 10) ('abd', 25135, 11) ('1936', 14054, 12) ('courses', 8637, 13) ### ### ('course', 6206, 14) ('dukes', 267, 15) ('##ere', 17806, 16) ('golf', 2186, 17) ('duke', 2, 18) ### ### ('bel', 16401, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('windsor', 0, 0) ('castle', 1, 1) ('duke', 2, 18) ('worried', 27, 20) ('¹', 4, 86) ### ### ('whilst', 109, 10) ('.', 3, 124) ('absently', 5, 119) ('curious', 29, 41) ('anyway', 67, 23) ### ### ('ம', 11, 121) ('খ', 7, 161) ('indeed', 54, 34) ('monarch', 61, 29) ('fort', 377, 9) ('幸', 14, 148) ### ### ('dukes', 267, 15) ('leaned', 91, 33) ('intently', 9, 226) ('貴', 8, 234) ### ############################################################################################################ [2023-11-25 18:32:31,553][root][INFO] - Epoch: 4: Step: 4701/19530, loss[v]=0.084016, lr=0.000009, acc@1[1]=997.0/1024=0.9736328125, acc@1[2]=1018.0/1024=0.994140625 [2023-11-25 18:34:20,423][root][INFO] - Train batch 4800 [2023-11-25 18:34:20,424][root][INFO] - Avg. loss per last 100 batches: 0.134898 [2023-11-25 18:34:21,130][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.2/29522=99.95% | mean: 0.02 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: gerald linford beckley ( born september 12, 1952 ) is an america n singer, songwriter, ### ### and musician, and a founding member of the band america. ### ### [P_TEXT]: til i hear it from you issued january 23, 1996 as the headline track on a four - track cd ### ### maxi single which featured " til i hear it from you " plus two tracks from the " til i hear it from ### ### you " cd maxi single release of september 1995 : " seeing stars " and " idiot summer ". ( " idiot ### ### summer " was also included on the soundtrack to " wayne's world 2 ". ) america remade " til i hear ### ### it from you " for their 2012 release, " back pages ", a cover album that according to group member ### ### gerry beckley comprises " killer songs that are great examples that come from our ### ### =============================================== v_q ================================================ ### ### ('america', 0, 12) ('beck', 1, 5) ('¹', 2, 329) ('##ley', 3, 15) ('gerald', 4, 1934) ### ### ('lin', 5, 17950) ('³', 6, 742) ('band', 7, 84) ('幸', 8, 110) ('##ப', 9, 137) ('ம', 10, 98) ### ### ('খ', 11, 120) ('stevie', 12, 18879) ('intently', 13, 107) ('musician', 14, 704) ### ### ('absently', 15, 56) ('##ford', 16, 23734) ('貴', 17, 148) ('pmid', 18, 358) ('exclaimed', 19, 1365) ### ### =============================================== v_p ================================================ ### ### ('pages', 15449, 0) ('idiot', 4655, 1) ('til', 11958, 2) ('seeing', 5490, 3) ('hear', 7248, 4) ### ### ('beck', 1, 5) ('wayne', 2142, 6) ('gerry', 849, 7) ('soundtrack', 902, 8) ('maxi', 1502, 9) ### ### ('stars', 7396, 10) ('remade', 25510, 11) ('america', 0, 12) ('″', 672, 13) ('remake', 5936, 14) ### ### ('##ley', 3, 15) ('summer', 2540, 16) ('back', 5193, 17) ('group', 107, 18) ('world', 2490, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('beck', 1, 5) ('america', 0, 12) ('##ley', 3, 15) ('headline', 21, 20) ('band', 7, 84) ### ### ('¹', 2, 329) ('幸', 8, 110) ('group', 107, 18) ('absently', 15, 56) ('yeah', 45, 30) ### ### ('cover', 123, 21) ('ம', 10, 98) ('album', 49, 38) ('খ', 11, 120) ('intently', 13, 107) ### ### ('gerry', 849, 7) ('##ப', 9, 137) ('soundtrack', 902, 8) ('″', 672, 13) ('stared', 95, 44) ### ############################################################################################################ [2023-11-25 18:34:21,635][root][INFO] - Epoch: 4: Step: 4801/19530, loss[v]=0.175051, lr=0.000009, acc@1[1]=972.5/1024=0.94970703125, acc@1[2]=1001.0/1024=0.9775390625 [2023-11-25 18:36:10,342][root][INFO] - Train batch 4900 [2023-11-25 18:36:10,343][root][INFO] - Avg. loss per last 100 batches: 0.135092 [2023-11-25 18:36:11,057][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.94 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.3/29522=99.95% | mean: 0.02 | max: 5.77 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: producer jack douglas later insisted : the album's opening track, " back in the saddle ", ### ### recalls the gene autry song " back in the saddle again " ( vocalist steven tyler yodels on the fade ### ### ) and features the sound of a whip by whirling a thirty - foot cord in the middle of six neumann ### ### mikes and adding a cap gun for the cracking sound effect. ### ### [P_TEXT]: jack douglas ( record producer ) york dolls'first album that douglas was encouraged by ### ### producer bob ezrin to also consider becoming a record producer. douglas engineered and produced ### ### many of aerosmith's albums in the 1970s, including " get your wings " ( 1974 ), " toys in the attic ### ### " ( 1975 ), " rocks " ( 1976 ) and " draw the line " ( 1977 ), all of which have gone multi - ### ### platinum. " toys in the attic " and " rocks " broke aerosmith into the mainstream and have become ### ### highly influential, with both albums ranking among " rolling stone "'s list ### ### =============================================== v_q ================================================ ### ### ('saddle', 0, 18132) ('whip', 1, 11843) ('douglas', 2, 1) ('neumann', 3, 17649) ('cord', 4, 17316) ### ### ('##del', 5, 27657) ('steven', 6, 15579) ('back', 7, 7344) ('cap', 8, 19715) ('jack', 9, 6) ### ### ('##try', 10, 27706) ('absently', 11, 75) ('fade', 12, 2370) ('yo', 13, 5294) ('gun', 14, 12245) ### ### ('tyler', 15, 4962) ('##½', 16, 166) ('¹', 17, 85) ('au', 18, 19771) ('gene', 19, 13521) ### ### =============================================== v_p ================================================ ### ### ('rocks', 1932, 0) ('douglas', 2, 1) ('wings', 4033, 2) ('toys', 1033, 3) ('attic', 20243, 4) ### ### ('″', 22, 5) ('jack', 9, 6) ('##rin', 21873, 7) ('draw', 15666, 8) ('get', 2564, 9) ### ### ('dolls', 425, 10) ('##smith', 2096, 11) ('rolling', 48, 12) ('album', 24, 13) ('bob', 2817, 14) ### ### ('line', 9893, 15) ('engineered', 485, 16) ('anyway', 310, 17) ('your', 2701, 18) ### ### ('mastering', 321, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('douglas', 2, 1) ('jack', 9, 6) ('″', 22, 5) ('album', 24, 13) ('rolling', 48, 12) ('"', 47, 21) ### ### ('albums', 27, 36) ('absently', 11, 75) ('幸', 20, 59) ('producer', 95, 22) ('¹', 17, 85) ### ### ('yeah', 42, 39) ('rocks', 1932, 0) ('toys', 1033, 3) ('dolls', 425, 10) ('খ', 21, 98) ### ### ('intently', 29, 67) ('stared', 51, 37) ('##ப', 23, 108) ('anyway', 310, 17) ### ############################################################################################################ [2023-11-25 18:36:11,564][root][INFO] - Epoch: 4: Step: 4901/19530, loss[v]=0.059293, lr=0.000009, acc@1[1]=1003.0/1024=0.9794921875, acc@1[2]=1018.0/1024=0.994140625 [2023-11-25 18:38:00,365][root][INFO] - Train batch 5000 [2023-11-25 18:38:00,365][root][INFO] - Avg. loss per last 100 batches: 0.127799 [2023-11-25 18:38:01,065][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.3/29522=99.93% | mean: 0.02 | max: 5.59 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 26.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: however, the comparative difficulty of the approach to the climbs makes it a much less ### ### popular climb than the grand teton and other peaks to the south. ### ### [P_TEXT]: teton range says the range is named for the teton sioux ( from thithuŋwaŋ ), also known ### ### as the lakota people. it is likely that the local shoshone people once called the whole range ', ### ### meaning " many pinnacles ". the principal summits of the central massif, sometimes referred to as ### ### the cathedral group, are grand teton ( ), mount owen ( ), teewinot ( ), middle teton ( ) and south ### ### teton ( ). other peaks in the range include mount moran ( ), mount wister ( ), buck mountain ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 57) ('te', 1, 1) ('##ton', 2, 2) ('³', 3, 181) ('grand', 4, 22) ('幸', 5, 118) ### ### ('貴', 6, 149) ('climbed', 7, 988) ('intently', 8, 92) ('everest', 9, 3343) ('খ', 10, 148) ### ### ('##ப', 11, 152) ('ம', 12, 82) ('peak', 13, 35) ('##¨', 14, 729) ('.', 15, 931) ### ### ('mountain', 16, 27) ('climb', 17, 74) ('ascent', 18, 339) ('peaks', 19, 21) ### ### =============================================== v_p ================================================ ### ### ('moran', 1380, 0) ('te', 1, 1) ('##ton', 2, 2) ('owen', 10759, 3) ('sioux', 1098, 4) ### ### ('##kota', 16107, 5) ('cathedral', 4887, 6) ('pinnacle', 774, 7) ('buck', 8896, 8) ### ### ('range', 228, 9) ('tee', 2009, 10) ('##ster', 23325, 11) ('##win', 26600, 12) ('mount', 20, 13) ### ### ('##sho', 1702, 14) ('massif', 168, 15) ('sho', 12323, 16) ('summit', 25, 17) ('anyway', 26, 18) ### ### ('hissed', 74, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('te', 1, 1) ('##ton', 2, 2) ('grand', 4, 22) ('¹', 0, 57) ('mount', 20, 13) ('mountain', 16, 27) ### ### ('peaks', 19, 21) ('peak', 13, 35) ('summit', 25, 17) ('anyway', 26, 18) ('幸', 5, 118) ### ### ('intently', 8, 92) ('³', 3, 181) ('ம', 12, 82) ('climb', 17, 74) ('circling', 64, 20) ### ### ('hissed', 74, 19) ('range', 228, 9) ('貴', 6, 149) ('খ', 10, 148) ### ############################################################################################################ [2023-11-25 18:38:01,571][root][INFO] - Epoch: 4: Step: 5001/19530, loss[v]=0.234048, lr=0.000009, acc@1[1]=982.0/1024=0.958984375, acc@1[2]=1002.0/1024=0.978515625 [2023-11-25 18:39:50,035][root][INFO] - Train batch 5100 [2023-11-25 18:39:50,035][root][INFO] - Avg. loss per last 100 batches: 0.140630 [2023-11-25 18:39:50,733][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.3/29522=99.94% | mean: 0.02 | max: 5.68 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: on september 2016, the 32 year - old asian elephant named kaavan, turned mentally ill due ### ### to being chain bound for the last two decades. ### ### [P_TEXT]: kaavan kaavan ( ; born 1985 ) is a male asian elephant known as the'world's loneliest ### ### elephant'after his partner saheli died in 2012. kaavan was gifted by the government of sri lanka to ### ### pakistan in 1985 as a result of the positive relationship between the two countries. he remained at ### ### the islamabad zoo until november 2020, when he was moved to a sanctuary in cambodia in response to ### ### a campaign launched by local and international animal rights activists led by american singer cher. ### ### in may 2020, the islamabad high court ordered kaavan to be released. kaavan was born in sri lanka ### ### in 1985 and upon his ### ### =============================================== v_q ================================================ ### ### ('##ava', 0, 0) ('chain', 1, 12529) ('bound', 2, 8342) ('elephant', 3, 6) ('ka', 4, 12) ### ### ('2016', 5, 2575) ('mental', 6, 11816) ('intently', 7, 331) ('¹', 8, 1012) ('貴', 9, 317) ### ### ('asian', 10, 28) ('absently', 11, 258) ('幸', 12, 271) ('##ன', 13, 129) ('ம', 14, 228) ### ### ('³', 15, 4617) ('খ', 16, 401) ('##n', 17, 31) ('crore', 18, 18762) ('2017', 19, 1093) ### ### =============================================== v_p ================================================ ### ### ('##ava', 0, 0) ('islamabad', 1798, 1) ('cher', 21281, 2) ('zoo', 34, 3) ('##liest', 913, 4) ### ### ('cambodia', 1747, 5) ('elephant', 3, 6) ('##hel', 26383, 7) ('gifted', 945, 8) ### ### ('elephants', 185, 9) ('worried', 65, 10) ('lanka', 113, 11) ('ka', 4, 12) ('males', 5040, 13) ### ### ('leaned', 2930, 14) ('sanctuary', 89, 15) ('court', 10803, 16) ('pakistan', 1930, 17) ### ### ('high', 7725, 18) ('relationship', 8683, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ava', 0, 0) ('elephant', 3, 6) ('ka', 4, 12) ('asian', 10, 28) ('zoo', 34, 3) ('##n', 17, 31) ### ### ('anyway', 27, 22) ('worried', 65, 10) ('sanctuary', 89, 15) ('lanka', 113, 11) ### ### ('elephants', 185, 9) ('##ன', 13, 129) ('animal', 258, 20) ('islamabad', 1798, 1) ### ### ('##liest', 913, 4) ('sri', 85, 47) ('absently', 11, 258) ('貴', 9, 317) ('gifted', 945, 8) ### ### ('intently', 7, 331) ### ############################################################################################################ [2023-11-25 18:39:51,238][root][INFO] - Epoch: 4: Step: 5101/19530, loss[v]=0.080532, lr=0.000009, acc@1[1]=999.0/1024=0.9755859375, acc@1[2]=1015.0/1024=0.9912109375 [2023-11-25 18:41:40,015][root][INFO] - Train batch 5200 [2023-11-25 18:41:40,016][root][INFO] - Avg. loss per last 100 batches: 0.134455 [2023-11-25 18:41:40,729][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.82 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.4/29522=99.95% | mean: 0.02 | max: 5.69 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the drugs known as angiotensin converting enzyme inhibitors ( ace inhibitors ), which are ### ### used for the treatment of hypertension and some types of congestive heart failure, were developed ### ### from a peptide found in the venom of this species in 1965 by brazilian scientist sergio henrique ### ### ferreira. ### ### [P_TEXT]: ace inhibitor renoprotective effect in patients afflicted with diabetic nephropathy, and ### ### pediatric iga nephropathy. leonard t. skeggs and his colleagues ( including norman shumway ) ### ### discovered ace in plasma in 1956. it was also noted that those who worked in banana plantations in ### ### south - western brazil collapsed after being bitten by a pit viper, leading to a search for a blood ### ### pressure lowering component in its venom. brazilian scientist sergio henrique ferreira reported a ### ### bradykinin - potentiating factor ( bpf ) present in the venom of " bothrops jararaca ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 275) ('ferreira', 1, 4) ('ace', 2, 0) ('sergio', 3, 24) ('absently', 4, 342) ('幸', 5, 247) ### ### ('খ', 6, 286) ('貴', 7, 312) ('intently', 8, 379) ('.', 9, 1697) ('##ப', 10, 294) ('³', 11, 2729) ### ### ('ang', 12, 27904) ('##tension', 13, 10362) ('venom', 14, 11) ('inhibitor', 15, 2) ('ம', 16, 179) ### ### ('ɐ', 17, 2269) ('##iot', 18, 29216) ('peptide', 19, 680) ### ### =============================================== v_p ================================================ ### ### ('ace', 2, 0) ('brady', 16027, 1) ('inhibitor', 15, 2) ('viper', 124, 3) ('ferreira', 1, 4) ### ### ('shu', 10078, 5) ('##way', 11846, 6) ('plantations', 26508, 7) ('pit', 2045, 8) ('bp', 10948, 9) ### ### ('##bet', 10968, 10) ('venom', 14, 11) ('jar', 12116, 12) ('banana', 3555, 13) ### ### ('littered', 5091, 14) ('potent', 324, 15) ('bananas', 24366, 16) ('collapsed', 19425, 17) ### ### ('henri', 30, 18) ('norman', 22340, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ferreira', 1, 4) ('ace', 2, 0) ('inhibitor', 15, 2) ('sergio', 3, 24) ('venom', 14, 11) ### ### ('henri', 30, 18) ('inhibitors', 28, 22) ('brazil', 29, 27) ('brazilian', 33, 36) ('viper', 124, 3) ### ### ('blood', 52, 35) ('anyway', 67, 40) ('¹', 0, 275) ('##in', 23, 100) ('potent', 324, 15) ### ### ('幸', 5, 247) ('ம', 16, 179) ('খ', 6, 286) ('##que', 49, 81) ('absently', 4, 342) ### ############################################################################################################ [2023-11-25 18:41:41,236][root][INFO] - Epoch: 4: Step: 5201/19530, loss[v]=0.122196, lr=0.000009, acc@1[1]=999.5/1024=0.97607421875, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 18:43:30,341][root][INFO] - Train batch 5300 [2023-11-25 18:43:30,341][root][INFO] - Avg. loss per last 100 batches: 0.137941 [2023-11-25 18:43:31,048][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.82 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29505.4/29522=99.94% | mean: 0.02 | max: 5.71 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it states that bjorn was the son of ragnar and aslaug and that his brothers were ### ### hvitserk, ivar the boneless, and sigurd snake - in - the - eye. ### ### [P_TEXT]: tale of ragnar's sons serpent's head, and when the people found out what had happened, he ### ### married thora. then, he proceeded to liberate his kingdom. after a few years lagertha left ragnar. ### ### he then married aslaug, also known as randalin, the daughter of sigurd and brynhildr. they had 5 ### ### sons, ivar the boneless, ubbe, hvitserk, bjorn ironside, and sigurd snake - in - the - eye, so ### ### named because his iris had the image of a snake encircling the pupil. ragnar's sons grow up and in ### ### order to show themselves ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 319) ('bjorn', 1, 8) ('##lau', 2, 4) ('খ', 3, 207) ('rag', 4, 11) ('幸', 5, 144) ### ### ('貴', 6, 217) ('##ப', 7, 165) ('##k', 8, 68) ('snake', 9, 10) ('¹', 10, 59) ('intently', 11, 210) ### ### ('ம', 12, 102) ('##ts', 13, 23) ('##gur', 14, 7) ('iv', 15, 9) ('tolkien', 16, 10178) ### ### ('cursed', 17, 1672) ('##部', 18, 132) ('##nar', 19, 21) ### ### =============================================== v_p ================================================ ### ### ('iris', 6590, 0) ('thor', 95, 1) ('tale', 243, 2) ('bryn', 8529, 3) ('##lau', 2, 4) ### ### ('##bbe', 11694, 5) ('serpent', 167, 6) ('##gur', 14, 7) ('bjorn', 1, 8) ('iv', 15, 9) ### ### ('snake', 9, 10) ('rag', 4, 11) ('rand', 5004, 12) ('##ali', 26676, 13) ('pupil', 4371, 14) ### ### ('sons', 83, 15) ('eye', 25, 16) ('si', 35, 17) ('##hil', 14944, 18) ('bone', 50, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##lau', 2, 4) ('bjorn', 1, 8) ('rag', 4, 11) ('snake', 9, 10) ('##gur', 14, 7) ('iv', 15, 9) ### ### ('##ts', 13, 23) ('eye', 25, 16) ('##nar', 19, 21) ('si', 35, 17) ('h', 26, 26) ('thor', 95, 1) ### ### ('bone', 50, 19) ('##vi', 39, 31) ('¹', 10, 59) ('##less', 42, 35) ('##k', 8, 68) ('sons', 83, 15) ### ### ('serpent', 167, 6) ('brother', 57, 25) ### ############################################################################################################ [2023-11-25 18:43:31,554][root][INFO] - Epoch: 4: Step: 5301/19530, loss[v]=0.132114, lr=0.000009, acc@1[1]=988.5/1024=0.96533203125, acc@1[2]=1013.5/1024=0.98974609375 [2023-11-25 18:45:19,995][root][INFO] - Train batch 5400 [2023-11-25 18:45:19,995][root][INFO] - Avg. loss per last 100 batches: 0.136808 [2023-11-25 18:45:20,702][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.1/29522=99.93% | mean: 0.02 | max: 5.61 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the clarendonian is preceded by the barstovian and followed by the hemphillian nalma ### ### stages. ### ### [P_TEXT]: barstovian the barstovian north american stage on the geologic timescale is the north ### ### american faunal stage according to the north american land mammal ages chronology ( nalma ), ### ### typically set from 16, 300, 000 to 13, 600, 000 years bp, a period of. it is usually considered to ### ### overlap the langhian and serravallian stages of the middle miocene. the barstovian is preceded by ### ### the hemingfordian and followed by the clarendonian nalma stages. ### ### =============================================== v_q ================================================ ### ### ('##ndon', 0, 5) ('##lma', 1, 2) ('##phi', 2, 25466) ('bars', 3, 0) ('hem', 4, 8) ('clare', 5, 14) ### ### ('¹', 6, 185) ('na', 7, 19) ('貴', 8, 170) ('幸', 9, 133) ('খ', 10, 152) ('##lli', 11, 22527) ### ### ('##tov', 12, 1) ('absently', 13, 144) ('intently', 14, 136) ('##ian', 15, 33) ('##ப', 16, 138) ### ### ('ம', 17, 122) ('anyway', 18, 31) ('extinct', 19, 1879) ### ### =============================================== v_p ================================================ ### ### ('bars', 3, 0) ('##tov', 12, 1) ('##lma', 1, 2) ('serra', 2427, 3) ('miocene', 111, 4) ### ### ('##ndon', 0, 5) ('stages', 273, 6) ('lang', 17978, 7) ('hem', 4, 8) ('ages', 307, 9) ### ### ('##hian', 591, 10) ('stage', 370, 11) ('chronology', 1243, 12) ('##ford', 10508, 13) ### ### ('clare', 5, 14) ('land', 3048, 15) ('mammal', 124, 16) ('preceded', 296, 17) ('##cal', 15701, 18) ### ### ('na', 7, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ndon', 0, 5) ('bars', 3, 0) ('##lma', 1, 2) ('##tov', 12, 1) ('hem', 4, 8) ('clare', 5, 14) ### ### ('na', 7, 19) ('##ian', 15, 33) ('anyway', 18, 31) ('fauna', 29, 20) ('miocene', 111, 4) ### ### ('geologic', 54, 23) ('geological', 81, 22) ('mammal', 124, 16) ('stages', 273, 6) ('ages', 307, 9) ### ### ('幸', 9, 133) ('middle', 116, 29) ('¹', 6, 185) ('ம', 17, 122) ### ############################################################################################################ [2023-11-25 18:45:21,207][root][INFO] - Epoch: 4: Step: 5401/19530, loss[v]=0.084316, lr=0.000009, acc@1[1]=996.0/1024=0.97265625, acc@1[2]=1011.5/1024=0.98779296875 [2023-11-25 18:47:10,169][root][INFO] - Train batch 5500 [2023-11-25 18:47:10,170][root][INFO] - Avg. loss per last 100 batches: 0.135845 [2023-11-25 18:47:10,922][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.5/29522=99.94% | mean: 0.02 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: a centrepiece of the larger singapore sports hub complex, the stadium hosts major ### ### association football matches, including the home matches of the singapore national team. ### ### [P_TEXT]: kallang airport park, a large public park created as part of " project lung " ; later, ### ### the old national stadium was built on the same site ; today, the singapore sports hub ( including ### ### the new national stadium and singapore indoor stadium ) is a major national landmark. the eastern ### ### portion of the kallang airport was redeveloped into one of singapore's first modern residential ### ### precinct, now called the " old kallang airport estate ". it was one of the locations that hosted ### ### the 2011 singapore biennale art festival. there are plans to transform the kallang airport ### ### =============================================== v_q ================================================ ### ### ('singapore', 0, 4) ('stadium', 1, 3) ('hub', 2, 7) ('¹', 3, 223) ('ɐ', 4, 3038) ('貴', 5, 134) ### ### ('ம', 6, 99) ('##ப', 7, 139) ('³', 8, 355) ('complex', 9, 1994) ('.', 10, 1251) ('幸', 11, 124) ### ### ('sports', 12, 13) ('intently', 13, 64) ('unveiled', 14, 2342) ('matches', 15, 176) ### ### ('national', 16, 16) ('games', 17, 104) ('held', 18, 522) ('footballer', 19, 1108) ### ### =============================================== v_p ================================================ ### ### ('lung', 8686, 0) ('biennale', 5404, 1) ('##llan', 20472, 2) ('stadium', 1, 3) ('singapore', 0, 4) ### ### ('indoor', 138, 5) ('airport', 424, 6) ('hub', 2, 7) ('anyway', 243, 8) ('2011', 1797, 9) ### ### ('project', 3992, 10) ('festival', 10349, 11) ('ka', 3319, 12) ('sports', 12, 13) ### ### ('precinct', 514, 14) ('landmark', 340, 15) ('national', 16, 16) ('″', 5317, 17) ('art', 6671, 18) ### ### ('hissed', 345, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('singapore', 0, 4) ('stadium', 1, 3) ('hub', 2, 7) ('sports', 12, 13) ('national', 16, 16) ### ### ('indoor', 138, 5) ('park', 35, 26) ('sport', 61, 23) ('airport', 424, 6) ('anyway', 243, 8) ### ### ('ம', 6, 99) ('intently', 13, 64) ('¹', 3, 223) ('貴', 5, 134) ('landmark', 340, 15) ('##ப', 7, 139) ### ### ('幸', 11, 124) ('games', 17, 104) ('hissed', 345, 19) ('precinct', 514, 14) ### ############################################################################################################ [2023-11-25 18:47:11,432][root][INFO] - Epoch: 4: Step: 5501/19530, loss[v]=0.204434, lr=0.000009, acc@1[1]=972.5/1024=0.94970703125, acc@1[2]=1005.0/1024=0.9814453125 [2023-11-25 18:49:00,279][root][INFO] - Train batch 5600 [2023-11-25 18:49:00,280][root][INFO] - Avg. loss per last 100 batches: 0.137902 [2023-11-25 18:49:00,984][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.03 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.6/29522=99.95% | mean: 0.03 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: team is led by the hangman, who calls on satannish. ### ### [P_TEXT]: satannish demonic energies. satannish was revealed to have hidden in the shadow of ### ### mephista, daughter of mephisto, in order to be taken into mephisto's domain. there, he battled ### ### mephisto, mephista, and doctor strange, and was vanquished when strange mystically connected ### ### mephisto's and satannish's life essences. satannish also once empowered the night shift. during the ### ### " secret invasion " storyline, satannish is one of many dark powers imprisoned within avalon. he is ### ### freed, along with all the others, when pete wisdom releases merlin. after hearing the invading ### ### skrull s boast that they will ### ### =============================================== v_q ================================================ ### ### ('hang', 0, 8395) ('##nish', 1, 2) ('satan', 2, 3) ('##man', 3, 21948) ('absently', 4, 183) ### ### ('！', 5, 1396) ('貴', 6, 185) ('¹', 7, 96) ('幸', 8, 120) ('freaking', 9, 289) ('character', 10, 88) ### ### ('comics', 11, 406) ('ம', 12, 128) ('খ', 13, 167) ('characters', 14, 98) ('.', 15, 70) ### ### ('³', 16, 1741) ('intently', 17, 217) ('##ப', 18, 129) ('hanging', 19, 7023) ### ### =============================================== v_p ================================================ ### ### ('avalon', 2579, 0) ('merlin', 13618, 1) ('##nish', 1, 2) ('satan', 2, 3) ('wisdom', 10487, 4) ### ### ('shift', 2946, 5) ('invasion', 3131, 6) ('##phi', 18653, 7) ('pete', 2076, 8) ('strange', 1588, 9) ### ### ('##sta', 7542, 10) ('sk', 15041, 11) ('mystical', 2065, 12) ('secret', 374, 13) ### ### ('##sto', 22230, 14) ('night', 1576, 15) ('storyline', 94, 16) ('shadow', 582, 17) ### ### ('essence', 2186, 18) ('powers', 463, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##nish', 1, 2) ('satan', 2, 3) ('storyline', 94, 16) ('anyway', 52, 24) ('¹', 7, 96) ### ### ('secret', 374, 13) ('.', 15, 70) ('character', 10, 88) ('幸', 8, 120) ('absently', 4, 183) ### ### ('characters', 14, 98) ('avalon', 2579, 0) ('貴', 6, 185) ('powers', 463, 19) ('ம', 12, 128) ### ### ('appears', 192, 26) ('shadow', 582, 17) ('villains', 38, 66) ('strange', 1588, 9) ### ### ('worried', 107, 41) ### ############################################################################################################ [2023-11-25 18:49:01,489][root][INFO] - Epoch: 4: Step: 5601/19530, loss[v]=0.095275, lr=0.000009, acc@1[1]=993.5/1024=0.97021484375, acc@1[2]=1010.5/1024=0.98681640625 [2023-11-25 18:50:50,035][root][INFO] - Train batch 5700 [2023-11-25 18:50:50,036][root][INFO] - Avg. loss per last 100 batches: 0.137299 [2023-11-25 18:50:50,740][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29505.7/29522=99.94% | mean: 0.02 | max: 5.76 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.5/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 1969, menon contested a seat in the lok sabha from the bengal constituency of ### ### midnapore, running as an independent in a by - election, and defeating his congress rival by a ### ### margin of 106, 767 votes in may of that year. ### ### [P_TEXT]: pranab mukherjee and contested and won a seat in the lok sabha ( lower house ). he served ### ### there until mid - 2012, when he ran for president of india. mukherjee started his political career ### ### in 1967 as a founding member of the bangla congress. ahead of the 1967 election, he played a ### ### crucial role in forging the united front alliance against the indian national congress. in 1969, he ### ### managed the successful midnapore by - election campaign of an independent candidate, v. k. krishna ### ### menon. he became a member of ### ### =============================================== v_q ================================================ ### ### ('menon', 0, 2) ('¹', 1, 223) ('congress', 2, 4) ('##pore', 3, 7) ('bengal', 4, 832) ('mid', 5, 9) ### ### ('1969', 6, 22) ('.', 7, 252) ('sabha', 8, 6) ('lok', 9, 33) ('³', 10, 1966) ('india', 11, 27) ### ### ('খ', 12, 151) ('貴', 13, 173) ('##na', 14, 48) ('ம', 15, 107) ('election', 16, 31) ### ### ('intently', 17, 164) ('##ப', 18, 130) ('sonia', 19, 1105) ### ### =============================================== v_p ================================================ ### ### ('mukherjee', 125, 0) ('bang', 1798, 1) ('menon', 0, 2) ('1967', 127, 3) ('congress', 2, 4) ### ### ('front', 3117, 5) ('sabha', 8, 6) ('##pore', 3, 7) ('##ana', 6104, 8) ('mid', 5, 9) ### ### ('krishna', 2008, 10) ('leaned', 11583, 11) ('turnout', 12845, 12) ('##la', 5974, 13) ### ### ('alliance', 1697, 14) ('united', 14530, 15) ('pr', 1703, 16) ('indian', 38, 17) ### ### ('elections', 60, 18) ('founders', 4427, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('menon', 0, 2) ('congress', 2, 4) ('##pore', 3, 7) ('sabha', 8, 6) ('mid', 5, 9) ('1969', 6, 22) ### ### ('india', 11, 27) ('lok', 9, 33) ('election', 16, 31) ('mukherjee', 125, 0) ('littered', 25, 26) ### ### ('1967', 127, 3) ('indian', 38, 17) ('independent', 39, 24) ('##na', 14, 48) ('elections', 60, 18) ### ### ('¹', 1, 223) ('bjp', 59, 29) ('ம', 15, 107) ('খ', 12, 151) ### ############################################################################################################ [2023-11-25 18:50:51,247][root][INFO] - Epoch: 4: Step: 5701/19530, loss[v]=0.192631, lr=0.000009, acc@1[1]=980.5/1024=0.95751953125, acc@1[2]=1002.5/1024=0.97900390625 [2023-11-25 18:52:40,282][root][INFO] - Train batch 5800 [2023-11-25 18:52:40,282][root][INFO] - Avg. loss per last 100 batches: 0.139596 [2023-11-25 18:52:41,003][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.75 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.2/29522=99.95% | mean: 0.02 | max: 5.68 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in mid - march 2003, a tunnel under the western scheldt opened, the western scheldt ### ### tunnel, running from ellewoutsdijk in the municipality of borsele ( zuid - beveland ) to terneuzen ### ### in zeelandic flanders. ### ### [P_TEXT]: louis van kuyck to animal painting ; predominantly horses with stables. this would ### ### eventually come to dominate his oeuvre. an unusual incident was crucial to promoting his career. in ### ### 1852, queen victoria was sailing back to england, following a state visit with king leopold i of ### ### belgium. her ship encountered a storm, and was forced to anchor on the western scheldt, near ### ### terneuzen. when the storm had subsided a bit, the royal party went to shore in a dinghy where, ### ### among other things, they toured a local farm. she was apparently ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 188) ('¹', 1, 174) ('tunnel', 2, 8387) ('##uts', 3, 20571) ('##rse', 4, 23252) ### ### ('খ', 5, 258) ('幸', 6, 294) ('##uze', 7, 3) ('##vel', 8, 15048) ('flanders', 9, 855) ('貴', 10, 273) ### ### ('elle', 11, 15435) ('##ப', 12, 305) ('ம', 13, 211) ('zee', 14, 19177) ('ter', 15, 13) ### ### ('intently', 16, 449) ('netherlands', 17, 463) ('##and', 18, 25070) ('western', 19, 25) ### ### =============================================== v_p ================================================ ### ### ('##held', 28, 0) ('leopold', 1779, 1) ('##yck', 3599, 2) ('##uze', 7, 3) ('ding', 14391, 4) ### ### ('horses', 9475, 5) ('1852', 4827, 6) ('animal', 11230, 7) ('storms', 26496, 8) ('louis', 3293, 9) ### ### ('outskirts', 584, 10) ('ku', 20069, 11) ('victoria', 10805, 12) ('ter', 15, 13) ### ### ('belgium', 29, 14) ('storm', 2904, 15) ('leaned', 92, 16) ('worried', 351, 17) ('belgian', 43, 18) ### ### ('painting', 6311, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##uze', 7, 3) ('##held', 28, 0) ('ter', 15, 13) ('belgium', 29, 14) ('western', 19, 25) ### ### ('belgian', 43, 18) ('absently', 0, 188) ('leaned', 92, 16) ('sc', 23, 92) ('¹', 1, 174) ### ### ('anyway', 96, 24) ('##ne', 44, 68) ('stared', 32, 93) ('van', 142, 21) ('খ', 5, 258) ### ### ('dutch', 21, 139) ('幸', 6, 294) ('貴', 10, 273) ('ம', 13, 211) ('brussels', 27, 182) ### ############################################################################################################ [2023-11-25 18:52:41,511][root][INFO] - Epoch: 4: Step: 5801/19530, loss[v]=0.184695, lr=0.000009, acc@1[1]=978.0/1024=0.955078125, acc@1[2]=1002.0/1024=0.978515625 [2023-11-25 18:54:30,247][root][INFO] - Train batch 5900 [2023-11-25 18:54:30,248][root][INFO] - Avg. loss per last 100 batches: 0.131459 [2023-11-25 18:54:30,953][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.0/29522=99.95% | mean: 0.02 | max: 5.71 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the borough borders bloomingdale, wanaque and west milford in passaic county ; mahwah and ### ### oakland in bergen county ; tuxedo and warwick in orange county, new york ; and ramapo in rockland ### ### county, new york. ### ### [P_TEXT]: oakland, new jersey borough include ramapo lake and rotten pond. the borough borders ### ### franklin lakes and mahwah in bergen county ; and pompton lakes, ringwood, wanaque and wayne in ### ### passaic county. the census bureau's 2006 - 2010 american community survey showed that ( in 2010 ### ### inflation - adjusted dollars ) median household income was $ 111, 390 ( with a margin of error of + ### ### / - $ 6, 160 ) and the median family income was $ 114, 973 ( + / - $ 7, 378 ). males had a median ### ### income of $ 82, 750 ( + / - $ 6, 931 ) versus $ 59, 349 ( + / - $ 7, 903 ) for females. the per ### ### capita income for the borough was $ 43, 651 ( + / - $ 3, 082 ). about 0. 7 % of families and 1. 9 % ### ### of the population were below ### ### =============================================== v_q ================================================ ### ### ('jersey', 0, 15) ('absently', 1, 79) ('##xed', 2, 7276) ('warwick', 3, 14030) ('##hwa', 4, 2) ### ### ('bloom', 5, 4946) ('milford', 6, 821) ('.', 7, 75) ('幸', 8, 93) ('oakland', 9, 0) ('貴', 10, 106) ### ### ('খ', 11, 104) ('##ப', 12, 115) ('##po', 13, 14) ('##部', 14, 153) ('ம', 15, 86) ### ### ('##dale', 16, 19260) ('intently', 17, 89) ('rama', 18, 7) ('¹', 19, 39) ### ### =============================================== v_p ================================================ ### ### ('oakland', 9, 0) ('rotten', 9723, 1) ('##hwa', 4, 2) ('wayne', 619, 3) ('ring', 6456, 4) ### ### ('##mpton', 420, 5) ('lakes', 5001, 6) ('rama', 18, 7) ('pond', 108, 8) ('##aq', 28, 9) ### ### ('##wood', 3367, 10) ('bergen', 26, 11) ('inflation', 2806, 12) ('borough', 51, 13) ### ### ('##po', 13, 14) ('jersey', 0, 15) ('surveyed', 6495, 16) ('anyway', 214, 17) ### ### ('neighboring', 384, 18) ('wan', 21, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('oakland', 9, 0) ('##hwa', 4, 2) ('jersey', 0, 15) ('rama', 18, 7) ('##po', 13, 14) ### ### ('bergen', 26, 11) ('##aq', 28, 9) ('wan', 21, 19) ('ma', 20, 25) ('¹', 19, 39) ('##ue', 36, 24) ### ### ('borough', 51, 13) ('absently', 1, 79) ('pond', 108, 8) ('##h', 59, 22) ('residents', 45, 34) ### ### ('.', 7, 75) ('幸', 8, 93) ('wayne', 619, 3) ('ம', 15, 86) ### ############################################################################################################ [2023-11-25 18:54:31,457][root][INFO] - Epoch: 4: Step: 5901/19530, loss[v]=0.209840, lr=0.000008, acc@1[1]=977.5/1024=0.95458984375, acc@1[2]=1008.0/1024=0.984375 [2023-11-25 18:56:20,093][root][INFO] - Train batch 6000 [2023-11-25 18:56:20,094][root][INFO] - Avg. loss per last 100 batches: 0.137832 [2023-11-25 18:56:20,795][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.8/29522=99.94% | mean: 0.02 | max: 5.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: along with his teaching career, he led the modernist movement in gujarati literature. ### ### [P_TEXT]: gujarati literature with ethical values and its main writers were gulabdas broker, ### ### mansukhlal jhaveri, vishnuprasad trivedi and others. existentialism, surrealism and symbolism ### ### influenced the latter. the modernists also wanted to do away with moral values and religious ### ### beliefs. eminent writers of this trend comprise niranjan bhagat, chandrakant bakshi, suresh joshi, ### ### madhu rye, raghuveer chaudhari, dhiruben patel, saroj pathak, and others. there was also a ### ### noticeable segment of popular writers like vithal pandya, sarang barot, dinkar joshi, harkisan ### ### mehta and ashwinee bhatt whose novels found a place in the hearts of common people. their novels ### ### reached every corner ### ### =============================================== v_q ================================================ ### ### ('literature', 0, 2) ('gujarati', 1, 0) ('modernist', 2, 24) ('modernism', 3, 79) ### ### ('teaching', 4, 1526) ('¹', 5, 88) ('.', 6, 308) ('teacher', 7, 2748) ('littered', 8, 25) ### ### ('movement', 9, 459) ('##jee', 10, 19466) ('gujarat', 11, 62) ('pained', 12, 124) ### ### ('proponent', 13, 539) ('literary', 14, 72) ('##ि', 15, 53) ('ahmedabad', 16, 198) ### ### ('pmid', 17, 629) ('intently', 18, 165) ('খ', 19, 182) ### ### =============================================== v_p ================================================ ### ### ('gujarati', 1, 0) ('rye', 19101, 1) ('literature', 0, 2) ('broker', 16827, 3) ('##wine', 24212, 4) ### ### ('##dya', 5252, 5) ('surreal', 570, 6) ('joshi', 749, 7) ('##gat', 5248, 8) ('##isan', 16602, 9) ### ### ('sure', 3991, 10) ('patel', 23, 11) ('##ved', 13659, 12) ('path', 2899, 13) ('##kshi', 11242, 14) ### ### ('anyway', 2181, 15) ('chandra', 69, 16) ('##lab', 23800, 17) ('##hta', 163, 18) ('din', 687, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('gujarati', 1, 0) ('literature', 0, 2) ('modernist', 2, 24) ('littered', 8, 25) ('patel', 23, 11) ### ### ('modernism', 3, 79) ('gujarat', 11, 62) ('##ि', 15, 53) ('chandra', 69, 16) ('¹', 5, 88) ### ### ('literary', 14, 72) ('##hta', 163, 18) ('surreal', 570, 6) ('hissed', 113, 44) ('##ism', 106, 49) ### ### ('joshi', 749, 7) ('pained', 12, 124) ('##kar', 81, 69) ('poet', 37, 89) ('din', 687, 19) ### ############################################################################################################ [2023-11-25 18:56:21,301][root][INFO] - Epoch: 4: Step: 6001/19530, loss[v]=0.068346, lr=0.000008, acc@1[1]=1002.5/1024=0.97900390625, acc@1[2]=1019.0/1024=0.9951171875 [2023-11-25 18:58:10,370][root][INFO] - Train batch 6100 [2023-11-25 18:58:10,370][root][INFO] - Avg. loss per last 100 batches: 0.138313 [2023-11-25 18:58:11,089][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.9/29522=99.94% | mean: 0.02 | max: 5.68 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 2004 his performance in " four shades of brown " earned him a guldbagge award for best ### ### actor in a leading role. ### ### [P_TEXT]: four shades of brown four shades of brown ( ) is a 2004 swedish film written by the ### ### comedy group killingganget and directed by their member tomas alfredson. the film stars robert ### ### gustafsson, johan rheborg, henrik schyffert, jonas inde, maria kulle and ulf brunnberg. the film ### ### consists of four interwoven stories about life tragedies, each related to fatherhood as the common ### ### theme. it was produced by sveriges television. the film won four guldbagge awards including best ### ### director and best actor for gustafsson. in 2005 it was aired on television as four mid - length ### ### television films, which had been reedited with ### ### =============================================== v_q ================================================ ### ### ('shades', 0, 0) ('brown', 1, 3) ('¹', 2, 106) ('four', 3, 36) ('swedish', 4, 18) ('2004', 5, 15) ### ### ('stockholm', 6, 140) ('norwegian', 7, 319) ('littered', 8, 42) ('film', 9, 64) ('##bag', 10, 4) ### ### ('##ld', 11, 34) ('貴', 12, 181) ('gu', 13, 66) ('leads', 14, 490) ('actor', 15, 33) ### ### ('whoever', 16, 108) ('convincing', 17, 56) ('starring', 18, 10711) ('henrik', 19, 24) ### ### =============================================== v_p ================================================ ### ### ('shades', 0, 0) ('##gang', 7385, 1) ('tomas', 70, 2) ('brown', 1, 3) ('##bag', 10, 4) ### ### ('gust', 6099, 5) ('worried', 455, 6) ('ind', 17318, 7) ('##lle', 9318, 8) ('alfred', 5342, 9) ### ### ('killing', 4135, 10) ('##nberg', 1263, 11) ('##ffer', 7973, 12) ('##sson', 277, 13) ### ### ('##aged', 16552, 14) ('2004', 5, 15) ('jonas', 163, 16) ('sweden', 28, 17) ('swedish', 4, 18) ### ### ('ul', 16900, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('shades', 0, 0) ('brown', 1, 3) ('##bag', 10, 4) ('2004', 5, 15) ('swedish', 4, 18) ### ### ('four', 3, 36) ('##ld', 11, 34) ('henrik', 19, 24) ('littered', 8, 42) ('actor', 15, 33) ### ### ('sweden', 28, 17) ('tomas', 70, 2) ('film', 9, 64) ('convincing', 17, 56) ('awards', 54, 21) ### ### ('gu', 13, 66) ('¹', 2, 106) ('finest', 36, 39) ('##ge', 31, 47) ('2005', 52, 35) ### ############################################################################################################ [2023-11-25 18:58:11,595][root][INFO] - Epoch: 4: Step: 6101/19530, loss[v]=0.141215, lr=0.000008, acc@1[1]=982.5/1024=0.95947265625, acc@1[2]=1007.5/1024=0.98388671875 [2023-11-25 19:00:00,320][root][INFO] - Train batch 6200 [2023-11-25 19:00:00,321][root][INFO] - Avg. loss per last 100 batches: 0.130641 [2023-11-25 19:00:01,023][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.5/29522=99.95% | mean: 0.02 | max: 5.81 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 1933, gurtner came into conflict with one of his subordinates, roland freisler, over ### ### the issues of " rassenschande " ( literally : " racial disgrace " ), or sexual relationship between ### ### an " aryan " and a " non - aryan ", which freisler wanted immediately criminalized. ### ### [P_TEXT]: roland freisler prohibition of " mixed - blood " sexual intercourse, which met with ### ### expressions of public unease in the dying elements of the german free press and non - nazi ### ### political classes and, at the time, lacked public authorization from the policy of the nazi party, ### ### which had only just obtained dictatorial control of the state. it also led to a clash with his ### ### superior franz gurtner, but freisler's ideological views reflected things to come, as was shown by ### ### the enactment of the nuremberg laws within two years. in october 1939, he introduced the concept ### ### of'precocious juvenile criminal'in the " juvenile ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 646) ('##eis', 1, 2) ('gu', 2, 21) ('##rt', 3, 6) ('roland', 4, 0) ('##ler', 5, 12) ### ### ('##ʂ', 6, 2518) ('³', 7, 4663) ('##ner', 8, 26) ('幸', 9, 380) ('fr', 10, 16) ('intently', 11, 429) ### ### ('খ', 12, 226) ('##ப', 13, 268) ('貴', 14, 356) ('ம', 15, 227) ('aryan', 16, 1613) ### ### ('nazis', 17, 104) ('absently', 18, 186) ('pained', 19, 708) ### ### =============================================== v_p ================================================ ### ### ('roland', 4, 0) ('juvenile', 4599, 1) ('##eis', 1, 2) ('authorization', 16212, 3) ### ### ('laws', 2576, 4) ('##cious', 14717, 5) ('##rt', 3, 6) ('nuremberg', 74, 7) ('sexual', 38, 8) ### ### ('press', 3742, 9) ('nazi', 28, 10) ('classes', 1688, 11) ('##ler', 5, 12) ('prohibition', 509, 13) ### ### ('criminal', 50, 14) ('intercourse', 1538, 15) ('fr', 10, 16) ('mixed', 5413, 17) ### ### ('dictator', 1441, 18) ('franz', 243, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##eis', 1, 2) ('roland', 4, 0) ('##rt', 3, 6) ('gu', 2, 21) ('##ler', 5, 12) ('fr', 10, 16) ### ### ('##ner', 8, 26) ('nazi', 28, 10) ('german', 22, 25) ('sexual', 38, 8) ('criminal', 50, 14) ### ### ('nuremberg', 74, 7) ('non', 35, 39) ('hitler', 37, 36) ('germany', 52, 51) ('1939', 161, 27) ### ### ('franz', 243, 19) ('anyway', 140, 30) ('reich', 83, 43) ('nazis', 17, 104) ### ############################################################################################################ [2023-11-25 19:00:01,527][root][INFO] - Epoch: 4: Step: 6201/19530, loss[v]=0.100022, lr=0.000008, acc@1[1]=996.5/1024=0.97314453125, acc@1[2]=1017.0/1024=0.9931640625 [2023-11-25 19:01:50,142][root][INFO] - Train batch 6300 [2023-11-25 19:01:50,143][root][INFO] - Avg. loss per last 100 batches: 0.139882 [2023-11-25 19:01:50,859][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.7/29522=99.94% | mean: 0.02 | max: 5.89 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the film is loosely based on the true story of tom and eileen lonergan, who in 1998 went ### ### out with a scuba diving group, outer edge dive company, on the great barrier reef, and were ### ### accidentally left behind because ### ### [P_TEXT]: disappearance of tom and eileen lonergan edge dive, was fined after pleading guilty to ### ### negligence and went out of business. queensland's government also introduced stiffer regulations, ### ### among which was the requirement that captains and dive masters independently confirm each head ### ### count. the 2003 movie " open water " was inspired by the lonergans'disappearance. the film is set ### ### in the caribbean and shows two divers getting left at sea due to an incorrect headcount. the ### ### documentary " crime stories : deep secret " also featured the story of the lonergans. their ### ### disappearances were referenced in a game titled " stranded deep ", in which a ### ### =============================================== v_q ================================================ ### ### ('eileen', 0, 2) ('lone', 1, 10) ('absently', 2, 126) ('edge', 3, 20) ('tom', 4, 35) ### ### ('##rgan', 5, 11) ('outer', 6, 19655) ('loosely', 7, 3150) ('¹', 8, 1144) ('film', 9, 37) ### ### ('intently', 10, 320) ('1998', 11, 4198) ('খ', 12, 286) ('貴', 13, 364) ('reef', 14, 4092) ### ### ('³', 15, 1165) ('″', 16, 17) ('幸', 17, 325) ('true', 18, 9250) ('ம', 19, 231) ### ### =============================================== v_p ================================================ ### ### ('negligence', 13828, 0) ('stranded', 956, 1) ('eileen', 0, 2) ('dive', 20, 3) ('count', 23651, 4) ### ### ('open', 8817, 5) ('stiff', 10209, 6) ('deep', 6126, 7) ('water', 413, 8) ### ### ('disappearance', 1908, 9) ('lone', 1, 10) ('##rgan', 5, 11) ('queensland', 1554, 12) ### ### ('##unt', 28527, 13) ('caribbean', 2191, 14) ('head', 7637, 15) ('divers', 3965, 16) ('″', 16, 17) ### ### ('pleading', 23412, 18) ('regulations', 16116, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('eileen', 0, 2) ('lone', 1, 10) ('##rgan', 5, 11) ('edge', 3, 20) ('tom', 4, 35) ('dive', 20, 3) ### ### ('″', 16, 17) ('film', 9, 37) ('diving', 34, 23) ('movie', 32, 39) ('documentary', 56, 34) ### ### ('absently', 2, 126) ('underwater', 45, 44) ('water', 413, 8) ('stranded', 956, 1) ### ### ('anyway', 142, 36) ('remake', 29, 78) ('stories', 232, 29) ('sea', 64, 52) ('films', 50, 60) ### ############################################################################################################ [2023-11-25 19:01:51,365][root][INFO] - Epoch: 4: Step: 6301/19530, loss[v]=0.056594, lr=0.000008, acc@1[1]=1010.0/1024=0.986328125, acc@1[2]=1019.0/1024=0.9951171875 [2023-11-25 19:03:40,456][root][INFO] - Train batch 6400 [2023-11-25 19:03:40,457][root][INFO] - Avg. loss per last 100 batches: 0.132961 [2023-11-25 19:03:41,180][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.83 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.3/29522=99.93% | mean: 0.02 | max: 5.63 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: poor design and continuity in elements such as clothing, props, and proper proportions ### ### between characters and their environments, with industry columnist peter david responding to ### ### liefeld's 1996 work on the " heroes reborn " " captain america " by proclaiming liefeld the " ed ### ### wood of comics ". ### ### [P_TEXT]: heroes reborn ( comics ) also, marvel announced that jeph loeb and rob liefeld would ### ### reunite in november 2006 for " onslaught reborn ", a five - part weekly limited series that would ### ### feature both onslaught and a world similar to the " heroes reborn " universe. this was planned to ### ### tie in with the memorial scholarship fund established in honor of loeb's son sam. ### ### =============================================== v_q ================================================ ### ### ('lie', 0, 12) ('##feld', 1, 4) ('absently', 2, 86) ('¹', 3, 171) ('wood', 4, 16368) ### ### ('peter', 5, 5938) ('david', 6, 12492) ('幸', 7, 118) ('captain', 8, 3711) ('খ', 9, 142) ### ### ('³', 10, 1321) ('reborn', 11, 1) ('intently', 12, 124) ('##ப', 13, 95) ('comics', 14, 29) ### ### ('貴', 15, 147) ('america', 16, 8652) ('freaking', 17, 77) ('！', 18, 1379) ('heroes', 19, 2) ### ### =============================================== v_p ================================================ ### ### ('onslaught', 1868, 0) ('reborn', 11, 1) ('heroes', 19, 2) ('sam', 8773, 3) ('##feld', 1, 4) ### ### ('littered', 1533, 5) ('##eb', 24142, 6) ('scholarships', 29348, 7) ('fund', 13185, 8) ### ### ('##ph', 26874, 9) ('scholarship', 6367, 10) ('universe', 48, 11) ('lie', 0, 12) ### ### ('memorial', 11800, 13) ('worried', 131, 14) ('anyway', 265, 15) ('marvel', 279, 16) ### ### ('rob', 7520, 17) ('je', 15200, 18) ('lo', 23139, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('reborn', 11, 1) ('##feld', 1, 4) ('lie', 0, 12) ('heroes', 19, 2) ('comics', 14, 29) ### ### ('absently', 2, 86) ('universe', 48, 11) ('freaking', 17, 77) ('幸', 7, 118) ('¹', 3, 171) ### ### ('##ப', 13, 95) ('worried', 131, 14) ('onslaught', 1868, 0) ('intently', 12, 124) ('খ', 9, 142) ### ### ('ம', 20, 111) ('貴', 15, 147) ('stared', 27, 51) ('anyway', 265, 15) ('##部', 23, 113) ### ############################################################################################################ [2023-11-25 19:03:41,687][root][INFO] - Epoch: 4: Step: 6401/19530, loss[v]=0.099400, lr=0.000008, acc@1[1]=995.5/1024=0.97216796875, acc@1[2]=1010.5/1024=0.98681640625 [2023-11-25 19:05:30,297][root][INFO] - Train batch 6500 [2023-11-25 19:05:30,298][root][INFO] - Avg. loss per last 100 batches: 0.121326 [2023-11-25 19:05:30,998][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.2/29522=99.94% | mean: 0.02 | max: 5.82 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: during the original broadcast, despite the general lack of interest, leiji matsumoto ### ### received a " very surprising " amount of fan mail from women who had watched the show, which came ### ### as a surprise due to the fact he had ### ### [P_TEXT]: leiji matsumoto, to become a mangaka. matsumoto is famous for his space opera s such as " ### ### space battleship yamato " and " galaxy express 999 ". his style is characterized by mythological ### ### and often tragic storylines with strong moral themes, noble heroes, feminine heroines, and a love ### ### of strange worlds and melancholy atmosphere. matsumoto made his debut under his real name, akira ### ### matsumoto, in 1953. his wife is shojo manga artist miyako maki ( better known as the creator of the ### ### doll licca - chan, the japanese equivalent of ### ### =============================================== v_q ================================================ ### ### ('mats', 0, 6) ('lei', 1, 1) ('absently', 2, 156) ('##um', 3, 34) ('##oto', 4, 8) ('¹', 5, 158) ### ### ('貴', 6, 167) ('fan', 7, 2128) ('intently', 8, 185) ('！', 9, 1667) ('##ji', 10, 57) ('খ', 11, 151) ### ### ('##ப', 12, 153) ('幸', 13, 166) ('ம', 14, 146) ('mail', 15, 11551) ('³', 16, 903) ### ### ('fans', 17, 1525) ('show', 18, 4867) ('japanese', 19, 37) ### ### =============================================== v_p ================================================ ### ### ('999', 11612, 0) ('lei', 1, 1) ('##cca', 9409, 2) ('doll', 3391, 3) ('manga', 3461, 4) ### ### ('opera', 4504, 5) ('mats', 0, 6) ('battleship', 7749, 7) ('##oto', 4, 8) ('space', 892, 9) ### ### ('sho', 7088, 10) ('##mat', 13816, 11) ('battleships', 16761, 12) ('express', 6393, 13) ### ### ('akira', 3745, 14) ('myths', 27413, 15) ('galaxy', 418, 16) ('chan', 3238, 17) ('ma', 21098, 18) ### ### ('″', 36, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('lei', 1, 1) ('mats', 0, 6) ('##oto', 4, 8) ('##um', 3, 34) ('japanese', 19, 37) ('##ji', 10, 57) ### ### ('anyway', 26, 26) ('″', 36, 19) ('absently', 2, 156) ('¹', 5, 158) ('貴', 6, 167) ('খ', 11, 151) ### ### ('##ப', 12, 153) ('worried', 68, 43) ('ம', 14, 146) ('intently', 8, 185) ('幸', 13, 166) ### ### ('universe', 22, 116) ('##子', 74, 56) ('galaxy', 418, 16) ### ############################################################################################################ [2023-11-25 19:05:31,503][root][INFO] - Epoch: 4: Step: 6501/19530, loss[v]=0.057398, lr=0.000008, acc@1[1]=1000.5/1024=0.97705078125, acc@1[2]=1018.0/1024=0.994140625 [2023-11-25 19:07:20,064][root][INFO] - Train batch 6600 [2023-11-25 19:07:20,065][root][INFO] - Avg. loss per last 100 batches: 0.131691 [2023-11-25 19:07:20,785][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.76 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.0/29522=99.94% | mean: 0.02 | max: 5.76 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.5/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the bernoulli numbers are related to fermat's last theorem ( flt ) by kummer's theorem, ### ### which says : prime numbers with this property are called regular prime s. another classical result ### ### of kummer are the following congruences. ### ### [P_TEXT]: herbrand – ribet theorem in mathematics, the herbrand – ribet theorem is a result on the ### ### class group of certain number fields. it is a strengthening of ernst kummer's theorem to the effect ### ### that the prime " p " divides the class number of the cyclotomic field of " p " - th roots of unity ### ### if and only if " p " divides the numerator of the " n " - th bernoulli number " b " " n " for some ### ### " n ", 0 < " n " < " p " − 1. the herbrand – ribet theorem specifies what, in particular, it means ### ### when " p " divides such an " b " " n ". the galois group δ of the cyclotomic field of " p " th ### ### roots of unity for an odd prime " p ", q ( ζ ) ### ### =============================================== v_q ================================================ ### ### ('##mmer', 0, 1) ('absently', 1, 162) ('regular', 2, 17939) ('prime', 3, 34) ('貴', 4, 106) ### ### ('¹', 5, 131) ('##ப', 6, 149) ('幸', 7, 144) ('খ', 8, 115) ('ம', 9, 104) ('intently', 10, 166) ### ### ('theorem', 11, 3) ('ku', 12, 29) ('bern', 13, 12) ('.', 14, 1676) ('³', 15, 1051) ('##ou', 16, 31) ### ### ('number', 17, 35) ('mathematician', 18, 1940) ('cong', 19, 19839) ### ### =============================================== v_p ================================================ ### ### ('class', 680, 0) ('##mmer', 0, 1) ('herb', 6527, 2) ('theorem', 11, 3) ('fields', 138, 4) ### ### ('rib', 15261, 5) ('unity', 2900, 6) ('gal', 4913, 7) ('##ator', 5599, 8) ('##rand', 2819, 9) ### ### ('##oto', 25148, 10) ('##ois', 19763, 11) ('bern', 13, 12) ('ζ', 602, 13) ('ernst', 2311, 14) ### ### ('##ators', 25837, 15) ('δ', 418, 16) ('roots', 1096, 17) ('field', 252, 18) ('leaned', 1688, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##mmer', 0, 1) ('theorem', 11, 3) ('prime', 3, 34) ('bern', 13, 12) ('ku', 12, 29) ### ### ('##ou', 16, 31) ('number', 17, 35) ('##lli', 23, 38) ('numbers', 22, 39) ('anyway', 35, 27) ### ### ('worried', 27, 40) ('fields', 138, 4) ('貴', 4, 106) ('absently', 1, 162) ('ம', 9, 104) ### ### ('খ', 8, 115) ('mathematics', 33, 48) ('¹', 5, 131) ('幸', 7, 144) ('##ப', 6, 149) ### ############################################################################################################ [2023-11-25 19:07:21,292][root][INFO] - Epoch: 4: Step: 6601/19530, loss[v]=0.159808, lr=0.000008, acc@1[1]=979.0/1024=0.9560546875, acc@1[2]=1004.0/1024=0.98046875 [2023-11-25 19:09:10,265][root][INFO] - Train batch 6700 [2023-11-25 19:09:10,266][root][INFO] - Avg. loss per last 100 batches: 0.127273 [2023-11-25 19:09:10,987][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.7/29522=99.94% | mean: 0.02 | max: 5.84 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: on 25 june 1975 following an adverse court decision against her, indira gandhi declared a ### ### national emergency, delayed elections, censored ### ### [P_TEXT]: the emergency ( india ) and actions taken during them, also likely played a role in this ### ### lack of success. although special tribunals were organised and scores of senior congress party and ### ### government officials arrested and charged, including mrs gandhi and sanjay gandhi, police were ### ### unable to submit sufficient evidence for most cases, and only a few low - level officials were ### ### convicted of any abuses. the emergency lasted 21 months, and its legacy remains intensely ### ### controversial. a few days after the emergency was imposed, the bombay edition of " the times of ### ### india " carried an obituary that read a few ### ### =============================================== v_q ================================================ ### ### ('indira', 0, 1037) ('1975', 1, 19855) ('adverse', 2, 8606) ('absently', 3, 183) ('gandhi', 4, 1) ### ### ('貴', 5, 218) ('emergency', 6, 0) ('intently', 7, 322) ('.', 8, 110) ('india', 9, 8) ('খ', 10, 167) ### ### ('##ப', 11, 236) ('##nsor', 12, 19748) ('幸', 13, 274) ('ம', 14, 164) ('！', 15, 592) ('¹', 16, 245) ### ### ('##子', 17, 163) ('sonia', 18, 1755) ('decision', 19, 2217) ### ### =============================================== v_p ================================================ ### ### ('emergency', 6, 0) ('gandhi', 4, 1) ('obituary', 17664, 2) ('tribunal', 204, 3) ### ### ('sanjay', 1494, 4) ('bombay', 698, 5) ('mrs', 1204, 6) ('edition', 11047, 7) ('india', 9, 8) ### ### ('congress', 838, 9) ('times', 744, 10) ('police', 1330, 11) ('emergencies', 13840, 12) ### ### ('anyway', 97, 13) ('reads', 24603, 14) ('acquitted', 18647, 15) ('special', 2141, 16) ### ### ('editions', 15350, 17) ('senior', 6965, 18) ('worried', 120, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('emergency', 6, 0) ('gandhi', 4, 1) ('india', 9, 8) ('tribunal', 204, 3) ('herself', 24, 28) ### ### ('indian', 23, 54) ('.', 8, 110) ('anyway', 97, 13) ('bombay', 698, 5) ('absently', 3, 183) ### ### ('worried', 120, 19) ('খ', 10, 167) ('貴', 5, 218) ('ம', 14, 164) ('sanjay', 1494, 4) ### ### ('##ப', 11, 236) ('##子', 17, 163) ('mrs', 1204, 6) ('intently', 7, 322) ('court', 21, 138) ### ############################################################################################################ [2023-11-25 19:09:11,492][root][INFO] - Epoch: 4: Step: 6701/19530, loss[v]=0.068792, lr=0.000008, acc@1[1]=1002.5/1024=0.97900390625, acc@1[2]=1021.0/1024=0.9970703125 [2023-11-25 19:11:00,627][root][INFO] - Train batch 6800 [2023-11-25 19:11:00,628][root][INFO] - Avg. loss per last 100 batches: 0.140657 [2023-11-25 19:11:01,340][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.9/29522=99.94% | mean: 0.02 | max: 5.67 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: between the two clubs, sport donnelly, ed malley and pudge heffelfinger of the chicago ### ### athletic association football team appeared in the allegheny line - up. ### ### [P_TEXT]: pittsburgh athletic club ( football ) the pittsburgh athletic club or the pittsburg ### ### athletic club football team, established in 1891, was based in pittsburgh, pennsylvania. in 1892 ### ### the intense competition between two pittsburgh - area clubs, the allegheny athletic association and ### ### the pittsburgh athletic club, led to william ( pudge ) heffelfinger becoming the first known ### ### professional football player. heffelfinger was paid $ 500 by allegheny to play in a game against ### ### pittsburgh on november 12, 1892. as a result, heffelfinger became the first person to be paid to ### ### play football. allegheny would go on to win the game, 4 - 0, when heffelfinger picked up a ### ### =============================================== v_q ================================================ ### ### ('chicago', 0, 1081) ('donnelly', 1, 10863) ('sport', 2, 194) ('allegheny', 3, 2) ### ### ('absently', 4, 97) ('malley', 5, 6947) ('ed', 6, 20373) ('##ffe', 7, 8) ('athletic', 8, 3) ### ### ('football', 9, 9) ('খ', 10, 144) ('¹', 11, 174) ('貴', 12, 117) ('pu', 13, 18) ('##lf', 14, 11) ### ### ('幸', 15, 168) ('intently', 16, 121) ('##ப', 17, 162) ('ம', 18, 137) ('fame', 19, 2940) ### ### =============================================== v_p ================================================ ### ### ('pittsburgh', 32, 0) ('worried', 854, 1) ('allegheny', 3, 2) ('athletic', 8, 3) ('pitt', 10872, 4) ### ### ('anyway', 200, 5) ('##sburg', 1065, 6) ('##dge', 22, 7) ('##ffe', 7, 8) ('football', 9, 9) ### ### ('##inger', 27, 10) ('##lf', 14, 11) ('1892', 343, 12) ('club', 111, 13) ('william', 4209, 14) ### ### ('association', 21, 15) ('professionals', 2884, 16) ('clubs', 534, 17) ('pu', 13, 18) ### ### ('$', 870, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('allegheny', 3, 2) ('athletic', 8, 3) ('##ffe', 7, 8) ('football', 9, 9) ('##lf', 14, 11) ### ### ('pittsburgh', 32, 0) ('##dge', 22, 7) ('pu', 13, 18) ('##inger', 27, 10) ('association', 21, 15) ### ### ('anyway', 200, 5) ('club', 111, 13) ('absently', 4, 97) ('shirt', 36, 30) ('worried', 854, 1) ### ### ('sport', 2, 194) ('貴', 12, 117) ('teams', 98, 21) ('1892', 343, 12) ('intently', 16, 121) ### ############################################################################################################ [2023-11-25 19:11:01,846][root][INFO] - Epoch: 4: Step: 6801/19530, loss[v]=0.177225, lr=0.000008, acc@1[1]=978.0/1024=0.955078125, acc@1[2]=1008.0/1024=0.984375 [2023-11-25 19:12:51,163][root][INFO] - Train batch 6900 [2023-11-25 19:12:51,163][root][INFO] - Avg. loss per last 100 batches: 0.133601 [2023-11-25 19:12:51,894][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 3.94 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.1/29522=99.94% | mean: 0.02 | max: 5.82 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it extends for from interstate 90 ( i - 90 ) exit 8 in north greenbush to the ### ### massachusetts state line, where it continues into williamstown as massachusetts route 43. ### ### [P_TEXT]: massachusetts route 43 route 43 is a extension of new york state route 43 ( ny 43 ) into ### ### massachusetts. unlike ny 43, which is signed east - west, massachusetts route 43 is a north - south ### ### route. route 43's northern terminus is at route 2 in williamstown. along the way it intersects u. ### ### s. route 7 ( us 7 ) in williamstown. massachusetts route 43 begins at the eastern end of new york ### ### state route 43 in hancock. it turns northward through the kinderhook creek valley in the taconic ### ### mountains, passing near jiminy peak, and passing through the northern half of hancock before ### ### entering williamstown. in williamstown, the ### ### =============================================== v_q ================================================ ### ### ('43', 0, 1) ('massachusetts', 1, 11) ('##bush', 2, 20778) ('90', 3, 14409) ('williams', 4, 3) ### ### ('green', 5, 11186) ('¹', 6, 119) ('##town', 7, 15) ('貴', 8, 116) ('absently', 9, 92) ### ### ('intently', 10, 113) ('##ப', 11, 130) ('.', 12, 243) ('খ', 13, 163) ('幸', 14, 147) ('³', 15, 723) ### ### ('route', 16, 5) ('ம', 17, 110) ('boston', 18, 53) ('pond', 19, 2817) ### ### =============================================== v_p ================================================ ### ### ('hancock', 1105, 0) ('43', 0, 1) ('jimi', 18250, 2) ('williams', 4, 3) ('ny', 243, 4) ### ### ('route', 16, 5) ('kind', 14610, 6) ('worried', 302, 7) ('peak', 3200, 8) ('##ic', 23427, 9) ### ### ('7', 7829, 10) ('massachusetts', 1, 11) ('anyway', 310, 12) ('##ok', 15338, 13) ### ### ('##con', 24533, 14) ('##town', 7, 15) ('2', 10357, 16) ('ta', 21461, 17) ('mountains', 4082, 18) ### ### ('routes', 91, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('43', 0, 1) ('williams', 4, 3) ('massachusetts', 1, 11) ('##town', 7, 15) ('route', 16, 5) ### ### ('boston', 18, 53) ('##43', 22, 45) ('¹', 6, 119) ('absently', 9, 92) ('ny', 243, 4) ### ### ('hancock', 1105, 0) ('routes', 91, 19) ('貴', 8, 116) ('intently', 10, 113) ('worried', 302, 7) ### ### ('creek', 134, 20) ('driveway', 34, 48) ('##ப', 11, 130) ('anyway', 310, 12) ('york', 84, 24) ### ############################################################################################################ [2023-11-25 19:12:52,401][root][INFO] - Epoch: 4: Step: 6901/19530, loss[v]=0.078675, lr=0.000008, acc@1[1]=1001.5/1024=0.97802734375, acc@1[2]=1015.5/1024=0.99169921875 [2023-11-25 19:14:41,787][root][INFO] - Train batch 7000 [2023-11-25 19:14:41,788][root][INFO] - Avg. loss per last 100 batches: 0.129415 [2023-11-25 19:14:42,498][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.3/29522=99.93% | mean: 0.02 | max: 5.66 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 29.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: pamela s. chasek ( born 1961 ) is a professor in the department of political science at ### ### manhattan college, and editor of the " earth negotiations bulletin ". ### ### [P_TEXT]: david l. downie author of numerous publications on a variety of topics, his co - authored ### ### book, " global environmental politics ", written with professor pamela chasek ( a co - creator of ### ### the earth negotiations bulletin ), is one of the most widely used in the field. david downie joined ### ### the politics department at fairfield university in 2008. he had been recruited to fairfield to ### ### become the first director of the revamped environmental studies program. prior to moving to ### ### fairfield university, downie taught courses in international environmental politics at columbia ### ### university from 1994 - 2008. while at columbia, dr. downie held ### ### =============================================== v_q ================================================ ### ### ('chase', 0, 1) ('pamela', 1, 3) ('negotiations', 2, 4) ('manhattan', 3, 431) ('¹', 4, 481) ### ### ('college', 5, 82) ('##k', 6, 39) ('earth', 7, 11) ('bulletin', 8, 2) ('³', 9, 3710) ### ### ('editor', 10, 95) ('貴', 11, 161) ('##ப', 12, 113) ('talks', 13, 24) ('幸', 14, 189) ### ### ('intently', 15, 205) ('political', 16, 57) ('negotiating', 17, 171) ('খ', 18, 200) ### ### ('university', 19, 15) ### ### =============================================== v_p ================================================ ### ### ('fairfield', 16144, 0) ('chase', 0, 1) ('bulletin', 8, 2) ('pamela', 1, 3) ('negotiations', 2, 4) ### ### ('down', 11212, 5) ('environmental', 1095, 6) ('politics', 44, 7) ('##ie', 21156, 8) ### ### ('david', 835, 9) ('columbia', 521, 10) ('earth', 7, 11) ('studies', 991, 12) ('global', 145, 13) ### ### ('semester', 240, 14) ('university', 19, 15) ('l', 6229, 16) ('anyway', 788, 17) ### ### ('herself', 644, 18) ('negotiation', 21, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('chase', 0, 1) ('pamela', 1, 3) ('negotiations', 2, 4) ('bulletin', 8, 2) ('earth', 7, 11) ### ### ('##k', 6, 39) ('university', 19, 15) ('politics', 44, 7) ('talks', 13, 24) ('negotiation', 21, 19) ### ### ('college', 5, 82) ('global', 145, 13) ('political', 16, 57) ('semester', 240, 14) ### ### ('editor', 10, 95) ('faculty', 79, 27) ('columbia', 521, 10) ('manhattan', 3, 431) ('##ப', 12, 113) ### ### ('professor', 59, 41) ### ############################################################################################################ [2023-11-25 19:14:43,002][root][INFO] - Epoch: 4: Step: 7001/19530, loss[v]=0.146672, lr=0.000008, acc@1[1]=981.0/1024=0.9580078125, acc@1[2]=1001.5/1024=0.97802734375 [2023-11-25 19:16:31,908][root][INFO] - Train batch 7100 [2023-11-25 19:16:31,909][root][INFO] - Avg. loss per last 100 batches: 0.131493 [2023-11-25 19:16:32,625][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.8/29522=99.94% | mean: 0.02 | max: 5.85 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: east of the tacoma dome area, i - 5 intersects sr 167 and crosses over the puyallup river ### ### and a railroad carrying sounder commuter trains. ### ### [P_TEXT]: washington state route 167 to a freeway since the mid - 20th century. wsdot is performing ### ### advance design and engineering work. the proposed highway would bypass the snarled traffic at ### ### meridian way in puyallup and continue across interstate 5 to sr 509 in the city of fife. the ### ### project would construct a new six mile ( 10 km ) freeway north of the puyallup river and complete ### ### the partial interchange at meridian way ( sr 161 ) with a single point urban interchange ( spui ). ### ### the new freeway would be four lanes with adequate space for future hov lanes. the interchange with ### ### i - 5 would ### ### =============================================== v_q ================================================ ### ### ('5', 0, 11) ('tacoma', 1, 386) ('¹', 2, 163) ('dome', 3, 12305) ('sound', 4, 14513) ('167', 5, 0) ### ### ('i', 6, 65) ('pu', 7, 17) ('幸', 8, 187) ('##ப', 9, 130) ('##er', 10, 23458) ('খ', 11, 185) ### ### ('seattle', 12, 54) ('absently', 13, 111) ('washington', 14, 13) ('intently', 15, 135) ### ### ('interstate', 16, 27) ('ம', 17, 114) ('貴', 18, 144) ('##lu', 19, 9) ### ### =============================================== v_p ================================================ ### ### ('167', 5, 0) ('fife', 6234, 1) ('meridian', 10814, 2) ('161', 11090, 3) ('##sd', 25038, 4) ### ### ('##yal', 22, 5) ('##9', 18771, 6) ('##ui', 20968, 7) ('50', 1712, 8) ('##lu', 19, 9) ### ### ('lanes', 537, 10) ('5', 0, 11) ('way', 465, 12) ('washington', 14, 13) ('anyway', 167, 14) ### ### ('urban', 573, 15) ('snarled', 13190, 16) ('pu', 7, 17) ('sp', 8575, 18) ('##v', 8129, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('167', 5, 0) ('5', 0, 11) ('##lu', 19, 9) ('##yal', 22, 5) ('washington', 14, 13) ('pu', 7, 17) ### ### ('sr', 21, 20) ('interstate', 16, 27) ('##p', 25, 21) ('seattle', 12, 54) ('i', 6, 65) ### ### ('route', 30, 33) ('¹', 2, 163) ('routes', 41, 34) ('absently', 13, 111) ('5th', 29, 59) ### ### ('##ப', 9, 130) ('ம', 17, 114) ('river', 49, 35) ('anyway', 167, 14) ### ############################################################################################################ [2023-11-25 19:16:33,130][root][INFO] - Epoch: 4: Step: 7101/19530, loss[v]=0.052690, lr=0.000008, acc@1[1]=1001.0/1024=0.9775390625, acc@1[2]=1019.0/1024=0.9951171875 [2023-11-25 19:18:22,690][root][INFO] - Train batch 7200 [2023-11-25 19:18:22,691][root][INFO] - Avg. loss per last 100 batches: 0.129581 [2023-11-25 19:18:23,415][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.9/29522=99.95% | mean: 0.03 | max: 5.83 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the rias baixas, found south of fisterra, include corcubion, muros e noia, arousa, ### ### pontevedra and vigo. ### ### [P_TEXT]: serra da barbanza the serra da barbanza ( spanish : " sierra del barbanza " ) is a ### ### mountain range in north - west spain, extending in a south - westerly direction from noia in ### ### galicia. the range forms the peninsula between the two ria s of ria de muros e noia and ria de ### ### arousa in the extreme west of galicia. due to the relatively large height of over with respect to ### ### the proximity to the two rias it forms an impressive mountain range though small in length. highest ### ### peaks ### ### =============================================== v_q ================================================ ### ### ('pont', 0, 21622) ('##ub', 1, 29155) ('fist', 2, 12703) ('##ous', 3, 6) ('absently', 4, 100) ### ### ('##dra', 5, 22153) ('##ros', 6, 9) ('ri', 7, 2) ('vi', 8, 20302) ('##rc', 9, 27344) ### ### ('##go', 10, 22083) ('leaned', 11, 21) ('খ', 12, 142) ('##ப', 13, 119) ('幸', 14, 164) ### ### ('ம', 15, 105) ('貴', 16, 96) ('##ion', 17, 10034) ('sardinia', 18, 1316) ('.', 19, 847) ### ### =============================================== v_p ================================================ ### ### ('galicia', 42, 0) ('no', 26, 1) ('ri', 7, 2) ('##ia', 45, 3) ('worried', 78, 4) ('serra', 513, 5) ### ### ('##ous', 3, 6) ('##ban', 17245, 7) ('##as', 34, 8) ('##ros', 6, 9) ('anyway', 86, 10) ### ### ('spain', 63, 11) ('peninsula', 232, 12) ('peaks', 10087, 13) ('ar', 22, 14) ('bar', 12151, 15) ### ### ('spanish', 125, 16) ('mu', 20, 17) ('sierra', 754, 18) ('##a', 87, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ri', 7, 2) ('##ous', 3, 6) ('##ros', 6, 9) ('no', 26, 1) ('galicia', 42, 0) ('##ia', 45, 3) ### ### ('##as', 34, 8) ('leaned', 11, 21) ('ar', 22, 14) ('mu', 20, 17) ('worried', 78, 4) ### ### ('spain', 63, 11) ('anyway', 86, 10) ('absently', 4, 100) ('##a', 87, 19) ('貴', 16, 96) ### ### ('ம', 15, 105) ('##ப', 13, 119) ('galician', 66, 33) ('inhabitants', 76, 28) ### ############################################################################################################ [2023-11-25 19:18:23,923][root][INFO] - Epoch: 4: Step: 7201/19530, loss[v]=0.196345, lr=0.000008, acc@1[1]=972.5/1024=0.94970703125, acc@1[2]=999.0/1024=0.9755859375 [2023-11-25 19:20:13,046][root][INFO] - Train batch 7300 [2023-11-25 19:20:13,047][root][INFO] - Avg. loss per last 100 batches: 0.131712 [2023-11-25 19:20:13,752][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.3/29522=99.94% | mean: 0.02 | max: 5.89 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: time, it is possible that mussato began his tutorship under lovato lovati. ### ### [P_TEXT]: lovato lovati correct them and ascertain their meaning, and a desire to imitate them. ### ### noted by petrarch, it was lovato's study of law which introduced him to the ancient literature. ### ### however while drawing on ancient roman sources for their own work, lovato and his disciple ### ### albertino mussato, showed little interest in investigating the nature of the society, that produced ### ### them. lovato's poetry exposed his confused notion of the ancient past, where greek and roman ### ### history and mythology mingle without distinction as their basis in reality. yet, lovati's success ### ### demonstrated his profound intensive study of ancient law, ### ### =============================================== v_q ================================================ ### ### ('##ssa', 0, 2) ('intently', 1, 251) ('¹', 2, 176) ('##to', 3, 20) ('lo', 4, 4) ('##vati', 5, 0) ### ### ('貴', 6, 214) ('##vat', 7, 1) ('幸', 8, 190) ('খ', 9, 185) ('##ப', 10, 155) ('mu', 11, 22) ### ### ('ம', 12, 144) ('³', 13, 3312) ('absently', 14, 139) ('sparhawk', 15, 1375) ('tutor', 16, 282) ### ### ('anyway', 17, 21) ('italian', 18, 46) ('belgarath', 19, 839) ### ### =============================================== v_p ================================================ ### ### ('##vati', 5, 0) ('##vat', 7, 1) ('##ssa', 0, 2) ('albert', 22574, 3) ('lo', 4, 4) ### ### ('petra', 3353, 5) ('law', 7106, 6) ('ming', 13291, 7) ('disciple', 733, 8) ('myths', 19952, 9) ### ### ('greek', 28, 10) ('past', 7644, 11) ('ancient', 46, 12) ('roman', 715, 13) ('leaned', 5621, 14) ### ### ('myth', 4138, 15) ('greeks', 2191, 16) ('confused', 1995, 17) ('literature', 199, 18) ### ### ('poetry', 1242, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ssa', 0, 2) ('##vati', 5, 0) ('##vat', 7, 1) ('lo', 4, 4) ('##to', 3, 20) ('mu', 11, 22) ### ### ('anyway', 17, 21) ('greek', 28, 10) ('ancient', 46, 12) ('italian', 18, 46) ('italy', 32, 24) ### ### ('¹', 2, 176) ('study', 52, 28) ('intently', 1, 251) ('ம', 12, 144) ('absently', 14, 139) ### ### ('##ப', 10, 155) ('貴', 6, 214) ('幸', 8, 190) ('খ', 9, 185) ### ############################################################################################################ [2023-11-25 19:20:14,256][root][INFO] - Epoch: 4: Step: 7301/19530, loss[v]=0.057676, lr=0.000008, acc@1[1]=1005.0/1024=0.9814453125, acc@1[2]=1018.5/1024=0.99462890625 [2023-11-25 19:22:03,324][root][INFO] - Train batch 7400 [2023-11-25 19:22:03,324][root][INFO] - Avg. loss per last 100 batches: 0.135333 [2023-11-25 19:22:04,038][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.78 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.3/29522=99.94% | mean: 0.02 | max: 5.66 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it is also known as a signature song for carlos toshiki's time as the lead vocalist ### ### alongside the single " kimi wa 1000 %. " ### ### [P_TEXT]: carlos toshiki carlos toshiki ( born as carlos toshiki takahashi on 7 april 1964 ) is a ### ### japanese brazilian singer - songwriter who was the lead vocalist of the popular japanese city pop ### ### band 1986 omega tribe and its successor band carlos toshiki & omega tribe during the mid to late ### ### 80's. he was a popular vocalist during the band's era, with the singles " kimi wa 1000 % " and " ### ### aquamarine no mama de ite " being hit singles on the oricon charts. born in southern brazil, ### ### toshiki imitated songs that his father used ### ### =============================================== v_q ================================================ ### ### ('carlos', 0, 1) ('1000', 1, 5) ('kim', 2, 9) ('%', 3, 20) ('″', 4, 22) ('##ki', 5, 15) ### ### ('##shi', 6, 6) ('¹', 7, 159) ('」', 8, 399) ('³', 9, 539) ('幸', 10, 132) ('##i', 11, 140) ### ### ('貴', 12, 118) ('intently', 13, 114) ('##ப', 14, 126) ('japan', 15, 23) ('খ', 16, 141) ### ### ('！', 17, 529) ('percent', 18, 98) ('popularity', 19, 41) ### ### =============================================== v_p ================================================ ### ### ('omega', 4228, 0) ('carlos', 0, 1) ('tribe', 2159, 2) ('##marine', 21403, 3) ('takahashi', 431, 4) ### ### ('1000', 1, 5) ('##shi', 6, 6) ('aqua', 2180, 7) ('brazil', 4568, 8) ('kim', 2, 9) ('ט', 27, 10) ### ### ('brazilian', 2271, 11) ('1986', 5035, 12) ('japanese', 25, 13) ('worried', 345, 14) ### ### ('##ki', 5, 15) ('mama', 785, 16) ('anyway', 79, 17) ('city', 9121, 18) ('wa', 20, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('carlos', 0, 1) ('1000', 1, 5) ('kim', 2, 9) ('##shi', 6, 6) ('%', 3, 20) ('##ki', 5, 15) ### ### ('″', 4, 22) ('japanese', 25, 13) ('wa', 20, 19) ('ט', 27, 10) ('japan', 15, 23) ### ### ('popularity', 19, 41) ('anyway', 79, 17) ('##い', 28, 48) ('singles', 30, 39) ('¹', 7, 159) ### ### ('songs', 33, 36) ('##き', 29, 52) ('song', 35, 31) ('intently', 13, 114) ### ############################################################################################################ [2023-11-25 19:22:04,544][root][INFO] - Epoch: 4: Step: 7401/19530, loss[v]=0.160072, lr=0.000008, acc@1[1]=983.5/1024=0.96044921875, acc@1[2]=1001.5/1024=0.97802734375 [2023-11-25 19:23:54,008][root][INFO] - Train batch 7500 [2023-11-25 19:23:54,009][root][INFO] - Avg. loss per last 100 batches: 0.140928 [2023-11-25 19:23:54,731][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 3.69 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.4/29522=99.94% | mean: 0.02 | max: 5.60 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: his neighbours there included another world champion, hennie kuiper. ### ### [P_TEXT]: sport in the netherlands. in 2017, tom dumoulin won the giro d'italia, becoming the first ### ### dutchman to win the race and the first dutch grand tour winner since zoetemelk's tour win in 1980. ### ### seven dutch men have been world champion on the road ( theo middelkamp ( 1947 ), jan janssen ( 1964 ### ### ), harm ottenbros ( 1969 ), hennie kuiper ( 1975 ), gerrie knetemann ( 1978 ), jan raas ( 1979 ), ### ### joop zoetemelk ( 1985 ) ), but only one dutchman has won the time trial world championship tom ### ### dumoulin ( 2017 ). eight ### ### =============================================== v_q ================================================ ### ### ('##ipe', 0, 2) ('貴', 1, 93) ('¹', 2, 55) ('hen', 3, 32) ('intently', 4, 113) ('##nie', 5, 48) ### ### ('speedway', 6, 900) ('##ப', 7, 106) ('幸', 8, 104) ('ku', 9, 80) ('world', 10, 39) ('খ', 11, 124) ### ### ('³', 12, 396) ('chess', 13, 2325) ('anyway', 14, 7) ('ம', 15, 103) ('ɐ', 16, 2140) ### ### ('##r', 17, 223) ('handball', 18, 5640) ('##¨', 19, 1525) ### ### =============================================== v_p ================================================ ### ### ('##ulin', 26175, 0) ('##ssen', 14128, 1) ('##ipe', 0, 2) ('netherlands', 337, 3) ('joo', 14865, 4) ### ### ('##tem', 28238, 5) ('giro', 790, 6) ('anyway', 14, 7) ('trial', 3114, 8) ('##kam', 25466, 9) ### ### ('1980', 232, 10) ('##rrie', 8728, 11) ('dutch', 705, 12) ('theo', 13853, 13) ('zoe', 13251, 14) ### ### ('harm', 16319, 15) ('tom', 1488, 16) ('##as', 8776, 17) ('2017', 21814, 18) ('littered', 25, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ipe', 0, 2) ('anyway', 14, 7) ('hen', 3, 32) ('world', 10, 39) ('littered', 25, 19) ### ### ('¹', 2, 55) ('##nie', 5, 48) ('貴', 1, 93) ('sport', 86, 23) ('ku', 9, 80) ('netherlands', 337, 3) ### ### ('1980', 232, 10) ('intently', 4, 113) ('worlds', 29, 57) ('##ப', 7, 106) ('championships', 27, 62) ### ### ('幸', 8, 104) ('ம', 15, 103) ('খ', 11, 124) ('race', 218, 30) ### ############################################################################################################ [2023-11-25 19:23:55,238][root][INFO] - Epoch: 4: Step: 7501/19530, loss[v]=0.174714, lr=0.000008, acc@1[1]=984.5/1024=0.96142578125, acc@1[2]=1002.0/1024=0.978515625 [2023-11-25 19:25:44,398][root][INFO] - Train batch 7600 [2023-11-25 19:25:44,399][root][INFO] - Avg. loss per last 100 batches: 0.129976 [2023-11-25 19:25:45,102][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.3/29522=99.95% | mean: 0.03 | max: 5.87 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: cats, dogs, mice and rat s were the next selection for the menu. ### ### [P_TEXT]: siege of paris ( 1870 – 71 ) declared, " the frenchmen of 1870 are the sons of those ### ### gauls for whom battles were holidays. " due to a severe shortage of food, parisians were forced to ### ### slaughter whatever animals were at hand. rats, dogs, cats, and horses were the first to be ### ### slaughtered and became regular fare on restaurant menus. once the supply of those animals ran low, ### ### the citizens of paris turned on the zoo animals residing at jardin des plantes. even castor and ### ### pollux, the only pair of elephants in paris, were slaughtered ### ### =============================================== v_q ================================================ ### ### ('menu', 0, 15) ('.', 1, 251) ('rat', 2, 57) ('absently', 3, 314) ('貴', 4, 360) ('mice', 5, 151) ### ### ('intently', 6, 392) ('cats', 7, 25) ('##ப', 8, 227) ('ம', 9, 248) ('খ', 10, 323) ('幸', 11, 354) ### ### ('¹', 12, 331) ('mouse', 13, 261) ('##部', 14, 280) ('game', 15, 5753) ('anyway', 16, 22) ### ### ('worried', 17, 61) ('cat', 18, 192) ('commodore', 19, 22194) ### ### =============================================== v_p ================================================ ### ### ('siege', 25085, 0) ('poll', 11680, 1) ('rats', 114, 2) ('paris', 4618, 3) ('elephants', 23011, 4) ### ### ('besieged', 29229, 5) ('cast', 17751, 6) ('holidays', 13260, 7) ('slaughtered', 28502, 8) ### ### ('plant', 2404, 9) ('##ux', 20385, 10) ('##din', 26815, 11) ('1870', 10909, 12) ('gaul', 22747, 13) ### ### ('zoo', 377, 14) ('menu', 0, 15) ('##or', 17747, 16) ('elephant', 1434, 17) ('animals', 388, 18) ### ### ('horses', 12252, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('menu', 0, 15) ('cats', 7, 25) ('rat', 2, 57) ('anyway', 16, 22) ('dog', 23, 32) ('rats', 114, 2) ### ### ('dogs', 32, 24) ('worried', 17, 61) ('mice', 5, 151) ('.', 1, 251) ('##ப', 8, 227) ### ### ('restaurant', 62, 42) ('absently', 3, 314) ('ம', 9, 248) ('貴', 4, 360) ('zoo', 377, 14) ### ### ('food', 56, 55) ('animal', 112, 40) ('খ', 10, 323) ('intently', 6, 392) ### ############################################################################################################ [2023-11-25 19:25:45,606][root][INFO] - Epoch: 4: Step: 7601/19530, loss[v]=0.179404, lr=0.000008, acc@1[1]=973.5/1024=0.95068359375, acc@1[2]=1006.0/1024=0.982421875 [2023-11-25 19:27:34,852][root][INFO] - Train batch 7700 [2023-11-25 19:27:34,853][root][INFO] - Avg. loss per last 100 batches: 0.135005 [2023-11-25 19:27:35,590][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.99 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.9/29522=99.95% | mean: 0.03 | max: 5.93 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: george town was, since colonial times, a free port - until its sudden revocation by the ### ### malaysian federal government in 1969. ### ### [P_TEXT]: george town, penang by the british, george town's economy is now dominated by other ### ### tertiary sub - sectors ranging from manufacturing to finance, whilst newer industries, including ### ### entrepreneurial startups, are taking root within the city as well. in addition, george town serves ### ### as the economic pole of northern malaysia, with relatively wide logistical connectivity. the penang ### ### international airport is one of the nation's busiest, whilst swettenham pier has cemented the ### ### city's reputation as a popular destination for cruise shipping. since the 1970s, manufacturing ### ### formed the backbone of penang's economy, generating 44. 8 % of ### ### =============================================== v_q ================================================ ### ### ('george', 0, 1) ('town', 1, 2) ('penang', 2, 0) ('¹', 3, 95) ('malaysia', 4, 18) ('貴', 5, 114) ### ### ('port', 6, 505) ('খ', 7, 134) ('##ப', 8, 109) ('幸', 9, 143) ('intently', 10, 149) ('ம', 11, 70) ### ### ('free', 12, 10559) ('absently', 13, 54) ('.', 14, 226) ('kuala', 15, 1672) ('malaysian', 16, 35) ### ### ('##部', 17, 188) ('rev', 18, 22314) ('malay', 19, 239) ### ### =============================================== v_p ================================================ ### ### ('penang', 2, 0) ('george', 0, 1) ('town', 1, 2) ('pier', 2925, 3) ('airport', 175, 4) ### ### ('startup', 22093, 5) ('pole', 16952, 6) ('cruise', 647, 7) ('worried', 67, 8) ('##ette', 27660, 9) ### ### ('shipping', 81, 10) ('anyway', 114, 11) ('##nham', 22467, 12) ('economy', 1563, 13) ### ### ('tertiary', 117, 14) ('busiest', 9286, 15) ('sw', 17234, 16) ('northern', 6438, 17) ### ### ('malaysia', 4, 18) ('manufacturing', 12560, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('penang', 2, 0) ('george', 0, 1) ('town', 1, 2) ('malaysia', 4, 18) ('malaysian', 16, 35) ### ### ('worried', 67, 8) ('shipping', 81, 10) ('absently', 13, 54) ('airport', 175, 4) ('ம', 11, 70) ### ### ('¹', 3, 95) ('anyway', 114, 11) ('tertiary', 117, 14) ('貴', 5, 114) ('##ப', 8, 109) ### ### ('ship', 40, 33) ('খ', 7, 134) ('幸', 9, 143) ('economic', 149, 20) ('intently', 10, 149) ### ############################################################################################################ [2023-11-25 19:27:36,099][root][INFO] - Epoch: 4: Step: 7701/19530, loss[v]=0.071284, lr=0.000008, acc@1[1]=999.0/1024=0.9755859375, acc@1[2]=1018.0/1024=0.994140625 [2023-11-25 19:29:25,622][root][INFO] - Train batch 7800 [2023-11-25 19:29:25,623][root][INFO] - Avg. loss per last 100 batches: 0.134083 [2023-11-25 19:29:26,339][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.5/29522=99.93% | mean: 0.02 | max: 5.64 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the film is often considered the first of a trilogy of stillman films, followed by " ### ### barcelona " ( 1994, but written before " metropolitan " ) and " the last days of disco " ( 1998 ). ### ### [P_TEXT]: whit stillman john whitney stillman ( born january 25, 1952 ) is an american writer - ### ### director known for his 1990 film " metropolitan ", which earned him a nomination for the academy ### ### award for best original screenplay. he is also known for his other films, " barcelona " ( 1994 ), " ### ### the last days of disco " ( 1998 ), " damsels in distress " ( 2011 ), as well as his most recent ### ### film, " love & friendship ", which was released in 2016. stillman was born in 1952 in ### ### =============================================== v_q ================================================ ### ### ('barcelona', 0, 0) ('metropolitan', 1, 1) ('disco', 2, 8) ('″', 3, 6) ('still', 4, 5) ### ### ('1994', 5, 28) ('¹', 6, 116) ('days', 7, 16) ('"', 8, 85) ('##man', 9, 12) ('幸', 10, 78) ### ### ('film', 11, 14) ('##ப', 12, 53) ('1998', 13, 29) ('##¨', 14, 1285) ('last', 15, 40) ('貴', 16, 102) ### ### ('films', 17, 26) ('ம', 18, 55) ('godfather', 19, 1741) ### ### =============================================== v_p ================================================ ### ### ('barcelona', 0, 0) ('metropolitan', 1, 1) ('##hit', 27005, 2) ('whitney', 13304, 3) ### ### ('distress', 12210, 4) ('still', 4, 5) ('″', 3, 6) ('friendship', 3553, 7) ('disco', 2, 8) ### ### ('dams', 18956, 9) ('nominees', 9635, 10) ('leaned', 1043, 11) ('##man', 9, 12) ('1990', 217, 13) ### ### ('film', 11, 14) ('screenplay', 252, 15) ('days', 7, 16) ('##els', 15577, 17) ('w', 22611, 18) ### ### ('love', 434, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('barcelona', 0, 0) ('metropolitan', 1, 1) ('″', 3, 6) ('still', 4, 5) ('disco', 2, 8) ### ### ('##man', 9, 12) ('film', 11, 14) ('days', 7, 16) ('1994', 5, 28) ('1998', 13, 29) ### ### ('films', 17, 26) ('last', 15, 40) ('absently', 23, 30) ('##ப', 12, 53) ('filmmakers', 28, 33) ### ### ('"', 8, 85) ('幸', 10, 78) ('ம', 18, 55) ('¹', 6, 116) ('貴', 16, 102) ### ############################################################################################################ [2023-11-25 19:29:26,844][root][INFO] - Epoch: 4: Step: 7801/19530, loss[v]=0.140464, lr=0.000008, acc@1[1]=987.0/1024=0.9638671875, acc@1[2]=1006.5/1024=0.98291015625 [2023-11-25 19:31:15,977][root][INFO] - Train batch 7900 [2023-11-25 19:31:15,978][root][INFO] - Avg. loss per last 100 batches: 0.136683 [2023-11-25 19:31:16,691][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.6/29522=99.94% | mean: 0.02 | max: 5.85 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: straight between the eyes is the sixth studio album by the british hard rock band ### ### rainbow, released in 1982. ### ### [P_TEXT]: rainbow ( rock band ) " difficult to cure " ( 1981 ), " straight between the eyes " ( ### ### 1982 ) and " bent out of shape " ( 1983 ). after breaking up in 1984, blackmore reformed rainbow in ### ### 1993 with a new lineup, which recorded their eighth and last studio album to date " stranger in us ### ### all " ( 1995 ). blackmore's change in direction, from rock to renaissance and medieval - influenced ### ### music, led to rainbow's second dissolution in 1997. he revived the band once again in 2015, and ### ### they still perform live occasionally. over the years rainbow went through many personnel ### ### =============================================== v_q ================================================ ### ### ('rainbow', 0, 0) ('straight', 1, 6) ('eyes', 2, 5) ('¹', 3, 212) ('between', 4, 18) ### ### ('1982', 5, 34) ('″', 6, 10) ('hard', 7, 162) ('³', 8, 937) ('sixth', 9, 1500) ('band', 10, 23) ### ### ('album', 11, 21) ('##½', 12, 425) ('幸', 13, 125) ('খ', 14, 121) ('pained', 15, 802) ('」', 16, 95) ### ### ('"', 17, 170) ('##ப', 18, 85) ('fleetwood', 19, 3684) ### ### =============================================== v_p ================================================ ### ### ('rainbow', 0, 0) ('renaissance', 17220, 1) ('stranger', 4688, 2) ('medieval', 11153, 3) ### ### ('bent', 10293, 4) ('eyes', 2, 5) ('straight', 1, 6) ('difficult', 746, 7) ('shape', 10274, 8) ### ### ('cure', 128, 9) ('″', 6, 10) ('##more', 4575, 11) ('us', 8309, 12) ('eighth', 868, 13) ### ### ('anyway', 787, 14) ('leaned', 6065, 15) ('1997', 7690, 16) ('music', 173, 17) ('between', 4, 18) ### ### ('lineup', 4593, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('rainbow', 0, 0) ('eyes', 2, 5) ('straight', 1, 6) ('″', 6, 10) ('between', 4, 18) ('1982', 5, 34) ### ### ('band', 10, 23) ('album', 11, 21) ('¹', 3, 212) ('cure', 128, 9) ('rock', 25, 33) ### ### ('bands', 51, 27) ('hard', 7, 162) ('1983', 82, 25) ('ம', 21, 62) ('1984', 73, 32) ### ### ('albums', 31, 52) ('difficult', 746, 7) ('幸', 13, 125) ('1981', 46, 42) ### ############################################################################################################ [2023-11-25 19:31:17,196][root][INFO] - Epoch: 4: Step: 7901/19530, loss[v]=0.103249, lr=0.000008, acc@1[1]=997.0/1024=0.9736328125, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 19:33:06,429][root][INFO] - Train batch 8000 [2023-11-25 19:33:06,430][root][INFO] - Avg. loss per last 100 batches: 0.123761 [2023-11-25 19:33:07,159][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.83 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.2/29522=99.94% | mean: 0.02 | max: 5.82 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.4/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it was one of five new piers opened in 2000 on the thames funded by the millennium ### ### commission as part of the thames 2000 project ( the others being blackfriars millennium pier, ### ### millbank millennium pier, tower millennium pier, and westminster millennium pier ), as part of an ### ### integrated transport and regeneration strategy ### ### [P_TEXT]: tower millennium pier - greenwich tourist boats ( operated by city cruises ). private - ### ### charter entertainment boats also use tower pier. tower millennium pier was opened on 14 july by ### ### london mayor ken livingstone. its creation was funded by the millennium commission as part of the ### ### thames 2000 project, and was one of five new piers provided by the commission on the thames ( the ### ### others being blackfriars millennium pier, london eye pier, westminster millennium pier and millbank ### ### millennium pier ). the pier has at its upstream end a small ### ### =============================================== v_q ================================================ ### ### ('millennium', 0, 0) ('pier', 1, 1) ('thames', 2, 8) ('2000', 3, 7) ('westminster', 4, 11) ### ### ('貴', 5, 87) ('幸', 6, 97) ('london', 7, 16) ('##ப', 8, 94) ('intently', 9, 61) ('খ', 10, 133) ### ### ('¹', 11, 54) ('ம', 12, 67) ('##fr', 13, 14) ('absently', 14, 92) ('mill', 15, 25) ### ### ('refurbishment', 16, 2626) ('commission', 17, 4) ('tower', 18, 5) ('.', 19, 3201) ### ### =============================================== v_p ================================================ ### ### ('millennium', 0, 0) ('pier', 1, 1) ('cruises', 3191, 2) ('livingstone', 1822, 3) ### ### ('commission', 17, 4) ('tower', 18, 5) ('eye', 1338, 6) ('2000', 3, 7) ('thames', 2, 8) ### ### ('greenwich', 86, 9) ('anyway', 97, 10) ('westminster', 4, 11) ('project', 28, 12) ### ### ('boats', 2400, 13) ('##fr', 13, 14) ('mayor', 850, 15) ('london', 7, 16) ('hissed', 31, 17) ### ### ('commissions', 7219, 18) ('##bank', 20, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('millennium', 0, 0) ('pier', 1, 1) ('thames', 2, 8) ('2000', 3, 7) ('commission', 17, 4) ### ### ('westminster', 4, 11) ('tower', 18, 5) ('london', 7, 16) ('##fr', 13, 14) ('##bank', 20, 19) ### ### ('mill', 15, 25) ('project', 28, 12) ('hissed', 31, 17) ('piers', 25, 26) ('greenwich', 86, 9) ### ### ('anyway', 97, 10) ('¹', 11, 54) ('black', 22, 34) ('intently', 9, 61) ('貴', 5, 87) ### ############################################################################################################ [2023-11-25 19:33:07,668][root][INFO] - Epoch: 4: Step: 8001/19530, loss[v]=0.147398, lr=0.000008, acc@1[1]=985.0/1024=0.9619140625, acc@1[2]=1003.5/1024=0.97998046875 [2023-11-25 19:34:57,257][root][INFO] - Train batch 8100 [2023-11-25 19:34:57,257][root][INFO] - Avg. loss per last 100 batches: 0.138986 [2023-11-25 19:34:57,963][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.74 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.1/29522=99.93% | mean: 0.02 | max: 5.68 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.3/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the communications and electronics branch ( ) is a personnel branch of the canadian armed ### ### forces ( caf ). ### ### [P_TEXT]: personnel branch uniform ( i. e. " navy ", " army ", " air force " or " special ### ### operations " ) is a function of military occupation, not personnel branch. for example, within the ### ### communications and electronics branch, all signal operators are uniformed army and all aerospace ### ### telecommunication and information systems technicians are designated as air force. on the other ### ### hand, within the royal canadian logistics service, supply technicians may be designated as any ### ### environment. each branch has its own branch advisor — normally a colonel or naval captain — to the ### ### chief of the defence staff. branches have carried forward many of the traditions inherited ### ### =============================================== v_q ================================================ ### ### ('branch', 0, 1) ('canadian', 1, 7) ('communications', 2, 27) ('¹', 3, 191) ('electronics', 4, 6) ### ### ('armed', 5, 370) ('##ப', 6, 124) ('幸', 7, 127) ('canada', 8, 42) ('貴', 9, 130) ### ### ('intently', 10, 190) ('ம', 11, 83) ('personnel', 12, 3) ('³', 13, 1606) ('খ', 14, 129) ### ### ('caf', 15, 23905) ('forces', 16, 163) ('electronic', 17, 74) ('communication', 18, 75) ### ### ('##部', 19, 106) ### ### =============================================== v_p ================================================ ### ### ('uniformed', 650, 0) ('branch', 0, 1) ('technicians', 9635, 2) ('personnel', 12, 3) ### ### ('occupation', 7027, 4) ('defence', 94, 5) ('electronics', 4, 6) ('canadian', 1, 7) ### ### ('aerospace', 1513, 8) ('telecommunications', 121, 9) ('staff', 59, 10) ('logistics', 1240, 11) ### ### ('technician', 1072, 12) ('chief', 574, 13) ('signal', 193, 14) ('captain', 642, 15) ### ### ('air', 370, 16) ('special', 1031, 17) ('uniform', 802, 18) ('anyway', 77, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('branch', 0, 1) ('canadian', 1, 7) ('electronics', 4, 6) ('personnel', 12, 3) ### ### ('communications', 2, 27) ('canada', 8, 42) ('branches', 23, 23) ('staff', 59, 10) ### ### ('telecommunication', 22, 30) ('defence', 94, 5) ('telecommunications', 121, 9) ('army', 31, 31) ### ### ('ம', 11, 83) ('¹', 3, 191) ('military', 52, 25) ('##ப', 6, 124) ('information', 69, 21) ### ### ('uniformed', 650, 0) ('anyway', 77, 19) ('幸', 7, 127) ### ############################################################################################################ [2023-11-25 19:34:58,466][root][INFO] - Epoch: 4: Step: 8101/19530, loss[v]=0.072101, lr=0.000008, acc@1[1]=1000.5/1024=0.97705078125, acc@1[2]=1018.0/1024=0.994140625 [2023-11-25 19:36:47,450][root][INFO] - Train batch 8200 [2023-11-25 19:36:47,450][root][INFO] - Avg. loss per last 100 batches: 0.137178 [2023-11-25 19:36:48,162][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.3/29522=99.93% | mean: 0.02 | max: 5.89 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: sr 522 connects several of the metropolitan area's major highways, including interstate 5 ### ### ( i - 5 ), i - 405, sr 9, and u. s. route 2 ( us 2 ). ### ### [P_TEXT]: interstate 5 in washington i - 5 intersects ravenna boulevard and sr 522, a major highway ### ### that travels along the north side of lake washington. further north, the freeway reaches northgate ### ### and the express lanes merge back with the mainline, forming a set of hov lanes. i - 5 passes to the ### ### west of northgate mall and the northgate light rail station along 1st avenue before moving back ### ### east to 5th avenue near haller lake. at jackson park, freeway intersects sr 523, which runs on ### ### 145th street and forms the northern city ### ### =============================================== v_q ================================================ ### ### ('52', 0, 6) ('2', 1, 1757) ('##2', 2, 26) ('absently', 3, 88) ('intently', 4, 108) ('9', 5, 4370) ### ### ('貴', 6, 94) ('##ப', 7, 100) ('405', 8, 21775) ('幸', 9, 110) ('¹', 10, 69) ('খ', 11, 143) ### ### ('ம', 12, 89) ('5', 13, 4) ('california', 14, 2127) ('sr', 15, 24) ('utah', 16, 807) ### ### ('maine', 17, 1762) ('.', 18, 210) ('florida', 19, 1520) ### ### =============================================== v_p ================================================ ### ### ('halle', 14909, 0) ('raven', 3075, 1) ('##gate', 16718, 2) ('jackson', 1127, 3) ('5', 13, 4) ### ### ('washington', 38, 5) ('52', 0, 6) ('lanes', 445, 7) ('5th', 251, 8) ('lake', 1022, 9) ### ### ('mall', 2550, 10) ('interstate', 27, 11) ('145', 12749, 12) ('1st', 20031, 13) ('##v', 17857, 14) ### ### ('anyway', 73, 15) ('ho', 19174, 16) ('avenue', 116, 17) ('park', 623, 18) ('avenues', 12526, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('52', 0, 6) ('5', 13, 4) ('##2', 2, 26) ('washington', 38, 5) ('interstate', 27, 11) ### ### ('sr', 15, 24) ('worried', 30, 29) ('absently', 3, 88) ('anyway', 73, 15) ('¹', 10, 69) ### ### ('貴', 6, 94) ('i', 34, 30) ('intently', 4, 108) ('##ப', 7, 100) ('幸', 9, 110) ('##₂', 31, 42) ### ### ('5th', 251, 8) ('ம', 12, 89) ('avenue', 116, 17) ('driveway', 40, 46) ### ############################################################################################################ [2023-11-25 19:36:48,668][root][INFO] - Epoch: 4: Step: 8201/19530, loss[v]=0.088155, lr=0.000008, acc@1[1]=999.0/1024=0.9755859375, acc@1[2]=1016.5/1024=0.99267578125 [2023-11-25 19:38:38,107][root][INFO] - Train batch 8300 [2023-11-25 19:38:38,108][root][INFO] - Avg. loss per last 100 batches: 0.133484 [2023-11-25 19:38:38,833][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29505.2/29522=99.94% | mean: 0.02 | max: 5.66 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the dictionary of the scots language ( dsl ) (, ) is an online scots - english ### ### dictionary, run by scottish language dictionaries. ### ### [P_TEXT]: marc alexander ( academic ) corpus, " a linguistically - annotated version of the records ### ### of british parliamentary speech from 1803 to the present. he is also convener of the board of ### ### directors of scottish language dictionaries, which produces the dictionary of the scots language. ### ### =============================================== v_q ================================================ ### ### ('scots', 0, 0) ('dictionary', 1, 5) ('¹', 2, 1240) ('scottish', 3, 7) ('ds', 4, 24705) ### ### ('language', 5, 10) ('english', 6, 1560) ('online', 7, 3846) ('##l', 8, 26775) ('³', 9, 7006) ### ### ('貴', 10, 277) ('##aries', 11, 6) ('ד', 12, 2602) ('幸', 13, 298) ('##ப', 14, 168) ### ### ('irish', 15, 703) ('curious', 16, 1102) ('খ', 17, 237) ('sparhawk', 18, 3296) ('scotland', 19, 29) ### ### =============================================== v_p ================================================ ### ### ('scots', 0, 0) ('marc', 17130, 1) ('corpus', 2166, 2) ('alexander', 2490, 3) ('speech', 531, 4) ### ### ('dictionary', 1, 5) ('##aries', 11, 6) ('scottish', 3, 7) ('directors', 10179, 8) ### ### ('academic', 627, 9) ('language', 5, 10) ('parliamentary', 7819, 11) ('1803', 8208, 12) ### ### ('leaned', 3235, 13) ('linguistic', 1312, 14) ('anyway', 1531, 15) ('copy', 1077, 16) ### ### ('languages', 61, 17) ('1804', 18905, 18) ('board', 558, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('scots', 0, 0) ('dictionary', 1, 5) ('scottish', 3, 7) ('language', 5, 10) ('##aries', 11, 6) ### ### ('##ction', 21, 20) ('scotland', 19, 29) ('languages', 61, 17) ('speech', 531, 4) ('di', 22, 57) ### ### ('academic', 627, 9) ('hissed', 24, 78) ('¹', 2, 1240) ('british', 172, 25) ('##ப', 14, 168) ### ### ('encyclopedia', 44, 51) ('貴', 10, 277) ('corpus', 2166, 2) ('absently', 23, 106) ### ### ('alexander', 2490, 3) ### ############################################################################################################ [2023-11-25 19:38:39,340][root][INFO] - Epoch: 4: Step: 8301/19530, loss[v]=0.171919, lr=0.000008, acc@1[1]=976.5/1024=0.95361328125, acc@1[2]=1003.5/1024=0.97998046875 [2023-11-25 19:40:28,610][root][INFO] - Train batch 8400 [2023-11-25 19:40:28,611][root][INFO] - Avg. loss per last 100 batches: 0.129132 [2023-11-25 19:40:29,322][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.83 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.9/29522=99.94% | mean: 0.02 | max: 5.68 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.1/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he was a descendant of charles applegate, who blazed the oregon trail in 1843 with his ### ### brothers jesse and lindsay and established the applegate trail. ### ### [P_TEXT]: shannon applegate his brothers, jesse applegate and lindsay applegate. her father was ### ### colonel rex applegate, the author of acclaimed military and police training manuals. she lives with ### ### her husband, daniel robertson, on of the land that was a part of charles applegate's original ### ### donation land claim. their home is the " oldest house in the state of oregon still occupied by its ### ### original family ". her daughter, jessica applegate brown, owns and manages applegate house ### ### vineyards at the charles applegate home, often referred to as applegate house. ### ### =============================================== v_q ================================================ ### ### ('apple', 0, 9) ('##gate', 1, 8) ('oregon', 2, 3) ('jesse', 3, 6) ('lindsay', 4, 10) ### ### ('exclaimed', 5, 134) ('charles', 6, 20) ('¹', 7, 199) ('trail', 8, 4232) ('貴', 9, 195) ### ### ('.', 10, 9548) ('intently', 11, 137) ('幸', 12, 197) ('descendant', 13, 9905) ('littered', 14, 37) ### ### ('খ', 15, 218) ('##ப', 16, 173) ('bellowed', 17, 146) ('absently', 18, 119) ### ### ('townspeople', 19, 7177) ### ### =============================================== v_p ================================================ ### ### ('shannon', 7386, 0) ('robertson', 2189, 1) ('rex', 5216, 2) ('oregon', 2, 3) ('donation', 2311, 4) ### ### ('jessica', 13301, 5) ('jesse', 3, 6) ('vineyards', 24877, 7) ('##gate', 1, 8) ('apple', 0, 9) ### ### ('lindsay', 4, 10) ('claim', 2013, 11) ('daniel', 660, 12) ('claims', 1828, 13) ### ### ('training', 10763, 14) ('brother', 34, 15) ('anyway', 277, 16) ('police', 1006, 17) ### ### ('colonel', 952, 18) ('bedrooms', 13394, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##gate', 1, 8) ('oregon', 2, 3) ('apple', 0, 9) ('jesse', 3, 6) ('lindsay', 4, 10) ### ### ('charles', 6, 20) ('brothers', 20, 27) ('littered', 14, 37) ('brother', 34, 15) ### ### ('exclaimed', 5, 134) ('intently', 11, 137) ('anyway', 277, 16) ('¹', 7, 199) ('worried', 137, 30) ### ### ('貴', 9, 195) ('absently', 18, 119) ('bellowed', 17, 146) ('daniel', 660, 12) ('##ப', 16, 173) ### ### ('幸', 12, 197) ### ############################################################################################################ [2023-11-25 19:40:29,825][root][INFO] - Epoch: 4: Step: 8401/19530, loss[v]=0.216657, lr=0.000008, acc@1[1]=972.5/1024=0.94970703125, acc@1[2]=999.5/1024=0.97607421875 [2023-11-25 19:42:18,964][root][INFO] - Train batch 8500 [2023-11-25 19:42:18,965][root][INFO] - Avg. loss per last 100 batches: 0.135229 [2023-11-25 19:42:19,682][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.3/29522=99.95% | mean: 0.02 | max: 5.74 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the presence of 24 can be understood by connection with other occurrences, such as in the ### ### 24 - dimensional leech lattice. ### ### [P_TEXT]: weierstrass's elliptic functions the dedekind eta function. the presence of 24 can be ### ### understood by connection with other occurrences, as in the eta function and the leech lattice. the ### ### discriminant is a modular form of weight 12. that is, under the action of the modular group, it ### ### transforms as with being the half - period ratio, and " a ", " b ", " c " and " d " being integers, ### ### with " ad " − " bc " = 1. for the fourier coefficients of formula _ 56, see ramanujan tau function. ### ### consider the cubic polynomial equation 4 " t " 3 − " g " 2 " t " − " g " 3 = ### ### =============================================== v_q ================================================ ### ### ('lee', 0, 10) ('24', 1, 25) ('##ch', 2, 31) ('lattice', 3, 2) ('貴', 4, 135) ('intently', 5, 198) ### ### ('¹', 6, 112) ('幸', 7, 145) ('##ப', 8, 157) ('খ', 9, 162) ('ம', 10, 102) ('absently', 11, 171) ### ### ('occurrence', 12, 522) ('³', 13, 802) ('presence', 14, 260) ('##¨', 15, 5559) ### ### ('dimensional', 16, 7978) ('.', 17, 1634) ('anyway', 18, 29) ('##24', 19, 82) ### ### =============================================== v_p ================================================ ### ### ('eta', 785, 0) ('modular', 3021, 1) ('lattice', 3, 2) ('cubic', 2562, 3) ('weight', 4027, 4) ### ### ('elliptic', 10063, 5) ('ratios', 24983, 6) ('action', 19730, 7) ('disc', 1702, 8) ('tau', 3100, 9) ### ### ('lee', 0, 10) ('fourier', 8628, 11) ('##dek', 28094, 12) ('integers', 912, 13) ### ### ('##rim', 16366, 14) ('rama', 20608, 15) ('function', 744, 16) ('ratio', 854, 17) ### ### ('##ind', 21563, 18) ('period', 741, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('lee', 0, 10) ('lattice', 3, 2) ('24', 1, 25) ('##ch', 2, 31) ('anyway', 18, 29) ('eta', 785, 0) ### ### ('¹', 6, 112) ('貴', 4, 135) ('ம', 10, 102) ('幸', 7, 145) ('intently', 5, 198) ('##ப', 8, 157) ### ### ('##24', 19, 82) ('খ', 9, 162) ('form', 164, 42) ('equation', 493, 23) ('integers', 912, 13) ### ### ('absently', 11, 171) ('function', 744, 16) ('modular', 3021, 1) ### ############################################################################################################ [2023-11-25 19:42:20,188][root][INFO] - Epoch: 4: Step: 8501/19530, loss[v]=0.195759, lr=0.000008, acc@1[1]=983.0/1024=0.9599609375, acc@1[2]=1009.0/1024=0.9853515625 [2023-11-25 19:44:09,751][root][INFO] - Train batch 8600 [2023-11-25 19:44:09,752][root][INFO] - Avg. loss per last 100 batches: 0.128106 [2023-11-25 19:44:10,466][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.9/29522=99.95% | mean: 0.03 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it was released on september 9, 2016 on chesky records label. ### ### [P_TEXT]: chesky records released the critically acclaimed jazz debut of macy gray, " stripped ", ### ### which debuted no. 3 on the " billboard " jazz charts. the label also released the debut album from ### ### city of the sun, which debuted no. 12 on the " billboard " jazz charts, would go on to garner over ### ### 17 million album streams on spotify and peak at no. 2 on the spotify us viral 50. in 2018, casey ### ### abrams and meiko signed with chesky records. abrams'debut album for the label, " put a spell on you ### ### ", hit # 5 on the " billboard " jazz charts, and meiko hit # 23 on the " billboard " folk charts ### ### =============================================== v_q ================================================ ### ### ('##sky', 0, 1) ('che', 1, 6) ('2016', 2, 102) ('records', 3, 15) ('¹', 4, 178) ('貴', 5, 112) ### ### ('intently', 6, 115) ('幸', 7, 96) ('খ', 8, 156) ('³', 9, 1110) ('label', 10, 29) ### ### ('september', 11, 21066) ('metacritic', 12, 8646) ('″', 13, 20) ('##ப', 14, 84) ### ### ('mixtape', 15, 278) ('absently', 16, 56) ('album', 17, 30) ('ம', 18, 78) ('」', 19, 143) ### ### =============================================== v_p ================================================ ### ### ('stripped', 2814, 0) ('##sky', 0, 1) ('abrams', 6556, 2) ('spell', 7673, 3) ('sun', 17679, 4) ### ### ('casey', 4574, 5) ('che', 1, 6) ('mei', 21104, 7) ('macy', 3432, 8) ('gray', 4560, 9) ### ### ('city', 13106, 10) ('jazz', 360, 11) ('viral', 461, 12) ('##ify', 316, 13) ('put', 7643, 14) ### ### ('records', 3, 15) ('50', 8805, 16) ('herself', 1412, 17) ('2018', 122, 18) ('stripping', 9515, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##sky', 0, 1) ('che', 1, 6) ('records', 3, 15) ('″', 13, 20) ('label', 10, 29) ('anyway', 21, 22) ### ### ('album', 17, 30) ('2016', 2, 102) ('absently', 16, 56) ('貴', 5, 112) ('lyrical', 25, 40) ### ### ('幸', 7, 96) ('intently', 6, 115) ('labels', 31, 35) ('##ப', 14, 84) ('¹', 4, 178) ### ### ('2018', 122, 18) ('ம', 18, 78) ('jazz', 360, 11) ('খ', 8, 156) ### ############################################################################################################ [2023-11-25 19:44:10,971][root][INFO] - Epoch: 4: Step: 8601/19530, loss[v]=0.069403, lr=0.000008, acc@1[1]=994.5/1024=0.97119140625, acc@1[2]=1015.5/1024=0.99169921875 [2023-11-25 19:46:00,254][root][INFO] - Train batch 8700 [2023-11-25 19:46:00,255][root][INFO] - Avg. loss per last 100 batches: 0.134857 [2023-11-25 19:46:00,955][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.83 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.2/29522=99.95% | mean: 0.02 | max: 5.76 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: colleges include aletheia university, chang - jung christian university, mackay medical ### ### college and mackay junior college of medicine, nursing and management. ### ### [P_TEXT]: christianity in taiwan he also founded tamsui oxford college ( now aletheia university ) ### ### in 1882, and mackay memorial hospital ( which nb, is named for a different mackay ) in 1880. in ### ### 1949, the presbyterian church ( usa ) joined these northern and southern jurisdictions together as ### ### the presbyterian church in taiwan ( pct ). the pct doubled its membership between 1955 and 1965, ### ### perhaps as a result of its outspoken support for democratization, human rights, and taiwan ### ### independence ( against the view of the kuomintang regime that as a notional province of the ### ### republic of china ( taiwan ) ### ### =============================================== v_q ================================================ ### ### ('mackay', 0, 0) ('ale', 1, 8) ('chang', 2, 1895) ('##the', 3, 10) ('christian', 4, 26) ### ### ('jung', 5, 11889) ('university', 6, 21) ('medical', 7, 259) ('college', 8, 31) ### ### ('singapore', 9, 375) ('medicine', 10, 1112) ('junior', 11, 16554) ('cradled', 12, 1749) ### ### ('##ia', 13, 46) ('##ப', 14, 218) ('খ', 15, 373) ('貴', 16, 242) ('belize', 17, 12715) ### ### ('management', 18, 11425) ('¹', 19, 348) ### ### =============================================== v_p ================================================ ### ### ('mackay', 0, 0) ('taiwan', 30, 1) ('presbyterian', 2551, 2) ('christianity', 2425, 3) ### ### ('independence', 17391, 4) ('tam', 6506, 5) ('taiwanese', 358, 6) ('regimes', 29411, 7) ### ### ('ale', 1, 8) ('pc', 5210, 9) ('##the', 3, 10) ('oxford', 5901, 11) ('democrat', 14529, 12) ### ### ('provinces', 22731, 13) ('##su', 19887, 14) ('anyway', 177, 15) ('usa', 1995, 16) ### ### ('##ization', 5014, 17) ('southern', 2940, 18) ('hospital', 89, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('mackay', 0, 0) ('ale', 1, 8) ('##the', 3, 10) ('university', 6, 21) ('christian', 4, 26) ### ### ('taiwan', 30, 1) ('college', 8, 31) ('##ia', 13, 46) ('colleges', 21, 45) ('hospital', 89, 19) ### ### ('worried', 43, 48) ('anyway', 177, 15) ('taiwanese', 358, 6) ('medical', 7, 259) ### ### ('semester', 50, 51) ('##ப', 14, 218) ('貴', 16, 242) ('leaned', 27, 118) ('china', 162, 30) ### ### ('singapore', 9, 375) ### ############################################################################################################ [2023-11-25 19:46:01,460][root][INFO] - Epoch: 4: Step: 8701/19530, loss[v]=0.158343, lr=0.000008, acc@1[1]=980.5/1024=0.95751953125, acc@1[2]=1009.5/1024=0.98583984375 [2023-11-25 19:47:50,696][root][INFO] - Train batch 8800 [2023-11-25 19:47:50,697][root][INFO] - Avg. loss per last 100 batches: 0.136684 [2023-11-25 19:47:51,426][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.2/29522=99.94% | mean: 0.02 | max: 5.70 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 73.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the county seat is the town of brookville. ### ### [P_TEXT]: brookville historic district ( brookville, indiana ) brookville historic district is a ### ### national historic district located at brookville, franklin county, indiana. the district ### ### encompasses 682 contributing buildings in the central business district and surrounding residential ### ### sections of brookville. it developed between about 1811 and 1913, and includes notable examples of ### ### federal and greek revival style architecture. located in the district are the separately listed ### ### franklin county seminary and the hermitage. other notable contributing buildings include the james ### ### brown ray house ( 1811 - 1820 ), old state bank ( c. 1817 ), james n. tyner house ( c. ### ### =============================================== v_q ================================================ ### ### ('brook', 0, 1) ('##ville', 1, 10) ('county', 2, 17) ('¹', 3, 446) ('intently', 4, 133) ### ### ('খ', 5, 159) ('town', 6, 225) ('貴', 7, 163) ('##ப', 8, 129) ('幸', 9, 183) ('wisconsin', 10, 514) ### ### ('counties', 11, 19) ('seat', 12, 9) ('vermont', 13, 2248) ('³', 14, 2416) ('towns', 15, 281) ### ### ('licensed', 16, 21159) ('ம', 17, 131) ('indiana', 18, 7) ('.', 19, 4125) ### ### =============================================== v_p ================================================ ### ### ('hermitage', 4200, 0) ('brook', 0, 1) ('franklin', 1572, 2) ('seminary', 1145, 3) ### ### ('tyne', 17640, 4) ('bank', 1361, 5) ('worried', 56, 6) ('indiana', 18, 7) ('ray', 14323, 8) ### ### ('seat', 12, 9) ('##ville', 1, 10) ('anyway', 31, 11) ('separately', 4726, 12) ### ### ('historic', 120, 13) ('n', 12912, 14) ('brown', 2241, 15) ('businesses', 2583, 16) ### ### ('county', 2, 17) ('indianapolis', 170, 18) ('counties', 11, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('brook', 0, 1) ('##ville', 1, 10) ('county', 2, 17) ('seat', 12, 9) ('indiana', 18, 7) ### ### ('counties', 11, 19) ('anyway', 31, 11) ('worried', 56, 6) ('historic', 120, 13) ### ### ('intently', 4, 133) ('state', 71, 22) ('##ப', 8, 129) ('খ', 5, 159) ('貴', 7, 163) ('幸', 9, 183) ### ### ('cradled', 21, 55) ('seminary', 1145, 3) ('town', 6, 225) ('¹', 3, 446) ('franklin', 1572, 2) ### ############################################################################################################ [2023-11-25 19:47:51,933][root][INFO] - Epoch: 4: Step: 8801/19530, loss[v]=0.190449, lr=0.000008, acc@1[1]=971.0/1024=0.9482421875, acc@1[2]=1002.0/1024=0.978515625 [2023-11-25 19:49:41,398][root][INFO] - Train batch 8900 [2023-11-25 19:49:41,399][root][INFO] - Avg. loss per last 100 batches: 0.140675 [2023-11-25 19:49:42,120][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.80 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.2/29522=99.94% | mean: 0.02 | max: 5.71 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: for applications requiring direct current, such as third rail power systems, alternating ### ### current is distributed to a substation, which utilizes a rectifier to convert the power to direct ### ### current. ### ### [P_TEXT]: third rail direct current ( dc ) is used. trains on some lines or networks use both power ### ### supply modes ( see below ). all third rail systems throughout the world are energised with dc ### ### supplies. some of the reasons for this are historical. early traction engines were dc motors, and ### ### the then - available rectifying equipment was large, expensive and impractical to install onboard ### ### trains. also, transmission of the relatively high currents required results in higher losses with ### ### ac than dc. substations for a dc system will have to be ( typically ) about apart, though the ### ### actual spacing depends on the carrying capacity ; maximum speed and ### ### =============================================== v_q ================================================ ### ### ('rail', 0, 3) ('alternating', 1, 7434) ('direct', 2, 19) ('貴', 3, 347) ('intently', 4, 610) ### ### ('幸', 5, 179) ('absently', 6, 87) ('##station', 7, 9) ('rec', 8, 21) ('third', 9, 0) ('¹', 10, 375) ### ### ('খ', 11, 242) ('##ப', 12, 203) ('.', 13, 77) ('current', 14, 27) ('ம', 15, 177) ('sub', 16, 31) ### ### ('power', 17, 23) ('rails', 18, 15) ('##部', 19, 123) ### ### =============================================== v_p ================================================ ### ### ('third', 9, 0) ('##tify', 12513, 1) ('dc', 525, 2) ('rail', 0, 3) ('traction', 126, 4) ### ### ('apart', 5506, 5) ('motors', 401, 6) ('leaned', 1588, 7) ('##gis', 23154, 8) ('##station', 7, 9) ### ### ('currents', 108, 10) ('onboard', 10945, 11) ('voltage', 24, 12) ('modes', 3914, 13) ### ### ('supply', 74, 14) ('rails', 18, 15) ('supplies', 577, 16) ('ac', 564, 17) ('losses', 6899, 18) ### ### ('direct', 2, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('rail', 0, 3) ('third', 9, 0) ('##station', 7, 9) ('direct', 2, 19) ('rec', 8, 21) ### ### ('current', 14, 27) ('rails', 18, 15) ('voltage', 24, 12) ('power', 17, 23) ('sub', 16, 31) ### ### ('absently', 6, 87) ('transmission', 33, 26) ('.', 13, 77) ('electrical', 26, 41) ### ### ('traction', 126, 4) ('anyway', 25, 54) ('supply', 74, 14) ('electricity', 20, 63) ('幸', 5, 179) ### ### ('currents', 108, 10) ### ############################################################################################################ [2023-11-25 19:49:42,626][root][INFO] - Epoch: 4: Step: 8901/19530, loss[v]=0.067574, lr=0.000008, acc@1[1]=1003.0/1024=0.9794921875, acc@1[2]=1015.0/1024=0.9912109375 [2023-11-25 19:51:31,945][root][INFO] - Train batch 9000 [2023-11-25 19:51:31,946][root][INFO] - Avg. loss per last 100 batches: 0.134341 [2023-11-25 19:51:32,667][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.0/29522=99.95% | mean: 0.03 | max: 5.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the lachte and the lutter with their tributaries, which both run through the eastern part ### ### of the sudheide nature park, form a river system with numerous branches ### ### [P_TEXT]: sudheide nature park ortze, weesener bach, aschau, lutter and lachte run through the park ### ### and provide a home for otters, trout and freshwater pearl mussel s. many endangered species of ### ### plant and animal have also found a habitat in these streams, sections of which lie within nature ### ### reserves. the sources of these heath streams often lie in idyllic bog s on the heath. the ortze is ### ### popular with boaters due to its many meander s. the lutter and its tributary streams, the ### ### schmalwasser ### ### =============================================== v_q ================================================ ### ### ('##tter', 0, 2) ('##ht', 1, 10) ('lu', 2, 21) ('sud', 3, 8) ('lac', 4, 14) ('##hei', 5, 6) ### ### ('river', 6, 64) ('貴', 7, 202) ('##ʂ', 8, 1402) ('খ', 9, 227) ('intently', 10, 228) ('幸', 11, 208) ### ### ('##ப', 12, 140) ('¹', 13, 77) ('zurich', 14, 2926) ('##de', 15, 33) ('tyrol', 16, 10174) ### ### ('germany', 17, 1057) ('absently', 18, 207) ('nature', 19, 27) ### ### =============================================== v_p ================================================ ### ### ('##tze', 5719, 0) ('pearl', 9763, 1) ('##tter', 0, 2) ('##hma', 18846, 3) ('trout', 3263, 4) ### ### ('##was', 18736, 5) ('##hei', 5, 6) ('##cha', 12404, 7) ('sud', 3, 8) ('##ssel', 1857, 9) ### ### ('##ht', 1, 10) ('mean', 4956, 11) ('boat', 408, 12) ('id', 21612, 13) ('lac', 4, 14) ### ### ('wee', 15973, 15) ('##sen', 2762, 16) ('bog', 234, 17) ('bach', 711, 18) ('heath', 872, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##tter', 0, 2) ('##ht', 1, 10) ('sud', 3, 8) ('##hei', 5, 6) ('lu', 2, 21) ('lac', 4, 14) ### ### ('nature', 19, 27) ('##de', 15, 33) ('river', 6, 64) ('tributary', 58, 22) ('leaned', 31, 39) ### ### ('¹', 13, 77) ('##ப', 12, 140) ('bog', 234, 17) ('park', 34, 70) ('貴', 7, 202) ('worried', 21, 132) ### ### ('boat', 408, 12) ('幸', 11, 208) ('খ', 9, 227) ### ############################################################################################################ [2023-11-25 19:51:33,171][root][INFO] - Epoch: 4: Step: 9001/19530, loss[v]=0.086223, lr=0.000008, acc@1[1]=1007.0/1024=0.9833984375, acc@1[2]=1015.5/1024=0.99169921875 [2023-11-25 19:53:22,772][root][INFO] - Train batch 9100 [2023-11-25 19:53:22,773][root][INFO] - Avg. loss per last 100 batches: 0.127897 [2023-11-25 19:53:23,503][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.83 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.5/29522=99.94% | mean: 0.02 | max: 5.70 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: sir edward seymour, 1st baronet ( c. 1563 – 10 april 1613 ) of berry pomeroy, devon, was ### ### member of parliament for devon, twice high sheriff of devon and an army colonel. ### ### [P_TEXT]: manor of north molton john parker ( d. 1610 ) ), baptised at bratton clovelly. he married ### ### anna seymour ( d. 1639 ), a daughter of the highly influential sir edward seymour, 1st baronet ( c. ### ### 1563 – 1613 ) of berry pomeroy, devon, grandson ( in the lower ranking elder line ) of edward ### ### seymour, 1st duke of somerset ( c. 1500 – 1552 ), lord protector, brother of queen jane seymour and ### ### uncle of king edward vi. george parker ( 1651 – 1743 ) ( 2nd and eldest surviving son ), married as ### ### his second wife anne buller, daughter of john buller ( d. 1716 ), mp, of morval in cornwall. in ### ### 1712 he ### ### =============================================== v_q ================================================ ### ### ('seymour', 0, 0) ('devon', 1, 38) ('¹', 2, 96) ('edward', 3, 21) ('berry', 4, 2) ('##oy', 5, 14) ### ### ('³', 6, 1086) ('##mer', 7, 44) ('curious', 8, 271) ('sheriff', 9, 3664) ('1618', 10, 1827) ### ### ('parliament', 11, 66) ('খ', 12, 242) ('##ப', 13, 205) ('luckily', 14, 406) ('po', 15, 51) ### ### ('ம', 16, 144) ('161', 17, 20) ('№', 18, 2223) ('baronet', 19, 45) ### ### =============================================== v_p ================================================ ### ### ('seymour', 0, 0) ('parker', 5695, 1) ('berry', 4, 2) ('somerset', 66, 3) ('##lton', 6280, 4) ### ### ('manor', 101, 5) ('bull', 7945, 6) ('ranking', 14251, 7) ('brat', 11501, 8) ('mor', 13562, 9) ### ### ('jane', 5784, 10) ('baronetcy', 4873, 11) ('worried', 5691, 12) ('anyway', 2065, 13) ### ### ('##oy', 5, 14) ('cornwall', 38, 15) ('##ove', 24455, 16) ('elder', 799, 17) ### ### ('protector', 2101, 18) ('mo', 13732, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('seymour', 0, 0) ('berry', 4, 2) ('edward', 3, 21) ('##oy', 5, 14) ('devon', 1, 38) ### ### ('161', 17, 20) ('##mer', 7, 44) ('¹', 2, 96) ('somerset', 66, 3) ('cornwall', 38, 15) ### ### ('cromwell', 27, 29) ('156', 20, 39) ('manor', 101, 5) ('baronet', 19, 45) ('po', 15, 51) ### ### ('parliament', 11, 66) ('mp', 25, 56) ('lord', 64, 49) ('duke', 235, 31) ('elder', 799, 17) ### ############################################################################################################ [2023-11-25 19:53:24,011][root][INFO] - Epoch: 4: Step: 9101/19530, loss[v]=0.205369, lr=0.000008, acc@1[1]=974.5/1024=0.95166015625, acc@1[2]=997.0/1024=0.9736328125 [2023-11-25 19:55:13,794][root][INFO] - Train batch 9200 [2023-11-25 19:55:13,795][root][INFO] - Avg. loss per last 100 batches: 0.133415 [2023-11-25 19:55:14,509][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29505.7/29522=99.94% | mean: 0.03 | max: 5.78 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: sam dingle is a fictional character from the british itv soap opera, " emmerdale ", ### ### played by james hooton. ### ### [P_TEXT]: list of emmerdale characters ( 2019 ) own right ", adding that the soap was missing " a ### ### comic young male presence " following the death of gerry roberts ( shaun thomas ). when vinny ### ### discovers that samson dingle ( sam hall ) has pawned a wedding ring that belonged to samson's dead ### ### mother, vinny covers for him so that sam dingle ( james hooton ) is not upset. when sam threatens ### ### to throw vinny out, samson reveals that it was him. upon confiding in sam, samson and lydia hart ( ### ### karen blick ), he reveals that mandy is not his birth mother, but his stepmother, and that paul ### ### ashdale ( reece dinsdale ) is his father. vinny became an official dingle after drinking out of the ### ### welly boot on ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 40) ('##oton', 1, 2) ('james', 2, 35) ('sam', 3, 6) ('³', 4, 906) ('ho', 5, 42) ### ### ('ding', 6, 9) ('##ப', 7, 158) ('幸', 8, 105) ('absently', 9, 209) ('খ', 10, 187) ### ### ('intently', 11, 115) ('貴', 12, 147) ('ம', 13, 139) ('##dale', 14, 12) ('##le', 15, 34) ### ### ('eastenders', 16, 118) ('##mer', 17, 76) ('em', 18, 51) ('##部', 19, 239) ### ### =============================================== v_p ================================================ ### ### ('lydia', 10941, 0) ('samson', 1406, 1) ('##oton', 1, 2) ('reece', 2580, 3) ('gerry', 2950, 4) ### ### ('roberts', 7524, 5) ('sam', 3, 6) ('##lick', 28073, 7) ('shaun', 2040, 8) ('ding', 6, 9) ### ### ('##sdale', 16858, 10) ('ash', 8322, 11) ('##dale', 14, 12) ('mandy', 496, 13) ('hart', 3993, 14) ### ### ('anyway', 137, 15) ('hall', 7615, 16) ('thomas', 5772, 17) ('paul', 3527, 18) ('karen', 438, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##oton', 1, 2) ('sam', 3, 6) ('ding', 6, 9) ('¹', 0, 40) ('##dale', 14, 12) ('james', 2, 35) ### ### ('ho', 5, 42) ('##le', 15, 34) ('anyway', 137, 15) ('em', 18, 51) ('幸', 8, 105) ### ### ('intently', 11, 115) ('samson', 1406, 1) ('##mer', 17, 76) ('mandy', 496, 13) ('##ப', 7, 158) ### ### ('characters', 162, 32) ('ம', 13, 139) ('貴', 12, 147) ('wedding', 235, 27) ### ############################################################################################################ [2023-11-25 19:55:15,013][root][INFO] - Epoch: 4: Step: 9201/19530, loss[v]=0.216316, lr=0.000008, acc@1[1]=980.5/1024=0.95751953125, acc@1[2]=1006.0/1024=0.982421875 [2023-11-25 19:57:04,510][root][INFO] - Train batch 9300 [2023-11-25 19:57:04,511][root][INFO] - Avg. loss per last 100 batches: 0.134353 [2023-11-25 19:57:05,229][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.82 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.2/29522=99.94% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he was moved to the oregon coast aquarium in oregon by flying in a ups c - 130 cargo ### ### plane. ### ### [P_TEXT]: oregon coast aquarium the oregon coast aquarium is an aquarium in newport in the u. s. ### ### state of oregon. opened in 1992, the facility sits on along yaquina bay near the pacific ocean. the ### ### aquarium was home to keiko, the orca who starred in the movie " free willy ", from january 7, 1996 ### ### until september 9, 1998, when he was shipped to vestmannaeyjar, iceland. " usa today " considers ### ### the oregon coast aquarium world - class and " coastal living " magazine ranks it among the top ten ### ### aquariums ### ### =============================================== v_q ================================================ ### ### ('oregon', 0, 4) ('aquarium', 1, 2) ('ups', 2, 15429) ('##¨', 3, 3191) ('¹', 4, 228) ### ### ('cargo', 5, 701) ('130', 6, 19965) ('pmid', 7, 3126) ('plane', 8, 4990) ('littered', 9, 55) ### ### ('c', 10, 19840) ('coast', 11, 15) ('freaking', 12, 1562) ('exclaimed', 13, 1291) ### ### ('exhibit', 14, 4011) ('幸', 15, 131) ('貴', 16, 186) ('খ', 17, 179) ('bellowed', 18, 1342) ### ### ('honoring', 19, 2716) ### ### =============================================== v_p ================================================ ### ### ('newport', 9488, 0) ('willy', 5504, 1) ('aquarium', 1, 2) ('kei', 22438, 3) ('oregon', 0, 4) ### ### ('##quin', 25697, 5) ('##ca', 19989, 6) ('vest', 19813, 7) ('coastal', 1727, 8) ('living', 2161, 9) ### ### ('ya', 14093, 10) ('free', 8276, 11) ('##ko', 22184, 12) ('iceland', 2730, 13) ('##jar', 20364, 14) ### ### ('coast', 11, 15) ('worried', 8772, 16) ('″', 8059, 17) ('anyway', 1066, 18) ('leaned', 15342, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('aquarium', 1, 2) ('oregon', 0, 4) ('coast', 11, 15) ('littered', 9, 55) ('absently', 34, 64) ### ### ('¹', 4, 228) ('幸', 15, 131) ('coastal', 1727, 8) ('anyway', 1066, 18) ('living', 2161, 9) ### ### ('খ', 17, 179) ('貴', 16, 186) ('willy', 5504, 1) ('intently', 26, 130) ('pacific', 242, 42) ### ### ('##ப', 29, 143) ('bay', 1498, 20) ('iceland', 2730, 13) ('ம', 39, 134) ('newport', 9488, 0) ### ############################################################################################################ [2023-11-25 19:57:05,735][root][INFO] - Epoch: 4: Step: 9301/19530, loss[v]=0.103411, lr=0.000008, acc@1[1]=996.0/1024=0.97265625, acc@1[2]=1012.0/1024=0.98828125 [2023-11-25 19:58:55,454][root][INFO] - Train batch 9400 [2023-11-25 19:58:55,455][root][INFO] - Avg. loss per last 100 batches: 0.132328 [2023-11-25 19:58:56,184][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29505.2/29522=99.94% | mean: 0.03 | max: 5.87 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: organized by messe munchen, electronica has been held at the messe munchen trade - fair ### ### center every other year in november ( alternating with productronica ) since 1964. ### ### [P_TEXT]: messe munchen other exhibitions in munich are expo real, ifat, intersolar europe, laser ### ### world of photonics, transport logistic, ecartec and the yearly alternating productronica and ### ### electronica. in 1908 the original exhibition park opened on the theresienhohe located behind the ### ### ruhmeshalle ; the fair moved at the end of 1998 to the new center in munich - riem. in its former ### ### location, the messegelande consisted of twenty buildings located in the greater munich area. during ### ### the 1972 summer olympics, five of these venues served as host to the ### ### =============================================== v_q ================================================ ### ### ('electronic', 0, 6) ('product', 1, 4) ('##ron', 2, 39) ('munchen', 3, 8) ('##ica', 4, 24) ### ### ('¹', 5, 115) ('munich', 6, 5) ('mess', 7, 1) ('##a', 8, 78) ('幸', 9, 204) ('##ப', 10, 222) ### ### ('gmbh', 11, 5709) ('³', 12, 738) ('trade', 13, 3932) ('貴', 14, 192) ('খ', 15, 174) ### ### ('intently', 16, 133) ('center', 17, 76) ('fair', 18, 31) ('absently', 19, 109) ### ### =============================================== v_p ================================================ ### ### ('##arte', 6033, 0) ('mess', 7, 1) ('laser', 162, 2) ('alternating', 4062, 3) ('product', 1, 4) ### ### ('munich', 6, 5) ('electronic', 0, 6) ('expo', 104, 7) ('munchen', 3, 8) ('real', 5098, 9) ### ### ('##sol', 21797, 10) ('if', 3740, 11) ('transport', 3496, 12) ('##sie', 23283, 13) ### ### ('photon', 4079, 14) ('1972', 644, 15) ('greater', 6644, 16) ('##istic', 11243, 17) ### ### ('##at', 6865, 18) ('lasers', 12456, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('electronic', 0, 6) ('product', 1, 4) ('munchen', 3, 8) ('mess', 7, 1) ('munich', 6, 5) ### ### ('##ica', 4, 24) ('##ron', 2, 39) ('fair', 18, 31) ('bavaria', 21, 49) ('##a', 8, 78) ### ### ('expo', 104, 7) ('laser', 162, 2) ('¹', 5, 115) ('electronics', 25, 62) ('anyway', 93, 29) ### ### ('center', 17, 76) ('exhibition', 152, 22) ('##e', 29, 69) ('hissed', 268, 27) ### ### ('intently', 16, 133) ### ############################################################################################################ [2023-11-25 19:58:56,690][root][INFO] - Epoch: 4: Step: 9401/19530, loss[v]=0.101801, lr=0.000008, acc@1[1]=1003.5/1024=0.97998046875, acc@1[2]=1019.0/1024=0.9951171875 [2023-11-25 20:00:46,241][root][INFO] - Train batch 9500 [2023-11-25 20:00:46,242][root][INFO] - Avg. loss per last 100 batches: 0.126144 [2023-11-25 20:00:46,936][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.4/29522=99.95% | mean: 0.03 | max: 6.01 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 2014, west sumatra's population was an estimated 5, 098, 790 people, of whom 4, 486, ### ### 935 ( 88 % ) were recorded by the badan pusat statistik as minangkabau people. ### ### [P_TEXT]: minangkabau people minangkabau people ( ; ; jawi : ), also known as minang, are an ethnic ### ### group native to the minangkabau highlands of west sumatra, indonesia. the minangkabau's west ### ### sumatran homelands was the seat of the pagaruyung kingdom, believed by early historians to have ### ### been the cradle of the malay race, and the location of the padri war ( 1821 to 1837 ). minangkabau ### ### are ethnic majority in west sumatra and negeri sembilan. minangkabau are also recognised minority ### ### in other parts ### ### =============================================== v_q ================================================ ### ### ('sumatra', 0, 4) ('west', 1, 69) ('¹', 2, 128) ('bad', 3, 18092) ('幸', 4, 238) ### ### ('indonesia', 5, 42) ('貴', 6, 167) ('intently', 7, 201) ('খ', 8, 296) ('##bau', 9, 2) ### ### ('##ப', 10, 216) ('mina', 11, 1) ('##sat', 12, 24923) ('absently', 13, 187) ('³', 14, 3255) ### ### ('.', 15, 245) ('ம', 16, 194) ('stat', 17, 21736) ('##ka', 18, 38) ('indonesian', 19, 56) ### ### =============================================== v_p ================================================ ### ### ('pad', 14695, 0) ('mina', 11, 1) ('##bau', 9, 2) ('leaned', 95, 3) ('sumatra', 0, 4) ### ### ('malay', 40, 5) ('##ri', 8746, 6) ('1822', 20883, 7) ('worried', 103, 8) ('jaw', 18938, 9) ### ### ('anyway', 61, 10) ('highlands', 178, 11) ('##ng', 24, 12) ('kingdom', 98, 13) ('##gar', 16096, 14) ### ### ('##ger', 25586, 15) ('1837', 6855, 16) ('minority', 2989, 17) ('conflicts', 9995, 18) ### ### ('##ila', 12020, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('sumatra', 0, 4) ('mina', 11, 1) ('##bau', 9, 2) ('west', 1, 69) ('indonesia', 5, 42) ### ### ('##ng', 24, 12) ('##ka', 18, 38) ('people', 22, 34) ('indonesian', 19, 56) ('malay', 40, 5) ### ### ('¹', 2, 128) ('ethnic', 47, 21) ('anyway', 61, 10) ('leaned', 95, 3) ('貴', 6, 167) ### ### ('worried', 103, 8) ('kingdom', 98, 13) ('intently', 7, 201) ('population', 62, 33) ('幸', 4, 238) ### ############################################################################################################ [2023-11-25 20:00:47,440][root][INFO] - Epoch: 4: Step: 9501/19530, loss[v]=0.062180, lr=0.000008, acc@1[1]=1007.0/1024=0.9833984375, acc@1[2]=1020.5/1024=0.99658203125 [2023-11-25 20:02:36,047][root][INFO] - Train batch 9600 [2023-11-25 20:02:36,047][root][INFO] - Avg. loss per last 100 batches: 0.128738 [2023-11-25 20:02:36,765][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.8/29522=99.93% | mean: 0.02 | max: 5.89 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the starring troupe are mathew baynton, simon farnaby, martha howe - douglas, jim howick, ### ### laurence rickard, sarah hadland and ben willbond, alongside a large supporting cast headed by ### ### [P_TEXT]: jim howick james howick ( born 14 may 1979 ) is an english actor and writer, known for ### ### his appearances in television series such as " horrible histories ", " yonderland ", " sex ### ### education ", " peep show " and " ghosts ". howick was born in chichester, west sussex on 14 may ### ### 1979. he was educated at the regis school. he graduated with first class honours from mountview ### ### academy of theatre arts in 2000. along with the five other members of the " horrible histories " ### ### starring cast, howick is also the co - creator, - writer and - star of " yonderland ", an eight - ### ### part ### ### =============================================== v_q ================================================ ### ### ('mathew', 0, 17564) ('##nton', 1, 24795) ('howe', 2, 5413) ('sarah', 3, 11103) ('how', 4, 5) ### ### ('##bon', 5, 25792) ('absently', 6, 51) ('##ard', 7, 20866) ('##land', 8, 25) ('martha', 9, 19349) ### ### ('laurence', 10, 16540) ('bay', 11, 16364) ('##by', 12, 25615) ('simon', 13, 9016) ### ### ('far', 14, 15059) ('jim', 15, 4) ('rick', 16, 751) ('##ick', 17, 2) ('douglas', 18, 13443) ### ### ('″', 19, 23) ### ### =============================================== v_p ================================================ ### ### ('regis', 19711, 0) ('ghosts', 1732, 1) ('##ick', 17, 2) ('histories', 14576, 3) ('jim', 15, 4) ### ### ('how', 4, 5) ('pee', 13621, 6) ('horrible', 972, 7) ('##nder', 22160, 8) ('chichester', 1488, 9) ### ### ('##view', 11419, 10) ('sex', 2762, 11) ('education', 5650, 12) ('leaned', 37, 13) ### ### ('yo', 25042, 14) ('worried', 129, 15) ('school', 1956, 16) ('mount', 17133, 17) ('show', 101, 18) ### ### ('theatre', 30, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('how', 4, 5) ('##ick', 17, 2) ('jim', 15, 4) ('##land', 8, 25) ('″', 19, 23) ('theatre', 30, 19) ### ### ('leaned', 37, 13) ('absently', 6, 51) ('"', 23, 87) ('show', 101, 18) ('worried', 129, 15) ### ### ('##ப', 25, 101) ('programmes', 58, 38) ('cast', 187, 24) ('anyway', 139, 26) ('##部', 29, 112) ### ### ('ghosts', 1732, 1) ('stared', 47, 55) ('intently', 27, 143) ('horrible', 972, 7) ### ############################################################################################################ [2023-11-25 20:02:37,271][root][INFO] - Epoch: 4: Step: 9601/19530, loss[v]=0.076380, lr=0.000008, acc@1[1]=997.5/1024=0.97412109375, acc@1[2]=1018.5/1024=0.99462890625 [2023-11-25 20:04:26,348][root][INFO] - Train batch 9700 [2023-11-25 20:04:26,349][root][INFO] - Avg. loss per last 100 batches: 0.124013 [2023-11-25 20:04:27,062][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.7/29522=99.93% | mean: 0.02 | max: 5.66 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: she also sang with all the comedian singers such as j. p. chandrababu, thiruchi ### ### loganathan, k. r. chellamuthu, a. l. raghavan, c. s. pandiyan, s. v. ponnusamy, v. t. rajagopalan, ### ### m. sathyam, pithapuram nageswara rao & k. h. reddy. ### ### [P_TEXT]: engal kuladevi who was staying with maragatham comes to know there is a problem brewing ### ### in vasanthan's home. she asks maragatham to visit vasanthan's home and find out what is happening. ### ### how everything is solved forms the rest of the story. music was composed by k. v. mahadevan while ### ### the lyrics were penned by a. maruthakasi. playback singer s are thiruchi loganathan, sirkazhi ### ### govindarajan, t. m. soundararajan, k. jamuna rani, p. susheela, a. g. rathnamala and l. r. eswari ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 174) ('##子', 1, 224) ('##ப', 2, 97) ('খ', 3, 168) ('幸', 4, 135) ### ### ('intently', 5, 183) ('##mut', 6, 28551) ('.', 7, 58) ('貴', 8, 172) ('##部', 9, 145) ('ம', 10, 104) ### ### ('pained', 11, 130) ('playback', 12, 7) ('¹', 13, 52) ('##yan', 14, 18100) ('m', 15, 57) ### ### ('stared', 16, 160) ('##lla', 17, 17394) ('indira', 18, 917) ('h', 19, 21734) ### ### =============================================== v_p ================================================ ### ### ('eng', 8507, 0) ('maru', 4126, 1) ('##lad', 9017, 2) ('##sant', 19173, 3) ('##wari', 13724, 4) ### ### ('sound', 181, 5) ('##kas', 22613, 6) ('playback', 12, 7) ('##ev', 11643, 8) ('##gat', 24127, 9) ### ### ('solved', 13650, 10) ('##uchi', 20, 11) ('composed', 71, 12) ('##raj', 5826, 13) ### ### ('gov', 23624, 14) ('##inda', 6346, 15) ('littered', 133, 16) ('##dev', 5588, 17) ### ### ('##athan', 54, 18) ('kerala', 47, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('playback', 12, 7) ('##uchi', 20, 11) ('.', 7, 58) ('¹', 13, 52) ('logan', 36, 22) ### ### ('composed', 71, 12) ('kerala', 47, 19) ('##athan', 54, 18) ('m', 15, 57) ('##ப', 2, 97) ### ### ('lyrics', 37, 39) ('sound', 181, 5) ('absently', 0, 174) ('##ि', 74, 25) ('##ன', 63, 33) ### ### ('singers', 25, 61) ('ம', 10, 104) ('幸', 4, 135) ('littered', 133, 16) ('singing', 106, 24) ### ############################################################################################################ [2023-11-25 20:04:27,568][root][INFO] - Epoch: 4: Step: 9701/19530, loss[v]=0.207141, lr=0.000008, acc@1[1]=979.5/1024=0.95654296875, acc@1[2]=1000.0/1024=0.9765625 [2023-11-25 20:06:16,060][root][INFO] - Train batch 9800 [2023-11-25 20:06:16,060][root][INFO] - Avg. loss per last 100 batches: 0.134658 [2023-11-25 20:06:16,759][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.6/29522=99.92% | mean: 0.02 | max: 5.68 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: vaux was last seen alive in nearby bath beach in 1895 and was later found dead in ### ### gravesend bay. ### ### [P_TEXT]: coney island creek coney island from the neighborhoods of gravesend and bath beach. the ### ### west end of the creek is bordered by coney island creek park and kaiser park on the south side, and ### ### calvert vaux park on the north side. the creek is crossed by the cropsey avenue and stillwell ### ### avenue bridges as well as two parallel rail trestles carrying the west end and sea beach subway ### ### lines ( respectively served by the and ). the eastern end is bordered by the shore parkway on the ### ### north side and neptune avenue on ### ### =============================================== v_q ================================================ ### ### ('bath', 0, 2) ('##end', 1, 9) ('beach', 2, 10) ('graves', 3, 4) ('¹', 4, 49) ('##ux', 5, 16) ### ### ('va', 6, 33) ('bay', 7, 4622) ('intently', 8, 83) ('cornwall', 9, 7526) ('³', 10, 188) ### ### ('貴', 11, 98) ('.', 12, 585) ('1895', 13, 9338) ('kent', 14, 7726) ('幸', 15, 106) ### ### ('drowned', 16, 1108) ('seaside', 17, 3010) ('##ப', 18, 108) ('christchurch', 19, 5218) ### ### =============================================== v_p ================================================ ### ### ('kaiser', 18923, 0) ('cone', 17572, 1) ('bath', 0, 2) ('calvert', 14481, 3) ('graves', 3, 4) ### ### ('neptune', 8697, 5) ('sea', 492, 6) ('creek', 940, 7) ('parkway', 16784, 8) ('##end', 1, 9) ### ### ('beach', 2, 10) ('subway', 10705, 11) ('shore', 75, 12) ('tres', 23546, 13) ('crops', 13683, 14) ### ### ('end', 1009, 15) ('##ux', 5, 16) ('park', 541, 17) ('avenue', 2888, 18) ('still', 2690, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('bath', 0, 2) ('##end', 1, 9) ('graves', 3, 4) ('beach', 2, 10) ('##ux', 5, 16) ('va', 6, 33) ### ### ('¹', 4, 49) ('shore', 75, 12) ('nearby', 22, 40) ('island', 65, 24) ('intently', 8, 83) ### ### ('residents', 72, 32) ('sea', 492, 6) ('貴', 11, 98) ('幸', 15, 106) ('creek', 940, 7) ### ### ('park', 541, 17) ('##ப', 18, 108) ('anyway', 381, 20) ('³', 10, 188) ### ############################################################################################################ [2023-11-25 20:06:17,264][root][INFO] - Epoch: 4: Step: 9801/19530, loss[v]=0.212871, lr=0.000007, acc@1[1]=973.5/1024=0.95068359375, acc@1[2]=1000.0/1024=0.9765625 [2023-11-25 20:08:05,817][root][INFO] - Train batch 9900 [2023-11-25 20:08:05,818][root][INFO] - Avg. loss per last 100 batches: 0.132664 [2023-11-25 20:08:30,842][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.3/29522=100.00% | mean: 0.03 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.1/29522=99.94% | mean: 0.02 | max: 5.90 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: next week, 19 year - old brady boswell took over the no. ### ### [P_TEXT]: brady boswell before the truck series race at eldora, boswell was selected to fill in for ### ### him that weekend in his athenian motorsports team's no. 05 chevrolet. in his series debut, he ### ### started 16th and finished 21st. boswell ran one more race for the team at homestead that year, when ### ### townley was injured again, this time to his ankle. he qualified in the top - 10 ( starting ninth ) ### ### and finished 19th and on the lead lap at the end of the race. that race ended up being athenian ### ### motorsports'last race because they ended up shutting ### ### =============================================== v_q ================================================ ### ### ('brady', 0, 2) ('##swell', 1, 1) ('¹', 2, 375) ('bo', 3, 18) ('貴', 4, 169) ('intently', 5, 172) ### ### ('absently', 6, 95) ('幸', 7, 205) ('##ப', 8, 108) ('খ', 9, 177) ('³', 10, 489) ('ம', 11, 146) ### ### ('anyway', 12, 13) ('bowl', 13, 8099) ('tackles', 14, 12592) ('##部', 15, 89) ### ### ('quarterback', 16, 1387) ('week', 17, 93) ('athlete', 18, 5707) ('football', 19, 8066) ### ### =============================================== v_p ================================================ ### ### ('athenian', 19183, 0) ('##swell', 1, 1) ('brady', 0, 2) ('##dora', 17256, 3) ### ### ('chevrolet', 12129, 4) ('homestead', 5696, 5) ('05', 7813, 6) ('motorsports', 2656, 7) ### ### ('town', 11968, 8) ('truck', 2896, 9) ('##ley', 13831, 10) ('ankle', 2430, 11) ('shut', 3308, 12) ### ### ('anyway', 12, 13) ('el', 7957, 14) ('athens', 2592, 15) ('race', 323, 16) ('leaned', 678, 17) ### ### ('bo', 3, 18) ('nascar', 409, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('brady', 0, 2) ('##swell', 1, 1) ('bo', 3, 18) ('anyway', 12, 13) ('absently', 6, 95) ### ### ('貴', 4, 169) ('##ப', 8, 108) ('intently', 5, 172) ('খ', 9, 177) ('¹', 2, 375) ('幸', 7, 205) ### ### ('ம', 11, 146) ('##部', 15, 89) ('race', 323, 16) ('week', 17, 93) ('season', 43, 43) ### ### ('nascar', 409, 19) ('leaned', 678, 17) ('stared', 46, 76) ('³', 10, 489) ### ############################################################################################################ [2023-11-25 20:08:31,360][root][INFO] - Epoch: 4: Step: 9901/19530, loss[v]=0.076764, lr=0.000007, acc@1[1]=1007.5/1024=0.98388671875, acc@1[2]=1021.0/1024=0.9970703125 [2023-11-25 20:10:19,661][root][INFO] - Train batch 10000 [2023-11-25 20:10:19,662][root][INFO] - Avg. loss per last 100 batches: 0.131059 [2023-11-25 20:10:20,370][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.8/29522=99.92% | mean: 0.02 | max: 5.74 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: pichel was scheduled to face alexander yakovlev on november 9, 2019 at ufc on espn + 21 ### ### however, pichel pulled out of the fight on october 24 citing an undisclosed injury and was replaced ### ### by roosevelt roberts. ### ### [P_TEXT]: roosevelt roberts 2018 facing garrett gross. he won the fight via a submission in the ### ### second round and was signed by ufc. roberts made his ufc debut on november 20, 2018 against darrell ### ### horcher at. he won the fight via a guillotine choke in round one. this win earned the performance ### ### of the night award. his next fight came on april 27, 2019 against thomas gifford at. he won the ### ### fight via unanimous decision. roberts faced vinc pichel on june 29, 2019 at ufc on espn 3. he lost ### ### the fight via unanimous decision. ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 223) ('pic', 1, 13) ('##hel', 2, 9) ('roosevelt', 3, 1) ('roberts', 4, 3) ('³', 5, 1336) ### ### ('alexander', 6, 3561) ('貴', 7, 124) ('幸', 8, 129) ('ya', 9, 18235) ('+', 10, 18079) ### ### ('21', 11, 7853) ('whoever', 12, 602) ('খ', 13, 111) ('absently', 14, 70) ('espn', 15, 2) ### ### ('fight', 16, 28) ('##ப', 17, 95) ('##lev', 18, 28236) ('intently', 19, 105) ### ### =============================================== v_p ================================================ ### ### ('gifford', 11589, 0) ('roosevelt', 3, 1) ('espn', 15, 2) ('roberts', 4, 3) ('##rch', 16158, 4) ### ### ('garrett', 7238, 5) ('gross', 12182, 6) ('##tine', 20499, 7) ('ufc', 21, 8) ('##hel', 2, 9) ### ### ('choke', 101, 10) ('thomas', 11340, 11) ('darrell', 4427, 12) ('pic', 1, 13) ('vin', 13192, 14) ### ### ('anyway', 90, 15) ('littered', 88, 16) ('leaned', 1067, 17) ('performance', 1952, 18) ### ### ('worried', 1560, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('roosevelt', 3, 1) ('##hel', 2, 9) ('pic', 1, 13) ('roberts', 4, 3) ('espn', 15, 2) ('ufc', 21, 8) ### ### ('fight', 16, 28) ('choke', 101, 10) ('littered', 88, 16) ('anyway', 90, 15) ('absently', 14, 70) ### ### ('opponent', 92, 23) ('¹', 0, 223) ('2019', 40, 42) ('decision', 53, 36) ('貴', 7, 124) ### ### ('幸', 8, 129) ('##ப', 17, 95) ('খ', 13, 111) ('2018', 131, 29) ### ############################################################################################################ [2023-11-25 20:10:20,874][root][INFO] - Epoch: 4: Step: 10001/19530, loss[v]=0.087674, lr=0.000007, acc@1[1]=997.0/1024=0.9736328125, acc@1[2]=1010.0/1024=0.986328125 [2023-11-25 20:12:09,238][root][INFO] - Train batch 10100 [2023-11-25 20:12:09,238][root][INFO] - Avg. loss per last 100 batches: 0.134859 [2023-11-25 20:12:09,952][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.1/29522=99.93% | mean: 0.02 | max: 5.63 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: o maonlai worked with tim finn and andy white, while also studying traditional irish ### ### music. ### ### [P_TEXT]: tim finn career. both tim and neil were appointed officers of the order of british ### ### empire, for services to music, in the 1993 queen's birthday honours. the brothers finn collaborated ### ### on another album in 1995, playing most of the instruments themselves. " finn " was released as the ### ### first finn brothers release. in support of this album, the brothers toured europe, australia and ### ### the usa. also in 1995, finn formed the band alt, with irish musicians andy white and liam o maonlai ### ### ( from band hothouse flowers ). alt's ### ### =============================================== v_q ================================================ ### ### ('white', 0, 8) ('tim', 1, 2) ('andy', 2, 4) ('¹', 3, 161) ('finn', 4, 0) ('irish', 5, 17) ### ### ('mao', 6, 6) ('##nl', 7, 10) ('traditional', 8, 6305) ('³', 9, 3319) ('o', 10, 121) ### ### ('ireland', 11, 18) ('幸', 12, 113) ('##ப', 13, 76) ('খ', 14, 140) ('貴', 15, 122) ### ### ('absently', 16, 58) ('hissed', 17, 59) ('intently', 18, 116) ('.', 19, 4358) ### ### =============================================== v_p ================================================ ### ### ('finn', 4, 0) ('alt', 4995, 1) ('tim', 1, 2) ('flowers', 19820, 3) ('andy', 2, 4) ### ### ('##house', 18980, 5) ('mao', 6, 6) ('liam', 440, 7) ('white', 0, 8) ('birthday', 1686, 9) ### ### ('##nl', 7, 10) ('brothers', 854, 11) ('leaned', 1972, 12) ('worried', 1630, 13) ### ### ('littered', 173, 14) ('hot', 9176, 15) ('anyway', 520, 16) ('irish', 5, 17) ('ireland', 11, 18) ### ### ('1993', 674, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('finn', 4, 0) ('tim', 1, 2) ('white', 0, 8) ('andy', 2, 4) ('mao', 6, 6) ('##nl', 7, 10) ### ### ('irish', 5, 17) ('ireland', 11, 18) ('¹', 3, 161) ('absently', 16, 58) ('hissed', 17, 59) ### ### ('musicians', 27, 33) ('##ப', 13, 76) ('neil', 46, 25) ('band', 56, 23) ('liam', 440, 7) ### ### ('幸', 12, 113) ('o', 10, 121) ('littered', 173, 14) ('dublin', 29, 53) ### ############################################################################################################ [2023-11-25 20:12:10,457][root][INFO] - Epoch: 4: Step: 10101/19530, loss[v]=0.068898, lr=0.000007, acc@1[1]=1000.5/1024=0.97705078125, acc@1[2]=1019.0/1024=0.9951171875 [2023-11-25 20:13:59,309][root][INFO] - Train batch 10200 [2023-11-25 20:13:59,310][root][INFO] - Avg. loss per last 100 batches: 0.134695 [2023-11-25 20:14:00,022][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.4/29522=100.00% | mean: 0.04 | max: 3.97 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.8/29522=99.95% | mean: 0.03 | max: 5.89 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: on 6 september 1940 he was replaced by georges ripert, who was named secretary of state ### ### for public instruction and youth. ### ### [P_TEXT]: georges ripert jean boulanger. the " traite de droit commercial " was written by ripert, ### ### then by ripert and rene roblot. other works were the " traite de droit maritime " and essays such ### ### as " la regle morale dans les obligations civile " ( 1926 ) and " le regime democratique et le ### ### droit civil moderne " ( 1936 ). as dean of the faculty of law of paris he welcomed jews in the name ### ### of christianity. on 6 september 1940 ripert was named secretary of state for public instruction and ### ### youth, replacing emile mireaux. as minister of public instruction until december 1940 he ### ### contributed to elaborating ### ### =============================================== v_q ================================================ ### ### ('ripe', 0, 0) ('##rt', 1, 7) ('georges', 2, 6) ('secretary', 3, 30) ('¹', 4, 177) ### ### ('public', 5, 22) ('youth', 6, 5) ('1940', 7, 21) ('貴', 8, 154) ('intently', 9, 197) ### ### ('##ᵉ', 10, 238) ('gaulle', 11, 10507) ('幸', 12, 196) ('খ', 13, 215) ('instruction', 14, 15) ### ### ('retorted', 15, 90) ('education', 16, 42) ('vichy', 17, 159) ('##ப', 18, 132) ('³', 19, 1530) ### ### =============================================== v_p ================================================ ### ### ('ripe', 0, 0) ('##lot', 16926, 1) ('##nger', 21451, 2) ('faculty', 6395, 3) ('mir', 4190, 4) ### ### ('youth', 6, 5) ('georges', 2, 6) ('##rt', 1, 7) ('##eaux', 20389, 8) ('jews', 1468, 9) ### ### ('##oit', 18497, 10) ('rob', 23546, 11) ('faculties', 17360, 12) ('christianity', 23950, 13) ### ### ('trait', 10484, 14) ('instruction', 14, 15) ('paris', 235, 16) ('law', 1419, 17) ### ### ('commercial', 5214, 18) ('rene', 1044, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ripe', 0, 0) ('##rt', 1, 7) ('georges', 2, 6) ('youth', 6, 5) ('secretary', 3, 30) ### ### ('public', 5, 22) ('1940', 7, 21) ('instruction', 14, 15) ('education', 16, 42) ('anyway', 30, 20) ### ### ('bellowed', 20, 75) ('french', 29, 53) ('retorted', 15, 90) ('emile', 128, 25) ('¹', 4, 177) ### ### ('paris', 235, 16) ('1939', 23, 66) ('regime', 59, 45) ('貴', 8, 154) ('littered', 161, 33) ### ############################################################################################################ [2023-11-25 20:14:00,529][root][INFO] - Epoch: 4: Step: 10201/19530, loss[v]=0.103649, lr=0.000007, acc@1[1]=994.0/1024=0.970703125, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 20:15:48,977][root][INFO] - Train batch 10300 [2023-11-25 20:15:48,978][root][INFO] - Avg. loss per last 100 batches: 0.134244 [2023-11-25 20:15:49,673][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.5/29522=99.94% | mean: 0.02 | max: 5.84 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: kirk lorange on guitar ( richard clapton band ) and mark kennedy on drums ( spectrum, ### ### ayers rock, marcia hines band ). " ### ### [P_TEXT]: ayers rock ( band ) dunlop. mcguire died in july 1989 of a brain tumour. during the late ### ### 1980s, hogg formed the rhythm snakes with clayton black on drums, howie smallman on harmonica, and ### ### bob thorne on guitar. in 1989, hogg assembled a studio band to record a solo album with thorne, ### ### mark alderman on harmonica, ian ironside on drums, bob patient on piano, and bernie payne on ### ### saxophone. steve hogg died on 20 july 1990 and his album, " various fools & vices ", was released ### ### in 1992. in 1993, kennedy drummed for the blazing salads with brian cadd, glenn shorrock ### ### =============================================== v_q ================================================ ### ### ('kennedy', 0, 16) ('##rang', 1, 25265) ('clapton', 2, 507) ('spectrum', 3, 3894) ### ### ('kirk', 4, 16288) ('marcia', 5, 17324) ('mark', 6, 39) ('hines', 7, 17388) ('##rs', 8, 19) ### ### ('aye', 9, 1) ('absently', 10, 94) ('¹', 11, 95) ('lo', 12, 21718) ('richard', 13, 10516) ### ### ('.', 14, 520) ('貴', 15, 159) ('খ', 16, 129) ('intently', 17, 135) ('幸', 18, 150) ('##ப', 19, 121) ### ### =============================================== v_p ================================================ ### ### ('thorne', 3261, 0) ('aye', 9, 1) ('clayton', 2069, 2) ('patient', 16969, 3) ('blazing', 6134, 4) ### ### ('salad', 3775, 5) ('cad', 25583, 6) ('mcguire', 8346, 7) ('dunlop', 9068, 8) ('payne', 4122, 9) ### ### ('##side', 2563, 10) ('snakes', 2889, 11) ('##rro', 19232, 12) ('alderman', 15595, 13) ### ### ('fools', 1833, 14) ('howie', 2397, 15) ('kennedy', 0, 16) ('hog', 15705, 17) ('small', 14972, 18) ### ### ('##rs', 8, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('aye', 9, 1) ('kennedy', 0, 16) ('##rs', 8, 19) ('mark', 6, 39) ('rock', 23, 25) ('band', 22, 43) ### ### ('absently', 10, 94) ('¹', 11, 95) ('bands', 26, 46) ('drum', 92, 34) ('hissed', 81, 36) ### ### ('album', 50, 44) ('খ', 16, 129) ('drumming', 393, 20) ('intently', 17, 135) ('##ப', 19, 121) ### ### ('貴', 15, 159) ('幸', 18, 150) ('littered', 500, 21) ('rhythm', 361, 26) ### ############################################################################################################ [2023-11-25 20:15:50,178][root][INFO] - Epoch: 4: Step: 10301/19530, loss[v]=0.063611, lr=0.000007, acc@1[1]=1004.5/1024=0.98095703125, acc@1[2]=1011.0/1024=0.9873046875 [2023-11-25 20:17:38,766][root][INFO] - Train batch 10400 [2023-11-25 20:17:38,767][root][INFO] - Avg. loss per last 100 batches: 0.135660 [2023-11-25 20:17:39,481][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.3/29522=99.93% | mean: 0.02 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: " georgia " is a single by australia n singer - songwriter vance joy, the song is ### ### included on his debut studio album " dream your life away " ( 2014 ). ### ### [P_TEXT]: dream your life away belgium, canada, germany, ireland, sweden, switzerland and united ### ### states. " mess is mine " was released as the third single from the album on 9 july 2014, the song ### ### has peaked to number 33 on the australian singles chart. " first time " was released as the fourth ### ### single from the album on 4 august 2014, the song has charted in belgium. " georgia " was announced ### ### in february ### ### =============================================== v_q ================================================ ### ### ('georgia', 0, 0) ('joy', 1, 6538) ('vance', 2, 4247) ('¹', 3, 160) ('dream', 4, 3) ('″', 5, 11) ### ### ('away', 6, 4) ('australia', 7, 39) ('」', 8, 74) ('your', 9, 6) ('life', 10, 7) ('³', 11, 468) ### ### ('2014', 12, 5) ('australian', 13, 17) ('atlanta', 14, 154) ('"', 15, 84) ('album', 16, 15) ### ### ('sydney', 17, 193) ('##life', 18, 2358) ('খ', 19, 127) ### ### =============================================== v_p ================================================ ### ### ('georgia', 0, 0) ('mine', 1638, 1) ('mess', 4555, 2) ('dream', 4, 3) ('away', 6, 4) ### ### ('2014', 12, 5) ('your', 9, 6) ('life', 10, 7) ('time', 10312, 8) ('initial', 5919, 9) ### ### ('first', 21988, 10) ('″', 5, 11) ('single', 32, 12) ('lyrical', 31, 13) ('##time', 13961, 14) ### ### ('album', 16, 15) ('belgium', 7349, 16) ('australian', 13, 17) ('33', 13908, 18) ### ### ('anyway', 1999, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('georgia', 0, 0) ('dream', 4, 3) ('away', 6, 4) ('″', 5, 11) ('your', 9, 6) ('life', 10, 7) ### ### ('2014', 12, 5) ('australian', 13, 17) ('australia', 7, 39) ('album', 16, 15) ('lyrical', 31, 13) ### ### ('single', 32, 12) ('singles', 26, 24) ('」', 8, 74) ('¹', 3, 160) ('song', 28, 29) ### ### ('track', 64, 27) ('"', 15, 84) ('airplay', 74, 31) ('mine', 1638, 1) ### ############################################################################################################ [2023-11-25 20:17:39,988][root][INFO] - Epoch: 4: Step: 10401/19530, loss[v]=0.072114, lr=0.000007, acc@1[1]=998.5/1024=0.97509765625, acc@1[2]=1013.5/1024=0.98974609375 [2023-11-25 20:19:28,776][root][INFO] - Train batch 10500 [2023-11-25 20:19:28,776][root][INFO] - Avg. loss per last 100 batches: 0.130153 [2023-11-25 20:19:29,485][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.0/29522=99.93% | mean: 0.02 | max: 5.68 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in the late 19th century and into the 20th, keswick was the focus of several important ### ### initiatives by the growing conservation movement, often led by hardwicke rawnsley, vicar of the ### ### nearby crosthwaite parish and co - founder of the national trust, which has ### ### [P_TEXT]: keswick school of industrial art keswick school of industrial art ( ksia ) was founded in ### ### 1884 by canon hardwicke rawnsley and his wife edith as an evening class in woodwork and repousse ### ### metalwork at the crosthwaite parish rooms, in keswick, cumbria. the enterprise, designed to ### ### alleviate unemployment, prospered, and within ten years more than a hundred men were attending ### ### classes. a new building was erected for the school at a nearby site. the school closed in 1984 and ### ### the building became a restaurant. rawnsley was the vicar of crosthwaite, at the edge of keswick, ### ### from 1883 to 1917. he ### ### =============================================== v_q ================================================ ### ### ('ke', 0, 16) ('##ost', 1, 5) ('¹', 2, 537) ('##swick', 3, 0) ('##hwa', 4, 8) ('³', 5, 3657) ### ### ('raw', 6, 7) ('conservation', 7, 4168) ('幸', 8, 247) ('trust', 9, 11855) ('貴', 10, 233) ### ### ('hard', 11, 24) ('##wick', 12, 25) ('##ns', 13, 31) ('খ', 14, 285) ('intently', 15, 274) ### ### ('##ப', 16, 215) ('nursery', 17, 2726) ('cr', 18, 38) ('ம', 19, 133) ### ### =============================================== v_p ================================================ ### ### ('##swick', 3, 0) ('enterprise', 15415, 1) ('edith', 10232, 2) ('rooms', 4084, 3) ### ### ('industrial', 1145, 4) ('##ost', 1, 5) ('canon', 16054, 6) ('raw', 6, 7) ('##hwa', 4, 8) ### ### ('bedrooms', 10268, 9) ('evening', 1330, 10) ('##ous', 12920, 11) ('unemployment', 1759, 12) ### ### ('cumbria', 41, 13) ('rep', 21415, 14) ('worried', 430, 15) ('ke', 0, 16) ('school', 585, 17) ### ### ('art', 9979, 18) ('ks', 15955, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##swick', 3, 0) ('##ost', 1, 5) ('##hwa', 4, 8) ('ke', 0, 16) ('raw', 6, 7) ('hard', 11, 24) ### ### ('##wick', 12, 25) ('##ns', 13, 31) ('##ite', 22, 20) ('cr', 18, 38) ('cumbria', 41, 13) ### ### ('vicar', 28, 26) ('churchyard', 20, 41) ('parish', 39, 28) ('##ley', 21, 43) ('nearby', 43, 46) ### ### ('outskirts', 35, 50) ('worried', 430, 15) ('anyway', 405, 23) ('industrial', 1145, 4) ### ############################################################################################################ [2023-11-25 20:19:29,991][root][INFO] - Epoch: 4: Step: 10501/19530, loss[v]=0.181953, lr=0.000007, acc@1[1]=982.0/1024=0.958984375, acc@1[2]=1005.0/1024=0.9814453125 [2023-11-25 20:21:18,235][root][INFO] - Train batch 10600 [2023-11-25 20:21:18,236][root][INFO] - Avg. loss per last 100 batches: 0.121925 [2023-11-25 20:21:18,940][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.1/29522=99.94% | mean: 0.02 | max: 5.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: many of his building projects were continued by his son john wood, the younger including ### ### ; royal crescent, bath assembly rooms and buckland house. ### ### [P_TEXT]: royal crescent street to the circus which had been designed by john wood, the elder. the ### ### land on which the royal crescent stands was bought from sir benet garrard of the garrard baronets, ### ### who were the landlords, in december 1766. between 1767 and 1775 john wood designed the great curved ### ### facade with ionic column s on a rusticated ground floor. each original purchaser bought a length of ### ### the facade, and then employed their own architect to build a house behind the facade to their own ### ### specifications ; hence what can appear ### ### =============================================== v_q ================================================ ### ### ('wood', 0, 4) ('john', 1, 15) ('crescent', 2, 0) ('bath', 3, 3792) ('assembly', 4, 13131) ### ### ('rooms', 5, 38) ('buck', 6, 18814) ('royal', 7, 9) ('¹', 8, 559) ('##land', 9, 17543) ### ### ('younger', 10, 127) ('##ப', 11, 148) ('intently', 12, 582) ('幸', 13, 296) ('貴', 14, 249) ### ### ('son', 15, 7312) ('খ', 16, 226) ('³', 17, 7479) ('retorted', 18, 116) ('1820s', 19, 4851) ### ### =============================================== v_p ================================================ ### ### ('crescent', 2, 0) ('ionic', 8572, 1) ('circus', 3036, 2) ('##rrard', 2576, 3) ('wood', 0, 4) ### ### ('rustic', 8439, 5) ('column', 1038, 6) ('elder', 641, 7) ('columns', 3149, 8) ('royal', 7, 9) ### ### ('facade', 159, 10) ('baronetcy', 11851, 11) ('baronet', 1034, 12) ('landlord', 40, 13) ### ### ('curved', 2120, 14) ('john', 1, 15) ('elders', 4397, 16) ('ben', 9822, 17) ('1768', 9375, 18) ### ### ('##et', 21230, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('wood', 0, 4) ('crescent', 2, 0) ('royal', 7, 9) ('john', 1, 15) ('rooms', 5, 38) ### ### ('landlord', 40, 13) ('architect', 20, 29) ('architects', 38, 25) ('facade', 159, 10) ### ### ('designed', 46, 32) ('architectural', 36, 41) ('littered', 75, 26) ('younger', 10, 127) ### ### ('elder', 641, 7) ('design', 61, 42) ('column', 1038, 6) ('##ப', 11, 148) ('worried', 178, 28) ### ### ('house', 30, 61) ('architecture', 69, 47) ### ############################################################################################################ [2023-11-25 20:21:19,445][root][INFO] - Epoch: 4: Step: 10601/19530, loss[v]=0.087613, lr=0.000007, acc@1[1]=997.0/1024=0.9736328125, acc@1[2]=1014.5/1024=0.99072265625 [2023-11-25 20:23:08,206][root][INFO] - Train batch 10700 [2023-11-25 20:23:08,207][root][INFO] - Avg. loss per last 100 batches: 0.134275 [2023-11-25 20:23:08,926][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.4/29522=99.94% | mean: 0.02 | max: 5.73 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 73.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: since 1997, dominic brunt has played the role of paddy kirk in the itv soap opera " ### ### emmerdale ". ### ### [P_TEXT]: jackson walsh a gay bar in hotten. after aaron leaves, jackson finds his mobile phone and ### ### he comes to emmerdale to return it. he speaks to paddy kirk ( dominic brunt ) and gives aaron his ### ### number. aaron sees jackson again and they agree to meet for a drink, but aaron stands him up. ### ### jackson tries to encourage aaron to come out and they go on a date and share their first kiss. ### ### whilst in the woolpack, aaron believes jackson is going to out him and he punches him. aaron is ### ### arrested and jackson presses charges. adam barton ( adam thomas ) tells jackson that aaron ### ### attempted suicide and jackson goes to his court ### ### =============================================== v_q ================================================ ### ### ('##unt', 0, 12) ('paddy', 1, 1) ('dominic', 2, 10) ('¹', 3, 95) ('kirk', 4, 8) ('absently', 5, 87) ### ### ('br', 6, 20) ('貴', 7, 178) ('幸', 8, 150) ('##dale', 9, 42) ('##ப', 10, 117) ('খ', 11, 155) ### ### ('intently', 12, 222) ('³', 13, 1967) ('ம', 14, 100) ('em', 15, 933) ('guiding', 16, 9763) ### ### ('itv', 17, 67) ('realises', 18, 1424) ('##部', 19, 169) ### ### =============================================== v_p ================================================ ### ### ('walsh', 1825, 0) ('paddy', 1, 1) ('jackson', 6135, 2) ('barton', 8239, 3) ('##ten', 13188, 4) ### ### ('adam', 1032, 5) ('##pack', 3122, 6) ('hot', 15681, 7) ('kirk', 4, 8) ('thomas', 7215, 9) ### ### ('dominic', 2, 10) ('littered', 3847, 11) ('##unt', 0, 12) ('worried', 597, 13) ('wool', 7377, 14) ### ### ('gay', 1710, 15) ('anyway', 118, 16) ('soap', 20, 17) ('portrayed', 279, 18) ('role', 476, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('paddy', 1, 1) ('##unt', 0, 12) ('kirk', 4, 8) ('dominic', 2, 10) ('br', 6, 20) ('soap', 20, 17) ### ### ('##dale', 9, 42) ('¹', 3, 95) ('eastenders', 21, 34) ('absently', 5, 87) ('anyway', 118, 16) ### ### ('##mer', 23, 37) ('whilst', 24, 38) ('##ப', 10, 117) ('ம', 14, 100) ('幸', 8, 150) ### ### ('adam', 1032, 5) ('itv', 17, 67) ('portrayed', 279, 18) ('貴', 7, 178) ### ############################################################################################################ [2023-11-25 20:23:09,436][root][INFO] - Epoch: 4: Step: 10701/19530, loss[v]=0.173294, lr=0.000007, acc@1[1]=974.5/1024=0.95166015625, acc@1[2]=999.5/1024=0.97607421875 [2023-11-25 20:24:58,613][root][INFO] - Train batch 10800 [2023-11-25 20:24:58,614][root][INFO] - Avg. loss per last 100 batches: 0.134932 [2023-11-25 20:24:59,325][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.8/29522=99.93% | mean: 0.02 | max: 5.81 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: nations had granted the south seas mandate ( " nanyo " ) to japan ( which had joined the ### ### allied powers in the first world war ) which included the already japanese - held micronesia ### ### islands north of the equator that were part of the former colony of german new guinea of the german ### ### empire ; these include the modern nation / states of palau, federated states of micronesia, ### ### northern mariana islands and marshall islands. ### ### [P_TEXT]: micronesia. the sovereignty of wake island is contested : it is claimed both by the ### ### united states and by the republic of the marshall islands. the united states has actual possession ### ### of wake island, which is under the immediate administration of the united states air force. human ### ### settlement of micronesia began several millennia ago. there are competing theories about where the ### ### first micronesians came from, and when they arrived. the earliest known contact of europeans with ### ### micronesia was in 1521, when spanish ships landed in the marianas. jules dumont d'urville is ### ### usually credited ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 69) ('খ', 1, 233) ('貴', 2, 402) ('##nesia', 3, 1) ('.', 4, 105) ('seas', 5, 1144) ### ### ('幸', 6, 220) ('guinea', 7, 2723) ('mandate', 8, 1970) ('marshall', 9, 4) ('intently', 10, 430) ### ### ('##au', 11, 24331) ('##ப', 12, 290) ('mariana', 13, 8) ('german', 14, 8074) ('ம', 15, 157) ### ### ('equator', 16, 3098) ('japan', 17, 1085) ('micro', 18, 2) ('islands', 19, 37) ### ### =============================================== v_p ================================================ ### ### ('wake', 4953, 0) ('##nesia', 3, 1) ('micro', 18, 2) ('dumont', 5003, 3) ('marshall', 9, 4) ### ### ('sovereignty', 134, 5) ('jules', 3545, 6) ('152', 8935, 7) ('mariana', 13, 8) ('ur', 16719, 9) ### ### ('littered', 1251, 10) ('actual', 1617, 11) ('island', 73, 12) ('hypothesis', 14501, 13) ### ### ('states', 31, 14) ('anyway', 37, 15) ('air', 2295, 16) ('theories', 14786, 17) ### ### ('republic', 83, 18) ('millennia', 6237, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##nesia', 3, 1) ('marshall', 9, 4) ('micro', 18, 2) ('mariana', 13, 8) ('absently', 0, 69) ### ### ('islands', 19, 37) ('states', 31, 14) ('anyway', 37, 15) ('.', 4, 105) ('worried', 48, 22) ### ### ('island', 73, 12) ('sovereignty', 134, 5) ('খ', 1, 233) ('幸', 6, 220) ('stared', 28, 74) ### ### ('republic', 83, 18) ('ம', 15, 157) ('¹', 22, 106) ('貴', 2, 402) ('##ப', 12, 290) ### ############################################################################################################ [2023-11-25 20:24:59,835][root][INFO] - Epoch: 4: Step: 10801/19530, loss[v]=0.111127, lr=0.000007, acc@1[1]=998.5/1024=0.97509765625, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 20:26:48,710][root][INFO] - Train batch 10900 [2023-11-25 20:26:48,711][root][INFO] - Avg. loss per last 100 batches: 0.128885 [2023-11-25 20:26:49,419][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.9/29522=99.94% | mean: 0.02 | max: 5.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: mittelwerk gmbh also headed sites for v - 2 rocket development and testing at schlier ( ### ### project zement ) and lehesten. ### ### [P_TEXT]: redl - zipf plant in a nearby tunnel. after an august 1944 explosion at the liquid oxygen ### ### plant stopped schlier production, the third v - 2 liquid oxygen plant ( 5000 tons / month ) was ### ### built at a slate quarry at lehesten near the mittelwerk ( turbopump / chamber compatibility testing ### ### for mittelwerk production was also performed at the lehesten facility ). karl heimberg, who had ### ### worked at peenemunde test stand 7, was transferred to " vorwerk sued " at redl - zipf and then, for ### ### the period from late 1944 - early april 1945, to lehesten ( he later returned to peenemunde with ### ### walter riedel iii to burn design office ### ### =============================================== v_q ================================================ ### ### ('##hes', 0, 9) ('##ment', 1, 22920) ('##hli', 2, 2) ('absently', 3, 130) ('##tel', 4, 24) ### ### ('¹', 5, 118) ('gmbh', 6, 2342) ('v', 7, 19) ('ze', 8, 19009) ('##ten', 9, 34) ('幸', 10, 167) ### ### ('貴', 11, 213) ('খ', 12, 159) ('le', 13, 38) ('##ப', 14, 134) ('intently', 15, 326) ('2', 16, 73) ### ### ('germany', 17, 168) ('rocket', 18, 1856) ('##werk', 19, 5) ### ### =============================================== v_p ================================================ ### ### ('compatibility', 10204, 0) ('zip', 17768, 1) ('##hli', 2, 2) ('sued', 4200, 3) ('slate', 11403, 4) ### ### ('##werk', 19, 5) ('explosion', 1613, 6) ('quarry', 4869, 7) ('oxygen', 595, 8) ('##hes', 0, 9) ### ### ('##ede', 11204, 10) ('##nem', 25038, 11) ('burn', 12546, 12) ('liquid', 789, 13) ### ### ('chamber', 2733, 14) ('pee', 25857, 15) ('littered', 1730, 16) ('mit', 25, 17) ('1944', 1324, 18) ### ### ('v', 7, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##hes', 0, 9) ('##hli', 2, 2) ('##tel', 4, 24) ('v', 7, 19) ('##werk', 19, 5) ('##ten', 9, 34) ### ### ('le', 13, 38) ('mit', 25, 17) ('anyway', 30, 36) ('2', 16, 73) ('##₂', 26, 57) ('¹', 5, 118) ### ### ('absently', 3, 130) ('testing', 80, 29) ('##ப', 14, 134) ('幸', 10, 167) ('খ', 12, 159) ### ### ('ம', 20, 125) ('leaned', 69, 54) ('design', 170, 31) ### ############################################################################################################ [2023-11-25 20:26:49,928][root][INFO] - Epoch: 4: Step: 10901/19530, loss[v]=0.174680, lr=0.000007, acc@1[1]=983.0/1024=0.9599609375, acc@1[2]=1003.5/1024=0.97998046875 [2023-11-25 20:28:39,349][root][INFO] - Train batch 11000 [2023-11-25 20:28:39,350][root][INFO] - Avg. loss per last 100 batches: 0.131689 [2023-11-25 20:28:40,085][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.83 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.0/29522=99.94% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: amir asadollah alam ( 24 july 1919 – 14 april 1978 ) was an iranian politician who was ### ### prime minister during the shah's regime from 1962 to 1964. ### ### [P_TEXT]: government of ali amini ali amini was appointed to rule by decree as the prime minister ### ### of iran on 5 may 1961, succeeding jafar sharif - emami. his cabinet was approved on 9 may 1961. ### ### though amini was considered a " maverick aristocrat " and " too independent of the personal control ### ### of the monarch ", appointment of ministers of foreign affairs, war, the interior was made at the ### ### behest of the shah. all of the three portfolios, plus agriculture ministry were left unchanged in ### ### the next administration under asadollah alam. most controversially, amini gave three ministries to ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 104) ('asa', 1, 9) ('amir', 2, 12141) ('alam', 3, 2) ('³', 4, 1216) ('iranian', 5, 8) ### ### ('幸', 6, 178) ('shah', 7, 21) ('iran', 8, 6) ('貴', 9, 133) ('খ', 10, 118) ('intently', 11, 142) ### ### ('##ப', 12, 124) ('prime', 13, 18) ('##llah', 14, 20) ('ம', 15, 103) ('regime', 16, 143) ### ### ('1964', 17, 449) ('1962', 18, 41) ('absently', 19, 158) ### ### =============================================== v_p ================================================ ### ### ('amin', 97, 0) ('sharif', 529, 1) ('alam', 3, 2) ('decree', 640, 3) ('maverick', 14609, 4) ### ### ('agriculture', 5652, 5) ('iran', 8, 6) ('anyway', 185, 7) ('iranian', 5, 8) ('asa', 1, 9) ### ### ('1961', 102, 10) ('interior', 3009, 11) ('ali', 384, 12) ('##ami', 16611, 13) ### ### ('littered', 137, 14) ('ministry', 563, 15) ('monarch', 4959, 16) ('worried', 2086, 17) ### ### ('prime', 13, 18) ('cabinet', 234, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('alam', 3, 2) ('asa', 1, 9) ('iranian', 5, 8) ('iran', 8, 6) ('shah', 7, 21) ('prime', 13, 18) ### ### ('##llah', 14, 20) ('amin', 97, 0) ('1962', 18, 41) ('##do', 22, 31) ('¹', 0, 104) ### ### ('persian', 26, 34) ('1961', 102, 10) ('minister', 27, 47) ('ministers', 50, 25) ### ### ('littered', 137, 14) ('anyway', 185, 7) ('খ', 10, 118) ('government', 92, 24) ('sharif', 529, 1) ### ############################################################################################################ [2023-11-25 20:28:40,593][root][INFO] - Epoch: 4: Step: 11001/19530, loss[v]=0.197179, lr=0.000007, acc@1[1]=980.5/1024=0.95751953125, acc@1[2]=999.0/1024=0.9755859375 [2023-11-25 20:30:29,887][root][INFO] - Train batch 11100 [2023-11-25 20:30:29,888][root][INFO] - Avg. loss per last 100 batches: 0.140620 [2023-11-25 20:30:30,590][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.5/29522=99.93% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: joel skousen is the older brother of mark skousen and the younger brother of royal ### ### skousen. ### ### [P_TEXT]: mark skousen in economics from george washington university in 1977. skousen, like his ### ### family, is a devout member of the church of jesus christ of latter - day saints. conservative ### ### political commentator and survival strategist joel skousen and linguist royal skousen are his older ### ### brothers. he is the nephew of w. cleon skousen, the late conservative author of the " legendarily ### ### bizarre " john birch society tract the naked communist which claimed the soviet union would attempt ### ### to take over the united states by, among other things, " eliminating ### ### =============================================== v_q ================================================ ### ### ('##ouse', 0, 0) ('¹', 1, 563) ('joel', 2, 4) ('sk', 3, 8) ('mark', 4, 14) ('³', 5, 3452) ### ### ('royal', 6, 18) ('##n', 7, 117) ('freaking', 8, 2212) ('幸', 9, 323) ('intently', 10, 370) ### ### ('brother', 11, 35) ('貴', 12, 390) ('##ப', 13, 259) ('bumped', 14, 10715) ('খ', 15, 300) ### ### ('rapper', 16, 24034) ('！', 17, 3903) ('hissed', 18, 154) ('brothers', 19, 46) ### ### =============================================== v_p ================================================ ### ### ('##ouse', 0, 0) ('naked', 4145, 1) ('survival', 3490, 2) ('bizarre', 2610, 3) ('joel', 2, 4) ### ### ('conservative', 4856, 5) ('legend', 3441, 6) ('tract', 24940, 7) ('sk', 3, 8) ('##eon', 19685, 9) ### ### ('birch', 19775, 10) ('communist', 18336, 11) ('##ari', 27296, 12) ('devout', 26649, 13) ### ### ('mark', 4, 14) ('soviets', 28707, 15) ('society', 9847, 16) ('tracts', 24126, 17) ('royal', 6, 18) ### ### ('economics', 15866, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ouse', 0, 0) ('joel', 2, 4) ('sk', 3, 8) ('mark', 4, 14) ('royal', 6, 18) ('brother', 11, 35) ### ### ('##n', 7, 117) ('brothers', 19, 46) ('anyway', 50, 26) ('littered', 68, 24) ('¹', 1, 563) ### ### ('hissed', 18, 154) ('幸', 9, 323) ('##ப', 13, 259) ('intently', 10, 370) ('ம', 20, 178) ### ### ('nephew', 283, 36) ('খ', 15, 300) ('貴', 12, 390) ('bizarre', 2610, 3) ### ############################################################################################################ [2023-11-25 20:30:31,095][root][INFO] - Epoch: 4: Step: 11101/19530, loss[v]=0.104439, lr=0.000007, acc@1[1]=996.0/1024=0.97265625, acc@1[2]=1015.0/1024=0.9912109375 [2023-11-25 20:32:20,158][root][INFO] - Train batch 11200 [2023-11-25 20:32:20,158][root][INFO] - Avg. loss per last 100 batches: 0.131111 [2023-11-25 20:32:20,876][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.7/29522=99.93% | mean: 0.02 | max: 5.74 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: as a thought experiment, the phenomenon is usually referred to as molyneux's problem. ### ### [P_TEXT]: molyneux's problem molyneux's problem is a thought experiment in philosophy concerning ### ### immediate recovery from blindness. it was first formulated by william molyneux, and notably ### ### referred to in john locke's " an essay concerning human understanding " ( 1689 ). the problem can ### ### be stated in brief, " if a man born blind can feel the differences between shape s such as sphere s ### ### and cube s, could he, if given the ability to see, distinguish those objects by sight alone, in ### ### reference to the tactile schemata ### ### =============================================== v_q ================================================ ### ### ('##eux', 0, 1) ('thought', 1, 19) ('##lyn', 2, 15) ('¹', 3, 288) ('貴', 4, 407) ('problem', 5, 9) ### ### ('mo', 6, 31) ('experiment', 7, 17) ('intently', 8, 447) ('phenomenon', 9, 391) ('幸', 10, 271) ### ### ('absently', 11, 218) ('##ப', 12, 225) ('খ', 13, 273) ('ம', 14, 210) ('anyway', 15, 37) ### ### ('³', 16, 3353) ('effect', 17, 4500) ('.', 18, 130) ('worried', 19, 66) ### ### =============================================== v_p ================================================ ### ### ('cube', 1262, 0) ('##eux', 0, 1) ('sphere', 855, 2) ('understanding', 498, 3) ### ### ('blindness', 9261, 4) ('##hema', 4105, 5) ('shape', 2322, 6) ('locke', 815, 7) ('1689', 17193, 8) ### ### ('problem', 5, 9) ('recovery', 5166, 10) ('spheres', 7726, 11) ('blind', 3868, 12) ### ### ('immediate', 7520, 13) ('concerning', 255, 14) ('##lyn', 2, 15) ('experiments', 762, 16) ### ### ('experiment', 7, 17) ('sight', 1572, 18) ('thought', 1, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##eux', 0, 1) ('problem', 5, 9) ('##lyn', 2, 15) ('thought', 1, 19) ('experiment', 7, 17) ### ### ('mo', 6, 31) ('anyway', 15, 37) ('understanding', 498, 3) ('concerning', 255, 14) ### ### ('cube', 1262, 0) ('worried', 19, 66) ('sphere', 855, 2) ('¹', 3, 288) ('philosophy', 115, 36) ### ### ('absently', 11, 218) ('##ப', 12, 225) ('幸', 10, 271) ('ம', 14, 210) ('locke', 815, 7) ### ### ('貴', 4, 407) ### ############################################################################################################ [2023-11-25 20:32:21,386][root][INFO] - Epoch: 4: Step: 11201/19530, loss[v]=0.075634, lr=0.000007, acc@1[1]=1003.5/1024=0.97998046875, acc@1[2]=1017.0/1024=0.9931640625 [2023-11-25 20:34:10,940][root][INFO] - Train batch 11300 [2023-11-25 20:34:10,941][root][INFO] - Avg. loss per last 100 batches: 0.133616 [2023-11-25 20:34:11,662][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.3/29522=99.94% | mean: 0.02 | max: 5.66 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.6/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: within gmina drawno is a large forest ( puszcza drawska ) which is part of the drawno ### ### national park. ### ### [P_TEXT]: pszczewko pszczewko ( ) is a village in the administrative district of gmina drawno, ### ### within choszczno county, west pomeranian voivodeship, in north - western poland. it lies ### ### approximately east of choszczno and east of the regional capital szczecin. before 1945 the village ### ### was german - settled and part of the german state of prussia. ### ### =============================================== v_q ================================================ ### ### ('drawn', 0, 1) ('##o', 1, 17) ('¹', 2, 381) ('##ʂ', 3, 57) ('draws', 4, 43) ('##ो', 5, 53) ### ### ('drew', 6, 31) ('forest', 7, 7404) ('貴', 8, 138) ('幸', 9, 152) ('poland', 10, 91) ### ### ('park', 11, 14239) ('intently', 12, 211) ('³', 13, 1824) ('খ', 14, 108) ('national', 15, 11572) ### ### ('pu', 16, 16511) ('drawing', 17, 30) ('##ப', 18, 71) ('gmina', 19, 226) ### ### =============================================== v_p ================================================ ### ### ('ps', 7377, 0) ('drawn', 0, 1) ('leaned', 69, 2) ('##ew', 15966, 3) ('cho', 7068, 4) ### ### ('prussia', 40, 5) ('##cz', 388, 6) ('worried', 126, 7) ('##sz', 33, 8) ('##no', 206, 9) ### ### ('##ko', 276, 10) ('##zcz', 243, 11) ('seat', 5653, 12) ('german', 298, 13) ('pomeranian', 23, 14) ### ### ('germans', 359, 15) ('poles', 642, 16) ('##o', 1, 17) ('prussian', 100, 18) ('anyway', 62, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('drawn', 0, 1) ('##o', 1, 17) ('drew', 6, 31) ('pomeranian', 23, 14) ('leaned', 69, 2) ### ### ('##ʂ', 3, 57) ('##sz', 33, 8) ('draws', 4, 43) ('prussia', 40, 5) ('drawing', 17, 30) ### ### ('##ो', 5, 53) ('outskirts', 21, 33) ('polish', 26, 25) ('pomerania', 30, 26) ('worried', 126, 7) ### ### ('poland', 10, 91) ('##ப', 18, 71) ('anyway', 62, 19) ('খ', 14, 108) ('ம', 20, 83) ### ############################################################################################################ [2023-11-25 20:34:12,174][root][INFO] - Epoch: 4: Step: 11301/19530, loss[v]=0.142191, lr=0.000007, acc@1[1]=987.0/1024=0.9638671875, acc@1[2]=1011.0/1024=0.9873046875 [2023-11-25 20:36:01,310][root][INFO] - Train batch 11400 [2023-11-25 20:36:01,310][root][INFO] - Avg. loss per last 100 batches: 0.124890 [2023-11-25 20:36:02,016][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.97 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.2/29522=99.94% | mean: 0.02 | max: 6.00 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the former governor of the reserve bank of india, raghuram rajan, had predicted the ### ### crisis in 2005 when he became chief economist at the international monetary fund. ### ### [P_TEXT]: raghuram rajan of the time. these risks are known as tail risks. but perhaps the most ### ### important concern is whether banks will be able to provide liquidity to financial markets so that ### ### if the tail risk does materialise, financial positions can be unwound and losses allocated so that ### ### the consequences to the real economy are minimised. " the response to rajan's paper at the time was ### ### negative. for example, former u. s. treasury secretary and former harvard president lawrence ### ### summers called the warnings " misguided " and rajan himself a " luddite ". however, following the ### ### financial crisis of 2007 – 2008, rajan's views came ### ### =============================================== v_q ================================================ ### ### ('raja', 0, 1) ('¹', 1, 339) ('absently', 2, 97) ('monetary', 3, 1667) ('貴', 4, 276) ### ### ('intently', 5, 287) ('bank', 6, 35) ('crisis', 7, 11) ('reserve', 8, 8110) ('幸', 9, 204) ### ### ('rag', 10, 8) ('খ', 11, 177) ('##ப', 12, 124) ('##hur', 13, 10) ('india', 14, 83) ('ம', 15, 138) ### ### ('governor', 16, 1175) ('##am', 17, 30) ('2005', 18, 921) ('fund', 19, 1161) ### ### =============================================== v_p ================================================ ### ### ('summers', 27185, 0) ('raja', 0, 1) ('tail', 10331, 2) ('harvard', 9947, 3) ('##dd', 27949, 4) ### ### ('##guide', 15382, 5) ('lawrence', 5396, 6) ('liquid', 4715, 7) ('rag', 10, 8) ('financial', 51, 9) ### ### ('##hur', 13, 10) ('crisis', 7, 11) ('risk', 198, 12) ('##ite', 8834, 13) ('mini', 18129, 14) ### ### ('economy', 126, 15) ('crises', 5237, 16) ('##ity', 1734, 17) ('littered', 1861, 18) ### ### ('2007', 120, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('raja', 0, 1) ('crisis', 7, 11) ('rag', 10, 8) ('##hur', 13, 10) ('bank', 6, 35) ('##am', 17, 30) ### ### ('##n', 20, 45) ('financial', 51, 9) ('absently', 2, 97) ('india', 14, 83) ('economy', 126, 15) ### ### ('risk', 198, 12) ('2007', 120, 19) ('##ப', 12, 124) ('anyway', 63, 36) ('ம', 15, 138) ### ### ('recession', 115, 26) ('খ', 11, 177) ('##部', 24, 91) ('economic', 35, 63) ### ############################################################################################################ [2023-11-25 20:36:02,519][root][INFO] - Epoch: 4: Step: 11401/19530, loss[v]=0.092821, lr=0.000007, acc@1[1]=1001.5/1024=0.97802734375, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 20:37:51,547][root][INFO] - Train batch 11500 [2023-11-25 20:37:51,548][root][INFO] - Avg. loss per last 100 batches: 0.114943 [2023-11-25 20:37:52,257][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.04 | max: 4.01 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.2/29522=99.94% | mean: 0.02 | max: 5.81 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: sprecher's films have starred matthew mcconaughey, lisa kudrow, and greg kinnear. ### ### [P_TEXT]: clockwatchers clockwatchers is an american comedy - drama film released in 1997. directed ### ### by jill sprecher, it stars parker posey, lisa kudrow, toni collette and alanna ubach as temporary ### ### office staffers in an office complex. the four become misfit friends in an office environment where ### ### they are ignored and mistrusted by their co - workers. iris ( collette ) is a shy young ### ### professional who doesn't want to rock the boat at the office where she temps. margaret ( posey ) is ### ### the polar opposite, and serves as ### ### =============================================== v_q ================================================ ### ### ('##dro', 0, 14) ('##cher', 1, 7) ('sp', 2, 18) ('¹', 3, 92) ('##re', 4, 28) ('kin', 5, 17364) ### ### ('lisa', 6, 16) ('##nea', 7, 27165) ('matthew', 8, 18722) ('greg', 9, 9923) ('幸', 10, 133) ### ### ('貴', 11, 197) ('mcc', 12, 26001) ('absently', 13, 66) ('intently', 14, 192) ('film', 15, 38) ### ### ('##ugh', 16, 25420) ('³', 17, 1675) ('##ப', 18, 91) ('ku', 19, 25) ### ### =============================================== v_p ================================================ ### ### ('##watch', 1385, 0) ('clock', 16136, 1) ('pose', 19530, 2) ('iris', 4839, 3) ('##bach', 23514, 4) ### ### ('##lette', 16450, 5) ('jill', 1264, 6) ('##cher', 1, 7) ('1997', 869, 8) ('toni', 5021, 9) ### ### ('##ers', 18475, 10) ('worried', 144, 11) ('margaret', 10743, 12) ('parker', 1686, 13) ### ### ('##dro', 0, 14) ('littered', 256, 15) ('lisa', 6, 16) ('opposite', 7016, 17) ('sp', 2, 18) ### ### ('clocks', 28082, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##cher', 1, 7) ('##dro', 0, 14) ('sp', 2, 18) ('lisa', 6, 16) ('##re', 4, 28) ('ku', 19, 25) ### ### ('film', 15, 38) ('¹', 3, 92) ('absently', 13, 66) ('##w', 24, 51) ('worried', 144, 11) ### ### ('anyway', 39, 23) ('##ப', 18, 91) ('filmmakers', 28, 47) ('ம', 23, 86) ('幸', 10, 133) ### ### ('##watch', 1385, 0) ('খ', 20, 126) ('littered', 256, 15) ('films', 54, 36) ### ############################################################################################################ [2023-11-25 20:37:52,768][root][INFO] - Epoch: 4: Step: 11501/19530, loss[v]=0.204609, lr=0.000007, acc@1[1]=979.5/1024=0.95654296875, acc@1[2]=1003.0/1024=0.9794921875 [2023-11-25 20:39:42,459][root][INFO] - Train batch 11600 [2023-11-25 20:39:42,459][root][INFO] - Avg. loss per last 100 batches: 0.125495 [2023-11-25 20:39:43,170][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.0/29522=99.94% | mean: 0.02 | max: 5.84 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: produced by sunjoy waddhwa under sphere origins, it starred avinash sachdev and shrenu ### ### parikh as shlok and aastha, respectively. ### ### [P_TEXT]: avinash sachdev organised by kanal 7 for the fans of " tatlı bela " to meet avinash ### ### sachdev as shlok agnihotri from " iss pyaar ko kya naam doon? ek baar phir ". in 2015, sachdev ### ### joined the cast of " qubool hai 4 " ( 2015 – 2016 ) portraying armaan raza sheikh, a wealthy ### ### landowner. sachdev was also a part of " balika vadhu – lamhe pyaar ke " in 2016. this was his third ### ### television project with the production company, sphere origins. in 2019, sachdev starred in & tv's ### ### " main bhi ardhangini ### ### =============================================== v_q ================================================ ### ### ('##ikh', 0, 12688) ('absently', 1, 150) ('##lok', 2, 2) ('##st', 3, 27457) ('sh', 4, 31) ### ### ('sphere', 5, 1) ('幸', 6, 182) ('##ren', 7, 27487) ('¹', 8, 172) ('貴', 9, 224) ('##inas', 10, 0) ### ### ('##ப', 11, 119) ('খ', 12, 204) ('par', 13, 19912) ('aa', 14, 19214) ('intently', 15, 238) ### ### ('″', 16, 30) ('sac', 17, 4) ('av', 18, 10) ('origins', 19, 15) ### ### =============================================== v_p ================================================ ### ### ('##inas', 10, 0) ('sphere', 5, 1) ('##lok', 2, 2) ('iss', 15821, 3) ('sac', 17, 4) ### ### ('##aan', 17340, 5) ('bali', 5753, 6) ('phi', 10719, 7) ('lam', 8891, 8) ('bela', 16601, 9) ### ### ('av', 18, 10) ('##hd', 22, 11) ('kan', 17047, 12) ('&', 4092, 13) ('sheikh', 684, 14) ### ### ('origins', 19, 15) ('7', 15145, 16) ('##tl', 26538, 17) ('##ı', 2767, 18) ('##ol', 21506, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('sphere', 5, 1) ('##inas', 10, 0) ('##lok', 2, 2) ('sac', 17, 4) ('av', 18, 10) ('##hd', 22, 11) ### ### ('sh', 4, 31) ('origins', 19, 15) ('″', 16, 30) ('##ev', 25, 28) ('absently', 1, 150) ### ### ('##ப', 11, 119) ('hissed', 41, 41) ('anyway', 46, 38) ('¹', 8, 172) ('幸', 6, 182) ('tv', 93, 24) ### ### ('##部', 26, 111) ('ம', 23, 128) ('leaned', 69, 37) ### ############################################################################################################ [2023-11-25 20:39:43,676][root][INFO] - Epoch: 4: Step: 11601/19530, loss[v]=0.071257, lr=0.000007, acc@1[1]=1002.0/1024=0.978515625, acc@1[2]=1016.5/1024=0.99267578125 [2023-11-25 20:41:32,760][root][INFO] - Train batch 11700 [2023-11-25 20:41:32,761][root][INFO] - Avg. loss per last 100 batches: 0.127681 [2023-11-25 20:41:33,461][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.8/29522=99.94% | mean: 0.02 | max: 5.66 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.9/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in the 1st round, wayne jones 13 – 7 in the 2nd round, wes newton 16 – 5 in the quarter - ### ### final, andy hamilton 17 – 9 in the semi - final and james wade 18 – 8 in the final. ### ### [P_TEXT]: 2013 championship league darts van barneveld ( 10 ) and vincent van der voort ( 29 ) ### ### chose not to compete in the tournament. played on tuesday 24 september and was won by phil taylor. ### ### he also hit a nine - dart finish during a 6 – 5 league stage defeat to adrian lewis. simon whitlock ### ### and james wade were eliminated. played on wednesday 25 september and was won by andy hamilton. ### ### robert thornton and justin pipe were eliminated. played on thursday 26 september and was won by wes ### ### newton. brendan dolan and adrian lewis were eliminated. played on tuesday 15 october and ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 116) ('wes', 1, 6) ('貴', 2, 101) ('.', 3, 94) ('幸', 4, 99) ('andy', 5, 25) ### ### ('##ப', 6, 77) ('wayne', 7, 13663) ('jones', 8, 8941) ('খ', 9, 103) ('james', 10, 93) ('ம', 11, 80) ### ### ('hamilton', 12, 2) ('intently', 13, 79) ('wade', 14, 4) ('darts', 15, 7) ('##部', 16, 107) ### ### ('snooker', 17, 66) ('newton', 18, 0) ('round', 19, 24) ### ### =============================================== v_p ================================================ ### ### ('newton', 18, 0) ('pipe', 3201, 1) ('hamilton', 12, 2) ('thornton', 6061, 3) ('wade', 14, 4) ### ### ('dart', 830, 5) ('wes', 1, 6) ('darts', 15, 7) ('2013', 214, 8) ('##lock', 15688, 9) ### ### ('##lan', 14111, 10) ('##ort', 20075, 11) ('anyway', 23, 12) ('championship', 31, 13) ### ### ('taylor', 3931, 14) ('justin', 8396, 15) ('littered', 270, 16) ('brendan', 3315, 17) ### ### ('lewis', 2388, 18) ('adrian', 2139, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('wes', 1, 6) ('hamilton', 12, 2) ('wade', 14, 4) ('newton', 18, 0) ('darts', 15, 7) ### ### ('andy', 5, 25) ('anyway', 23, 12) ('round', 19, 24) ('championship', 31, 13) ('he', 28, 26) ### ### ('absently', 0, 116) ('final', 30, 28) ('##ப', 6, 77) ('.', 3, 94) ('ம', 11, 80) ('貴', 2, 101) ### ### ('intently', 13, 79) ('幸', 4, 99) ('james', 10, 93) ('¹', 21, 44) ### ############################################################################################################ [2023-11-25 20:41:33,972][root][INFO] - Epoch: 4: Step: 11701/19530, loss[v]=0.175737, lr=0.000007, acc@1[1]=973.0/1024=0.9501953125, acc@1[2]=1007.5/1024=0.98388671875 [2023-11-25 20:43:23,088][root][INFO] - Train batch 11800 [2023-11-25 20:43:23,089][root][INFO] - Avg. loss per last 100 batches: 0.133381 [2023-11-25 20:43:23,812][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.75 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.0/29522=99.94% | mean: 0.02 | max: 5.68 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: coached by bennett and captained under wally lewis at the broncos, langer developed ### ### further skill with the ball and learnt much from his mentors, and his two tries in game i of the ### ### 1988 state of origin series saw him named man of the match again. ### ### [P_TEXT]: craig grauf the preliminary final. he was then signed by the brisbane broncos for their ### ### inaugural season, coming off the bench in the club ’ s maiden premiership game and producing the ### ### bomb that led to brisbane's first - ever try ( scored by brett le man ) in the nswrl competition ### ### against manly in round one at lang park. after playing only five first - grade games over two ### ### seasons at the broncos, due to being stuck behind the likes of allan langer and wally lewis in his ### ### chosen positions, grauf went back to norths in the ### ### =============================================== v_q ================================================ ### ### ('lange', 0, 4) ('¹', 1, 106) ('##r', 2, 45) ('1988', 3, 16986) ('wally', 4, 1) ('lewis', 5, 6) ### ### ('bennett', 6, 2702) ('³', 7, 1478) ('ɐ', 8, 741) ('origin', 9, 3141) ('খ', 10, 140) ('幸', 11, 161) ### ### ('##ப', 12, 97) ('broncos', 13, 11) ('貴', 14, 134) ('ம', 15, 88) ('absently', 16, 92) ### ### ('ashes', 17, 1043) ('intently', 18, 126) ('rugby', 19, 31) ### ### =============================================== v_p ================================================ ### ### ('bomb', 9407, 0) ('wally', 4, 1) ('lang', 8941, 2) ('craig', 102, 3) ('lange', 0, 4) ### ### ('brett', 74, 5) ('lewis', 5, 6) ('allan', 872, 7) ('le', 15804, 8) ('preliminary', 1013, 9) ### ### ('premiership', 96, 10) ('broncos', 13, 11) ('park', 1463, 12) ('worried', 1067, 13) ### ### ('anyway', 419, 14) ('nsw', 456, 15) ('##au', 20887, 16) ('gr', 8062, 17) ('manly', 126, 18) ### ### ('bombs', 14875, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('lange', 0, 4) ('wally', 4, 1) ('lewis', 5, 6) ('broncos', 13, 11) ('##r', 2, 45) ### ### ('brisbane', 20, 22) ('littered', 22, 20) ('craig', 102, 3) ('rugby', 19, 31) ('brett', 74, 5) ### ### ('¹', 1, 106) ('premiership', 96, 10) ('man', 54, 26) ('manly', 126, 18) ('tries', 24, 47) ### ### ('##ப', 12, 97) ('sydney', 36, 41) ('ம', 15, 88) ('try', 59, 36) ('absently', 16, 92) ### ############################################################################################################ [2023-11-25 20:43:24,325][root][INFO] - Epoch: 4: Step: 11801/19530, loss[v]=0.141326, lr=0.000007, acc@1[1]=975.5/1024=0.95263671875, acc@1[2]=1006.5/1024=0.98291015625 [2023-11-25 20:45:13,934][root][INFO] - Train batch 11900 [2023-11-25 20:45:13,935][root][INFO] - Avg. loss per last 100 batches: 0.136989 [2023-11-25 20:45:14,652][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.9/29522=99.94% | mean: 0.02 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: throughout the second half of the 19th century and leading up to the first world war, the ### ### port of hull played a major role in the emigration of northern european settlers to the new world, ### ### with thousands of emigrants sailing to the city and ### ### [P_TEXT]: harry dale ( rugby league ) as a or, i. e. number 6 or 7. harry dale was born in kingston ### ### upon hull, east riding of yorkshire, england, he was a pupil at buckingham street school ( now ### ### buckingham primary academy ), kingston upon hull, he worked as a stevedore at the port of hull, he ### ### sustained a serious head injury while working with a royal engineers bomb disposal unit during ### ### world war ii, he experienced ill - health in later life and following several spells in hospital, ### ### he died aged ### ### =============================================== v_q ================================================ ### ### ('hull', 0, 3) ('¹', 1, 1750) ('emigration', 2, 11545) ('port', 3, 6) ('.', 4, 1526) ### ### ('yorkshire', 5, 26) ('wakefield', 6, 1082) ('settler', 7, 21789) ('খ', 8, 340) ('幸', 9, 455) ### ### ('europeans', 10, 17235) ('³', 11, 8949) ('northern', 12, 7577) ('england', 13, 2296) ### ### ('halifax', 14, 628) ('european', 15, 12187) ('##ப', 16, 287) ('ம', 17, 225) ('absently', 18, 141) ### ### ('ontario', 19, 8968) ### ### =============================================== v_p ================================================ ### ### ('buckingham', 8604, 0) ('dale', 563, 1) ('##dor', 26582, 2) ('hull', 0, 3) ('harry', 9735, 4) ### ### ('kingston', 38, 5) ('port', 3, 6) ('worried', 61, 7) ('bomb', 2137, 8) ('steve', 23529, 9) ### ### ('injury', 14666, 10) ('engineers', 6694, 11) ('bombed', 26728, 12) ('rugby', 620, 13) ### ### ('anyway', 1256, 14) ('ports', 535, 15) ('riding', 986, 16) ('pupils', 12443, 17) ### ### ('numbers', 821, 18) ('disposal', 4697, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('hull', 0, 3) ('port', 3, 6) ('kingston', 38, 5) ('yorkshire', 5, 26) ('worried', 61, 7) ### ### ('dale', 563, 1) ('outskirts', 28, 67) ('wwii', 54, 59) ('leaned', 167, 27) ('ports', 535, 15) ### ### ('rugby', 620, 13) ('absently', 18, 141) ('leeds', 26, 106) ('settlement', 40, 80) ('war', 39, 90) ### ### ('towns', 80, 70) ('numbers', 821, 18) ('town', 87, 68) ('riding', 986, 16) ('খ', 8, 340) ### ############################################################################################################ [2023-11-25 20:45:15,163][root][INFO] - Epoch: 4: Step: 11901/19530, loss[v]=0.235993, lr=0.000007, acc@1[1]=974.5/1024=0.95166015625, acc@1[2]=995.5/1024=0.97216796875 [2023-11-25 20:47:04,293][root][INFO] - Train batch 12000 [2023-11-25 20:47:04,293][root][INFO] - Avg. loss per last 100 batches: 0.144859 [2023-11-25 20:47:05,002][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.7/29522=99.94% | mean: 0.02 | max: 5.77 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: ruler of pavia from 1315, five years later he was " podesta " of vigevano, where he ### ### erected the castle that is still visible. ### ### [P_TEXT]: visconti - sforza castle ( vigevano ) the visconti - sforza castle is a mediaeval castle ### ### located in the centre of the city of vigevano, lombardy, northern italy. in the 14th and 15th ### ### centuries, members of the visconti and sforza houses, lords and dukes of milan, transformed a ### ### previous fortification into a vast family resort. the castle was part of a wider plan of urban ### ### development for vigevano, which included the erection of other buildings and the construction of ### ### the central piazza ducale. luchino visconti, lord of milan and ruler of vigevano ### ### =============================================== v_q ================================================ ### ### ('##van', 0, 9) ('##ge', 1, 15) ('castle', 2, 5) ('খ', 3, 146) ('##via', 4, 21709) ### ### ('absently', 5, 115) ('¹', 6, 164) ('pa', 7, 21661) ('intently', 8, 172) ('##est', 9, 23642) ### ### ('貴', 10, 158) ('##ப', 11, 129) ('131', 12, 296) ('pod', 13, 17730) ('幸', 14, 175) ('vi', 15, 12) ### ### ('ம', 16, 118) ('##ᵉ', 17, 1710) ('³', 18, 2881) ('##ʂ', 19, 2880) ### ### =============================================== v_p ================================================ ### ### ('piazza', 880, 0) ('sf', 14644, 1) ('dukes', 4525, 2) ('worried', 252, 3) ('##ale', 21294, 4) ### ### ('castle', 2, 5) ('luc', 1382, 6) ('duc', 482, 7) ('castles', 3776, 8) ('##van', 0, 9) ### ### ('milan', 292, 10) ('vis', 5134, 11) ('vi', 15, 12) ('##con', 11049, 13) ('leaned', 2480, 14) ### ### ('##ge', 1, 15) ('anyway', 24, 16) ('15th', 638, 17) ('##hin', 17084, 18) ('plan', 2604, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##van', 0, 9) ('castle', 2, 5) ('##ge', 1, 15) ('vi', 15, 12) ('anyway', 24, 16) ('##o', 22, 31) ### ### ('lombardy', 28, 20) ('duke', 43, 30) ('worried', 252, 3) ('absently', 5, 115) ('medieval', 34, 43) ### ### ('milan', 292, 10) ('italy', 46, 39) ('খ', 3, 146) ('##ப', 11, 129) ('ம', 16, 118) ('duc', 482, 7) ### ### ('¹', 6, 164) ('貴', 10, 158) ('piazza', 880, 0) ### ############################################################################################################ [2023-11-25 20:47:05,506][root][INFO] - Epoch: 4: Step: 12001/19530, loss[v]=0.149101, lr=0.000007, acc@1[1]=980.5/1024=0.95751953125, acc@1[2]=1008.0/1024=0.984375 [2023-11-25 20:48:54,857][root][INFO] - Train batch 12100 [2023-11-25 20:48:54,858][root][INFO] - Avg. loss per last 100 batches: 0.123834 [2023-11-25 20:48:55,597][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.2/29522=99.93% | mean: 0.02 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he worked for lanvin and balmain. ### ### [P_TEXT]: duckie thot her darker complexion. she made the decision to move to brooklyn, new york in ### ### hopes that she could become a more successful model in america. after scheduling meetings with ### ### multiple modeling agencies, thot flew to new york. in new york, she got multiple offers to sign ### ### with agencies and ultimately chose to sign with new york model management. thot has done campaigns ### ### for fenty x puma, fenty beauty, moschino, balmain, and oscar de la renta. she debuted in the 2018 ### ### victoria's secret fashion show. she appeared on the cover art for chic's album " it's about time " ### ### =============================================== v_q ================================================ ### ### ('##main', 0, 0) ('lan', 1, 17206) ('bal', 2, 17) ('##vin', 3, 22359) ('¹', 4, 782) ### ### ('cerambycidae', 5, 898) ('貴', 6, 204) ('.', 7, 6284) ('##½', 8, 1382) ('anyway', 9, 37) ### ### ('幸', 10, 231) ('খ', 11, 178) ('littered', 12, 63) ('intently', 13, 320) ('³', 14, 3075) ### ### ('pained', 15, 1027) ('##ப', 16, 150) ('№', 17, 9921) ('worked', 18, 4276) ('bellowed', 19, 907) ### ### =============================================== v_p ================================================ ### ### ('##main', 0, 0) ('tho', 11250, 1) ('duck', 11150, 2) ('rent', 8882, 3) ('fen', 3681, 4) ### ### ('oscar', 4834, 5) ('##ino', 28163, 6) ('fashion', 1263, 7) ('chi', 16991, 8) ('about', 24850, 9) ### ### ('##sch', 25643, 10) ('##ty', 15300, 11) ('modelling', 14972, 12) ('pu', 24959, 13) ### ### ('brooklyn', 1754, 14) ('victoria', 10479, 15) ('secret', 17214, 16) ('bal', 2, 17) ### ### ('time', 8200, 18) ('model', 4868, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##main', 0, 0) ('bal', 2, 17) ('anyway', 9, 37) ('littered', 12, 63) ('worried', 260, 29) ### ### ('貴', 6, 204) ('absently', 31, 71) ('fashion', 1263, 7) ('খ', 11, 178) ('ம', 21, 123) ### ### ('hissed', 85, 51) ('##ப', 16, 150) ('management', 855, 20) ('幸', 10, 231) ('fen', 3681, 4) ### ### ('stared', 83, 64) ('brooklyn', 1754, 14) ('intently', 13, 320) ('york', 1825, 23) ### ### ('concerning', 1829, 26) ### ############################################################################################################ [2023-11-25 20:48:56,105][root][INFO] - Epoch: 4: Step: 12101/19530, loss[v]=0.219739, lr=0.000007, acc@1[1]=977.0/1024=0.9541015625, acc@1[2]=1006.0/1024=0.982421875 [2023-11-25 20:50:45,265][root][INFO] - Train batch 12200 [2023-11-25 20:50:45,266][root][INFO] - Avg. loss per last 100 batches: 0.130872 [2023-11-25 20:50:45,976][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.9/29522=99.94% | mean: 0.02 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: at iu she was a pupil of renowned soprano virginia zeani. ### ### [P_TEXT]: virginia zeani continued teaching there after rossi - lemeni's death in 1991 and was ### ### awarded the title of distinguished professor of music in 1994. among her many students at the ### ### jacobs school who went on to international careers as opera singers are angela brown, nicole ### ### chevalier, vivica genaux, sylvia mcnair, marilyn mims, mark nicolson, susan patterson, elizabeth ### ### futral and ailyn perez. zeani retired to west palm beach, florida in 2004, but has continued to ### ### teach students privately. she was awarded indiana ### ### =============================================== v_q ================================================ ### ### ('ze', 0, 10) ('##ani', 1, 11) ('virginia', 2, 15) ('##子', 3, 69) ('university', 4, 1419) ### ### ('pupil', 5, 79) ('¹', 6, 75) ('herself', 7, 60) ('soprano', 8, 65) ('貴', 9, 129) ### ### ('intently', 10, 106) ('opera', 11, 21) ('и', 12, 3092) ('খ', 13, 135) ('幸', 14, 143) ### ### ('³', 15, 839) ('pupils', 16, 74) ('singers', 17, 52) ('helen', 18, 3830) ('##ப', 19, 122) ### ### =============================================== v_p ================================================ ### ### ('jacobs', 3363, 0) ('chevalier', 547, 1) ('angela', 389, 2) ('##nai', 11091, 3) ### ### ('##lson', 10832, 4) ('patterson', 6005, 5) ('nicole', 861, 6) ('rossi', 940, 7) ### ### ('##tral', 25203, 8) ('##aux', 18106, 9) ('ze', 0, 10) ('##ani', 1, 11) ('##lyn', 2970, 12) ### ### ('marilyn', 799, 13) ('nico', 6026, 14) ('virginia', 2, 15) ('sylvia', 28, 16) ('##ms', 21495, 17) ### ### ('brown', 7161, 18) ('susan', 497, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ze', 0, 10) ('##ani', 1, 11) ('virginia', 2, 15) ('opera', 11, 21) ('sylvia', 28, 16) ### ### ('##子', 3, 69) ('anyway', 34, 31) ('herself', 7, 60) ('soprano', 8, 65) ('angela', 389, 2) ### ### ('¹', 6, 75) ('pupil', 5, 79) ('chevalier', 547, 1) ('singers', 17, 52) ('she', 96, 36) ### ### ('hissed', 129, 34) ('susan', 497, 19) ('nicole', 861, 6) ('intently', 10, 106) ('rossi', 940, 7) ### ############################################################################################################ [2023-11-25 20:50:46,480][root][INFO] - Epoch: 4: Step: 12201/19530, loss[v]=0.075911, lr=0.000007, acc@1[1]=1003.0/1024=0.9794921875, acc@1[2]=1017.0/1024=0.9931640625 [2023-11-25 20:52:34,910][root][INFO] - Train batch 12300 [2023-11-25 20:52:34,911][root][INFO] - Avg. loss per last 100 batches: 0.124695 [2023-11-25 20:52:35,621][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29505.7/29522=99.94% | mean: 0.03 | max: 5.91 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: with the assistance of her husband, francis minor ( a lawyer ), she brought an action in ### ### state courts against reese happersett, the registrar who had rejected her application to register ### ### [P_TEXT]: francis minor case to the supreme court of the united states, argued the case, and lost ### ### in a unanimous decision. minor v. happersett, 88 u. s. 162 ( 1874 ). when francis minor died at age ### ### 71, on february 19, 1892 ( two years before his wife ), susan b. anthony wrote about him, “ no man ### ### has contributed to the woman suffrage movement so much valuable constitutional argument and proof ### ### as mr. minor. ” francis is buried at bellefontaine cemetery next to his wife and their only child. ### ### " coincidentally, in a unmarked grave just across the cemetery road, ### ### =============================================== v_q ================================================ ### ### ('minor', 0, 0) ('francis', 1, 1) ('registrar', 2, 4458) ('##sett', 3, 5) ('##pper', 4, 4) ### ### ('reese', 5, 4440) ('¹', 6, 617) ('register', 7, 10928) ('virginia', 8, 2732) ('carolina', 9, 8298) ### ### ('intently', 10, 407) ('##ப', 11, 157) ('ha', 12, 107) ('applicant', 13, 24204) ### ### ('townspeople', 14, 3429) ('registry', 15, 13292) ('貴', 16, 385) ('幸', 17, 411) ('maine', 18, 8676) ### ### ('case', 19, 31) ### ### =============================================== v_p ================================================ ### ### ('minor', 0, 0) ('francis', 1, 1) ('unanimous', 14781, 2) ('proof', 11899, 3) ('##pper', 4, 4) ### ### ('##sett', 3, 5) ('##nta', 22384, 6) ('162', 16318, 7) ('cemetery', 2024, 8) ('belle', 481, 9) ### ### ('argument', 3239, 10) ('woman', 372, 11) ('anthony', 6991, 12) ('suffrage', 82, 13) ### ### ('unmarked', 22979, 14) ('constitution', 480, 15) ('constitutional', 2409, 16) ### ### ('driveway', 1395, 17) ('arguments', 5795, 18) ('worried', 579, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('minor', 0, 0) ('francis', 1, 1) ('##pper', 4, 4) ('##sett', 3, 5) ('court', 33, 20) ### ### ('case', 19, 31) ('suffrage', 82, 13) ('herself', 21, 54) ('husband', 23, 60) ('wife', 53, 38) ### ### ('woman', 372, 11) ('belle', 481, 9) ('cases', 100, 34) ('women', 158, 25) ('anyway', 131, 33) ### ### ('ha', 12, 107) ('##子', 22, 85) ('constitution', 480, 15) ('defendant', 159, 36) ### ### ('absently', 39, 76) ### ############################################################################################################ [2023-11-25 20:52:36,126][root][INFO] - Epoch: 4: Step: 12301/19530, loss[v]=0.094698, lr=0.000007, acc@1[1]=997.0/1024=0.9736328125, acc@1[2]=1015.0/1024=0.9912109375 [2023-11-25 20:54:24,922][root][INFO] - Train batch 12400 [2023-11-25 20:54:24,923][root][INFO] - Avg. loss per last 100 batches: 0.127615 [2023-11-25 20:54:25,647][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.4/29522=99.94% | mean: 0.02 | max: 5.74 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: a live - action adaptation of carroll's works and the animated film, " alice in ### ### wonderland ", directed by tim burton, was released in 2010. ### ### [P_TEXT]: alice through the looking glass ( 2016 film ) alice through the looking glass is a 2016 ### ### american live - action / animated fantasy adventure film directed by james bobin, written by linda ### ### woolverton and produced by tim burton, joe roth, suzanne todd, and jennifer todd. it is based on ### ### the characters created by lewis carroll and is the sequel to the 2010 film " alice in wonderland ", ### ### a live - action reimagining of disney's 1951 animated film of the same name. the film ### ### =============================================== v_q ================================================ ### ### ('wonderland', 0, 12) ('alice', 1, 1) ('absently', 2, 40) ('carroll', 3, 16) ('貴', 4, 188) ### ### ('¹', 5, 120) ('幸', 6, 127) ('ம', 7, 95) ('খ', 8, 193) ('burton', 9, 8) ('##ப', 10, 111) ### ### ('film', 11, 27) ('2010', 12, 33) ('.', 13, 350) ('intently', 14, 136) ('curious', 15, 147) ### ### ('##部', 16, 131) ('tim', 17, 2814) ('adaptation', 18, 64) ('realizes', 19, 123) ### ### =============================================== v_p ================================================ ### ### ('roth', 8771, 0) ('alice', 1, 1) ('todd', 7886, 2) ('suzanne', 16015, 3) ('##vert', 10304, 4) ### ### ('1951', 12470, 5) ('2016', 3695, 6) ('jennifer', 6740, 7) ('burton', 9, 8) ('glass', 10021, 9) ### ### ('looking', 295, 10) ('wool', 8928, 11) ('wonderland', 0, 12) ('worried', 38, 13) ('bob', 7058, 14) ### ### ('live', 83, 15) ('carroll', 3, 16) ('james', 8866, 17) ('joe', 5893, 18) ('##in', 17288, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('alice', 1, 1) ('wonderland', 0, 12) ('carroll', 3, 16) ('burton', 9, 8) ('absently', 2, 40) ### ### ('film', 11, 27) ('2010', 12, 33) ('disney', 27, 21) ('worried', 38, 13) ('animated', 23, 28) ### ### ('live', 83, 15) ('anyway', 71, 23) ('films', 39, 30) ('ம', 7, 95) ('adaptation', 18, 64) ### ### ('¹', 5, 120) ('##ப', 10, 111) ('looking', 295, 10) ('幸', 6, 127) ('貴', 4, 188) ### ############################################################################################################ [2023-11-25 20:54:26,155][root][INFO] - Epoch: 4: Step: 12401/19530, loss[v]=0.169498, lr=0.000007, acc@1[1]=986.5/1024=0.96337890625, acc@1[2]=1008.0/1024=0.984375 [2023-11-25 20:56:15,105][root][INFO] - Train batch 12500 [2023-11-25 20:56:15,106][root][INFO] - Avg. loss per last 100 batches: 0.132081 [2023-11-25 20:56:15,809][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 3.84 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.9/29522=99.94% | mean: 0.02 | max: 5.87 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: thus is born the idea of the " philosopher - king ", the wise person who accepts the ### ### power thrust upon him by the people who are wise enough to choose a good master. ### ### [P_TEXT]: philosopher king according to plato, a philosopher king is a ruler who possesses both a ### ### love of wisdom, as well as intelligence, reliability, and a willingness to live a simple life. such ### ### are the rulers of his utopian city " kallipolis ". for such a community to ever come into being, " ### ### philosophers [ must ] become kings... or those now called kings [ must ]... genuinely and ### ### adequately philosophize " ( plato, " the republic ", 5. 473d ). plato defined a philosopher firstly ### ### as its eponymous occupation : " wisdom - lover. " he then distinguishes between one who loves true ### ### knowledge ### ### =============================================== v_q ================================================ ### ### ('philosopher', 0, 1) ('king', 1, 6) ('master', 2, 2841) ('intently', 3, 337) ('貴', 4, 335) ### ### ('absently', 5, 164) ('幸', 6, 358) ('ம', 7, 122) ('খ', 8, 344) ('##ப', 9, 219) ### ### ('sparhawk', 10, 4619) ('¹', 11, 168) ('anyway', 12, 23) ('idea', 13, 1518) ('.', 14, 91) ### ### ('philosophy', 15, 24) ('־', 16, 2438) ('concept', 17, 365) ('³', 18, 3342) ('worried', 19, 33) ### ### =============================================== v_p ================================================ ### ### ('reliability', 7974, 0) ('philosopher', 0, 1) ('##3d', 24106, 2) ('##polis', 4401, 3) ### ### ('republic', 5733, 4) ('wisdom', 389, 5) ('king', 1, 6) ('plato', 21, 7) ('eponymous', 9924, 8) ### ### ('philosophers', 535, 9) ('utopia', 1285, 10) ('intelligence', 976, 11) ('simple', 547, 12) ### ### ('occupation', 18489, 13) ('##lli', 28448, 14) ('knowledge', 163, 15) ('lover', 5741, 16) ### ### ('47', 16608, 17) ('willingness', 587, 18) ('must', 280, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('philosopher', 0, 1) ('king', 1, 6) ('plato', 21, 7) ('anyway', 12, 23) ('philosophy', 15, 24) ### ### ('worried', 19, 33) ('kings', 78, 21) ('wisdom', 389, 5) ('knowledge', 163, 15) ('.', 14, 91) ### ### ('ம', 7, 122) ('absently', 5, 164) ('philosophers', 535, 9) ('¹', 11, 168) ('must', 280, 19) ### ### ('simple', 547, 12) ('##ப', 9, 219) ('intently', 3, 337) ('貴', 4, 335) ('##部', 20, 128) ### ############################################################################################################ [2023-11-25 20:56:16,313][root][INFO] - Epoch: 4: Step: 12501/19530, loss[v]=0.074150, lr=0.000007, acc@1[1]=1000.5/1024=0.97705078125, acc@1[2]=1017.5/1024=0.99365234375 [2023-11-25 20:58:04,909][root][INFO] - Train batch 12600 [2023-11-25 20:58:04,909][root][INFO] - Avg. loss per last 100 batches: 0.132618 [2023-11-25 20:58:05,617][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 3.77 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.4/29522=99.94% | mean: 0.02 | max: 5.56 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.2/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the house is composed of representatives who sit in congressional districts allocated to ### ### each state on a basis of population as measured by the u. s. census, with each district ### ### [P_TEXT]: 2014 united states house of representatives elections in west virginia the 2014 united ### ### states house of representatives election in west virginia was held on november 4, 2014, to elect ### ### the three u. s. representatives from west virginia, one from each of the state's three ### ### congressional districts. republicans won control of every congressional district in west virginia ### ### for the first time since 1924. representatives are elected for two - year terms. those elected ### ### served in the 114th congress from january 2015 until january 2017. results of the 2014 united ### ### states house of representatives elections in west virginia by district : republican david mckinley, ### ### who had represented west virginia's 1st congressional district since 2011, was re - elected in ### ### 2012. while ### ### =============================================== v_q ================================================ ### ### ('congressional', 0, 14) ('貴', 1, 137) ('absently', 2, 76) ('intently', 3, 106) ('.', 4, 315) ### ### ('¹', 5, 166) ('ம', 6, 75) ('house', 7, 16) ('খ', 8, 125) ('幸', 9, 159) ('allocated', 10, 2640) ### ### ('##ப', 11, 84) ('districts', 12, 8) ('statehood', 13, 1799) ('states', 14, 60) ('³', 15, 997) ### ### ('anyway', 16, 13) ('representatives', 17, 32) ('worried', 18, 15) ('##部', 19, 102) ### ### =============================================== v_p ================================================ ### ### ('mckinley', 1723, 0) ('114', 17521, 1) ('2014', 8747, 2) ('virginia', 89, 3) ('1st', 6539, 4) ### ### ('reelected', 17207, 5) ('republicans', 3918, 6) ('1924', 11241, 7) ('districts', 12, 8) ### ### ('littered', 1893, 9) ('david', 18842, 10) ('congress', 45, 11) ('district', 25, 12) ### ### ('anyway', 16, 13) ('congressional', 0, 14) ('worried', 18, 15) ('house', 7, 16) ('every', 67, 17) ### ### ('2011', 7366, 18) ('reelection', 810, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('congressional', 0, 14) ('house', 7, 16) ('districts', 12, 8) ('anyway', 16, 13) ### ### ('worried', 18, 15) ('district', 25, 12) ('each', 20, 20) ('representatives', 17, 32) ### ### ('congress', 45, 11) ('virginia', 89, 3) ('absently', 2, 76) ('ம', 6, 75) ('every', 67, 17) ### ### ('intently', 3, 106) ('貴', 1, 137) ('##ப', 11, 84) ('states', 14, 60) ('republican', 78, 23) ### ### ('খ', 8, 125) ('¹', 5, 166) ### ############################################################################################################ [2023-11-25 20:58:06,124][root][INFO] - Epoch: 4: Step: 12601/19530, loss[v]=0.159969, lr=0.000007, acc@1[1]=984.5/1024=0.96142578125, acc@1[2]=1009.0/1024=0.9853515625 [2023-11-25 20:59:55,263][root][INFO] - Train batch 12700 [2023-11-25 20:59:55,264][root][INFO] - Avg. loss per last 100 batches: 0.131286 [2023-11-25 20:59:55,980][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.2/29522=99.95% | mean: 0.02 | max: 5.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.4/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: by 2006, thanks to the efforts of former supremes singer mary wilson, legislation had ### ### been launched in 33 states via the truth in music act to prevent performers from using the name of ### ### a group that didn't have at least one original member, causing the groups who marshak had hired as ### ### marvelettes to bill themselves as " tribute to the ### ### [P_TEXT]: mickey's monkey ( song ) smokey then requested that lamont record it for the miracles, at ### ### the time motown's top group, to which lamont agreed. recorded in the motown studios with an ### ### intended " live party " feel, the song has smokey start with the now - famous line : " alright... ### ### is everybody " ready? " " followed by the people saying “ yeah! ”. in addition to the ### ### miracles'contribution, " mickey's monkey " also featured background vocals by mary wilson of the ### ### supremes, famed detroit dee jay " jockey jack " gibson, martha & the vandellas, and members of the ### ### temptations and the marvelettes ### ### =============================================== v_q ================================================ ### ### ('truth', 0, 7806) ('##ettes', 1, 7) ('mary', 2, 0) ('supreme', 3, 10) ('wilson', 4, 3) ### ### ('marvel', 5, 14) ('¹', 6, 140) ('absently', 7, 89) ('##ak', 8, 23692) ('intently', 9, 110) ### ### ('marsh', 10, 16135) ('খ', 11, 125) ('³', 12, 3067) ('幸', 13, 144) ('貴', 14, 146) ('act', 15, 4510) ### ### ('##ப', 16, 108) ('ம', 17, 80) ('！', 18, 248) ('legislation', 19, 13208) ### ### =============================================== v_p ================================================ ### ### ('mary', 2, 0) ('mickey', 7555, 1) ('monkey', 6622, 2) ('wilson', 4, 3) ('jockey', 929, 4) ### ### ('gibson', 4771, 5) ('dee', 866, 6) ('##ettes', 1, 7) ('alright', 4602, 8) ('everybody', 2618, 9) ### ### ('supreme', 3, 10) ('jay', 1861, 11) ('detroit', 410, 12) ('smoke', 1182, 13) ('marvel', 5, 14) ### ### ('martha', 1991, 15) ('yeah', 342, 16) ('motown', 32, 17) ('anyway', 118, 18) ('jack', 10649, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('mary', 2, 0) ('##ettes', 1, 7) ('wilson', 4, 3) ('supreme', 3, 10) ('marvel', 5, 14) ### ### ('motown', 32, 17) ('group', 26, 29) ('groups', 23, 30) ('absently', 7, 89) ('temptation', 71, 21) ### ### ('anyway', 118, 18) ('¹', 6, 140) ('intently', 9, 110) ('detroit', 410, 12) ('dee', 866, 6) ### ### ('worried', 187, 23) ('yeah', 342, 16) ('jockey', 929, 4) ('খ', 11, 125) ('ம', 17, 80) ### ############################################################################################################ [2023-11-25 20:59:56,487][root][INFO] - Epoch: 4: Step: 12701/19530, loss[v]=0.191505, lr=0.000007, acc@1[1]=981.0/1024=0.9580078125, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-25 21:01:45,323][root][INFO] - Train batch 12800 [2023-11-25 21:01:45,323][root][INFO] - Avg. loss per last 100 batches: 0.136983 [2023-11-25 21:01:46,018][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 3.75 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.6/29522=99.93% | mean: 0.02 | max: 5.59 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.3/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: graphs of degree at most 2 are easy to classify : a 0 - regular graph consists of ### ### disconnected vertices, a 1 - regular graph consists of disconnected edges, and a 2 - regular graph ### ### consists of a disjoint union of cycle s and infinite chains. ### ### [P_TEXT]: oberwolfach problem graph theory, the pairs of people sitting next to each other at a ### ### single meal can be represented as a disjoint union of cycle graph s formula _ 4 of the specified ### ### lengths, with one cycle for each of the dining tables. this union of cycles is a 2 - regular graph, ### ### and every 2 - regular graph has this form. if formula _ 5 is this 2 - regular graph and has formula ### ### _ 6 vertices, the question is whether the complete graph formula _ 7 can be represented as an edge ### ### - disjoint union of copies of formula _ 5. in order for a solution to exist, the total ### ### =============================================== v_q ================================================ ### ### ('graph', 0, 3) ('regular', 1, 5) ('幸', 2, 152) ('absently', 3, 122) ('貴', 4, 208) ('খ', 5, 157) ### ### ('##ப', 6, 118) ('ம', 7, 95) ('union', 8, 6) ('cycle', 9, 2) ('intently', 10, 263) ('¹', 11, 191) ### ### ('.', 12, 2056) ('##部', 13, 107) ('chains', 14, 17618) ('##jo', 15, 22) ('2', 16, 27) ### ### ('vertices', 17, 19) ('degree', 18, 1702) ('is', 19, 148) ### ### =============================================== v_p ================================================ ### ### ('meal', 18249, 0) ('meals', 28851, 1) ('cycle', 9, 2) ('graph', 0, 3) ('dining', 11060, 4) ### ### ('regular', 1, 5) ('union', 8, 6) ('obe', 23635, 7) ('problem', 120, 8) ('tables', 2720, 9) ### ### ('complete', 233, 10) ('vertex', 104, 11) ('cycles', 2654, 12) ('unions', 272, 13) ### ### ('anyway', 33, 14) ('leaned', 529, 15) ('##lf', 26433, 16) ('##wo', 14495, 17) ('graphs', 111, 18) ### ### ('vertices', 17, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('graph', 0, 3) ('regular', 1, 5) ('cycle', 9, 2) ('union', 8, 6) ('vertices', 17, 19) ### ### ('##jo', 15, 22) ('2', 16, 27) ('anyway', 33, 14) ('##int', 29, 20) ('worried', 21, 36) ### ### ('problem', 120, 8) ('vertex', 104, 11) ('ம', 7, 95) ('absently', 3, 122) ('幸', 2, 152) ### ### ('##ப', 6, 118) ('graphs', 111, 18) ('খ', 5, 157) ('complete', 233, 10) ('##部', 13, 107) ### ############################################################################################################ [2023-11-25 21:01:46,522][root][INFO] - Epoch: 4: Step: 12801/19530, loss[v]=0.215781, lr=0.000007, acc@1[1]=973.5/1024=0.95068359375, acc@1[2]=997.5/1024=0.97412109375 [2023-11-25 21:03:34,995][root][INFO] - Train batch 12900 [2023-11-25 21:03:34,996][root][INFO] - Avg. loss per last 100 batches: 0.133487 [2023-11-25 21:03:35,700][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29505.3/29522=99.94% | mean: 0.02 | max: 5.71 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: inky bloaters is the third solo album by danielle dax, an english experimental musician ### ### and former member of the lemon kittens. ### ### [P_TEXT]: falcon stuart 1980. stuart and ant parted company in 1981. during the 1980s, stuart ### ### remained active in the music business, managing and touring globally with the new romantic band ### ### classix nouveaux, formed by former x - ray spex members jak airport and b. p. hurding, as well as ### ### ska / pop act amazulu, and swedish rock pop band trance dance, while also setting up the ### ### independent label awesome records, distributed by rough trade and the cartel. danielle dax released ### ### the acclaimed albums " jesus egg that wept ", " inky bloaters " and " cat - house " on awesome ### ### before signing to sire records. stuart remained involved with ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 107) ('dax', 1, 1) ('danielle', 2, 9) ('##lo', 3, 62) ('##ater', 4, 5) ('ink', 5, 26) ### ### ('³', 6, 1316) ('″', 7, 42) ('kitten', 8, 1204) ('幸', 9, 144) ('lemon', 10, 9770) ### ### ('intently', 11, 155) ('ம', 12, 104) ('খ', 13, 161) ('album', 14, 45) ('貴', 15, 184) ### ### ('##ப', 16, 113) ('absently', 17, 132) ('hissed', 18, 54) ('b', 19, 94) ### ### =============================================== v_p ================================================ ### ### ('cartel', 4542, 0) ('dax', 1, 1) ('falcon', 12729, 2) ('awesome', 124, 3) ('stuart', 3285, 4) ### ### ('##ater', 4, 5) ('nouveau', 2014, 6) ('trance', 99, 7) ('rough', 9693, 8) ('danielle', 2, 9) ### ### ('airport', 15257, 10) ('ama', 13947, 11) ('##rdi', 25379, 12) ('##ex', 9994, 13) ### ### ('romantic', 5795, 14) ('wept', 24365, 15) ('##zu', 25253, 16) ('sire', 3345, 17) ('egg', 7136, 18) ### ### ('cat', 1575, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('dax', 1, 1) ('danielle', 2, 9) ('##ater', 4, 5) ('ink', 5, 26) ('″', 7, 42) ('##lo', 3, 62) ### ### ('¹', 0, 107) ('album', 14, 45) ('trance', 99, 7) ('awesome', 124, 3) ('hissed', 18, 54) ### ### ('ம', 12, 104) ('band', 24, 40) ('幸', 9, 144) ('acts', 40, 41) ('##ப', 16, 113) ### ### ('intently', 11, 155) ('bands', 82, 37) ('absently', 17, 132) ('albums', 25, 61) ### ############################################################################################################ [2023-11-25 21:03:36,206][root][INFO] - Epoch: 4: Step: 12901/19530, loss[v]=0.205087, lr=0.000007, acc@1[1]=976.0/1024=0.953125, acc@1[2]=1002.0/1024=0.978515625 [2023-11-25 21:05:25,299][root][INFO] - Train batch 13000 [2023-11-25 21:05:25,300][root][INFO] - Avg. loss per last 100 batches: 0.134694 [2023-11-25 21:05:26,013][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.0/29522=99.95% | mean: 0.03 | max: 5.92 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: his 1924 – 25 team was retroactively named the national champion by the helms athletic ### ### foundation and the premo - porretta power poll. ### ### [P_TEXT]: 1924 – 25 princeton tigers men's basketball team the 1924 – 25 princeton tigers men's ### ### basketball team represented princeton university in intercollegiate college basketball during the ### ### 1924 – 25 ncaa men's basketball season. the head coach was albert wittmer and the team captain was ### ### stephen cleaves. the team played its home games in the university gymnasium on the university ### ### campus in princeton, new jersey. the team was the winner of the eastern intercollegiate basketball ### ### league ( eibl ) and is considered the retroactive national champion by the helms athletic ### ### foundation and the premo - porretta power poll ### ### =============================================== v_q ================================================ ### ### ('1924', 0, 4) ('25', 1, 16) ('absently', 2, 93) ('##ret', 3, 7) ('power', 4, 25) ('貴', 5, 185) ### ### ('intently', 6, 125) ('খ', 7, 196) ('幸', 8, 204) ('football', 9, 1072) ('prem', 10, 6) ### ### ('1925', 11, 33) ('##ப', 12, 167) ('ம', 13, 138) ('##¨', 14, 3156) ('¹', 15, 427) ('poll', 16, 9) ### ### ('bellowed', 17, 583) ('.', 18, 954) ('helm', 19, 2) ### ### =============================================== v_p ================================================ ### ### ('gymnasium', 1419, 0) ('princeton', 368, 1) ('helm', 19, 2) ('##tm', 25856, 3) ('1924', 0, 4) ### ### ('##ib', 27543, 5) ('prem', 10, 6) ('##ret', 3, 7) ('intercollegiate', 864, 8) ('poll', 16, 9) ### ### ('worried', 71, 10) ('wit', 16915, 11) ('##ves', 13457, 12) ('basketball', 22, 13) ### ### ('foundation', 41, 14) ('stephen', 9403, 15) ('25', 1, 16) ('albert', 2098, 17) ('por', 24, 18) ### ### ('anyway', 26, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('1924', 0, 4) ('25', 1, 16) ('##ret', 3, 7) ('prem', 10, 6) ('helm', 19, 2) ('poll', 16, 9) ### ### ('power', 4, 25) ('basketball', 22, 13) ('por', 24, 18) ('anyway', 26, 19) ('retro', 25, 20) ### ### ('1925', 11, 33) ('foundation', 41, 14) ('worried', 71, 10) ('princeton', 368, 1) ### ### ('nationally', 47, 26) ('littered', 56, 24) ('athletic', 35, 31) ('absently', 2, 93) ### ### ('1923', 20, 51) ### ############################################################################################################ [2023-11-25 21:05:26,518][root][INFO] - Epoch: 4: Step: 13001/19530, loss[v]=0.132725, lr=0.000007, acc@1[1]=993.5/1024=0.97021484375, acc@1[2]=1008.5/1024=0.98486328125 [2023-11-25 21:07:15,164][root][INFO] - Train batch 13100 [2023-11-25 21:07:15,165][root][INFO] - Avg. loss per last 100 batches: 0.123767 [2023-11-25 21:07:15,867][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29507.1/29522=99.95% | mean: 0.03 | max: 5.84 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: when carl j. burckhardt became high commissioner in february 1937, both poles and germans ### ### openly welcomed his withdrawal, and polish minister of foreign affairs jozef beck notified him not ### ### to " count on the support of the polish state " in the case of difficulties ### ### [P_TEXT]: jozef beck its main trade route, effectively making the polish economy dependent on ### ### germany. he also rejected demands for an extraterritorial rail and highway corridor that was to run ### ### to east prussia and the free city of danzig, in exchange for vague promises regarding trade and ### ### annexation of territories inhabited by ukrainians and belorussians in the soviet union after a ### ### future war. while hitler had been planning to annex polish territory for several years, he finally ### ### decided to go ahead with his plans for war by early september 1939. beck was surprised when ### ### britain, looking for a ### ### =============================================== v_q ================================================ ### ### ('beck', 0, 0) ('¹', 1, 138) ('jozef', 2, 1) ('carl', 3, 17287) ('foreign', 4, 10428) ### ### ('##khar', 5, 27438) ('##rc', 6, 23774) ('##ʂ', 7, 1476) ('absently', 8, 248) ('##dt', 9, 24561) ### ### ('幸', 10, 294) ('খ', 11, 183) ('貴', 12, 290) ('bu', 13, 22513) ('intently', 14, 357) ### ### ('##ப', 15, 217) ('withdrawal', 16, 2741) ('commissioner', 17, 11310) ('³', 18, 1968) ### ### ('affairs', 19, 10442) ### ### =============================================== v_p ================================================ ### ### ('beck', 0, 0) ('jozef', 2, 1) ('##rito', 18152, 2) ('corridor', 5887, 3) ('ukrainian', 1303, 4) ### ### ('danzig', 176, 5) ('##sian', 14828, 6) ('##orus', 3644, 7) ('hitler', 40, 8) ('economy', 1863, 9) ### ### ('bel', 9169, 10) ('prussia', 72, 11) ('extra', 13857, 12) ('vague', 1597, 13) ### ### ('demands', 6652, 14) ('highway', 14137, 15) ('leaned', 2064, 16) ('corridors', 18382, 17) ### ### ('##rial', 9618, 18) ('polish', 21, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('beck', 0, 0) ('jozef', 2, 1) ('polish', 21, 19) ('poland', 22, 37) ('hitler', 40, 8) ### ### ('prussia', 72, 11) ('¹', 1, 138) ('danzig', 176, 5) ('anyway', 45, 30) ('1939', 32, 43) ### ### ('german', 31, 45) ('soviet', 54, 33) ('poles', 34, 51) ('খ', 11, 183) ('ம', 20, 143) ### ### ('absently', 8, 248) ('germans', 99, 41) ('##ப', 15, 217) ('幸', 10, 294) ('貴', 12, 290) ### ############################################################################################################ [2023-11-25 21:07:16,372][root][INFO] - Epoch: 4: Step: 13101/19530, loss[v]=0.084289, lr=0.000007, acc@1[1]=995.5/1024=0.97216796875, acc@1[2]=1010.0/1024=0.986328125 [2023-11-25 21:09:05,109][root][INFO] - Train batch 13200 [2023-11-25 21:09:05,109][root][INFO] - Avg. loss per last 100 batches: 0.129198 [2023-11-25 21:09:05,823][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.6/29522=99.94% | mean: 0.02 | max: 5.87 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the volume channeled through kiva's peer - to - peer platform is about $ 100 million as ### ### of november 2009 ( kiva facilitates approximately $ 5m in loans each month ). ### ### [P_TEXT]: kiva ( organization ) initial stage, flannery and jackley founded kiva as a non - profit. ### ### in 2006, notable entrepreneurs and businessmen joined kiva's staff, including premal shah from ### ### paypal and reid hoffman, ceo and founder of linkedin. shortly after its first anniversary in ### ### october 2006, kiva reached $ 1 million in facilitated loans and acquired its twentieth field ### ### partner. to the present day, kiva has continued to grow and expand its field partners while ### ### acquiring support from the media and the public. kiva works with more than 300 microfinance ### ### institutions, social impact businesses, schools and non - profit ### ### =============================================== v_q ================================================ ### ### ('##va', 0, 5) ('ki', 1, 4) ('peer', 2, 8695) ('¹', 3, 272) ('intently', 4, 309) ('幸', 5, 176) ### ### ('貴', 6, 276) ('খ', 7, 173) ('ம', 8, 106) ('absently', 9, 77) ('platform', 10, 333) ### ### ('##ப', 11, 118) ('³', 12, 2405) ('channel', 13, 5730) ('anyway', 14, 46) ('encourage', 15, 300) ### ### ('encourages', 16, 1162) ('$', 17, 96) ('internet', 18, 864) ('provider', 19, 440) ### ### =============================================== v_p ================================================ ### ### ('prem', 8719, 0) ('facilitated', 79, 1) ('reid', 8777, 2) ('impact', 874, 3) ('ki', 1, 4) ### ### ('##va', 0, 5) ('hoffman', 1217, 6) ('##anne', 24185, 7) ('shah', 16108, 8) ('loans', 138, 9) ### ### ('linked', 651, 10) ('twentieth', 9176, 11) ('impacts', 20903, 12) ('jack', 11082, 13) ### ### ('field', 14170, 14) ('##ley', 26551, 15) ('##fin', 23314, 16) ('ceo', 487, 17) ('pay', 990, 18) ### ### ('##ance', 9894, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##va', 0, 5) ('ki', 1, 4) ('facilitated', 79, 1) ('anyway', 14, 46) ('loans', 138, 9) ### ### ('absently', 9, 77) ('ம', 8, 106) ('social', 91, 28) ('##ப', 11, 118) ('impact', 874, 3) ### ### ('幸', 5, 176) ('loan', 81, 39) ('organization', 178, 23) ('linked', 651, 10) ('খ', 7, 173) ### ### ('¹', 3, 272) ('hoffman', 1217, 6) ('$', 17, 96) ('ceo', 487, 17) ('media', 111, 44) ### ############################################################################################################ [2023-11-25 21:09:06,329][root][INFO] - Epoch: 4: Step: 13201/19530, loss[v]=0.067724, lr=0.000007, acc@1[1]=1002.0/1024=0.978515625, acc@1[2]=1017.5/1024=0.99365234375 [2023-11-25 21:10:55,512][root][INFO] - Train batch 13300 [2023-11-25 21:10:55,512][root][INFO] - Avg. loss per last 100 batches: 0.119439 [2023-11-25 21:10:56,230][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.2/29522=99.94% | mean: 0.02 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: murphy and two other seals, danny dietz and matthew axelson, were killed in the fighting, ### ### in addition to 16 u. s. special operations soldiers, who were killed when their helicopter was shot ### ### down while attempting to extract the seal team. ### ### [P_TEXT]: erik s. kristensen hindu kush mountains. the seal team was made up of matthew axelson, ### ### danny dietz, marcus luttrell, and michael p. murphy. luttrell was rear medical, axelson the team's ### ### sniper ; dietz was the communications officer and murphy the team leader. the four seals engaged ### ### local taliban forces, they were engaged in an intense gun battle against a force of approximately ### ### 30 – 40 enemy fighters. murphy risked his life to get off an emergency message to his command. of ### ### the four - man team, only luttrell would survive. upon hearing the distress call, an mh - 47 ### ### =============================================== v_q ================================================ ### ### ('murphy', 0, 0) ('seals', 1, 1) ('seal', 2, 11) ('absently', 3, 227) ('axel', 4, 3) ### ### ('danny', 5, 13) ('diet', 6, 5) ('¹', 7, 88) ('intently', 8, 293) ('matthew', 9, 22) ('খ', 10, 191) ### ### ('幸', 11, 160) ('貴', 12, 317) ('##ப', 13, 153) ('helicopter', 14, 3350) ('special', 15, 11296) ### ### ('ம', 16, 121) ('##z', 17, 37) ('³', 18, 2826) ('freaking', 19, 99) ### ### =============================================== v_p ================================================ ### ### ('murphy', 0, 0) ('seals', 1, 1) ('hindu', 17562, 2) ('axel', 4, 3) ('##rell', 11259, 4) ### ### ('diet', 6, 5) ('erik', 8666, 6) ('47', 1112, 7) ('distress', 4393, 8) ('taliban', 1057, 9) ### ### ('kristen', 5932, 10) ('seal', 2, 11) ('marcus', 4046, 12) ('danny', 5, 13) ('hinduism', 23408, 14) ### ### ('communications', 724, 15) ('rear', 6996, 16) ('sniper', 29, 17) ('##sen', 11154, 18) ### ### ('m', 6472, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('murphy', 0, 0) ('seals', 1, 1) ('axel', 4, 3) ('diet', 6, 5) ('seal', 2, 11) ('danny', 5, 13) ### ### ('matthew', 9, 22) ('##z', 17, 37) ('sniper', 29, 17) ('##son', 23, 39) ('¹', 7, 88) ### ### ('team', 36, 23) ('anyway', 50, 31) ('absently', 3, 227) ('officer', 127, 21) ('幸', 11, 160) ### ### ('ம', 16, 121) ('shooting', 26, 69) ('খ', 10, 191) ('hissed', 51, 47) ### ############################################################################################################ [2023-11-25 21:10:56,736][root][INFO] - Epoch: 4: Step: 13301/19530, loss[v]=0.174968, lr=0.000007, acc@1[1]=977.5/1024=0.95458984375, acc@1[2]=1004.0/1024=0.98046875 [2023-11-25 21:12:45,430][root][INFO] - Train batch 13400 [2023-11-25 21:12:45,430][root][INFO] - Avg. loss per last 100 batches: 0.135431 [2023-11-25 21:12:46,139][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.83 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.9/29522=99.94% | mean: 0.02 | max: 5.88 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: established on 20 december 1958, belu regency has its seat ( capital ) in the large town ### ### of atambua. ### ### [P_TEXT]: atambua planted more trees in the place now called " lapangan umum ". the most notable ### ### tree planted there is the banyan tree. in september 1999, more than 250, 000 refugees arrived here ### ### from east timor, after their vote for independence and the following violence. as late as 2002, an ### ### estimated 60, 000 refugees remained in camps. as like belu regency, the television in this town is ### ### limited, which are : the town's transport system relies mainly on minibuses, usually called " bemo ### ### " or " mikrolet ", and " motorcycle taxi " ( ojek ) provide an alternative. there are only four ### ### routes in the town served ### ### =============================================== v_q ================================================ ### ### ('bel', 0, 1) ('##mb', 1, 5) ('regency', 2, 2) ('ata', 3, 3) ('¹', 4, 367) ('##ua', 5, 19) ### ### ('##u', 6, 27) ('indonesia', 7, 70) ('幸', 8, 382) ('³', 9, 2443) ('sulawesi', 10, 419) ### ### ('##ப', 11, 233) ('ம', 12, 234) ('貴', 13, 452) ('intently', 14, 465) ('indonesian', 15, 73) ### ### ('খ', 16, 381) ('papua', 17, 731) ('java', 18, 187) ('timor', 19, 0) ### ### =============================================== v_p ================================================ ### ### ('timor', 19, 0) ('bel', 0, 1) ('regency', 2, 2) ('ata', 3, 3) ('taxi', 12717, 4) ('##mb', 1, 5) ### ### ('##ole', 7622, 6) ('tree', 7197, 7) ('motorcycle', 3020, 8) ('violence', 6673, 9) ### ### ('##yan', 12680, 10) ('camps', 8640, 11) ('transport', 1931, 12) ('independence', 3414, 13) ### ### ('refugees', 4107, 14) ('1999', 5065, 15) ('##kr', 17818, 16) ('lap', 18035, 17) ('mini', 8877, 18) ### ### ('##ua', 5, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('bel', 0, 1) ('regency', 2, 2) ('ata', 3, 3) ('##mb', 1, 5) ('##ua', 5, 19) ('timor', 19, 0) ### ### ('##u', 6, 27) ('indonesia', 7, 70) ('indonesian', 15, 73) ('leaned', 52, 30) ('¹', 4, 367) ### ### ('town', 55, 46) ('anyway', 148, 21) ('##ப', 11, 233) ('ம', 12, 234) ('worried', 84, 43) ### ### ('java', 18, 187) ('幸', 8, 382) ('sulawesi', 10, 419) ('towns', 156, 42) ### ############################################################################################################ [2023-11-25 21:12:46,643][root][INFO] - Epoch: 4: Step: 13401/19530, loss[v]=0.069899, lr=0.000007, acc@1[1]=1001.5/1024=0.97802734375, acc@1[2]=1014.5/1024=0.99072265625 [2023-11-25 21:14:35,521][root][INFO] - Train batch 13500 [2023-11-25 21:14:35,522][root][INFO] - Avg. loss per last 100 batches: 0.118404 [2023-11-25 21:14:36,246][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.8/29522=99.95% | mean: 0.02 | max: 5.81 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: sociologists also use the term in reference to conservative christian beliefs which ### ### include not just literalism but also biblical inerrancy. ### ### [P_TEXT]: basic theology basic theology ( 1986, 1999 ) is a systematic theology book written by ### ### christian author charles caldwell ryrie and published by moody publishers. written for the layman, ### ### the book makes a conscious effort to use simple language and examples, many illustrations, and few ### ### footnotes. " basic theology " is written from a theologically conservative perspective. it assumes ### ### a number of presuppositions, including the inerrancy of scripture, a normal / plain interpretation ### ### of scripture, and the legitimacy of proof texts. ryrie's theology stems from an approach to ### ### scripture, claiming that " when objective ### ### =============================================== v_q ================================================ ### ### ('literal', 0, 251) ('##ran', 1, 5) ('absently', 2, 149) ('##er', 3, 22) ('conservative', 4, 24) ### ### ('intently', 5, 430) ('貴', 6, 487) ('¹', 7, 479) ('幸', 8, 347) ('christian', 9, 32) ('ம', 10, 166) ### ### ('.', 11, 4020) ('খ', 12, 306) ('biblical', 13, 76) ('##ப', 14, 243) ('worried', 15, 52) ### ### ('sparhawk', 16, 491) ('³', 17, 1920) ('christianity', 18, 46) ('！', 19, 3625) ### ### =============================================== v_p ================================================ ### ### ('moody', 6015, 0) ('basic', 5270, 1) ('normal', 2585, 2) ('theology', 169, 3) ### ### ('caldwell', 13372, 4) ('##ran', 1, 5) ('proof', 5823, 6) ('scripture', 510, 7) ('plain', 6068, 8) ### ### ('systematic', 1123, 9) ('ry', 19496, 10) ('legitimacy', 61, 11) ('publishers', 13804, 12) ### ### ('lay', 446, 13) ('interpretation', 250, 14) ('##rie', 25893, 15) ('##notes', 20539, 16) ### ### ('simple', 4737, 17) ('interpretations', 7179, 18) ('littered', 3328, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ran', 1, 5) ('##er', 3, 22) ('conservative', 4, 24) ('christian', 9, 32) ('bible', 21, 26) ### ### ('anyway', 25, 21) ('worried', 15, 52) ('legitimacy', 61, 11) ('christianity', 18, 46) ### ### ('theology', 169, 3) ('biblical', 13, 76) ('absently', 2, 149) ('interpretation', 250, 14) ### ### ('literal', 0, 251) ('scripture', 510, 7) ('lay', 446, 13) ('belief', 20, 90) ('ம', 10, 166) ### ### ('##cy', 121, 30) ('christians', 27, 70) ### ############################################################################################################ [2023-11-25 21:14:36,752][root][INFO] - Epoch: 4: Step: 13501/19530, loss[v]=0.162923, lr=0.000007, acc@1[1]=983.0/1024=0.9599609375, acc@1[2]=1006.5/1024=0.98291015625 [2023-11-25 21:16:25,789][root][INFO] - Train batch 13600 [2023-11-25 21:16:25,790][root][INFO] - Avg. loss per last 100 batches: 0.132334 [2023-11-25 21:16:26,503][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.79 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.6/29522=99.93% | mean: 0.02 | max: 5.57 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.6/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 1897, her sister, duchess sophie in bavaria, died in an accidental fire at the " bazar ### ### de la charite " in paris. ### ### [P_TEXT]: duchess sophie charlotte in bavaria " sopherl " within the family. she was also a sister ### ### of empress elisabeth of austria and queen maria sophie of the two sicilies. after the 1861 marriage ### ### of her older sister, duchess mathilde ludovika in bavaria, to the neapolitan prince luis of the two ### ### sicilies, her parents then looked for a suitable husband for sophie charlotte. she was engaged to ### ### her cousin king ludwig ii of bavaria. their engagement was publicised on 22 january 1867, but after ### ### having repeatedly postponed the wedding date, ludwig finally cancelled it ### ### =============================================== v_q ================================================ ### ### ('sophie', 0, 3) ('¹', 1, 87) ('duchess', 2, 30) ('##ite', 3, 20589) ('absently', 4, 216) ### ### ('1897', 5, 1372) ('char', 6, 21284) ('intently', 7, 111) ('貴', 8, 154) ('bavaria', 9, 8) ### ### ('幸', 10, 134) ('খ', 11, 137) ('.', 12, 236) ('³', 13, 453) ('ம', 14, 88) ('princess', 15, 82) ### ### ('sister', 16, 21) ('##ப', 17, 114) ('empress', 18, 27) ('freud', 19, 8429) ### ### =============================================== v_p ================================================ ### ### ('ludwig', 238, 0) ('luis', 6049, 1) ('elisabeth', 316, 2) ('sophie', 0, 3) ('##pher', 23508, 4) ### ### ('charlotte', 812, 5) ('##vik', 25708, 6) ('1861', 4550, 7) ('bavaria', 9, 8) ('anyway', 96, 9) ### ### ('sic', 10493, 10) ('1867', 2347, 11) ('##ilde', 648, 12) ('austria', 131, 13) ('math', 19971, 14) ### ### ('##ili', 24530, 15) ('austrians', 8767, 16) ('##ap', 28547, 17) ('worried', 120, 18) ### ### ('maria', 186, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('sophie', 0, 3) ('bavaria', 9, 8) ('duchess', 2, 30) ('sister', 16, 21) ('empress', 18, 27) ### ### ('ludwig', 238, 0) ('queen', 26, 31) ('anyway', 96, 9) ('elisabeth', 316, 2) ('¹', 1, 87) ### ### ('austria', 131, 13) ('sisters', 50, 32) ('worried', 120, 18) ('prince', 47, 37) ### ### ('intently', 7, 111) ('princess', 15, 82) ('ம', 14, 88) ('charlotte', 812, 5) ('maria', 186, 19) ### ### ('her', 140, 24) ### ############################################################################################################ [2023-11-25 21:16:27,008][root][INFO] - Epoch: 4: Step: 13601/19530, loss[v]=0.071983, lr=0.000007, acc@1[1]=1005.0/1024=0.9814453125, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 21:18:15,980][root][INFO] - Train batch 13700 [2023-11-25 21:18:15,980][root][INFO] - Avg. loss per last 100 batches: 0.129185 [2023-11-25 21:18:16,698][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.6/29522=99.93% | mean: 0.02 | max: 5.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: they challenged the right of the mississippi democratic party's delegation to participate ### ### in the convention, claiming that the regulars had been illegally elected in a completely segregated ### ### process that violated both party regulations and federal law, and that the regulars had no ### ### intention to support president lyndon b. johnson, the party's candidate, in the november election. ### ### [P_TEXT]: lyndon b. johnson through a gauntlet of interviews to guarantee his absolute loyalty and ### ### having made the decision, he kept the announcement from the press until the last moment to maximize ### ### media speculation and coverage. in preparation for the democratic convention, johnson requested the ### ### fbi send a squad of 30 agents to cover convention activities ; the objective of the squad was to ### ### inform the white house staff of any disruptive activities on the floor. the squad's focus narrowed ### ### upon the mississippi freedom democratic party ( mfdp ) delegation, which sought to displace the ### ### white segregationist delegation regularly selected in the state. the squad's ### ### =============================================== v_q ================================================ ### ### ('mississippi', 0, 3) ('absently', 1, 59) ('lyndon', 2, 1) ('convention', 3, 7) ### ### ('democratic', 4, 16) ('খ', 5, 129) ('regulars', 6, 2912) ('幸', 7, 210) ('johnson', 8, 0) ### ### ('intently', 9, 239) ('b', 10, 10) ('貴', 11, 345) ('ம', 12, 90) ('¹', 13, 150) ### ### ('segregation', 14, 2) ('segregated', 15, 45) ('##ப', 16, 132) ('exclaimed', 17, 287) ### ### ('ד', 18, 305) ('angrily', 19, 302) ### ### =============================================== v_p ================================================ ### ### ('johnson', 8, 0) ('lyndon', 2, 1) ('segregation', 14, 2) ('mississippi', 0, 3) ### ### ('freedom', 1506, 4) ('squad', 12096, 5) ('disrupt', 1859, 6) ('convention', 3, 7) ('fbi', 437, 8) ### ### ('##dp', 6612, 9) ('b', 10, 10) ('gaunt', 19764, 11) ('floor', 1054, 12) ('loyalty', 764, 13) ### ### ('house', 1522, 14) ('anyway', 40, 15) ('democratic', 4, 16) ('speculation', 894, 17) ### ### ('squads', 19664, 18) ('white', 1144, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('mississippi', 0, 3) ('lyndon', 2, 1) ('johnson', 8, 0) ('segregation', 14, 2) ### ### ('convention', 3, 7) ('b', 10, 10) ('democratic', 4, 16) ('party', 22, 28) ('absently', 1, 59) ### ### ('segregated', 15, 45) ('delegation', 23, 39) ('democrats', 28, 27) ('anyway', 40, 15) ### ### ('খ', 5, 129) ('ம', 12, 90) ('.', 20, 75) ('stared', 33, 50) ('fbi', 437, 8) ('##ப', 16, 132) ### ### ('¹', 13, 150) ### ############################################################################################################ [2023-11-25 21:18:17,203][root][INFO] - Epoch: 4: Step: 13701/19530, loss[v]=0.077256, lr=0.000006, acc@1[1]=1000.0/1024=0.9765625, acc@1[2]=1019.5/1024=0.99560546875 [2023-11-25 21:20:06,642][root][INFO] - Train batch 13800 [2023-11-25 21:20:06,642][root][INFO] - Avg. loss per last 100 batches: 0.129455 [2023-11-25 21:20:07,405][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.7/29522=99.94% | mean: 0.02 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the range stretches from the far eastern czech republic ( 3 % ) in the northwest through ### ### slovakia ( 17 % ), poland ( 10 % ), hungary ( 4 % ), ukraine ( 10 % ), romania ( 50 % ) to serbia ( ### ### 5 % ) in the south. ### ### [P_TEXT]: slovakia their territories east of the carpathian mountains, invaded the carpathian basin ### ### and started to occupy the territory gradually around 896. their armies'advance may have been ### ### promoted by continuous wars among the countries of the region whose rulers still hired them ### ### occasionally to intervene in their struggles. it is not known what happened with both mojmir ii and ### ### svatopluk ii because they are not mentioned in written sources after 906. in three battles ( 4 – 5 ### ### july and 9 august 907 ) near bratislava, the magyars routed bavaria n armies. some historians put ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 101) ('range', 1, 851) ('.', 2, 73) ('খ', 3, 237) ('幸', 4, 255) ('romania', 5, 848) ### ### ('ம', 6, 176) ('貴', 7, 319) ('mountains', 8, 82) ('slovakia', 9, 0) ('hungary', 10, 49) ### ### ('##ப', 11, 268) ('poland', 12, 571) ('serbia', 13, 254) ('intently', 14, 442) ('¹', 15, 90) ### ### ('ukraine', 16, 76) ('curious', 17, 5) ('##部', 18, 444) ('czech', 19, 1031) ### ### =============================================== v_p ================================================ ### ### ('slovakia', 9, 0) ('##jm', 14429, 1) ('bavaria', 1551, 2) ('intervened', 27790, 3) ### ### ('89', 18732, 4) ('curious', 17, 5) ('ii', 11779, 6) ('90', 17417, 7) ('##ir', 12834, 8) ### ### ('##pl', 20766, 9) ('##ato', 24874, 10) ('invaded', 27689, 11) ('anyway', 33, 12) ### ### ('invasions', 19981, 13) ('wars', 12216, 14) ('##uk', 18397, 15) ('conquered', 17172, 16) ### ### ('slovak', 84, 17) ('skirmish', 10509, 18) ('sv', 5213, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('slovakia', 9, 0) ('curious', 17, 5) ('hungary', 10, 49) ('anyway', 33, 12) ('.', 2, 73) ### ### ('absently', 0, 101) ('carpathian', 21, 42) ('mountains', 8, 82) ('ukraine', 16, 76) ('¹', 15, 90) ### ### ('slovak', 84, 17) ('ம', 6, 176) ('খ', 3, 237) ('stared', 29, 75) ('bratislava', 100, 22) ### ### ('幸', 4, 255) ('czechoslovakia', 104, 29) ('worried', 69, 51) ('##ப', 11, 268) ('serbia', 13, 254) ### ############################################################################################################ [2023-11-25 21:20:07,918][root][INFO] - Epoch: 4: Step: 13801/19530, loss[v]=0.072898, lr=0.000006, acc@1[1]=995.0/1024=0.9716796875, acc@1[2]=1015.0/1024=0.9912109375 [2023-11-25 21:21:57,313][root][INFO] - Train batch 13900 [2023-11-25 21:21:57,314][root][INFO] - Avg. loss per last 100 batches: 0.136731 [2023-11-25 21:21:58,028][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.0/29522=99.95% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the city is located between the willamette and tualatin rivers ; it includes the former ### ### townsites / developments of willamette, bolton, multnomah city, sunset city, and west oregon city. ### ### [P_TEXT]: willamette river oregon route 34 ( corvallis – lebanon highway ) over the river upstream ### ### of rm 131 ( rk 211 ) in corvallis. the machinery to operate the swing span was removed in the ### ### 1950s. the oregon city bridge, built in 1922, replaced a suspension span constructed at the site in ### ### 1888. it carries oregon route 43 over the river at about rm 26 ( rk 42 ) between oregon city and ### ### west linn. the ross island bridge carries u. s. route 26 ( mount hood highway ) over the river at ### ### rm ### ### =============================================== v_q ================================================ ### ### ('bolton', 0, 3862) ('oregon', 1, 6) ('absently', 2, 97) ('cradled', 3, 252) ('sunset', 4, 4842) ### ### ('খ', 5, 206) ('幸', 6, 172) ('.', 7, 1635) ('ம', 8, 107) ('city', 9, 29) ('##ப', 10, 163) ### ### ('leaned', 11, 57) ('##部', 12, 159) ('will', 13, 15) ('貴', 14, 213) ('##ala', 15, 25785) ### ### ('intently', 16, 116) ('##lt', 17, 19922) ('river', 18, 22) ('snuggled', 19, 121) ### ### =============================================== v_p ================================================ ### ### ('##rval', 17040, 0) ('ross', 5019, 1) ('43', 4341, 2) ('rm', 12344, 3) ('lin', 12471, 4) ### ### ('##ame', 30, 5) ('oregon', 1, 6) ('lebanon', 16340, 7) ('34', 5353, 8) ('211', 10954, 9) ### ### ('26', 5261, 10) ('suspension', 20954, 11) ('hood', 2037, 12) ('swing', 12128, 13) ### ### ('131', 9364, 14) ('will', 13, 15) ('bridge', 3501, 16) ('anyway', 103, 17) ('island', 4826, 18) ### ### ('highway', 401, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('oregon', 1, 6) ('will', 13, 15) ('##ame', 30, 5) ('city', 9, 29) ('river', 18, 22) ### ### ('absently', 2, 97) ('leaned', 11, 57) ('##tte', 36, 24) ('worried', 35, 31) ('anyway', 103, 17) ### ### ('ம', 8, 107) ('stared', 21, 68) ('幸', 6, 172) ('west', 27, 58) ('residents', 87, 28) ('খ', 5, 206) ### ### ('##ப', 10, 163) ('intently', 16, 116) ('cradled', 3, 252) ('##部', 12, 159) ### ############################################################################################################ [2023-11-25 21:21:58,532][root][INFO] - Epoch: 4: Step: 13901/19530, loss[v]=0.063499, lr=0.000006, acc@1[1]=1007.0/1024=0.9833984375, acc@1[2]=1020.0/1024=0.99609375 [2023-11-25 21:23:47,763][root][INFO] - Train batch 14000 [2023-11-25 21:23:47,763][root][INFO] - Avg. loss per last 100 batches: 0.123466 [2023-11-25 21:23:48,474][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29505.0/29522=99.94% | mean: 0.02 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the greenbrier military school was a boys - only, private, military, boarding high school ### ### and junior college, located in lewisburg, west virginia. ### ### [P_TEXT]: homer a. holt homer adams holt ( march 1, 1898january 16, 1975 ) was a west virginia ### ### lawyer and politician who served as that state's 20th governor from 1937 to 1941. born in ### ### lewisburg, west virginia, he attended the greenbrier military school there and then went on to ### ### graduate from washington and lee university in 1918, where he was a member of the phi kappa psi ### ### fraternity. after serving in the army during world war i, he returned to washington and lee in 1920 ### ### and studied law, receiving his degree in 1923. in 1924, ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 1186) ('military', 1, 11) ('##bri', 2, 3) ('green', 3, 14) ('virginia', 4, 9) ### ### ('³', 5, 4140) ('lewis', 6, 2) ('##burg', 7, 4) ('school', 8, 26) ('幸', 9, 334) ### ### ('exclaimed', 10, 1147) ('খ', 11, 269) ('intently', 12, 261) ('##ப', 13, 266) ('seminary', 14, 123) ### ### ('ம', 15, 215) ('貴', 16, 399) ('curious', 17, 2291) ('west', 18, 35) ('##ᵉ', 19, 1389) ### ### =============================================== v_p ================================================ ### ### ('holt', 9058, 0) ('homer', 19077, 1) ('lewis', 6, 2) ('##bri', 2, 3) ('##burg', 7, 4) ### ### ('psi', 14420, 5) ('lee', 1236, 6) ('kappa', 4364, 7) ('##uary', 3506, 8) ('virginia', 4, 9) ### ### ('governor', 1400, 10) ('military', 1, 11) ('20th', 1547, 12) ('worried', 2771, 13) ### ### ('green', 3, 14) ('washington', 248, 15) ('phi', 1916, 16) ('fraternity', 229, 17) ### ### ('1898', 983, 18) ('governors', 9243, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##bri', 2, 3) ('lewis', 6, 2) ('##burg', 7, 4) ('military', 1, 11) ('virginia', 4, 9) ### ### ('green', 3, 14) ('school', 8, 26) ('west', 18, 35) ('schooling', 57, 24) ('soldiers', 29, 45) ### ### ('university', 60, 29) ('schoolhouse', 39, 42) ('seminary', 14, 123) ('washington', 248, 15) ### ### ('anyway', 147, 22) ('fraternity', 229, 17) ('খ', 11, 269) ('intently', 12, 261) ('army', 35, 62) ### ### ('幸', 9, 334) ### ############################################################################################################ [2023-11-25 21:23:48,979][root][INFO] - Epoch: 4: Step: 14001/19530, loss[v]=0.188091, lr=0.000006, acc@1[1]=975.0/1024=0.9521484375, acc@1[2]=1003.5/1024=0.97998046875 [2023-11-25 21:25:38,656][root][INFO] - Train batch 14100 [2023-11-25 21:25:38,657][root][INFO] - Avg. loss per last 100 batches: 0.123351 [2023-11-25 21:25:39,383][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 4.06 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29505.3/29522=99.94% | mean: 0.02 | max: 6.00 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: oxon hill is located in prince george's county along maryland route 210 ( indian head ### ### highway ) and maryland route 414 ### ### [P_TEXT]: maryland route 210 maryland route 210 ( md 210 ) is a state highway in the u. s. state of ### ### maryland. known as indian head highway, the highway runs from potomac avenue in indian head north ### ### to the district of columbia boundary in forest heights, where the highway continues into washington ### ### as south capitol street. md 210 is a four - to six - lane divided highway that connects washington ### ### with the suburban communities of oxon hill, fort washington, and accokeek in southwestern prince ### ### george's county, and bryans road and indian head in northwestern charles county. the highway ### ### =============================================== v_q ================================================ ### ### ('ox', 0, 1) ('¹', 1, 106) ('maryland', 2, 10) ('hill', 3, 21) ('³', 4, 315) ('210', 5, 0) ### ### ('prince', 6, 20) ('খ', 7, 149) ('##on', 8, 27) ('幸', 9, 139) ('intently', 10, 76) ### ### ('##4', 11, 23605) ('george', 12, 23) ('head', 13, 9) ('##ப', 14, 124) ('baltimore', 15, 33) ### ### ('41', 16, 21439) ('townspeople', 17, 742) ('ம', 18, 101) ('貴', 19, 128) ### ### =============================================== v_p ================================================ ### ### ('210', 5, 0) ('ox', 0, 1) ('bryan', 18720, 2) ('fort', 277, 3) ('capitol', 248, 4) ### ### ('worried', 717, 5) ('md', 1401, 6) ('forest', 3742, 7) ('##oke', 10211, 8) ('head', 13, 9) ### ### ('maryland', 2, 10) ('washington', 31, 11) ('indian', 21, 12) ('acc', 8037, 13) ('##ek', 25729, 14) ### ### ('heights', 421, 15) ('potomac', 25, 16) ('anyway', 137, 17) ('charles', 8262, 18) ### ### ('avenue', 370, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ox', 0, 1) ('210', 5, 0) ('maryland', 2, 10) ('head', 13, 9) ('hill', 3, 21) ('prince', 6, 20) ### ### ('indian', 21, 12) ('george', 12, 23) ('##on', 8, 27) ('washington', 31, 11) ('potomac', 25, 16) ### ### ('baltimore', 15, 33) ('¹', 1, 106) ('route', 30, 28) ('county', 53, 22) ('intently', 10, 76) ### ### ('capitol', 248, 4) ('fort', 277, 3) ('anyway', 137, 17) ('counties', 59, 26) ### ############################################################################################################ [2023-11-25 21:25:39,890][root][INFO] - Epoch: 4: Step: 14101/19530, loss[v]=0.100109, lr=0.000006, acc@1[1]=995.0/1024=0.9716796875, acc@1[2]=1013.5/1024=0.98974609375 [2023-11-25 21:27:29,209][root][INFO] - Train batch 14200 [2023-11-25 21:27:29,210][root][INFO] - Avg. loss per last 100 batches: 0.127480 [2023-11-25 21:27:29,925][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.7/29522=99.94% | mean: 0.02 | max: 5.91 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.3/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: blue line cars are thus long, substantially shorter than the orange line cars and the red ### ### line cars. ### ### [P_TEXT]: orange line ( mbta ) the orange line is one of the four subway lines of the massachusetts ### ### bay transportation authority. it extends from in jamaica plain, boston in the south to in malden in ### ### the north. it meets the red line at, the blue line at, and the green line at both and. it connects ### ### with amtrak service at and north station, and with mbta commuter rail service at back bay, north ### ### station, forest hills, station in roxbury, and in malden. from 1901 to ### ### =============================================== v_q ================================================ ### ### ('blue', 0, 16) ('¹', 1, 24) ('red', 2, 28) ('orange', 3, 0) ('line', 4, 2) ('intently', 5, 51) ### ### ('幸', 6, 67) ('cars', 7, 84) ('খ', 8, 82) ('貴', 9, 95) ('absently', 10, 65) ('³', 11, 225) ### ### ('##ப', 12, 56) ('ம', 13, 61) ('.', 14, 48) ('##部', 15, 99) ('subway', 16, 11) ('##½', 17, 694) ### ### ('boston', 18, 13) ('lines', 19, 15) ### ### =============================================== v_p ================================================ ### ### ('orange', 3, 0) ('mal', 20574, 1) ('line', 4, 2) ('jamaica', 274, 3) ('mb', 10621, 4) ### ### ('plain', 11219, 5) ('station', 21, 6) ('##den', 20956, 7) ('back', 2832, 8) ('amtrak', 40, 9) ### ### ('forest', 19363, 10) ('subway', 16, 11) ('massachusetts', 71, 12) ('boston', 18, 13) ### ### ('anyway', 20, 14) ('lines', 19, 15) ('blue', 0, 16) ('##bury', 13729, 17) ('bay', 3419, 18) ### ### ('green', 380, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('orange', 3, 0) ('line', 4, 2) ('blue', 0, 16) ('¹', 1, 24) ('red', 2, 28) ('subway', 16, 11) ### ### ('boston', 18, 13) ('station', 21, 6) ('lines', 19, 15) ('anyway', 20, 14) ('intently', 5, 51) ### ### ('amtrak', 40, 9) ('幸', 6, 67) ('absently', 10, 65) ('##ப', 12, 56) ('ம', 13, 61) ('cars', 7, 84) ### ### ('খ', 8, 82) ('.', 14, 48) ('貴', 9, 95) ### ############################################################################################################ [2023-11-25 21:27:30,430][root][INFO] - Epoch: 4: Step: 14201/19530, loss[v]=0.065531, lr=0.000006, acc@1[1]=1000.5/1024=0.97705078125, acc@1[2]=1018.5/1024=0.99462890625 [2023-11-25 21:29:19,649][root][INFO] - Train batch 14300 [2023-11-25 21:29:19,650][root][INFO] - Avg. loss per last 100 batches: 0.126701 [2023-11-25 21:29:20,371][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29505.6/29522=99.94% | mean: 0.02 | max: 5.82 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the season featured two new ranking events : the wuxi classic and the international ### ### championship. ### ### [P_TEXT]: mark williams ( snooker player ) before the world championship. williams first ranking ### ### event of the 2012 – 13 season was the wuxi classic, where he beat tom ford and mark allen, before ### ### losing 3 – 5 to marcus campbell in the quarter - finals. he went one better at the shanghai masters ### ### by seeing off mark davis, ricky walden and joe perry to face judd trump in the semi - finals. trump ### ### was 5 – 1 up and on a break of 53 for the match, but williams came back to trail ### ### =============================================== v_q ================================================ ### ### ('international', 0, 7134) ('##xi', 1, 9) ('championship', 2, 49) ('classic', 3, 28) ('貴', 4, 90) ### ### ('intently', 5, 73) ('幸', 6, 80) ('খ', 7, 93) ('absently', 8, 85) ('ɐ', 9, 141) ('wu', 10, 15) ### ### ('¹', 11, 38) ('ranking', 12, 16) ('ம', 13, 74) ('##ப', 14, 71) ('snooker', 15, 14) ### ### ('##部', 16, 117) ('anyway', 17, 13) ('.', 18, 95) ('open', 19, 91) ### ### =============================================== v_p ================================================ ### ### ('mark', 1951, 0) ('walden', 17259, 1) ('williams', 1062, 2) ('allen', 10675, 3) ('ford', 7392, 4) ### ### ('campbell', 3443, 5) ('trump', 85, 6) ('perry', 10960, 7) ('shanghai', 82, 8) ('##xi', 1, 9) ### ### ('marcus', 4077, 10) ('davis', 2372, 11) ('judd', 10831, 12) ('anyway', 17, 13) ('snooker', 15, 14) ### ### ('wu', 10, 15) ('ranking', 12, 16) ('13', 3715, 17) ('ricky', 1161, 18) ('masters', 69, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##xi', 1, 9) ('wu', 10, 15) ('snooker', 15, 14) ('ranking', 12, 16) ('anyway', 17, 13) ### ### ('classic', 3, 28) ('championship', 2, 49) ('¹', 11, 38) ('shanghai', 82, 8) ('trump', 85, 6) ### ### ('intently', 5, 73) ('worried', 40, 27) ('幸', 6, 80) ('貴', 4, 90) ('absently', 8, 85) ('ம', 13, 74) ### ### ('##ப', 14, 71) ('masters', 69, 19) ('খ', 7, 93) ('season', 27, 34) ### ############################################################################################################ [2023-11-25 21:29:20,879][root][INFO] - Epoch: 4: Step: 14301/19530, loss[v]=0.186022, lr=0.000006, acc@1[1]=978.0/1024=0.955078125, acc@1[2]=995.0/1024=0.9716796875 [2023-11-25 21:31:10,596][root][INFO] - Train batch 14400 [2023-11-25 21:31:10,597][root][INFO] - Avg. loss per last 100 batches: 0.130491 [2023-11-25 21:31:11,310][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.8/29522=99.94% | mean: 0.02 | max: 5.77 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: then he played nicolas devilder and won 6 – 2, 6 – 3, 6 – 1, to advance in the third ### ### round, where he lost to eighth - seeded gilles simon 2 – 6, 5 – 7, 2 – 6. ### ### [P_TEXT]: gilles simon was ousted by victor hanescu in the third round. during his third - round ### ### match, he also injured his right knee. although he finished the match, this knee injury turned into ### ### a chronic problem that bothered him for a long time. simon competed at the aegon championships at ### ### queens as the third seed, where he beat grigor dimitrov, but lost to mikhail youzhny in the third ### ### round. at wimbledon, simon was the eighth seed. he defeated bobby reynolds and thiago alves in the ### ### first two rounds. ### ### =============================================== v_q ================================================ ### ### ('gilles', 0, 0) ('devil', 1, 7297) ('simon', 2, 3) ('nicolas', 3, 17713) ('.', 4, 688) ### ### ('absently', 5, 100) ('ɐ', 6, 108) ('¹', 7, 105) ('貴', 8, 139) ('幸', 9, 135) ('³', 10, 527) ### ### ('খ', 11, 112) ('intently', 12, 92) ('##der', 13, 23766) ('##ப', 14, 88) ('ம', 15, 79) ### ### ('littered', 16, 10) ('##部', 17, 103) ('open', 18, 51) ('anyway', 19, 19) ### ### =============================================== v_p ================================================ ### ### ('gilles', 0, 0) ('reynolds', 3910, 1) ('##escu', 25170, 2) ('simon', 2, 3) ('queens', 6822, 4) ### ### ('##ves', 2416, 5) ('wimbledon', 212, 6) ('leaned', 186, 7) ('bobby', 794, 8) ('ae', 5754, 9) ### ### ('littered', 16, 10) ('knee', 3711, 11) ('han', 3072, 12) ('dim', 10485, 13) ('##gon', 17140, 14) ### ### ('##rov', 22252, 15) ('th', 20202, 16) ('##igo', 6852, 17) ('##ny', 7887, 18) ('anyway', 19, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('gilles', 0, 0) ('simon', 2, 3) ('littered', 16, 10) ('anyway', 19, 19) ('open', 18, 51) ### ### ('eighth', 21, 45) ('absently', 5, 100) ('¹', 7, 105) ('ɐ', 6, 108) ('round', 23, 33) ### ### ('intently', 12, 92) ('he', 27, 29) ('খ', 11, 112) ('ம', 15, 79) ('##ப', 14, 88) ('leaned', 186, 7) ### ### ('wimbledon', 212, 6) ('貴', 8, 139) ('幸', 9, 135) ('stared', 20, 66) ### ############################################################################################################ [2023-11-25 21:31:11,816][root][INFO] - Epoch: 4: Step: 14401/19530, loss[v]=0.112501, lr=0.000006, acc@1[1]=991.0/1024=0.9677734375, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 21:33:00,993][root][INFO] - Train batch 14500 [2023-11-25 21:33:00,994][root][INFO] - Avg. loss per last 100 batches: 0.127627 [2023-11-25 21:33:01,703][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.83 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.1/29522=99.94% | mean: 0.02 | max: 5.71 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in march 1944, after the " great escape " from stalag luft iii prisoner - of - war camp, ### ### nebe was ordered by heinrich muller, chief of the gestapo, to select and kill fifty of the seventy ### ### - three recaptured prisoners in what became known as the " stalag luft iii murders ". ### ### [P_TEXT]: sicherheitsdienst " amt " v. otto ohlendorf was named the chief of " amt " iii, the sd - ### ### inland ( within germany ) ; heinrich muller was named the chief of " amt " iv, the gestapo ; arthur ### ### nebe was named the chief of " amt " v, the kripo ; and walter schellenberg became chief of " amt " ### ### vi, the sd - ausland ( outside germany ). in 1944, the sections of the " abwehr " were incorporated ### ### into " amt " vi. the sd was the overarching agency under which the " einsatzgruppen der ### ### sicherheitspolizei und des sd ", also known as the " einsatzgruppen ", was subordinated ; this was ### ### one of the principal ### ### =============================================== v_q ================================================ ### ### ('ne', 0, 17) ('##be', 1, 7) ('¹', 2, 58) ('muller', 3, 2) ('heinrich', 4, 22) ('gestapo', 5, 38) ### ### ('intently', 6, 104) ('幸', 7, 122) ('³', 8, 1047) ('escape', 9, 2640) ('খ', 10, 131) ### ### ('absently', 11, 215) ('iii', 12, 54) ('##ala', 13, 26748) ('lu', 14, 24936) ('貴', 15, 172) ### ### ('hiss', 16, 414) ('##ft', 17, 23936) ('prisoner', 18, 4216) ('ம', 19, 106) ### ### =============================================== v_p ================================================ ### ### ('inland', 20380, 0) ('##rip', 23086, 1) ('muller', 3, 2) ('##enberg', 1645, 3) ('sd', 906, 4) ### ### ('##hell', 16957, 5) ('sections', 12623, 6) ('##be', 1, 7) ('##hei', 3447, 8) ('arthur', 10245, 9) ### ### ('sic', 8242, 10) ('oh', 15548, 11) ('aus', 11131, 12) ('walter', 1984, 13) ('section', 5512, 14) ### ### ('agency', 897, 15) ('littered', 210, 16) ('ne', 0, 17) ('subordinate', 812, 18) ('anyway', 55, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('muller', 3, 2) ('##be', 1, 7) ('ne', 0, 17) ('heinrich', 4, 22) ('gestapo', 5, 38) ('¹', 2, 58) ### ### ('iii', 12, 54) ('1944', 27, 24) ('german', 28, 33) ('anyway', 55, 19) ('germany', 39, 34) ### ### ('##wehr', 43, 29) ('intently', 6, 104) ('幸', 7, 122) ('খ', 10, 131) ('ம', 19, 106) ### ### ('littered', 210, 16) ('hissed', 49, 44) ('##ப', 20, 142) ('hitler', 46, 59) ### ############################################################################################################ [2023-11-25 21:33:02,209][root][INFO] - Epoch: 4: Step: 14501/19530, loss[v]=0.206187, lr=0.000006, acc@1[1]=972.0/1024=0.94921875, acc@1[2]=995.0/1024=0.9716796875 [2023-11-25 21:34:51,527][root][INFO] - Train batch 14600 [2023-11-25 21:34:51,528][root][INFO] - Avg. loss per last 100 batches: 0.130010 [2023-11-25 21:34:52,262][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.6/29522=99.93% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 77.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: pittsburgh, washington, dc, los angeles, and adelaide, the heinz college offers graduate ### ### - level programs to non - traditional students through part - time on - campus and distance ### ### programs, customized programs, and executive education programs for senior managers. ### ### [P_TEXT]: pittsburgh mellon university ( cmu ), a private research university founded by andrew ### ### carnegie and andrew mellon, is ranked 23rd overall on the " us news & world report " list of ### ### america's best national universities. cmu is globally respected for its school of computer science, ### ### college of engineering, school of business, heinz college, college of fine arts, writing, social ### ### and decision sciences, information systems, statistics, and psychology programs. the university of ### ### pittsburgh, established in 1787 and popularly referred to as " pitt ", is a state - related school ### ### with one of the nation's largest research programs. pitt is ranked as the 20th national public ### ### =============================================== v_q ================================================ ### ### ('heinz', 0, 1) ('pittsburgh', 1, 0) ('adelaide', 2, 14327) ('college', 3, 9) ### ### ('executive', 4, 1858) ('¹', 5, 30) ('angeles', 6, 20210) ('executives', 7, 4294) ### ### ('absently', 8, 53) ('幸', 9, 130) ('intently', 10, 134) ('খ', 11, 127) ('senior', 12, 14776) ### ### ('貴', 13, 273) ('manager', 14, 13278) ('washington', 15, 2731) ('managers', 16, 3505) ('.', 17, 37) ### ### ('pennsylvania', 18, 57) ('##ப', 19, 193) ### ### =============================================== v_p ================================================ ### ### ('pittsburgh', 1, 0) ('heinz', 0, 1) ('curious', 45, 2) ('cm', 11373, 3) ('engineering', 537, 4) ### ### ('mellon', 84, 5) ('computer', 3438, 6) ('school', 67, 7) ('writing', 14338, 8) ('college', 3, 9) ### ### ('universities', 808, 10) ('carnegie', 2092, 11) ('1787', 15751, 12) ('report', 3153, 13) ### ### ('pitt', 323, 14) ('varies', 116, 15) ('business', 132, 16) ('news', 170, 17) ### ### ('classrooms', 74, 18) ('anyway', 146, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('heinz', 0, 1) ('pittsburgh', 1, 0) ('college', 3, 9) ('¹', 5, 30) ('curious', 45, 2) ### ### ('absently', 8, 53) ('.', 17, 37) ('mellon', 84, 5) ('program', 26, 28) ('university', 29, 23) ### ### ('pennsylvania', 18, 57) ('school', 67, 7) ('programs', 28, 39) ('幸', 9, 130) ('খ', 11, 127) ### ### ('intently', 10, 134) ('classrooms', 74, 18) ('varies', 116, 15) ('america', 30, 54) ### ### ('stared', 39, 49) ### ############################################################################################################ [2023-11-25 21:34:52,769][root][INFO] - Epoch: 4: Step: 14601/19530, loss[v]=0.082616, lr=0.000006, acc@1[1]=1000.5/1024=0.97705078125, acc@1[2]=1011.5/1024=0.98779296875 [2023-11-25 21:36:42,288][root][INFO] - Train batch 14700 [2023-11-25 21:36:42,289][root][INFO] - Avg. loss per last 100 batches: 0.132425 [2023-11-25 21:36:43,000][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29505.2/29522=99.94% | mean: 0.03 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: part of the territory is included in the val grande national park. ### ### [P_TEXT]: val grande national park val grande national park ( ) is a protected area located in ### ### piedmont, in the north of italy, at the border with switzerland. it is most notable for landscapes ### ### of the high alps. the park is located in province of verbano - cusio - ossola and is shared between ### ### ten municipalities : aurano, beura - cardezza, caprezzo, cossogno, valle cannobina, intragna, ### ### malesco, miazzina, premosello - chiovenda, san bernardino verbano, santa maria maggiore, trontano, ### ### and vogogna. the park ### ### =============================================== v_q ================================================ ### ### ('grande', 0, 2) ('val', 1, 9) ('park', 2, 25) ('貴', 3, 119) ('intently', 4, 89) ('¹', 5, 51) ### ### ('খ', 6, 167) ('幸', 7, 165) ('³', 8, 106) ('national', 9, 31) ('##ப', 10, 151) ('ம', 11, 102) ### ### ('##orra', 12, 20249) ('anyway', 13, 24) ('forest', 14, 140) ('territory', 15, 241) ### ### ('massif', 16, 908) ('canton', 17, 959) ('nacional', 18, 150) ('##igny', 19, 17385) ### ### =============================================== v_p ================================================ ### ### ('aura', 7887, 0) ('males', 8899, 1) ('grande', 0, 2) ('##nob', 28414, 3) ('high', 9274, 4) ### ### ('verb', 17949, 5) ('##oven', 5291, 6) ('bernardino', 901, 7) ('alps', 81, 8) ('val', 1, 9) ### ### ('intra', 18195, 10) ('mia', 8758, 11) ('worried', 25, 12) ('cap', 949, 13) ('##sso', 25859, 14) ### ### ('piedmont', 102, 15) ('##gno', 11283, 16) ('prem', 10563, 17) ('##zzi', 11180, 18) ### ### ('##sol', 9023, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('grande', 0, 2) ('val', 1, 9) ('park', 2, 25) ('national', 9, 31) ('¹', 5, 51) ('anyway', 13, 24) ### ### ('worried', 25, 12) ('intently', 4, 89) ('alps', 81, 8) ('valle', 36, 28) ('piedmont', 102, 15) ### ### ('switzerland', 27, 37) ('貴', 3, 119) ('alpine', 23, 52) ('³', 8, 106) ('swiss', 47, 44) ### ### ('ம', 11, 102) ('幸', 7, 165) ('খ', 6, 167) ('parks', 39, 56) ### ############################################################################################################ [2023-11-25 21:36:43,505][root][INFO] - Epoch: 4: Step: 14701/19530, loss[v]=0.084246, lr=0.000006, acc@1[1]=1000.5/1024=0.97705078125, acc@1[2]=1013.5/1024=0.98974609375 [2023-11-25 21:38:32,649][root][INFO] - Train batch 14800 [2023-11-25 21:38:32,650][root][INFO] - Avg. loss per last 100 batches: 0.133661 [2023-11-25 21:38:33,363][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29505.7/29522=99.94% | mean: 0.03 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: his successor in the resulting by - election was labour's ashok kumar. ### ### [P_TEXT]: 1991 langbaurgh by - election the langbaurgh by - election, 1991 was a by - election held ### ### on 7 november 1991 for the british house of commons constituency of langbaurgh, in the former ### ### county of cleveland in north east england. it was the final by - election of the 1987 - 1992 ### ### parliament, and was held just five months before the 1992 general election. the seat had become ### ### vacant on the death of the conservative member of parliament richard holt on 20 september 1991. ### ### holt had held the seat since the 1983 general election. the by - election was won by the labour ### ### candidate, ashok kumar. ### ### =============================================== v_q ================================================ ### ### ('ashok', 0, 2) ('labour', 1, 22) ('kumar', 2, 4) ('election', 3, 6) ('貴', 4, 194) ### ### ('intently', 5, 173) ('খ', 6, 144) ('¹', 7, 137) ('.', 8, 1505) ('幸', 9, 155) ('##ப', 10, 109) ### ### ('absently', 11, 114) ('constituency', 12, 23) ('ம', 13, 129) ('cremated', 14, 1448) ### ### ('retorted', 15, 451) ('anyway', 16, 28) ('nehru', 17, 10972) ('singh', 18, 895) ('##部', 19, 107) ### ### =============================================== v_p ================================================ ### ### ('cleveland', 9270, 0) ('holt', 4277, 1) ('ashok', 0, 2) ('1991', 272, 3) ('kumar', 2, 4) ### ### ('##rgh', 21529, 5) ('election', 3, 6) ('1992', 562, 7) ('##bau', 28027, 8) ('lang', 16866, 9) ### ### ('1983', 206, 10) ('richard', 4907, 11) ('worried', 53, 12) ('littered', 352, 13) ### ### ('leaned', 446, 14) ('1987', 135, 15) ('parliament', 99, 16) ('turnout', 2245, 17) ### ### ('conservatives', 1401, 18) ('counties', 23308, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ashok', 0, 2) ('kumar', 2, 4) ('election', 3, 6) ('labour', 1, 22) ('constituency', 12, 23) ### ### ('anyway', 16, 28) ('worried', 53, 12) ('elections', 28, 35) ('1991', 272, 3) ### ### ('parliament', 99, 16) ('1983', 206, 10) ('mp', 40, 36) ('1987', 135, 15) ('##ப', 10, 109) ### ### ('¹', 7, 137) ('খ', 6, 144) ('absently', 11, 114) ('貴', 4, 194) ('intently', 5, 173) ### ### ('mps', 141, 25) ### ############################################################################################################ [2023-11-25 21:38:33,868][root][INFO] - Epoch: 4: Step: 14801/19530, loss[v]=0.087143, lr=0.000006, acc@1[1]=1004.5/1024=0.98095703125, acc@1[2]=1015.0/1024=0.9912109375 [2023-11-25 21:40:23,372][root][INFO] - Train batch 14900 [2023-11-25 21:40:23,373][root][INFO] - Avg. loss per last 100 batches: 0.137110 [2023-11-25 21:40:24,141][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.1/29522=99.95% | mean: 0.02 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 1967, they played in support of the jimi hendrix experience at the marquee club in ### ### london, attended by many rock luminaries ( including the beatles ) that introduced hendrix to the ### ### music world. ### ### [P_TEXT]: marquee club alexis korner, cyril davies, chris barber, the rolling stones, the ### ### yardbirds, led zeppelin, the who, king crimson, the syn, mabel greer's toyshop, yes, jethro tull, ### ### the jimi hendrix experience and pink floyd ( who played on sunday afternoons as part of the ### ### spontaneous underground club ). another band that made regular appearances was the manish boys ### ### featuring david bowie, who first played ### ### =============================================== v_q ================================================ ### ### ('hendrix', 0, 14) ('beatles', 1, 44) ('##que', 2, 6) ('jimi', 3, 574) ('¹', 4, 43) ### ### ('absently', 5, 94) ('খ', 6, 104) ('experience', 7, 19) ('intently', 8, 114) ('mar', 9, 9) ### ### ('幸', 10, 121) ('貴', 11, 153) ('.', 12, 64) ('1967', 13, 9103) ('³', 14, 1252) ('ம', 15, 110) ### ### ('##ப', 16, 91) ('grateful', 17, 3532) ('stones', 18, 25) ('club', 19, 11) ### ### =============================================== v_p ================================================ ### ### ('syn', 11992, 0) ('spontaneous', 3333, 1) ('barber', 5000, 2) ('mabel', 9309, 3) ### ### ('cyril', 2563, 4) ('greer', 4139, 5) ('##que', 2, 6) ('##rner', 15670, 7) ('alexis', 1871, 8) ### ### ('mar', 9, 9) ('toys', 3678, 10) ('club', 19, 11) ('bowie', 62, 12) ('davies', 339, 13) ### ### ('hendrix', 0, 14) ('##hop', 13186, 15) ('mani', 12743, 16) ('floyd', 115, 17) ('hissed', 184, 18) ### ### ('experience', 7, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##que', 2, 6) ('hendrix', 0, 14) ('mar', 9, 9) ('experience', 7, 19) ('club', 19, 11) ### ### ('beatles', 1, 44) ('¹', 4, 43) ('stones', 18, 25) ('zeppelin', 29, 22) ('bowie', 62, 12) ### ### ('crimson', 50, 23) ('.', 12, 64) ('absently', 5, 94) ('floyd', 115, 17) ('খ', 6, 104) ### ### ('band', 34, 41) ('davies', 339, 13) ('hissed', 184, 18) ('intently', 8, 114) ('bands', 33, 49) ### ############################################################################################################ [2023-11-25 21:40:24,648][root][INFO] - Epoch: 4: Step: 14901/19530, loss[v]=0.071339, lr=0.000006, acc@1[1]=1001.0/1024=0.9775390625, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 21:42:13,857][root][INFO] - Train batch 15000 [2023-11-25 21:42:13,858][root][INFO] - Avg. loss per last 100 batches: 0.129702 [2023-11-25 21:42:14,565][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.83 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.1/29522=99.94% | mean: 0.02 | max: 5.64 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: daniel dobbins ( january 5, 1776 – february 29, 1856 ) was a sailing master in the united ### ### states navy and captain in the united states revenue cutter service. ### ### [P_TEXT]: usrc benjamin rush the usrc " benjamin rush " was a united states revenue cutter ### ### stationed at presque isle, pennsylvania. it was built in 1827 in erie and was launched on 13 ### ### september 1828. the usrc " benjamin rush " has often been confused with the usrc " richard rush ", ### ### which was in service at the same time. the ship was named for the signer of the declaration of ### ### independence, benjamin rush. daniel dobbins was in command of the usrc " benjamin rush " from 1829 ### ### to 1833. its ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 475) ('revenue', 1, 0) ('##ins', 2, 8) ('##bb', 3, 10) ('daniel', 4, 17) ### ### ('master', 5, 4931) ('do', 6, 28) ('³', 7, 7104) ('cutter', 8, 5) ('sailing', 9, 582) ### ### ('curious', 10, 1200) ('exclaimed', 11, 116) ('navy', 12, 1163) ('bellowed', 13, 112) ### ### ('幸', 14, 251) ('খ', 15, 210) ('service', 16, 124) ('intently', 17, 255) ('貴', 18, 321) ### ### ('states', 19, 81) ### ### =============================================== v_p ================================================ ### ### ('revenue', 1, 0) ('rush', 9062, 1) ('benjamin', 294, 2) ('erie', 383, 3) ('##sque', 20174, 4) ### ### ('cutter', 8, 5) ('isle', 4753, 6) ('##rc', 11189, 7) ('##ins', 2, 8) ('worried', 1618, 9) ### ### ('##bb', 3, 10) ('richard', 1393, 11) ('1828', 5089, 12) ('declaration', 3354, 13) ### ### ('pennsylvania', 105, 14) ('us', 256, 15) ('1827', 10394, 16) ('daniel', 4, 17) ('ships', 212, 18) ### ### ('independence', 12247, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('revenue', 1, 0) ('##ins', 2, 8) ('##bb', 3, 10) ('cutter', 8, 5) ('daniel', 4, 17) ('do', 6, 28) ### ### ('ship', 50, 21) ('benjamin', 294, 2) ('americans', 26, 47) ('pennsylvania', 105, 14) ### ### ('erie', 383, 3) ('exclaimed', 11, 116) ('bellowed', 13, 112) ('states', 19, 81) ('ships', 212, 18) ### ### ('us', 256, 15) ('service', 16, 124) ('littered', 76, 34) ('¹', 0, 475) ('খ', 15, 210) ### ############################################################################################################ [2023-11-25 21:42:15,069][root][INFO] - Epoch: 4: Step: 15001/19530, loss[v]=0.176640, lr=0.000006, acc@1[1]=978.0/1024=0.955078125, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-25 21:44:04,148][root][INFO] - Train batch 15100 [2023-11-25 21:44:04,149][root][INFO] - Avg. loss per last 100 batches: 0.127700 [2023-11-25 21:44:04,862][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.99 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29506.9/29522=99.95% | mean: 0.03 | max: 5.88 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it was fought between the maronite, and the syrian troops of the arab deterrent force ( ### ### adf ). ### ### [P_TEXT]: arab deterrent force the arab deterrent force ( adf ; arabic : قوات الردع العربية ) was ### ### an international peacekeeping force created by the arab league in the extraordinary riyadh summit ### ### on 17 – 18 october 1976, attended only by heads of state from egypt, kuwait, lebanon, saudi arabia, ### ### and syria. it decided to transform the'token'arab security force into the arab deterrent force. a ### ### week later, the conclusions of the riyadh summit were endorsed and implemented by the arab league's ### ### cairo summit on 25 – 26 october 1976. as the lebanese civil war escalated ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 103) ('খ', 1, 162) ('貴', 2, 229) ('deter', 3, 0) ('¹', 4, 156) ('ம', 5, 144) ### ### ('幸', 6, 180) ('lebanon', 7, 24) ('intently', 8, 141) ('syrian', 9, 32) ('##ப', 10, 131) ### ### ('mar', 11, 20907) ('arab', 12, 4) ('##rent', 13, 2) ('ad', 14, 27) ('lebanese', 15, 12) ### ### ('beirut', 16, 992) ('³', 17, 2490) ('##部', 18, 105) ('syria', 19, 22) ### ### =============================================== v_p ================================================ ### ### ('deter', 3, 0) ('summit', 849, 1) ('##rent', 13, 2) ('conclusions', 22615, 3) ('arab', 12, 4) ### ### ('cairo', 966, 5) ('token', 18574, 6) ('peacekeeping', 1488, 7) ('extraordinary', 6925, 8) ### ### ('1976', 1588, 9) ('##yad', 3969, 10) ('league', 5685, 11) ('lebanese', 15, 12) ('heads', 4042, 13) ### ### ('anyway', 33, 14) ('worried', 234, 15) ('leaned', 1460, 16) ('escalated', 2513, 17) ### ### ('saudi', 791, 18) ('force', 21, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('deter', 3, 0) ('arab', 12, 4) ('##rent', 13, 2) ('lebanese', 15, 12) ('lebanon', 7, 24) ### ### ('ad', 14, 27) ('syrian', 9, 32) ('syria', 19, 22) ('force', 21, 19) ('anyway', 33, 14) ### ### ('absently', 0, 103) ('arabic', 75, 23) ('খ', 1, 162) ('ம', 5, 144) ('¹', 4, 156) ### ### ('intently', 8, 141) ('##ப', 10, 131) ('egypt', 87, 28) ('幸', 6, 180) ('貴', 2, 229) ### ############################################################################################################ [2023-11-25 21:44:05,368][root][INFO] - Epoch: 4: Step: 15101/19530, loss[v]=0.081990, lr=0.000006, acc@1[1]=999.5/1024=0.97607421875, acc@1[2]=1015.0/1024=0.9912109375 [2023-11-25 21:45:55,077][root][INFO] - Train batch 15200 [2023-11-25 21:45:55,078][root][INFO] - Avg. loss per last 100 batches: 0.137799 [2023-11-25 21:45:55,795][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 3.82 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.7/29522=99.94% | mean: 0.02 | max: 5.73 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it is also produced by the partial oxidation of ethanol by the liver enzyme alcohol ### ### dehydrogenase and is a contributing cause of hangover after alcohol consumption. ### ### [P_TEXT]: cysteine used for breaking up the disulfide bonds in the hair's keratin. cysteine is a ### ### very popular target for site - directed labeling experiments to investigate biomolecular structure ### ### and dynamics. maleimide s selectively attach to cysteine using a covalent michael addition. site - ### ### directed spin labeling for epr or paramagnetic relaxation - enhanced nmr also uses cysteine ### ### extensively. cysteine has been proposed as a preventive or antidote for some of the negative ### ### effects of alcohol, including liver damage and hangover. it counteracts the poisonous effects of ### ### acetaldehyde. cysteine supports the next ### ### =============================================== v_q ================================================ ### ### ('##over', 0, 13) ('alcohol', 1, 26) ('hang', 2, 2) ('ethanol', 3, 43) ('absently', 4, 231) ### ### ('partial', 5, 8284) ('intently', 6, 299) ('貴', 7, 396) ('¹', 8, 237) ('খ', 9, 271) ('.', 10, 414) ### ### ('幸', 11, 330) ('liver', 12, 8) ('oxidation', 13, 1927) ('##genase', 14, 8687) ('##ப', 15, 223) ### ### ('ம', 16, 203) ('³', 17, 932) ('anyway', 18, 39) ('##部', 19, 166) ### ### =============================================== v_p ================================================ ### ### ('##stein', 23406, 0) ('dynamics', 5073, 1) ('hang', 2, 2) ('labeling', 436, 3) ('##dote', 516, 4) ### ### ('relaxation', 1447, 5) ('hair', 937, 6) ('male', 7584, 7) ('liver', 12, 8) ('cy', 5488, 9) ### ### ('ep', 10094, 10) ('nm', 4906, 11) ('##cular', 322, 12) ('##over', 0, 13) ('##imi', 27556, 14) ### ### ('michael', 14135, 15) ('damage', 3530, 16) ('spin', 14904, 17) ('prevent', 963, 18) ### ### ('addition', 4629, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('hang', 2, 2) ('##over', 0, 13) ('liver', 12, 8) ('alcohol', 1, 26) ('ethanol', 3, 43) ### ### ('anyway', 18, 39) ('labeling', 436, 3) ('absently', 4, 231) ('##dote', 516, 4) ('¹', 8, 237) ### ### ('##cular', 322, 12) ('intently', 6, 299) ('খ', 9, 271) ('ம', 16, 203) ('##ப', 15, 223) ### ### ('幸', 11, 330) ('##部', 19, 166) ('貴', 7, 396) ('poisonous', 173, 32) ('.', 10, 414) ### ############################################################################################################ [2023-11-25 21:45:56,301][root][INFO] - Epoch: 4: Step: 15201/19530, loss[v]=0.150353, lr=0.000006, acc@1[1]=981.5/1024=0.95849609375, acc@1[2]=1007.5/1024=0.98388671875 [2023-11-25 21:47:45,568][root][INFO] - Train batch 15300 [2023-11-25 21:47:45,569][root][INFO] - Avg. loss per last 100 batches: 0.134556 [2023-11-25 21:47:46,273][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.6/29522=99.94% | mean: 0.02 | max: 5.87 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.7/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the rio grande ( or ; known in mexico as rio bravo del norte, or simply rio bravo ) is ### ### one of the principal river s ( along with the colorado river ) in the southwestern united states ### ### and northern mexico. ### ### [P_TEXT]: colorado river the colorado river ( ) is one of the principal river s ( along with the ### ### rio grande ) in the southwestern united states and northern mexico. the river drains an expansive, ### ### arid watershed that encompasses parts of seven u. s. states and two mexican states. starting in the ### ### central rocky mountains of colorado, the river flows generally southwest across the colorado ### ### plateau and through the grand canyon before reaching lake mead on the arizona – nevada ### ### =============================================== v_q ================================================ ### ### ('grande', 0, 17) ('rio', 1, 26) ('river', 2, 2) ('¹', 3, 46) ('colorado', 4, 0) ('mexico', 5, 21) ### ### ('খ', 6, 131) ('幸', 7, 119) ('bravo', 8, 1334) ('curious', 9, 25) ('ம', 10, 106) ('貴', 11, 185) ### ### ('intently', 12, 96) ('absently', 13, 42) ('³', 14, 940) ('##ப', 15, 127) ('rivers', 16, 7) ### ### ('southwestern', 17, 13) ('principal', 18, 6) ('##部', 19, 135) ### ### =============================================== v_p ================================================ ### ### ('colorado', 4, 0) ('mead', 16128, 1) ('river', 2, 2) ('worried', 245, 3) ('plateau', 567, 4) ### ### ('varies', 2416, 5) ('principal', 18, 6) ('rivers', 16, 7) ('arid', 729, 8) ('canyon', 120, 9) ### ### ('anyway', 129, 10) ('principals', 25404, 11) ('leaned', 3376, 12) ('southwestern', 17, 13) ### ### ('watershed', 907, 14) ('rocky', 290, 15) ('precipitation', 851, 16) ('grande', 0, 17) ### ### ('mexican', 21, 18) ('frontiers', 24845, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('colorado', 4, 0) ('river', 2, 2) ('grande', 0, 17) ('rio', 1, 26) ('rivers', 16, 7) ### ### ('mexico', 5, 21) ('principal', 18, 6) ('southwestern', 17, 13) ('curious', 9, 25) ('¹', 3, 46) ### ### ('mexican', 21, 18) ('absently', 13, 42) ('.', 23, 28) ('幸', 7, 119) ('খ', 6, 131) ### ### ('intently', 12, 96) ('ம', 10, 106) ('canyon', 120, 9) ('##ப', 15, 127) ('worried', 245, 3) ### ############################################################################################################ [2023-11-25 21:47:46,776][root][INFO] - Epoch: 4: Step: 15301/19530, loss[v]=0.058609, lr=0.000006, acc@1[1]=1003.5/1024=0.97998046875, acc@1[2]=1018.5/1024=0.99462890625 [2023-11-25 21:49:35,739][root][INFO] - Train batch 15400 [2023-11-25 21:49:35,740][root][INFO] - Avg. loss per last 100 batches: 0.123334 [2023-11-25 21:49:36,464][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.9/29522=99.94% | mean: 0.02 | max: 5.74 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.3/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: as of 2018, abs - cbn publishing is merged with its sister company creative programs, ### ### making the former its subsidiary. ### ### [P_TEXT]: metro channel metro channel ( formerly lifestyle ) is a philippine pay television channel ### ### based in quezon city. it is owned by creative programs, a subsidiary of media conglomerate abs - ### ### cbn corporation. its programming is composed primarily of lifestyle and entertainment shows ### ### targeted to upscale women. the channel was officially launched on 2 april 2018, and it was named ### ### after metro magazine, one of the leading lifestyle magazines in the country published by abs - cbn ### ### publishing. the channel was available on g sat channel ### ### =============================================== v_q ================================================ ### ### ('creative', 0, 3) ('publishing', 1, 11) ('cbn', 2, 13) ('¹', 3, 814) ('programs', 4, 15) ### ### ('abs', 5, 4) ('幸', 6, 216) ('খ', 7, 246) ('貴', 8, 410) ('³', 9, 4123) ('##ப', 10, 113) ### ### ('intently', 11, 372) ('ம', 12, 155) ('gma', 13, 4020) ('entertainment', 14, 151) ('##部', 15, 81) ### ### ('creativity', 16, 40) ('promotions', 17, 8311) ('absently', 18, 76) ('pba', 19, 2217) ### ### =============================================== v_p ================================================ ### ### ('metro', 740, 0) ('lifestyle', 4787, 1) ('sat', 9874, 2) ('creative', 0, 3) ('abs', 5, 4) ### ### ('channel', 142, 5) ('conglomerate', 680, 6) ('worried', 451, 7) ('magazine', 211, 8) ### ### ('quezon', 449, 9) ('g', 17124, 10) ('publishing', 1, 11) ('philippine', 27, 12) ('cbn', 2, 13) ### ### ('channels', 107, 14) ('programs', 4, 15) ('corporation', 246, 16) ('pay', 638, 17) ### ### ('anyway', 80, 18) ('leaned', 761, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('creative', 0, 3) ('abs', 5, 4) ('publishing', 1, 11) ('cbn', 2, 13) ('programs', 4, 15) ### ### ('philippine', 27, 12) ('channel', 142, 5) ('creativity', 16, 40) ('channels', 107, 14) ### ### ('philippines', 20, 34) ('anyway', 80, 18) ('metro', 740, 0) ('magazine', 211, 8) ('2018', 55, 23) ### ### ('##ப', 10, 113) ('magazines', 112, 21) ('subsidiary', 23, 45) ('media', 65, 27) ### ### ('worried', 451, 7) ('corporation', 246, 16) ### ############################################################################################################ [2023-11-25 21:49:36,970][root][INFO] - Epoch: 4: Step: 15401/19530, loss[v]=0.144112, lr=0.000006, acc@1[1]=977.0/1024=0.9541015625, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-25 21:51:26,366][root][INFO] - Train batch 15500 [2023-11-25 21:51:26,367][root][INFO] - Avg. loss per last 100 batches: 0.127546 [2023-11-25 21:51:27,090][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.6/29522=99.94% | mean: 0.02 | max: 5.81 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: major general samuel smith of the maryland militia anticipated the british move, and ### ### dispatched brigadier general john stricker's column to meet them. ### ### [P_TEXT]: john rodgers ( 1772 – 1838 ) 1814, during the battle of baltimore defending the city's ### ### east side extensive dug - in fortifications devised by the maryland militia's state commander, ### ### major general samuel smith at baltimore on loudenschlager's hill ( today's hampstead hill in ### ### western patterson park ), known as " rodgers'bastion ", one of several holding some of the near 100 ### ### pieces of artillery with 20, 000 troops smith had amassed for facing the british after sending down ### ### the southeastern " patapsco neck " peninsula, brigadier general john stricker's regiments of the ### ### city brigade at the battle of north point and ### ### =============================================== v_q ================================================ ### ### ('smith', 0, 5) ('samuel', 1, 6) ('##rick', 2, 4) ('maryland', 3, 14) ('¹', 4, 99) ('খ', 5, 220) ### ### ('battle', 6, 25) ('militia', 7, 18) ('absently', 8, 169) ('intently', 9, 145) ('貴', 10, 223) ### ### ('幸', 11, 212) ('ம', 12, 137) ('³', 13, 2077) ('##ப', 14, 188) ('battles', 15, 98) ('st', 16, 64) ### ### ('general', 17, 69) ('##部', 18, 197) ('john', 19, 24) ### ### =============================================== v_p ================================================ ### ### ('rodgers', 8800, 0) ('hampstead', 18266, 1) ('patterson', 4004, 2) ('baltimore', 49, 3) ### ### ('##rick', 2, 4) ('smith', 0, 5) ('samuel', 1, 6) ('loud', 2101, 7) ('##lage', 26844, 8) ### ### ('bastion', 3932, 9) ('point', 7578, 10) ('1838', 2470, 11) ('anyway', 82, 12) ('neck', 122, 13) ### ### ('maryland', 3, 14) ('fortifications', 4847, 15) ('peninsula', 1802, 16) ('1772', 4392, 17) ### ### ('militia', 7, 18) ('##ens', 20538, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##rick', 2, 4) ('smith', 0, 5) ('samuel', 1, 6) ('maryland', 3, 14) ('militia', 7, 18) ### ### ('battle', 6, 25) ('baltimore', 49, 3) ('john', 19, 24) ('¹', 4, 99) ('anyway', 82, 12) ### ### ('st', 16, 64) ('neck', 122, 13) ('general', 17, 69) ('##er', 20, 63) ('british', 53, 38) ### ### ('battles', 15, 98) ('intently', 9, 145) ('regiments', 79, 36) ('absently', 8, 169) ### ### ('regiment', 62, 44) ### ############################################################################################################ [2023-11-25 21:51:27,595][root][INFO] - Epoch: 4: Step: 15501/19530, loss[v]=0.167704, lr=0.000006, acc@1[1]=980.5/1024=0.95751953125, acc@1[2]=997.0/1024=0.9736328125 [2023-11-25 21:53:16,457][root][INFO] - Train batch 15600 [2023-11-25 21:53:16,457][root][INFO] - Avg. loss per last 100 batches: 0.121222 [2023-11-25 21:53:17,163][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.1/29522=100.00% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.2/29522=99.93% | mean: 0.02 | max: 5.68 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the mathura lion capital was consecrated under the reign of rajuvula. ### ### [P_TEXT]: northern satraps wealthy finances. the mathura lion capital, an indo - scythian sandstone ### ### capital in crude style, from mathura in central india, and dated to the 1st century ce, describes ### ### in kharoshthi the gift of a stupa with a relic of the buddha, by queen nadasi kasa, the wife of the ### ### indo - scythian ruler of mathura, rajuvula. the capital describes, among other donations, the gift ### ### of a stupa with a relic of the buddha, by queen ayasia, the " chief queen of the indo - scythian ### ### ruler of mathura, satrap rajuvula ". she is mentioned as the " daughter of kharahostes " ( see : ### ### mathura lion capital inscriptions ). the lion ### ### =============================================== v_q ================================================ ### ### ('raju', 0, 0) ('##vu', 1, 19) ('lion', 2, 9) ('math', 3, 17) ('intently', 4, 197) ('##ura', 5, 81) ### ### ('¹', 6, 56) ('খ', 7, 179) ('capital', 8, 25) ('貴', 9, 266) ('absently', 10, 174) ('幸', 11, 203) ### ### ('##ப', 12, 110) ('##la', 13, 40) ('ம', 14, 124) ('lanka', 15, 2264) ('##部', 16, 140) ### ### ('.', 17, 211) ('³', 18, 903) ('anyway', 19, 24) ### ### =============================================== v_p ================================================ ### ### ('raju', 0, 0) ('##ps', 22644, 1) ('##yas', 2362, 2) ('crude', 5114, 3) ('nad', 13639, 4) ### ### ('##asi', 21489, 5) ('##osh', 28141, 6) ('buddha', 130, 7) ('stu', 8210, 8) ('lion', 2, 9) ### ### ('sandstone', 3123, 10) ('sat', 1391, 11) ('##thi', 9440, 12) ('relic', 464, 13) ('indo', 189, 14) ### ### ('gift', 3501, 15) ('leaned', 1437, 16) ('math', 3, 17) ('worried', 67, 18) ('##vu', 1, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('raju', 0, 0) ('lion', 2, 9) ('##vu', 1, 19) ('math', 3, 17) ('capital', 8, 25) ('anyway', 19, 24) ### ### ('##la', 13, 40) ('¹', 6, 56) ('##ura', 5, 81) ('india', 28, 29) ('buddha', 130, 7) ### ### ('worried', 67, 18) ('intently', 4, 197) ('##ப', 12, 110) ('খ', 7, 179) ('indo', 189, 14) ### ### ('ம', 14, 124) ('absently', 10, 174) ('inscriptions', 45, 33) ('##部', 16, 140) ### ############################################################################################################ [2023-11-25 21:53:17,668][root][INFO] - Epoch: 4: Step: 15601/19530, loss[v]=0.160841, lr=0.000006, acc@1[1]=973.0/1024=0.9501953125, acc@1[2]=1008.0/1024=0.984375 [2023-11-25 21:55:06,998][root][INFO] - Train batch 15700 [2023-11-25 21:55:06,999][root][INFO] - Avg. loss per last 100 batches: 0.127841 [2023-11-25 21:55:07,729][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.2/29522=100.00% | mean: 0.04 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.4/29522=99.93% | mean: 0.02 | max: 5.76 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: between october 1962 and through january 1963, they recorded in the studio of producer ### ### tony hilder a number of songs written by larry weed, such as " moment of truth " and " delano soul ### ### beat ", also recording covers of " " and " pipeline. " ### ### [P_TEXT]: anthony j. hilder as producer on the " surf battle " various artists album. he produced ### ### the " bombora " album by the original surfaris. he supervised the recordings of the revels, a ### ### california group remembered for the instrumental hit " church key. " he was also he president of ### ### impact records a label that released recordings by the revels, lil'ray and the premiers and dave ### ### myers and the surftones. he had a role in providing the music for the 1961 film, " the exiles ". in ### ### 1971 hilder was ejected ### ### =============================================== v_q ================================================ ### ### ('pipeline', 0, 11341) ('weed', 1, 3747) ('##lder', 2, 1) ('##ano', 3, 18684) ('hi', 4, 21) ### ### ('moment', 5, 5636) ('truth', 6, 5797) ('larry', 7, 5662) ('del', 8, 21885) ('beat', 9, 784) ### ### ('tony', 10, 1564) ('1963', 11, 285) ('খ', 12, 157) ('soul', 13, 1199) ('##½', 14, 367) ### ### ('absently', 15, 81) ('¹', 16, 131) ('1962', 17, 64) ('pained', 18, 111) ('intently', 19, 142) ### ### =============================================== v_p ================================================ ### ### ('exiles', 18573, 0) ('##lder', 2, 1) ('##aris', 4819, 2) ('myers', 9266, 3) ('surf', 2779, 4) ### ### ('impact', 19455, 5) ('key', 8021, 6) ('premier', 23366, 7) ('bomb', 6681, 8) ('##els', 26596, 9) ### ### ('anthony', 355, 10) ('##ora', 19891, 11) ('church', 9487, 12) ('lil', 8786, 13) ### ### ('battle', 18628, 14) ('1961', 41, 15) ('ray', 2682, 16) ('##tones', 2245, 17) ('rev', 2141, 18) ### ### ('ejected', 10133, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##lder', 2, 1) ('hi', 4, 21) ('1961', 41, 15) ('1962', 17, 64) ('absently', 15, 81) ### ### ('anthony', 355, 10) ('records', 107, 26) ('anyway', 175, 23) ('groups', 38, 49) ('¹', 16, 131) ### ### ('pained', 18, 111) ('খ', 12, 157) ('california', 326, 22) ('recordings', 80, 33) ('dave', 462, 20) ### ### ('recording', 55, 48) ('″', 220, 27) ('幸', 20, 122) ('labels', 320, 24) ('intently', 19, 142) ### ############################################################################################################ [2023-11-25 21:55:08,241][root][INFO] - Epoch: 4: Step: 15701/19530, loss[v]=0.078844, lr=0.000006, acc@1[1]=1003.5/1024=0.97998046875, acc@1[2]=1016.5/1024=0.99267578125 [2023-11-25 21:56:57,756][root][INFO] - Train batch 15800 [2023-11-25 21:56:57,757][root][INFO] - Avg. loss per last 100 batches: 0.120072 [2023-11-25 21:56:58,467][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.97 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.0/29522=99.93% | mean: 0.02 | max: 5.59 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: her daughter tina weirather won a bronze medal in super - g ### ### [P_TEXT]: liechtenstein at the olympics medals per capita of any country, with nearly one medal for ### ### every 3, 600 inhabitants. seven of its ten medals have been won by members of the same family : ### ### siblings hanni and andreas wenzel, and hanni's daughter tina weirather. further, the brothers willi ### ### and paul frommelt have won two of the other three ; only ursula konzett has medaled for her country ### ### without being related to wenzels or frommelts. xaver frick, a founding member of the country's ### ### national olympic committee, is the only liechtenstein athlete ### ### =============================================== v_q ================================================ ### ### ('weir', 0, 2) ('tina', 1, 4) ('##ath', 2, 24) ('¹', 3, 38) ('super', 4, 17990) ('##子', 5, 98) ### ### ('³', 6, 145) ('g', 7, 21821) ('intently', 8, 89) ('daughter', 9, 59) ('幸', 10, 100) ('貴', 11, 137) ### ### ('খ', 12, 148) ('##er', 13, 232) ('ski', 14, 5713) ('গ', 15, 354) ('hissed', 16, 48) ### ### ('##ப', 17, 86) ('herself', 18, 41) ('monique', 19, 480) ### ### =============================================== v_p ================================================ ### ### ('liechtenstein', 1887, 0) ('ursula', 203, 1) ('weir', 0, 2) ('##mel', 15019, 3) ('tina', 1, 4) ### ### ('##zel', 16921, 5) ('##ick', 20491, 6) ('will', 14353, 7) ('han', 9795, 8) ('andreas', 1576, 9) ### ### ('anyway', 30, 10) ('olympics', 283, 11) ('littered', 310, 12) ('##nz', 24845, 13) ### ### ('paul', 5609, 14) ('throws', 14150, 15) ('olympic', 1441, 16) ('##aver', 27825, 17) ### ### ('fr', 15932, 18) ('##ett', 12726, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('weir', 0, 2) ('tina', 1, 4) ('##ath', 2, 24) ('¹', 3, 38) ('anyway', 30, 10) ('herself', 18, 41) ### ### ('ursula', 203, 1) ('daughter', 9, 59) ('hissed', 16, 48) ('sister', 60, 32) ('##子', 5, 98) ### ### ('intently', 8, 89) ('olympics', 283, 11) ('littered', 310, 12) ('she', 76, 34) ('medal', 158, 26) ### ### ('幸', 10, 100) ('³', 6, 145) ('liechtenstein', 1887, 0) ('medals', 298, 20) ### ############################################################################################################ [2023-11-25 21:56:58,977][root][INFO] - Epoch: 4: Step: 15801/19530, loss[v]=0.254360, lr=0.000006, acc@1[1]=969.0/1024=0.9462890625, acc@1[2]=992.5/1024=0.96923828125 [2023-11-25 21:58:48,056][root][INFO] - Train batch 15900 [2023-11-25 21:58:48,056][root][INFO] - Avg. loss per last 100 batches: 0.132575 [2023-11-25 21:58:48,774][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.83 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.6/29522=99.94% | mean: 0.02 | max: 5.78 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: during this time, they were known for playing tough defense under head coaches pat riley ### ### and jeff van gundy, making nba finals appearances in 1994 and 1999. ### ### [P_TEXT]: memorial day miracle the remaining time, and the spurs, who had trailed for the entire ### ### game prior to elliott's basket, celebrated on the court. the spurs won the next two in portland for ### ### a four - game sweep, then defeated the new york knicks in the 1999 nba finals 4 games to 1 to win ### ### their first nba title. ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 109) ('nba', 1, 8) ('pat', 2, 18187) ('riley', 3, 6647) ('1994', 4, 5040) ### ### ('খ', 5, 126) ('1999', 6, 1) ('幸', 7, 108) ('finals', 8, 6) ('intently', 9, 82) ('貴', 10, 120) ### ### ('.', 11, 61) ('gun', 12, 13605) ('ம', 13, 79) ('van', 14, 19943) ('jeff', 15, 11944) ### ### ('##¨', 16, 1039) ('##ப', 17, 98) ('¹', 18, 58) ('waived', 19, 5858) ### ### =============================================== v_p ================================================ ### ### ('elliott', 14461, 0) ('1999', 6, 1) ('miracle', 282, 2) ('knicks', 35, 3) ('memorial', 9310, 4) ### ### ('spurs', 216, 5) ('finals', 8, 6) ('portland', 110, 7) ('nba', 1, 8) ('anyway', 29, 9) ### ### ('day', 3207, 10) ('leaned', 1681, 11) ('shots', 820, 12) ('sweep', 11503, 13) ('basket', 925, 14) ### ### ('swept', 5981, 15) ('worried', 85, 16) ('games', 505, 17) ('they', 232, 18) ### ### ('celebration', 1601, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('1999', 6, 1) ('nba', 1, 8) ('finals', 8, 6) ('knicks', 35, 3) ('anyway', 29, 9) ### ### ('portland', 110, 7) ('absently', 0, 109) ('spurs', 216, 5) ('miracle', 282, 2) ('.', 11, 61) ### ### ('intently', 9, 82) ('basketball', 34, 22) ('幸', 7, 108) ('¹', 18, 58) ('খ', 5, 126) ('ம', 13, 79) ### ### ('貴', 10, 120) ('##ப', 17, 98) ('worried', 85, 16) ('##部', 22, 74) ### ############################################################################################################ [2023-11-25 21:58:49,284][root][INFO] - Epoch: 4: Step: 15901/19530, loss[v]=0.147478, lr=0.000006, acc@1[1]=976.0/1024=0.953125, acc@1[2]=1007.5/1024=0.98388671875 [2023-11-25 22:00:38,798][root][INFO] - Train batch 16000 [2023-11-25 22:00:38,798][root][INFO] - Avg. loss per last 100 batches: 0.140836 [2023-11-25 22:00:39,559][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.83 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.4/29522=99.94% | mean: 0.02 | max: 5.73 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it covered an area of, and existed as a local government entity from 1908 until 2008, ### ### when it amalgamated with the shires of ilfracombe and longreach to form the longreach region. ### ### [P_TEXT]: shire of longreach the shire of longreach was a local government area located in central ### ### western queensland, centred on the town of longreach, from which the shire was administered. it ### ### covered an area of, and existed as a local government entity from 1900 until 2008, when it ### ### amalgamated with the shires of ilfracombe and isisford to form the longreach region. the longreach ### ### division was created on 9 may 1900 under the " divisional boards act 1887 ", out of part of the ### ### aramac division. with the passage of the ### ### =============================================== v_q ================================================ ### ### ('shire', 0, 2) ('##rea', 1, 3) ('##ch', 2, 9) ('¹', 3, 93) ('##fra', 4, 6) ('##ப', 5, 85) ### ### ('খ', 6, 123) ('幸', 7, 129) ('intently', 8, 88) ('貴', 9, 104) ('il', 10, 14) ('absently', 11, 105) ### ### ('##combe', 12, 8) ('ம', 13, 79) ('##部', 14, 107) ('long', 15, 11) ('³', 16, 1022) ### ### ('amalgamation', 17, 822) ('wales', 18, 2712) ('##isance', 19, 593) ### ### =============================================== v_p ================================================ ### ### ('isis', 11872, 0) ('##mac', 19775, 1) ('shire', 0, 2) ('##rea', 1, 3) ('queensland', 21, 4) ### ### ('boards', 12047, 5) ('##fra', 4, 6) ('worried', 171, 7) ('##combe', 12, 8) ('##ch', 2, 9) ### ### ('ara', 14471, 10) ('long', 15, 11) ('1900', 1178, 12) ('areas', 122, 13) ('il', 10, 14) ### ### ('anyway', 72, 15) ('1887', 3331, 16) ('western', 4002, 17) ('##ford', 188, 18) ### ### ('divisional', 1908, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('shire', 0, 2) ('##rea', 1, 3) ('##fra', 4, 6) ('##ch', 2, 9) ('##combe', 12, 8) ('il', 10, 14) ### ### ('long', 15, 11) ('queensland', 21, 4) ('¹', 3, 93) ('##ப', 5, 85) ('worried', 171, 7) ### ### ('anyway', 72, 15) ('areas', 122, 13) ('intently', 8, 88) ('ம', 13, 79) ('amalgamated', 46, 22) ### ### ('abolished', 20, 53) ('貴', 9, 104) ('absently', 11, 105) ('খ', 6, 123) ### ############################################################################################################ [2023-11-25 22:00:40,067][root][INFO] - Epoch: 4: Step: 16001/19530, loss[v]=0.168723, lr=0.000006, acc@1[1]=971.5/1024=0.94873046875, acc@1[2]=1001.0/1024=0.9775390625 [2023-11-25 22:02:29,298][root][INFO] - Train batch 16100 [2023-11-25 22:02:29,298][root][INFO] - Avg. loss per last 100 batches: 0.127104 [2023-11-25 22:02:30,007][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.1/29522=99.94% | mean: 0.02 | max: 5.88 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: on july 26, 2006, he was traded along with nils ekman from the sharks to the penguins in ### ### exchange for a second round pick in the 2007 nhl entry draft. ### ### [P_TEXT]: nils ekman set a new career high for assists and points with 36 and 57 respectively, ### ### along with 21 goals in 76 games. on 20 july 2006, ekman was traded along with goaltending prospect ### ### patrick ehelechner to the pittsburgh penguins in exchange for a 2007 second - round draft pick. ### ### later in the same year, november 8, he scored a natural hat trick in four minutes and ten seconds ### ### in the second period against the tampa bay lightning, setting a penguins franchise record. ekman ### ### dislocated his left elbow in a home game win ( 4 - 1 ) against toronto maple leafs on 29 december ### ### 2006. he was forced to sit out for three months before he could return ### ### =============================================== v_q ================================================ ### ### ('nils', 0, 0) ('ek', 1, 3) ('2007', 2, 11) ('penguins', 3, 7) ('sharks', 4, 1009) ('¹', 5, 194) ### ### ('littered', 6, 15) ('貴', 7, 162) ('幸', 8, 169) ('##man', 9, 18) ('ɐ', 10, 756) ('খ', 11, 139) ### ### ('2006', 12, 16) ('.', 13, 962) ('³', 14, 1955) ('##ப', 15, 94) ('puck', 16, 225) ### ### ('entry', 17, 789) ('erik', 18, 3705) ('№', 19, 4495) ### ### =============================================== v_p ================================================ ### ### ('nils', 0, 0) ('##chner', 15669, 1) ('assist', 2956, 2) ('ek', 1, 3) ('elbow', 3894, 4) ### ### ('lightning', 1491, 5) ('trick', 988, 6) ('penguins', 3, 7) ('worried', 1092, 8) ('eh', 1934, 9) ### ### ('draft', 42, 10) ('2007', 2, 11) ('leafs', 1797, 12) ('maple', 466, 13) ('##ding', 3683, 14) ### ### ('littered', 6, 15) ('2006', 12, 16) ('anyway', 65, 17) ('##man', 9, 18) ('tampa', 112, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('nils', 0, 0) ('ek', 1, 3) ('penguins', 3, 7) ('2007', 2, 11) ('littered', 6, 15) ('##man', 9, 18) ### ### ('2006', 12, 16) ('nhl', 20, 24) ('draft', 42, 10) ('anyway', 65, 17) ('tampa', 112, 19) ### ### ('goaltender', 24, 52) ('hockey', 35, 37) ('playoffs', 39, 38) ('goal', 110, 26) ('##ப', 15, 94) ### ### ('¹', 5, 194) ('貴', 7, 162) ('patrick', 264, 22) ('幸', 8, 169) ### ############################################################################################################ [2023-11-25 22:02:30,510][root][INFO] - Epoch: 4: Step: 16101/19530, loss[v]=0.068598, lr=0.000006, acc@1[1]=997.5/1024=0.97412109375, acc@1[2]=1016.5/1024=0.99267578125 [2023-11-25 22:04:19,677][root][INFO] - Train batch 16200 [2023-11-25 22:04:19,678][root][INFO] - Avg. loss per last 100 batches: 0.126669 [2023-11-25 22:04:20,388][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.8/29522=99.93% | mean: 0.02 | max: 5.87 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: if a club is promoted back to the premier league during the parachute payment period, ### ### then it no longer receives parachute payments. ### ### [P_TEXT]: premier league revenues of £2. 4 billion in 2016 – 17, with a further £343 million in ### ### solidarity payments to english football league ( efl ) clubs. the premier league is the most - ### ### watched sports league in the world, broadcast in 212 territories to 643 million homes and a ### ### potential tv audience of 4. 7 billion people. for the 2018 – 19 season average premier league match ### ### attendance was at 38, 181, second to the bundesliga's 43, 500, while aggregated attendance across ### ### all matches is the highest of any league at 14, 508, 981. most stadium occupancies are near ### ### capacity. the premier league ranks second in the uefa coefficient s of leagues based on ### ### performances in european competitions over the past five seasons as of 2021, ### ### =============================================== v_q ================================================ ### ### ('premier', 0, 1) ('parachute', 1, 9312) ('貴', 2, 417) ('intently', 3, 753) ('absently', 4, 47) ### ### ('league', 5, 3) ('¹', 6, 89) ('幸', 7, 202) ('খ', 8, 197) ('payment', 9, 28) ('.', 10, 110) ### ### ('ம', 11, 154) ('##ப', 12, 137) ('anyway', 13, 234) ('##部', 14, 177) ('³', 15, 914) ### ### ('worried', 16, 93) ('clubs', 17, 29) ('if', 18, 1752) ('promotion', 19, 804) ### ### =============================================== v_p ================================================ ### ### ('coefficient', 3706, 0) ('premier', 0, 1) ('bundesliga', 781, 2) ('league', 5, 3) ### ### ('solidarity', 11668, 4) ('212', 14853, 5) ('territories', 1852, 6) ('payments', 68, 7) ### ### ('revenue', 604, 8) ('ranks', 4073, 9) ('##43', 9952, 10) ('aggregate', 457, 11) ### ### ('leaned', 2998, 12) ('shirt', 51, 13) ('2021', 754, 14) ('sports', 353, 15) ('billion', 4647, 16) ### ### ('##cies', 11380, 17) ('homes', 10242, 18) ('broadcasts', 14426, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('premier', 0, 1) ('league', 5, 3) ('payment', 9, 28) ('absently', 4, 47) ('clubs', 17, 29) ### ### ('¹', 6, 89) ('payments', 68, 7) ('.', 10, 110) ('shirt', 51, 13) ('##ப', 12, 137) ('ம', 11, 154) ### ### ('worried', 16, 93) ('leagues', 25, 42) ('খ', 8, 197) ('幸', 7, 202) ('season', 22, 72) ### ### ('stared', 21, 73) ('##部', 14, 177) ('貴', 2, 417) ('anyway', 13, 234) ### ############################################################################################################ [2023-11-25 22:04:20,898][root][INFO] - Epoch: 4: Step: 16201/19530, loss[v]=0.086821, lr=0.000006, acc@1[1]=1001.5/1024=0.97802734375, acc@1[2]=1012.5/1024=0.98876953125 [2023-11-25 22:06:10,646][root][INFO] - Train batch 16300 [2023-11-25 22:06:10,647][root][INFO] - Avg. loss per last 100 batches: 0.134941 [2023-11-25 22:06:11,365][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 3.80 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.6/29522=99.93% | mean: 0.02 | max: 5.66 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in january 2007, entrepreneur warren alpert donated $ 100 million to brown medical ### ### school, tying sidney frank for the largest single monetary contribution ever made to the ### ### university. ### ### [P_TEXT]: history of brown university brown re - established its medical school, which had been ### ### suspended since 1827. in 2007, the school was renamed the warren alpert medical school of brown ### ### university in honor of benefactor warren alpert. in august of 2011, the alpert medical school ### ### opened a new campus in providence's jewelry district. in 2010, brown established its division of ### ### engineering as a new school of engineering. in 2013, brown transitioned the alpert medical school's ### ### department of community health into an independent school of public health. the current president ### ### of the university ### ### =============================================== v_q ================================================ ### ### ('brown', 0, 2) ('sidney', 1, 13954) ('university', 2, 19) ('warren', 3, 3) ('frank', 4, 18872) ### ### ('¹', 5, 192) ('##per', 6, 6) ('medical', 7, 10) ('al', 8, 13) ('intently', 9, 164) ('³', 10, 2446) ### ### ('absently', 11, 72) ('幸', 12, 167) ('##ப', 13, 112) ('খ', 14, 173) ('貴', 15, 170) ('ம', 16, 102) ### ### ('##t', 17, 36) ('semester', 18, 22) ('school', 19, 21) ### ### =============================================== v_p ================================================ ### ### ('jewelry', 6451, 0) ('providence', 736, 1) ('brown', 0, 2) ('warren', 3, 3) ### ### ('engineering', 543, 4) ('littered', 5024, 5) ('##per', 6, 6) ('health', 37, 7) ### ### ('benefactor', 268, 8) ('history', 3258, 9) ('medical', 7, 10) ('community', 2307, 11) ### ### ('jewellery', 20989, 12) ('al', 8, 13) ('1827', 20330, 14) ('anyway', 86, 15) ### ### ('suspended', 5060, 16) ('independent', 11282, 17) ('worried', 197, 18) ('university', 2, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('brown', 0, 2) ('warren', 3, 3) ('##per', 6, 6) ('medical', 7, 10) ('university', 2, 19) ### ### ('al', 8, 13) ('health', 37, 7) ('school', 19, 21) ('semester', 18, 22) ('campus', 24, 27) ### ### ('##t', 17, 36) ('absently', 11, 72) ('anyway', 86, 15) ('medicine', 21, 52) ('benefactor', 268, 8) ### ### ('¹', 5, 192) ('##ப', 13, 112) ('providence', 736, 1) ('engineering', 543, 4) ### ### ('graduation', 32, 50) ### ############################################################################################################ [2023-11-25 22:06:11,872][root][INFO] - Epoch: 4: Step: 16301/19530, loss[v]=0.181652, lr=0.000006, acc@1[1]=973.0/1024=0.9501953125, acc@1[2]=998.5/1024=0.97509765625 [2023-11-25 22:08:00,814][root][INFO] - Train batch 16400 [2023-11-25 22:08:00,815][root][INFO] - Avg. loss per last 100 batches: 0.123965 [2023-11-25 22:08:01,519][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.5/29522=99.92% | mean: 0.02 | max: 5.76 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.8/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the primary members of this group are pantalone, il dottore and il capitano. ### ### [P_TEXT]: il dottore il dottore (, " the doctor " ; commonly known in italian as " dottor balan " ### ### or simply " balanzone " ; bolognese ) is a " commedia dell'arte " stock character, one of the " ### ### vecchi ", or " old men ", whose function in a scenario is to be an obstacle to the young lovers. il ### ### dottore and pantalone are the comic foil of each other, pantalone being the decadent wealthy ### ### merchant, and il dottore being the decadent erudite. he has been part of the main canon of ### ### characters ### ### =============================================== v_q ================================================ ### ### ('dot', 0, 0) ('##tal', 1, 13) ('capita', 2, 9944) ('pan', 3, 14) ('il', 4, 9) ('##tore', 5, 4) ### ### ('##one', 6, 32) ('貴', 7, 285) ('幸', 8, 204) ('italian', 9, 23) ('¹', 10, 104) ### ### ('intently', 11, 370) ('খ', 12, 226) ('maltese', 13, 3820) ('absently', 14, 160) ('anyway', 15, 26) ### ### ('##no', 16, 11882) ('ம', 17, 150) ('##ப', 18, 132) ('savoy', 19, 661) ### ### =============================================== v_p ================================================ ### ### ('dot', 0, 0) ('doctor', 1330, 1) ('##nzo', 2083, 2) ('decade', 4527, 3) ('##tore', 5, 4) ### ### ('canon', 7685, 5) ('stock', 3055, 6) ('##cchi', 6546, 7) ('bala', 17190, 8) ('il', 4, 9) ### ### ('##udi', 22526, 10) ('arte', 13448, 11) ('old', 19769, 12) ('##tal', 1, 13) ('pan', 3, 14) ### ### ('##media', 13238, 15) ('worried', 40, 16) ('leaned', 1023, 17) ('##log', 22530, 18) ### ### ('##nese', 1137, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('dot', 0, 0) ('##tore', 5, 4) ('##tal', 1, 13) ('il', 4, 9) ('pan', 3, 14) ('italian', 9, 23) ### ### ('anyway', 15, 26) ('##one', 6, 32) ('worried', 40, 16) ('italy', 22, 59) ('¹', 10, 104) ### ### ('venetian', 96, 31) ('.', 21, 98) ('##ப', 18, 132) ('ம', 17, 150) ('absently', 14, 160) ### ### ('幸', 8, 204) ('貴', 7, 285) ('doctor', 1330, 1) ('খ', 12, 226) ### ############################################################################################################ [2023-11-25 22:08:02,023][root][INFO] - Epoch: 4: Step: 16401/19530, loss[v]=0.088943, lr=0.000006, acc@1[1]=997.0/1024=0.9736328125, acc@1[2]=1013.5/1024=0.98974609375 [2023-11-25 22:09:51,195][root][INFO] - Train batch 16500 [2023-11-25 22:09:51,196][root][INFO] - Avg. loss per last 100 batches: 0.134967 [2023-11-25 22:09:51,922][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.1/29522=99.94% | mean: 0.03 | max: 5.91 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: pratap made his directorial debut with " current " ( 2009 ) and later directed " kumari ### ### 21f " ( 2015 ). ### ### [P_TEXT]: kumari 21f hyderabad and the inspector plans to close the case because he could not find ### ### the trio. it is revealed that siddhu has chained them in the ruins and has been feeding and ### ### torturing them for the past three years ; they beg him to kill them. the film ends with siddhu ### ### beating the trio after feeding them and saying he is not mature enough to forgive them. in october ### ### 2014, sukumar announced he would co - produce a film along with vijay bandreddi and thomas reddy ### ### under the banner p. a. motion pictures. palnati surya pratap, who made his directorial debut ### ### =============================================== v_q ================================================ ### ### ('current', 0, 6010) ('21', 1, 2) ('¹', 2, 156) ('##f', 3, 8) ('##ata', 4, 12) ('kumar', 5, 3) ### ### ('pr', 6, 39) ('film', 7, 41) ('幸', 8, 221) ('貴', 9, 249) ('intently', 10, 371) ('খ', 11, 168) ### ### ('##p', 12, 49) ('##ப', 13, 96) ('ம', 14, 114) ('absently', 15, 135) ('³', 16, 2062) ### ### ('crore', 17, 4943) ('2009', 18, 2962) ('directed', 19, 361) ### ### =============================================== v_p ================================================ ### ### ('reddy', 3047, 0) ('##red', 22833, 1) ('21', 1, 2) ('kumar', 5, 3) ('inspector', 543, 4) ### ### ('##nat', 12892, 5) ('sid', 9959, 6) ('hyderabad', 59, 7) ('##f', 3, 8) ('motion', 2253, 9) ### ### ('thomas', 14610, 10) ('vijay', 209, 11) ('##ata', 4, 12) ('feeding', 21634, 13) ### ### ('worried', 77, 14) ('p', 2248, 15) ('2014', 433, 16) ('littered', 174, 17) ('chained', 6314, 18) ### ### ('directorial', 31, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('21', 1, 2) ('kumar', 5, 3) ('##f', 3, 8) ('##ata', 4, 12) ('pr', 6, 39) ('film', 7, 41) ### ### ('##p', 12, 49) ('directorial', 31, 19) ('₹', 28, 23) ('hyderabad', 59, 7) ('¹', 2, 156) ### ### ('anyway', 29, 27) ('##ப', 13, 96) ('worried', 77, 14) ('##i', 23, 44) ('ம', 14, 114) ### ### ('absently', 15, 135) ('খ', 11, 168) ('##ि', 41, 36) ('幸', 8, 221) ### ############################################################################################################ [2023-11-25 22:09:52,428][root][INFO] - Epoch: 4: Step: 16501/19530, loss[v]=0.079387, lr=0.000006, acc@1[1]=1001.5/1024=0.97802734375, acc@1[2]=1015.0/1024=0.9912109375 [2023-11-25 22:11:41,786][root][INFO] - Train batch 16600 [2023-11-25 22:11:41,787][root][INFO] - Avg. loss per last 100 batches: 0.123201 [2023-11-25 22:11:42,526][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.6/29522=99.94% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: ranked third in the short program and second in the free skate, they finished first ### ### overall, outscoring the silver medallists, aleksandra boikova / dmitrii kozlovskii of russia, by 2. ### ### 05 points. ### ### [P_TEXT]: aleksandra boikova, france. they won the bronze medal with a personal best score of 159. ### ### 72. later in the same month, they finished sixth competing on the senior level at the russian ### ### championships. they won the junior national title in february 2017. in march, boikova / kozlovskii ### ### won silver at the 2017 world junior championships in taipei, taiwan. ranked first in the short ### ### program and fourth in the free skate, they finished second to australia's ekaterina alexandrovskaya ### ### / harley windsor with a deficit of 2. 05 points. boikova / kozlovskii won two medals on ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 171) ('aleksandr', 1, 1) ('幸', 2, 137) ('খ', 3, 133) ('貴', 4, 132) ('##lov', 5, 7) ### ### ('¹', 6, 62) ('intently', 7, 122) ('##ப', 8, 87) ('ɐ', 9, 201) ('ம', 10, 88) ('##iko', 11, 3) ### ### ('##部', 12, 109) ('dmitri', 13, 5584) ('skate', 14, 24) ('placed', 15, 120) ('bo', 16, 29) ### ### ('³', 17, 739) ('skating', 18, 17) ('anyway', 19, 16) ### ### =============================================== v_p ================================================ ### ### ('harley', 18009, 0) ('aleksandr', 1, 1) ('windsor', 11278, 2) ('##iko', 11, 3) ('2017', 259, 4) ### ### ('junior', 528, 5) ('championships', 52, 6) ('##lov', 5, 7) ('taiwan', 4700, 8) ### ### ('throws', 10775, 9) ('##skaya', 6006, 10) ('australia', 3978, 11) ('leaned', 1598, 12) ### ### ('taipei', 6087, 13) ('russian', 97, 14) ('entries', 9246, 15) ('anyway', 19, 16) ### ### ('skating', 18, 17) ('ek', 2999, 18) ('skaters', 829, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('aleksandr', 1, 1) ('##lov', 5, 7) ('##iko', 11, 3) ('skating', 18, 17) ('anyway', 19, 16) ### ### ('skate', 14, 24) ('bo', 16, 29) ('##ski', 23, 23) ('championships', 52, 6) ('free', 25, 31) ### ### ('russia', 40, 20) ('¹', 6, 62) ('points', 29, 34) ('russian', 97, 14) ('##ப', 8, 87) ### ### ('ko', 30, 41) ('ம', 10, 88) ('hissed', 47, 33) ('short', 51, 32) ('2017', 259, 4) ### ############################################################################################################ [2023-11-25 22:11:43,031][root][INFO] - Epoch: 4: Step: 16601/19530, loss[v]=0.082078, lr=0.000006, acc@1[1]=999.0/1024=0.9755859375, acc@1[2]=1015.5/1024=0.99169921875 [2023-11-25 22:13:31,792][root][INFO] - Train batch 16700 [2023-11-25 22:13:31,792][root][INFO] - Avg. loss per last 100 batches: 0.126918 [2023-11-25 22:13:32,510][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.3/29522=99.94% | mean: 0.02 | max: 5.87 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: reaction on condensed tannins from douglas fir bark produces epi catechin and catechin ### ### thioglycolates. ### ### [P_TEXT]: catechin 2, 3 - cis - epicatechin or ( 2 " r ", 3 " r " ) - ( - ) - epicatechin ). the ### ### different epimers can be distinguished using chiral column chromatography. making reference to no ### ### particular isomer, the molecule can just be called catechin. mixtures of the different enantiomers ### ### can be called ( + / - ) - catechin or dl - catechin and ( + / - ) - epicatechin or dl - ### ### epicatechin. catechin and epicatechin are the building blocks of the proanthocyanidin s, a type of ### ### condensed tannin. moreover, the flexibility of the c - ring allows for two conformation isomers, ### ### putting the b ring either in a pseudoequatorial position ( e conformer ) or in a pseudoaxial ### ### position ( a conformer ). studies confirmed that ( + ) - catechin adopts a ### ### =============================================== v_q ================================================ ### ### ('##ech', 0, 0) ('fir', 1, 22475) ('ep', 2, 19) ('douglas', 3, 16376) ('cat', 4, 6) ('¹', 5, 131) ### ### ('幸', 6, 157) ('貴', 7, 199) ('tan', 8, 5) ('খ', 9, 165) ('intently', 10, 227) ('##ப', 11, 126) ### ### ('bark', 12, 11200) ('##cola', 13, 21276) ('##in', 14, 18) ('ம', 15, 137) ('absently', 16, 196) ### ### ('condensed', 17, 3) ('##gly', 18, 24670) ('.', 19, 4555) ### ### =============================================== v_p ================================================ ### ### ('##ech', 0, 0) ('conform', 5282, 1) ('##chin', 9038, 2) ('condensed', 17, 3) ('##din', 3243, 4) ### ### ('tan', 8, 5) ('cat', 4, 6) ('iso', 5028, 7) ('flexibility', 2505, 8) ('column', 3644, 9) ### ### ('pseudo', 5524, 10) ('##nin', 24, 11) ('dl', 18728, 12) ('ring', 2257, 13) ('epic', 13773, 14) ### ### ('##tori', 21783, 15) ('molecules', 2793, 16) ('##ant', 3239, 17) ('##in', 14, 18) ('ep', 2, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ech', 0, 0) ('cat', 4, 6) ('tan', 8, 5) ('condensed', 17, 3) ('ep', 2, 19) ('##in', 14, 18) ### ### ('##nin', 24, 11) ('anyway', 26, 44) ('¹', 5, 131) ('幸', 6, 157) ('##ப', 11, 126) ('খ', 9, 165) ### ### ('ம', 15, 137) ('worried', 31, 57) ('貴', 7, 199) ('intently', 10, 227) ('##部', 23, 112) ### ### ('absently', 16, 196) ('hissed', 97, 42) ('##₈', 20, 208) ### ############################################################################################################ [2023-11-25 22:13:33,014][root][INFO] - Epoch: 4: Step: 16701/19530, loss[v]=0.060624, lr=0.000006, acc@1[1]=997.5/1024=0.97412109375, acc@1[2]=1018.5/1024=0.99462890625 [2023-11-25 22:15:22,035][root][INFO] - Train batch 16800 [2023-11-25 22:15:22,036][root][INFO] - Avg. loss per last 100 batches: 0.131742 [2023-11-25 22:15:22,765][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.94 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29505.3/29522=99.94% | mean: 0.03 | max: 5.89 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: md 201 intersects riverdale road and md 410 ( east – west highway ), then follows the ### ### eastern boundary of riverdale park, intersecting md 594e ( river road ). ### ### [P_TEXT]: maryland route 594 from us 1 in college park east to md 201 in riverdale park in ### ### northwestern prince george's county. md 594f runs from md 410 east to md 450 in new carrollton in ### ### northwestern prince george's county. the three auxiliary routes are designated md 594b, md 594c, ### ### and md 594d, and are located in montgomery county. the sections of the route in montgomery county ### ### were transferred to the state in 2016 while the sections of the route in prince george's county ### ### were transferred to the state in ### ### =============================================== v_q ================================================ ### ### ('201', 0, 1) ('maryland', 1, 9) ('410', 2, 0) ('##dale', 3, 7) ('¹', 4, 52) ('river', 5, 19) ### ### ('59', 6, 3) ('route', 7, 20) ('driveway', 8, 23) ('baltimore', 9, 58) ('##ப', 10, 75) ### ### ('幸', 11, 87) ('absently', 12, 90) ('貴', 13, 88) ('খ', 14, 107) ('park', 15, 11) ('ம', 16, 57) ### ### ('³', 17, 248) ('intently', 18, 62) ('##4', 19, 10) ### ### =============================================== v_p ================================================ ### ### ('410', 2, 0) ('201', 0, 1) ('carroll', 8959, 2) ('59', 6, 3) ('md', 37, 4) ('450', 2807, 5) ### ### ('college', 4700, 6) ('##dale', 3, 7) ('montgomery', 230, 8) ('maryland', 1, 9) ('##4', 19, 10) ### ### ('park', 15, 11) ('1', 6832, 12) ('prince', 10623, 13) ('worried', 209, 14) ('anyway', 378, 15) ### ### ('##ton', 8718, 16) ('george', 4793, 17) ('auxiliary', 4326, 18) ('river', 5, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('201', 0, 1) ('410', 2, 0) ('maryland', 1, 9) ('##dale', 3, 7) ('59', 6, 3) ('river', 5, 19) ### ### ('park', 15, 11) ('##4', 19, 10) ('route', 7, 20) ('driveway', 8, 23) ('md', 37, 4) ('¹', 4, 52) ### ### ('baltimore', 9, 58) ('##ப', 10, 75) ('幸', 11, 87) ('montgomery', 230, 8) ('ம', 16, 57) ### ### ('routes', 57, 21) ('intently', 18, 62) ('absently', 12, 90) ### ############################################################################################################ [2023-11-25 22:15:23,271][root][INFO] - Epoch: 4: Step: 16801/19530, loss[v]=0.141595, lr=0.000006, acc@1[1]=997.0/1024=0.9736328125, acc@1[2]=1011.5/1024=0.98779296875 [2023-11-25 22:17:12,496][root][INFO] - Train batch 16900 [2023-11-25 22:17:12,496][root][INFO] - Avg. loss per last 100 batches: 0.120775 [2023-11-25 22:17:13,212][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.9/29522=99.94% | mean: 0.02 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: as a reflection of its cultural wealth, italy is home to the world's largest number of ### ### world heritage site s ( 55 ), and is the fifth - most visited country. ### ### [P_TEXT]: national language co - official language and became recognized as a language with " ### ### special status " used in state institutions. the italian language is the de jure and de facto ### ### official language of italy. italian is also referred to as national language for historical and ### ### cultural reasons, because since the 15th century, italian became the language used in the courts of ### ### nearly every state in italy and in general among educated italians ( scholars, writers, poets, ### ### philosophers, scientists, composers and artists ) who contributed to what is nowadays the culture ### ### of italy. furthermore, italian was often an official language of the various italian states before ### ### unification, slowly ### ### =============================================== v_q ================================================ ### ### ('italy', 0, 3) ('site', 1, 9700) ('貴', 2, 288) ('¹', 3, 305) ('world', 4, 12407) ('ம', 5, 136) ### ### ('幸', 6, 213) ('খ', 7, 241) ('absently', 8, 112) ('intently', 9, 355) ('heritage', 10, 2511) ### ### ('italian', 11, 2) ('##ப', 12, 181) ('visited', 13, 3684) ('cultural', 14, 29) ('sites', 15, 3683) ### ### ('.', 16, 77) ('italia', 17, 1533) ('wealth', 18, 1948) ('worried', 19, 16) ### ### =============================================== v_p ================================================ ### ### ('national', 3344, 0) ('language', 1505, 1) ('italian', 11, 2) ('italy', 0, 3) ('italians', 213, 4) ### ### ('regimes', 15686, 5) ('languages', 2683, 6) ('leaned', 2122, 7) ('culture', 31, 8) ### ### ('gradually', 1900, 9) ('philosophers', 25045, 10) ('composers', 8303, 11) ('persisted', 4254, 12) ### ### ('15th', 2765, 13) ('scholars', 8133, 14) ('unification', 1093, 15) ('worried', 19, 16) ### ### ('cultures', 3705, 17) ('nationals', 4293, 18) ('ju', 21396, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('italy', 0, 3) ('italian', 11, 2) ('worried', 19, 16) ('cultural', 14, 29) ('culture', 31, 8) ### ### ('countries', 20, 39) ('anyway', 29, 23) ('.', 16, 77) ('absently', 8, 112) ('ம', 5, 136) ### ### ('italians', 213, 4) ('貴', 2, 288) ('幸', 6, 213) ('varied', 72, 25) ('##ப', 12, 181) ('খ', 7, 241) ### ### ('¹', 3, 305) ('intently', 9, 355) ('language', 1505, 1) ('##部', 22, 187) ### ############################################################################################################ [2023-11-25 22:17:13,717][root][INFO] - Epoch: 4: Step: 16901/19530, loss[v]=0.170886, lr=0.000006, acc@1[1]=979.0/1024=0.9560546875, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-25 22:19:02,335][root][INFO] - Train batch 17000 [2023-11-25 22:19:02,336][root][INFO] - Avg. loss per last 100 batches: 0.127598 [2023-11-25 22:19:03,044][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 4.02 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29505.7/29522=99.94% | mean: 0.03 | max: 5.89 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 26.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he resigned under pressure of mass demonstrations known as rose revolution on 23 november ### ### 2003. ### ### [P_TEXT]: giorgi kvirikashvili in georgia from 1993 to 1999 and as a deputy head of fiscal and ### ### monetary office at the state chancellery of the president of georgia in 1999. from 1999 to 2004, he ### ### was member of the parliament of georgia on the ticket of union of citizens of georgia party. after ### ### the rose revolution swept mikheil saakashvili to the presidency of georgia, kvirikashvili returned ### ### to his businesses. from 2006 to 2011, he was director general of cartu bank, owned by the multi - ### ### billionaire tycoon bidzina ivanishvili. with ivanishvili's coming in politics ### ### =============================================== v_q ================================================ ### ### ('rose', 0, 2) ('revolution', 1, 5) ('¹', 2, 244) ('2003', 3, 662) ('！', 4, 1062) ('貴', 5, 185) ### ### ('resign', 6, 737) ('resignation', 7, 1001) ('littered', 8, 15) ('幸', 9, 176) ('##ʂ', 10, 2427) ### ### ('intently', 11, 152) ('.', 12, 444) ('ɐ', 13, 1913) ('খ', 14, 228) ('³', 15, 1238) ('ம', 16, 135) ### ### ('anyway', 17, 12) ('resigning', 18, 2170) ('##¨', 19, 1515) ### ### =============================================== v_p ================================================ ### ### ('chancel', 20643, 0) ('cart', 22091, 1) ('rose', 0, 2) ('georgia', 9387, 3) ('citizens', 973, 4) ### ### ('revolution', 1, 5) ('fiscal', 7779, 6) ('bid', 8971, 7) ('parliament', 2789, 8) ### ### ('president', 40, 9) ('worried', 144, 10) ('##org', 26533, 11) ('anyway', 17, 12) ### ### ('##vili', 5633, 13) ('##kas', 13822, 14) ('littered', 8, 15) ('georgian', 1143, 16) ### ### ('##eil', 27618, 17) ('##iri', 14102, 18) ('bank', 1185, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('rose', 0, 2) ('revolution', 1, 5) ('littered', 8, 15) ('anyway', 17, 12) ('president', 40, 9) ### ### ('protests', 22, 37) ('worried', 144, 10) ('貴', 5, 185) ('2004', 135, 34) ('1999', 397, 20) ### ### ('¹', 2, 244) ('citizens', 973, 4) ('intently', 11, 152) ('ம', 16, 135) ('幸', 9, 176) ### ### ('coup', 43, 77) ('presidential', 53, 74) ('georgian', 1143, 16) ('##ப', 20, 173) ('খ', 14, 228) ### ############################################################################################################ [2023-11-25 22:19:03,551][root][INFO] - Epoch: 4: Step: 17001/19530, loss[v]=0.197995, lr=0.000006, acc@1[1]=972.0/1024=0.94921875, acc@1[2]=1005.5/1024=0.98193359375 [2023-11-25 22:20:52,764][root][INFO] - Train batch 17100 [2023-11-25 22:20:52,764][root][INFO] - Avg. loss per last 100 batches: 0.128558 [2023-11-25 22:20:53,494][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 3.94 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.1/29522=99.93% | mean: 0.02 | max: 5.72 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: including four reigns as pwg world tag team champions as well as being the only team to ### ### win pwg's annual dynamite duumvirate tag team title tournament on three occasions ( 2009, 2011, and ### ### 2013 ). ### ### [P_TEXT]: trent beretta and taylor defeated three other teams, including the inner city machine ### ### guns ( rich swann and ricochet ) in the finals, to win pwg's 2014 dynamite duumvirate tag team ### ### title tournament ( ddt4 ). as a result of winning the 2014 ddt4, trent? and chuck taylor received a ### ### shot at the pwg world tag team championship on march 28, but were defeated by the defending ### ### champions, the young bucks ( matt and nick jackson ). in march 2016 during pwg's all star weekend ### ### 12, trent? picked up his first singles win against drew galloway, then he ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 138) ('幸', 1, 94) ('p', 2, 15) ('intently', 3, 101) ('貴', 4, 97) ('খ', 5, 118) ### ### ('##ப', 6, 104) ('##wg', 7, 5) ('ம', 8, 80) ('dynamite', 9, 6) ('¹', 10, 26) ('³', 11, 395) ### ### ('tag', 12, 22) ('whoever', 13, 399) ('freaking', 14, 173) ('.', 15, 85) ('！', 16, 749) ### ### ('2009', 17, 9621) ('##部', 18, 124) ('anyway', 19, 20) ### ### =============================================== v_p ================================================ ### ### ('trent', 8758, 0) ('##ret', 19199, 1) ('swan', 13553, 2) ('bucks', 536, 3) ('##chet', 15221, 4) ### ### ('##wg', 7, 5) ('dynamite', 9, 6) ('taylor', 676, 7) ('dd', 3334, 8) ('galloway', 1213, 9) ### ### ('inner', 16020, 10) ('jackson', 1830, 11) ('rico', 335, 12) ('weekend', 3387, 13) ('matt', 44, 14) ### ### ('p', 2, 15) ('rich', 13781, 16) ('guns', 5802, 17) ('chuck', 791, 18) ('nick', 666, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##wg', 7, 5) ('dynamite', 9, 6) ('p', 2, 15) ('tag', 12, 22) ('¹', 10, 26) ('anyway', 19, 20) ### ### ('matt', 44, 14) ('##vira', 32, 21) ('幸', 1, 94) ('ம', 8, 80) ('貴', 4, 97) ('absently', 0, 138) ### ### ('intently', 3, 101) ('tournament', 31, 32) ('##ப', 6, 104) ('wrestling', 27, 41) ('খ', 5, 118) ### ### ('worried', 46, 27) ('match', 30, 50) ('.', 15, 85) ### ############################################################################################################ [2023-11-25 22:20:54,003][root][INFO] - Epoch: 4: Step: 17101/19530, loss[v]=0.146912, lr=0.000006, acc@1[1]=991.5/1024=0.96826171875, acc@1[2]=1011.5/1024=0.98779296875 [2023-11-25 22:22:43,019][root][INFO] - Train batch 17200 [2023-11-25 22:22:43,019][root][INFO] - Avg. loss per last 100 batches: 0.127931 [2023-11-25 22:22:43,723][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.9/29522=99.94% | mean: 0.02 | max: 5.84 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the brother of saint apronia ( evronie ), saint aprus was born near trier. ### ### [P_TEXT]: apronia of toul apronia of toul, also called evronie of troyes, was a nun and saint of ### ### the 6th century. she was born at tranquille, a village in trier, germany. her brother was saint ### ### aprus of toul, a bishop in toul in northeastern france, from whom she received the veil. ### ### hagiographer sabine baring - gould said of apronia : " drawn by her love of christ to a religious ### ### life, she led on earth a virginal and angelic life, in imitation of her brother, a man of the ### ### highest sanctity. during her life ### ### =============================================== v_q ================================================ ### ### ('apr', 0, 5) ('apron', 1, 0) ('¹', 2, 475) ('##ron', 3, 21) ('saint', 4, 29) ('ev', 5, 9) ### ### ('trier', 6, 2) ('幸', 7, 263) ('খ', 8, 245) ('##us', 9, 105) ('³', 10, 5422) ('intently', 11, 378) ### ### ('貴', 12, 254) ('absently', 13, 114) ('##ᵉ', 14, 1344) ('##ப', 15, 209) ('sparhawk', 16, 2063) ### ### ('##ia', 17, 31) ('ம', 18, 165) ('frankish', 19, 285) ### ### =============================================== v_p ================================================ ### ### ('apron', 1, 0) ('##ul', 14423, 1) ('trier', 6, 2) ('gould', 14222, 3) ('tran', 20666, 4) ### ### ('apr', 0, 5) ('sabine', 225, 6) ('veil', 1541, 7) ('troy', 105, 8) ('ev', 5, 9) ### ### ('virgin', 218, 10) ('ט', 10798, 11) ('bari', 5437, 12) ('##quil', 11624, 13) ('worried', 912, 14) ### ### ('imitation', 6267, 15) ('##gio', 906, 16) ('france', 59, 17) ('leaned', 6205, 18) ### ### ('##le', 16761, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('apron', 1, 0) ('apr', 0, 5) ('trier', 6, 2) ('ev', 5, 9) ('##ron', 3, 21) ('saint', 4, 29) ### ### ('##ia', 17, 31) ('saints', 26, 26) ('brother', 25, 33) ('troy', 105, 8) ('france', 59, 17) ### ### ('sabine', 225, 6) ('##us', 9, 105) ('virgin', 218, 10) ('absently', 13, 114) ('bishop', 52, 36) ### ### ('##ie', 22, 82) ('anyway', 114, 24) ('mainz', 49, 49) ('ம', 18, 165) ### ############################################################################################################ [2023-11-25 22:22:44,226][root][INFO] - Epoch: 4: Step: 17201/19530, loss[v]=0.093476, lr=0.000006, acc@1[1]=996.5/1024=0.97314453125, acc@1[2]=1016.0/1024=0.9921875 [2023-11-25 22:24:32,931][root][INFO] - Train batch 17300 [2023-11-25 22:24:32,932][root][INFO] - Avg. loss per last 100 batches: 0.128885 [2023-11-25 22:24:33,645][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.8/29522=99.94% | mean: 0.02 | max: 5.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: line 1 of the zhengzhou metro ( ) is a rapid transit line running from west to east ### ### zhengzhou. ### ### [P_TEXT]: xisanhuan station xisanhuan ( ) is a metro station of zhengzhou metro line 1. the station ### ### is named after the west 3rd ring road ( xisanhuan in pinyin ). since the road is over long, it is ### ### hard to locate the station according to its name. the naming of the station is considered to be ### ### confusing and thus criticized by some people on local forums. the station has 2 floors underground. ### ### the b1 floor is for the station concourse and the b2 floor is for the platforms and tracks. the ### ### station has one island platform and two tracks for line 1. ### ### =============================================== v_q ================================================ ### ### ('zheng', 0, 0) ('metro', 1, 9) ('¹', 2, 392) ('##zhou', 3, 13) ('1', 4, 6) ('henan', 5, 67) ### ### ('##₁', 6, 21) ('幸', 7, 141) ('³', 8, 2908) ('খ', 9, 156) ('貴', 10, 125) ('line', 11, 8) ### ### ('ம', 12, 86) ('china', 13, 31) ('station', 14, 34) ('##ப', 15, 101) ('beijing', 16, 102) ### ### ('##部', 17, 103) ('chinese', 18, 26) ('cheng', 19, 405) ### ### =============================================== v_p ================================================ ### ### ('zheng', 0, 0) ('forums', 25015, 1) ('##hua', 587, 2) ('b1', 3567, 3) ('##san', 17563, 4) ### ### ('concourse', 544, 5) ('1', 4, 6) ('ring', 3159, 7) ('line', 11, 8) ('metro', 1, 9) ### ### ('3rd', 4247, 10) ('worried', 1138, 11) ('pinyin', 357, 12) ('##zhou', 3, 13) ('floor', 267, 14) ### ### ('xi', 126, 15) ('lines', 78, 16) ('stations', 96, 17) ('road', 394, 18) ('anyway', 196, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('zheng', 0, 0) ('metro', 1, 9) ('1', 4, 6) ('##zhou', 3, 13) ('line', 11, 8) ('##₁', 6, 21) ### ### ('china', 13, 31) ('henan', 5, 67) ('station', 14, 34) ('chinese', 18, 26) ('lines', 78, 16) ### ### ('xi', 126, 15) ('stations', 96, 17) ('幸', 7, 141) ('ம', 12, 86) ('貴', 10, 125) ('¹', 2, 392) ### ### ('খ', 9, 156) ('##hua', 587, 2) ('floor', 267, 14) ### ############################################################################################################ [2023-11-25 22:24:34,151][root][INFO] - Epoch: 4: Step: 17301/19530, loss[v]=0.131862, lr=0.000006, acc@1[1]=984.0/1024=0.9609375, acc@1[2]=1008.0/1024=0.984375 [2023-11-25 22:26:23,443][root][INFO] - Train batch 17400 [2023-11-25 22:26:23,443][root][INFO] - Avg. loss per last 100 batches: 0.125466 [2023-11-25 22:26:24,163][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 3.94 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29505.9/29522=99.95% | mean: 0.03 | max: 5.92 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: north from the water douglas passes the beacon lodge, then forms the western border of ### ### beacon hill park before passing behind the empress hotel and the royal british columbia museum. ### ### [P_TEXT]: beacon lodge the beacon lodge is a unique apartment building in victoria, british ### ### columbia, canada. located at 30 douglas street, the building is directly across from the start of ### ### the trans - canada highway, and just down the street from the famous beacon drive in. the building ### ### has 59 rental suites located on 4 separate floors, with the majority of them being small bachelors ### ### and a few one bedrooms. many suites have views of the olympic mountains across the juan de fuca ### ### strait in the pacific ocean and ### ### =============================================== v_q ================================================ ### ### ('douglas', 0, 4) ('empress', 1, 24854) ('beacon', 2, 0) ('columbia', 3, 49) ('museum', 4, 6323) ### ### ('hotel', 5, 173) ('park', 6, 6024) ('hill', 7, 2051) ('¹', 8, 466) ('lodge', 9, 1) ### ### ('vancouver', 10, 21) ('intently', 11, 232) ('##ப', 12, 143) ('british', 13, 246) ### ### ('water', 14, 863) ('খ', 15, 162) ('³', 16, 2809) ('幸', 17, 185) ('貴', 18, 194) ('yukon', 19, 186) ### ### =============================================== v_p ================================================ ### ### ('beacon', 2, 0) ('lodge', 9, 1) ('olympic', 10239, 2) ('victoria', 51, 3) ('douglas', 0, 4) ### ### ('strait', 1942, 5) ('bedrooms', 6034, 6) ('trans', 12316, 7) ('drive', 10229, 8) ### ### ('bachelor', 7238, 9) ('rental', 3822, 10) ('worried', 808, 11) ('suites', 12165, 12) ### ### ('highway', 58, 13) ('fu', 24240, 14) ('canada', 22, 15) ('juan', 6645, 16) ('mountains', 1402, 17) ### ### ('street', 538, 18) ('anyway', 1214, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('beacon', 2, 0) ('douglas', 0, 4) ('lodge', 9, 1) ('vancouver', 10, 21) ('columbia', 3, 49) ### ### ('canada', 22, 15) ('victoria', 51, 3) ('highway', 58, 13) ('hotel', 5, 173) ('canadian', 52, 44) ### ### ('worried', 808, 11) ('driveway', 233, 22) ('##ப', 12, 143) ('bc', 23, 89) ('streets', 174, 32) ### ### ('fraser', 31, 70) ('leaned', 344, 26) ('খ', 15, 162) ('street', 538, 18) ('strait', 1942, 5) ### ############################################################################################################ [2023-11-25 22:26:24,669][root][INFO] - Epoch: 4: Step: 17401/19530, loss[v]=0.068632, lr=0.000006, acc@1[1]=1002.5/1024=0.97900390625, acc@1[2]=1019.0/1024=0.9951171875 [2023-11-25 22:28:13,395][root][INFO] - Train batch 17500 [2023-11-25 22:28:13,395][root][INFO] - Avg. loss per last 100 batches: 0.133776 [2023-11-25 22:28:14,101][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29505.2/29522=99.94% | mean: 0.02 | max: 5.76 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: has a mouthpiece, shank s with both purchase and lever arms, a ring for rein attachment ### ### at the bottom of the shank, and a curb chain. ### ### [P_TEXT]: bit shank a rein ring at the bottom for the curb rein and a cheek ring at the top to ### ### attach the headstall. some shanks may also add rings or slots to attach a snaffle rein at the ### ### mouthpiece, allowing the bit to be used with two sets of reins, making it a pelham bit. some ### ### shanks, especially on the weymouth, have small rings placed midway down the shank to attach a lip ### ### strap, a helpful addition to the bit for preventing a horse from grabbing at the shanks with its ### ### lips. some shanks on western - style bits ### ### =============================================== v_q ================================================ ### ### ('shan', 0, 3) ('##k', 1, 30) ('chain', 2, 2582) ('mouth', 3, 20) ('curb', 4, 7) ('##piece', 5, 9) ### ### ('horse', 6, 19) ('intently', 7, 503) ('lever', 8, 2367) ('ring', 9, 12) ('purchase', 10, 5794) ### ### ('¹', 11, 262) ('rodeo', 12, 804) ('幸', 13, 211) ('貴', 14, 327) ('handle', 15, 1936) ### ### ('rider', 16, 232) ('weighs', 17, 13142) ('breed', 18, 181) ('accessory', 19, 10199) ### ### =============================================== v_p ================================================ ### ### ('cheek', 4773, 0) ('##lham', 13055, 1) ('bit', 3999, 2) ('shan', 0, 3) ('bits', 11903, 4) ### ### ('##ffle', 6091, 5) ('rein', 21, 6) ('curb', 4, 7) ('reins', 50, 8) ('##piece', 5, 9) ### ### ('lip', 4777, 10) ('midway', 11099, 11) ('ring', 9, 12) ('strap', 87, 13) ('horses', 148, 14) ### ### ('##outh', 16195, 15) ('lips', 1088, 16) ('cheeks', 10893, 17) ('heads', 7424, 18) ('horse', 6, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('shan', 0, 3) ('curb', 4, 7) ('##piece', 5, 9) ('mouth', 3, 20) ('horse', 6, 19) ('ring', 9, 12) ### ### ('rein', 21, 6) ('##k', 1, 30) ('reins', 50, 8) ('strap', 87, 13) ('horses', 148, 14) ### ### ('anyway', 49, 33) ('equestrian', 118, 29) ('fastened', 32, 80) ('riding', 59, 54) ### ### ('tack', 28, 107) ('breed', 18, 181) ('¹', 11, 262) ('lips', 1088, 16) ('absently', 27, 127) ### ############################################################################################################ [2023-11-25 22:28:14,605][root][INFO] - Epoch: 4: Step: 17501/19530, loss[v]=0.077666, lr=0.000006, acc@1[1]=1002.5/1024=0.97900390625, acc@1[2]=1016.0/1024=0.9921875 [2023-11-25 22:30:03,311][root][INFO] - Train batch 17600 [2023-11-25 22:30:03,312][root][INFO] - Avg. loss per last 100 batches: 0.124471 [2023-11-25 22:30:04,018][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.5/29522=99.93% | mean: 0.02 | max: 5.61 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.4/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he stated : in a press conference on december 19 held by both attorney general gonzales ### ### and general michael hayden, the principal deputy director for national intelligence, general hayden ### ### claimed, " this program has been successful in detecting and ### ### [P_TEXT]: robert deitz his former boss michael hayden. deitz's teaching at george mason was the ### ### subject of criticism from the intercept's ken silverstein, who argued that deitz was unfit to teach ### ### courses on ethical challenges in public policy on accord of his legal work in defense of nsa ### ### warrantless surveillance programs and what silverstein suggested was a politically motivated ### ### investigation deitz conducted into former cia inspector general john helgerson's review of the ### ### agency's extraordinary rendition, black site s, and enhanced interrogation practices at the behest ### ### of director michael hayden. in the 2016 presidential ### ### =============================================== v_q ================================================ ### ### ('hayden', 0, 0) ('gonzales', 1, 2375) ('absently', 2, 109) ('¹', 3, 114) ('intelligence', 4, 341) ### ### ('intently', 5, 154) ('michael', 6, 20) ('貴', 7, 137) ('幸', 8, 127) ('খ', 9, 120) ('##ப', 10, 97) ### ### ('general', 11, 31) ('national', 12, 15378) ('attorney', 13, 2386) ('ம', 14, 76) ('³', 15, 2273) ### ### ('freaking', 16, 559) ('.', 17, 1280) ('deputy', 18, 2939) ('##部', 19, 101) ### ### =============================================== v_p ================================================ ### ### ('hayden', 0, 0) ('##rson', 21003, 1) ('intercept', 492, 2) ('dei', 27912, 3) ('##lge', 26447, 4) ### ### ('site', 6629, 5) ('rendition', 351, 6) ('mason', 1396, 7) ('enhanced', 4673, 8) ('black', 6535, 9) ### ### ('##stein', 10752, 10) ('extraordinary', 1646, 11) ('surveillance', 26, 12) ('silver', 8275, 13) ### ### ('ken', 17684, 14) ('inspector', 3173, 15) ('##tz', 13646, 16) ('sites', 5877, 17) ### ### ('interrogation', 145, 18) ('warrant', 185, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('hayden', 0, 0) ('michael', 6, 20) ('general', 11, 31) ('surveillance', 26, 12) ('cia', 21, 23) ### ### ('director', 29, 26) ('interrogation', 145, 18) ('absently', 2, 109) ('rendition', 351, 6) ### ### ('intercept', 492, 2) ('anyway', 77, 27) ('warrant', 185, 19) ('¹', 3, 114) ('2016', 98, 30) ### ### ('ம', 14, 76) ('##ப', 10, 97) ('intently', 5, 154) ('খ', 9, 120) ('幸', 8, 127) ('貴', 7, 137) ### ############################################################################################################ [2023-11-25 22:30:04,525][root][INFO] - Epoch: 4: Step: 17601/19530, loss[v]=0.221390, lr=0.000005, acc@1[1]=970.5/1024=0.94775390625, acc@1[2]=994.5/1024=0.97119140625 [2023-11-25 22:31:53,879][root][INFO] - Train batch 17700 [2023-11-25 22:31:53,880][root][INFO] - Avg. loss per last 100 batches: 0.130149 [2023-11-25 22:31:54,589][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.3/29522=99.94% | mean: 0.02 | max: 5.84 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 77.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he was the son of john smoke johnson, a mohawk bear clan chief, and helen martin, a ### ### mohawk whose dutch mother catherine rolleston had been captured as a girl and adopted and ### ### assimilated into the wolf clan. ### ### [P_TEXT]: john smoke johnson been captured as a girl by the mohawk, adopted by a family in the wolf ### ### clan, and became assimilated as mohawk. the johnsons had seven children, who were considered to be ### ### born into the wolf clan : mary, aaron, joseph, william, margret, susannah and george henry martin ( ### ### " onwanonsyshon " ). johnson became a leader in canada. he fought for the british crown in the war ### ### of 1812 and was honored by his tribal council when chosen as a " pine tree chief ", a non - ### ### hereditary position. he was influential in the mohawk and british communities of ontario, canada. ### ### =============================================== v_q ================================================ ### ### ('johnson', 0, 3) ('smoke', 1, 0) ('martin', 2, 14) ('helen', 3, 9794) ('bear', 4, 4619) ### ### ('¹', 5, 226) ('wolf', 6, 1) ('bellowed', 7, 437) ('mohawk', 8, 2) ('³', 9, 3175) ('john', 10, 33) ### ### ('exclaimed', 11, 631) ('son', 12, 172) ('##est', 13, 26471) ('##ப', 14, 270) ('খ', 15, 277) ### ### ('absently', 16, 320) ('幸', 17, 299) ('littered', 18, 61) ('indians', 19, 561) ### ### =============================================== v_p ================================================ ### ### ('smoke', 1, 0) ('wolf', 6, 1) ('mohawk', 8, 2) ('johnson', 0, 3) ('susannah', 2958, 4) ### ### ('aaron', 4532, 5) ('clan', 32, 6) ('tribal', 498, 7) ('assimilation', 155, 8) ('pine', 6345, 9) ### ### ('hereditary', 1765, 10) ('##gre', 9283, 11) ('tree', 1800, 12) ('clans', 10598, 13) ### ### ('martin', 2, 14) ('adoption', 78, 15) ('communities', 428, 16) ('adopt', 6271, 17) ### ### ('ontario', 52, 18) ('worried', 1312, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('smoke', 1, 0) ('johnson', 0, 3) ('wolf', 6, 1) ('mohawk', 8, 2) ('martin', 2, 14) ('clan', 32, 6) ### ### ('john', 10, 33) ('littered', 18, 61) ('canada', 31, 31) ('iroquois', 23, 60) ('ontario', 52, 18) ### ### ('adoption', 78, 15) ('chief', 35, 70) ('assimilation', 155, 8) ('¹', 5, 226) ('adopted', 84, 36) ### ### ('son', 12, 172) ('leaders', 119, 39) ('tribal', 498, 7) ('settlers', 104, 54) ### ############################################################################################################ [2023-11-25 22:31:55,095][root][INFO] - Epoch: 4: Step: 17701/19530, loss[v]=0.099708, lr=0.000005, acc@1[1]=997.0/1024=0.9736328125, acc@1[2]=1018.0/1024=0.994140625 [2023-11-25 22:33:43,778][root][INFO] - Train batch 17800 [2023-11-25 22:33:43,779][root][INFO] - Avg. loss per last 100 batches: 0.127954 [2023-11-25 22:33:44,474][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 4.00 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.2/29522=99.94% | mean: 0.02 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: theme song " more " by italian composer riz ortolani was nominated for the academy award ### ### for best original song in 1963, the year of its premier in the united states. ### ### [P_TEXT]: riz ortolani conservatorio statale di musica " gioachino rossini " in his hometown of ### ### pesaro before moving to rome in 1948 and finding work with the rai orchestra. though the chronology ### ### is unclear, he also likely served as a musician in the italian air force orchestra, formed a jazz ### ### ensemble, and came to the united states as a jazz musician in hollywood, all before scoring his ### ### first film. in the early 1950s, ortolani was founder and member of a well - known italian jazz ### ### band. one of his early film scores was for paolo cavara and gualtiero jacopetti's 1962 pseudo - ### ### documentary " mondo cane ", whose main title - song " more " earned ### ### =============================================== v_q ================================================ ### ### ('more', 0, 28) ('##lani', 1, 0) ('absently', 2, 135) ('ri', 3, 15) ('farther', 4, 42) ### ### ('幸', 5, 202) ('貴', 6, 270) ('intently', 7, 257) ('খ', 8, 198) ('¹', 9, 355) ('academy', 10, 11065) ### ### ('1963', 11, 93) ('##to', 12, 13) ('##½', 13, 290) ('theme', 14, 1287) ('ம', 15, 130) ### ### ('##ப', 16, 183) ('italian', 17, 9) ('original', 18, 13122) ('exclaimed', 19, 488) ### ### =============================================== v_p ================================================ ### ### ('##lani', 1, 0) ('cane', 13400, 1) ('##vara', 25560, 2) ('rai', 3221, 3) ('##cope', 1232, 4) ### ### ('##tier', 16215, 5) ('##sar', 23879, 6) ('rossi', 7024, 7) ('paolo', 1025, 8) ('italian', 17, 9) ### ### ('orchestra', 418, 10) ('musica', 3034, 11) ('jazz', 2202, 12) ('##to', 12, 13) ### ### ('pseudo', 9196, 14) ('ri', 3, 15) ('″', 81, 16) ('air', 3809, 17) ('##chin', 20777, 18) ### ### ('mon', 21638, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##lani', 1, 0) ('more', 0, 28) ('ri', 3, 15) ('italian', 17, 9) ('##to', 12, 13) ### ### ('farther', 4, 42) ('1962', 31, 21) ('##z', 21, 52) ('worried', 36, 27) ('song', 22, 55) ### ### ('hollywood', 29, 37) ('anyway', 53, 25) ('″', 81, 16) ('absently', 2, 135) ('1963', 11, 93) ### ### ('italy', 55, 38) ('or', 27, 54) ('film', 34, 50) ('ம', 15, 130) ('幸', 5, 202) ### ############################################################################################################ [2023-11-25 22:33:44,978][root][INFO] - Epoch: 4: Step: 17801/19530, loss[v]=0.091751, lr=0.000005, acc@1[1]=1004.0/1024=0.98046875, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 22:35:33,886][root][INFO] - Train batch 17900 [2023-11-25 22:35:33,887][root][INFO] - Avg. loss per last 100 batches: 0.131846 [2023-11-25 22:35:34,613][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.1/29522=99.93% | mean: 0.02 | max: 5.70 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: on january 1, 2010, takatsuki, along with the towns of kohoku and torahime ( both from ### ### higashiazai district ), and the towns of kinomoto, nishiazai, and yogo ( all from ika district ), ### ### was merged ### ### [P_TEXT]: torahime, shiga was a town located in higashiazai district, shiga prefecture, japan. as ### ### of 2003, the town had an estimated population of 5, 737 and a density of 607. 09 persons per km2. ### ### the total area was 9. 45 km2. on january 1, 2010, torahime, along with the towns of kohoku ( also ### ### from higashiazai district ), and the towns of kinomoto, nishiazai, takatsuki and yogo ( all from ### ### ika district ), was merged into the ### ### =============================================== v_q ================================================ ### ### ('##uki', 0, 10) ('¹', 1, 33) ('tak', 2, 34) ('幸', 3, 64) ('ik', 4, 1) ('ம', 5, 56) ('##ப', 6, 61) ### ### ('intently', 7, 75) ('খ', 8, 90) ('³', 9, 324) ('absently', 10, 99) ('貴', 11, 70) ('##ats', 12, 12) ### ### ('##部', 13, 74) ('district', 14, 16) ('##gas', 15, 4) ('japanese', 16, 20) ('##hia', 17, 8) ### ### ('##omo', 18, 9) ('##き', 19, 113) ### ### =============================================== v_p ================================================ ### ### ('torah', 76, 0) ('ik', 4, 1) ('##ime', 49, 2) ('##zai', 98, 3) ('##gas', 15, 4) ('shi', 4785, 5) ### ### ('##ga', 4946, 6) ('worried', 97, 7) ('##hia', 17, 8) ('##omo', 18, 9) ('##uki', 0, 10) ### ### ('yo', 28, 11) ('##ats', 12, 12) ('##go', 69, 13) ('kin', 40, 14) ('##ho', 22, 15) ### ### ('district', 14, 16) ('##az', 21, 17) ('japan', 24, 18) ('anyway', 61, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ik', 4, 1) ('##uki', 0, 10) ('##gas', 15, 4) ('##ats', 12, 12) ('##hia', 17, 8) ('##omo', 18, 9) ### ### ('district', 14, 16) ('japanese', 16, 20) ('##ho', 22, 15) ('¹', 1, 33) ('tak', 2, 34) ### ### ('##az', 21, 17) ('##ime', 49, 2) ('yo', 28, 11) ('torah', 76, 0) ('japan', 24, 18) ('kin', 40, 14) ### ### ('ம', 5, 56) ('幸', 3, 64) ('##ப', 6, 61) ### ############################################################################################################ [2023-11-25 22:35:35,120][root][INFO] - Epoch: 4: Step: 17901/19530, loss[v]=0.215871, lr=0.000005, acc@1[1]=969.5/1024=0.94677734375, acc@1[2]=1001.5/1024=0.97802734375 [2023-11-25 22:37:24,173][root][INFO] - Train batch 18000 [2023-11-25 22:37:24,174][root][INFO] - Avg. loss per last 100 batches: 0.125256 [2023-11-25 22:37:24,879][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29521.0/29522=100.00% | mean: 0.03 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.4/29522=99.94% | mean: 0.02 | max: 5.89 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the first and second season have been released on disney + since its launch, but users ### ### have noted episodes being out of order. ### ### [P_TEXT]: disney + " the simpsons ", " x - men ", " ducktales ", " phineas and ferb ", " kim ### ### possible ", and " " are presented almost entirely out of order, while some series are missing ### ### episodes. others have questioned why an extensive catalog of disney - owned material is missing ### ### from the platform, including marvel content, disney junior titles, various muppets media, and " ### ### star wars ". on june 26, 2020, the 2017 " ducktales " series had its episodes arranged in the ### ### proper order while co - creator of " phineas and ferb ", jeff " swampy " marsh, has stated ### ### =============================================== v_q ================================================ ### ### ('+', 0, 0) ('disney', 1, 2) ('season', 2, 231) ('##₂', 3, 1606) ('intently', 4, 117) ('貴', 5, 147) ### ### ('¹', 6, 32) ('absently', 7, 92) ('幸', 8, 88) ('##ப', 9, 86) ('খ', 10, 116) ('ம', 11, 73) ### ### ('##₁', 12, 974) ('worried', 13, 174) ('2', 14, 15646) ('initial', 15, 4485) ('anyway', 16, 30) ### ### ('series', 17, 27) ('netflix', 18, 1247) ('"', 19, 47) ### ### =============================================== v_p ================================================ ### ### ('+', 0, 0) ('##tale', 5526, 1) ('disney', 1, 2) ('possible', 2697, 3) ('kim', 10764, 4) ### ### ('marsh', 18733, 5) ('swamp', 2493, 6) ('duck', 2326, 7) ('simpsons', 178, 8) ('content', 67, 9) ### ### ('missing', 1322, 10) ('x', 811, 11) ('media', 318, 12) ('junior', 3068, 13) ('wars', 3433, 14) ### ### ('″', 41, 15) ('##rb', 21297, 16) ('marvel', 1365, 17) ('##neas', 11595, 18) ('creator', 144, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('+', 0, 0) ('disney', 1, 2) ('¹', 6, 32) ('series', 17, 27) ('anyway', 16, 30) ('″', 41, 15) ### ### ('content', 67, 9) ('episodes', 29, 23) ('absently', 7, 92) ('幸', 8, 88) ('2020', 23, 40) ### ### ('ம', 11, 73) ('##ப', 9, 86) ('intently', 4, 117) ('##ppet', 49, 21) ('order', 31, 39) ### ### ('"', 19, 47) ('simpsons', 178, 8) ('貴', 5, 147) ('খ', 10, 116) ### ############################################################################################################ [2023-11-25 22:37:25,383][root][INFO] - Epoch: 4: Step: 18001/19530, loss[v]=0.068956, lr=0.000005, acc@1[1]=1005.0/1024=0.9814453125, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 22:39:13,935][root][INFO] - Train batch 18100 [2023-11-25 22:39:13,936][root][INFO] - Avg. loss per last 100 batches: 0.125138 [2023-11-25 22:39:14,641][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.5/29522=99.93% | mean: 0.02 | max: 5.91 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: jack. org ( formerly known as the jack project under the charitable organization kids ### ### help phone ) is a canadian non - profit organization focused on youth mental health and suicide ### ### prevention. ### ### [P_TEXT]: kids help phone the jack project @ kids help phone " ( re - branded to jack. org in 2013 ### ### ) to develop programs and workshops targeted at young people. with their assistance, kids help ### ### phone was able to create an online chat service and mobile app to reach young people. it then ### ### launched brotalk, a dedicated online service for teenage boys, in 2015 as an outreach program to ### ### reduce stigma around men seeking mental health support. since taking the first call on may 16, ### ### 1989, kids help phone continues to deliver mental health services to young people across canada. ### ### the organization ’ s founding partners are bell canada, ### ### =============================================== v_q ================================================ ### ### ('jack', 0, 5) ('org', 1, 9) ('¹', 2, 1967) ('phone', 3, 12) ('kids', 4, 2) ('mental', 5, 6) ### ### ('suicide', 6, 3398) ('help', 7, 8) ('³', 8, 4421) ('health', 9, 14) ('canadian', 10, 23) ### ### ('youth', 11, 52) ('prevention', 12, 280) ('project', 13, 27) ('幸', 14, 271) ('canada', 15, 7) ### ### ('teenagers', 16, 114) ('hissed', 17, 30) ('貴', 18, 442) ('curious', 19, 3063) ### ### =============================================== v_p ================================================ ### ### ('##talk', 6752, 0) ('chat', 11077, 1) ('kids', 4, 2) ('stigma', 1479, 3) ('bell', 10773, 4) ### ### ('jack', 0, 5) ('mental', 5, 6) ('canada', 15, 7) ('help', 7, 8) ('org', 1, 9) ('bro', 20119, 10) ### ### ('teenage', 60, 11) ('phone', 3, 12) ('outreach', 35, 13) ('health', 9, 14) ('workshops', 8484, 15) ### ### ('@', 152, 16) ('psychiatric', 354, 17) ('leaned', 5228, 18) ('boys', 342, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('jack', 0, 5) ('kids', 4, 2) ('org', 1, 9) ('phone', 3, 12) ('mental', 5, 6) ('help', 7, 8) ### ### ('health', 9, 14) ('canada', 15, 7) ('canadian', 10, 23) ('project', 13, 27) ('outreach', 35, 13) ### ### ('hissed', 17, 30) ('youth', 11, 52) ('teenage', 60, 11) ('organization', 39, 33) ('@', 152, 16) ### ### ('online', 112, 20) ('teenagers', 16, 114) ('prevention', 12, 280) ('stigma', 1479, 3) ### ############################################################################################################ [2023-11-25 22:39:15,145][root][INFO] - Epoch: 4: Step: 18101/19530, loss[v]=0.107151, lr=0.000005, acc@1[1]=996.0/1024=0.97265625, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-25 22:41:04,086][root][INFO] - Train batch 18200 [2023-11-25 22:41:04,086][root][INFO] - Avg. loss per last 100 batches: 0.132254 [2023-11-25 22:41:04,812][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.3/29522=99.94% | mean: 0.02 | max: 5.74 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: puig runs the red bull motogp academy, designed to find and train promising grand prix ### ### racers and, he is credited with starting the careers of casey stoner, dani pedrosa and toni elias. ### ### [P_TEXT]: dani pedrosa minibike championship and ended his debut season in second place, scoring ### ### his first podium finish in the second race of the season. the next year, pedrosa entered the same ### ### championship, but health problems prevented him from improving his results and he ended that season ### ### in 3rd position. in 2001, pedrosa made his world championship debut in the 125cc class after being ### ### selected from the movistar activa cup, a series designed to promote fresh racing talent in spain, ### ### back in 1999. under the guidance of alberto puig, pedrosa scored two podium finishes in the first ### ### season and won ### ### =============================================== v_q ================================================ ### ### ('##ig', 0, 3) ('pu', 1, 4) ('¹', 2, 177) ('absently', 3, 131) ('幸', 4, 139) ('casey', 5, 4461) ### ### ('toni', 6, 16840) ('academy', 7, 10009) ('খ', 8, 114) ('##ப', 9, 90) ('stone', 10, 17173) ### ### ('dani', 11, 0) ('貴', 12, 175) ('³', 13, 1334) ('ம', 14, 92) ('elias', 15, 9556) ### ### ('intently', 16, 231) ('bull', 17, 1104) ('.', 18, 576) ('red', 19, 18477) ### ### =============================================== v_p ================================================ ### ### ('dani', 11, 0) ('pedro', 52, 1) ('alberto', 369, 2) ('##ig', 0, 3) ('pu', 1, 4) ### ### ('##iva', 20823, 5) ('2001', 1717, 6) ('##sa', 22, 7) ('leaned', 403, 8) ('act', 17996, 9) ### ### ('125', 6734, 10) ('littered', 26, 11) ('mini', 3102, 12) ('spain', 244, 13) ('1999', 1820, 14) ### ### ('##vis', 25369, 15) ('worried', 98, 16) ('anyway', 71, 17) ('##bi', 13966, 18) ('podium', 377, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ig', 0, 3) ('pu', 1, 4) ('dani', 11, 0) ('##sa', 22, 7) ('littered', 26, 11) ('pedro', 52, 1) ### ### ('race', 34, 20) ('racing', 27, 31) ('¹', 2, 177) ('absently', 3, 131) ('##ப', 9, 90) ('幸', 4, 139) ### ### ('খ', 8, 114) ('anyway', 71, 17) ('worried', 98, 16) ('ம', 14, 92) ('alberto', 369, 2) ### ### ('貴', 12, 175) ('stared', 25, 78) ('spain', 244, 13) ### ############################################################################################################ [2023-11-25 22:41:05,318][root][INFO] - Epoch: 4: Step: 18201/19530, loss[v]=0.081154, lr=0.000005, acc@1[1]=1000.0/1024=0.9765625, acc@1[2]=1017.5/1024=0.99365234375 [2023-11-25 22:42:54,369][root][INFO] - Train batch 18300 [2023-11-25 22:42:54,370][root][INFO] - Avg. loss per last 100 batches: 0.129251 [2023-11-25 22:42:55,073][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.4/29522=99.93% | mean: 0.02 | max: 5.88 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the highest peak in the tian shan is jengish chokusu ( also called victory peak ) on the ### ### border of china. ### ### [P_TEXT]: jengish chokusu jengish chokusu (,,, ; ;, " pik pobedy ", ) is the highest mountain in ### ### the tian shan mountain system at. it lies on the kyrgyzstan – china border between the raion of ak ### ### - suu, in the issyk - kul region of far eastern kyrgyzstan and wensu county, xinjiang, china. it is ### ### part of the kakshaal too, the highest part of the tian shan and located southeast of lake issyk ### ### kul. the mountain's official name in kyrgyz is " jengish choqusu ", ### ### =============================================== v_q ================================================ ### ### ('shan', 0, 17) ('tian', 1, 8) ('victory', 2, 14864) ('peak', 3, 126) ('¹', 4, 74) ('##kus', 5, 1) ### ### ('jen', 6, 4) ('absently', 7, 143) ('খ', 8, 164) ('幸', 9, 161) ('china', 10, 13) ('貴', 11, 122) ### ### ('ம', 12, 118) ('##ப', 13, 145) ('.', 14, 169) ('curious', 15, 947) ('mountain', 16, 21) ### ### ('intently', 17, 136) ('##部', 18, 141) ('mountains', 19, 25) ### ### =============================================== v_p ================================================ ### ### ('##ksha', 21457, 0) ('##kus', 5, 1) ('xinjiang', 165, 2) ('iss', 3416, 3) ('jen', 6, 4) ### ### ('##gis', 22, 5) ('wen', 9231, 6) ('##su', 3901, 7) ('tian', 1, 8) ('raion', 21383, 9) ### ### ('ak', 8379, 10) ('worried', 70, 11) ('frontiers', 8923, 12) ('china', 10, 13) ### ### ('kyrgyzstan', 2761, 14) ('##yk', 22154, 15) ('lake', 540, 16) ('shan', 0, 17) ('cho', 23, 18) ### ### ('leaned', 541, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##kus', 5, 1) ('tian', 1, 8) ('jen', 6, 4) ('shan', 0, 17) ('china', 10, 13) ('##gis', 22, 5) ### ### ('mountain', 16, 21) ('cho', 23, 18) ('mountains', 19, 25) ('¹', 4, 74) ('chinese', 21, 50) ### ### ('worried', 70, 11) ('xinjiang', 165, 2) ('peak', 3, 126) ('absently', 7, 143) ('highest', 34, 35) ### ### ('peaks', 41, 32) ('貴', 11, 122) ('খ', 8, 164) ('幸', 9, 161) ### ############################################################################################################ [2023-11-25 22:42:55,577][root][INFO] - Epoch: 4: Step: 18301/19530, loss[v]=0.065314, lr=0.000005, acc@1[1]=1003.5/1024=0.97998046875, acc@1[2]=1015.5/1024=0.99169921875 [2023-11-25 22:44:44,176][root][INFO] - Train batch 18400 [2023-11-25 22:44:44,176][root][INFO] - Avg. loss per last 100 batches: 0.128822 [2023-11-25 22:44:44,883][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.99 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.6/29522=99.92% | mean: 0.02 | max: 5.84 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: some of huemer's most creative work was done in partnership with joe grant ; examples ### ### include " fantasia " ( story director ), " dumbo " ( screenplay ), and several propaganda films to ### ### advance the u. s. war effort during world war ii. ### ### [P_TEXT]: fantasia ( 1940 film ) one day after a series of meetings began to select the musical ### ### pieces for " the concert feature ". disney made story writers joe grant and dick huemer gather a ### ### preliminary selection of music and along with stokowski, taylor, and the heads of various ### ### departments, discussed their ideas. each meeting was recorded verbatim by stenographers with ### ### participants being given a copy of the entire conversation for review. as selections were ### ### considered, a recording of the piece was located and played back at the next gathering. disney did ### ### not contribute much to early discussions ; he admitted that his knowledge ### ### =============================================== v_q ================================================ ### ### ('grant', 0, 3) ('hue', 1, 5) ('fantasia', 2, 0) ('¹', 3, 236) ('##mer', 4, 10) ('dumb', 5, 11168) ### ### ('joe', 6, 4) ('##½', 7, 330) ('³', 8, 2092) ('film', 9, 23) ('##o', 10, 24915) ('幸', 11, 162) ### ### ('propaganda', 12, 1298) ('welles', 13, 8879) ('exclaimed', 14, 64) ('pictures', 15, 411) ### ### ('puzzled', 16, 96) ('খ', 17, 142) ('##¨', 18, 447) ('##ப', 19, 119) ### ### =============================================== v_p ================================================ ### ### ('fantasia', 2, 0) ('concert', 17459, 1) ('1940', 170, 2) ('grant', 0, 3) ('joe', 6, 4) ### ### ('hue', 1, 5) ('feature', 5359, 6) ('verb', 24261, 7) ('disney', 63, 8) ('taylor', 8146, 9) ### ### ('##mer', 4, 10) ('##oko', 22351, 11) ('writers', 419, 12) ('##wski', 19114, 13) ### ### ('pieces', 4086, 14) ('worried', 3815, 15) ('dick', 204, 16) ('littered', 449, 17) ### ### ('meetings', 3430, 18) ('conversation', 2217, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('fantasia', 2, 0) ('grant', 0, 3) ('hue', 1, 5) ('joe', 6, 4) ('##mer', 4, 10) ('film', 9, 23) ### ### ('disney', 63, 8) ('1940', 170, 2) ('¹', 3, 236) ('dick', 204, 16) ('exclaimed', 14, 64) ### ### ('story', 97, 28) ('writers', 419, 12) ('movie', 143, 29) ('幸', 11, 162) ('##½', 7, 330) ### ### ('films', 40, 50) ('littered', 449, 17) ('puzzled', 16, 96) ('1939', 319, 24) ### ############################################################################################################ [2023-11-25 22:44:45,388][root][INFO] - Epoch: 4: Step: 18401/19530, loss[v]=0.075360, lr=0.000005, acc@1[1]=999.5/1024=0.97607421875, acc@1[2]=1013.5/1024=0.98974609375 [2023-11-25 22:46:34,673][root][INFO] - Train batch 18500 [2023-11-25 22:46:34,673][root][INFO] - Avg. loss per last 100 batches: 0.125597 [2023-11-25 22:46:35,400][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.5/29522=99.93% | mean: 0.02 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: smith signed as a free agent with the bc lions in may, 2006, and replaced antonio warren ### ### as their starting running back when warren was released in ### ### [P_TEXT]: bc lions the lions in the off - season also acquired former roughrider paul mccallum to ### ### stabilize the kicking game which had not recovered from lui passaglia's retirement in 2000. after a ### ### sluggish 2 – 3 start in 2006, coach buono signalled that no starting job was safe by releasing ### ### runningback antonio warren and defensive back sam young. the team responded by winning the next six ### ### games. new running back joe smith ran for over 100 yards in his first game. brent johnson, barron ### ### miles, and mark washington led ### ### =============================================== v_q ================================================ ### ### ('smith', 0, 6) ('warren', 1, 3) ('bc', 2, 0) ('¹', 3, 33) ('antonio', 4, 19) ('lions', 5, 1) ### ### ('³', 6, 214) ('cfl', 7, 184) ('幸', 8, 91) ('ɐ', 9, 1824) ('football', 10, 125) ('貴', 11, 147) ### ### ('2006', 12, 18) ('offensive', 13, 192) ('##ப', 14, 87) ('lacrosse', 15, 1057) ('খ', 16, 126) ### ### ('intently', 17, 133) ('ம', 18, 74) ('grey', 19, 13767) ### ### =============================================== v_p ================================================ ### ### ('bc', 2, 0) ('lions', 5, 1) ('barron', 20890, 2) ('warren', 1, 3) ('mccall', 1458, 4) ### ### ('##ride', 14510, 5) ('smith', 0, 6) ('kicking', 260, 7) ('littered', 27, 8) ('johnson', 1208, 9) ### ### ('washington', 2265, 10) ('sam', 1370, 11) ('joe', 215, 12) ('rough', 186, 13) ('brent', 239, 14) ### ### ('young', 5324, 15) ('##ono', 19739, 16) ('miles', 9443, 17) ('2006', 12, 18) ('antonio', 4, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('bc', 2, 0) ('smith', 0, 6) ('warren', 1, 3) ('lions', 5, 1) ('antonio', 4, 19) ('¹', 3, 33) ### ### ('2006', 12, 18) ('littered', 27, 8) ('pass', 33, 31) ('幸', 8, 91) ('kicking', 260, 7) ### ### ('rough', 186, 13) ('game', 25, 46) ('touchdown', 37, 39) ('joe', 215, 12) ('coach', 68, 29) ### ### ('brent', 239, 14) ('season', 44, 41) ('hissed', 92, 27) ('cfl', 7, 184) ### ############################################################################################################ [2023-11-25 22:46:35,907][root][INFO] - Epoch: 4: Step: 18501/19530, loss[v]=0.155085, lr=0.000005, acc@1[1]=980.0/1024=0.95703125, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-25 22:48:24,486][root][INFO] - Train batch 18600 [2023-11-25 22:48:24,487][root][INFO] - Avg. loss per last 100 batches: 0.134086 [2023-11-25 22:48:25,191][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29505.0/29522=99.94% | mean: 0.03 | max: 5.90 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: robeston wathen is a rural village and parish in pembrokeshire, wales, west of narberth, ### ### on the former narberth to haverfordwest turnpike subsequently designated the a40 which bypassed the ### ### village in 2011. ### ### [P_TEXT]: a40 road majority. the political party at the centre of the row instead directed the ### ### project deferment to damning environmental statements by friends of the earth cymru. the scheme was ### ### officially scrapped in 2008 after a welsh assembly committee decided to abandon the proposals. ### ### instead, it recommended upgrades to the existing route including bypasses around robeston wathen ### ### and llanddewi velfrey using a three - lane option. this was welcomed by the friends of the earth, ### ### saying " the dual carriageway on the a40 would increase traffic levels, increase the emission of ### ### greenhouse gases, it would be harmful for ### ### =============================================== v_q ================================================ ### ### ('##hen', 0, 7) ('wat', 1, 3) ('¹', 2, 339) ('robes', 3, 1) ('##rber', 4, 27200) ('³', 5, 616) ### ### ('##ton', 6, 59) ('curious', 7, 1418) ('##40', 8, 0) ('幸', 9, 266) ('##th', 10, 24279) ### ### ('na', 11, 26854) ('খ', 12, 293) ('luckily', 13, 71) ('intently', 14, 315) ('##ப', 15, 185) ### ### ('貴', 16, 286) ('villagers', 17, 270) ('ம', 18, 119) ('pembroke', 19, 4794) ### ### =============================================== v_p ================================================ ### ### ('##40', 8, 0) ('robes', 3, 1) ('friends', 4047, 2) ('wat', 1, 3) ('##rey', 3750, 4) ### ### ('welsh', 30, 5) ('earth', 11323, 6) ('##hen', 0, 7) ('##ru', 8160, 8) ('##lf', 13371, 9) ### ### ('carriage', 968, 10) ('anyway', 178, 11) ('road', 140, 12) ('lanes', 3220, 13) ('bypass', 84, 14) ### ### ('committee', 10690, 15) ('assembly', 7695, 16) ('ve', 14824, 17) ('option', 9524, 18) ### ### ('greenhouse', 9236, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('robes', 3, 1) ('wat', 1, 3) ('##40', 8, 0) ('##hen', 0, 7) ('welsh', 30, 5) ('##ton', 6, 59) ### ### ('hissed', 24, 22) ('luckily', 13, 71) ('wales', 26, 39) ('¹', 2, 339) ('bypass', 84, 14) ### ### ('outskirts', 27, 75) ('road', 140, 12) ('anyway', 178, 11) ('ம', 18, 119) ('##ப', 15, 185) ### ### ('nearby', 57, 44) ('幸', 9, 266) ('village', 34, 69) ('roundabout', 48, 52) ### ############################################################################################################ [2023-11-25 22:48:25,695][root][INFO] - Epoch: 4: Step: 18601/19530, loss[v]=0.076265, lr=0.000005, acc@1[1]=999.0/1024=0.9755859375, acc@1[2]=1016.0/1024=0.9921875 [2023-11-25 22:50:14,334][root][INFO] - Train batch 18700 [2023-11-25 22:50:14,335][root][INFO] - Avg. loss per last 100 batches: 0.132664 [2023-11-25 22:50:15,046][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.4/29522=99.94% | mean: 0.02 | max: 5.69 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: underglaze decoration is applied before the glaze, usually to unfired pottery ( " raw " ### ### or " greenware " ) but sometimes to " biscuit " - fired ( an ### ### [P_TEXT]: wemyss ware the griselda hill pottery in ceres, fife. esther weeks taught the painters at ### ### ceres techniques she learned from joseph nekola at bovey. after moulding, the pottery was first ### ### fired at a low temperature to produce a porous biscuit body onto which paint could be applied. the ### ### colours were then applied, followed by a lead ceramic glaze, a technique known as underglazing. the ### ### pottery was then fired again at a low temperature in order to preserve the colour, making the ### ### product soft and fragile and contributing to the scarcity of original wemyss. ### ### =============================================== v_q ================================================ ### ### ('pottery', 0, 3) ('green', 1, 12600) ('貴', 2, 308) ('intently', 3, 240) ('খ', 4, 207) ### ### ('##gl', 5, 8) ('¹', 6, 743) ('幸', 7, 267) ('##az', 8, 7825) ('bis', 9, 31) ('absently', 10, 148) ### ### ('##cuit', 11, 12) ('.', 12, 9769) ('##ware', 13, 1257) ('##laze', 14, 6) ('##ப', 15, 177) ### ### ('ம', 16, 127) ('g', 17, 53) ('##fire', 18, 14622) ('raw', 19, 14115) ### ### =============================================== v_p ================================================ ### ### ('##vey', 17258, 0) ('##res', 15132, 1) ('esther', 9979, 2) ('pottery', 0, 3) ('##lda', 28773, 4) ### ### ('weeks', 7562, 5) ('##laze', 14, 6) ('ware', 40, 7) ('##gl', 5, 8) ('fife', 1448, 9) ### ### ('ce', 2166, 10) ('##ss', 21290, 11) ('##cuit', 11, 12) ('##azi', 7738, 13) ('##ko', 27017, 14) ### ### ('littered', 2002, 15) ('joseph', 12288, 16) ('ne', 23666, 17) ('ceramic', 53, 18) ### ### ('##my', 20445, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('pottery', 0, 3) ('##gl', 5, 8) ('##laze', 14, 6) ('##cuit', 11, 12) ('bis', 9, 31) ### ### ('worried', 20, 33) ('ware', 40, 7) ('anyway', 23, 27) ('g', 17, 53) ('underneath', 25, 41) ### ### ('ceramic', 53, 18) ('under', 22, 72) ('paint', 71, 30) ('ceramics', 111, 23) ('absently', 10, 148) ### ### ('ம', 16, 127) ('খ', 4, 207) ('intently', 3, 240) ('##ப', 15, 177) ('fired', 44, 65) ### ############################################################################################################ [2023-11-25 22:50:15,552][root][INFO] - Epoch: 4: Step: 18701/19530, loss[v]=0.152417, lr=0.000005, acc@1[1]=988.5/1024=0.96533203125, acc@1[2]=1006.0/1024=0.982421875 [2023-11-25 22:52:04,882][root][INFO] - Train batch 18800 [2023-11-25 22:52:04,882][root][INFO] - Avg. loss per last 100 batches: 0.125361 [2023-11-25 22:52:05,593][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.76 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.8/29522=99.93% | mean: 0.02 | max: 5.66 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: hiba nawab is an indian television actress known for her portrayal of anushka in channel ### ### v india's " crazy stupid ishq ", amaya in star plus's " tere sheher mein " and elaichi in sab tv's ### ### " jijaji chhat per hain ". ### ### [P_TEXT]: jijaji chhat parr koii hai jijaji chhat parr koii hai ( ) is an upcoming indian ### ### television series which is scheduled to premiere on 8 march 2021 on sony sab. it is a sequel to the ### ### series " jijaji chhat per hain ". produced by edit ii production, it stars hiba nawab, shubhashish ### ### jha and anup upadhyay. " jijaji chhat parr koii hai " is a classic tale of an age - old family feud ### ### between the jaldirams and the jindals over the claims on the ancestral property. while the property ### ### feud stirs the rivalry between the families, there ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 306) ('nawab', 1, 1) ('hi', 2, 5) ('##ba', 3, 19) ('³', 4, 2875) ('stupid', 5, 5783) ### ### ('whoever', 6, 94) ('幸', 7, 183) ('mein', 8, 21194) ('##ush', 9, 22580) ('el', 10, 26139) ### ### ('##ப', 11, 168) ('খ', 12, 249) ('貴', 13, 237) ('ter', 14, 25957) ('##her', 15, 21484) ### ### ('##q', 16, 27066) ('ama', 17, 25030) ('ம', 18, 148) ('intently', 19, 318) ### ### =============================================== v_p ================================================ ### ### ('edit', 16431, 0) ('nawab', 1, 1) ('sony', 2151, 2) ('jin', 2204, 3) ('##ld', 26426, 4) ### ### ('hi', 2, 5) ('##dal', 13488, 6) ('worried', 1695, 7) ('ji', 22, 8) ('##ira', 7090, 9) ### ### ('shu', 9844, 10) ('##hya', 4926, 11) ('##ii', 14498, 12) ('##bha', 5603, 13) ('sequel', 6950, 14) ### ### ('feud', 3439, 15) ('##up', 7080, 16) ('littered', 131, 17) ('##b', 24, 18) ('##ba', 3, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('nawab', 1, 1) ('hi', 2, 5) ('##ba', 3, 19) ('ji', 22, 8) ('##b', 24, 18) ('##ja', 39, 21) ### ### ('##hat', 38, 24) ('##ji', 30, 38) ('ch', 42, 25) ('whoever', 6, 94) ('hai', 49, 26) ### ### ('hissed', 25, 55) ('¹', 0, 306) ('sa', 55, 36) ('india', 62, 33) ('littered', 131, 17) ### ### ('幸', 7, 183) ('indian', 102, 40) ('##ப', 11, 168) ('dramas', 92, 42) ### ############################################################################################################ [2023-11-25 22:52:06,099][root][INFO] - Epoch: 4: Step: 18801/19530, loss[v]=0.206678, lr=0.000005, acc@1[1]=973.5/1024=0.95068359375, acc@1[2]=998.0/1024=0.974609375 [2023-11-25 22:53:54,675][root][INFO] - Train batch 18900 [2023-11-25 22:53:54,676][root][INFO] - Avg. loss per last 100 batches: 0.138396 [2023-11-25 22:53:55,378][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.8/29522=99.93% | mean: 0.02 | max: 5.77 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the lounge was used in the 1981 tv series " brideshead revisited " as the interior of an ### ### ocean liner. ### ### [P_TEXT]: brideshead revisited ( tv serial ) the exterior of marchmain house, and its interiors ### ### were filmed in tatton hall. rex and julia's wedding was filmed in the chapel at lyme park. venice ### ### locations included the basilica di santa maria gloriosa dei frari, the scuola di san rocco, and the ### ### palazzi barbaro. the ocean - liner deck scenes were filmed on the qe2 during an actual storm, but ### ### the ship's interiors were either sets or public rooms in the adelphi hotel in liverpool and the ### ### park lane hotel in london. the riot in the general strike ### ### =============================================== v_q ================================================ ### ### ('liner', 0, 25) ('lounge', 1, 1032) ('revisited', 2, 8) ('bride', 3, 19) ('¹', 4, 168) ### ### ('##sh', 5, 54) ('1981', 6, 19301) ('##ead', 7, 10) ('intently', 8, 196) ('貴', 9, 311) ### ### ('幸', 10, 249) ('³', 11, 661) ('ocean', 12, 53) ('খ', 13, 265) ('##ப', 14, 234) ### ### ('cruise', 15, 2900) ('thames', 16, 2484) ('ம', 17, 188) ('interior', 18, 317) ### ### ('absently', 19, 194) ### ### =============================================== v_p ================================================ ### ### ('rocco', 14620, 0) ('##main', 21022, 1) ('rex', 12020, 2) ('##tton', 2132, 3) ('##yme', 18651, 4) ### ### ('julia', 6385, 5) ('strike', 11805, 6) ('basilica', 6658, 7) ('revisited', 2, 8) ('q', 14434, 9) ### ### ('##ead', 7, 10) ('hotel', 28, 11) ('deck', 161, 12) ('storms', 24467, 13) ('fra', 25400, 14) ### ### ('##bar', 22191, 15) ('##az', 27110, 16) ('chapel', 2606, 17) ('##uo', 27196, 18) ('bride', 3, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('liner', 0, 25) ('revisited', 2, 8) ('bride', 3, 19) ('##ead', 7, 10) ('##sh', 5, 54) ### ### ('hotel', 28, 11) ('ocean', 12, 53) ('ship', 27, 40) ('liverpool', 102, 22) ('deck', 161, 12) ### ### ('anyway', 51, 41) ('¹', 4, 168) ('intently', 8, 196) ('tv', 50, 76) ('luckily', 30, 107) ### ### ('filmed', 355, 38) ('hall', 277, 48) ('location', 98, 68) ('幸', 10, 249) ('site', 114, 67) ### ############################################################################################################ [2023-11-25 22:53:55,882][root][INFO] - Epoch: 4: Step: 18901/19530, loss[v]=0.130812, lr=0.000005, acc@1[1]=993.0/1024=0.9697265625, acc@1[2]=1011.0/1024=0.9873046875 [2023-11-25 22:55:44,643][root][INFO] - Train batch 19000 [2023-11-25 22:55:44,644][root][INFO] - Avg. loss per last 100 batches: 0.131789 [2023-11-25 22:55:45,364][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.8/29522=99.93% | mean: 0.02 | max: 5.70 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: together with vlora, berat and elbasan they provide almost 90 % of olive oil production. ### ### [P_TEXT]: vlore vlore. a motorway is being constructed linking the city with fier and albania as a ### ### whole. one of the most panoramic routes of the albanian riviera starts to the south of town ### ### stretching up to sarande in southernmost albania. vlore has a mediterranean climate ( koppen : " ### ### csa " ) with cool wet winters and hot, dry summers with temperatures exceeding in july and august. ### ### vlore has a relatively high precipitation for the dry subtropical zone that is found, the values ### ### can be high in the end of the spring, but dry ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 294) ('olive', 1, 7055) ('##lor', 2, 3925) ('幸', 3, 227) ('ம', 4, 112) ('খ', 5, 182) ### ### ('貴', 6, 177) ('v', 7, 5) ('absently', 8, 75) ('el', 9, 15112) ('albania', 10, 3) ### ### ('intently', 11, 133) ('##ப', 12, 202) ('##bas', 13, 24043) ('.', 14, 65) ('³', 15, 1062) ### ### ('##rat', 16, 27018) ('adriatic', 17, 2583) ('##部', 18, 208) ('kosovo', 19, 258) ### ### =============================================== v_p ================================================ ### ### ('fi', 17286, 0) ('##lore', 2547, 1) ('csa', 8302, 2) ('albania', 10, 3) ('##vier', 6322, 4) ### ### ('v', 7, 5) ('sara', 3995, 6) ('motorway', 254, 7) ('worried', 300, 8) ('##nde', 18057, 9) ### ### ('##er', 19930, 10) ('subtropical', 6193, 11) ('albanians', 666, 12) ('albanian', 21, 13) ### ### ('anyway', 35, 14) ('ri', 16269, 15) ('precipitation', 1644, 16) ('mediterranean', 86, 17) ### ### ('climatic', 16745, 18) ('dry', 9371, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('v', 7, 5) ('albania', 10, 3) ('albanian', 21, 13) ('anyway', 35, 14) ('city', 26, 27) ### ### ('mediterranean', 86, 17) ('motorway', 254, 7) ('.', 14, 65) ('absently', 8, 75) ### ### ('worried', 300, 8) ('outskirts', 36, 31) ('ம', 4, 112) ('town', 50, 28) ('intently', 11, 133) ### ### ('খ', 5, 182) ('貴', 6, 177) ('leaned', 207, 20) ('幸', 3, 227) ('¹', 0, 294) ('##lore', 2547, 1) ### ############################################################################################################ [2023-11-25 22:55:45,870][root][INFO] - Epoch: 4: Step: 19001/19530, loss[v]=0.058959, lr=0.000005, acc@1[1]=1005.0/1024=0.9814453125, acc@1[2]=1019.5/1024=0.99560546875 [2023-11-25 22:57:34,954][root][INFO] - Train batch 19100 [2023-11-25 22:57:34,955][root][INFO] - Avg. loss per last 100 batches: 0.124867 [2023-11-25 22:57:35,669][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.8/29522=99.93% | mean: 0.02 | max: 5.73 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: warnermedia stated that the service would have over 2, 000 films available at - launch, ### ### including hbo's existing pay television rights that are sourced from studios including sister ### ### company warner bros. pictures, universal pictures, and 20th century studios ( the latter two have ### ### respective output deals with hbo until 2022 ; films from 20th century will move to disney + and ### ### hulu once its output deal expire, while future rights to universal films are still up for ### ### negotiation despite the launch of peacock ### ### [P_TEXT]: the fresh prince of bel - air with the lifestyle of his upper - class relatives. " the ### ### fresh prince of bel - air " ran for 148 episodes over six seasons. a reunion special / ### ### retrospective reuniting the original cast debuted on hbo max on november 18, 2020, one day ahead of ### ### schedule. a more dramatic reboot based on the fan film " bel - air " is in active development, with ### ### a two - season order for peacock. the theme song and opening sequence set the premise of the show. ### ### will smith is a street - smart teenager, west philadelphia " born and raised ". while playing ### ### street basketball, will misses a shot and ### ### =============================================== v_q ================================================ ### ### ('peacock', 0, 0) ('absently', 1, 67) ('hbo', 2, 6) ('universal', 3, 12432) ('pictures', 4, 566) ### ### ('##media', 5, 15679) ('幸', 6, 226) ('貴', 7, 289) ('+', 8, 8331) ('bros', 9, 23583) ### ### ('intently', 10, 275) ('warner', 11, 14565) ('খ', 12, 221) ('disney', 13, 5694) ('hu', 14, 23704) ### ### ('##ப', 15, 166) ('¹', 16, 831) ('ம', 17, 117) ('films', 18, 121) ('.', 19, 1578) ### ### =============================================== v_p ================================================ ### ### ('peacock', 0, 0) ('bel', 16613, 1) ('##boot', 5480, 2) ('fresh', 2595, 3) ### ### ('philadelphia', 1952, 4) ('max', 340, 5) ('hbo', 2, 6) ('smith', 15246, 7) ('prince', 12966, 8) ### ### ('basketball', 5173, 9) ('air', 1307, 10) ('fan', 11918, 11) ('active', 4593, 12) ### ### ('born', 5755, 13) ('reunion', 8173, 14) ('will', 31, 15) ('smart', 5597, 16) ### ### ('lifestyle', 2563, 17) ('street', 7814, 18) ('worried', 32, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('peacock', 0, 0) ('hbo', 2, 6) ('absently', 1, 67) ('will', 31, 15) ('2020', 28, 25) ### ### ('worried', 32, 19) ('film', 25, 41) ('anyway', 42, 33) ('ம', 17, 117) ('films', 18, 121) ### ### ('max', 340, 5) ('##ப', 15, 166) ('幸', 6, 226) ('খ', 12, 221) ('##部', 21, 138) ('stared', 30, 88) ### ### ('intently', 10, 275) ('貴', 7, 289) ('raised', 151, 36) ('schedule', 126, 48) ### ############################################################################################################ [2023-11-25 22:57:36,174][root][INFO] - Epoch: 4: Step: 19101/19530, loss[v]=0.207940, lr=0.000005, acc@1[1]=982.5/1024=0.95947265625, acc@1[2]=1005.0/1024=0.9814453125 [2023-11-25 22:59:24,833][root][INFO] - Train batch 19200 [2023-11-25 22:59:24,833][root][INFO] - Avg. loss per last 100 batches: 0.132286 [2023-11-25 22:59:25,539][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.2/29522=99.92% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: manufactured by rocky mountain construction, the roller coaster is both the fastest ### ### wooden coaster in europe, and second tallest wooden coaster in the world. ### ### [P_TEXT]: rocky mountain construction in sweden announced plans to build the first rocky mountain ### ### coaster in europe called wildfire, which opened on june 28, 2016. to handle increased demand, rocky ### ### mountain construction opened a new factory, twice the size of their existing one, in august 2014. ### ### as of 2019, rocky mountain construction has built or refurbished 23 roller coaster s around the ### ### world. 18 of them are currently operating, 5 are being constructed. ### ### =============================================== v_q ================================================ ### ### ('rocky', 0, 0) ('coaster', 1, 2) ('wooden', 2, 1767) ('¹', 3, 821) ('construction', 4, 3) ### ### ('mountain', 5, 5) ('幸', 6, 178) ('absently', 7, 45) ('³', 8, 4464) ('貴', 9, 251) ('খ', 10, 198) ### ### ('europe', 11, 13) ('tallest', 12, 1351) ('##ப', 13, 124) ('intently', 14, 241) ### ### ('fastest', 15, 4414) ('ride', 16, 48) ('roller', 17, 14) ('.', 18, 4130) ('disney', 19, 3038) ### ### =============================================== v_p ================================================ ### ### ('rocky', 0, 0) ('##fire', 25435, 1) ('coaster', 1, 2) ('construction', 4, 3) ('factory', 966, 4) ### ### ('mountain', 5, 5) ('wild', 2202, 6) ('sweden', 161, 7) ('demand', 7334, 8) ('anyway', 223, 9) ### ### ('leaned', 3592, 10) ('build', 335, 11) ('swedish', 1149, 12) ('europe', 11, 13) ('roller', 17, 14) ### ### ('climbs', 22447, 15) ('operating', 5985, 16) ('factories', 5675, 17) ('worried', 252, 18) ### ### ('size', 1798, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('rocky', 0, 0) ('coaster', 1, 2) ('construction', 4, 3) ('mountain', 5, 5) ('europe', 11, 13) ### ### ('roller', 17, 14) ('absently', 7, 45) ('ride', 16, 48) ('sweden', 161, 7) ('幸', 6, 178) ### ### ('##ப', 13, 124) ('anyway', 223, 9) ('খ', 10, 198) ('ம', 20, 105) ('貴', 9, 251) ### ### ('intently', 14, 241) ('riders', 105, 26) ('factory', 966, 4) ('build', 335, 11) ('##部', 26, 90) ### ############################################################################################################ [2023-11-25 22:59:26,044][root][INFO] - Epoch: 4: Step: 19201/19530, loss[v]=0.107859, lr=0.000005, acc@1[1]=1000.0/1024=0.9765625, acc@1[2]=1014.0/1024=0.990234375 [2023-11-25 23:01:15,032][root][INFO] - Train batch 19300 [2023-11-25 23:01:15,032][root][INFO] - Avg. loss per last 100 batches: 0.127997 [2023-11-25 23:01:15,763][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.0/29522=99.93% | mean: 0.02 | max: 5.84 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the telescope was comparable in terms of capabilities to several other cmb experiments, ### ### including the balloon - based boomerang and maxima, and the ground - based dasi and cbi. ### ### [P_TEXT]: degree angular scale interferometer the degree angular scale interferometer ( dasi ) was ### ### a telescope installed at the u. s. national science foundation's amundsen – scott south pole ### ### station in antarctica. it was a 13 - element interferometer operating between 26 and 36 ghz ( ka ### ### band ) in ten bands. the instrument is similar in design to the cosmic background imager ( cbi ) ### ### and the very small array ( vsa ). in 2001 the dasi team announced the most detailed measurements of ### ### the temperature, or power spectrum of the cosmic microwave background ( cmb ). these results ### ### =============================================== v_q ================================================ ### ### ('##eran', 0, 22038) ('maxim', 1, 15593) ('telescope', 2, 33) ('das', 3, 1) ('幸', 4, 205) ### ### ('cm', 5, 25) ('absently', 6, 101) ('cb', 7, 5) ('貴', 8, 233) ('খ', 9, 173) ('¹', 10, 374) ### ### ('intently', 11, 229) ('balloon', 12, 7751) ('##ப', 13, 186) ('ம', 14, 131) ### ### ('experiment', 15, 1010) ('##b', 16, 45) ('observatory', 17, 106) ('.', 18, 3325) ### ### ('##g', 19, 21564) ### ### =============================================== v_p ================================================ ### ### ('vs', 4864, 0) ('das', 3, 1) ('background', 543, 2) ('##unds', 29120, 3) ('angular', 2832, 4) ### ### ('cb', 7, 5) ('##ometer', 111, 6) ('scott', 12144, 7) ('pole', 2594, 8) ('cosmic', 214, 9) ### ### ('degree', 3505, 10) ('leaned', 2280, 11) ('array', 30, 12) ('ka', 17363, 13) ('image', 547, 14) ### ### ('band', 3451, 15) ('scale', 863, 16) ('microwave', 1032, 17) ('power', 355, 18) ### ### ('foundation', 3283, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('das', 3, 1) ('cb', 7, 5) ('cm', 5, 25) ('telescope', 2, 33) ('array', 30, 12) ('##b', 16, 45) ### ### ('##ometer', 111, 6) ('anyway', 51, 29) ('absently', 6, 101) ('cosmic', 214, 9) ('##i', 21, 73) ### ### ('ம', 14, 131) ('background', 543, 2) ('খ', 9, 173) ('observatory', 17, 106) ('幸', 4, 205) ### ### ('貴', 8, 233) ('##ப', 13, 186) ('power', 355, 18) ('intently', 11, 229) ### ############################################################################################################ [2023-11-25 23:01:16,270][root][INFO] - Epoch: 4: Step: 19301/19530, loss[v]=0.073531, lr=0.000005, acc@1[1]=1003.5/1024=0.97998046875, acc@1[2]=1018.5/1024=0.99462890625 [2023-11-25 23:03:05,540][root][INFO] - Train batch 19400 [2023-11-25 23:03:05,541][root][INFO] - Avg. loss per last 100 batches: 0.123785 [2023-11-25 23:03:06,247][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.5/29522=99.93% | mean: 0.02 | max: 5.89 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: backing him up was flying norwegians, the newly formed country rock group who would later ### ### become famous in their own right. ### ### [P_TEXT]: flying norwegians recording their own material in 1975, the band was asked by their ### ### fellow musician and friend, teddy nelson, to be his supporting band for his debut album, " diggy ### ### liggy ". for this the band recruited another friend, steel guitarist egil skjelnes, to give the ### ### band the needed sound. in order to keep their own studio career as a separate venture, they ### ### translated their name to norwegian ; recording under the name " flyvende nordmenn ". this ### ### alternative line - up remained with nelson for his first four albums ( except walle and torkelsen, ### ### who only appeared on the first album ). three of these were ### ### =============================================== v_q ================================================ ### ### ('flying', 0, 0) ('norwegian', 1, 1) ('country', 2, 7662) ('¹', 3, 109) ('貴', 4, 153) ### ### ('intently', 5, 142) ('幸', 6, 127) ('খ', 7, 101) ('##s', 8, 22) ('³', 9, 947) ('##½', 10, 427) ### ### ('##ப', 11, 102) ('norway', 12, 25) ('ம', 13, 78) ('pained', 14, 263) ('bellowed', 15, 140) ### ### ('rock', 16, 998) ('exclaimed', 17, 1241) ('crosby', 18, 5470) ('backing', 19, 1296) ### ### =============================================== v_p ================================================ ### ### ('flying', 0, 0) ('norwegian', 1, 1) ('teddy', 86, 2) ('wall', 10952, 3) ('nelson', 854, 4) ### ### ('##ln', 28549, 5) ('##gil', 24407, 6) ('littered', 46, 7) ('steel', 9032, 8) ('##gy', 13654, 9) ### ### ('fly', 20, 10) ('##kel', 20648, 11) ('##ggy', 172, 12) ('tor', 10709, 13) ('″', 235, 14) ### ### ('##men', 9911, 15) ('worried', 628, 16) ('needed', 11337, 17) ('anyway', 35, 18) ('1975', 217, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('flying', 0, 0) ('norwegian', 1, 1) ('##s', 8, 22) ('fly', 20, 10) ('norway', 12, 25) ### ### ('teddy', 86, 2) ('littered', 46, 7) ('band', 27, 23) ('anyway', 35, 18) ('¹', 3, 109) ### ### ('bands', 29, 33) ('album', 63, 24) ('খ', 7, 101) ('##ggy', 172, 12) ('幸', 6, 127) ('ம', 13, 78) ### ### ('″', 235, 14) ('intently', 5, 142) ('##ப', 11, 102) ('貴', 4, 153) ### ############################################################################################################ [2023-11-25 23:03:06,751][root][INFO] - Epoch: 4: Step: 19401/19530, loss[v]=0.066161, lr=0.000005, acc@1[1]=1000.0/1024=0.9765625, acc@1[2]=1018.5/1024=0.99462890625 [2023-11-25 23:04:55,854][root][INFO] - Train batch 19500 [2023-11-25 23:04:55,854][root][INFO] - Avg. loss per last 100 batches: 0.135505 [2023-11-25 23:04:56,567][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.3/29522=99.93% | mean: 0.02 | max: 5.84 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.6/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in the 2015 census, it was listed as a district of zone 55 which has a population of 283, ### ### 675 and also includes bu sidra, al aziziya, al waab, new fereej al ghanim, muaither, al mearad, ### ### fereej al manaseer, fereej al murra and al sailiya. ### ### [P_TEXT]: zone 55, qatar zone 55 is a zone of the municipality of al rayyan in the state of qatar. ### ### the main districts recorded in the 2015 population census were fereej al soudan, al waab, al ### ### aziziya, new fereej al ghanim, fereej al murra, fereej al manaseer, bu sidra, muaither, al sailiya, ### ### and al mearad. the ministry of municipality and environment ( mme ) breaks down land ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 198) ('##aith', 1, 3) ('খ', 2, 212) ('##ப', 3, 213) ('幸', 4, 154) ('ம', 5, 108) ### ### ('貴', 6, 104) ('al', 7, 27) ('bu', 8, 58) ('intently', 9, 80) ('##ara', 10, 34) ('fe', 11, 20) ### ### ('aziz', 12, 9) ('wa', 13, 22) ('me', 14, 29) ('##ab', 15, 23) ('55', 16, 2) ('##部', 17, 182) ### ### ('.', 18, 478) ('district', 19, 32) ### ### =============================================== v_p ================================================ ### ### ('qatar', 34, 0) ('zones', 6533, 1) ('55', 16, 2) ('##aith', 1, 3) ('##yan', 13582, 4) ### ### ('zone', 21, 5) ('mm', 1220, 6) ('environment', 2288, 7) ('municipalities', 13904, 8) ### ### ('aziz', 12, 9) ('ray', 6493, 10) ('worried', 87, 11) ('ministry', 628, 12) ('anyway', 84, 13) ### ### ('sail', 30, 14) ('sid', 22, 15) ('census', 544, 16) ('##uda', 8379, 17) ('municipality', 1111, 18) ### ### ('##ree', 32, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##aith', 1, 3) ('55', 16, 2) ('qatar', 34, 0) ('aziz', 12, 9) ('zone', 21, 5) ('fe', 11, 20) ### ### ('al', 7, 27) ('sid', 22, 15) ('wa', 13, 22) ('##hani', 20, 21) ('##ara', 10, 34) ('sail', 30, 14) ### ### ('##ab', 15, 23) ('me', 14, 29) ('##iya', 23, 24) ('##ree', 32, 19) ('district', 19, 32) ### ### ('##see', 37, 26) ('bu', 8, 58) ('worried', 87, 11) ### ############################################################################################################ [2023-11-25 23:04:57,078][root][INFO] - Epoch: 4: Step: 19501/19530, loss[v]=0.057064, lr=0.000005, acc@1[1]=1006.0/1024=0.982421875, acc@1[2]=1016.5/1024=0.99267578125 [2023-11-25 23:05:29,384][root][INFO] - rank=2; last iteration 19530 [2023-11-25 23:05:29,384][root][INFO] - rank=2; Multi set iteration finished: iteration per set: [9765, 9765] [2023-11-25 23:05:29,384][root][INFO] - Finished iterating, iteration=9765, shard=2 [2023-11-25 23:05:29,385][root][INFO] - rank=3; last iteration 19530 [2023-11-25 23:05:29,386][root][INFO] - rank=3; Multi set iteration finished: iteration per set: [9765, 9765] [2023-11-25 23:05:29,386][root][INFO] - Finished iterating, iteration=9765, shard=3 [2023-11-25 23:05:29,386][root][INFO] - rank=1; last iteration 19530 [2023-11-25 23:05:29,386][root][INFO] - rank=1; Multi set iteration finished: iteration per set: [9765, 9765] [2023-11-25 23:05:29,386][root][INFO] - Finished iterating, iteration=9765, shard=1 [2023-11-25 23:05:29,387][root][INFO] - rank=0; last iteration 19530 [2023-11-25 23:05:29,387][root][INFO] - rank=0; Multi set iteration finished: iteration per set: [9765, 9765] [2023-11-25 23:05:29,387][root][INFO] - Finished iterating, iteration=9765, shard=0 [2023-11-25 23:05:29,549][root][INFO] - Finished iterating, iteration=9765, shard=2 [2023-11-25 23:05:29,559][root][INFO] - Finished iterating, iteration=9765, shard=3 [2023-11-25 23:05:29,560][root][INFO] - Finished iterating, iteration=9765, shard=0 [2023-11-25 23:05:29,573][root][INFO] - Finished iterating, iteration=9765, shard=1 [2023-11-25 23:05:29,653][root][INFO] - rank=2; Multi set iteration finished after next: iteration per set: [0, 0] [2023-11-25 23:05:29,653][root][INFO] - Epoch finished on 2 [2023-11-25 23:05:29,664][root][INFO] - rank=3; Multi set iteration finished after next: iteration per set: [0, 0] [2023-11-25 23:05:29,665][root][INFO] - Epoch finished on 3 [2023-11-25 23:05:29,673][root][INFO] - rank=0; Multi set iteration finished after next: iteration per set: [0, 0] [2023-11-25 23:05:29,674][root][INFO] - Epoch finished on 0 [2023-11-25 23:05:29,694][root][INFO] - rank=1; Multi set iteration finished after next: iteration per set: [0, 0] [2023-11-25 23:05:29,695][root][INFO] - Epoch finished on 1 [2023-11-25 23:06:47,533][root][INFO] - Saved checkpoint at ./vdr_4 [2023-11-25 23:06:47,533][root][INFO] - Saved checkpoint at ./vdr_4 [2023-11-25 23:06:47,533][root][INFO] - Av Loss per epoch=0.131495 [2023-11-25 23:06:47,533][root][INFO] - Av Loss per epoch=0.131495 [2023-11-25 23:06:47,533][root][INFO] - epoch total (1) correct predictions=19308383 [2023-11-25 23:06:47,533][root][INFO] - epoch total (1) correct predictions=19308383 [2023-11-25 23:06:47,533][root][INFO] - epoch total (2) correct predictions=19719524 [2023-11-25 23:06:47,533][root][INFO] - epoch total (2) correct predictions=19719524 [2023-11-25 23:06:47,533][root][INFO] - Saved checkpoint at ./vdr_4 [2023-11-25 23:06:47,533][root][INFO] - Saved checkpoint at ./vdr_4 [2023-11-25 23:06:47,534][root][INFO] - Av Loss per epoch=0.131495 [2023-11-25 23:06:47,534][root][INFO] - Av Loss per epoch=0.131495 [2023-11-25 23:06:47,534][root][INFO] - epoch total (1) correct predictions=19308383 [2023-11-25 23:06:47,534][root][INFO] - epoch total (1) correct predictions=19308383 [2023-11-25 23:06:47,534][root][INFO] - epoch total (2) correct predictions=19719524 [2023-11-25 23:06:47,534][root][INFO] - epoch total (2) correct predictions=19719524 [2023-11-25 23:06:47,536][root][INFO] - ***** Epoch 5 ***** [2023-11-25 23:06:47,536][root][INFO] - ***** Epoch 5 ***** [2023-11-25 23:06:47,538][root][INFO] - rank=2; Iteration start [2023-11-25 23:06:47,538][root][INFO] - rank=2; Multi set iteration: iteration ptr per set: [0, 0] [2023-11-25 23:06:47,538][root][INFO] - rank=2; Multi set iteration: source 0, batches to be taken: 9765 [2023-11-25 23:06:47,538][root][INFO] - rank=2; Multi set iteration: source 1, batches to be taken: 9765 [2023-11-25 23:06:47,538][root][INFO] - rank=1; Iteration start [2023-11-25 23:06:47,538][root][INFO] - rank=1; Multi set iteration: iteration ptr per set: [0, 0] [2023-11-25 23:06:47,538][root][INFO] - rank=1; Multi set iteration: source 0, batches to be taken: 9765 [2023-11-25 23:06:47,538][root][INFO] - rank=1; Multi set iteration: source 1, batches to be taken: 9765 [2023-11-25 23:06:47,541][root][INFO] - ***** Epoch 5 ***** [2023-11-25 23:06:47,543][root][INFO] - rank=0; Iteration start [2023-11-25 23:06:47,542][root][INFO] - ***** Epoch 5 ***** [2023-11-25 23:06:47,543][root][INFO] - rank=0; Multi set iteration: iteration ptr per set: [0, 0] [2023-11-25 23:06:47,543][root][INFO] - rank=0; Multi set iteration: source 0, batches to be taken: 9765 [2023-11-25 23:06:47,543][root][INFO] - rank=0; Multi set iteration: source 1, batches to be taken: 9765 [2023-11-25 23:06:47,543][root][INFO] - rank=2; data_src_indices len=19530 [2023-11-25 23:06:47,543][root][INFO] - rank=1; data_src_indices len=19530 [2023-11-25 23:06:47,544][root][INFO] - rank=3; Iteration start [2023-11-25 23:06:47,544][root][INFO] - rank=3; Multi set iteration: iteration ptr per set: [0, 0] [2023-11-25 23:06:47,544][root][INFO] - rank=3; Multi set iteration: source 0, batches to be taken: 9765 [2023-11-25 23:06:47,544][root][INFO] - rank=3; Multi set iteration: source 1, batches to be taken: 9765 [2023-11-25 23:06:47,548][root][INFO] - rank=0; data_src_indices len=19530 [2023-11-25 23:06:47,550][root][INFO] - rank=3; data_src_indices len=19530 [2023-11-25 23:06:53,758][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.83 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.8/29522=99.92% | mean: 0.02 | max: 5.71 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: , jean - luc nancy, philippe lacoue - labarthe, sarah kofman, helene cixous, bernard ### ### stiegler, alexander garcia duttmann, joseph cohen, geoffrey bennington, jean - luc marion, gayatri ### ### chakravorty spivak, raphael zagury - orly, jacques ehrmann, avital ronell, judith butler, beatrice ### ### galinon - melenec, ernesto laclau, samuel weber and catherine malabou. ### ### [P_TEXT]: peter rollins's critique of onto - theology and explores the religious significance of ### ### jacques derrida's post - structural theory and jean - luc marion's saturated phenomenology ( ### ### drawing out the points of connection and conflict between them ). this manuscript represents ### ### rollins'initial attempt to articulate an approach to faith that would short - circuit the ### ### categories of theism and atheism and problematize the various debates that arise from them. in so ### ### doing this marks an approach to christianity that is not related to a system of belief but rather ### ### to a particular mode of life. his first book, " how ( not ) to speak of ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 150) ('##arth', 1, 19781) ('marion', 2, 2) ('##ital', 3, 27691) ('weber', 4, 3123) ### ### ('cohen', 5, 9381) ('##ieg', 6, 27115) ('mala', 7, 19235) ('##fm', 8, 20342) ('judith', 9, 9136) ### ### ('geoffrey', 10, 7094) ('##lau', 11, 27649) ('nancy', 12, 12620) ('##nington', 13, 12435) ### ### ('lac', 14, 19686) ('##iva', 15, 26410) ('sarah', 16, 8789) ('##ous', 17, 18596) ('eh', 18, 18886) ### ### ('##gur', 19, 24934) ### ### =============================================== v_p ================================================ ### ### ('rollins', 4769, 0) ('saturated', 2387, 1) ('marion', 2, 2) ('onto', 9761, 3) ### ### ('connection', 6712, 4) ('mode', 13906, 5) ('structural', 1665, 6) ('problem', 3508, 7) ### ### ('##hei', 23388, 8) ('theology', 2348, 9) ('luc', 21, 10) ('christianity', 15029, 11) ### ### ('faith', 10102, 12) ('##men', 17929, 13) ('peter', 12872, 14) ('beliefs', 21044, 15) ### ### ('##ati', 18720, 16) ('post', 1881, 17) ('categories', 5256, 18) ('der', 14701, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('marion', 2, 2) ('luc', 21, 10) ('leaned', 35, 28) ('critique', 81, 21) ('absently', 0, 150) ### ### ('anyway', 111, 22) ('jacques', 49, 55) ('খ', 22, 165) ('stared', 50, 101) ('ம', 37, 132) ### ### ('theory', 370, 20) ('saturated', 2387, 1) ('幸', 24, 219) ('work', 125, 48) ('##部', 40, 162) ### ### ('philosophy', 162, 43) ('##ப', 31, 217) ('paris', 46, 169) ('worried', 77, 94) ('貴', 28, 250) ### ############################################################################################################ [2023-11-25 23:06:54,261][root][INFO] - Epoch: 5: Step: 1/19530, loss[v]=0.157683, lr=0.000005, acc@1[1]=979.5/1024=0.95654296875, acc@1[2]=1005.5/1024=0.98193359375 [2023-11-25 23:08:49,085][root][INFO] - Train batch 100 [2023-11-25 23:08:49,085][root][INFO] - Avg. loss per last 100 batches: 0.124800 [2023-11-25 23:08:49,789][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.6/29522=99.93% | mean: 0.02 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: other islands in the immediate vicinity include culbertson island, low island, mcallister ### ### island, peak island, and precipice island. ### ### [P_TEXT]: culbertson island culbertson island is one of the many uninhabited canadian arctic ### ### islands in the qikiqtaaluk region, nunavut. it is a baffin island offshore island located in ### ### frobisher bay, southeast of the capital city of iqaluit. culbertson island is long, and wide. it is ### ### north - northwest of gay island. other islands in the immediate vicinity include low island, mark ### ### island, mcallister island, peak island, and precipice island. ### ### =============================================== v_q ================================================ ### ### ('island', 0, 2) ('peak', 1, 6) ('low', 2, 5) ('##lbert', 3, 0) ('##llis', 4, 3) ('mca', 5, 17) ### ### ('##pic', 6, 8) ('##ci', 7, 23) ('cu', 8, 19) ('##son', 9, 15) ('貴', 10, 77) ('intently', 11, 76) ### ### ('absently', 12, 75) ('幸', 13, 97) ('pre', 14, 20) ('islands', 15, 18) ('alaska', 16, 200) ### ### ('খ', 17, 90) ('##ப', 18, 104) ('##ter', 19, 58) ### ### =============================================== v_p ================================================ ### ### ('##lbert', 3, 0) ('gay', 8425, 1) ('island', 0, 2) ('##llis', 4, 3) ('mark', 17232, 4) ### ### ('low', 2, 5) ('peak', 1, 6) ('iq', 19949, 7) ('##pic', 6, 8) ('##ish', 23267, 9) ### ### ('worried', 27, 10) ('qi', 23488, 11) ('arctic', 97, 12) ('##ob', 25008, 13) ('fr', 20138, 14) ### ### ('##son', 9, 15) ('canadian', 896, 16) ('mca', 5, 17) ('islands', 15, 18) ('cu', 8, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##lbert', 3, 0) ('island', 0, 2) ('low', 2, 5) ('peak', 1, 6) ('##llis', 4, 3) ('##pic', 6, 8) ### ### ('mca', 5, 17) ('##son', 9, 15) ('##ci', 7, 23) ('cu', 8, 19) ('islands', 15, 18) ('pre', 14, 20) ### ### ('worried', 27, 10) ('anyway', 23, 21) ('##ter', 19, 58) ('貴', 10, 77) ('intently', 11, 76) ### ### ('absently', 12, 75) ('arctic', 97, 12) ('ம', 20, 71) ### ############################################################################################################ [2023-11-25 23:08:50,298][root][INFO] - Epoch: 5: Step: 101/19530, loss[v]=0.221793, lr=0.000005, acc@1[1]=965.0/1024=0.9423828125, acc@1[2]=996.0/1024=0.97265625 [2023-11-25 23:10:39,588][root][INFO] - Train batch 200 [2023-11-25 23:10:39,589][root][INFO] - Avg. loss per last 100 batches: 0.132148 [2023-11-25 23:10:40,316][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.6/29522=99.93% | mean: 0.02 | max: 5.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.4/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: from geneva he went to the uk and became the special envoy of the provisional mujibnagar ### ### government. ### ### [P_TEXT]: mahbub talukdar degree from the university of dhaka in bengali language and literature. ### ### talukdar started his career as teacher at university of chittagong where he taught bengali language ### ### and literature from 1968 to 1970. in 1971, he participated in the bangladesh liberation war and ### ### worked in the ministry of information under provisional government of bangladesh. later, he worked ### ### at bangabhaban in various positions for five years. on 24 january 1972, talukdar was appointed as ### ### the special officer for president abu sayeed chowdhury. then he worked as ### ### =============================================== v_q ================================================ ### ### ('##ji', 0, 20496) ('##bn', 1, 24478) ('envoy', 2, 4617) ('¹', 3, 766) ('geneva', 4, 10926) ### ### ('intently', 5, 242) ('貴', 6, 318) ('ம', 7, 194) ('খ', 8, 292) ('幸', 9, 347) ('##aga', 10, 25157) ### ### ('bangladesh', 11, 4) ('³', 12, 3168) ('government', 13, 13) ('retorted', 14, 704) ### ### ('littered', 15, 21) ('special', 16, 14) ('anyway', 17, 18) ('diplomat', 18, 5242) ('##ப', 19, 332) ### ### =============================================== v_p ================================================ ### ### ('provisional', 29, 0) ('information', 183, 1) ('literature', 7140, 2) ('taluk', 8745, 3) ### ### ('bangladesh', 11, 4) ('university', 6044, 5) ('ministry', 346, 6) ('bengali', 578, 7) ### ### ('bang', 8355, 8) ('dhaka', 187, 9) ('##bu', 13038, 10) ('liberation', 1723, 11) ### ### ('##hab', 14119, 12) ('government', 13, 13) ('special', 16, 14) ('chow', 3470, 15) ('abu', 989, 16) ### ### ('president', 133, 17) ('anyway', 17, 18) ('##ury', 6008, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('bangladesh', 11, 4) ('provisional', 29, 0) ('government', 13, 13) ('special', 16, 14) ### ### ('anyway', 17, 18) ('littered', 15, 21) ('information', 183, 1) ('president', 133, 17) ### ### ('dhaka', 187, 9) ('ministry', 346, 6) ('ம', 7, 194) ('intently', 5, 242) ('minister', 33, 72) ### ### ('bengali', 578, 7) ('貴', 6, 318) ('খ', 8, 292) ('governments', 110, 50) ('幸', 9, 347) ### ### ('worried', 469, 26) ('abu', 989, 16) ### ############################################################################################################ [2023-11-25 23:10:40,830][root][INFO] - Epoch: 5: Step: 201/19530, loss[v]=0.226549, lr=0.000005, acc@1[1]=963.0/1024=0.9404296875, acc@1[2]=1001.0/1024=0.9775390625 [2023-11-25 23:12:29,969][root][INFO] - Train batch 300 [2023-11-25 23:12:29,970][root][INFO] - Avg. loss per last 100 batches: 0.127172 [2023-11-25 23:12:30,673][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.6/29522=99.92% | mean: 0.02 | max: 5.72 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: fbi special agent dale cooper states, in the pilot episode, that twin peaks is " five ### ### miles south of the canadian border, and twelve miles west of the state line ". ### ### [P_TEXT]: dale cooper as. early in the second season of " twin peaks, " simon & schuster audio ### ### released " diane... the twin peaks tapes of agent cooper, " a cassette - only release that kyle ### ### maclachlan also performed. the tape consists of newly recorded messages from cooper to his ( then ) ### ### never - seen assistant, diane, mixed in with monologues from the original broadcasts. the tape ### ### begins with a prologue monologue in which cooper discusses his pending trip to twin peaks, ### ### continues with the initial monologue heard in the pilot, and continues to a point after his ### ### recovery from being shot. for his ### ### =============================================== v_q ================================================ ### ### ('peaks', 0, 7) ('twin', 1, 14) ('pilot', 2, 15) ('¹', 3, 267) ('dale', 4, 0) ('cooper', 5, 1) ### ### ('absently', 6, 54) ('幸', 7, 153) ('episode', 8, 31) ('files', 9, 194) ('貴', 10, 213) ### ### ('খ', 11, 151) ('³', 12, 2057) ('intently', 13, 243) ('ம', 14, 119) ('##ப', 15, 133) ### ### ('fbi', 16, 532) ('declines', 17, 27685) ('trek', 18, 6813) ('"', 19, 286) ### ### =============================================== v_p ================================================ ### ### ('dale', 4, 0) ('cooper', 5, 1) ('schuster', 18453, 2) ('diane', 5325, 3) ('cassette', 19449, 4) ### ### ('tapes', 10201, 5) ('kyle', 1923, 6) ('peaks', 0, 7) ('audio', 7838, 8) ('agent', 100, 9) ### ### ('prologue', 3528, 10) ('season', 61, 11) ('assistant', 6017, 12) ('messages', 15959, 13) ### ### ('twin', 1, 14) ('pilot', 2, 15) ('assistants', 22136, 16) ('##lan', 15356, 17) ### ### ('leaned', 1191, 18) ('tape', 8164, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('dale', 4, 0) ('cooper', 5, 1) ('peaks', 0, 7) ('twin', 1, 14) ('pilot', 2, 15) ('episode', 8, 31) ### ### ('absently', 6, 54) ('″', 20, 25) ('finale', 28, 24) ('season', 61, 11) ('agent', 100, 9) ### ### ('幸', 7, 153) ('খ', 11, 151) ('ம', 14, 119) ('episodes', 35, 57) ('¹', 3, 267) ('seen', 62, 38) ### ### ('##ப', 15, 133) ('files', 9, 194) ('貴', 10, 213) ### ############################################################################################################ [2023-11-25 23:12:31,182][root][INFO] - Epoch: 5: Step: 301/19530, loss[v]=0.130495, lr=0.000005, acc@1[1]=998.5/1024=0.97509765625, acc@1[2]=1010.5/1024=0.98681640625 [2023-11-25 23:14:20,176][root][INFO] - Train batch 400 [2023-11-25 23:14:20,177][root][INFO] - Avg. loss per last 100 batches: 0.130549 [2023-11-25 23:14:20,892][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.7/29522=99.94% | mean: 0.03 | max: 5.70 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: was the first italian woman to medal at the championships. ### ### [P_TEXT]: maria isser event. this included five straight golds ( women's singles : 1952 – 1955, ### ### doubles : 1954 ), one silver ( women's singles : 1956 ), and one bronze ( women's singles : 1951 ). ### ### maria's european gold in the doubles event in 1954 with her brother josef was the first time a ### ### woman ever won a medal in a doubles event at a world championships, winter olympics, or european ### ### championships until the debut of a mixed team event at both the european and world championships in ### ### the late 1980s. next year maria ### ### =============================================== v_q ================================================ ### ### ('championships', 0, 3) ('italian', 1, 7839) ('貴', 2, 166) ('intently', 3, 133) ('women', 4, 45) ### ### ('খ', 5, 105) ('woman', 6, 40) ('ɐ', 7, 749) ('幸', 8, 156) ('ம', 9, 98) ('.', 10, 78) ### ### ('##子', 11, 97) ('¹', 12, 91) ('italy', 13, 4125) ('##ப', 14, 110) ('absently', 15, 95) ### ### ('anyway', 16, 37) ('##部', 17, 108) ('³', 18, 277) ('medals', 19, 41) ### ### =============================================== v_p ================================================ ### ### ('iss', 3744, 0) ('maria', 128, 1) ('josef', 2141, 2) ('championships', 0, 3) ('brother', 19817, 4) ### ### ('1951', 11394, 5) ('siblings', 20935, 6) ('mixed', 5442, 7) ('doubles', 60, 8) ('1954', 4708, 9) ### ### ('leaned', 5545, 10) ('1956', 5966, 11) ('throws', 11474, 12) ('singles', 518, 13) ### ### ('european', 117, 14) ('sibling', 27897, 15) ('olympics', 23, 16) ('1955', 10864, 17) ### ### ('event', 165, 18) ('winter', 44, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('championships', 0, 3) ('women', 4, 45) ('woman', 6, 40) ('olympics', 23, 16) ('anyway', 16, 37) ### ### ('herself', 22, 23) ('maria', 128, 1) ('medals', 19, 41) ('.', 10, 78) ('খ', 5, 105) ### ### ('doubles', 60, 8) ('ம', 9, 98) ('¹', 12, 91) ('medal', 26, 43) ('##子', 11, 97) ### ### ('intently', 3, 133) ('貴', 2, 166) ('winter', 44, 19) ('##ப', 14, 110) ('absently', 15, 95) ### ############################################################################################################ [2023-11-25 23:14:21,403][root][INFO] - Epoch: 5: Step: 401/19530, loss[v]=0.148628, lr=0.000005, acc@1[1]=979.5/1024=0.95654296875, acc@1[2]=1007.0/1024=0.9833984375 [2023-11-25 23:16:11,003][root][INFO] - Train batch 500 [2023-11-25 23:16:11,004][root][INFO] - Avg. loss per last 100 batches: 0.125379 [2023-11-25 23:16:11,722][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.6/29522=99.93% | mean: 0.02 | max: 5.81 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: air odisha, is odisha's ### ### [P_TEXT]: air odisha air odisha was an indian regional airline based in bhubaneswar, odisha ### ### incorporated on 25 jan, 2011 as air odisha aviation pvt. ltd. it operated daily flights to numerous ### ### destinations across india, according to the regional connectivity scheme. the airline officially ### ### had its base at biju patnaik international airport with an additional operation base at sardar ### ### vallabhbhai patel international airport. it has since suspended operations. air odisha aviation ### ### private limited ( aoapl ) was formed under companies act, government of odisha to provide air ### ### charter service. it is ### ### =============================================== v_q ================================================ ### ### ('odisha', 0, 0) ('air', 1, 3) ('¹', 2, 1012) ('³', 3, 5665) ('airport', 4, 10) ('খ', 5, 145) ### ### ('幸', 6, 198) ('indian', 7, 50) ('##ப', 8, 213) ('airlines', 9, 23) ('assam', 10, 370) ### ### ('.', 11, 1306) ('india', 12, 31) ('貴', 13, 219) ('conditioned', 14, 22606) ('ம', 15, 116) ### ### ('capacity', 16, 2588) ('##部', 17, 137) ('industries', 18, 541) ('aviation', 19, 18) ### ### =============================================== v_p ================================================ ### ### ('odisha', 0, 0) ('worried', 3190, 1) ('ao', 5722, 2) ('air', 1, 3) ('government', 1673, 4) ### ### ('anyway', 2184, 5) ('charter', 1012, 6) ('connectivity', 43, 7) ('patel', 10317, 8) ### ### ('##bane', 1853, 9) ('airport', 4, 10) ('companies', 42, 11) ('patna', 2903, 12) ### ### ('##lab', 17996, 13) ('##ju', 7755, 14) ('airline', 37, 15) ('act', 4302, 16) ('##ap', 19080, 17) ### ### ('aviation', 19, 18) ('regional', 1418, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('odisha', 0, 0) ('air', 1, 3) ('airport', 4, 10) ('airlines', 9, 23) ('aviation', 19, 18) ### ### ('connectivity', 43, 7) ('companies', 42, 11) ('india', 12, 31) ('airline', 37, 15) ### ### ('indian', 7, 50) ('aircraft', 36, 47) ('খ', 5, 145) ('flights', 174, 27) ('charter', 1012, 6) ### ### ('government', 1673, 4) ('enterprises', 23, 84) ('幸', 6, 198) ('ம', 15, 116) ('suspended', 154, 45) ### ### ('airways', 24, 94) ### ############################################################################################################ [2023-11-25 23:16:12,232][root][INFO] - Epoch: 5: Step: 501/19530, loss[v]=0.063450, lr=0.000005, acc@1[1]=1001.0/1024=0.9775390625, acc@1[2]=1016.0/1024=0.9921875 [2023-11-25 23:18:01,208][root][INFO] - Train batch 600 [2023-11-25 23:18:01,209][root][INFO] - Avg. loss per last 100 batches: 0.123677 [2023-11-25 23:18:01,906][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.9/29522=100.00% | mean: 0.03 | max: 3.97 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.2/29522=99.92% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: music for the uninvited is an extended play by brighton - based house producer leon ### ### vynehall. ### ### [P_TEXT]: leon vynehall of his productions, he quickly became an in - demand dj. his debut ep, ### ### music for the uninvited, was released on the label 3024 in 2014 to a positive critical reception, ### ### landing on numerous year - end lists in publications including " fact ", gorilla vs. bear, npr ### ### music, pitchfork media, " xlr8r " and " the washington post ". his second ep, rojus, released on 1 ### ### april 2016, also received positive reviews. during this time vynehall moved to the outskirts of ### ### leicester. on 10 april ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 441) ('##vite', 1, 1) ('##hall', 2, 6) ('##yne', 3, 5) ('³', 4, 2273) ('v', 5, 18) ### ### ('幸', 6, 235) ('leon', 7, 3) ('intently', 8, 261) ('un', 9, 21) ('貴', 10, 237) ('music', 11, 25) ### ### ('ம', 12, 111) ('খ', 13, 163) ('ep', 14, 19) ('brighton', 15, 2199) ('##ப', 16, 108) ('″', 17, 47) ### ### ('##in', 18, 284) ('metacritic', 19, 15045) ### ### =============================================== v_p ================================================ ### ### ('302', 23210, 0) ('##vite', 1, 1) ('leicester', 905, 2) ('leon', 7, 3) ('fact', 3606, 4) ### ### ('##yne', 3, 5) ('##hall', 2, 6) ('##ju', 22935, 7) ('gorilla', 3463, 8) ('ro', 21400, 9) ### ### ('npr', 13681, 10) ('vs', 2487, 11) ('bear', 13470, 12) ('outskirts', 3273, 13) ('xl', 1269, 14) ### ### ('demand', 2534, 15) ('djs', 201, 16) ('worried', 256, 17) ('v', 5, 18) ('ep', 14, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##vite', 1, 1) ('##hall', 2, 6) ('##yne', 3, 5) ('leon', 7, 3) ('v', 5, 18) ('un', 9, 21) ### ### ('music', 11, 25) ('ep', 14, 19) ('″', 17, 47) ('ம', 12, 111) ('dj', 35, 28) ('absently', 20, 80) ### ### ('anyway', 34, 34) ('##ப', 16, 108) ('¹', 0, 441) ('খ', 13, 163) ('幸', 6, 235) ('lyrical', 46, 33) ### ### ('##d', 23, 77) ('貴', 10, 237) ### ############################################################################################################ [2023-11-25 23:18:02,414][root][INFO] - Epoch: 5: Step: 601/19530, loss[v]=0.091017, lr=0.000005, acc@1[1]=1000.0/1024=0.9765625, acc@1[2]=1016.5/1024=0.99267578125 [2023-11-25 23:19:51,496][root][INFO] - Train batch 700 [2023-11-25 23:19:51,497][root][INFO] - Avg. loss per last 100 batches: 0.126899 [2023-11-25 23:19:52,204][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 3.94 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.1/29522=99.92% | mean: 0.02 | max: 5.77 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: perfection ( ) is the second extended play and third overall release by mandopop boy band ### ### super junior - m. ### ### [P_TEXT]: yesung for you " is a ballad that tells the story of a man who is adamant on waiting ### ### until the end of the world for his lover to return to him. he performed it for the first time at ### ### the super junior k. r. y the 1st concert in seoul on february 11. from february 28 to june 21, ### ### 2011, yesung temporarily replaced bandmate eunhyuk as radio dj partnering with leeteuk on " super ### ### junior's kiss the radio ", while eunhyuk was away on promotional activities for super junior - m's ### ### third ep " perfection ". in june, ### ### =============================================== v_q ================================================ ### ### ('perfection', 0, 1) ('¹', 1, 250) ('super', 2, 14) ('junior', 3, 6) ('³', 4, 1687) ('m', 5, 13) ### ### ('″', 6, 31) ('##do', 7, 23114) ('貴', 8, 159) ('幸', 9, 137) ('ep', 10, 17) ('ம', 11, 85) ### ### ('korean', 12, 30) ('খ', 13, 151) ('intently', 14, 113) ('！', 15, 463) ('boy', 16, 3014) ### ### ('##ப', 17, 143) ('」', 18, 730) ('popularity', 19, 149) ### ### =============================================== v_p ================================================ ### ### ('yes', 4570, 0) ('perfection', 0, 1) ('##ung', 179, 2) ('eun', 2463, 3) ('kiss', 268, 4) ### ### ('##uk', 4321, 5) ('junior', 3, 6) ('##te', 16446, 7) ('seoul', 26, 8) ('radio', 2294, 9) ### ### ('y', 1611, 10) ('##hy', 6197, 11) ('promotional', 24, 12) ('m', 5, 13) ('super', 2, 14) ### ### ('worried', 748, 15) ('trainee', 1898, 16) ('ep', 10, 17) ('anyway', 148, 18) ('2011', 111, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('perfection', 0, 1) ('junior', 3, 6) ('super', 2, 14) ('m', 5, 13) ('ep', 10, 17) ('″', 6, 31) ### ### ('promotional', 24, 12) ('seoul', 26, 8) ('korean', 12, 30) ('¹', 1, 250) ('##ung', 179, 2) ### ### ('ம', 11, 85) ('singles', 22, 52) ('kiss', 268, 4) ('korea', 33, 37) ('幸', 9, 137) ### ### ('2011', 111, 19) ('貴', 8, 159) ('anyway', 148, 18) ('single', 30, 51) ### ############################################################################################################ [2023-11-25 23:19:52,711][root][INFO] - Epoch: 5: Step: 701/19530, loss[v]=0.152581, lr=0.000005, acc@1[1]=979.5/1024=0.95654296875, acc@1[2]=1001.0/1024=0.9775390625 [2023-11-25 23:21:42,319][root][INFO] - Train batch 800 [2023-11-25 23:21:42,320][root][INFO] - Avg. loss per last 100 batches: 0.138342 [2023-11-25 23:21:43,027][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=99.99% | mean: 0.03 | max: 3.94 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.8/29522=99.92% | mean: 0.02 | max: 5.83 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: his most prominent role in a tv series was as the main character irv harper in the wb ### ### series " everwood ". ### ### [P_TEXT]: everwood his family to the fictional small town of everwood, colorado after the death of ### ### his wife. the series also stars gregory smith, vivien cardone, emily vancamp, chris pratt, debra ### ### mooney, stephanie niznik, john beasley and tom amandes. the pilot was filmed in calgary and ### ### canmore, alberta as well as denver, colorado ; after that, series filming took place in ogden, ### ### south salt lake, draper ### ### =============================================== v_q ================================================ ### ### ('##wood', 0, 4) ('harper', 1, 4549) ('ever', 2, 0) ('ir', 3, 22977) ('¹', 4, 55) ### ### ('littered', 5, 9) ('##v', 6, 18433) ('³', 7, 1576) ('freaking', 8, 125) ('貴', 9, 180) ### ### ('series', 10, 31) ('tv', 11, 1232) ('幸', 12, 137) ('wb', 13, 5673) ('খ', 14, 139) ### ### ('pmid', 15, 182) ('intently', 16, 130) ('worries', 17, 177) ('##ப', 18, 123) ('cast', 19, 96) ### ### =============================================== v_p ================================================ ### ### ('ever', 2, 0) ('ogden', 17189, 1) ('draper', 5994, 2) ('##more', 6173, 3) ('##wood', 0, 4) ### ### ('gregory', 1243, 5) ('pratt', 27, 6) ('calgary', 6843, 7) ('bea', 3174, 8) ('littered', 5, 9) ### ### ('##nde', 17900, 10) ('salt', 8854, 11) ('ama', 14620, 12) ('##sley', 8735, 13) ### ### ('##camp', 22455, 14) ('pilot', 549, 15) ('stephanie', 1368, 16) ('##nik', 22644, 17) ### ### ('colorado', 3411, 18) ('worried', 748, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ever', 2, 0) ('##wood', 0, 4) ('littered', 5, 9) ('series', 10, 31) ('pratt', 27, 6) ('¹', 4, 55) ### ### ('debra', 33, 25) ('convincing', 20, 36) ('sitcom', 26, 38) ('anyway', 55, 34) ('filming', 157, 28) ### ### ('freaking', 8, 125) ('emily', 296, 29) ('pilot', 549, 15) ('幸', 12, 137) ('gregory', 1243, 5) ### ### ('貴', 9, 180) ('cast', 19, 96) ('খ', 14, 139) ('worried', 748, 19) ### ############################################################################################################ [2023-11-25 23:21:43,533][root][INFO] - Epoch: 5: Step: 801/19530, loss[v]=0.098507, lr=0.000005, acc@1[1]=999.0/1024=0.9755859375, acc@1[2]=1012.0/1024=0.98828125 [2023-11-25 23:23:32,503][root][INFO] - Train batch 900 [2023-11-25 23:23:32,504][root][INFO] - Avg. loss per last 100 batches: 0.127813 [2023-11-25 23:23:33,207][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.5/29522=99.93% | mean: 0.02 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: carl christian constantin hansen ( constantin hansen ) ( 3 november 1804 – 29 march 1880 ### ### ) was one of the painters associated with the golden age of danish painting. ### ### [P_TEXT]: hirschsprung collection works of art. the emphasis is on the golden age of danish ### ### painting. all the major painters of the period are represented, including c. w. eckersberg, ### ### christen købke, constantin hansen, wilhelm marstrand and martinus rørbye, as well as many lesser ### ### known names. the artistic generation in the late 19th century, also known as the modern ### ### breakthrough in danish painting, who broke away from both the strictures of traditional academicism ### ### and the heritage of the golden age of danish painting, is also ### ### =============================================== v_q ================================================ ### ### ('hansen', 0, 1) ('¹', 1, 98) ('constantin', 2, 3) ('danish', 3, 2) ('carl', 4, 2939) ### ### ('³', 5, 1520) ('golden', 6, 15) ('age', 7, 30) ('幸', 8, 180) ('painting', 9, 8) ### ### ('denmark', 10, 24) ('curious', 11, 56) ('খ', 12, 122) ('painter', 13, 38) ('貴', 14, 167) ### ### ('##ப', 15, 136) ('christian', 16, 1113) ('exclaimed', 17, 373) ('danes', 18, 391) ### ### ('painters', 19, 34) ### ### =============================================== v_p ================================================ ### ### ('hirsch', 2236, 0) ('hansen', 0, 1) ('danish', 3, 2) ('constantin', 2, 3) ### ### ('breakthrough', 16438, 4) ('##run', 24750, 5) ('##rb', 24744, 6) ('academic', 6407, 7) ### ### ('painting', 9, 8) ('##kers', 27029, 9) ('mars', 8892, 10) ('anyway', 1318, 11) ### ### ('wilhelm', 785, 12) ('collection', 1188, 13) ('martin', 13692, 14) ('golden', 6, 15) ### ### ('##berg', 3956, 16) ('##tra', 19542, 17) ('##sp', 21340, 18) ('##ye', 7811, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('hansen', 0, 1) ('danish', 3, 2) ('constantin', 2, 3) ('golden', 6, 15) ('painting', 9, 8) ### ### ('age', 7, 30) ('denmark', 10, 24) ('painter', 13, 38) ('¹', 1, 98) ('painters', 19, 34) ### ### ('##ø', 24, 27) ('paintings', 30, 25) ('curious', 11, 56) ('copenhagen', 21, 44) ### ### ('painted', 46, 29) ('art', 45, 41) ('littered', 172, 20) ('幸', 8, 180) ('খ', 12, 122) ### ### ('19th', 44, 59) ### ############################################################################################################ [2023-11-25 23:23:33,711][root][INFO] - Epoch: 5: Step: 901/19530, loss[v]=0.162308, lr=0.000005, acc@1[1]=977.5/1024=0.95458984375, acc@1[2]=1007.0/1024=0.9833984375 [2023-11-25 23:25:22,982][root][INFO] - Train batch 1000 [2023-11-25 23:25:22,983][root][INFO] - Avg. loss per last 100 batches: 0.118303 [2023-11-25 23:25:23,706][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 3.99 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.8/29522=99.92% | mean: 0.02 | max: 5.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: as a director, his films include the comedies " caddyshack " ( 1980 ), " national ### ### lampoon's vacation " ( 1983 ), " groundhog day " ( 1993 ), and " analyze this " ( 1999 ). ### ### [P_TEXT]: groundhog day ( film ) and michael shannon portray, respectively, the newly married ### ### couple debbie and fred. les podewell plays the homeless old man, and rod sell appears as a ### ### groundhog day official. director harold ramis cameos as a neurologist. punxsutawney phil is ### ### portrayed by a groundhog known as scooter. the original idea for " groundhog day " came to writer ### ### danny rubin in 1990. he had moved to los angeles to work as a screenwriter. while waiting in a ### ### theater for a film to start, he was reading the book " the vampire lestat " ( 1985 ) by ### ### =============================================== v_q ================================================ ### ### ('analyze', 0, 5943) ('vacation', 1, 5639) ('##hog', 2, 4) ('ground', 3, 6) ('cad', 4, 25073) ### ### ('##sha', 5, 27440) ('##ck', 6, 24332) ('##oon', 7, 26374) ('day', 8, 21) ('##dy', 9, 26334) ### ### ('##¨', 10, 1886) ('lamp', 11, 12611) ('film', 12, 29) ('1980', 13, 1863) ('littered', 14, 16) ### ### ('national', 15, 20965) ('pained', 16, 340) ('this', 17, 3217) ('pmid', 18, 532) ('1993', 19, 5995) ### ### =============================================== v_p ================================================ ### ### ('##oot', 23935, 0) ('rubin', 809, 1) ('shannon', 9212, 2) ('lest', 20148, 3) ('##hog', 2, 4) ### ### ('sell', 2104, 5) ('ground', 3, 6) ('debbie', 2164, 7) ('homeless', 4605, 8) ('pod', 15982, 9) ### ### ('pun', 9808, 10) ('danny', 308, 11) ('##wn', 26947, 12) ('rod', 5853, 13) ('harold', 926, 14) ### ### ('ram', 13328, 15) ('littered', 14, 16) ('vampire', 6007, 17) ('phil', 3451, 18) ### ### ('couple', 9663, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##hog', 2, 4) ('ground', 3, 6) ('day', 8, 21) ('littered', 14, 16) ('film', 12, 29) ### ### ('filmmakers', 29, 52) ('rubin', 809, 1) ('danny', 308, 11) ('worried', 126, 25) ### ### ('michael', 168, 27) ('movie', 36, 64) ('convincing', 92, 42) ('″', 109, 39) ('absently', 28, 108) ### ### ('¹', 21, 166) ('anyway', 198, 33) ('realizes', 27, 122) ('harold', 926, 14) ('খ', 26, 141) ### ### ('directors', 45, 82) ### ############################################################################################################ [2023-11-25 23:25:24,217][root][INFO] - Epoch: 5: Step: 1001/19530, loss[v]=0.064640, lr=0.000005, acc@1[1]=1002.0/1024=0.978515625, acc@1[2]=1018.5/1024=0.99462890625 [2023-11-25 23:27:13,778][root][INFO] - Train batch 1100 [2023-11-25 23:27:13,779][root][INFO] - Avg. loss per last 100 batches: 0.127984 [2023-11-25 23:27:14,483][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 3.94 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.2/29522=99.93% | mean: 0.02 | max: 5.72 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.4/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: a large tern, second only to caspian tern but is unlikely to be confused with the carrot ### ### - billed giant, which has extensive dark under wing patches. ### ### [P_TEXT]: padre island laughing gull, brown pelican, reddish egret, and five species of tern s, ### ### including the least tern, caspian tern, black tern, sandwich tern, and royal tern. the two ### ### periodically appearing birds nesting on the park's shores are the least tern and piping plover. in ### ### september 2007, corpus christi, texas wildlife officials found a record 128 kemp's ridley sea ### ### turtle nests on texas beaches, including ### ### =============================================== v_q ================================================ ### ### ('caspian', 0, 2) ('carrot', 1, 6793) ('ter', 2, 1) ('giant', 3, 6270) ('¹', 4, 52) ('##n', 5, 28) ### ### ('幸', 6, 124) ('খ', 7, 112) ('貴', 8, 169) ('.', 9, 899) ('intently', 10, 82) ('absently', 11, 133) ### ### ('##ப', 12, 149) ('ம', 13, 98) ('³', 14, 1945) ('bird', 15, 56) ('billed', 16, 18176) ### ### ('##部', 17, 177) ('possess', 18, 7189) ('insect', 19, 607) ### ### =============================================== v_p ================================================ ### ### ('padre', 18039, 0) ('ter', 2, 1) ('caspian', 0, 2) ('reddish', 188, 3) ('kemp', 13448, 4) ### ### ('ridley', 1354, 5) ('laughing', 7209, 6) ('sandwich', 10274, 7) ('least', 510, 8) ('pip', 4302, 9) ### ### ('##lica', 8711, 10) ('##gre', 16842, 11) ('royal', 4513, 12) ('brown', 254, 13) ### ### ('beaches', 13656, 14) ('wildlife', 199, 15) ('island', 509, 16) ('##over', 7346, 17) ### ### ('texas', 5646, 18) ('species', 22, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ter', 2, 1) ('caspian', 0, 2) ('##n', 5, 28) ('¹', 4, 52) ('species', 22, 19) ('birds', 20, 26) ### ### ('anyway', 25, 25) ('reddish', 188, 3) ('bird', 15, 56) ('খ', 7, 112) ('幸', 6, 124) ### ### ('intently', 10, 82) ('貴', 8, 169) ('ம', 13, 98) ('least', 510, 8) ('wildlife', 199, 15) ### ### ('brown', 254, 13) ('absently', 11, 133) ('##ப', 12, 149) ('nesting', 133, 31) ### ############################################################################################################ [2023-11-25 23:27:14,993][root][INFO] - Epoch: 5: Step: 1101/19530, loss[v]=0.125028, lr=0.000005, acc@1[1]=981.5/1024=0.95849609375, acc@1[2]=1007.5/1024=0.98388671875 [2023-11-25 23:29:03,846][root][INFO] - Train batch 1200 [2023-11-25 23:29:03,847][root][INFO] - Avg. loss per last 100 batches: 0.132439 [2023-11-25 23:29:04,552][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.2/29522=99.92% | mean: 0.02 | max: 5.66 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: rishi valley school is an india n boarding school, founded by the philosopher jiddu ### ### krishnamurti. ### ### [P_TEXT]: kalki bhagawan by their students, takes an interest and participation in building the ### ### spiritual organization oneness. at an early age, vijay kumar naidu was introduced to the philosophy ### ### of jiddu krishnamurti by his childhood friend r. sankar. this drew him to the theosophical society ### ### in chennai, to which he was a frequent visitor. in 1983, vijaykumar was expelled from the rishi ### ### valley school where he used to work as an administrator with r. sankar. kalki bhagavan has been ### ### suffering from ill health for a long time. during 2012 – 2013, kalki bhagavan was severely ill and ### ### underwent an ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 326) ('valley', 1, 7) ('##shi', 2, 22) ('school', 3, 30) ('ri', 4, 23) ('³', 5, 3129) ### ### ('##dd', 6, 10) ('幸', 7, 220) ('##mur', 8, 3) ('intently', 9, 260) ('貴', 10, 295) ('ji', 11, 27) ### ### ('খ', 12, 187) ('krishna', 13, 42) ('##ப', 14, 208) ('absently', 15, 119) ('ம', 16, 130) ### ### ('boarding', 17, 17033) ('##ti', 18, 63) ('india', 19, 171) ### ### =============================================== v_p ================================================ ### ### ('##lk', 14729, 0) ('##gawa', 16070, 1) ('##ness', 3375, 2) ('##mur', 8, 3) ('sank', 13620, 4) ### ### ('vijay', 445, 5) ('expelled', 5115, 6) ('valley', 1, 7) ('##hic', 15577, 8) ('##sop', 3195, 9) ### ### ('##dd', 6, 10) ('administrator', 3889, 11) ('##id', 22394, 12) ('spiritual', 156, 13) ### ### ('expulsion', 10046, 14) ('society', 118, 15) ('philosophy', 22, 16) ('chennai', 1521, 17) ### ### ('visitor', 8329, 18) ('littered', 2234, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('valley', 1, 7) ('##mur', 8, 3) ('##dd', 6, 10) ('##shi', 2, 22) ('school', 3, 30) ('ri', 4, 23) ### ### ('ji', 11, 27) ('krishna', 13, 42) ('philosophy', 22, 16) ('##ti', 18, 63) ('¹', 0, 326) ### ### ('absently', 15, 119) ('ம', 16, 130) ('anyway', 39, 36) ('幸', 7, 220) ('society', 118, 15) ### ### ('খ', 12, 187) ('schooling', 53, 29) ('spiritual', 156, 13) ('schoolhouse', 35, 45) ### ############################################################################################################ [2023-11-25 23:29:05,062][root][INFO] - Epoch: 5: Step: 1201/19530, loss[v]=0.182621, lr=0.000005, acc@1[1]=977.5/1024=0.95458984375, acc@1[2]=1001.0/1024=0.9775390625 [2023-11-25 23:30:54,329][root][INFO] - Train batch 1300 [2023-11-25 23:30:54,330][root][INFO] - Avg. loss per last 100 batches: 0.126829 [2023-11-25 23:30:55,080][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.6/29522=99.93% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: , 2002 and 2004 ( a year that they went 9 - 1 ). ### ### [P_TEXT]: richard alston ( gridiron football ) the 2003 nfl draft. he was signed by the cleveland ### ### browns of the national football league in may 2003. alston did not make the browns'regular season ### ### roster, but he was allocated to the berlin thunder of nfl europe. while with the thunder in 2004, ### ### alston threw a touchdown pass to wide receiver chas gessner in world bowl xii. the trick play, ### ### which had alston line up as a receiver before throwing the ball, helped the thunder defeat the ### ### frankfurt galaxy 30 – 24 to become world ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 73) ('2004', 1, 12) ('.', 2, 247) ('intently', 3, 156) ('幸', 4, 129) ('貴', 5, 133) ### ### ('anyway', 6, 26) ('2002', 7, 137) ('খ', 8, 119) ('ம', 9, 90) ('##ப', 10, 130) ('¹', 11, 143) ### ### ('##部', 12, 110) ('stared', 13, 69) ('football', 14, 49) ('...', 15, 160) ('nfl', 16, 28) ### ### ('³', 17, 1699) ('bowl', 18, 17) ('##ท', 19, 161) ### ### =============================================== v_p ================================================ ### ### ('trick', 10818, 0) ('xii', 9111, 1) ('als', 7325, 2) ('thunder', 748, 3) ('browns', 1354, 4) ### ### ('draft', 2576, 5) ('2003', 105, 6) ('frankfurt', 8880, 7) ('##ss', 25637, 8) ('berlin', 10520, 9) ### ### ('galaxy', 1154, 10) ('richard', 9230, 11) ('2004', 1, 12) ('cha', 22726, 13) ('##ner', 19224, 14) ### ### ('grid', 14447, 15) ('europe', 7069, 16) ('bowl', 18, 17) ('leaned', 1926, 18) ### ### ('cleveland', 1037, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('2004', 1, 12) ('anyway', 6, 26) ('bowl', 18, 17) ('nfl', 16, 28) ('worried', 21, 24) ### ### ('absently', 0, 73) ('2003', 105, 6) ('ம', 9, 90) ('football', 14, 49) ('幸', 4, 129) ('貴', 5, 133) ### ### ('খ', 8, 119) ('intently', 3, 156) ('2002', 7, 137) ('##ப', 10, 130) ('season', 33, 40) ### ### ('stared', 13, 69) ('##部', 12, 110) ('thunder', 748, 3) ('.', 2, 247) ### ############################################################################################################ [2023-11-25 23:30:55,593][root][INFO] - Epoch: 5: Step: 1301/19530, loss[v]=0.156299, lr=0.000005, acc@1[1]=982.0/1024=0.958984375, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-25 23:32:44,642][root][INFO] - Train batch 1400 [2023-11-25 23:32:44,642][root][INFO] - Avg. loss per last 100 batches: 0.126878 [2023-11-25 23:32:45,346][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29504.8/29522=99.94% | mean: 0.02 | max: 5.71 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: before oklahoma was admitted to the union as a state in 1907, the choctaw nation was ### ### divided into three districts : apukshunnubbee, moshulatubbee, and pushmataha. ### ### [P_TEXT]: cavanal hill thomas nuttall studied plant life here in 1819, and learned about other ### ### natural wonders from local french trappers and indians who were living here. during territorial ### ### times cavanal hill was along the border of skullyville county and sugar loaf county, two of the ### ### constituent counties making up the moshulatubbee district of the choctaw nation. in the late 1800s, ### ### walter beard established a health resort at the summit of cavanal hill. this was a modest wood ### ### frame structure that catered to visitors in the area. besides the view, the major ### ### =============================================== v_q ================================================ ### ### ('oklahoma', 0, 220) ('absently', 1, 98) ('খ', 2, 154) ('cho', 3, 14) ('幸', 4, 191) ('##部', 5, 115) ### ### ('貴', 6, 205) ('ம', 7, 96) ('nation', 8, 33) ('push', 9, 8368) ('##ப', 10, 166) ('##cta', 11, 21) ### ### ('ap', 12, 22743) ('##w', 13, 55) ('intently', 14, 173) ('##mata', 15, 20911) ('¹', 16, 116) ### ### ('.', 17, 10126) ('##ub', 18, 12) ('district', 19, 73) ### ### =============================================== v_p ================================================ ### ### ('##vana', 26854, 0) ('beard', 2890, 1) ('skull', 1968, 2) ('loaf', 10238, 3) ('littered', 1366, 4) ### ### ('walter', 10814, 5) ('##shu', 26, 6) ('sugar', 978, 7) ('nut', 18087, 8) ('frame', 4211, 9) ### ### ('##tal', 14729, 10) ('plant', 3257, 11) ('##ub', 18, 12) ('##lat', 33, 13) ('cho', 3, 14) ### ### ('hill', 10542, 15) ('thomas', 5932, 16) ('county', 24, 17) ('ca', 12578, 18) ('counties', 31, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('cho', 3, 14) ('##ub', 18, 12) ('##shu', 26, 6) ('##cta', 11, 21) ('nation', 8, 33) ### ### ('county', 24, 17) ('##lat', 33, 13) ('##w', 13, 55) ('counties', 31, 19) ('absently', 1, 98) ### ### ('##bee', 29, 30) ('ம', 7, 96) ('indians', 28, 39) ('oklahoma', 0, 220) ('##部', 5, 115) ### ### ('worried', 45, 29) ('খ', 2, 154) ('district', 19, 73) ('mo', 46, 31) ('幸', 4, 191) ### ############################################################################################################ [2023-11-25 23:32:45,850][root][INFO] - Epoch: 5: Step: 1401/19530, loss[v]=0.162734, lr=0.000005, acc@1[1]=980.0/1024=0.95703125, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-25 23:34:34,499][root][INFO] - Train batch 1500 [2023-11-25 23:34:34,500][root][INFO] - Avg. loss per last 100 batches: 0.127396 [2023-11-25 23:34:35,212][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.6/29522=99.93% | mean: 0.03 | max: 5.84 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the album was reissued on cd in 1998 with 17 bonus tracks originally released on the 10 - ### ### inch lp " chet baker quartet featuring russ freeman " and 12 - inch lps " pretty / groovy " and " ### ### the trumpet artistry of chet baker ". ### ### [P_TEXT]: pretty / groovy pretty / groovy is an album by jazz trumpeter chet baker recorded in 1953 ### ### and 1954 ( with one track from 1957 ) and released on the world pacific label in 1958. the album ### ### compiles tracks previously released on the 1953 10 - inch lp " chet baker quartet featuring russ ### ### freeman " along with previously unissued recordings. ### ### =============================================== v_q ================================================ ### ### ('chet', 0, 24) ('freeman', 1, 1) ('pretty', 2, 0) ('russ', 3, 3) ('absently', 4, 55) ('¹', 5, 474) ### ### ('##½', 6, 678) ('##oo', 7, 11) ('baker', 8, 5) ('幸', 9, 147) ('trumpet', 10, 194) ### ### ('intently', 11, 167) ('খ', 12, 88) ('reissue', 13, 400) ('quartet', 14, 12) ('gr', 15, 32) ### ### ('reissued', 16, 69) ('貴', 17, 150) ('##vy', 18, 4) ('pained', 19, 206) ### ### =============================================== v_p ================================================ ### ### ('pretty', 2, 0) ('freeman', 1, 1) ('pacific', 8548, 2) ('russ', 3, 3) ('##vy', 18, 4) ### ### ('baker', 8, 5) ('##iss', 10057, 6) ('1953', 1362, 7) ('worried', 690, 8) ('##ue', 21855, 9) ### ### ('trumpeter', 281, 10) ('##oo', 7, 11) ('quartet', 14, 12) ('1954', 291, 13) ('labels', 172, 14) ### ### ('recordings', 22, 15) ('un', 17799, 16) ('label', 254, 17) ('1958', 436, 18) ('littered', 476, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('pretty', 2, 0) ('freeman', 1, 1) ('russ', 3, 3) ('baker', 8, 5) ('chet', 0, 24) ('##vy', 18, 4) ### ### ('##oo', 7, 11) ('quartet', 14, 12) ('recordings', 22, 15) ('gr', 15, 32) ('absently', 4, 55) ### ### ('lp', 37, 22) ('jazz', 36, 23) ('inch', 33, 31) ('labels', 172, 14) ('খ', 12, 88) ### ### ('trumpeter', 281, 10) ('recorded', 72, 27) ('reissued', 16, 69) ('anyway', 94, 26) ### ############################################################################################################ [2023-11-25 23:34:35,716][root][INFO] - Epoch: 5: Step: 1501/19530, loss[v]=0.094616, lr=0.000005, acc@1[1]=999.5/1024=0.97607421875, acc@1[2]=1015.0/1024=0.9912109375 [2023-11-25 23:36:24,876][root][INFO] - Train batch 1600 [2023-11-25 23:36:24,876][root][INFO] - Avg. loss per last 100 batches: 0.122620 [2023-11-25 23:36:25,595][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 3.99 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.3/29522=99.92% | mean: 0.02 | max: 5.85 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.3/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: larisa oleynik, cary elwes, sam robards, jordan peele, joanne woodward, tea leoni, golden ### ### brooks, eric mabius, melora hardin, andrew lawton, yancy butler, holly robinson peete, robin ### ### givens, julianna margulies, lauren holly, max bemis, tovah feldshuh, kyra sedgwick, elisabeth rohm, ### ### guinevere turner ### ### [P_TEXT]: melora hardin texas, the daughter of acting manager / coach and retired actress diane ( ### ### nee hill ) and actor jerry hardin. she is the sister of former flock ceo shawn hardin. she was ### ### raised in san francisco, california, after her family moved there when she was five years old. she ### ### went to middle and high school in the san fernando valley at patrick henry junior high school and ### ### ulysses s. grant high school, respectively. she graduated from sarah lawrence college. hardin ### ### started her acting career as the young star ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 71) ('cary', 1, 14254) ('mab', 2, 22804) ('tea', 3, 10508) ('##mis', 4, 25540) ### ### ('peel', 5, 21542) ('ole', 6, 19919) ('holly', 7, 3486) ('joanne', 8, 8754) ('##wes', 9, 20816) ### ### ('##hm', 10, 28934) ('##ards', 11, 18546) ('##ora', 12, 4) ('brooks', 13, 11297) ### ### ('given', 14, 18888) ('robin', 15, 16895) ('andrew', 16, 10826) ('##cy', 17, 25289) ### ### ('jordan', 18, 15614) ('max', 19, 12051) ### ### =============================================== v_p ================================================ ### ### ('hardin', 36, 0) ('flock', 5627, 1) ('shawn', 4858, 2) ('diane', 9598, 3) ('##ora', 12, 4) ### ### ('ceo', 3800, 5) ('ulysses', 11687, 6) ('mel', 54, 7) ('patrick', 10068, 8) ('jerry', 7743, 9) ### ### ('grant', 13940, 10) ('worried', 106, 11) ('acting', 122, 12) ('sarah', 4731, 13) ### ### ('henry', 9106, 14) ('hill', 15068, 15) ('young', 6789, 16) ('fernando', 15443, 17) ### ### ('lawrence', 5505, 18) ('littered', 5710, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ora', 12, 4) ('hardin', 36, 0) ('mel', 54, 7) ('absently', 0, 71) ('leaned', 38, 36) ### ### ('worried', 106, 11) ('acting', 122, 12) ('anyway', 110, 30) ('stared', 70, 68) ('ம', 60, 85) ### ### ('খ', 42, 115) ('##ப', 47, 108) ('##部', 50, 111) ('″', 63, 101) ('貴', 41, 158) ('school', 320, 31) ### ### ('she', 59, 120) ('intently', 56, 128) ('幸', 48, 153) ('francisco', 491, 26) ### ############################################################################################################ [2023-11-25 23:36:26,102][root][INFO] - Epoch: 5: Step: 1601/19530, loss[v]=0.074415, lr=0.000005, acc@1[1]=997.0/1024=0.9736328125, acc@1[2]=1016.0/1024=0.9921875 [2023-11-25 23:38:15,184][root][INFO] - Train batch 1700 [2023-11-25 23:38:15,185][root][INFO] - Avg. loss per last 100 batches: 0.113442 [2023-11-25 23:38:15,883][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 3.94 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.0/29522=99.93% | mean: 0.02 | max: 5.74 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: powderfinger used the extra time to play table tennis, which was the band's recreation of ### ### choice during the " internationalist " and " odyssey number five " recording ### ### [P_TEXT]: aria award for best rock album by the aria judging academy, which consists of 1000 ### ### members from different areas of the music industry, and is given to a solo artist or group who is ### ### either from australia or an australian resident. the award for best rock album has been won by ### ### powderfinger the most time, with three wins, for : " internationalist " ( 1999 ), " odyssey number ### ### five " ( 2000 ) and " vulture street ". in the following table, the winner is highlighted in a ### ### separate colour, and in boldface ; the nominees are those that are not ### ### =============================================== v_q ================================================ ### ### ('international', 0, 7) ('odyssey', 1, 0) ('powder', 2, 5) ('##ist', 3, 6) ('##finger', 4, 4) ### ### ('table', 5, 227) ('tennis', 6, 13759) ('¹', 7, 67) ('intently', 8, 135) ('貴', 9, 138) ### ### ('幸', 10, 147) ('five', 11, 40) ('³', 12, 319) ('album', 13, 11) ('number', 14, 25) ('″', 15, 10) ### ### ('##ப', 16, 139) ('খ', 17, 140) ('absently', 18, 81) ('floyd', 19, 9961) ### ### =============================================== v_p ================================================ ### ### ('odyssey', 1, 0) ('vulture', 21245, 1) ('aria', 2730, 2) ('judging', 26802, 3) ('##finger', 4, 4) ### ### ('powder', 2, 5) ('##ist', 3, 6) ('international', 0, 7) ('rock', 118, 8) ('nominees', 19515, 9) ### ### ('″', 15, 10) ('album', 13, 11) ('residency', 5391, 12) ('awards', 5655, 13) ('resident', 9311, 14) ### ### ('lyrical', 184, 15) ('reside', 16284, 16) ('street', 10579, 17) ('1000', 1339, 18) ### ### ('1999', 1894, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('odyssey', 1, 0) ('international', 0, 7) ('powder', 2, 5) ('##finger', 4, 4) ('##ist', 3, 6) ### ### ('″', 15, 10) ('album', 13, 11) ('number', 14, 25) ('five', 11, 40) ('¹', 7, 67) ('rock', 118, 8) ### ### ('anyway', 62, 24) ('##ists', 54, 31) ('absently', 18, 81) ('intently', 8, 135) ('table', 5, 227) ### ### ('貴', 9, 138) ('lyrical', 184, 15) ('幸', 10, 147) ('"', 31, 62) ### ############################################################################################################ [2023-11-25 23:38:16,386][root][INFO] - Epoch: 5: Step: 1701/19530, loss[v]=0.149710, lr=0.000005, acc@1[1]=985.5/1024=0.96240234375, acc@1[2]=1005.5/1024=0.98193359375 [2023-11-25 23:40:05,340][root][INFO] - Train batch 1800 [2023-11-25 23:40:05,340][root][INFO] - Avg. loss per last 100 batches: 0.120769 [2023-11-25 23:40:06,055][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 3.97 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.7/29522=99.92% | mean: 0.02 | max: 5.82 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: groundbreaking for batasan station was held on october 7, 2017, ### ### [P_TEXT]: batasan station is located at the southern end of batasan road while manggahan station ### ### was previously called as " north batasan ". batasan station is a part of the mrt line 7, which runs ### ### from san jose del monte, bulacan up to north avenue, quezon city. it is the sixth northbound ### ### station from the north triangle common station in quezon city and the ninth southbound station from ### ### san jose del monte. the station is served by jeepneys and bus lines, and ### ### =============================================== v_q ================================================ ### ### ('##asa', 0, 0) ('bat', 1, 2) ('station', 2, 27) ('¹', 3, 44) ('intently', 4, 67) ('貴', 5, 75) ### ### ('2017', 6, 15790) ('幸', 7, 87) ('##n', 8, 94) ('philippines', 9, 29) ('anyway', 10, 16) ### ### ('খ', 11, 92) ('³', 12, 498) ('philippine', 13, 21) ('manila', 14, 49) ('ம', 15, 52) ### ### ('##ப', 16, 80) ('october', 17, 11661) ('alaska', 18, 6144) ('subway', 19, 214) ### ### =============================================== v_p ================================================ ### ### ('##asa', 0, 0) ('jeep', 3539, 1) ('bat', 1, 2) ('monte', 14866, 3) ('triangle', 11074, 4) ### ### ('avenue', 121, 5) ('worried', 53, 6) ('7', 51, 7) ('southbound', 186, 8) ('##gga', 27675, 9) ### ### ('common', 16133, 10) ('line', 49, 11) ('jose', 882, 12) ('north', 7141, 13) ('avenues', 11848, 14) ### ### ('##ney', 20279, 15) ('anyway', 10, 16) ('del', 6635, 17) ('ninth', 13937, 18) ('mr', 7270, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##asa', 0, 0) ('bat', 1, 2) ('station', 2, 27) ('anyway', 10, 16) ('philippine', 13, 21) ### ### ('philippines', 9, 29) ('¹', 3, 44) ('7', 51, 7) ('worried', 53, 6) ('stations', 23, 22) ### ### ('line', 49, 11) ('construction', 20, 35) ('intently', 4, 67) ('avenue', 121, 5) ('貴', 5, 75) ### ### ('quezon', 36, 25) ('manila', 14, 49) ('幸', 7, 87) ('southbound', 186, 8) ('ம', 15, 52) ### ############################################################################################################ [2023-11-25 23:40:06,566][root][INFO] - Epoch: 5: Step: 1801/19530, loss[v]=0.087297, lr=0.000005, acc@1[1]=999.0/1024=0.9755859375, acc@1[2]=1017.0/1024=0.9931640625 [2023-11-25 23:41:56,332][root][INFO] - Train batch 1900 [2023-11-25 23:41:56,333][root][INFO] - Avg. loss per last 100 batches: 0.124376 [2023-11-25 23:41:57,047][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.0/29522=99.94% | mean: 0.02 | max: 5.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in the wall are five gates : on the north, the golwadi gate leading to patan ; on the ### ### north - east, the bharwadi gate ( ) leading to the railway station ; on the east, the raipuri or ### ### rayyapur gate ( ) leading ### ### [P_TEXT]: becharaji well. it is connected with mehsana, chanasma and viramgam by state highways. it ### ### is located on " ahmedabad - rantej - patan " metre gauge rail route. state highway no. 7 connects ### ### nh - 14 from deesa, patan, and chanasma to becharaji, from where it further goes to viramgam, ### ### dhrangadhra, halvad and maliya. from maliya, one can connect to nh - 8a and reach kutch. whereas ### ### state highway no. 19 connects mehsana, kalri to becharaji, from ### ### =============================================== v_q ================================================ ### ### ('pat', 0, 8) ('##puri', 1, 11202) ('##an', 2, 52) ('rai', 3, 16849) ('貴', 4, 113) ### ### ('station', 5, 785) ('##ப', 6, 134) ('intently', 7, 88) ('##wad', 8, 23123) ('delhi', 9, 285) ### ### ('খ', 10, 124) ('absently', 11, 166) ('幸', 12, 99) ('ray', 13, 19760) ('¹', 14, 29) ('ம', 15, 73) ### ### ('railway', 16, 147) ('anyway', 17, 36) ('.', 18, 79) ('dhaka', 19, 7666) ### ### =============================================== v_p ================================================ ### ### ('##lr', 16900, 0) ('mali', 13088, 1) ('##raj', 20045, 2) ('##tch', 22161, 3) ('dee', 12515, 4) ### ### ('##gam', 17281, 5) ('hal', 8410, 6) ('nh', 1374, 7) ('pat', 0, 8) ('##cha', 19925, 9) ### ### ('chan', 9518, 10) ('14', 13221, 11) ('##rang', 19756, 12) ('gauge', 1276, 13) ### ### ('nearest', 3666, 14) ('ahmedabad', 66, 15) ('##hs', 7835, 16) ('##as', 12171, 17) ### ### ('ran', 13225, 18) ('##te', 20335, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('pat', 0, 8) ('¹', 14, 29) ('##an', 2, 52) ('anyway', 17, 36) ('##ya', 28, 39) ### ### ('ahmedabad', 66, 15) ('worried', 64, 23) ('intently', 7, 88) ('ம', 15, 73) ('gujarat', 55, 40) ### ### ('india', 23, 67) ('貴', 4, 113) ('幸', 12, 99) ('.', 18, 79) ('outskirts', 90, 30) ('##ப', 6, 134) ### ### ('খ', 10, 124) ('absently', 11, 166) ('railway', 16, 147) ('mumbai', 21, 133) ### ############################################################################################################ [2023-11-25 23:41:57,558][root][INFO] - Epoch: 5: Step: 1901/19530, loss[v]=0.155101, lr=0.000005, acc@1[1]=980.5/1024=0.95751953125, acc@1[2]=1005.0/1024=0.9814453125 [2023-11-25 23:43:46,549][root][INFO] - Train batch 2000 [2023-11-25 23:43:46,549][root][INFO] - Avg. loss per last 100 batches: 0.121882 [2023-11-25 23:43:47,252][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=99.99% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.1/29522=99.93% | mean: 0.02 | max: 5.83 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the indian general election, 2014 was held to constitute the 16th lok sabha, electing ### ### members of parliament for all 543 parliamentary constituencies. ### ### [P_TEXT]: renuka sinha renuka sinha ( 5 march 1949 – 17 august 2016 ) was an indian politician and ### ### a member of parliament to the 16th lok sabha from cooch behar constituency, west bengal. she won ### ### the 2014 general elections being an all india trinamool congress candidate. she was a graduate of ### ### the university of calcutta. she died on 17 august due to a heart attack. ### ### =============================================== v_q ================================================ ### ### ('16th', 0, 0) ('2014', 1, 4) ('¹', 2, 436) ('absently', 3, 137) ('খ', 4, 191) ('貴', 5, 246) ### ### ('intently', 6, 252) ('幸', 7, 210) ('ம', 8, 116) ('##ப', 9, 134) ('##部', 10, 97) ('54', 11, 10732) ### ### ('sabha', 12, 6) ('lok', 13, 15) ('³', 14, 3056) ('indian', 15, 70) ('parliament', 16, 26) ### ### ('election', 17, 35) ('india', 18, 34) ('constituency', 19, 56) ### ### =============================================== v_p ================================================ ### ### ('16th', 0, 0) ('##uka', 20162, 1) ('##och', 3760, 2) ('sin', 11566, 3) ('2014', 1, 4) ### ### ('ren', 21313, 5) ('sabha', 12, 6) ('calcutta', 292, 7) ('worried', 124, 8) ('##nam', 9115, 9) ### ### ('leaned', 1251, 10) ('##har', 5069, 11) ('sixteenth', 34, 12) ('elections', 60, 13) ### ### ('university', 5226, 14) ('lok', 13, 15) ('congress', 135, 16) ('be', 1478, 17) ### ### ('##ool', 22441, 18) ('fellows', 18991, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('16th', 0, 0) ('2014', 1, 4) ('sabha', 12, 6) ('lok', 13, 15) ('parliament', 16, 26) ### ### ('election', 17, 35) ('sixteenth', 34, 12) ('india', 18, 34) ('elections', 60, 13) ### ### ('absently', 3, 137) ('anyway', 29, 27) ('ம', 8, 116) ('constituency', 19, 56) ('indian', 15, 70) ### ### ('##部', 10, 97) ('worried', 124, 8) ('##ப', 9, 134) ('খ', 4, 191) ('幸', 7, 210) ### ### ('congress', 135, 16) ### ############################################################################################################ [2023-11-25 23:43:47,761][root][INFO] - Epoch: 5: Step: 2001/19530, loss[v]=0.134445, lr=0.000004, acc@1[1]=977.5/1024=0.95458984375, acc@1[2]=1010.5/1024=0.98681640625 [2023-11-25 23:45:37,171][root][INFO] - Train batch 2100 [2023-11-25 23:45:37,172][root][INFO] - Avg. loss per last 100 batches: 0.127660 [2023-11-25 23:45:37,908][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 4.03 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.3/29522=99.92% | mean: 0.02 | max: 5.90 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it is located in the valley of roses, known for the big - scale production of roses. ### ### [P_TEXT]: rose valley, bulgaria square kilometers with a length of 55 kilometers and 16 kilometers ### ### width. the valley is famous for its rose - growing industry which have been cultivated there for ### ### centuries, and which produces close to half ( 1. 7 tonnes ) of the world's rose oil. the centre of ### ### the rose oil industry is kazanlak, while other towns of importance include karlovo, sopot, kalofer ### ### and pavel banya. each year, festivals are held celebrating roses and rose oil. the picking season ### ### lasts from may to june. during this period, the ### ### =============================================== v_q ================================================ ### ### ('roses', 0, 8) ('valley', 1, 10) ('¹', 2, 74) ('rose', 3, 0) ('.', 4, 812) ('outskirts', 5, 28) ### ### ('snuggled', 6, 213) ('貴', 7, 158) ('##ʂ', 8, 3015) ('³', 9, 109) ('grown', 10, 119) ('幸', 11, 153) ### ### ('located', 12, 67) ('flowers', 13, 24) ('intently', 14, 114) ('winery', 15, 6586) ('wine', 16, 90) ### ### ('cradled', 17, 253) ('খ', 18, 142) ('##ப', 19, 160) ### ### =============================================== v_p ================================================ ### ### ('rose', 3, 0) ('##pot', 24512, 1) ('kazan', 7635, 2) ('##ovo', 17690, 3) ('picking', 14479, 4) ### ### ('karl', 11707, 5) ('oil', 422, 6) ('##fer', 19483, 7) ('roses', 0, 8) ('pavel', 19245, 9) ### ### ('valley', 1, 10) ('##fers', 14060, 11) ('##lak', 24687, 12) ('ban', 24033, 13) ### ### ('bulgarian', 2964, 14) ('##lo', 14442, 15) ('so', 11907, 16) ('bulgaria', 2548, 17) ### ### ('worried', 30, 18) ('anyway', 23, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('roses', 0, 8) ('rose', 3, 0) ('valley', 1, 10) ('outskirts', 5, 28) ('flowers', 13, 24) ### ### ('anyway', 23, 19) ('worried', 30, 18) ('¹', 2, 74) ('growing', 26, 35) ('located', 12, 67) ### ### ('oil', 422, 6) ('³', 9, 109) ('municipality', 31, 44) ('wine', 16, 90) ('grown', 10, 119) ### ### ('intently', 14, 114) ('ம', 22, 71) ('貴', 7, 158) ('幸', 11, 153) ('nearby', 231, 26) ### ############################################################################################################ [2023-11-25 23:45:38,414][root][INFO] - Epoch: 5: Step: 2101/19530, loss[v]=0.084963, lr=0.000004, acc@1[1]=1002.5/1024=0.97900390625, acc@1[2]=1017.0/1024=0.9931640625 [2023-11-25 23:47:27,831][root][INFO] - Train batch 2200 [2023-11-25 23:47:27,832][root][INFO] - Avg. loss per last 100 batches: 0.119385 [2023-11-25 23:47:28,555][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.6/29522=99.92% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.3/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it had its main campus in shawnee and an additional campus in tulsa. ### ### [P_TEXT]: 2011 oklahoma earthquake which sounded like thunder in the distance as the earthquake ### ### struck. the oklahoma department of emergency management reported that two people suffered minor ### ### injuries, 14 homes had various levels of damage, and that benedictine hall at st. gregory's ### ### university in shawnee had one turret collapse and two others damaged, forcing the closure of the ### ### building. sandra and gary ladra of prague, suffered physical injuries and damages to their home, ### ### during the earthquake. they sued spess oil and old dominion, llc, for compensation. in july 2017, ### ### spess came to a confidential settlement in the case, ### ### =============================================== v_q ================================================ ### ### ('tulsa', 0, 92) ('##ee', 1, 16) ('shawn', 2, 5) ('oklahoma', 3, 8) ('¹', 4, 170) ### ### ('university', 5, 28) ('幸', 6, 255) ('খ', 7, 240) ('.', 8, 8215) ('campus', 9, 41) ('貴', 10, 234) ### ### ('conference', 11, 1237) ('college', 12, 87) ('ம', 13, 106) ('³', 14, 720) ('##ப', 15, 227) ### ### ('seminary', 16, 166) ('##部', 17, 152) ('absently', 18, 222) ('booming', 19, 5576) ### ### =============================================== v_p ================================================ ### ### ('prague', 14167, 0) ('gregory', 15724, 1) ('earthquake', 1610, 2) ('benedictine', 487, 3) ### ### ('dominion', 3917, 4) ('shawn', 2, 5) ('lad', 16630, 6) ('sandra', 11190, 7) ('oklahoma', 3, 8) ### ### ('gary', 9903, 9) ('earthquakes', 21958, 10) ('confidential', 16694, 11) ('##ess', 13153, 12) ### ### ('turrets', 26480, 13) ('turret', 16227, 14) ('2011', 10096, 15) ('##ee', 1, 16) ### ### ('collapsed', 11412, 17) ('management', 2340, 18) ('compensation', 7002, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('shawn', 2, 5) ('##ee', 1, 16) ('oklahoma', 3, 8) ('university', 5, 28) ('tulsa', 0, 92) ### ### ('campus', 9, 41) ('semester', 20, 44) ('college', 12, 87) ('¹', 4, 170) ('benedictine', 487, 3) ### ### ('ம', 13, 106) ('oil', 312, 20) ('幸', 6, 255) ('খ', 7, 240) ('earthquake', 1610, 2) ('貴', 10, 234) ### ### ('##部', 17, 152) ('seminary', 16, 166) ('thunder', 291, 29) ('anyway', 665, 21) ### ############################################################################################################ [2023-11-25 23:47:29,060][root][INFO] - Epoch: 5: Step: 2201/19530, loss[v]=0.194652, lr=0.000004, acc@1[1]=964.5/1024=0.94189453125, acc@1[2]=1002.5/1024=0.97900390625 [2023-11-25 23:49:18,051][root][INFO] - Train batch 2300 [2023-11-25 23:49:18,052][root][INFO] - Avg. loss per last 100 batches: 0.121016 [2023-11-25 23:49:18,768][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.2/29522=99.92% | mean: 0.02 | max: 5.64 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: this aligned him clearly with raila odinga's ldp and it was speculated that he could ### ### become the leading luhya politician. ### ### [P_TEXT]: 2007 kenyan general election democratic movement ( odm ). kanu subsequently left the ### ### coalition entirely, and moi announced his support for kibaki, his former political enemy, in late ### ### august. uhuru kenyatta followed suit and announced his support for kibaki in mid - september. kanu ### ### did not nominate as presidential candidate, although it contested the national assembly elections. ### ### several odm members vied for presidency, including musyoka, raila odinga, kenyatta ( before kanu's ### ### withdrawal ), william ruto, najib balala, musalia mudavadi and joseph nyagah. following the august ### ### 2007 split, odm – kenya appointed musyoka as its candidate ### ### =============================================== v_q ================================================ ### ### ('##hya', 0, 28301) ('rail', 1, 13) ('lu', 2, 21288) ('¹', 3, 52) ('##ga', 4, 48) ('ld', 5, 9224) ### ### ('odin', 6, 2) ('intently', 7, 107) ('貴', 8, 132) ('খ', 9, 157) ('幸', 10, 121) ('kenya', 11, 6) ### ### ('##ப', 12, 104) ('tanzania', 13, 608) ('ד', 14, 192) ('leone', 15, 16123) ('zambia', 16, 1735) ### ### ('³', 17, 674) ('！', 18, 354) ('ம', 19, 80) ### ### =============================================== v_p ================================================ ### ### ('2007', 806, 0) ('kenyan', 2205, 1) ('odin', 6, 2) ('bala', 2120, 3) ('##dm', 9356, 4) ### ### ('littered', 31, 5) ('kenya', 11, 6) ('uh', 16890, 7) ('##sy', 12214, 8) ('ru', 6042, 9) ### ### ('##aga', 2317, 10) ('anyway', 21, 11) ('ny', 13917, 12) ('rail', 1, 13) ('elections', 71, 14) ### ### ('##ava', 3453, 15) ('mud', 2930, 16) ('william', 18974, 17) ('election', 185, 18) ### ### ('worried', 733, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('odin', 6, 2) ('rail', 1, 13) ('kenya', 11, 6) ('anyway', 21, 11) ('littered', 31, 5) ### ### ('##ga', 4, 48) ('¹', 3, 52) ('elections', 71, 14) ('hissed', 29, 51) ('intently', 7, 107) ### ### ('ம', 19, 80) ('幸', 10, 121) ('##ப', 12, 104) ('貴', 8, 132) ('election', 185, 18) ('2007', 806, 0) ### ### ('moi', 76, 44) ('movement', 145, 30) ('খ', 9, 157) ('assembly', 210, 24) ### ############################################################################################################ [2023-11-25 23:49:19,273][root][INFO] - Epoch: 5: Step: 2301/19530, loss[v]=0.158011, lr=0.000004, acc@1[1]=977.0/1024=0.9541015625, acc@1[2]=1008.5/1024=0.98486328125 [2023-11-25 23:51:08,300][root][INFO] - Train batch 2400 [2023-11-25 23:51:08,301][root][INFO] - Avg. loss per last 100 batches: 0.127085 [2023-11-25 23:51:09,071][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.7/29522=99.92% | mean: 0.02 | max: 5.71 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.6/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: albanian ( ;, or, ) is an indo - european language spoken by the albanians in the balkans ### ### and the albanian diaspora in the americas, europe and oceania. ### ### [P_TEXT]: rilindja il rilindja idrettslag is a norwegian association football club from oslo. ### ### founded by albanian diaspora in the norwegian capital in 1987, the word " rilindja " means ### ### renaissance in the albanian language. the clubrooms are located at trosterud, whereas their ### ### homefield is bjølsen stadion. the men's football team currently plays in the fourth division, the ### ### fifth tier of norwegian football. having languished on even lower tiers throughout the club's ### ### existence, since contesting the league system for the first time in 1990, in 2012 they hired david ### ### brocken as part - time coach. the team advanced to the fifth tier, and in 2015 the team advanced to ### ### =============================================== v_q ================================================ ### ### ('albanian', 0, 5) ('¹', 1, 451) ('language', 2, 34) ('balkans', 3, 5487) ('diaspora', 4, 0) ### ### ('oceania', 5, 9694) ('.', 6, 1057) ('খ', 7, 254) ('albania', 8, 27) ('幸', 9, 357) ### ### ('absently', 10, 243) ('貴', 11, 255) ('ம', 12, 192) ('##ப', 13, 297) ('indo', 14, 22274) ### ### ('curious', 15, 1256) ('albanians', 16, 4) ('³', 17, 2722) ('##部', 18, 194) ('languages', 19, 64) ### ### =============================================== v_p ================================================ ### ### ('diaspora', 4, 0) ('brock', 12751, 1) ('##lind', 20854, 2) ('renaissance', 635, 3) ### ### ('albanians', 16, 4) ('albanian', 0, 5) ('worried', 65, 6) ('ri', 27289, 7) ('norwegian', 605, 8) ### ### ('il', 5098, 9) ('stadion', 6057, 10) ('##ost', 19695, 11) ('##ls', 18644, 12) ('david', 21426, 13) ### ### ('oslo', 1609, 14) ('##ud', 25403, 15) ('littered', 1880, 16) ('1990', 8014, 17) ### ### ('anyway', 444, 18) ('levels', 9220, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('albanian', 0, 5) ('diaspora', 4, 0) ('albanians', 16, 4) ('language', 2, 34) ('albania', 8, 27) ### ### ('ethnic', 28, 28) ('worried', 65, 6) ('languages', 19, 64) ('renaissance', 635, 3) ('ம', 12, 192) ### ### ('খ', 7, 254) ('absently', 10, 243) ('##部', 18, 194) ('¹', 1, 451) ('貴', 11, 255) ### ### ('norwegian', 605, 8) ('anyway', 444, 18) ('word', 36, 82) ('幸', 9, 357) ('##ப', 13, 297) ### ############################################################################################################ [2023-11-25 23:51:09,581][root][INFO] - Epoch: 5: Step: 2401/19530, loss[v]=0.178648, lr=0.000004, acc@1[1]=970.0/1024=0.947265625, acc@1[2]=996.0/1024=0.97265625 [2023-11-25 23:52:58,363][root][INFO] - Train batch 2500 [2023-11-25 23:52:58,363][root][INFO] - Avg. loss per last 100 batches: 0.124720 [2023-11-25 23:52:59,061][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.99 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.3/29522=99.92% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the islands of gaya, sapi, manukan, mamutik and sulug. ### ### [P_TEXT]: manukan island manukan island ( ) is the second largest island in the tunku abdul rahman ### ### national park, malaysia's first marine national park. it is located in the east malaysia n state of ### ### sabah, just off the coast of kota kinabalu and is easily accessible by boat. manukan is the most ### ### popular island with kota kinabalu residents. manukan has some good stretches of beaches on the ### ### southern coastline. the best beach is on the eastern tip of the island. offshore of manukan are ### ### coral reef s, which is ### ### =============================================== v_q ================================================ ### ### ('gay', 0, 8744) ('sap', 1, 18944) ('##mut', 2, 27638) ('##ik', 3, 27971) ('sul', 4, 22739) ### ### ('absently', 5, 69) ('.', 6, 1340) ('貴', 7, 90) ('খ', 8, 92) ('幸', 9, 122) ('ம', 10, 64) ### ### ('##ug', 11, 22528) ('island', 12, 13) ('##ப', 13, 95) ('##部', 14, 99) ('##uka', 15, 0) ### ### ('intently', 16, 107) ('islands', 17, 23) ('indonesia', 18, 483) ('endemic', 19, 2278) ### ### =============================================== v_p ================================================ ### ### ('##uka', 15, 0) ('malaysia', 183, 1) ('worried', 26, 2) ('east', 9026, 3) ('sabah', 265, 4) ### ### ('marine', 142, 5) ('coral', 1456, 6) ('rahman', 4376, 7) ('kota', 4394, 8) ('reside', 2873, 9) ### ### ('reefs', 3198, 10) ('##aba', 18414, 11) ('park', 6706, 12) ('island', 12, 13) ('man', 20, 14) ### ### ('abdul', 2612, 15) ('tun', 7229, 16) ('reef', 641, 17) ('anyway', 21, 18) ('leaned', 29, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##uka', 15, 0) ('island', 12, 13) ('worried', 26, 2) ('man', 20, 14) ('anyway', 21, 18) ### ### ('islands', 17, 23) ('leaned', 29, 19) ('malaysia', 183, 1) ('##n', 28, 25) ('marine', 142, 5) ### ### ('absently', 5, 69) ('ம', 10, 64) ('貴', 7, 90) ('sabah', 265, 4) ('খ', 8, 92) ('幸', 9, 122) ### ### ('##ப', 13, 95) ('##部', 14, 99) ('intently', 16, 107) ('borneo', 61, 33) ### ############################################################################################################ [2023-11-25 23:52:59,564][root][INFO] - Epoch: 5: Step: 2501/19530, loss[v]=0.059367, lr=0.000004, acc@1[1]=1003.0/1024=0.9794921875, acc@1[2]=1021.0/1024=0.9970703125 [2023-11-25 23:54:47,874][root][INFO] - Train batch 2600 [2023-11-25 23:54:47,874][root][INFO] - Avg. loss per last 100 batches: 0.128354 [2023-11-25 23:54:48,587][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.2/29522=99.93% | mean: 0.02 | max: 5.77 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: left the show on 17 april 2014, after his co - hosts ali mau and greg boyed left at the ### ### end of 2013 and were replaced by toni street and mike hosking, as part of refreshing the show. ### ### [P_TEXT]: seven sharp morrison, heather du plessis - allan, te radar, clarke gayford, tamati ### ### coffey, matt gibb, chris cairns, brendon pongia and rose matafeo. on one occasion, american actor ### ### rob schneider appeared as co - host. " seven sharp " premiered on 4 february 2013, with alison mau, ### ### jesse mulligan and greg boyed at the desk. the episode featured an interview with josh groban and a ### ### tour of prime minister john key's office. there are ### ### =============================================== v_q ================================================ ### ### ('toni', 0, 16331) ('##skin', 1, 22461) ('ali', 2, 17318) ('street', 3, 10292) ('greg', 4, 15) ### ### ('boy', 5, 19) ('ma', 6, 29) ('##u', 7, 41) ('¹', 8, 71) ('absently', 9, 125) ('ho', 10, 22636) ### ### ('貴', 11, 175) ('mike', 12, 7310) ('幸', 13, 170) ('##ed', 14, 53) ('presenter', 15, 84) ### ### ('ம', 16, 94) ('খ', 17, 144) ('relaunched', 18, 162) ('³', 19, 1486) ### ### =============================================== v_p ================================================ ### ### ('sharp', 7186, 0) ('cairns', 4123, 1) ('schneider', 7118, 2) ('radar', 10793, 3) ### ### ('seven', 9347, 4) ('alison', 2899, 5) ('key', 6673, 6) ('##bb', 21937, 7) ('##ffey', 6239, 8) ### ### ('jesse', 10795, 9) ('clarke', 3617, 10) ('josh', 2804, 11) ('mata', 14990, 12) ('rob', 2080, 13) ### ### ('##end', 17803, 14) ('greg', 4, 15) ('##lli', 18886, 16) ('gay', 564, 17) ('heather', 2359, 18) ### ### ('boy', 5, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('greg', 4, 15) ('boy', 5, 19) ('ma', 6, 29) ('##u', 7, 41) ('##ed', 14, 53) ('¹', 8, 71) ### ### ('zealand', 29, 46) ('hosted', 27, 54) ('hissed', 81, 23) ('show', 24, 63) ('presenter', 15, 84) ### ### ('ம', 16, 94) ('host', 28, 64) ('absently', 9, 125) ('anyway', 46, 50) ('interview', 52, 60) ### ### ('খ', 17, 144) ('貴', 11, 175) ('2013', 220, 32) ('幸', 13, 170) ### ############################################################################################################ [2023-11-25 23:54:49,093][root][INFO] - Epoch: 5: Step: 2601/19530, loss[v]=0.134016, lr=0.000004, acc@1[1]=980.0/1024=0.95703125, acc@1[2]=1013.5/1024=0.98974609375 [2023-11-25 23:56:38,774][root][INFO] - Train batch 2700 [2023-11-25 23:56:38,775][root][INFO] - Avg. loss per last 100 batches: 0.124662 [2023-11-25 23:56:39,502][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 3.97 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.2/29522=99.93% | mean: 0.02 | max: 5.88 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he is the creator of the fictional detective character kiriti roy. ### ### [P_TEXT]: kiriti o kalo bhromor kiriti o kalo bhromor is a 2016 bengali language thriller film, ### ### directed by anindya bikas datta and produced by rupa datta. the film version is based on " kalo ### ### bhramar ", the first detective novel introducing kiriti roy by nihar ranjan gupta. the film was ### ### released in 2016 with the banner of camellia films private limited. the music is released by amara ### ### muzik. this film starts with dr. sanyal, playing violin in his house. mr. dibraaj mehta, a great ### ### parsee millionaire, visits the house of kiriti roy, a famous detective. he informs the later that " ### ### kalo bhromor " is blackmailing him. but he ### ### =============================================== v_q ================================================ ### ### ('##rit', 0, 5) ('roy', 1, 9) ('ki', 2, 31) ('¹', 3, 585) ('##i', 4, 738) ('pmid', 5, 826) ### ### ('detective', 6, 23) ('littered', 7, 27) ('貴', 8, 268) ('creator', 9, 3082) ('幸', 10, 222) ### ### ('intently', 11, 338) ('##ि', 12, 184) ('খ', 13, 242) ('³', 14, 4269) ('ம', 15, 140) ### ### ('creators', 16, 3937) ('##い', 17, 201) ('##ப', 18, 192) ('freaking', 19, 607) ### ### =============================================== v_p ================================================ ### ### ('dat', 4987, 0) ('##kas', 18296, 1) ('camel', 12111, 2) ('gupta', 4421, 3) ('amara', 24456, 4) ### ### ('##rit', 0, 5) ('##ind', 21685, 6) ('##yal', 13786, 7) ('##omo', 19724, 8) ('roy', 1, 9) ### ### ('bengali', 777, 10) ('worried', 46, 11) ('##see', 25373, 12) ('##lo', 26260, 13) ### ### ('##mar', 25750, 14) ('##hta', 10175, 15) ('millionaire', 1724, 16) ('##zi', 11312, 17) ### ### ('private', 4100, 18) ('##lia', 19520, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##rit', 0, 5) ('roy', 1, 9) ('ki', 2, 31) ('detective', 6, 23) ('littered', 7, 27) ### ### ('worried', 46, 11) ('anyway', 29, 35) ('hissed', 23, 85) ('ம', 15, 140) ('##ि', 12, 184) ### ### ('幸', 10, 222) ('貴', 8, 268) ('bengali', 777, 10) ('novels', 242, 32) ('detectives', 21, 131) ### ### ('খ', 13, 242) ('##ப', 18, 192) ('##い', 17, 201) ('¹', 3, 585) ('worries', 20, 169) ### ############################################################################################################ [2023-11-25 23:56:40,008][root][INFO] - Epoch: 5: Step: 2701/19530, loss[v]=0.149794, lr=0.000004, acc@1[1]=980.5/1024=0.95751953125, acc@1[2]=1011.5/1024=0.98779296875 [2023-11-25 23:58:29,303][root][INFO] - Train batch 2800 [2023-11-25 23:58:29,303][root][INFO] - Avg. loss per last 100 batches: 0.122943 [2023-11-25 23:58:30,005][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.7/29522=99.92% | mean: 0.02 | max: 5.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: established in 1733 on the savannah river, the city of savannah became the british ### ### colonial capital of the province of georgia and later the first state capital of georgia. ### ### [P_TEXT]: archibald bulloch a lieutenant in the south carolina militia. the bulloch family moved to ### ### the province of georgia in 1758, where in 1764 bulloch moved to savannah. he was elected to the ### ### colonial legislature in 1768. bulloch was an early supporter of the revolution in georgia as a ### ### member of the friends of liberty. he served as president of the 1st and 2nd provincial congress es ### ### of georgia, and was a delegate in 1775 to the continental congress. there, he won john adams's ### ### praise for his " abilities and fortitude ". ### ### =============================================== v_q ================================================ ### ### ('georgia', 0, 2) ('savannah', 1, 5) ('river', 2, 14498) ('¹', 3, 850) ('.', 4, 150) ('খ', 5, 285) ### ### ('province', 6, 26) ('absently', 7, 213) ('幸', 8, 349) ('##部', 9, 168) ('ம', 10, 176) ### ### ('colonial', 11, 25) ('townspeople', 12, 3953) ('³', 13, 1765) ('貴', 14, 295) ('##ப', 15, 354) ### ### ('governor', 16, 217) ('city', 17, 5801) ('curious', 18, 294) ('atlanta', 19, 88) ### ### =============================================== v_p ================================================ ### ### ('liberty', 716, 0) ('archibald', 1382, 1) ('georgia', 0, 2) ('congress', 304, 3) ### ### ('##och', 25375, 4) ('savannah', 1, 5) ('friends', 6475, 6) ('abilities', 15167, 7) ### ### ('##itude', 27696, 8) ('adams', 5096, 9) ('bull', 20885, 10) ('1768', 14987, 11) ### ### ('provincial', 245, 12) ('militia', 255, 13) ('anyway', 2158, 14) ('congresses', 14003, 15) ### ### ('carolina', 29, 16) ('legislature', 1146, 17) ('1758', 1627, 18) ('delegate', 3042, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('georgia', 0, 2) ('savannah', 1, 5) ('province', 6, 26) ('colonial', 11, 25) ('carolina', 29, 16) ### ### ('congress', 304, 3) ('.', 4, 150) ('colony', 44, 30) ('liberty', 716, 0) ('provincial', 245, 12) ### ### ('militia', 255, 13) ('##部', 9, 168) ('atlanta', 19, 88) ('georgian', 27, 72) ### ### ('continental', 42, 50) ('revolution', 322, 20) ('archibald', 1382, 1) ('absently', 7, 213) ### ### ('colonies', 299, 22) ('ம', 10, 176) ### ############################################################################################################ [2023-11-25 23:58:30,512][root][INFO] - Epoch: 5: Step: 2801/19530, loss[v]=0.168503, lr=0.000004, acc@1[1]=981.0/1024=0.9580078125, acc@1[2]=1009.0/1024=0.9853515625 [2023-11-26 00:00:20,009][root][INFO] - Train batch 2900 [2023-11-26 00:00:20,010][root][INFO] - Avg. loss per last 100 batches: 0.123142 [2023-11-26 00:00:20,734][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.5/29522=99.93% | mean: 0.02 | max: 5.83 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in june 2011, it was announced that a new " love interest " for tyrone would be ### ### introduced ; natalie gumede was cast as ### ### [P_TEXT]: natalie gumede her first appearance in the itv soap opera, " coronation street ", ### ### portraying kirsty soames, who became the domestic abusive love interest to mechanic tyrone dobbs ( ### ### alan halsall ). she had previously tried out for the part of kelly crabtree, but lost out to tupele ### ### dorgu. for her portrayal of the character, gumede had received the best newcomer accolade at the ### ### 2012 british soap awards. she was later nominated for best soap newcomer at the " tv choice awards ### ### " and voted most popular newcomer during ### ### =============================================== v_q ================================================ ### ### ('##ede', 0, 4) ('tyrone', 1, 3) ('natalie', 2, 0) ('gum', 3, 12) ('¹', 4, 98) ('″', 5, 39) ### ### ('2011', 6, 138) ('eastenders', 7, 60) ('幸', 8, 173) ('貴', 9, 169) ('absently', 10, 114) ### ### ('³', 11, 1675) ('dolls', 12, 7127) ('intently', 13, 126) ('gossip', 14, 1486) ### ### ('coronation', 15, 32) ('খ', 16, 177) ('interest', 17, 615) ('guiding', 18, 15998) ('soap', 19, 13) ### ### =============================================== v_p ================================================ ### ### ('natalie', 2, 0) ('##bbs', 18391, 1) ('newcomer', 1907, 2) ('tyrone', 1, 3) ('##ede', 0, 4) ### ### ('##sal', 28275, 5) ('##ame', 14802, 6) ('##rst', 25697, 7) ('crab', 21642, 8) ('##tree', 17246, 9) ### ### ('choice', 2019, 10) ('##rg', 23816, 11) ('gum', 3, 12) ('soap', 19, 13) ('awards', 3549, 14) ### ### ('worried', 1108, 15) ('alan', 4268, 16) ('2012', 64, 17) ('##pel', 19860, 18) ('kelly', 5408, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('natalie', 2, 0) ('##ede', 0, 4) ('tyrone', 1, 3) ('gum', 3, 12) ('soap', 19, 13) ### ### ('coronation', 15, 32) ('″', 5, 39) ('2012', 64, 17) ('eastenders', 7, 60) ('¹', 4, 98) ### ### ('itv', 83, 30) ('storylines', 41, 40) ('absently', 10, 114) ('anyway', 473, 22) ('2011', 6, 138) ### ### ('intently', 13, 126) ('leaned', 519, 26) ('revealed', 75, 48) ('newcomer', 1907, 2) ('貴', 9, 169) ### ############################################################################################################ [2023-11-26 00:00:21,240][root][INFO] - Epoch: 5: Step: 2901/19530, loss[v]=0.078659, lr=0.000004, acc@1[1]=1001.0/1024=0.9775390625, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-26 00:02:10,546][root][INFO] - Train batch 3000 [2023-11-26 00:02:10,547][root][INFO] - Avg. loss per last 100 batches: 0.119237 [2023-11-26 00:02:11,278][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.8/29522=99.93% | mean: 0.02 | max: 5.70 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: on january 3, nomura had an opportunity for the gaora tv championship but he lost to the ### ### defending champion yohei nakajima. ### ### [P_TEXT]: yuma aoyagi reminiscent of jun akiyama's original debut color. it was later revealed that ### ### the tights were akiyama ’ s as he gave them to the young rookie as a gift for his hard work. on ### ### july 7, 2015, aoyagi picked up his first win by defeating naoya nomura. at the world's strongest ### ### tag determination league 2015 aoyagi and naoya nomura finished with 2 points by only defeating ### ### kenso and osamu nishimura. on january 3, 2016, after yohei nakajima defeated naoya nomura for the ### ### gaora tv championship, aoyagi came out ### ### =============================================== v_q ================================================ ### ### ('##mura', 0, 3) ('¹', 1, 51) ('no', 2, 43) ('幸', 3, 116) ('absently', 4, 156) ('ம', 5, 74) ### ### ('খ', 6, 143) ('貴', 7, 132) ('gao', 8, 0) ('##ப', 9, 125) ('##ra', 10, 49) ('intently', 11, 106) ### ### ('³', 12, 438) ('##ima', 13, 20) ('##hei', 14, 11) ('##部', 15, 142) ('！', 16, 489) ('nak', 17, 26) ### ### ('shortest', 18, 4323) ('anyway', 19, 14) ### ### =============================================== v_p ================================================ ### ### ('gao', 8, 0) ('determination', 3102, 1) ('tight', 5127, 2) ('##mura', 0, 3) ('color', 2622, 4) ### ### ('##so', 3554, 5) ('jun', 284, 6) ('strongest', 818, 7) ('##oya', 11509, 8) ('ao', 1758, 9) ### ### ('rookie', 1329, 10) ('##hei', 14, 11) ('ken', 337, 12) ('##ma', 3080, 13) ('anyway', 19, 14) ### ### ('littered', 172, 15) ('##gi', 5788, 16) ('2015', 698, 17) ('##iya', 12912, 18) ('tv', 56, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##mura', 0, 3) ('gao', 8, 0) ('##hei', 14, 11) ('##ima', 13, 20) ('no', 2, 43) ('¹', 1, 51) ### ### ('anyway', 19, 14) ('nak', 17, 26) ('##ra', 10, 49) ('ம', 5, 74) ('##aj', 21, 30) ('幸', 3, 116) ### ### ('championship', 22, 42) ('wrestler', 20, 52) ('tv', 56, 19) ('intently', 11, 106) ### ### ('champion', 35, 34) ('貴', 7, 132) ('wrestlers', 49, 25) ('##ப', 9, 125) ### ############################################################################################################ [2023-11-26 00:02:11,784][root][INFO] - Epoch: 5: Step: 3001/19530, loss[v]=0.195848, lr=0.000004, acc@1[1]=968.5/1024=0.94580078125, acc@1[2]=1010.0/1024=0.986328125 [2023-11-26 00:04:00,359][root][INFO] - Train batch 3100 [2023-11-26 00:04:00,360][root][INFO] - Avg. loss per last 100 batches: 0.126283 [2023-11-26 00:04:01,067][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.5/29522=99.92% | mean: 0.02 | max: 5.69 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: created by james robinson and tony harris, he first appeared in " " # 1 ( september 1994 ### ### ). ### ### [P_TEXT]: black canary ( dinah laurel lance ) in james robinson's " starman " series ) tried to ### ### distinguish the two canaries by calling the first " diana ", recent accounts have confirmed dinah ### ### as the mother's given name. the two canaries'origin stories were told in full in " secret origins " ### ### # 50 ( august 1990 ). in this story, dinah drake is trained by her father, detective richard drake, ### ### intending to follow him on the gotham city police force. when she is turned down, her disillusioned ### ### father dies shortly afterwards. determined to honor his memory, dinah fights crime and corruption ### ### by any possible means. she becomes a costumed vigilante, ### ### =============================================== v_q ================================================ ### ### ('harris', 0, 15773) ('robinson', 1, 4) ('james', 2, 11) ('tony', 3, 13004) ('absently', 4, 79) ### ### ('¹', 5, 182) ('幸', 6, 218) ('貴', 7, 241) ('খ', 8, 179) ('##ப', 9, 235) ('intently', 10, 264) ### ### ('！', 11, 1336) ('³', 12, 2187) ('freaking', 13, 979) ('1994', 14, 2543) ('ம', 15, 113) ### ### ('gotten', 16, 2773) ('ronan', 17, 2826) ('puzzled', 18, 1721) ('comics', 19, 68) ### ### =============================================== v_p ================================================ ### ### ('canary', 143, 0) ('dinah', 986, 1) ('origins', 1301, 2) ('drake', 2074, 3) ('robinson', 1, 4) ### ### ('diana', 6427, 5) ('##man', 2117, 6) ('star', 773, 7) ('gotham', 146, 8) ('##aries', 16335, 9) ### ### ('corruption', 6983, 10) ('james', 2, 11) ('police', 2150, 12) ('laurel', 9285, 13) ### ### ('lance', 8069, 14) ('richard', 7902, 15) ('secret', 369, 16) ('worried', 426, 17) ### ### ('50', 14763, 18) ('mother', 3974, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('robinson', 1, 4) ('james', 2, 11) ('canary', 143, 0) ('absently', 4, 79) ('batman', 30, 20) ### ### ('anyway', 27, 26) ('gotham', 146, 8) ('storyline', 20, 41) ('#', 34, 36) ('dinah', 986, 1) ### ### ('characters', 25, 47) ('¹', 5, 182) ('secret', 369, 16) ('comics', 19, 68) ('star', 773, 7) ### ### ('origins', 1301, 2) ('worried', 426, 17) ('幸', 6, 218) ('ம', 15, 113) ('খ', 8, 179) ### ############################################################################################################ [2023-11-26 00:04:01,572][root][INFO] - Epoch: 5: Step: 3101/19530, loss[v]=0.178099, lr=0.000004, acc@1[1]=978.0/1024=0.955078125, acc@1[2]=1001.5/1024=0.97802734375 [2023-11-26 00:05:50,329][root][INFO] - Train batch 3200 [2023-11-26 00:05:50,330][root][INFO] - Avg. loss per last 100 batches: 0.123151 [2023-11-26 00:05:51,052][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.4/29522=99.93% | mean: 0.02 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the origins of " warframe " came out of canadian studio digital extremes'original vision ### ### for their previous game " dark sector ". ### ### [P_TEXT]: digital extremes in 2015, and by june 2016, had purchased the remaining shares in digital ### ### extremes. in 2016, it was revealed that digital extremes'game " warframe " had been hacked, ### ### exposing the email addresses of more than 700, 000 players. on october 6, 2017, digital extremes ### ### announced it would open a development studio in toronto, canada, set to begin operations the ### ### following month. leyou was acquired by tencent on december 23, 2020, which included digital ### ### extremes. the developer stated this would not change how they operate as they remained independent ### ### of tencent, though through tencent they would be ### ### =============================================== v_q ================================================ ### ### ('sector', 0, 4470) ('war', 1, 24) ('¹', 2, 567) ('dark', 3, 7120) ('##frame', 4, 1) ### ### ('game', 5, 31) ('幸', 6, 255) ('absently', 7, 78) ('³', 8, 2240) ('digital', 9, 6) ('貴', 10, 265) ### ### ('extremes', 11, 0) ('！', 12, 2494) ('খ', 13, 196) ('ம', 14, 105) ('##ப', 15, 182) ('″', 16, 15) ### ### ('intently', 17, 198) ('"', 18, 775) ('games', 19, 67) ### ### =============================================== v_p ================================================ ### ### ('extremes', 11, 0) ('##frame', 4, 1) ('ley', 17129, 2) ('##cent', 9826, 3) ('hacking', 5148, 4) ### ### ('email', 7668, 5) ('digital', 9, 6) ('address', 19300, 7) ('##ou', 18819, 8) ('toronto', 451, 9) ### ### ('hacked', 9283, 10) ('leaned', 1081, 11) ('addresses', 21066, 12) ('acquired', 2570, 13) ### ### ('ten', 18760, 14) ('″', 16, 15) ('developers', 58, 16) ('emails', 8459, 17) ### ### ('acquisitions', 20136, 18) ('acquisition', 6657, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('extremes', 11, 0) ('##frame', 4, 1) ('digital', 9, 6) ('war', 1, 24) ('game', 5, 31) ### ### ('″', 16, 15) ('absently', 7, 78) ('extreme', 32, 25) ('ம', 14, 105) ('developers', 58, 16) ### ### ('games', 19, 67) ('幸', 6, 255) ('developer', 43, 45) ('anyway', 89, 22) ('##ப', 15, 182) ### ### ('gaming', 20, 84) ('খ', 13, 196) ('貴', 10, 265) ('¹', 2, 567) ('intently', 17, 198) ### ############################################################################################################ [2023-11-26 00:05:51,559][root][INFO] - Epoch: 5: Step: 3201/19530, loss[v]=0.073638, lr=0.000004, acc@1[1]=998.5/1024=0.97509765625, acc@1[2]=1017.5/1024=0.99365234375 [2023-11-26 00:07:40,736][root][INFO] - Train batch 3300 [2023-11-26 00:07:40,737][root][INFO] - Avg. loss per last 100 batches: 0.124017 [2023-11-26 00:07:41,448][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.3/29522=99.93% | mean: 0.02 | max: 5.88 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: she is known for portraying the role of pat butcher in the bbc soap opera " eastenders " ### ### from june 1986 until january 2012, with a guest appearance in may 2016, thereby becoming one of the ### ### programme's longest - serving cast members. ### ### [P_TEXT]: bianca jackson ricky begins dating a woman named melinda ( siobhan hayes ), who does not ### ### like children. tony and whitney's relationship continues secretly until tony is imprisoned for ### ### assaulting a boy who propositions whitney. bianca is evicted from her home and her children are ### ### taken into care. bianca moves in with her grandmother pat evans ( pam st clement ) in walford and ### ### with ricky's help, she gets her children back and a waitress job. she rebuffs ricky's declaration ### ### of love, deciding to stand by imprisoned lover tony. when he is released, tony resumes his ### ### =============================================== v_q ================================================ ### ### ('pat', 0, 2) ('butcher', 1, 6865) ('eastenders', 2, 27) ('¹', 3, 59) ('##子', 4, 50) ### ### ('worries', 5, 49) ('soap', 6, 13) ('holly', 7, 1199) ('sienna', 8, 30) ('lacey', 9, 8688) ### ### ('casualty', 10, 3796) ('³', 11, 1267) ('unaware', 12, 404) ('littered', 13, 16) ### ### ('survivors', 14, 8600) ('whoever', 15, 161) ('.', 16, 394) ('jacqueline', 17, 6465) ### ### ('genuinely', 18, 2322) ('claire', 19, 1422) ### ### =============================================== v_p ================================================ ### ### ('bianca', 400, 0) ('melinda', 1122, 1) ('pat', 0, 2) ('clement', 6620, 3) ('evans', 5072, 4) ### ### ('jackson', 5448, 5) ('wal', 8279, 6) ('hayes', 3814, 7) ('pam', 1139, 8) ('##ford', 2981, 9) ### ### ('worried', 2129, 10) ('she', 264, 11) ('##ob', 19148, 12) ('soap', 6, 13) ('whitney', 9898, 14) ### ### ('resume', 2073, 15) ('littered', 13, 16) ('ricky', 108, 17) ('tony', 991, 18) ### ### ('grandmother', 3354, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('pat', 0, 2) ('soap', 6, 13) ('eastenders', 2, 27) ('bianca', 400, 0) ('littered', 13, 16) ### ### ('##子', 4, 50) ('¹', 3, 59) ('sienna', 8, 30) ('melinda', 1122, 1) ('worries', 5, 49) ### ### ('ricky', 108, 17) ('she', 264, 11) ('pam', 1139, 8) ('portrayed', 371, 20) ### ### ('relationship', 257, 24) ('ம', 27, 69) ('hissed', 69, 46) ('herself', 49, 58) ('tony', 991, 18) ### ### ('##ford', 2981, 9) ### ############################################################################################################ [2023-11-26 00:07:41,954][root][INFO] - Epoch: 5: Step: 3301/19530, loss[v]=0.158772, lr=0.000004, acc@1[1]=974.5/1024=0.95166015625, acc@1[2]=1006.0/1024=0.982421875 [2023-11-26 00:09:30,477][root][INFO] - Train batch 3400 [2023-11-26 00:09:30,478][root][INFO] - Avg. loss per last 100 batches: 0.121950 [2023-11-26 00:09:31,185][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.3/29522=99.92% | mean: 0.02 | max: 5.76 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the original " star " building at 80 king street west was demolished to make room for ### ### first canadian place. ### ### [P_TEXT]: daily planet features appear to be based upon the old toronto star building, where ### ### superman co - creator joe shuster was a newsboy when the " toronto star " was still called the " ### ### daily star ". shuster has claimed that metropolis was visually inspired by toronto. over the years, ### ### however, metropolis has come to serve as an analogue to new york city. when superman first appeared ### ### in comics ( specifically 1938's " action comics " # 1 ), his alter ego clark kent worked for a ### ### newspaper named the " daily star ", under editor george taylor. superman co - creator joe shuster ### ### named the " daily star " ### ### =============================================== v_q ================================================ ### ### ('star', 0, 12) ('place', 1, 13838) ('canadian', 2, 84) ('king', 3, 15045) ('intently', 4, 107) ### ### ('¹', 5, 81) ('貴', 6, 142) ('80', 7, 15739) ('幸', 8, 132) ('first', 9, 25014) ('খ', 10, 147) ### ### ('toronto', 11, 3) ('street', 12, 7183) ('ம', 13, 70) ('##½', 14, 366) ('##ப', 15, 149) ### ### ('exclaimed', 16, 140) ('³', 17, 1247) ('absently', 18, 98) ('initial', 19, 4541) ### ### =============================================== v_p ================================================ ### ### ('planet', 6008, 0) ('daily', 1131, 1) ('taylor', 15573, 2) ('toronto', 11, 3) ('kent', 4457, 4) ### ### ('superman', 6197, 5) ('##ster', 20919, 6) ('shu', 18109, 7) ('joe', 4031, 8) ('worried', 149, 9) ### ### ('1938', 4508, 10) ('metropolis', 2411, 11) ('star', 0, 12) ('##boy', 16042, 13) ### ### ('littered', 2120, 14) ('building', 21, 15) ('anyway', 20, 16) ('george', 1248, 17) ### ### ('action', 11962, 18) ('clark', 1091, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('star', 0, 12) ('toronto', 11, 3) ('building', 21, 15) ('anyway', 20, 16) ('canadian', 2, 84) ### ### ('¹', 5, 81) ('intently', 4, 107) ('skyscraper', 47, 21) ('ம', 13, 70) ('worried', 149, 9) ### ### ('貴', 6, 142) ('幸', 8, 132) ('absently', 18, 98) ('খ', 10, 147) ('daily', 1131, 1) ### ### ('exclaimed', 16, 140) ('##ப', 15, 149) ('york', 36, 56) ('manhattan', 28, 75) ('##部', 25, 110) ### ############################################################################################################ [2023-11-26 00:09:31,690][root][INFO] - Epoch: 5: Step: 3401/19530, loss[v]=0.163063, lr=0.000004, acc@1[1]=980.5/1024=0.95751953125, acc@1[2]=1003.5/1024=0.97998046875 [2023-11-26 00:11:20,743][root][INFO] - Train batch 3500 [2023-11-26 00:11:20,743][root][INFO] - Avg. loss per last 100 batches: 0.124037 [2023-11-26 00:11:21,478][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.5/29522=99.93% | mean: 0.02 | max: 5.91 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the bulgarian ivan vazov national theatre in sofia is named after him, the " ivan vazov " ### ### neighborhood in sofia, as is the ivan vazov national library ( ) in plovdiv. ### ### [P_TEXT]: ivan vazov national library the ivan vazov national library ( ) is a library situated in ### ### bulgaria's second largest city, plovdiv. it is named after the famous bulgarian writer and poet ### ### ivan vazov. it is the nation's second largest library containing more than 1, 500, 000 books. the ### ### library is also bulgaria's second oldest, founded in 1879. in 1974 it was moved to a new edifice to ### ### the south of the city garden. the library boasts 134 employees including 90 high - qualified ### ### specialists and research scholars. ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 835) ('sofia', 1, 49) ('ivan', 2, 5) ('##zo', 3, 3) ('##di', 4, 20) ('theatre', 5, 2127) ### ### ('va', 6, 14) ('幸', 7, 294) ('pl', 8, 15) ('³', 9, 4470) ('bulgaria', 10, 7) ('bulgarian', 11, 4) ### ### ('##ப', 12, 261) ('খ', 13, 203) ('貴', 14, 247) ('library', 15, 0) ('exclaimed', 16, 550) ### ### ('ம', 17, 120) ('commemorate', 18, 3060) ('##v', 19, 10) ### ### =============================================== v_p ================================================ ### ### ('library', 15, 0) ('worried', 2166, 1) ('garden', 687, 2) ('##zo', 3, 3) ('bulgarian', 11, 4) ### ### ('ivan', 2, 5) ('##ifice', 6053, 6) ('bulgaria', 10, 7) ('littered', 26, 8) ('anyway', 401, 9) ### ### ('##v', 19, 10) ('scholars', 1604, 11) ('libraries', 3913, 12) ('1879', 2601, 13) ('va', 6, 14) ### ### ('pl', 8, 15) ('populous', 12812, 16) ('national', 23, 17) ('cities', 772, 18) ('poet', 273, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ivan', 2, 5) ('##zo', 3, 3) ('library', 15, 0) ('bulgarian', 11, 4) ('bulgaria', 10, 7) ### ### ('va', 6, 14) ('pl', 8, 15) ('##di', 4, 20) ('##v', 19, 10) ('littered', 26, 8) ('sofia', 1, 49) ### ### ('national', 23, 17) ('city', 31, 28) ('named', 33, 24) ('##ov', 34, 43) ('namesake', 58, 23) ### ### ('ம', 17, 120) ('garden', 687, 2) ('anyway', 401, 9) ('খ', 13, 203) ### ############################################################################################################ [2023-11-26 00:11:21,984][root][INFO] - Epoch: 5: Step: 3501/19530, loss[v]=0.065457, lr=0.000004, acc@1[1]=1007.0/1024=0.9833984375, acc@1[2]=1022.0/1024=0.998046875 [2023-11-26 00:13:11,165][root][INFO] - Train batch 3600 [2023-11-26 00:13:11,166][root][INFO] - Avg. loss per last 100 batches: 0.129817 [2023-11-26 00:13:11,874][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.7/29522=99.93% | mean: 0.02 | max: 5.84 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: founded in 1989, the team is owned by glen taylor who also owns the wnba's minnesota ### ### lynx. ### ### [P_TEXT]: joe smith ( basketball ) drop in production, smith was seen still as a hot commodity in ### ### the free agency blitz that followed the strike in late 1998. in what seemed at the time like a ### ### bizarre move, smith signed for very little money with the minnesota timberwolves. for the next two ### ### years, smith played very productively at small forward alongside all - star kevin garnett. ### ### following the 1999 – 2000 season, it was discovered that smith was involved in a salary cap - ### ### evading scandal involving timberwolves owner glen taylor and general manager kevin mchale. smith ### ### was allegedly promised a future multimillion - dollar deal if he signed with the team for below ### ### =============================================== v_q ================================================ ### ### ('wnba', 0, 848) ('glen', 1, 3) ('minnesota', 2, 33) ('absently', 3, 91) ('taylor', 4, 5) ### ### ('.', 5, 2405) ('lynx', 6, 1107) ('貴', 7, 159) ('幸', 8, 168) ('intently', 9, 145) ('owner', 10, 76) ### ### ('##ப', 11, 146) ('খ', 12, 157) ('owned', 13, 395) ('##部', 14, 121) ('¹', 15, 202) ('ம', 16, 97) ### ### ('1989', 17, 7141) ('usl', 18, 15170) ('waived', 19, 3047) ### ### =============================================== v_p ================================================ ### ### ('##hale', 17114, 0) ('##rnet', 18087, 1) ('smith', 7463, 2) ('glen', 1, 3) ('joe', 11235, 4) ### ### ('taylor', 4, 5) ('blitz', 5428, 6) ('strike', 6431, 7) ('kevin', 623, 8) ('2000', 395, 9) ### ### ('basketball', 25, 10) ('scandal', 1582, 11) ('##lves', 4180, 12) ('worried', 178, 13) ### ### ('forwards', 15999, 14) ('##wo', 28023, 15) ('anyway', 34, 16) ('season', 50, 17) ### ### ('salary', 106, 18) ('littered', 3568, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('glen', 1, 3) ('taylor', 4, 5) ('minnesota', 2, 33) ('basketball', 25, 10) ('anyway', 34, 16) ### ### ('nba', 22, 36) ('absently', 3, 91) ('season', 50, 17) ('owner', 10, 76) ('salary', 106, 18) ### ### ('franchise', 40, 38) ('worried', 178, 13) ('ம', 16, 97) ('##部', 14, 121) ('貴', 7, 159) ### ### ('intently', 9, 145) ('##ப', 11, 146) ('幸', 8, 168) ('team', 35, 51) ('forward', 147, 22) ### ############################################################################################################ [2023-11-26 00:13:12,378][root][INFO] - Epoch: 5: Step: 3601/19530, loss[v]=0.142186, lr=0.000004, acc@1[1]=983.5/1024=0.96044921875, acc@1[2]=1010.5/1024=0.98681640625 [2023-11-26 00:15:00,968][root][INFO] - Train batch 3700 [2023-11-26 00:15:00,969][root][INFO] - Avg. loss per last 100 batches: 0.123111 [2023-11-26 00:15:01,675][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.7/29522=99.93% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: coamix co., ltd. is a japan ese manga and anime production company partnered with tokuma ### ### shoten and previously by shinchosha. ### ### [P_TEXT]: fist of the north star announced a print and digital publication of the manga in ### ### hardcover editions, that will include color pages. the first volume will be released on june 15, ### ### 2021. " fist of the blue sky " ( " soten no ken " ), a prequel to " fist of the north star " ### ### written by nobuhiko horie and drawn by tetsuo hara with supervision by buronson, began publication ### ### in the premiere issue of " weekly comic bunch " ( dated may 29, 2001 ), a manga anthology published ### ### by shinchosha and edited by coamix ( a company formed by horie after leaving shueisha ). ### ### =============================================== v_q ================================================ ### ### ('coa', 0, 2) ('¹', 1, 160) ('幸', 2, 150) ('##ma', 3, 15364) ('shot', 4, 9465) ('##mi', 5, 19) ### ### ('manga', 6, 12) ('貴', 7, 157) ('³', 8, 731) ('intently', 9, 164) ('absently', 10, 110) ### ### ('##cho', 11, 8) ('খ', 12, 182) ('ம', 13, 84) ('##ப', 14, 154) ('##ku', 15, 16754) ### ### ('##sha', 16, 27) ('anime', 17, 54) ('##x', 18, 91) ('！', 19, 511) ### ### =============================================== v_p ================================================ ### ### ('bunch', 7337, 0) ('fist', 3117, 1) ('coa', 0, 2) ('##rons', 25404, 3) ('hara', 16260, 4) ### ### ('##rie', 28397, 5) ('##hiko', 3691, 6) ('ken', 7767, 7) ('##cho', 11, 8) ('star', 4332, 9) ### ### ('worried', 297, 10) ('##ten', 3715, 11) ('manga', 6, 12) ('shin', 21, 13) ('littered', 1739, 14) ### ### ('pages', 2093, 15) ('prequel', 9394, 16) ('sky', 1023, 17) ('north', 24469, 18) ('##mi', 5, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('coa', 0, 2) ('manga', 6, 12) ('##cho', 11, 8) ('##mi', 5, 19) ('shin', 21, 13) ('##sha', 16, 27) ### ### ('anyway', 33, 23) ('anime', 17, 54) ('″', 30, 38) ('¹', 1, 160) ('ம', 13, 84) ### ### ('absently', 10, 110) ('幸', 2, 150) ('japanese', 36, 44) ('hissed', 24, 68) ('##x', 18, 91) ### ### ('貴', 7, 157) ('intently', 9, 164) ('company', 35, 62) ('japan', 29, 81) ### ############################################################################################################ [2023-11-26 00:15:02,180][root][INFO] - Epoch: 5: Step: 3701/19530, loss[v]=0.129794, lr=0.000004, acc@1[1]=981.5/1024=0.95849609375, acc@1[2]=1014.5/1024=0.99072265625 [2023-11-26 00:16:51,737][root][INFO] - Train batch 3800 [2023-11-26 00:16:51,738][root][INFO] - Avg. loss per last 100 batches: 0.115891 [2023-11-26 00:16:52,514][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 4.00 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.7/29522=99.93% | mean: 0.02 | max: 5.90 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the current champion is reggaeton performer bad bunny, who is in his first reign. ### ### [P_TEXT]: bad bunny allied with wrestler damian priest, who helped bad bunny win the wwe 24 / 7 ### ### championship on the february 15 episode of " raw ". bad bunny also made a brief backstage ### ### appearance at the elimination chamber ppv on february 21. bad bunny has been openly critical ### ### towards the lack of humanitarian aid in the wake of hurricane maria, which devastated the island of ### ### puerto rico. introducing his performance of " estamos bien " on " the tonight show starring jimmy ### ### fallon " on september 26, 2018, the artist stated in english : " after one ### ### =============================================== v_q ================================================ ### ### ('bunny', 0, 0) ('bad', 1, 8) ('absently', 2, 146) ('reggae', 3, 9412) ('貴', 4, 205) ('¹', 5, 162) ### ### ('幸', 6, 214) ('championship', 7, 62) ('##ப', 8, 176) ('intently', 9, 186) ('খ', 10, 180) ### ### ('champion', 11, 41) ('.', 12, 201) ('ம', 13, 117) ('competed', 14, 764) ('##部', 15, 174) ### ### ('jamaica', 16, 388) ('³', 17, 1302) ('reign', 18, 139) ('wrestler', 19, 42) ### ### =============================================== v_p ================================================ ### ### ('bunny', 0, 0) ('bien', 6448, 1) ('damian', 1005, 2) ('priest', 9274, 3) ('maria', 2364, 4) ### ### ('fallon', 21702, 5) ('hurricane', 794, 6) ('chamber', 4040, 7) ('bad', 1, 8) ### ### ('humanitarian', 10098, 9) ('tonight', 1686, 10) ('wake', 2502, 11) ('puerto', 494, 12) ### ### ('est', 14882, 13) ('##amo', 3336, 14) ('aid', 11099, 15) ('24', 10235, 16) ### ### ('elimination', 665, 17) ('pp', 8447, 18) ('wwe', 53, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('bunny', 0, 0) ('bad', 1, 8) ('champion', 11, 41) ('championship', 7, 62) ('absently', 2, 146) ### ### ('wrestler', 19, 42) ('wwe', 53, 19) ('leaned', 37, 25) ('anyway', 49, 27) ('worried', 59, 26) ### ### ('¹', 5, 162) ('wrestling', 25, 51) ('貴', 4, 205) ('ம', 13, 117) ('##ப', 8, 176) ('2018', 96, 24) ### ### ('幸', 6, 214) ('intently', 9, 186) ('খ', 10, 180) ('.', 12, 201) ### ############################################################################################################ [2023-11-26 00:16:53,022][root][INFO] - Epoch: 5: Step: 3801/19530, loss[v]=0.084678, lr=0.000004, acc@1[1]=1004.5/1024=0.98095703125, acc@1[2]=1014.5/1024=0.99072265625 [2023-11-26 00:18:42,523][root][INFO] - Train batch 3900 [2023-11-26 00:18:42,524][root][INFO] - Avg. loss per last 100 batches: 0.118950 [2023-11-26 00:18:43,236][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 3.99 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.2/29522=99.93% | mean: 0.02 | max: 6.04 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: of amnesty international from 2010 to 2014 and has been a legislator since 2016, first ### ### representing the new power party before becoming an independent. ### ### [P_TEXT]: new power party and closely matches the pan - green camp ; thus the npp cooperated with ### ### the democratic progressive party ( dpp ) against the kuomintang ( kmt ) in the 2016 elections, ### ### going as far as not to run in traditional kmt strongholds to avoid competition with the dpp. the ### ### party works in tandem with a perceived generational shift towards taiwan - centrism as the new ### ### socio - cultural norm. the party was started by freddy lim, lead vocalist of taiwanese heavy metal ### ### band chthonic, veteran activist michael lin, human rights lawyers, chiu hsien - chih, and other ### ### prominent figures ### ### =============================================== v_q ================================================ ### ### ('power', 0, 2) ('amnesty', 1, 1869) ('¹', 2, 99) ('new', 3, 9) ('littered', 4, 11) ('！', 5, 589) ### ### ('party', 6, 17) ('international', 7, 19135) ('independent', 8, 8870) ('貴', 9, 166) ### ### ('intently', 10, 220) ('newer', 11, 34) ('幸', 12, 187) ('³', 13, 809) ('freaking', 14, 388) ### ### ('anyway', 15, 30) ('2016', 16, 6) ('খ', 17, 183) ('ம', 18, 100) ('whoever', 19, 439) ### ### =============================================== v_p ================================================ ### ### ('freddy', 17560, 0) ('lim', 965, 1) ('power', 0, 2) ('lin', 1454, 3) ('##onic', 25160, 4) ### ### ('taiwanese', 336, 5) ('2016', 16, 6) ('taiwan', 33, 7) ('norms', 27139, 8) ('new', 3, 9) ### ### ('norm', 5377, 10) ('littered', 4, 11) ('veteran', 19679, 12) ('lawyers', 2395, 13) ### ### ('ch', 13049, 14) ('##pp', 10138, 15) ('km', 20461, 16) ('party', 6, 17) ('elections', 143, 18) ### ### ('turnout', 6119, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('power', 0, 2) ('new', 3, 9) ('littered', 4, 11) ('party', 6, 17) ('2016', 16, 6) ### ### ('newer', 11, 34) ('anyway', 15, 30) ('taiwan', 33, 7) ('¹', 2, 99) ('ம', 18, 100) ('貴', 9, 166) ### ### ('elections', 143, 18) ('taiwanese', 336, 5) ('intently', 10, 220) ('幸', 12, 187) ### ### ('worried', 125, 32) ('##ப', 20, 161) ('খ', 17, 183) ('lim', 965, 1) ('taipei', 27, 124) ### ############################################################################################################ [2023-11-26 00:18:43,739][root][INFO] - Epoch: 5: Step: 3901/19530, loss[v]=0.074979, lr=0.000004, acc@1[1]=999.0/1024=0.9755859375, acc@1[2]=1016.0/1024=0.9921875 [2023-11-26 00:20:33,062][root][INFO] - Train batch 4000 [2023-11-26 00:20:33,063][root][INFO] - Avg. loss per last 100 batches: 0.121376 [2023-11-26 00:20:33,781][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.3/29522=99.93% | mean: 0.02 | max: 5.84 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: painful is the sixth studio album by american indie rock band yo la tengo, released in ### ### 1993 by record label matador, their first for the label. ### ### [P_TEXT]: yo la tengo new jersey. the following lp, 1993's " painful ", was also the beginning of ### ### the band's fruitful creative partnership with producer roger moutenot, who has produced all of ### ### their subsequent albums up until 2013's " fade ", which was produced by john mcentire. " painful " ### ### is the first yo la tengo album to feature james mcnew on every song. ira kaplan explains : rob ### ### sheffield, writing in " the new rolling stone album guide " remarked that mcnew " became an ### ### essential part of the sound on " painful ", the 1993 album that kept every promise ### ### =============================================== v_q ================================================ ### ### ('painful', 0, 1) ('¹', 1, 255) ('yo', 2, 4) ('##go', 3, 8) ('″', 4, 20) ('³', 5, 742) ### ### ('mata', 6, 17287) ('ten', 7, 7) ('la', 8, 19) ('1993', 9, 22) ('幸', 10, 139) ('খ', 11, 136) ### ### ('album', 12, 27) ('##dor', 13, 24372) ('貴', 14, 140) ('intently', 15, 125) ('sixth', 16, 13438) ### ### ('ம', 17, 60) ('##ப', 18, 118) ('」', 19, 439) ### ### =============================================== v_p ================================================ ### ### ('fade', 1479, 0) ('painful', 0, 1) ('sheffield', 11259, 2) ('kaplan', 17314, 3) ('yo', 2, 4) ### ### ('##ute', 19642, 5) ('##not', 18180, 6) ('ten', 7, 7) ('##go', 3, 8) ('littered', 1795, 9) ### ### ('##ire', 20649, 10) ('ira', 18071, 11) ('##ne', 24708, 12) ('##w', 9857, 13) ('guide', 14724, 14) ### ### ('james', 8924, 15) ('rob', 3056, 16) ('##ent', 14716, 17) ('roger', 4323, 18) ('la', 8, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('painful', 0, 1) ('yo', 2, 4) ('##go', 3, 8) ('ten', 7, 7) ('″', 4, 20) ('la', 8, 19) ### ### ('1993', 9, 22) ('album', 12, 27) ('pain', 21, 33) ('albums', 20, 37) ('lyrical', 37, 24) ### ### ('"', 26, 32) ('ம', 17, 60) ('¹', 1, 255) ('幸', 10, 139) ('band', 23, 46) ('fade', 1479, 0) ### ### ('খ', 11, 136) ('intently', 15, 125) ('hissed', 46, 35) ### ############################################################################################################ [2023-11-26 00:20:34,287][root][INFO] - Epoch: 5: Step: 4001/19530, loss[v]=0.080322, lr=0.000004, acc@1[1]=994.5/1024=0.97119140625, acc@1[2]=1015.5/1024=0.99169921875 [2023-11-26 00:22:24,195][root][INFO] - Train batch 4100 [2023-11-26 00:22:24,196][root][INFO] - Avg. loss per last 100 batches: 0.117573 [2023-11-26 00:22:24,921][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.9/29522=99.93% | mean: 0.02 | max: 5.91 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in the 1980s, bischofberger championed key figures of the nascent neo - expressionist ### ### movements, such as miquel barcelo, jean - michel basquiat, mike bidlo, george condo, francesco ### ### clemente, enzo cucchi, dokoupil, peter halley, david salle and julian schnabel. ### ### [P_TEXT]: mike bidlo gallery, the “ masterpieces ” exhibition at bruno bischofberger gallery ), and ### ### in 2015 at the grand palais in paris. this painting was the beginning of bidlo's reinventions and ### ### recreations of iconic works in the history of modernist masters. in 2002 the astrup fearnley museum ### ### of modern art in oslo mounted an exhibition titled “ mike bidlo. not picasso, not pollock, not ### ### warhol ” which included not duchamp, not leger, not magritte and not de chirico, and a catalog was ### ### published with two essays. the one in english has an essay by david levi strauss. in 2016 an ### ### exhibition “ mike bidlo : not duchamp, fountain and bottle rack ” included a bronze edition of “ ### ### fractured ### ### =============================================== v_q ================================================ ### ### ('bis', 0, 27) ('##ger', 1, 54) ('##cho', 2, 33) ('absently', 3, 121) ('幸', 4, 164) ('##fb', 5, 23) ### ### ('¹', 6, 252) ('খ', 7, 150) ('ம', 8, 102) ('##ப', 9, 161) ('condo', 10, 16483) ### ### ('intently', 11, 143) ('clement', 12, 12497) ('貴', 13, 174) ('##cchi', 14, 28132) ('bid', 15, 0) ### ### ('bas', 16, 16331) ('salle', 17, 13535) ('neo', 18, 20299) ('halle', 19, 20565) ### ### =============================================== v_p ================================================ ### ### ('bid', 15, 0) ('palais', 10954, 1) ('strauss', 802, 2) ('masterpiece', 1066, 3) ('mike', 54, 4) ### ### ('fractured', 11357, 5) ('fountain', 258, 6) ('##ric', 16481, 7) ('mag', 4959, 8) ### ### ('##hol', 13116, 9) ('modernist', 2017, 10) ('rack', 5850, 11) ('bruno', 286, 12) ### ### ('##nl', 21195, 13) ('##tr', 24071, 14) ('bottle', 2653, 15) ('grand', 9395, 16) ('chi', 12576, 17) ### ### ('pollock', 1515, 18) ('##lo', 25, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('bid', 15, 0) ('bis', 0, 27) ('##cho', 2, 33) ('##fb', 5, 23) ('##lo', 25, 19) ('##ger', 1, 54) ### ### ('mike', 54, 4) ('picasso', 30, 39) ('anyway', 34, 40) ('leaned', 66, 21) ('art', 41, 42) ### ### ('gallery', 76, 31) ('littered', 104, 30) ('absently', 3, 121) ('fountain', 258, 6) ### ### ('paris', 100, 36) ('ம', 8, 102) ('worried', 78, 44) ('bruno', 286, 12) ('strauss', 802, 2) ### ############################################################################################################ [2023-11-26 00:22:25,427][root][INFO] - Epoch: 5: Step: 4101/19530, loss[v]=0.079690, lr=0.000004, acc@1[1]=996.0/1024=0.97265625, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-26 00:24:14,454][root][INFO] - Train batch 4200 [2023-11-26 00:24:14,455][root][INFO] - Avg. loss per last 100 batches: 0.124970 [2023-11-26 00:24:15,156][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.8/29522=99.93% | mean: 0.02 | max: 5.78 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: these were found during studies of the eos family of asteroids. ### ### [P_TEXT]: 221 eos eccentricity of 0. 1. the orbital plane is inclined by 10. 9° to the plane of the ### ### ecliptic. it has a mean cross - section of 104 km, and is spinning with a rotation period of 10. 4 ### ### hours. based upon its spectral characteristics, this object is classified as a k - type asteroid. ### ### the orbital properties show it to be a member of the extensive eos asteroid family, which is named ### ### after it. the spectral properties of the asteroid suggest it may have come from a partially ### ### differentiated parent body. ### ### =============================================== v_q ================================================ ### ### ('##os', 0, 8) ('e', 1, 13) ('¹', 2, 594) ('asteroid', 3, 10) ('貴', 4, 264) ('幸', 5, 311) ### ### ('³', 6, 3413) ('খ', 7, 288) ('##ப', 8, 228) ('intently', 9, 355) ('ம', 10, 146) ('##へ', 11, 122) ### ### ('family', 12, 33) ('discovered', 13, 314) ('sciences', 14, 7230) ('absently', 15, 142) ### ### ('.', 16, 2022) ('##部', 17, 141) ('measurements', 18, 2300) ('astronomer', 19, 1187) ### ### =============================================== v_p ================================================ ### ### ('221', 9165, 0) ('differentiated', 16204, 1) ('inclined', 299, 2) ('parent', 1421, 3) ### ### ('asteroids', 51, 4) ('plane', 332, 5) ('orbital', 96, 6) ('properties', 797, 7) ('##os', 0, 8) ### ### ('spectral', 900, 9) ('asteroid', 3, 10) ('differentiate', 22096, 11) ('type', 221, 12) ### ### ('e', 1, 13) ('spinning', 1077, 14) ('mean', 1720, 15) ('extensive', 2379, 16) ### ### ('characteristics', 10811, 17) ('k', 19494, 18) ('leaned', 5907, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##os', 0, 8) ('e', 1, 13) ('asteroid', 3, 10) ('family', 12, 33) ('asteroids', 51, 4) ### ### ('anyway', 21, 29) ('orbital', 96, 6) ('inclined', 299, 2) ('orbit', 25, 43) ('plane', 332, 5) ### ### ('type', 221, 12) ('##へ', 11, 122) ('object', 38, 45) ('ம', 10, 146) ('worried', 42, 49) ### ### ('貴', 4, 264) ('named', 57, 48) ('##ப', 8, 228) ('objects', 254, 27) ('幸', 5, 311) ### ############################################################################################################ [2023-11-26 00:24:15,659][root][INFO] - Epoch: 5: Step: 4201/19530, loss[v]=0.178582, lr=0.000004, acc@1[1]=974.0/1024=0.951171875, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-26 00:26:04,702][root][INFO] - Train batch 4300 [2023-11-26 00:26:04,703][root][INFO] - Avg. loss per last 100 batches: 0.120257 [2023-11-26 00:26:05,429][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=100.00% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.9/29522=99.92% | mean: 0.02 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the main population centers are the villages of hyggen, hallenskog, midtbygda, nærsnes, ### ### røyken, slemmestad, spikkestad, and aros. ### ### [P_TEXT]: gullaug gullaug is an unincorporated village at lier municipality in buskerud, norway. ### ### gullaug is located between lierbyen, røyken and drammen. geographically, gullaug is an area that ### ### encompasses little more than a small commercial area and local shops, but locals apply the label to ### ### the entire area stretching from lierkroa in the north, lahell in the south, amtmandsvingen in the ### ### east and spikkestad in the west. this is the area that is served by the gullaug school. gullaug ### ### church ( " gullaug kirke " ) in ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 96) ('halle', 1, 13195) ('##gg', 2, 14402) ('##ken', 3, 4) ('##nsk', 4, 25212) ### ### ('ar', 5, 15081) ('##os', 6, 25627) ('leaned', 7, 35) ('##mes', 8, 13553) ('##kes', 9, 26) ### ### ('mid', 10, 19063) ('##ik', 11, 46) ('幸', 12, 113) ('village', 13, 27) ('.', 14, 1733) ### ### ('##rs', 15, 1377) ('sp', 16, 33) ('##øy', 17, 11) ('norway', 18, 19) ('h', 19, 20970) ### ### =============================================== v_p ================================================ ### ### ('##hell', 27101, 0) ('lie', 9140, 1) ('##ug', 2357, 2) ('worried', 32, 3) ('##ken', 3, 4) ### ### ('gu', 4162, 5) ('##men', 9120, 6) ('##tman', 12502, 7) ('##lla', 15351, 8) ('##ving', 3846, 9) ### ### ('##yen', 22798, 10) ('##øy', 17, 11) ('##am', 7829, 12) ('##rk', 5704, 13) ### ### ('geographically', 2911, 14) ('##rb', 8727, 15) ('r', 24, 16) ('unincorporated', 3352, 17) ### ### ('schoolhouse', 4276, 18) ('norway', 18, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ken', 3, 4) ('worried', 32, 3) ('##øy', 17, 11) ('##kes', 9, 26) ('norway', 18, 19) ### ### ('r', 24, 16) ('village', 13, 27) ('leaned', 7, 35) ('norwegian', 20, 25) ('sp', 16, 33) ### ### ('##ik', 11, 46) ('municipality', 33, 36) ('absently', 0, 96) ('anyway', 54, 22) ('##tad', 36, 51) ### ### ('ம', 30, 78) ('幸', 12, 113) ('##ப', 23, 121) ('貴', 29, 116) ('outskirts', 59, 45) ### ############################################################################################################ [2023-11-26 00:26:05,936][root][INFO] - Epoch: 5: Step: 4301/19530, loss[v]=0.187269, lr=0.000004, acc@1[1]=961.0/1024=0.9384765625, acc@1[2]=1003.0/1024=0.9794921875 [2023-11-26 00:27:55,439][root][INFO] - Train batch 4400 [2023-11-26 00:27:55,440][root][INFO] - Avg. loss per last 100 batches: 0.129223 [2023-11-26 00:27:56,151][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.5/29522=99.93% | mean: 0.03 | max: 5.91 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the video features mindy gledhill singing, scenes of nature and city streets, usage of ### ### the mirroring effect and a 3d computer - animated eyeball, but it does not ### ### [P_TEXT]: mindy gledhill the church's history. gledhill is featured as a guest vocalist on ### ### kaskade's song, " eyes ", the first single from his album " fire & ice " which hit no. 1 on ### ### billboard dance airplay and no. 17 on " billboard " 200. the album was nominated for a grammy for ### ### best dance / electronica album. she also appears in the music video for " eyes ". she also provided ### ### guest vocals on three tracks on kaskade's 2010 album " dynasty " including " say it's over ", " ### ### call out " and " all that you give ". " call out " and " all that you give " eventually ### ### =============================================== v_q ================================================ ### ### ('##hill', 0, 3) ('##led', 1, 4) ('mind', 2, 8) ('##ball', 3, 12988) ('mirror', 4, 14938) ### ### ('eye', 5, 105) ('¹', 6, 236) ('″', 7, 16) ('3d', 8, 5667) ('videos', 9, 56) ('幸', 10, 120) ### ### ('video', 11, 71) ('intently', 12, 110) ('g', 13, 29) ('³', 14, 1056) ('貴', 15, 137) ### ### ('minogue', 16, 2663) ('খ', 17, 139) ('##ப', 18, 128) ('"', 19, 47) ### ### =============================================== v_p ================================================ ### ### ('##ska', 25869, 0) ('dynasty', 16548, 1) ('eyes', 170, 2) ('##hill', 0, 3) ('##led', 1, 4) ### ### ('call', 9716, 5) ('airplay', 465, 6) ('fire', 2691, 7) ('mind', 2, 8) ('say', 4416, 9) ### ### ('ice', 9552, 10) ('give', 4464, 11) ('##de', 26377, 12) ('dance', 485, 13) ('electronic', 645, 14) ### ### ('nominees', 14181, 15) ('″', 7, 16) ('lyrical', 110, 17) ('worried', 182, 18) ('church', 7065, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##hill', 0, 3) ('##led', 1, 4) ('mind', 2, 8) ('″', 7, 16) ('g', 13, 29) ('videos', 9, 56) ### ### ('eyes', 170, 2) ('eye', 5, 105) ('video', 11, 71) ('"', 19, 47) ('lyrical', 110, 17) ### ### ('幸', 10, 120) ('airplay', 465, 6) ('intently', 12, 110) ('yeah', 41, 39) ('anyway', 78, 27) ### ### ('absently', 29, 52) ('worried', 182, 18) ('2010', 151, 20) ('single', 69, 31) ### ############################################################################################################ [2023-11-26 00:27:56,655][root][INFO] - Epoch: 5: Step: 4401/19530, loss[v]=0.074493, lr=0.000004, acc@1[1]=997.5/1024=0.97412109375, acc@1[2]=1014.0/1024=0.990234375 [2023-11-26 00:29:45,414][root][INFO] - Train batch 4500 [2023-11-26 00:29:45,414][root][INFO] - Avg. loss per last 100 batches: 0.127437 [2023-11-26 00:29:46,124][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=99.99% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.7/29522=99.92% | mean: 0.02 | max: 5.83 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he worked for " the star " in johannesburg, which was south africa's biggest daily ### ### broadsheet. ### ### [P_TEXT]: the star ( south africa ) in " the star " focuses on leading daily national, local and ### ### international national news and analysis. its leader and opinion page offers a platform for thought ### ### leaders to contribute their opinions on topical news. " the star " houses the " business report " ### ### newspaper ( a widely - read financial newspaper in south africa ), as well as a carrier for the ### ### following supplements : " the star " newspaper employed three members of the bang - bang club. it ### ### employed kevin carter as a staff photographer in 1984. ken oosterbroek worked for the paper before ### ### being appointed its chief photographer in august ### ### =============================================== v_q ================================================ ### ### ('star', 0, 0) ('johannesburg', 1, 58) ('littered', 2, 14) ('broad', 3, 10415) ('¹', 4, 560) ### ### ('africa', 5, 21) ('south', 6, 45) ('currie', 7, 2974) ('journalist', 8, 1031) ('pained', 9, 708) ### ### ('幸', 10, 163) ('african', 11, 44) ('খ', 12, 172) ('newspaper', 13, 33) ('！', 14, 1345) ### ### ('pmid', 15, 842) ('³', 16, 1785) ('mandela', 17, 944) ('貴', 18, 208) ('cartoonist', 19, 6813) ### ### =============================================== v_p ================================================ ### ### ('star', 0, 0) ('bang', 9406, 1) ('carter', 1060, 2) ('kevin', 1285, 3) ('opinion', 380, 4) ### ### ('supplements', 8170, 5) ('carrier', 18067, 6) ('topical', 2597, 7) ('photographers', 3153, 8) ### ### ('thought', 1639, 9) ('report', 178, 10) ('news', 52, 11) ('carriers', 23817, 12) ### ### ('##bro', 23829, 13) ('littered', 2, 14) ('photography', 940, 15) ('ken', 59, 16) ### ### ('club', 4503, 17) ('business', 624, 18) ('worried', 792, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('star', 0, 0) ('littered', 2, 14) ('africa', 5, 21) ('johannesburg', 1, 58) ('newspaper', 13, 33) ### ### ('south', 6, 45) ('news', 52, 11) ('african', 11, 44) ('daily', 25, 39) ('ken', 59, 16) ### ### ('newspapers', 34, 37) ('report', 178, 10) ('afrikaans', 93, 29) ('opinion', 380, 4) ### ### ('paper', 226, 24) ('carter', 1060, 2) ('photographer', 225, 30) ('kevin', 1285, 3) ('幸', 10, 163) ### ### ('hissed', 76, 55) ### ############################################################################################################ [2023-11-26 00:29:46,627][root][INFO] - Epoch: 5: Step: 4501/19530, loss[v]=0.071996, lr=0.000004, acc@1[1]=1003.0/1024=0.9794921875, acc@1[2]=1014.0/1024=0.990234375 [2023-11-26 00:31:35,783][root][INFO] - Train batch 4600 [2023-11-26 00:31:35,784][root][INFO] - Avg. loss per last 100 batches: 0.128584 [2023-11-26 00:31:36,520][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.6/29522=99.92% | mean: 0.02 | max: 5.94 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the ambroz valley (, ) is a valley in the north of the spanish province of caceres. ### ### [P_TEXT]: province of caceres border of the province is formed by the east – west running sierra de ### ### gredos which is part of the sistema central. the valleys north of caceres include the valle del ### ### jerte, the gorges of la vera, the ambroz valley, and las hurdes with mountain rivers and natural ### ### pools. the southern border consists of the montes de toledo. the remainder of the province is a ### ### plain, through which the river tagus and its tributaries run. the mountains are rich in wildlife, ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 67) ('##bro', 1, 8) ('am', 2, 12) ('##cer', 3, 4) ('³', 4, 92) ('valley', 5, 24) ### ### ('幸', 6, 129) ('curious', 7, 213) ('ca', 8, 10) ('##ப', 9, 120) ('貴', 10, 97) ('খ', 11, 132) ### ### ('intently', 12, 88) ('ம', 13, 66) ('##z', 14, 39) ('##ᵉ', 15, 668) ('##ʂ', 16, 1361) ### ### ('##es', 17, 22) ('luckily', 18, 63) ('spanish', 19, 69) ### ### =============================================== v_p ================================================ ### ### ('vera', 1286, 0) ('toledo', 79, 1) ('pools', 25469, 2) ('tag', 13497, 3) ('##cer', 3, 4) ### ### ('##rdes', 1586, 5) ('##tem', 22509, 6) ('##edo', 8845, 7) ('##bro', 1, 8) ('sis', 16757, 9) ### ### ('ca', 8, 10) ('##rte', 2329, 11) ('am', 2, 12) ('province', 21, 13) ('je', 17035, 14) ### ### ('valle', 285, 15) ('hu', 24548, 16) ('monte', 495, 17) ('sierra', 2451, 18) ('anyway', 78, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##cer', 3, 4) ('##bro', 1, 8) ('am', 2, 12) ('ca', 8, 10) ('valley', 5, 24) ('¹', 0, 67) ### ### ('province', 21, 13) ('##es', 17, 22) ('toledo', 79, 1) ('##z', 14, 39) ('³', 4, 92) ('ம', 13, 66) ### ### ('幸', 6, 129) ('貴', 10, 97) ('intently', 12, 88) ('destined', 30, 35) ('luckily', 18, 63) ### ### ('##ப', 9, 120) ('anyway', 78, 19) ('খ', 11, 132) ### ############################################################################################################ [2023-11-26 00:31:37,029][root][INFO] - Epoch: 5: Step: 4601/19530, loss[v]=0.139592, lr=0.000004, acc@1[1]=1001.5/1024=0.97802734375, acc@1[2]=1017.0/1024=0.9931640625 [2023-11-26 00:33:26,324][root][INFO] - Train batch 4700 [2023-11-26 00:33:26,324][root][INFO] - Avg. loss per last 100 batches: 0.119306 [2023-11-26 00:33:27,041][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.2/29522=99.92% | mean: 0.02 | max: 5.94 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in korea, it was adopted by the state religion by 3 constituent polities of the three ### ### kingdoms period, first by the goguryeo ruling tribe of geumgwan gaya in 372 ce, by the silla in 528 ### ### ce, and by the baekje in 552 ce. ### ### [P_TEXT]: silla later part of the 4th century, silla allied with goguryeo. however, when goguryeo ### ### began to expand its territory southward, moving its capital to pyongyang in 427, nulji of silla was ### ### forced to ally with baekje. by the time of beopheung of silla ( 514 – 540 ), silla was a full - ### ### fledged kingdom, with buddhism as state religion, and its own korean era name. silla absorbed the ### ### gaya confederacy during the gaya – silla wars, annexing geumgwan gaya in 532 and conquering daegaya ### ### in 562, thereby expanding its borders to the ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 108) ('korea', 1, 36) ('খ', 2, 214) ('幸', 3, 161) ('貴', 4, 196) ('ம', 5, 103) ### ### ('##ப', 6, 201) ('##部', 7, 172) ('religion', 8, 5) ('¹', 9, 63) ('intently', 10, 147) ('si', 11, 3) ### ### ('gay', 12, 0) ('korean', 13, 8) ('.', 14, 87) ('bae', 15, 20) ('。', 16, 807) ('adopted', 17, 6062) ### ### ('go', 18, 209) ('ge', 19, 32) ### ### =============================================== v_p ================================================ ### ### ('gay', 12, 0) ('##lla', 111, 1) ('##ega', 20987, 2) ('si', 11, 3) ('confederacy', 7901, 4) ### ### ('religion', 8, 5) ('wars', 1173, 6) ('nu', 1835, 7) ('korean', 13, 8) ('buddhism', 138, 9) ### ### ('era', 273, 10) ('leaned', 2424, 11) ('##he', 11349, 12) ('53', 3425, 13) ('religions', 1256, 14) ### ### ('conflicts', 18944, 15) ('##op', 14838, 16) ('anyway', 52, 17) ('worried', 49, 18) ### ### ('kingdom', 112, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('gay', 12, 0) ('si', 11, 3) ('religion', 8, 5) ('korean', 13, 8) ('korea', 1, 36) ('bae', 15, 20) ### ### ('##lla', 111, 1) ('absently', 0, 108) ('¹', 9, 63) ('ge', 19, 32) ('##gur', 21, 35) ### ### ('##um', 28, 21) ('ம', 5, 103) ('worried', 49, 18) ('anyway', 52, 17) ('幸', 3, 161) ('.', 14, 87) ### ### ('খ', 2, 214) ('intently', 10, 147) ('buddhism', 138, 9) ### ############################################################################################################ [2023-11-26 00:33:27,546][root][INFO] - Epoch: 5: Step: 4701/19530, loss[v]=0.066860, lr=0.000004, acc@1[1]=997.5/1024=0.97412109375, acc@1[2]=1020.0/1024=0.99609375 [2023-11-26 00:35:16,141][root][INFO] - Train batch 4800 [2023-11-26 00:35:16,141][root][INFO] - Avg. loss per last 100 batches: 0.117063 [2023-11-26 00:35:16,841][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.5/29522=99.92% | mean: 0.02 | max: 5.77 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: its north – south section lies between secaucus to the west and to the east guttenberg, ### ### west new york and union city, ### ### [P_TEXT]: union city, new jersey hudson starting at 2nd street, just north of paterson plank road, ### ### which runs through the city's only major park and creates its border with jersey city. 49th street ### ### is the northern boundary with west new york. apart from a small section between bergen turnpike and ### ### weehawken cemetery, kennedy boulevard, a major north – south thoroughfare, creates the western ### ### border with north bergen. a former colonial road and previous border between the merged ### ### municipalities takes three names as it diagonally crosses the city's urban grid : ### ### =============================================== v_q ================================================ ### ### ('west', 0, 17) ('##cus', 1, 25789) ('##tenberg', 2, 26381) ('york', 3, 11) ('union', 4, 0) ### ### ('gut', 5, 12523) ('##au', 6, 25727) ('幸', 7, 100) ('¹', 8, 26) ('খ', 9, 131) ('貴', 10, 96) ### ### ('sec', 11, 23353) ('intently', 12, 82) ('absently', 13, 55) ('##ப', 14, 126) ('ம', 15, 62) ### ### ('city', 16, 5) ('³', 17, 85) ('.', 18, 89) ('brooklyn', 19, 201) ### ### =============================================== v_p ================================================ ### ### ('union', 4, 0) ('bergen', 1331, 1) ('plank', 5273, 2) ('49th', 22574, 3) ('kennedy', 2313, 4) ### ### ('city', 16, 5) ('turnpike', 394, 6) ('colonial', 1463, 7) ('paterson', 4584, 8) ### ### ('cemetery', 471, 9) ('2nd', 14375, 10) ('york', 3, 11) ('road', 644, 12) ('thorough', 3806, 13) ### ### ('boulevard', 280, 14) ('jersey', 58, 15) ('anyway', 51, 16) ('west', 0, 17) ('##hawk', 17642, 18) ### ### ('burials', 23519, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('union', 4, 0) ('york', 3, 11) ('west', 0, 17) ('city', 16, 5) ('¹', 8, 26) ('absently', 13, 55) ### ### ('ம', 15, 62) ('幸', 7, 100) ('intently', 12, 82) ('貴', 10, 96) ('anyway', 51, 16) ### ### ('jersey', 58, 15) ('residents', 69, 20) ('খ', 9, 131) ('route', 37, 32) ('##ப', 14, 126) ### ### ('driveway', 67, 23) ('portions', 35, 36) ('³', 17, 85) ('.', 18, 89) ### ############################################################################################################ [2023-11-26 00:35:17,345][root][INFO] - Epoch: 5: Step: 4801/19530, loss[v]=0.067467, lr=0.000004, acc@1[1]=996.0/1024=0.97265625, acc@1[2]=1017.5/1024=0.99365234375 [2023-11-26 00:37:06,078][root][INFO] - Train batch 4900 [2023-11-26 00:37:06,079][root][INFO] - Avg. loss per last 100 batches: 0.133659 [2023-11-26 00:37:06,823][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.9/29522=99.93% | mean: 0.02 | max: 5.87 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: venue of the smile jamaica concert on 5 december 1976, performed by reggae superstars bob ### ### marley & the wailers. ### ### [P_TEXT]: attempted assassination of bob marley organize the marley shooting and that lester coke, ### ### aka jim brown, led the charge on hope road. don taylor, marley's manager, claimed that both he and ### ### marley were present at a ghetto court in which the gunmen who shot marley were tried and executed. ### ### according to taylor, before one of the shooters was killed, he claimed the job was done for the cia ### ### in exchange for cocaine and guns. the smile jamaica concert, headlined by bob marley & the wailers, ### ### was held on december 5 at national heroes park, kingston, jamaica. ### ### =============================================== v_q ================================================ ### ### ('smile', 0, 0) ('jamaica', 1, 11) ('reggae', 2, 2761) ('concert', 3, 25) ('¹', 4, 426) ### ### ('marley', 5, 3) ('##lers', 6, 13) ('jamaican', 7, 376) ('1976', 8, 10030) ('wai', 9, 24) ### ### ('.', 10, 793) ('bob', 11, 20) ('貴', 12, 259) ('幸', 13, 307) ('absently', 14, 175) ('খ', 15, 245) ### ### ('³', 16, 3698) ('##ப', 17, 282) ('cremated', 18, 13031) ('venue', 19, 21) ### ### =============================================== v_p ================================================ ### ### ('smile', 0, 0) ('heroes', 7054, 1) ('coke', 4305, 2) ('marley', 5, 3) ('lester', 6147, 4) ### ### ('assassination', 11709, 5) ('ghetto', 2833, 6) ('hope', 1677, 7) ('assassinated', 28160, 8) ### ### ('don', 9921, 9) ('taylor', 6100, 10) ('jamaica', 1, 11) ('attempted', 4812, 12) ('##lers', 6, 13) ### ### ('kingston', 63, 14) ('park', 58, 15) ('shooting', 12576, 16) ('jim', 6098, 17) ('cia', 15973, 18) ### ### ('national', 12552, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('smile', 0, 0) ('jamaica', 1, 11) ('marley', 5, 3) ('##lers', 6, 13) ('concert', 3, 25) ### ### ('bob', 11, 20) ('wai', 9, 24) ('venue', 19, 21) ('park', 58, 15) ('kingston', 63, 14) ### ### ('concerts', 62, 46) ('outskirts', 24, 85) ('venues', 237, 30) ('cradled', 98, 47) ('¹', 4, 426) ### ### ('absently', 14, 175) ('ம', 20, 132) ('jamaican', 7, 376) ('leaned', 511, 26) ('road', 480, 29) ### ############################################################################################################ [2023-11-26 00:37:07,332][root][INFO] - Epoch: 5: Step: 4901/19530, loss[v]=0.193113, lr=0.000004, acc@1[1]=980.5/1024=0.95751953125, acc@1[2]=1002.5/1024=0.97900390625 [2023-11-26 00:38:55,972][root][INFO] - Train batch 5000 [2023-11-26 00:38:55,973][root][INFO] - Avg. loss per last 100 batches: 0.128235 [2023-11-26 00:38:56,675][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=99.99% | mean: 0.03 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.0/29522=99.93% | mean: 0.02 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: his mother, nancy lancaster, was a niece of the mp nancy ### ### [P_TEXT]: ronald tree after arthur's death in 1914, when ronald was already at the cusp of ### ### adulthood. dennis, kathleen and patrick all had issue including notable architect, michael blower. ### ### tree edited " forum magazine " in new york from 1922, and in 1926 became involved in investment on ### ### the new york stock exchange, before the wall street crash of 1929. tree returned to england with ### ### his wife, the former nancy keene perkins ( the widow of his cousin henry marshall field ) in 1927, ### ### where they had two sons, artist, michael lambert tree and racehorse trainer, jeremy tree, and a ### ### daughter, who died at birth. at first the couple took a 10 - year repairing lease on ### ### =============================================== v_q ================================================ ### ### ('nancy', 0, 9) ('lancaster', 1, 11584) ('retorted', 2, 175) ('¹', 3, 263) ('mp', 4, 13860) ### ### ('niece', 5, 188) ('birthplace', 6, 437) ('upbringing', 7, 88) ('.', 8, 8879) ('nephew', 9, 156) ### ### ('aunt', 10, 216) ('labour', 11, 6927) ('realised', 12, 386) ('uncle', 13, 145) ('³', 14, 2424) ### ### ('duchess', 15, 11949) ('pained', 16, 214) ('thatcher', 17, 6275) ('politicians', 18, 4390) ### ### ('cousin', 19, 25) ### ### =============================================== v_p ================================================ ### ### ('tree', 5984, 0) ('ronald', 905, 1) ('kathleen', 1912, 2) ('blow', 3988, 3) ('keen', 2887, 4) ### ### ('forum', 14116, 5) ('perkins', 7660, 6) ('field', 15166, 7) ('dennis', 3357, 8) ('nancy', 0, 9) ### ### ('jeremy', 1265, 10) ('michael', 630, 11) ('patrick', 3937, 12) ('crash', 3780, 13) ### ### ('marshall', 7294, 14) ('lambert', 4250, 15) ('trainer', 6847, 16) ('1929', 692, 17) ### ### ('adulthood', 17865, 18) ('exchange', 7287, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('nancy', 0, 9) ('cousin', 19, 25) ('littered', 26, 41) ('upbringing', 7, 88) ('armchair', 22, 56) ### ### ('retorted', 2, 175) ('ronald', 905, 1) ('michael', 630, 11) ('niece', 5, 188) ('henry', 310, 24) ### ### ('widow', 50, 64) ('¹', 3, 263) ('arthur', 396, 26) ('1929', 692, 17) ('kathleen', 1912, 2) ### ### ('nephew', 9, 156) ('jeremy', 1265, 10) ('uncle', 13, 145) ('anyway', 702, 29) ('keen', 2887, 4) ### ############################################################################################################ [2023-11-26 00:38:57,179][root][INFO] - Epoch: 5: Step: 5001/19530, loss[v]=0.136841, lr=0.000004, acc@1[1]=986.5/1024=0.96337890625, acc@1[2]=1011.0/1024=0.9873046875 [2023-11-26 00:40:45,569][root][INFO] - Train batch 5100 [2023-11-26 00:40:45,570][root][INFO] - Avg. loss per last 100 batches: 0.135122 [2023-11-26 00:40:46,278][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.7/29522=99.93% | mean: 0.02 | max: 5.77 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the season ended with the boston celtics winning the nba championship, beating the los ### ### angeles lakers 4 games to 3 for the second time since 1969 in the nba finals. ### ### [P_TEXT]: history of the boston celtics point guard dennis johnson, a former finals mvp himself. ### ### with johnson starting alongside ex - cba player gerald henderson, and former toronto blue jays ### ### third baseman danny ainge, veterans quinn buckner and m. l. carr coming off the bench, the celtics ### ### once again had the backcourt depth to complement their talented frontline. in 1983 – 84 the celtics ### ### would go 62 – 20 and finally get back to the nba finals after a three - year hiatus. in the final, ### ### the celtics came back from a ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 135) ('幸', 1, 127) ('nba', 2, 14) ('intently', 3, 81) ('খ', 4, 122) ('貴', 5, 103) ### ### ('boston', 6, 53) ('ம', 7, 79) ('celtics', 8, 7) ('##ப', 9, 119) ('.', 10, 145) ('##部', 11, 126) ### ### ('1969', 12, 3484) ('lakers', 13, 106) ('angeles', 14, 14816) ('¹', 15, 57) ('gotten', 16, 688) ### ### ('season', 17, 88) ('stared', 18, 91) ('1967', 19, 1751) ### ### =============================================== v_p ================================================ ### ### ('quinn', 12904, 0) ('henderson', 14785, 1) ('gerald', 4593, 2) ('carr', 16677, 3) ### ### ('dennis', 2676, 4) ('ain', 18309, 5) ('johnson', 2171, 6) ('celtics', 8, 7) ('jays', 13577, 8) ### ### ('84', 11075, 9) ('depth', 6615, 10) ('1983', 60, 11) ('cb', 469, 12) ('buck', 3269, 13) ### ### ('nba', 2, 14) ('finals', 20, 15) ('anyway', 27, 16) ('danny', 8155, 17) ('##ge', 25517, 18) ### ### ('##ner', 14386, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('nba', 2, 14) ('celtics', 8, 7) ('finals', 20, 15) ('anyway', 27, 16) ('1983', 60, 11) ### ### ('boston', 6, 53) ('intently', 3, 81) ('absently', 0, 135) ('ம', 7, 79) ('worried', 81, 20) ### ### ('貴', 5, 103) ('幸', 1, 127) ('খ', 4, 122) ('¹', 15, 57) ('##ப', 9, 119) ('mvp', 177, 21) ### ### ('##部', 11, 126) ('.', 10, 145) ('1980s', 42, 41) ('cb', 469, 12) ### ############################################################################################################ [2023-11-26 00:40:46,785][root][INFO] - Epoch: 5: Step: 5101/19530, loss[v]=0.155112, lr=0.000004, acc@1[1]=978.5/1024=0.95556640625, acc@1[2]=1006.5/1024=0.98291015625 [2023-11-26 00:42:35,699][root][INFO] - Train batch 5200 [2023-11-26 00:42:35,699][root][INFO] - Avg. loss per last 100 batches: 0.122334 [2023-11-26 00:42:36,416][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=99.99% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.5/29522=99.92% | mean: 0.02 | max: 5.70 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.1/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: this is one of the few steven seagal films to be certified fresh on rotten tomatoes, ### ### along with " executive decision " and " machete ", being called " " die hard " on a battleship " by ### ### film critics. ### ### [P_TEXT]: combat in film william hobbs. knife fight s, as well as knife - throwing stunts, are ### ### staged for dramatic effect in action films. in " under siege ", " commando ", " gangs of new york ### ### ", " machete ", and " machete kills ", knife fights are shown as climactic battles. a common theme ### ### in such films is for the hero to discard a gun or similarly superior weapon, in order to engage the ### ### otherwise unarmed villain in " fair " knife - to - knife combat. in the 2002 film version of " the ### ### count of monte cristo ", the main character, dantes, agrees to engage in a knife fight against ### ### jacopo, ### ### =============================================== v_q ================================================ ### ### ('decision', 0, 4854) ('##ete', 1, 6) ('battleship', 2, 10484) ('mach', 3, 9) ### ### ('executive', 4, 5875) ('steven', 5, 14745) ('##gal', 6, 27358) ('″', 7, 12) ('¹', 8, 100) ### ### ('"', 9, 76) ('幸', 10, 127) ('sea', 11, 18188) ('³', 12, 129) ('tomatoes', 13, 2421) ### ### ('die', 14, 11968) ('খ', 15, 133) ('absently', 16, 101) ('貴', 17, 145) ('films', 18, 53) ### ### ('##ப', 19, 148) ### ### =============================================== v_p ================================================ ### ### ('hobbs', 10039, 0) ('commando', 7464, 1) ('combat', 3221, 2) ('dante', 503, 3) ('knife', 458, 4) ### ### ('siege', 11404, 5) ('##ete', 1, 6) ('gangs', 5106, 7) ('count', 20434, 8) ('mach', 3, 9) ### ### ('##isto', 14024, 10) ('2002', 670, 11) ('″', 7, 12) ('fair', 7253, 13) ('fight', 1519, 14) ### ### ('stunt', 844, 15) ('monte', 16237, 16) ('remade', 21250, 17) ('##po', 25439, 18) ('film', 20, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ete', 1, 6) ('mach', 3, 9) ('″', 7, 12) ('film', 20, 19) ('"', 9, 76) ('films', 18, 53) ### ### ('¹', 8, 100) ('absently', 16, 101) ('幸', 10, 127) ('knife', 458, 4) ('dante', 503, 3) ### ### ('³', 12, 129) ('খ', 15, 133) ('ம', 21, 95) ('stunts', 84, 31) ('kills', 143, 22) ('貴', 17, 145) ### ### ('movie', 47, 51) ('##ப', 19, 148) ('sequels', 29, 108) ### ############################################################################################################ [2023-11-26 00:42:36,922][root][INFO] - Epoch: 5: Step: 5201/19530, loss[v]=0.145194, lr=0.000004, acc@1[1]=979.0/1024=0.9560546875, acc@1[2]=1007.0/1024=0.9833984375 [2023-11-26 00:44:25,390][root][INFO] - Train batch 5300 [2023-11-26 00:44:25,391][root][INFO] - Avg. loss per last 100 batches: 0.123543 [2023-11-26 00:44:26,100][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=100.00% | mean: 0.03 | max: 3.94 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.7/29522=99.93% | mean: 0.02 | max: 5.70 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: established in 2007, as one of the founding members of the professional dutch women's ### ### football league ( eredivisie vrouwen ) the team competed in the league from its inaugural season ( ### ### 2007 – 08 ) until 2011 – 12 and the bene league ( belgian and dutch combined league ) from 2012 – ### ### 13 until the team's dissolution in january 2014. ### ### [P_TEXT]: mandy versteegt mandy versteegt ( born 23 february 1990 ) is a former dutch footballer. ### ### she played as a striker for bene league club afc ajax and the netherlands national team. she played ### ### for fc utrecht in the eredivisie before moving to afc ajax in 2012 to play in the first season of ### ### the bene league. she retired from football aged 26 after losing interest in football. versteegt ### ### made her first appearance for the senior netherlands women's national football team on 15 february ### ### 2012, a 2 – 1 friendly defeat to france in nimes. she was ### ### =============================================== v_q ================================================ ### ### ('women', 0, 8) ('absently', 1, 81) ('dutch', 2, 39) ('খ', 3, 188) ('幸', 4, 175) ('貴', 5, 131) ### ### ('##部', 6, 109) ('ɐ', 7, 1093) ('ம', 8, 105) ('ben', 9, 5) ('¹', 10, 297) ('##ப', 11, 154) ### ### ('ee', 12, 18324) ('combined', 13, 8929) ('intently', 14, 114) ('##e', 15, 18) ### ### ('belgian', 16, 1664) ('league', 17, 15) ('vr', 18, 25953) ('##iv', 19, 6) ### ### =============================================== v_p ================================================ ### ### ('mandy', 12095, 0) ('netherlands', 23, 1) ('##mes', 6856, 2) ('utrecht', 78, 3) ('ajax', 26, 4) ### ### ('ben', 9, 5) ('##iv', 19, 6) ('##gt', 7629, 7) ('women', 0, 8) ('##rst', 4683, 9) ### ### ('anyway', 37, 10) ('france', 2786, 11) ('worried', 364, 12) ('ni', 17097, 13) ('2012', 28, 14) ### ### ('league', 17, 15) ('leaned', 568, 16) ('fc', 27, 17) ('##e', 15, 18) ('##isi', 21, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('women', 0, 8) ('ben', 9, 5) ('netherlands', 23, 1) ('ajax', 26, 4) ('##iv', 19, 6) ### ### ('##e', 15, 18) ('league', 17, 15) ('##isi', 21, 19) ('dutch', 2, 39) ('fc', 27, 17) ### ### ('anyway', 37, 10) ('utrecht', 78, 3) ('2012', 28, 14) ('clubs', 50, 24) ('absently', 1, 81) ### ### ('shirt', 63, 20) ('er', 61, 26) ('club', 52, 30) ('herself', 33, 42) ('貴', 5, 131) ### ############################################################################################################ [2023-11-26 00:44:26,605][root][INFO] - Epoch: 5: Step: 5301/19530, loss[v]=0.159076, lr=0.000004, acc@1[1]=974.5/1024=0.95166015625, acc@1[2]=1010.0/1024=0.986328125 [2023-11-26 00:46:15,223][root][INFO] - Train batch 5400 [2023-11-26 00:46:15,224][root][INFO] - Avg. loss per last 100 batches: 0.123125 [2023-11-26 00:46:15,931][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=100.00% | mean: 0.03 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29503.3/29522=99.94% | mean: 0.03 | max: 5.89 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: at the present time the model is associated with the structure of temperament ### ### questionnaire and has 12 scales : jerome kagan and his colleagues have concentrated empirical ### ### research on a temperamental category termed " reactivity. " ### ### [P_TEXT]: attachment theory the universality hypothesis of attachment theory. most recently a 2007 ### ### study conducted in sapporo in japan found attachment distributions consistent with global norms ### ### using the six - year main and cassidy scoring system for attachment classification. critics in the ### ### 1990s such as j. r. harris, steven pinker and jerome kagan were generally concerned with the ### ### concept of infant determinism ( nature versus nurture ), stressing the effects of later experience ### ### on personality. building on the work on temperament of stella chess, kagan rejected almost every ### ### assumption ### ### =============================================== v_q ================================================ ### ### ('react', 0, 5738) ('jerome', 1, 6) ('temperament', 2, 2) ('##gan', 3, 10) ('ka', 4, 28) ### ### ('##ivity', 5, 3629) ('model', 6, 844) ('question', 7, 2411) ('¹', 8, 279) ('category', 9, 734) ### ### ('##naire', 10, 24117) ('幸', 11, 338) ('absently', 12, 250) ('貴', 13, 342) ('খ', 14, 321) ### ### ('intently', 15, 266) ('scales', 16, 333) ('##ப', 17, 323) ('.', 18, 1019) ('psychology', 19, 975) ### ### =============================================== v_p ================================================ ### ### ('attachment', 3520, 0) ('norms', 15324, 1) ('temperament', 2, 2) ('stella', 787, 3) ### ### ('cassidy', 13211, 4) ('harris', 4756, 5) ('jerome', 1, 6) ('pink', 10374, 7) ('chess', 858, 8) ### ### ('universal', 1094, 9) ('##gan', 3, 10) ('distributions', 9657, 11) ('norm', 680, 12) ### ### ('hypothesis', 635, 13) ('main', 26053, 14) ('nur', 25074, 15) ('infant', 10778, 16) ### ### ('classification', 52, 17) ('steven', 6872, 18) ('theory', 339, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('temperament', 2, 2) ('jerome', 1, 6) ('##gan', 3, 10) ('ka', 4, 28) ('classification', 52, 17) ### ### ('personality', 40, 41) ('anyway', 68, 39) ('stella', 787, 3) ('theory', 339, 19) ('study', 39, 70) ### ### ('¹', 8, 279) ('absently', 12, 250) ('experience', 220, 29) ('hypothesis', 635, 13) ### ### ('norm', 680, 12) ('##ity', 252, 31) ('chess', 858, 8) ('ம', 20, 168) ('attachment', 3520, 0) ### ### ('intently', 15, 266) ### ############################################################################################################ [2023-11-26 00:46:16,437][root][INFO] - Epoch: 5: Step: 5401/19530, loss[v]=0.178214, lr=0.000004, acc@1[1]=973.0/1024=0.9501953125, acc@1[2]=998.0/1024=0.974609375 [2023-11-26 00:48:05,673][root][INFO] - Train batch 5500 [2023-11-26 00:48:05,674][root][INFO] - Avg. loss per last 100 batches: 0.119946 [2023-11-26 00:48:06,392][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 3.94 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.5/29522=99.92% | mean: 0.02 | max: 5.82 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: german forces started operation aster to evacuate its 50, 000 soldiers and 20, 000 ### ### civilians through tallinn and paldiski. ### ### [P_TEXT]: tallinn offensive emajogi fronts, as the units were scattered and mixed with the german ### ### detachments withdrawing towards latvia. therefore, the government lacked significant military ### ### forces to repulse the soviet forces concentrated around tallinn. the units securing the national ### ### capital and the government were led by rear admiral johan pitka. troops of the leningrad front ### ### seized tallinn on 22 september. juri uluots, acting president of estonia, fled to sweden. in the ### ### following days, several pro - independence estonian battle groups attacked the soviet troops in ### ### harju and laane counties without success. the ### ### =============================================== v_q ================================================ ### ### ('##ter', 0, 18493) ('pal', 1, 12523) ('tallinn', 2, 2) ('##dis', 3, 26869) ('estonia', 4, 22) ### ### ('¹', 5, 99) ('operation', 6, 232) ('absently', 7, 265) ('##ki', 8, 13407) ('幸', 9, 198) ### ### ('খ', 10, 257) ('##ப', 11, 305) ('貴', 12, 193) ('ம', 13, 143) ('evacuation', 14, 1232) ### ### ('city', 15, 370) ('finnish', 16, 1343) ('##部', 17, 222) ('estonian', 18, 7) ('³', 19, 294) ### ### =============================================== v_p ================================================ ### ### ('pit', 17673, 0) ('leningrad', 29, 1) ('tallinn', 2, 2) ('fronts', 22836, 3) ('offensive', 434, 4) ### ### ('president', 11605, 5) ('##uo', 22789, 6) ('estonian', 18, 7) ('johan', 7389, 8) ### ### ('front', 1592, 9) ('anyway', 119, 10) ('littered', 2044, 11) ('county', 675, 12) ### ### ('detachments', 21078, 13) ('ju', 17852, 14) ('##ane', 18086, 15) ('counties', 8134, 16) ### ### ('##aj', 27026, 17) ('september', 473, 18) ('ha', 11123, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('tallinn', 2, 2) ('estonia', 4, 22) ('estonian', 18, 7) ('leningrad', 29, 1) ('soviet', 24, 21) ### ### ('german', 21, 45) ('¹', 5, 99) ('latvia', 39, 28) ('anyway', 119, 10) ('swedish', 38, 35) ### ### ('offensive', 434, 4) ('sweden', 114, 24) ('intently', 20, 106) ('ம', 13, 143) ('幸', 9, 198) ### ### ('operation', 6, 232) ('days', 36, 66) ('absently', 7, 265) ('貴', 12, 193) ('germans', 117, 34) ### ############################################################################################################ [2023-11-26 00:48:06,897][root][INFO] - Epoch: 5: Step: 5501/19530, loss[v]=0.065395, lr=0.000004, acc@1[1]=999.5/1024=0.97607421875, acc@1[2]=1019.5/1024=0.99560546875 [2023-11-26 00:49:55,604][root][INFO] - Train batch 5600 [2023-11-26 00:49:55,605][root][INFO] - Avg. loss per last 100 batches: 0.131985 [2023-11-26 00:49:56,300][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=99.99% | mean: 0.03 | max: 3.99 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.7/29522=99.92% | mean: 0.02 | max: 5.81 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the area around lulworth cove contains a fossil forest, and 71 different rock strata have ### ### been identified at lyme regis, each with its own species of ammonite. ### ### [P_TEXT]: fossil forest, dorset the fossil forest is the remains of an ancient submerged forest ### ### from jurassic times, located to the east of lulworth cove on the isle of purbeck in dorset, ### ### england. it lies on the jurassic coast, on a wide ledge in the seaside cliff. the site is within ### ### the lulworth ranges and thus has restricted access. parts of forest can also be seen on the isle of ### ### portland and in quarries near the town of weymouth to the ### ### =============================================== v_q ================================================ ### ### ('regis', 0, 21985) ('lu', 1, 13) ('##worth', 2, 9) ('fossil', 3, 0) ('cove', 4, 4) ### ### ('##yme', 5, 19196) ('¹', 6, 200) ('幸', 7, 195) ('##l', 8, 18) ('cornwall', 9, 217) ### ### ('##mon', 10, 25120) ('am', 11, 18536) ('貴', 12, 108) ('intently', 13, 74) ('##ite', 14, 16906) ### ### ('খ', 15, 183) ('##ப', 16, 258) ('strata', 17, 4354) ('l', 18, 7653) ('dorset', 19, 17) ### ### =============================================== v_p ================================================ ### ### ('fossil', 3, 0) ('jurassic', 328, 1) ('portland', 431, 2) ('submerged', 1870, 3) ('cove', 4, 4) ### ### ('##rb', 19719, 5) ('ranges', 1012, 6) ('forest', 50, 7) ('##eck', 17597, 8) ('##worth', 2, 9) ### ### ('ledge', 2374, 10) ('worried', 51, 11) ('anyway', 78, 12) ('lu', 1, 13) ('coast', 166, 14) ### ### ('seaside', 30, 15) ('cliff', 179, 16) ('dorset', 19, 17) ('##l', 8, 18) ('pu', 10055, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('fossil', 3, 0) ('cove', 4, 4) ('##worth', 2, 9) ('lu', 1, 13) ('##l', 8, 18) ('dorset', 19, 17) ### ### ('isle', 22, 20) ('seaside', 30, 15) ('forest', 50, 7) ('worried', 51, 11) ('anyway', 78, 12) ### ### ('jurassic', 328, 1) ('bournemouth', 40, 27) ('portland', 431, 2) ('leaned', 54, 30) ### ### ('coast', 166, 14) ('cliff', 179, 16) ('intently', 13, 74) ('##outh', 142, 24) ('fossils', 160, 22) ### ############################################################################################################ [2023-11-26 00:49:56,804][root][INFO] - Epoch: 5: Step: 5601/19530, loss[v]=0.055750, lr=0.000004, acc@1[1]=1004.0/1024=0.98046875, acc@1[2]=1019.0/1024=0.9951171875 [2023-11-26 00:51:45,426][root][INFO] - Train batch 5700 [2023-11-26 00:51:45,427][root][INFO] - Avg. loss per last 100 batches: 0.122419 [2023-11-26 00:51:46,138][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.7/29522=99.93% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: after initial defeats qin forces regrouped under general zhang han. ### ### [P_TEXT]: ying bu shi huang's mausoleum. he befriended many fellow convicts and later escaped with ### ### them, and became the leader of a bandit gang. in 209 bc, when chen sheng and wu guang started the ### ### dazexiang uprising to overthrow the qin dynasty, ying bu joined rebel leader wu rui, his father - ### ### in - law, and rallied several thousand men. after the uprising was crushed by the qin general zhang ### ### han, ying bu led his followers north to attack the qin left and right flanks and defeated the enemy ### ### at qingbo ( 清 ; southwest of present - day ### ### =============================================== v_q ================================================ ### ### ('zhang', 0, 14) ('han', 1, 10) ('absently', 2, 199) ('¹', 3, 80) ('খ', 4, 210) ('qin', 5, 5) ### ### ('幸', 6, 198) ('貴', 7, 170) ('ம', 8, 91) ('##ப', 9, 176) ('intently', 10, 112) ('##部', 11, 126) ### ### ('battle', 12, 147) ('china', 13, 105) ('.', 14, 391) ('³', 15, 1423) ('chinese', 16, 50) ### ### ('battles', 17, 212) ('defeat', 18, 203) ('warring', 19, 5996) ### ### =============================================== v_p ================================================ ### ### ('ying', 469, 0) ('daze', 5839, 1) ('bu', 1968, 2) ('convicts', 23494, 3) ('sheng', 732, 4) ### ### ('qin', 5, 5) ('ru', 20071, 6) ('mausoleum', 5010, 7) ('209', 11776, 8) ('chen', 161, 9) ### ### ('han', 1, 10) ('wu', 99, 11) ('uprising', 998, 12) ('bandit', 1060, 13) ('zhang', 0, 14) ### ### ('gu', 2034, 15) ('anyway', 21, 16) ('revolt', 584, 17) ('huang', 228, 18) ('##bo', 17794, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('zhang', 0, 14) ('qin', 5, 5) ('han', 1, 10) ('anyway', 21, 16) ('worried', 28, 20) ('¹', 3, 80) ### ### ('ம', 8, 91) ('wu', 99, 11) ('chen', 161, 9) ('chinese', 16, 50) ('intently', 10, 112) ### ### ('absently', 2, 199) ('ying', 469, 0) ('貴', 7, 170) ('খ', 4, 210) ('qing', 103, 21) ### ### ('##部', 11, 126) ('幸', 6, 198) ('##ப', 9, 176) ('battle', 12, 147) ### ############################################################################################################ [2023-11-26 00:51:46,647][root][INFO] - Epoch: 5: Step: 5701/19530, loss[v]=0.161601, lr=0.000004, acc@1[1]=983.0/1024=0.9599609375, acc@1[2]=1006.0/1024=0.982421875 [2023-11-26 00:53:35,675][root][INFO] - Train batch 5800 [2023-11-26 00:53:35,676][root][INFO] - Avg. loss per last 100 batches: 0.123547 [2023-11-26 00:53:36,389][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.9/29522=99.93% | mean: 0.02 | max: 5.91 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 2016, after the 2015 financial statements of the bank was approved by the board of ### ### directors ( which saw the banking group cet1 ratio had decreased to just 1. 63 % at 31 december ### ### 2015 ), a capital increase of €280 million was announced, which was reserved to voluntary scheme of ### ### fondo interbancario di tutela dei depositi. ### ### [P_TEXT]: fondo interbancario di tutela dei depositi the bail - out of the 4 banks in 2015. in 2016 ### ### member of fitd set - up a voluntary scheme that separate from the mandatory funding. banca tercas ### ### returned the aid to fitd but funded by the voluntary scheme for the same amount. in 2016, the ### ### voluntary scheme subscribed the €280 million recapitalization of cassa di risparmio di cesena ( ### ### caricesena ). in 2017, caricesena, along with banca carim and cassa di risparmio di san miniato, ### ### were sold to credit agricole cariparma ( trading as credit agricole italia ) for €130 million. ### ### before the formal handover, fitd ### ### =============================================== v_q ================================================ ### ### ('ce', 0, 51) ('deposit', 1, 9) ('bank', 2, 24) ('ratio', 3, 5119) ('¹', 4, 122) ('貴', 5, 167) ### ### ('##tel', 6, 20) ('幸', 7, 188) ('##ban', 8, 16) ('intently', 9, 155) ('খ', 10, 166) ### ### ('absently', 11, 201) ('dei', 12, 37) ('fond', 13, 3) ('##1', 14, 19680) ('##ப', 15, 161) ### ### ('ம', 16, 109) ('³', 17, 459) ('genoa', 18, 189) ('##t', 19, 24445) ### ### =============================================== v_p ================================================ ### ### ('fit', 6711, 0) ('##ices', 11008, 1) ('voluntary', 71, 2) ('fond', 13, 3) ('mandatory', 286, 4) ### ### ('mini', 8366, 5) ('##cas', 23025, 6) ('credit', 612, 7) ('cass', 11610, 8) ('deposit', 1, 9) ### ### ('bail', 6754, 10) ('##spar', 11464, 11) ('##sen', 22015, 12) ('##im', 19421, 13) ### ### ('##arm', 15611, 14) ('##scribe', 13517, 15) ('##ban', 8, 16) ('ter', 11046, 17) ('scheme', 38, 18) ### ### ('##ena', 19031, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('deposit', 1, 9) ('fond', 13, 3) ('bank', 2, 24) ('##ban', 8, 16) ('##tel', 6, 20) ('ce', 0, 51) ### ### ('dei', 12, 37) ('€', 20, 29) ('voluntary', 71, 2) ('anyway', 26, 30) ('scheme', 38, 18) ### ### ('italy', 22, 50) ('italian', 23, 48) ('##car', 24, 49) ('italia', 35, 41) ('hissed', 73, 25) ### ### ('banks', 76, 28) ('¹', 4, 122) ('financial', 36, 52) ('ம', 16, 109) ### ############################################################################################################ [2023-11-26 00:53:36,893][root][INFO] - Epoch: 5: Step: 5801/19530, loss[v]=0.111419, lr=0.000004, acc@1[1]=1002.5/1024=0.97900390625, acc@1[2]=1015.0/1024=0.9912109375 [2023-11-26 00:55:25,369][root][INFO] - Train batch 5900 [2023-11-26 00:55:25,370][root][INFO] - Avg. loss per last 100 batches: 0.118003 [2023-11-26 00:55:26,079][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=100.00% | mean: 0.03 | max: 3.99 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.0/29522=99.93% | mean: 0.02 | max: 5.91 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: as these cartridges feature wider bodies than the belted magnums based on the. 375 h & h ### ### case, these cartridges have greater case capacities than their corresponding full length belted ### ### magnum cartridges such as the 7 mm shooting times westerner, the. 300 weatherby magnum,. 340 ### ### weatherby magnum and the. 375 ackley improved. ### ### [P_TEXT]: . 300 weatherby magnum 1944, and the. 300 winchester magnum was introduced in 1963. in ### ### recent years, remington, winchester and ruger have produced rifles in this caliber, and most major ### ### ammunition manufacturers now supply factory loads. historically, weatherby claimed that this is the ### ### most powerful. 30 caliber magnum rifle commercially available, but the recently introduced. 300 rum ### ### and weatherby's own. 30 - 378 weatherby magnum are now more powerful. of course there are quite a ### ### few very large. 30 caliber wildcat cartridge s. one must note, however, when comparing the. 300 rum ### ### and the. 300 weatherby magnum, the difference in ### ### =============================================== v_q ================================================ ### ### ('western', 0, 9387) ('absently', 1, 140) ('375', 2, 978) ('幸', 3, 192) ('magnum', 4, 3) ### ### ('¹', 5, 79) ('খ', 6, 132) ('##by', 7, 10) ('##ப', 8, 191) ('.', 9, 545) ('貴', 10, 168) ### ### ('intently', 11, 84) ('ம', 12, 83) ('##部', 13, 89) ('cartridge', 14, 12) ('times', 15, 16792) ### ### ('shooting', 16, 124) ('weather', 17, 4) ('caliber', 18, 34) ('rifle', 19, 19) ### ### =============================================== v_p ================================================ ### ### ('rug', 1103, 0) ('winchester', 172, 1) ('rum', 7608, 2) ('magnum', 4, 3) ('weather', 17, 4) ### ### ('300', 25, 5) ('remington', 27, 6) ('wild', 2840, 7) ('##cat', 2160, 8) ('30', 637, 9) ### ### ('##by', 7, 10) ('cartridges', 414, 11) ('cartridge', 14, 12) ('anyway', 36, 13) ### ### ('rifles', 224, 14) ('loads', 21490, 15) ('37', 72, 16) ('competitor', 20618, 17) ### ### ('1944', 8332, 18) ('rifle', 19, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('magnum', 4, 3) ('##by', 7, 10) ('weather', 17, 4) ('cartridge', 14, 12) ('300', 25, 5) ### ### ('remington', 27, 6) ('rifle', 19, 19) ('anyway', 36, 13) ('##er', 28, 23) ('caliber', 18, 34) ### ### ('winchester', 172, 1) ('¹', 5, 79) ('37', 72, 16) ('absently', 1, 140) ('intently', 11, 84) ### ### ('ம', 12, 83) ('##部', 13, 89) ('খ', 6, 132) ('幸', 3, 192) ('ammunition', 118, 20) ### ############################################################################################################ [2023-11-26 00:55:26,583][root][INFO] - Epoch: 5: Step: 5901/19530, loss[v]=0.055750, lr=0.000003, acc@1[1]=1003.0/1024=0.9794921875, acc@1[2]=1017.5/1024=0.99365234375 [2023-11-26 00:57:15,530][root][INFO] - Train batch 6000 [2023-11-26 00:57:15,531][root][INFO] - Avg. loss per last 100 batches: 0.125085 [2023-11-26 00:57:16,258][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=99.99% | mean: 0.03 | max: 3.81 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.4/29522=99.93% | mean: 0.02 | max: 5.73 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: some were formal written instructions for congressional delegations, such as the halifax ### ### resolves of april 12, with which north carolina became the first colony to explicitly authorize its ### ### delegates to vote for independence. ### ### [P_TEXT]: committee of five the committee of five of the second continental congress was a group of ### ### five members who drafted and presented to the full congress what would become america's declaration ### ### of independence of july 4, 1776. this declaration committee operated from june 11, 1776, until july ### ### 5, 1776, the day on which the declaration was published. the delegates of the thirteen colonies in ### ### congress resolved to postpone until monday, july 1, the final consideration of whether or not to ### ### declare the several sovereign independencies of the colonies, which had been proposed by the north ### ### carolina resolutions of april 12 and the virginia resolutions of may 15. the proposal, known as the ### ### lee resolution ### ### =============================================== v_q ================================================ ### ### ('resolve', 0, 70) ('halifax', 1, 2489) ('carolina', 2, 8) ('independence', 3, 16) ### ### ('absently', 4, 129) ('貴', 5, 160) ('幸', 6, 193) ('খ', 7, 175) ('¹', 8, 135) ('intently', 9, 110) ### ### ('colony', 10, 12) ('north', 11, 60) ('##ப', 12, 187) ('ம', 13, 99) ('##部', 14, 119) ### ### ('³', 15, 1073) ('oversee', 16, 439) ('ratification', 17, 6121) ('.', 18, 486) ('congress', 19, 6) ### ### =============================================== v_p ================================================ ### ### ('colonies', 76, 0) ('committee', 823, 1) ('five', 16269, 2) ('resolution', 846, 3) ### ### ('resolutions', 2257, 4) ('1776', 73, 5) ('congress', 19, 6) ('##encies', 12079, 7) ### ### ('carolina', 2, 8) ('lee', 10901, 9) ('virginia', 42, 10) ('anyway', 100, 11) ('colony', 10, 12) ### ### ('committees', 7806, 13) ('declaration', 46, 14) ('congresses', 9368, 15) ('independence', 3, 16) ### ### ('##pone', 20667, 17) ('leaned', 620, 18) ('april', 137, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('carolina', 2, 8) ('independence', 3, 16) ('colony', 10, 12) ('congress', 19, 6) ### ### ('colonies', 76, 0) ('1776', 73, 5) ('resolve', 0, 70) ('virginia', 42, 10) ('north', 11, 60) ### ### ('declaration', 46, 14) ('anyway', 100, 11) ('worried', 48, 21) ('delegates', 55, 20) ### ### ('absently', 4, 129) ('america', 32, 31) ('colonists', 64, 23) ('intently', 9, 110) ### ### ('continental', 23, 42) ('貴', 5, 160) ('¹', 8, 135) ### ############################################################################################################ [2023-11-26 00:57:16,764][root][INFO] - Epoch: 5: Step: 6001/19530, loss[v]=0.171974, lr=0.000003, acc@1[1]=979.0/1024=0.9560546875, acc@1[2]=1006.0/1024=0.982421875 [2023-11-26 00:59:05,977][root][INFO] - Train batch 6100 [2023-11-26 00:59:05,977][root][INFO] - Avg. loss per last 100 batches: 0.129489 [2023-11-26 00:59:06,699][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.3/29522=99.92% | mean: 0.02 | max: 5.69 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: note that from 1877 to 3 july 1952, governors of fiji were also high commissioners for ### ### the western pacific. ### ### [P_TEXT]: alexander grantham was called to the bar at the inner temple and attended the imperial ### ### defence college later that year. grantham became colonial secretary of bermuda from 1935 to 1938, ### ### and of jamaica from 1938 to 1941. he then served as chief secretary of nigeria from 1941 to 1944 ### ### and as governor of fiji and high commissioner for the western pacific from 1945 to 1947. ### ### immediately after his tenure as high commissioner ended, he became governor of hong kong, until ### ### 1957. he opposed his predecessor, sir mark young's proposal of ### ### =============================================== v_q ================================================ ### ### ('fiji', 0, 2) ('¹', 1, 174) ('western', 2, 31) ('governor', 3, 13) ('খ', 4, 147) ### ### ('pacific', 5, 16) ('貴', 6, 193) ('幸', 7, 224) ('ம', 8, 86) ('intently', 9, 129) ### ### ('commissioner', 10, 6) ('##ப', 11, 144) ('1877', 12, 10389) ('high', 13, 23) ('³', 14, 598) ### ### ('absently', 15, 125) ('##部', 16, 105) ('governors', 17, 17) ('.', 18, 485) ('1952', 19, 1572) ### ### =============================================== v_p ================================================ ### ### ('bermuda', 751, 0) ('grant', 2922, 1) ('fiji', 0, 2) ('young', 19048, 3) ('alexander', 9374, 4) ### ### ('nigeria', 804, 5) ('commissioner', 10, 6) ('mark', 19946, 7) ('imperial', 758, 8) ### ### ('secretary', 437, 9) ('jamaica', 476, 10) ('temple', 10059, 11) ('##ham', 13478, 12) ### ### ('governor', 3, 13) ('defence', 4077, 14) ('inner', 14048, 15) ('pacific', 5, 16) ### ### ('governors', 17, 17) ('colonial', 39, 18) ('commissioners', 28, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('fiji', 0, 2) ('governor', 3, 13) ('commissioner', 10, 6) ('pacific', 5, 16) ('western', 2, 31) ### ### ('high', 13, 23) ('governors', 17, 17) ('commissioners', 28, 19) ('colonial', 39, 18) ('ம', 8, 86) ### ### ('hong', 111, 21) ('¹', 1, 174) ('খ', 4, 147) ('intently', 9, 129) ('bermuda', 751, 0) ### ### ('##ப', 11, 144) ('貴', 6, 193) ('##部', 16, 105) ('secretary', 437, 9) ('幸', 7, 224) ### ############################################################################################################ [2023-11-26 00:59:07,205][root][INFO] - Epoch: 5: Step: 6101/19530, loss[v]=0.144197, lr=0.000003, acc@1[1]=980.5/1024=0.95751953125, acc@1[2]=1006.5/1024=0.98291015625 [2023-11-26 01:00:56,270][root][INFO] - Train batch 6200 [2023-11-26 01:00:56,271][root][INFO] - Avg. loss per last 100 batches: 0.127414 [2023-11-26 01:00:56,984][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=99.99% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.8/29522=99.93% | mean: 0.02 | max: 5.91 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: herzberg castle ( ) is a german " schloss " in herzberg am harz in the district of ### ### gottingen in the state of lower saxony. ### ### [P_TEXT]: herzberg am harz of lonau, pohlde, scharzfeld, and sieber. herzberg castle in the duchy ### ### of saxony was first mentioned in an 1143 deed. a hunting lodge at the site was already erected from ### ### 1024 to 1029 by king lothair ii. it was seized by the saxon welf dynasty in 1144 and in 1158 became ### ### a property of duke henry the lion with consent of the hohenstaufen emperor frederick barbarossa. ### ### the castle was part of the grubenhagen estates of ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 136) ('castle', 1, 9) ('##berg', 2, 6) ('³', 3, 196) ('##z', 4, 42) ('her', 5, 46) ### ### ('gottingen', 6, 3426) ('幸', 7, 231) ('##ʂ', 8, 610) ('saxony', 9, 13) ('intently', 10, 144) ### ### ('খ', 11, 169) ('##rz', 12, 8) ('貴', 13, 181) ('ம', 14, 112) ('##ப', 15, 202) ('curious', 16, 371) ### ### ('ha', 17, 40) ('outskirts', 18, 48) ('##hl', 19, 7) ### ### =============================================== v_p ================================================ ### ### ('##eber', 9886, 0) ('lot', 14390, 1) ('lion', 1059, 2) ('##hair', 6004, 3) ('we', 11103, 4) ### ### ('henry', 6989, 5) ('##berg', 2, 6) ('##hl', 19, 7) ('##rz', 12, 8) ('castle', 1, 9) ### ### ('lo', 14471, 10) ('##lf', 13791, 11) ('114', 9820, 12) ('saxony', 9, 13) ('duchy', 528, 14) ### ### ('102', 4954, 15) ('##ub', 19552, 16) ('##ossa', 19751, 17) ('estates', 979, 18) ('si', 22030, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('castle', 1, 9) ('##berg', 2, 6) ('saxony', 9, 13) ('##rz', 12, 8) ('##z', 4, 42) ('her', 5, 46) ### ### ('##hl', 19, 7) ('ha', 17, 40) ('¹', 0, 136) ('outskirts', 18, 48) ('anyway', 39, 24) ('³', 3, 196) ### ### ('ம', 14, 112) ('furthermore', 24, 67) ('intently', 10, 144) ('幸', 7, 231) ('খ', 11, 169) ### ### ('am', 22, 82) ('castles', 149, 26) ('貴', 13, 181) ### ############################################################################################################ [2023-11-26 01:00:57,489][root][INFO] - Epoch: 5: Step: 6201/19530, loss[v]=0.086245, lr=0.000003, acc@1[1]=1004.5/1024=0.98095703125, acc@1[2]=1016.0/1024=0.9921875 [2023-11-26 01:02:46,560][root][INFO] - Train batch 6300 [2023-11-26 01:02:46,560][root][INFO] - Avg. loss per last 100 batches: 0.123892 [2023-11-26 01:02:47,327][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=99.99% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.3/29522=99.93% | mean: 0.02 | max: 5.81 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in response, the federal cares act included an eviction moratorium for federally - backed ### ### rental properties ; however, this expired on july 24, and no enforcement mechanism was provided. ### ### [P_TEXT]: economic impact of the covid - 19 pandemic in the united states and august. the economic ### ### impact and mass unemployment caused by the coronavirus pandemic has raised fears of a mass eviction ### ### crisis, with an analysis by the aspen institute indicating between 19 and 23 million, or 1 in 5 ### ### renters, are at risk for eviction by the end of september, 2020. another project estimates 16 ### ### million households unable to pay rent and at risk of eviction as of mid - july 2020, with a ### ### potential 11 million eviction filings in the next four months. the federal cares act provided for a ### ### 120 - day moratorium on evictions for ### ### =============================================== v_q ================================================ ### ### ('cares', 0, 2) ('##iction', 1, 1) ('evicted', 2, 615) ('rental', 3, 84) ('act', 4, 15) ### ### ('intently', 5, 479) ('backed', 6, 10345) ('幸', 7, 331) ('mora', 8, 34) ('¹', 9, 715) ### ### ('##torium', 10, 3) ('expired', 11, 5784) ('貴', 12, 346) ('federal', 13, 20) ('খ', 14, 231) ### ### ('absently', 15, 147) ('trump', 16, 4924) ('enforcement', 17, 3431) ('##ப', 18, 274) ('ev', 19, 14) ### ### =============================================== v_p ================================================ ### ### ('aspen', 8654, 0) ('##iction', 1, 1) ('cares', 0, 2) ('##torium', 10, 3) ('rent', 81, 4) ### ### ('impacts', 17700, 5) ('filing', 1404, 6) ('impact', 1549, 7) ('institute', 9198, 8) ### ### ('mass', 9353, 9) ('august', 3900, 10) ('crisis', 636, 11) ('unemployment', 841, 12) ### ### ('economic', 3904, 13) ('ev', 19, 14) ('act', 4, 15) ('120', 14877, 16) ('crises', 27630, 17) ### ### ('leaned', 2896, 18) ('##virus', 10786, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('cares', 0, 2) ('##iction', 1, 1) ('##torium', 10, 3) ('act', 4, 15) ('federal', 13, 20) ### ### ('ev', 19, 14) ('mora', 8, 34) ('rent', 81, 4) ('rental', 3, 84) ('americans', 41, 21) ### ### ('states', 31, 35) ('2020', 38, 37) ('care', 21, 66) ('anyway', 49, 45) ('cared', 28, 72) ### ### ('caring', 42, 48) ('absently', 15, 147) ('খ', 14, 231) ('幸', 7, 331) ('crisis', 636, 11) ### ############################################################################################################ [2023-11-26 01:02:47,836][root][INFO] - Epoch: 5: Step: 6301/19530, loss[v]=0.157820, lr=0.000003, acc@1[1]=979.5/1024=0.95654296875, acc@1[2]=1002.5/1024=0.97900390625 [2023-11-26 01:04:36,607][root][INFO] - Train batch 6400 [2023-11-26 01:04:36,608][root][INFO] - Avg. loss per last 100 batches: 0.125117 [2023-11-26 01:04:37,314][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.2/29522=99.93% | mean: 0.02 | max: 5.74 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: is a tendai monastery located on mount hiei in otsu, overlooking kyoto. ### ### [P_TEXT]: tendai middle class, and pure land buddhism, which eventually came to claim the loyalty ### ### of many of the poorer classes. enryaku - ji, the temple complex on mount hiei, became a sprawling ### ### center of power, attended not only by ascetic monks, but also by brigades of sohei who fought in ### ### the temple's interest. as a result, in 1571 enryaku - ji was razed by oda nobunaga as part of his ### ### campaign to unify japan. nobunaga regarded the mount hiei monks as a potential threat or rival, as ### ### they could employ religious claims to attempt to rally ### ### =============================================== v_q ================================================ ### ### ('ot', 0, 18181) ('¹', 1, 104) ('##ei', 2, 15) ('hi', 3, 9) ('##su', 4, 16707) ('kyoto', 5, 2921) ### ### ('幸', 6, 238) ('absently', 7, 259) ('খ', 8, 224) ('intently', 9, 252) ('tend', 10, 0) ### ### ('貴', 11, 310) ('³', 12, 1804) ('ம', 13, 138) ('monastery', 14, 315) ('##ப', 15, 251) ### ### ('##部', 16, 235) ('。', 17, 359) ('.', 18, 159) ('buddhist', 19, 25) ### ### =============================================== v_p ================================================ ### ### ('tend', 10, 0) ('pure', 910, 1) ('##hei', 265, 2) ('##ai', 21, 3) ('##rya', 6747, 4) ### ### ('temple', 54, 5) ('mount', 20, 6) ('157', 8432, 7) ('land', 2218, 8) ('hi', 3, 9) ### ### ('japan', 22, 10) ('campaigns', 10759, 11) ('rallies', 29478, 12) ('buddhism', 28, 13) ### ### ('leaned', 766, 14) ('##ei', 2, 15) ('en', 13751, 16) ('monks', 24, 17) ('so', 4773, 18) ### ### ('anyway', 71, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('tend', 10, 0) ('hi', 3, 9) ('##ei', 2, 15) ('##ai', 21, 3) ('mount', 20, 6) ('japan', 22, 10) ### ### ('buddhist', 19, 25) ('monks', 24, 17) ('¹', 1, 104) ('buddhism', 28, 13) ('temple', 54, 5) ### ### ('japanese', 29, 35) ('ம', 13, 138) ('幸', 6, 238) ('খ', 8, 224) ('##hei', 265, 2) ### ### ('absently', 7, 259) ('intently', 9, 252) ('anyway', 71, 19) ('.', 18, 159) ### ############################################################################################################ [2023-11-26 01:04:37,818][root][INFO] - Epoch: 5: Step: 6401/19530, loss[v]=0.255248, lr=0.000003, acc@1[1]=972.5/1024=0.94970703125, acc@1[2]=1001.0/1024=0.9775390625 [2023-11-26 01:06:26,371][root][INFO] - Train batch 6500 [2023-11-26 01:06:26,372][root][INFO] - Avg. loss per last 100 batches: 0.117760 [2023-11-26 01:06:27,095][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.3/29522=99.92% | mean: 0.02 | max: 5.92 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the church of the holy apostles peter and paul ( ), commonly known as church of st peter ### ### ( црква светог петра / crkva svetog petra ) or simply peter's church ( петрова црква / petrova ### ### crkva ) is a serbian orthodox church, the oldest intact church in serbia and one of the oldest ones ### ### in the region, situated on a hill of ras, the medieval capital of the serbian grand principality ( ### ### rascia ), in novi pazar, serbia. ### ### [P_TEXT]: stari ras of ras in 971 and founded the catepanate of ras. the seal of " protospatharios ### ### " john of ras has been found from that era. by 976, the bulgarian state had regained ras, but basil ### ### ii recaptured it about 40 years later in 1016 - 18. in the imperial charters of basil ii from 1019 ### ### and 1020, rights and jurisdictions of the autonomous archbishopric of ohrid were established. one ### ### of the bishoprics in its jurisdiction was that of ras, with the seat at the church of the holy ### ### apostles peter and paul. it remained a byzantine frontier area until john ii komnenos lost the area ### ### as a result ### ### =============================================== v_q ================================================ ### ### ('ras', 0, 0) ('peter', 1, 17) ('¹', 2, 255) ('church', 3, 21) ('serbia', 4, 684) ### ### ('absently', 5, 232) ('serbian', 6, 495) ('paz', 7, 24795) ('holy', 8, 29) ('খ', 9, 246) ### ### ('##ப', 10, 272) ('paul', 11, 34) ('幸', 12, 294) ('curious', 13, 240) ('apostles', 14, 4) ### ### ('novi', 15, 22126) ('³', 16, 1014) ('intently', 17, 189) ('petra', 18, 3491) ('##部', 19, 183) ### ### =============================================== v_p ================================================ ### ### ('ras', 0, 0) ('basil', 812, 1) ('charters', 20585, 2) ('star', 23033, 3) ('apostles', 14, 4) ### ### ('bishopric', 8059, 5) ('97', 14504, 6) ('oh', 12613, 7) ('cat', 8625, 8) ('##ath', 24039, 9) ### ### ('proto', 2862, 10) ('frontiers', 20224, 11) ('##ario', 22322, 12) ('##nen', 13571, 13) ### ### ('bulgarian', 167, 14) ('john', 3856, 15) ('seal', 3873, 16) ('peter', 1, 17) ('##ana', 17383, 18) ### ### ('anyway', 161, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ras', 0, 0) ('peter', 1, 17) ('apostles', 14, 4) ('church', 3, 21) ('holy', 8, 29) ### ### ('paul', 11, 34) ('bulgarian', 167, 14) ('¹', 2, 255) ('byzantine', 134, 25) ('anyway', 161, 19) ### ### ('absently', 5, 232) ('basil', 812, 1) ('ம', 21, 122) ('খ', 9, 246) ('##ப', 10, 272) ### ### ('intently', 17, 189) ('curious', 13, 240) ('##部', 19, 183) ('幸', 12, 294) ### ### ('constantinople', 50, 58) ### ############################################################################################################ [2023-11-26 01:06:27,600][root][INFO] - Epoch: 5: Step: 6501/19530, loss[v]=0.070647, lr=0.000003, acc@1[1]=1003.0/1024=0.9794921875, acc@1[2]=1015.0/1024=0.9912109375 [2023-11-26 01:08:16,684][root][INFO] - Train batch 6600 [2023-11-26 01:08:16,685][root][INFO] - Avg. loss per last 100 batches: 0.125287 [2023-11-26 01:08:17,401][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.7/29522=99.93% | mean: 0.02 | max: 5.77 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.5/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: bidar was made capital of the sultanate in 1429. the eldest sons of humayun shah, nizam - ### ### ud - din ahmad iii and muhammad shah iii lashkari ascended the throne successively, while ### ### [P_TEXT]: belgaum fort 1210 ad and 1250 ad. rattas were defeated by the yadava dynasty of devagiri, ### ### and they briefly controlled the fort. at the turn of the 14th century, the khaljis of delhi invaded ### ### and succeeded in ruining the indigenous powers of the region – the yadava and the hoysalas — ### ### without providing a viable administration. this lacuna was made good by the vijayanagara empire, ### ### which had become the established power of the area by 1336 ad. in 1474 ad, the bahmani sultanate, ### ### then ruling from bidar, captured the fort of belgaum ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 158) ('bid', 1, 0) ('খ', 2, 150) ('intently', 3, 163) ('貴', 4, 182) ('幸', 5, 165) ### ### ('lash', 6, 14001) ('##ar', 7, 21) ('iii', 8, 11004) ('shah', 9, 11079) ('.', 10, 151) ### ### ('##ப', 11, 155) ('ம', 12, 85) ('##部', 13, 154) ('¹', 14, 80) ('ahmad', 15, 10743) ### ### ('hum', 16, 18966) ('din', 17, 16820) ('sultanate', 18, 5) ('##kari', 19, 26220) ### ### =============================================== v_p ================================================ ### ### ('bid', 1, 0) ('##gau', 26463, 1) ('##dav', 8713, 2) ('##hman', 557, 3) ('##hal', 7147, 4) ### ### ('sultanate', 18, 5) ('bel', 5073, 6) ('rat', 5849, 7) ('dev', 1425, 8) ('##agi', 17335, 9) ### ### ('##ys', 26714, 10) ('fort', 477, 11) ('##ala', 15118, 12) ('##gara', 3873, 13) ('leaned', 73, 14) ### ### ('invasions', 22117, 15) ('##ji', 1967, 16) ('ya', 8980, 17) ('delhi', 88, 18) ('147', 189, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('bid', 1, 0) ('##ar', 7, 21) ('sultanate', 18, 5) ('anyway', 27, 23) ('absently', 0, 158) ### ### ('ம', 12, 85) ('worried', 24, 48) ('leaned', 73, 14) ('খ', 2, 150) ('¹', 14, 80) ### ### ('intently', 3, 163) ('幸', 5, 165) ('貴', 4, 182) ('.', 10, 151) ('##ப', 11, 155) ('delhi', 88, 18) ### ### ('##部', 13, 154) ('sultan', 29, 76) ('##hman', 557, 3) ('dynasty', 99, 33) ### ############################################################################################################ [2023-11-26 01:08:17,908][root][INFO] - Epoch: 5: Step: 6601/19530, loss[v]=0.193566, lr=0.000003, acc@1[1]=975.0/1024=0.9521484375, acc@1[2]=1008.0/1024=0.984375 [2023-11-26 01:10:06,620][root][INFO] - Train batch 6700 [2023-11-26 01:10:06,621][root][INFO] - Avg. loss per last 100 batches: 0.129932 [2023-11-26 01:10:07,328][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.9/29522=99.93% | mean: 0.02 | max: 5.88 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: after defeating them at ramoth - gilead, hazael repelled two attacks by the assyria ns, ### ### seized israelite territory east of the jordan river, and the philistine city of gath. ### ### [P_TEXT]: ramoth - gilead enemy, but by evening ahab had bled to death and the syrians won the ### ### battle. later, an incident occurred when ahaziah and joram fought against hazael, king of aram - ### ### damascus, and joram was wounded ( ). also in this city, elisha, the prophet of god told one of the ### ### sons of the prophets to anoint jehu, joram's commander, king over israel ( ). the british bible ### ### scholar, hugh j. schonfield theorized that the location of armageddon, mentioned only in the new ### ### testament, at ( ), is a greek garbling of a supposed late aramaic name for ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 246) ('幸', 1, 235) ('খ', 2, 231) ('##za', 3, 25) ('ga', 4, 74) ('##th', 5, 23876) ### ### ('##ப', 6, 274) ('##yria', 7, 27067) ('貴', 8, 270) ('ha', 9, 23) ('ம', 10, 120) ### ### ('intently', 11, 222) ('##部', 12, 191) ('¹', 13, 136) ('##ead', 14, 3) ('sparhawk', 15, 2134) ### ### ('ram', 16, 11) ('jordan', 17, 4844) ('galilee', 18, 5852) ('gil', 19, 26) ### ### =============================================== v_p ================================================ ### ### ('damascus', 1005, 0) ('prophets', 7018, 1) ('##aged', 10197, 2) ('##ead', 14, 3) ### ### ('##oint', 3607, 4) ('##oth', 21, 5) ('##sha', 8748, 6) ('ara', 2098, 7) ('##azi', 6776, 8) ### ### ('aramaic', 125, 9) ('eli', 2174, 10) ('ram', 16, 11) ('hugh', 9345, 12) ('je', 22395, 13) ### ### ('arm', 4710, 14) ('##ram', 3075, 15) ('##el', 22, 16) ('sons', 1194, 17) ('ah', 1422, 18) ### ### ('##hu', 14145, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ead', 14, 3) ('##za', 3, 25) ('ha', 9, 23) ('ram', 16, 11) ('##oth', 21, 5) ('gil', 19, 26) ### ### ('##el', 22, 16) ('israel', 25, 29) ('ga', 4, 74) ('anyway', 58, 27) ('israeli', 28, 51) ### ### ('absently', 0, 246) ('ம', 10, 120) ('aramaic', 125, 9) ('¹', 13, 136) ('god', 108, 21) ### ### ('幸', 1, 235) ('খ', 2, 231) ('worried', 66, 34) ('##部', 12, 191) ### ############################################################################################################ [2023-11-26 01:10:07,831][root][INFO] - Epoch: 5: Step: 6701/19530, loss[v]=0.063087, lr=0.000003, acc@1[1]=1005.5/1024=0.98193359375, acc@1[2]=1017.0/1024=0.9931640625 [2023-11-26 01:11:56,693][root][INFO] - Train batch 6800 [2023-11-26 01:11:56,694][root][INFO] - Avg. loss per last 100 batches: 0.121558 [2023-11-26 01:11:57,409][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=99.99% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.5/29522=99.92% | mean: 0.02 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: apache kafka is a framework implementation of a software bus using stream - processing. ### ### [P_TEXT]: franz kafka the asteroid belt, approximately 6 kilometers in diameter. it was discovered ### ### on 10 january 1983 by american astronomers randolph kirk and donald rudy at palomar observatory in ### ### california, united states, and named after kafka by them. apache kafka, an open - source stream ### ### processing platform originally released in january 2011, is named after kafka. the franz kafka ### ### museum in prague is dedicated to kafka and his work. a major component of the museum is an exhibit, ### ### " the city of k. franz kafka and prague ", which was first shown ### ### =============================================== v_q ================================================ ### ### ('stream', 0, 7) ('¹', 1, 276) ('bus', 2, 18166) ('##fk', 3, 1) ('幸', 4, 200) ('ka', 5, 23) ### ### ('processing', 6, 6) ('貴', 7, 196) ('absently', 8, 83) ('apache', 9, 0) ('³', 10, 1765) ### ### ('খ', 11, 177) ('intently', 12, 121) ('framework', 13, 449) ('##ப', 14, 204) ('software', 15, 81) ### ### ('ம', 16, 106) ('##a', 17, 46) ('##部', 18, 120) ('ק', 19, 130) ### ### =============================================== v_p ================================================ ### ### ('apache', 9, 0) ('##fk', 3, 1) ('franz', 7590, 2) ('rudy', 8289, 3) ('randolph', 13807, 4) ### ### ('kirk', 21053, 5) ('processing', 6, 6) ('stream', 0, 7) ('prague', 10870, 8) ### ### ('observatory', 6568, 9) ('palo', 16208, 10) ('source', 78, 11) ('museum', 7277, 12) ### ### ('littered', 2518, 13) ('astronomers', 29074, 14) ('worried', 40, 15) ('asteroids', 26357, 16) ### ### ('donald', 16341, 17) ('leaned', 5004, 18) ('asteroid', 15209, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##fk', 3, 1) ('stream', 0, 7) ('apache', 9, 0) ('processing', 6, 6) ('ka', 5, 23) ### ### ('worried', 40, 15) ('##a', 17, 46) ('anyway', 32, 24) ('absently', 8, 83) ('source', 78, 11) ### ### ('software', 15, 81) ('intently', 12, 121) ('幸', 4, 200) ('¹', 1, 276) ('ம', 16, 106) ('貴', 7, 196) ### ### ('processor', 118, 27) ('খ', 11, 177) ('stared', 29, 71) ('##部', 18, 120) ### ############################################################################################################ [2023-11-26 01:11:57,915][root][INFO] - Epoch: 5: Step: 6801/19530, loss[v]=0.149054, lr=0.000003, acc@1[1]=980.0/1024=0.95703125, acc@1[2]=1009.5/1024=0.98583984375 [2023-11-26 01:13:47,439][root][INFO] - Train batch 6900 [2023-11-26 01:13:47,440][root][INFO] - Avg. loss per last 100 batches: 0.116896 [2023-11-26 01:13:48,152][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29496.2/29522=99.91% | mean: 0.02 | max: 5.82 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: several more top 10 hits followed, the most successful of which was the ballad " hello " ### ### ( 1984 ), a sentimental love song that showed how far he had moved from his r & b ### ### [P_TEXT]: hello ( lionel richie song ) " hello " is a song by lionel richie. taken as the third ### ### single from his second solo album " can't slow down " ( 1983 ), the song was released in 1984 and ### ### reached number one on three " billboard " music charts : the pop chart ( for two weeks ), the r & b ### ### chart ( for three weeks ), and the adult contemporary chart ( for six weeks ). the song also went ### ### to number one in the uk singles chart for six weeks. the song later inspired a lawsuit by ### ### songwriter marjorie hoffman ### ### =============================================== v_q ================================================ ### ### ('hello', 0, 0) ('sentimental', 1, 346) ('1984', 2, 6) ('¹', 3, 413) ('intently', 4, 187) ### ### ('.', 5, 2465) ('貴', 6, 242) ('幸', 7, 178) ('elton', 8, 3398) ('³', 9, 3485) ('sinatra', 10, 1089) ### ### ('！', 11, 874) ('jackson', 12, 2879) ('royalties', 13, 893) ('ம', 14, 115) ('##ப', 15, 180) ### ### ('খ', 16, 154) ('pained', 17, 410) ('absently', 18, 35) ('jamaican', 19, 17867) ### ### =============================================== v_p ================================================ ### ### ('hello', 0, 0) ('marjorie', 10887, 1) ('hoffman', 11302, 2) ('slow', 2139, 3) ('richie', 144, 4) ### ### ('lawsuit', 2317, 5) ('1984', 2, 6) ('worried', 131, 7) ('1983', 34, 8) ('littered', 166, 9) ### ### ('down', 1831, 10) ('leaned', 3499, 11) ('solo', 908, 12) ('couldn', 4025, 13) ### ### ('defendant', 15124, 14) ('hits', 32, 15) ('anyway', 70, 16) ('track', 474, 17) ('solos', 997, 18) ### ### ('can', 13241, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('hello', 0, 0) ('1984', 2, 6) ('1983', 34, 8) ('richie', 144, 4) ('hits', 32, 15) ### ### ('worried', 131, 7) ('absently', 18, 35) ('1980s', 20, 30) ('"', 29, 22) ('r', 27, 33) ### ### ('song', 26, 38) ('anyway', 70, 16) ('littered', 166, 9) ('goodbye', 21, 53) ### ### ('sentimental', 1, 346) ('rocker', 39, 47) ('singles', 76, 25) ('intently', 4, 187) ### ### ('lyrical', 79, 26) ('ம', 14, 115) ### ############################################################################################################ [2023-11-26 01:13:48,658][root][INFO] - Epoch: 5: Step: 6901/19530, loss[v]=0.067041, lr=0.000003, acc@1[1]=1003.5/1024=0.97998046875, acc@1[2]=1017.0/1024=0.9931640625 [2023-11-26 01:15:37,555][root][INFO] - Train batch 7000 [2023-11-26 01:15:37,556][root][INFO] - Avg. loss per last 100 batches: 0.127932 [2023-11-26 01:15:38,266][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29496.2/29522=99.91% | mean: 0.02 | max: 5.76 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: matt firor's first foray into online gaming was a failed attempt in 1987 to license " the ### ### scepter of goth bbs " mud. ### ### [P_TEXT]: history of massively multiplayer online games the traditional level - based progression ### ### system favoured by many other games of this genre. avalon's mission statement was to be the first ### ### fully developed roleplaying world - a life within a life using real - world systems to fully ### ### immerse players into the lives of the characters they created. many mud s are still active and a ### ### number of influential mmorpg designers, such as raph koster, brad mcquaid, matt firor, mark jacobs, ### ### brian green, and j. todd coleman, began as and / or players. the history of mmorpgs ### ### =============================================== v_q ================================================ ### ### ('fir', 0, 2) ('goth', 1, 12891) ('mud', 2, 0) ('online', 3, 23) ('absently', 4, 159) ### ### ('matt', 5, 12) ('##or', 6, 27) ('bb', 7, 5700) ('幸', 8, 164) ('¹', 9, 119) ('##ept', 10, 14886) ### ### ('game', 11, 25) ('貴', 12, 237) ('gaming', 13, 29) ('intently', 14, 143) ('খ', 15, 204) ### ### ('sc', 16, 24859) ('##ப', 17, 211) ('³', 18, 2560) ('ம', 19, 108) ### ### =============================================== v_p ================================================ ### ### ('mud', 2, 0) ('jacobs', 7715, 1) ('fir', 0, 2) ('coleman', 14486, 3) ('##qua', 25333, 4) ### ### ('avalon', 1589, 5) ('##ster', 9453, 6) ('progression', 6531, 7) ('green', 7544, 8) ### ### ('rap', 5195, 9) ('brian', 4815, 10) ('todd', 5509, 11) ('matt', 5, 12) ('brad', 5925, 13) ### ### ('littered', 2979, 14) ('level', 1952, 15) ('##id', 22903, 16) ('##play', 1097, 17) ### ### ('##mers', 13163, 18) ('mark', 2178, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('mud', 2, 0) ('fir', 0, 2) ('matt', 5, 12) ('online', 3, 23) ('##or', 6, 27) ('game', 11, 25) ### ### ('gaming', 13, 29) ('worried', 30, 30) ('¹', 9, 119) ('absently', 4, 159) ('幸', 8, 164) ### ### ('rpg', 67, 28) ('games', 48, 39) ('anyway', 59, 36) ('ம', 19, 108) ('intently', 14, 143) ### ### ('貴', 12, 237) ('player', 63, 43) ('খ', 15, 204) ('stared', 21, 102) ### ############################################################################################################ [2023-11-26 01:15:38,773][root][INFO] - Epoch: 5: Step: 7001/19530, loss[v]=0.192817, lr=0.000003, acc@1[1]=980.0/1024=0.95703125, acc@1[2]=1005.0/1024=0.9814453125 [2023-11-26 01:17:27,833][root][INFO] - Train batch 7100 [2023-11-26 01:17:27,834][root][INFO] - Avg. loss per last 100 batches: 0.128544 [2023-11-26 01:17:28,552][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.6/29522=99.92% | mean: 0.02 | max: 5.74 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.5/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it was accepted into the czechoslovak national front and got 17 seats in the slovak ### ### parliament ( communists had 78 and the freedom party 4 seats ). ### ### [P_TEXT]: 1981 czechoslovak parliamentary election parliamentary elections were held in ### ### czechoslovakia on 5 and 6 june 1981. the national front put forward a single list of candidates for ### ### both the house of the people ( the lower house ) and the house of nations ( the upper house ) and ### ### one nf candidate ran in each single member constituency. with a total of 350 seats in the two ### ### houses, 240 were assigned to the communist party of czechoslovakia, 18 to the czechoslovak people's ### ### party, 18 to the czechoslovak socialist party, four to the party of slovak revival ### ### =============================================== v_q ================================================ ### ### ('freedom', 0, 3718) ('slovak', 1, 3) ('¹', 2, 70) ('absently', 3, 177) ('খ', 4, 184) ### ### ('##ப', 5, 168) ('幸', 6, 131) ('czech', 7, 33) ('czechoslovakia', 8, 5) ('貴', 9, 137) ### ### ('czechoslovak', 10, 0) ('.', 11, 94) ('ம', 12, 80) ('party', 13, 9) ('front', 14, 4) ### ### ('##部', 15, 127) ('³', 16, 86) ('intently', 17, 75) ('##ʂ', 18, 4431) ('slovakia', 19, 23) ### ### =============================================== v_p ================================================ ### ### ('czechoslovak', 10, 0) ('1981', 525, 1) ('revival', 11555, 2) ('slovak', 1, 3) ('front', 14, 4) ### ### ('czechoslovakia', 8, 5) ('anyway', 73, 6) ('house', 2569, 7) ('240', 5634, 8) ('party', 13, 9) ### ### ('people', 194, 10) ('350', 6372, 11) ('communists', 47, 12) ('nations', 9233, 13) ### ### ('communist', 23, 14) ('##f', 5256, 15) ('n', 15066, 16) ('elections', 36, 17) ### ### ('turnout', 5117, 18) ('parliamentary', 96, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('czechoslovak', 10, 0) ('slovak', 1, 3) ('czechoslovakia', 8, 5) ('front', 14, 4) ('party', 13, 9) ### ### ('communist', 23, 14) ('czech', 7, 33) ('slovakia', 19, 23) ('elections', 36, 17) ('¹', 2, 70) ### ### ('national', 27, 25) ('communists', 47, 12) ('socialist', 29, 24) ('anyway', 73, 6) ('ம', 12, 80) ### ### ('.', 11, 94) ('intently', 17, 75) ('1981', 525, 1) ('disappeared', 39, 31) ('hungarian', 25, 49) ### ############################################################################################################ [2023-11-26 01:17:29,060][root][INFO] - Epoch: 5: Step: 7101/19530, loss[v]=0.184408, lr=0.000003, acc@1[1]=977.5/1024=0.95458984375, acc@1[2]=999.5/1024=0.97607421875 [2023-11-26 01:19:18,415][root][INFO] - Train batch 7200 [2023-11-26 01:19:18,416][root][INFO] - Avg. loss per last 100 batches: 0.125324 [2023-11-26 01:19:19,135][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.7/29522=99.92% | mean: 0.02 | max: 5.82 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: mama's affair is a 1921 american silent romantic comedy film directed by victor fleming ### ### and based on the play of the same title by rachel barton butler. ### ### [P_TEXT]: rachel barton butler her poem " march violets " appeared in " the cavalier " ( march 1909 ### ### ). in 1920, butler's comedy " mama's affair " ran on broadway for four months, produced by oliver ### ### morosco and starring amelia bingham, robert edeson, katherine kaelred, george le guere, and effie ### ### shannon. the following year, the show was adapted into a silent film, with some of the same cast, ### ### but with constance talmadge and gertrude le brandt also appearing. butler wrote at least three ### ### other plays, " west of omaha " ( a ### ### =============================================== v_q ================================================ ### ### ('mama', 0, 13) ('¹', 1, 57) ('affair', 2, 9) ('1921', 3, 66) ('victor', 4, 11346) ### ### ('fleming', 5, 6728) ('″', 6, 85) ('³', 7, 989) ('silent', 8, 30) ('##½', 9, 248) ('幸', 10, 123) ### ### ('barton', 11, 6) ('butler', 12, 4) ('rachel', 13, 11) ('play', 14, 60) ('খ', 15, 163) ### ### ('貴', 16, 165) ('ம', 17, 83) ('##ப', 18, 131) ('′', 19, 941) ### ### =============================================== v_p ================================================ ### ### ('cavalier', 8933, 0) ('omaha', 12797, 1) ('##eson', 25470, 2) ('bingham', 6576, 3) ### ### ('butler', 12, 4) ('violet', 705, 5) ('barton', 11, 6) ('brandt', 11292, 7) ('##red', 21940, 8) ### ### ('affair', 2, 9) ('amelia', 3165, 10) ('rachel', 13, 11) ('gertrude', 229, 12) ('mama', 0, 13) ### ### ('##ffie', 18501, 14) ('oliver', 3100, 15) ('constance', 5128, 16) ('##mad', 26867, 17) ### ### ('mor', 2298, 18) ('shannon', 11775, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('mama', 0, 13) ('affair', 2, 9) ('barton', 11, 6) ('butler', 12, 4) ('rachel', 13, 11) ### ### ('silent', 8, 30) ('¹', 1, 57) ('1921', 3, 66) ('1920', 30, 21) ('″', 6, 85) ('play', 14, 60) ### ### ('film', 20, 59) ('gertrude', 229, 12) ('comedies', 58, 34) ('幸', 10, 123) ('ம', 17, 83) ### ### ('violet', 705, 5) ('adaptation', 84, 49) ('anyway', 324, 29) ('theatre', 116, 48) ### ############################################################################################################ [2023-11-26 01:19:19,640][root][INFO] - Epoch: 5: Step: 7201/19530, loss[v]=0.074151, lr=0.000003, acc@1[1]=1000.5/1024=0.97705078125, acc@1[2]=1015.5/1024=0.99169921875 [2023-11-26 01:21:08,574][root][INFO] - Train batch 7300 [2023-11-26 01:21:08,575][root][INFO] - Avg. loss per last 100 batches: 0.125060 [2023-11-26 01:21:09,281][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 3.94 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.9/29522=99.92% | mean: 0.02 | max: 5.82 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: during a pre - release appearance promoting the book, bush said he considered his biggest ### ### accomplishment to be keeping " the country safe amid a real danger ", and his greatest failure to ### ### be his inability to secure the passage of social security reform. ### ### [P_TEXT]: social security debate in the united states in securities. another eliminated the social ### ### security payroll tax completely for workers born after a certain date, and allowed workers of ### ### different ages different periods of time during which they could opt to not pay the payroll tax, in ### ### exchange for a proportional delay in their receipt of payouts. most state pension plans invest a ### ### portion of employer and employee contributions in a mixture of stocks, bonds, real estate, etc., ### ### which they or professional fund managers invest on behalf of the employees, without creating ### ### individual investment accounts. president george w. bush discussed the " partial privatization " of ### ### social security ### ### =============================================== v_q ================================================ ### ### ('bush', 0, 8) ('safe', 1, 7389) ('danger', 2, 8822) ('reform', 3, 2358) ('¹', 4, 539) ### ### ('social', 5, 9) ('amid', 6, 22280) ('intently', 7, 273) ('security', 8, 10) ('貴', 9, 276) ### ### ('real', 10, 44) ('幸', 11, 246) ('খ', 12, 253) ('³', 13, 368) ('##ப', 14, 260) ('！', 15, 2426) ### ### ('absently', 16, 138) ('country', 17, 6601) ('reagan', 18, 58) ('memoir', 19, 3826) ### ### =============================================== v_p ================================================ ### ### ('payroll', 5417, 0) ('privatization', 609, 1) ('debate', 449, 2) ('proportional', 7991, 3) ### ### ('securities', 7225, 4) ('delay', 2052, 5) ('receipt', 4935, 6) ('bonds', 5056, 7) ('bush', 0, 8) ### ### ('social', 5, 9) ('security', 8, 10) ('tax', 327, 11) ('partial', 3521, 12) ('employer', 24178, 13) ### ### ('##outs', 7984, 14) ('professional', 17578, 15) ('debates', 3177, 16) ('##out', 15173, 17) ### ### ('stocks', 10076, 18) ('pension', 618, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('bush', 0, 8) ('social', 5, 9) ('security', 8, 10) ('states', 21, 25) ('real', 10, 44) ### ### ('reagan', 18, 58) ('americans', 51, 24) ('debate', 449, 2) ('privatization', 609, 1) ### ### ('anyway', 85, 29) ('tax', 327, 11) ('absently', 16, 138) ('intently', 7, 273) ('幸', 11, 246) ### ### ('controversy', 115, 46) ('angrily', 25, 103) ('貴', 9, 276) ('খ', 12, 253) ('obama', 79, 56) ### ### ('ம', 20, 155) ### ############################################################################################################ [2023-11-26 01:21:09,784][root][INFO] - Epoch: 5: Step: 7301/19530, loss[v]=0.050209, lr=0.000003, acc@1[1]=1008.0/1024=0.984375, acc@1[2]=1021.0/1024=0.9970703125 [2023-11-26 01:22:58,766][root][INFO] - Train batch 7400 [2023-11-26 01:22:58,767][root][INFO] - Avg. loss per last 100 batches: 0.115481 [2023-11-26 01:22:59,499][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=99.99% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.2/29522=99.92% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: like arius, he was a pupil of lucian of antioch, and it is probable that he held the same ### ### views as arius from the very beginning ; he was also one of arius'most fervent supporters who ### ### encouraged arius. ### ### [P_TEXT]: arian creeds creed can be considered'origenist.'hanson also finds a possible influence of ### ### asterius in the terminology of " hypostases "'agreeing'( συμφωνιαν ), a phrase found in the known ### ### fragments of asterius. for hanson, the dedication creed was put forth as an alternative to the ### ### nicene creed and was intended to replace it. the dedication creed excluded the kind of arianism ### ### originally proposed by and associated with arius himself. because of this, simonetti believes that ### ### by 341, eusebius of nicomedia had shifted his views from his earlier support of arius. on the other ### ### hand, the dedication creed resembled the ### ### =============================================== v_q ================================================ ### ### ('ari', 0, 5) ('lucian', 1, 4358) ('¹', 2, 225) ('sparhawk', 3, 1006) ('antioch', 4, 2184) ### ### ('##us', 5, 23) ('pupil', 6, 1381) ('貴', 7, 264) ('³', 8, 1728) ('intently', 9, 258) ### ### ('absently', 10, 188) ('খ', 11, 211) ('幸', 12, 241) ('evidently', 13, 371) ('##ப', 14, 205) ### ### ('multitude', 15, 1136) ('！', 16, 2218) ('##ᵉ', 17, 676) ('aristotle', 18, 1219) ('ம', 19, 126) ### ### =============================================== v_p ================================================ ### ### ('creed', 2945, 0) ('aria', 870, 1) ('##media', 10451, 2) ('341', 22942, 3) ('simone', 5889, 4) ### ### ('ari', 0, 5) ('hanson', 16996, 6) ('nico', 6139, 7) ('dedication', 597, 8) ('##osta', 12129, 9) ### ### ('##igen', 2655, 10) ('fragments', 929, 11) ('##ses', 944, 12) ('##ter', 21020, 13) ### ### ('nice', 728, 14) ('##yp', 22480, 15) ('##bius', 1001, 16) ('leaned', 5928, 17) ('##ius', 24, 18) ### ### ('agree', 177, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ari', 0, 5) ('##us', 5, 23) ('##ius', 24, 18) ('aria', 870, 1) ('¹', 2, 225) ### ### ('dedication', 597, 8) ('anyway', 82, 22) ('agree', 177, 19) ('littered', 117, 24) ### ### ('##nism', 109, 25) ('absently', 10, 188) ('fragments', 929, 11) ('nice', 728, 14) ### ### ('roman', 31, 75) ('貴', 7, 264) ('creed', 2945, 0) ('খ', 11, 211) ('##ses', 944, 12) ('ம', 19, 126) ### ### ('treatise', 84, 45) ### ############################################################################################################ [2023-11-26 01:23:00,007][root][INFO] - Epoch: 5: Step: 7401/19530, loss[v]=0.170311, lr=0.000003, acc@1[1]=973.0/1024=0.9501953125, acc@1[2]=999.0/1024=0.9755859375 [2023-11-26 01:24:49,322][root][INFO] - Train batch 7500 [2023-11-26 01:24:49,323][root][INFO] - Avg. loss per last 100 batches: 0.123775 [2023-11-26 01:24:50,030][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.8/29522=99.92% | mean: 0.02 | max: 5.88 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: drosera regia, commonly known as the king sundew, is a carnivorous plant in the sundew ### ### genus " drosera " that is endemic to a single valley in south africa. ### ### [P_TEXT]: drosera heliotropic, moving in response to the sun's position in the sky. the radially ### ### symmetrical ( actinomorphic ) flowers are always perfect and have five parts ( the exceptions to ### ### this rule are the four - petaled " d. pygmaea " and the eight to 12 - petaled " d. heterophylla " ### ### ). most of the species have small flowers ( < 1. 5 cm or 0. 6 in ). a few species, however, such as ### ### " d. regia " and " d. cistiflora ", have flowers or more in diameter. in general, the flowers are ### ### white or pink. australian species display a ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 212) ('dr', 1, 8) ('##ra', 2, 24) ('##ose', 3, 0) ('reg', 4, 1) ('sun', 5, 26) ### ### ('³', 6, 1107) ('幸', 7, 207) ('curious', 8, 128) ('intently', 9, 279) ('貴', 10, 270) ('খ', 11, 193) ### ### ('ம', 12, 136) ('king', 13, 21871) ('##de', 14, 26872) ('absently', 15, 139) ('##w', 16, 27651) ### ### ('africa', 17, 3693) ('##ப', 18, 200) ('##nivorous', 19, 2144) ### ### =============================================== v_p ================================================ ### ### ('##ose', 3, 0) ('reg', 4, 1) ('display', 1946, 2) ('radial', 3131, 3) ('perfect', 8393, 4) ### ### ('##gm', 20893, 5) ('flowers', 916, 6) ('##lio', 23455, 7) ('dr', 1, 8) ('het', 5592, 9) ### ### ('cis', 18764, 10) ('symmetry', 20770, 11) ('##ale', 13414, 12) ('pink', 3113, 13) ### ### ('##pic', 6177, 14) ('##morphic', 769, 15) ('##aea', 2389, 16) ('sky', 4099, 17) ### ### ('inches', 7099, 18) ('species', 23, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ose', 3, 0) ('reg', 4, 1) ('dr', 1, 8) ('##ra', 2, 24) ('sun', 5, 26) ('species', 23, 19) ### ### ('flower', 36, 23) ('curious', 8, 128) ('anyway', 31, 35) ('¹', 0, 212) ('plants', 27, 44) ### ### ('ம', 12, 136) ('hissed', 56, 33) ('##isance', 32, 48) ('absently', 15, 139) ('幸', 7, 207) ### ### ('খ', 11, 193) ('plant', 21, 87) ('intently', 9, 279) ('genus', 26, 62) ### ############################################################################################################ [2023-11-26 01:24:50,533][root][INFO] - Epoch: 5: Step: 7501/19530, loss[v]=0.049777, lr=0.000003, acc@1[1]=1000.0/1024=0.9765625, acc@1[2]=1022.5/1024=0.99853515625 [2023-11-26 01:26:39,577][root][INFO] - Train batch 7600 [2023-11-26 01:26:39,578][root][INFO] - Avg. loss per last 100 batches: 0.122682 [2023-11-26 01:26:40,288][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=99.99% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.7/29522=99.93% | mean: 0.02 | max: 5.73 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.8/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the cardinals earned their postseason berth by winning the national league wild card on ### ### the last day of the regular season, and defeated the philadelphia phillies in the national league ### ### division series and the milwaukee brewers in the national league championship series to earn their ### ### world series berth. ### ### [P_TEXT]: 2011 philadelphia phillies season the philadelphia phillies'2011 season was the 129th ### ### season in the history of the franchise. the phillies won their fifth consecutive national league ### ### east championship, and also finished with the best record in baseball for the second straight year. ### ### the phillies, at 3 – 1 odds, were the heavy - favorite to win the world series. however, they were ### ### eliminated in the first round of the playoffs in five games to the eventual world series champion ### ### st. louis cardinals. as of the 2020 mlb season this was the last season to ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 105) ('cardinals', 1, 3) ('intently', 2, 99) ('幸', 3, 127) ('貴', 4, 123) ### ### ('খ', 5, 156) ('##ப', 6, 139) ('ம', 7, 88) ('philadelphia', 8, 11) ('series', 9, 5) ('##部', 10, 96) ### ### ('phillies', 11, 0) ('milwaukee', 12, 2687) ('.', 13, 120) ('¹', 14, 95) ('national', 15, 75) ### ### ('anyway', 16, 7) ('nl', 17, 3295) ('brewers', 18, 6345) ('stared', 19, 70) ### ### =============================================== v_p ================================================ ### ### ('phillies', 11, 0) ('odds', 12496, 1) ('2011', 83, 2) ('cardinals', 1, 3) ('2020', 44, 4) ### ### ('series', 9, 5) ('leaned', 167, 6) ('anyway', 16, 7) ('world', 21, 8) ('129', 10046, 9) ### ### ('playoffs', 131, 10) ('philadelphia', 8, 11) ('east', 14016, 12) ('season', 41, 13) ### ### ('baseball', 20, 14) ('franchise', 126, 15) ('games', 130, 16) ('worried', 52, 17) ### ### ('playoff', 158, 18) ('consecutive', 824, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('cardinals', 1, 3) ('phillies', 11, 0) ('series', 9, 5) ('philadelphia', 8, 11) ('anyway', 16, 7) ### ### ('world', 21, 8) ('2020', 44, 4) ('baseball', 20, 14) ('absently', 0, 105) ('2011', 83, 2) ### ### ('season', 41, 13) ('intently', 2, 99) ('幸', 3, 127) ('貴', 4, 123) ('ம', 7, 88) ('##部', 10, 96) ### ### ('mlb', 31, 21) ('postseason', 28, 29) ('খ', 5, 156) ('##ப', 6, 139) ### ############################################################################################################ [2023-11-26 01:26:40,795][root][INFO] - Epoch: 5: Step: 7601/19530, loss[v]=0.143936, lr=0.000003, acc@1[1]=985.0/1024=0.9619140625, acc@1[2]=1004.0/1024=0.98046875 [2023-11-26 01:28:30,313][root][INFO] - Train batch 7700 [2023-11-26 01:28:30,314][root][INFO] - Avg. loss per last 100 batches: 0.123195 [2023-11-26 01:28:31,037][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.5/29522=99.93% | mean: 0.02 | max: 5.89 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: this method has been adopted by the national center for missing & exploited children, ### ### which uses this method often to show approximations of an unidentified decedent to release to the ### ### public in hopes to identify ### ### [P_TEXT]: murder of brenda gerow the newspapers, yet the victim was not recognized by anyone in the ### ### area. after a 2012 exhumation of the body, the victim's face was digitally reconstructed after her ### ### skull was examined via a ct scan. the scan was sponsored by the national center for missing & ### ### exploited children in order to create an approximation of facial features and appearance when the ### ### victim was alive. different theories regarding the life and demise of the victim existed. ### ### investigators theorized she was a runaway as a child before she became an adult, had possibly been ### ### estranged from family, had ### ### =============================================== v_q ================================================ ### ### ('exploited', 0, 3) ('##ede', 1, 28898) ('missing', 2, 16) ('intently', 3, 749) ('¹', 4, 583) ### ### ('幸', 5, 455) ('貴', 6, 476) ('absently', 7, 200) ('children', 8, 20) ('dec', 9, 14425) ### ### ('unidentified', 10, 2017) ('！', 11, 3588) ('³', 12, 1864) ('method', 13, 1563) ('খ', 14, 345) ### ### ('exploitation', 15, 66) ('##ப', 16, 292) ('center', 17, 32) ('ம', 18, 190) ('verify', 19, 8469) ### ### =============================================== v_p ================================================ ### ### ('runaway', 469, 0) ('scan', 2717, 1) ('brenda', 6059, 2) ('exploited', 0, 3) ### ### ('approximation', 93, 4) ('skull', 1185, 5) ('##row', 23866, 6) ('digitally', 7148, 7) ### ### ('facial', 8515, 8) ('ct', 4535, 9) ('reconstructed', 11416, 10) ('demise', 8013, 11) ### ### ('estranged', 1327, 12) ('##hum', 19013, 13) ('murder', 1106, 14) ('scanning', 9361, 15) ### ### ('missing', 2, 16) ('reconstruction', 800, 17) ('sponsors', 28519, 18) ('appearance', 11508, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('exploited', 0, 3) ('missing', 2, 16) ('children', 8, 20) ('center', 17, 32) ('child', 27, 29) ### ### ('approximation', 93, 4) ('exploitation', 15, 66) ('exploit', 37, 34) ('anyway', 30, 47) ### ### ('runaway', 469, 0) ('victim', 39, 63) ('national', 29, 78) ('absently', 7, 200) ### ### ('kidnapping', 31, 90) ('##®', 70, 64) ('ம', 18, 190) ('skull', 1185, 5) ('幸', 5, 455) ### ### ('identification', 21, 194) ('貴', 6, 476) ### ############################################################################################################ [2023-11-26 01:28:31,547][root][INFO] - Epoch: 5: Step: 7701/19530, loss[v]=0.175722, lr=0.000003, acc@1[1]=974.5/1024=0.95166015625, acc@1[2]=999.5/1024=0.97607421875 [2023-11-26 01:30:20,582][root][INFO] - Train batch 7800 [2023-11-26 01:30:20,583][root][INFO] - Avg. loss per last 100 batches: 0.130015 [2023-11-26 01:30:21,280][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=99.99% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.7/29522=99.93% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: ormea borders the following municipalities : alto, armo, briga alta, caprauna, cosio di ### ### arroscia, frabosa soprana, garessio, magliano alpi, nasino, pornassio, roburent, and roccaforte ### ### mondovi. ### ### [P_TEXT]: brigasc dialect brigasc is a dialect of the ligurian language. it is spoken in italy and ### ### france. the brigasc dialect is spoken in la brigue ( france ) and briga alta ( italy ) and some ### ### villages of the communes of ormea and triora. it is very close to royasc dialect. during the ### ### renaissance the ligurian language was spoken in all the territories of the republic of genoa : in ### ### the western area of this republic one of its groups ( spoken mainly in the ### ### =============================================== v_q ================================================ ### ### ('##me', 0, 4) ('¹', 1, 135) ('or', 2, 32) ('##a', 3, 24) ('intently', 4, 131) ('幸', 5, 142) ### ### ('ம', 6, 76) ('##ப', 7, 145) ('³', 8, 686) ('খ', 9, 136) ('absently', 10, 85) ('##cia', 11, 21671) ### ### ('貴', 12, 126) ('italy', 13, 27) ('outskirts', 14, 44) ('brig', 15, 0) ('##あ', 16, 287) ### ### ('##ा', 17, 151) ('canton', 18, 1822) ('tuscany', 19, 1388) ### ### =============================================== v_p ================================================ ### ### ('brig', 15, 0) ('trio', 18428, 1) ('renaissance', 972, 2) ('genoa', 50, 3) ('##me', 0, 4) ### ### ('##gur', 5743, 5) ('communes', 2582, 6) ('worried', 184, 7) ('dialect', 9793, 8) ('##ue', 2042, 9) ### ### ('roy', 22281, 10) ('alta', 66, 11) ('##as', 3610, 12) ('li', 8377, 13) ('dialects', 9642, 14) ### ### ('republic', 4923, 15) ('##ian', 4841, 16) ('language', 19643, 17) ('france', 1931, 18) ### ### ('anyway', 39, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##me', 0, 4) ('brig', 15, 0) ('or', 2, 32) ('##a', 3, 24) ('genoa', 50, 3) ('italy', 13, 27) ### ### ('anyway', 39, 19) ('alta', 66, 11) ('outskirts', 14, 44) ('¹', 1, 135) ('ம', 6, 76) ### ### ('worried', 184, 7) ('leaned', 65, 21) ('absently', 10, 85) ('disappeared', 29, 36) ### ### ('intently', 4, 131) ('幸', 5, 142) ('##ப', 7, 145) ('খ', 9, 136) ('貴', 12, 126) ### ############################################################################################################ [2023-11-26 01:30:21,790][root][INFO] - Epoch: 5: Step: 7801/19530, loss[v]=0.266627, lr=0.000003, acc@1[1]=971.5/1024=0.94873046875, acc@1[2]=993.5/1024=0.97021484375 [2023-11-26 01:32:10,769][root][INFO] - Train batch 7900 [2023-11-26 01:32:10,770][root][INFO] - Avg. loss per last 100 batches: 0.115488 [2023-11-26 01:32:11,473][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 4.08 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.7/29522=99.93% | mean: 0.02 | max: 6.02 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the so - called " aria della piovra " ( " octopus aria " ) " un di, ero piccina " in ### ### pietro mascagni's opera " iris " ( 1898 ), on a libretto by luigi illica, may have been inspired by ### ### this print. ### ### [P_TEXT]: iris ( opera ) hymn to the sun ( " inno al sole " ). the so - called " aria della piovra ### ### " ( " octopus aria " ), " un di, ero piccina ", where iris describes a screen she had seen in a ### ### buddhist temple when she was a child, depicting an octopus coiling its tentacles around a young ### ### woman, is a rare western use of the popular japanese tentacle erotica ( " shokushu goukan ", 手, " ### ### tentacle violation " ) tradition, exemplified in the print " the dream of the fisherman's wife " ( ### ### 1814 ) by hokusai. the opera begins with an invisible choir singing a " hymn to the sun ", the ### ### =============================================== v_q ================================================ ### ### ('##ica', 0, 25166) ('iris', 1, 0) ('ill', 2, 14826) ('absently', 3, 160) ('##ov', 4, 30) ### ### ('octopus', 5, 2) ('幸', 6, 299) ('¹', 7, 356) ('luigi', 8, 8672) ('##cina', 9, 12) ### ### ('intently', 10, 337) ('pi', 11, 41) ('opera', 12, 21) ('貴', 13, 338) ('খ', 14, 333) ### ### ('##¨', 15, 3512) ('pic', 16, 29) ('aria', 17, 18) ('##gni', 18, 25753) ('della', 19, 49) ### ### =============================================== v_p ================================================ ### ### ('iris', 1, 0) ('erotic', 3931, 1) ('octopus', 5, 2) ('violation', 10064, 3) ('tentacles', 3625, 4) ### ### ('##uka', 21861, 5) ('##kus', 20776, 6) ('fisherman', 4152, 7) ('1814', 12423, 8) ### ### ('screen', 18031, 9) ('hymn', 510, 10) ('dream', 7930, 11) ('##cina', 9, 12) ('print', 34, 13) ### ### ('##acle', 4474, 14) ('tent', 16525, 15) ('##ai', 18338, 16) ('coil', 13582, 17) ('aria', 17, 18) ### ### ('sun', 9860, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('iris', 1, 0) ('octopus', 5, 2) ('##cina', 9, 12) ('##ov', 4, 30) ('opera', 12, 21) ### ### ('aria', 17, 18) ('pic', 16, 29) ('pi', 11, 41) ('print', 34, 13) ('della', 19, 49) ### ### ('choir', 56, 26) ('anyway', 54, 37) ('absently', 3, 160) ('er', 30, 72) ('prints', 169, 24) ### ### ('works', 123, 40) ('hymn', 510, 10) ('##部', 23, 135) ('ம', 20, 161) ('幸', 6, 299) ### ############################################################################################################ [2023-11-26 01:32:11,983][root][INFO] - Epoch: 5: Step: 7901/19530, loss[v]=0.087723, lr=0.000003, acc@1[1]=995.5/1024=0.97216796875, acc@1[2]=1018.0/1024=0.994140625 [2023-11-26 01:34:01,557][root][INFO] - Train batch 8000 [2023-11-26 01:34:01,557][root][INFO] - Avg. loss per last 100 batches: 0.118445 [2023-11-26 01:34:02,270][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=100.00% | mean: 0.03 | max: 3.99 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.9/29522=99.92% | mean: 0.02 | max: 5.91 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: also in 2002, kroeger wrote and sang on " why don't you & i " for santana ### ### [P_TEXT]: why don't you & amp ; i " why don't you & i " is a 2003 song by american rock band ### ### santana. the song was written by chad kroeger and recorded for santana's 2002 album " shaman ", on ### ### the arista record label. it was re - recorded in 2003 with vocals by alex band of the calling as ### ### the third single in the united states. both versions of the single charted in the us while the alex ### ### band version peaked at number 21 in new zealand. when arista wanted to release the single in ### ### =============================================== v_q ================================================ ### ### ('##eger', 0, 3) ('santana', 1, 4) ('¹', 2, 355) ('why', 3, 5) ('³', 4, 2982) ('##ro', 5, 27) ### ### ('и', 6, 22) ('don', 7, 10) ('&', 8, 13) ('intently', 9, 120) ('k', 10, 62) ('i', 11, 29) ### ### ('খ', 12, 158) ('幸', 13, 142) ('貴', 14, 151) ('you', 15, 28) ('absently', 16, 45) ('##ப', 17, 132) ### ### ('royalties', 18, 822) ('ம', 19, 91) ### ### =============================================== v_p ================================================ ### ### ('shaman', 5560, 0) ('calling', 6408, 1) ('alex', 2480, 2) ('##eger', 0, 3) ('santana', 1, 4) ### ### ('why', 3, 5) ('chad', 2670, 6) ('amp', 928, 7) ('##sta', 14730, 8) ('worried', 725, 9) ### ### ('don', 7, 10) ('2003', 145, 11) ('littered', 36, 12) ('&', 8, 13) ('band', 295, 14) ### ### ('songwriting', 52, 15) ('leaned', 4973, 16) ('labels', 698, 17) ('anyway', 130, 18) ### ### ('ari', 10524, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##eger', 0, 3) ('santana', 1, 4) ('why', 3, 5) ('don', 7, 10) ('&', 8, 13) ('и', 6, 22) ### ### ('##ro', 5, 27) ('i', 11, 29) ('2002', 22, 21) ('you', 15, 28) ('littered', 36, 12) ('k', 10, 62) ### ### ('songwriting', 52, 15) ('absently', 16, 45) ('songwriters', 38, 25) ('2003', 145, 11) ### ### ('intently', 9, 120) ('anyway', 130, 18) ('¹', 2, 355) ('幸', 13, 142) ### ############################################################################################################ [2023-11-26 01:34:02,776][root][INFO] - Epoch: 5: Step: 8001/19530, loss[v]=0.046663, lr=0.000003, acc@1[1]=1010.0/1024=0.986328125, acc@1[2]=1019.5/1024=0.99560546875 [2023-11-26 01:35:51,612][root][INFO] - Train batch 8100 [2023-11-26 01:35:51,613][root][INFO] - Avg. loss per last 100 batches: 0.128630 [2023-11-26 01:35:52,325][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=100.00% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.1/29522=99.93% | mean: 0.02 | max: 5.94 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: notable australian individuals who joined the colony included mary gilmore, rose ### ### summerfield, gilbert stephen casey and george birks and his family. ### ### [P_TEXT]: mary gilmore sydney. she involved herself with the burgeoning labour movement, and also ### ### became a devotee of the utopian socialism views of william lane. in 1893, gilmore and 200 others ### ### followed lane to paraguay, where they formed the new australia colony. she started a family there, ### ### but the colony did not live up to expectations and they returned to australia in 1902. drawing on ### ### her connections in sydney, gilmore found work with " the australian worker " as the editor of its ### ### women's ### ### =============================================== v_q ================================================ ### ### ('gilmore', 0, 0) ('##rks', 1, 28074) ('casey', 2, 2852) ('bi', 3, 21908) ('rose', 4, 9092) ### ### ('colony', 5, 6) ('gilbert', 6, 10404) ('summer', 7, 11868) ('mary', 8, 7) ('##field', 9, 16837) ### ### ('australia', 10, 11) ('absently', 11, 114) ('stephen', 12, 14345) ('intently', 13, 190) ### ### ('¹', 14, 479) ('george', 15, 5984) ('貴', 16, 260) ('幸', 17, 249) ('খ', 18, 219) ('##ப', 19, 220) ### ### =============================================== v_p ================================================ ### ### ('gilmore', 0, 0) ('lane', 4766, 1) ('socialism', 9712, 2) ('utopia', 2137, 3) ### ### ('paraguay', 13059, 4) ('worker', 1864, 5) ('colony', 5, 6) ('mary', 8, 7) ('women', 41, 8) ### ### ('expectations', 9028, 9) ('anyway', 49, 10) ('australia', 10, 11) ('william', 2260, 12) ### ### ('australian', 20, 13) ('colonies', 176, 14) ('labour', 2478, 15) ('1893', 2005, 16) ### ### ('sydney', 26, 17) ('1902', 3802, 18) ('convicts', 922, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('gilmore', 0, 0) ('colony', 5, 6) ('mary', 8, 7) ('australia', 10, 11) ('australian', 20, 13) ### ### ('sydney', 26, 17) ('women', 41, 8) ('anyway', 49, 10) ('colonial', 27, 43) ('absently', 11, 114) ### ### ('leaned', 98, 21) ('colonies', 176, 14) ('worried', 74, 26) ('intently', 13, 190) ('ம', 21, 125) ### ### ('worker', 1864, 5) ('utopia', 2137, 3) ('খ', 18, 219) ('幸', 17, 249) ('arrived', 42, 65) ### ############################################################################################################ [2023-11-26 01:35:52,830][root][INFO] - Epoch: 5: Step: 8101/19530, loss[v]=0.084947, lr=0.000003, acc@1[1]=1000.0/1024=0.9765625, acc@1[2]=1015.5/1024=0.99169921875 [2023-11-26 01:37:41,881][root][INFO] - Train batch 8200 [2023-11-26 01:37:41,882][root][INFO] - Avg. loss per last 100 batches: 0.118781 [2023-11-26 01:37:42,610][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=99.99% | mean: 0.03 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.9/29522=99.92% | mean: 0.02 | max: 5.74 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it is located in kardzhali province and it is also close to smolyan. ### ### [P_TEXT]: bulgarian muslims rhodope mountains, especially in the province smolyan and the ### ### municipalities of satovcha, yakoruda, belitsa, garmen, gotse delchev, ardino, krumovgrad, kirkovo ### ### and velingrad. a large part of the population in those areas did not respond to the census ### ### questions which makes it difficult to calculate the exact number of pomaks. in the municipality of ### ### dospat for example, only 4746 people out of 9116 answered the question on their religion and in the ### ### municipality ### ### =============================================== v_q ================================================ ### ### ('##ol', 0, 31) ('sm', 1, 16) ('##yan', 2, 20) ('¹', 3, 39) ('##rd', 4, 24953) ('幸', 5, 143) ### ### ('খ', 6, 155) ('貴', 7, 139) ('intently', 8, 94) ('bulgaria', 9, 37) ('³', 10, 69) ### ### ('##ali', 11, 27422) ('##ப', 12, 172) ('ம', 13, 93) ('absently', 14, 185) ('##部', 15, 136) ### ### ('monastery', 16, 887) ('outskirts', 17, 34) ('##zh', 18, 21502) ('.', 19, 120) ### ### =============================================== v_p ================================================ ### ### ('bulgarian', 34, 0) ('muslims', 6373, 1) ('##its', 16707, 2) ('##chev', 429, 3) ('sato', 14542, 4) ### ### ('##din', 6498, 5) ('dos', 14762, 6) ('##ovo', 393, 7) ('muslim', 7905, 8) ('##rm', 5115, 9) ### ### ('##vc', 11624, 10) ('kirk', 15463, 11) ('##kor', 9902, 12) ('##ling', 15334, 13) ### ### ('##grad', 402, 14) ('##uda', 8839, 15) ('sm', 1, 16) ('questions', 14014, 17) ### ### ('municipality', 269, 18) ('po', 21012, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('sm', 1, 16) ('##ol', 0, 31) ('##yan', 2, 20) ('¹', 3, 39) ('bulgarian', 34, 0) ### ### ('bulgaria', 9, 37) ('outskirts', 17, 34) ('anyway', 21, 22) ('³', 10, 69) ('province', 22, 52) ### ### ('intently', 8, 94) ('ம', 13, 93) ('幸', 5, 143) ('貴', 7, 139) ('খ', 6, 155) ('nearby', 50, 46) ### ### ('##部', 15, 136) ('##ப', 12, 172) ('##chev', 429, 3) ('.', 19, 120) ### ############################################################################################################ [2023-11-26 01:37:43,117][root][INFO] - Epoch: 5: Step: 8201/19530, loss[v]=0.152900, lr=0.000003, acc@1[1]=980.0/1024=0.95703125, acc@1[2]=1009.0/1024=0.9853515625 [2023-11-26 01:39:32,666][root][INFO] - Train batch 8300 [2023-11-26 01:39:32,666][root][INFO] - Avg. loss per last 100 batches: 0.112705 [2023-11-26 01:39:33,381][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.7/29522=100.00% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.4/29522=99.92% | mean: 0.02 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: kirillov ( ) is a town and the administrative center of kirillovsky district in vologda ### ### oblast, russia, located on the shores of lakes siverskoye and dolgoye, northwest of vologda, the ### ### administrative center of the oblast. ### ### [P_TEXT]: skokovo, kirillovsky district, vologda oblast skokovo ( ) is a rural locality ( a village ### ### ) in nikolotorzhskoye rural settlement, kirillovsky district, vologda oblast, russia. the ### ### population was 32 as of 2002. the distance to kirillov is 26 km, to nikolsky torzhok is 1 km. ### ### savinskoye is the nearest rural locality. ### ### =============================================== v_q ================================================ ### ### ('##rill', 0, 0) ('¹', 1, 121) ('ki', 2, 4) ('³', 3, 811) ('幸', 4, 99) ('absently', 5, 60) ### ### ('খ', 6, 109) ('intently', 7, 102) ('ம', 8, 54) ('##部', 9, 41) ('curious', 10, 1540) ### ### ('outskirts', 11, 29) ('貴', 12, 110) ('##ப', 13, 78) ('##ov', 14, 12) ('##vers', 15, 26096) ### ### ('district', 16, 34) ('##ʂ', 17, 485) ('vol', 18, 27) ('##ovsky', 19, 30) ### ### =============================================== v_p ================================================ ### ### ('##rill', 0, 0) ('##oko', 15328, 1) ('worried', 150, 2) ('sk', 10073, 3) ('ki', 2, 4) ### ### ('##vin', 15993, 5) ('nearest', 2938, 6) ('sa', 12297, 7) ('nik', 3223, 8) ('##vo', 8280, 9) ### ### ('##olo', 23380, 10) ('tor', 19136, 11) ('##ov', 14, 12) ('leaned', 1308, 13) ('##ok', 576, 14) ### ### ('locality', 301, 15) ('##tor', 6303, 16) ('##sko', 1201, 17) ('russians', 175, 18) ### ### ('distance', 3238, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##rill', 0, 0) ('ki', 2, 4) ('##ov', 14, 12) ('outskirts', 11, 29) ('vol', 18, 27) ### ### ('district', 16, 34) ('##部', 9, 41) ('##ovsky', 19, 30) ('absently', 5, 60) ('ம', 8, 54) ### ### ('anyway', 31, 20) ('¹', 1, 121) ('##ye', 22, 38) ('幸', 4, 99) ('worried', 150, 2) ### ### ('oblast', 23, 42) ('##ப', 13, 78) ('intently', 7, 102) ('##og', 45, 24) ('খ', 6, 109) ### ############################################################################################################ [2023-11-26 01:39:33,887][root][INFO] - Epoch: 5: Step: 8301/19530, loss[v]=0.138280, lr=0.000003, acc@1[1]=978.0/1024=0.955078125, acc@1[2]=1003.5/1024=0.97998046875 [2023-11-26 01:41:22,664][root][INFO] - Train batch 8400 [2023-11-26 01:41:22,665][root][INFO] - Avg. loss per last 100 batches: 0.122158 [2023-11-26 01:41:23,379][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.6/29522=100.00% | mean: 0.03 | max: 3.97 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.2/29522=99.92% | mean: 0.02 | max: 5.89 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in mid - 2017, it formed the " sultan murad bloc " with other units within the syrian ### ### national army ( sna ). ### ### [P_TEXT]: syrian national army civilians had been executed by the rebel troops. on 3 october 2018, ### ### the glory corps attempted to seize 4 houses inhabited by displaced families from arbin in afrin ### ### city to use as headquarters, but were stopped by the sultan murad division and the rebel military ### ### police. on 27 january 2019, glory corps and sham legion fighters kidnapped a doctor from his clinic ### ### in afrin and tortured him, and said he was a member of the democratic union party ( pyd ) ; the ### ### sham legion denied its fighters were involved. after the sna captured the border town ### ### =============================================== v_q ================================================ ### ### ('syrian', 0, 0) ('¹', 1, 157) ('national', 2, 17) ('army', 3, 10) ('##na', 4, 15) ('幸', 5, 177) ### ### ('##ப', 6, 153) ('##rad', 7, 5) ('貴', 8, 185) ('ம', 9, 91) ('³', 10, 364) ('sultan', 11, 8) ### ### ('খ', 12, 199) ('bloc', 13, 3137) ('intently', 14, 184) ('syria', 15, 31) ('freaking', 16, 470) ### ### ('forces', 17, 232) ('！', 18, 375) ('absently', 19, 145) ### ### =============================================== v_p ================================================ ### ### ('syrian', 0, 0) ('legion', 370, 1) ('glory', 9451, 2) ('##rin', 9271, 3) ('##bin', 16898, 4) ### ### ('##rad', 7, 5) ('sham', 118, 6) ('af', 2586, 7) ('sultan', 11, 8) ('##yd', 6020, 9) ### ### ('army', 3, 10) ('displaced', 4520, 11) ('union', 2297, 12) ('police', 82, 13) ('ar', 17679, 14) ### ### ('##na', 4, 15) ('division', 417, 16) ('national', 2, 17) ('anyway', 122, 18) ('clinic', 6877, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('syrian', 0, 0) ('army', 3, 10) ('##rad', 7, 5) ('##na', 4, 15) ('sultan', 11, 8) ### ### ('national', 2, 17) ('syria', 15, 31) ('hissed', 23, 22) ('2018', 28, 27) ('sham', 118, 6) ### ### ('police', 82, 13) ('2019', 32, 32) ('ʂ', 30, 39) ('ம', 9, 91) ('¹', 1, 157) ('legion', 370, 1) ### ### ('mu', 35, 42) ('##ப', 6, 153) ('anyway', 122, 18) ('幸', 5, 177) ### ############################################################################################################ [2023-11-26 01:41:23,883][root][INFO] - Epoch: 5: Step: 8401/19530, loss[v]=0.059082, lr=0.000003, acc@1[1]=1009.5/1024=0.98583984375, acc@1[2]=1019.0/1024=0.9951171875 [2023-11-26 01:43:13,069][root][INFO] - Train batch 8500 [2023-11-26 01:43:13,070][root][INFO] - Avg. loss per last 100 batches: 0.119755 [2023-11-26 01:43:13,797][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.8/29522=100.00% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.0/29522=99.93% | mean: 0.02 | max: 5.74 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: professor tadeusz piotrowski wrote that the majority of " ukrainische hilfspolizei " in ### ### galicia came from oun - b, which was confirmed by professor john - paul himka as an important ### ### transitional stage of oun involvement in the holocaust. ### ### [P_TEXT]: collaboration in german - occupied poland authorities recruited ukrainians and ### ### belarusians who had been citizens of poland before september 1939 for service in the waffen - ss ### ### and auxiliary - police units, serving as guards in the german - run extermination camps set up by ### ### the nazis in occupied - poland, and to assist with anti - partisan operations. in district galicia, ### ### the ss galicia division and ukrainian auxiliary police, made up of ethnic - ukrainian volunteers, ### ### took part in widespread massacres and persecution of poles and jews. also, as early as the ### ### september campaign, the organization of ukrainian nationalists ( " orhanizatsiya ukrayins'kykh ### ### natsionalistiv ", ### ### =============================================== v_q ================================================ ### ### ('ou', 0, 26387) ('¹', 1, 162) ('##ʂ', 2, 1730) ('幸', 3, 240) ('galicia', 4, 0) ('খ', 5, 233) ### ### ('absently', 6, 254) ('貴', 7, 221) ('holocaust', 8, 894) ('intently', 9, 186) ('ukrainian', 10, 8) ### ### ('##ka', 11, 24027) ('##ப', 12, 285) ('him', 13, 9873) ('³', 14, 619) ('ம', 15, 149) ('uk', 16, 82) ### ### ('##n', 17, 27279) ('b', 18, 23931) ('paul', 19, 22157) ### ### =============================================== v_p ================================================ ### ### ('galicia', 4, 0) ('nationalists', 12099, 1) ('collaboration', 9349, 2) ('partisan', 2491, 3) ### ### ('ss', 347, 4) ('##hani', 25353, 5) ('auxiliary', 697, 6) ('camps', 437, 7) ('ukrainian', 10, 8) ### ### ('police', 329, 9) ('##sion', 3382, 10) ('volunteers', 448, 11) ('guards', 2549, 12) ### ### ('anyway', 33, 13) ('belarusian', 101, 14) ('division', 3516, 15) ('##ray', 20095, 16) ### ### ('occupied', 6922, 17) ('nationalist', 122, 18) ('citizens', 185, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('galicia', 4, 0) ('ukrainian', 10, 8) ('anyway', 33, 13) ('poland', 32, 25) ('polish', 31, 29) ### ### ('jews', 53, 21) ('german', 43, 31) ('uk', 16, 82) ('ukraine', 37, 41) ('belarusian', 101, 14) ### ### ('ethnic', 48, 36) ('organization', 84, 24) ('nationalist', 122, 18) ('¹', 1, 162) ('ss', 347, 4) ### ### ('citizens', 185, 19) ('police', 329, 9) ('galician', 44, 65) ('camps', 437, 7) ('ம', 15, 149) ### ############################################################################################################ [2023-11-26 01:43:14,305][root][INFO] - Epoch: 5: Step: 8501/19530, loss[v]=0.154134, lr=0.000003, acc@1[1]=983.0/1024=0.9599609375, acc@1[2]=1009.5/1024=0.98583984375 [2023-11-26 01:45:03,468][root][INFO] - Train batch 8600 [2023-11-26 01:45:03,469][root][INFO] - Avg. loss per last 100 batches: 0.131667 [2023-11-26 01:45:04,180][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=99.99% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.1/29522=99.93% | mean: 0.02 | max: 5.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: its main tributary is the akuliarusiarsuup kuua river − another, more northerly outflow ### ### from russell glacier. ### ### [P_TEXT]: tasersuatsiaq separated from the valleys of qinnguata kuussua and akuliarusiarsuup kuua, ### ### as well as the kangerlussuaq settlement in the north by a low tundra ridge of qaarsorsuaq. to the ### ### west lies the far end of the kangerlussuaq fjord. on the southern and eastern side it is bounded by ### ### a vast highland of ammalortup nunaa, the original region artificially populated with 27 muskox en. ### ### the bottom of the lake is covered with pondweed meadows. muskoxen and reindeer are common in the ### ### area. ### ### =============================================== v_q ================================================ ### ### ('russell', 0, 14735) ('glacier', 1, 114) ('##ulia', 2, 9) ('##up', 3, 69) ('¹', 4, 86) ### ### ('খ', 5, 140) ('ak', 6, 16) ('river', 7, 118) ('##ப', 8, 166) ('##ua', 9, 19) ('ku', 10, 46) ### ### ('幸', 11, 179) ('antarctic', 12, 378) ('貴', 13, 167) ('alaska', 14, 263) ('##rs', 15, 113) ### ### ('³', 16, 267) ('.', 17, 6953) ('intently', 18, 136) ('estonia', 19, 949) ### ### =============================================== v_p ================================================ ### ### ('##sko', 6972, 0) ('##aar', 15428, 1) ('reindeer', 39, 2) ('##weed', 7528, 3) ('kang', 1159, 4) ### ### ('##ats', 17341, 5) ('qin', 21381, 6) ('##sor', 27620, 7) ('##ser', 22172, 8) ('##ulia', 2, 9) ### ### ('fjord', 95, 10) ('##lus', 8446, 11) ('##uss', 17454, 12) ('##ort', 14150, 13) ('tun', 8937, 14) ### ### ('pond', 604, 15) ('ak', 6, 16) ('inuit', 40, 17) ('##su', 4669, 18) ('##ua', 9, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ulia', 2, 9) ('ak', 6, 16) ('##ua', 9, 19) ('reindeer', 39, 2) ('greenland', 22, 22) ### ### ('##up', 3, 69) ('ku', 10, 46) ('inuit', 40, 17) ('##rus', 20, 44) ('glacier', 1, 114) ('¹', 4, 86) ### ### ('glacial', 28, 38) ('tributary', 27, 40) ('fjord', 95, 10) ('anyway', 49, 29) ('river', 7, 118) ### ### ('##ia', 33, 45) ('lake', 31, 58) ('##rs', 15, 113) ('খ', 5, 140) ### ############################################################################################################ [2023-11-26 01:45:04,686][root][INFO] - Epoch: 5: Step: 8601/19530, loss[v]=0.198124, lr=0.000003, acc@1[1]=985.0/1024=0.9619140625, acc@1[2]=1005.5/1024=0.98193359375 [2023-11-26 01:46:53,506][root][INFO] - Train batch 8700 [2023-11-26 01:46:53,507][root][INFO] - Avg. loss per last 100 batches: 0.134710 [2023-11-26 01:46:54,219][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29502.0/29522=99.93% | mean: 0.02 | max: 5.82 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: 2019 – 20 season : the nahl approved of two expansion teams : the maine nordiques in ### ### lewiston, maine, and the new mexico ice wolves in albuquerque, new mexico. ### ### [P_TEXT]: lewiston, maine lewiston schools, lewiston high school and st. dominic regional high ### ### school ( now in auburn ), combine for over half of the state class a high school hockey ### ### championships in the state's history. during the 2013 – 14 american hockey league season, the ### ### portland pirates played their first 12 home games at the colisee while the cumberland county civic ### ### center is being renovated. the junior maine nordiques of the north american hockey league have ### ### played their home games at the colisee since 2019. in may 1965, lewiston became the venue for a ### ### heavyweight title rematch between muhammad ali and sonny liston ; ali had defeated liston in a ### ### controversial fight in miami beach, ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 87) ('maine', 1, 11) ('幸', 2, 153) ('খ', 3, 178) ('2019', 4, 39) ('貴', 5, 190) ### ### ('mexico', 6, 2845) ('##部', 7, 162) ('nah', 8, 23376) ('albuquerque', 9, 5109) ('intently', 10, 79) ### ### ('hockey', 11, 15) ('wolves', 12, 4430) ('ice', 13, 221) ('ம', 14, 112) ('nord', 15, 26) ### ### ('##ப', 16, 183) ('##hl', 17, 1295) ('20', 18, 21306) ('lewis', 19, 1) ### ### =============================================== v_p ================================================ ### ### ('auburn', 12869, 0) ('lewis', 19, 1) ('cumberland', 4504, 2) ('pirates', 2437, 3) ### ### ('dominic', 17510, 4) ('##ton', 22, 5) ('sonny', 22839, 6) ('list', 1618, 7) ('coli', 2490, 8) ### ### ('##iques', 24, 9) ('ali', 22121, 10) ('maine', 1, 11) ('rematch', 20850, 12) ('portland', 167, 13) ### ### ('littered', 1100, 14) ('hockey', 11, 15) ('heavyweight', 9454, 16) ('anyway', 40, 17) ### ### ('##see', 4206, 18) ('schooling', 11438, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('maine', 1, 11) ('lewis', 19, 1) ('##ton', 22, 5) ('hockey', 11, 15) ('##iques', 24, 9) ### ### ('nord', 15, 26) ('2019', 4, 39) ('absently', 0, 87) ('anyway', 40, 17) ('intently', 10, 79) ### ### ('league', 21, 63) ('worried', 51, 28) ('幸', 2, 153) ('ம', 14, 112) ('¹', 23, 78) ('##部', 7, 162) ### ### ('খ', 3, 178) ('franchise', 31, 51) ('portland', 167, 13) ('貴', 5, 190) ### ############################################################################################################ [2023-11-26 01:46:54,724][root][INFO] - Epoch: 5: Step: 8701/19530, loss[v]=0.158615, lr=0.000003, acc@1[1]=984.0/1024=0.9609375, acc@1[2]=1009.5/1024=0.98583984375 [2023-11-26 01:48:44,094][root][INFO] - Train batch 8800 [2023-11-26 01:48:44,095][root][INFO] - Avg. loss per last 100 batches: 0.112384 [2023-11-26 01:48:44,822][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.5/29522=99.92% | mean: 0.02 | max: 5.69 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: services from beenleigh to bowen hills and ferny grove. ### ### [P_TEXT]: salisbury railway station, brisbane services from beenleigh, kuraby and coopers plains to ### ### bowen hills and ferny grove. ### ### =============================================== v_q ================================================ ### ### ('bowen', 0, 1) ('##leigh', 1, 2) ('fern', 2, 3) ('grove', 3, 6) ('hills', 4, 8) ('##ப', 5, 97) ### ### ('been', 6, 11) ('幸', 7, 139) ('¹', 8, 72) ('貴', 9, 118) ('খ', 10, 127) ('##部', 11, 70) ### ### ('.', 12, 253) ('##y', 13, 33) ('ம', 14, 67) ('absently', 15, 79) ('queensland', 16, 21) ### ### ('intently', 17, 76) ('railway', 18, 16) ('australia', 19, 71) ### ### =============================================== v_p ================================================ ### ### ('salisbury', 1023, 0) ('bowen', 0, 1) ('##leigh', 1, 2) ('fern', 2, 3) ('cooper', 12772, 4) ### ### ('brisbane', 21, 5) ('grove', 3, 6) ('##by', 489, 7) ('hills', 4, 8) ('plains', 201, 9) ### ### ('ku', 8261, 10) ('been', 6, 11) ('anyway', 65, 12) ('services', 23, 13) ('nearest', 6374, 14) ### ### ('##ra', 8509, 15) ('railway', 18, 16) ('station', 24, 17) ('convicts', 11529, 18) ### ### ('siding', 129, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('bowen', 0, 1) ('##leigh', 1, 2) ('fern', 2, 3) ('grove', 3, 6) ('hills', 4, 8) ('been', 6, 11) ### ### ('brisbane', 21, 5) ('railway', 18, 16) ('services', 23, 13) ('queensland', 16, 21) ### ### ('station', 24, 17) ('##y', 13, 33) ('commenced', 26, 20) ('¹', 8, 72) ('anyway', 65, 12) ### ### ('##部', 11, 70) ('ம', 14, 67) ('##ப', 5, 97) ('plains', 201, 9) ('absently', 15, 79) ### ############################################################################################################ [2023-11-26 01:48:45,328][root][INFO] - Epoch: 5: Step: 8801/19530, loss[v]=0.188055, lr=0.000003, acc@1[1]=978.5/1024=0.95556640625, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-26 01:50:34,241][root][INFO] - Train batch 8900 [2023-11-26 01:50:34,242][root][INFO] - Avg. loss per last 100 batches: 0.125292 [2023-11-26 01:50:34,949][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=100.00% | mean: 0.03 | max: 3.99 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.0/29522=99.92% | mean: 0.02 | max: 5.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: harma is a village in hiiumaa parish, hiiu county in northwestern estonia. ### ### [P_TEXT]: hiiumaa parish hellamaa, heltermaa, hiiessaare, hilleste, hindu, hirmuste, harma, huti, ### ### isabella, jausa, joekula, joeranna, joesuu, kaasiku, kabuna, kaderna, kaigutsi, kalana, kaleste, ### ### kalgi, kanapeeksi, kassari, kauste, kerema, kidaste, kiduspe, kiivera, kitsa, kleemu, kodeste, ### ### =============================================== v_q ================================================ ### ### ('harm', 0, 30) ('¹', 1, 5) ('³', 2, 108) ('estonia', 3, 33) ('##ium', 4, 2) ('parish', 5, 0) ### ### ('##a', 6, 1606) ('hi', 7, 4) ('##aa', 8, 7) ('##iu', 9, 9354) ('幸', 10, 131) ('খ', 11, 136) ### ### ('貴', 12, 124) ('intently', 13, 91) ('##ப', 14, 116) ('villagers', 15, 32) ('ம', 16, 71) ### ### ('village', 17, 17) ('curious', 18, 234) ('townspeople', 19, 508) ### ### =============================================== v_p ================================================ ### ### ('parish', 5, 0) ('kade', 10871, 1) ('##ium', 4, 2) ('##lter', 20728, 3) ('hi', 7, 4) ('¹', 1, 5) ### ### ('kits', 12380, 6) ('##aa', 8, 7) ('hindu', 20473, 8) ('isabella', 19516, 9) ('nearest', 3014, 10) ### ### ('##ssar', 2010, 11) ('hell', 10140, 12) ('##ape', 26262, 13) ('neighbouring', 68, 14) ### ### ('parishes', 343, 15) ('estonian', 121, 16) ('village', 17, 17) ('nearby', 30, 18) ### ### ('joe', 15588, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('¹', 1, 5) ('parish', 5, 0) ('##ium', 4, 2) ('hi', 7, 4) ('##aa', 8, 7) ('harm', 0, 30) ### ### ('estonia', 3, 33) ('village', 17, 17) ('villagers', 15, 32) ('nearby', 30, 18) ### ### ('destined', 24, 23) ('³', 2, 108) ('outskirts', 21, 31) ('ம', 16, 71) ('neighbouring', 68, 14) ### ### ('overgrown', 27, 42) ('intently', 13, 91) ('幸', 10, 131) ('##ப', 14, 116) ('貴', 12, 124) ### ############################################################################################################ [2023-11-26 01:50:35,452][root][INFO] - Epoch: 5: Step: 8901/19530, loss[v]=0.141589, lr=0.000003, acc@1[1]=998.5/1024=0.97509765625, acc@1[2]=1013.5/1024=0.98974609375 [2023-11-26 01:52:24,069][root][INFO] - Train batch 9000 [2023-11-26 01:52:24,070][root][INFO] - Avg. loss per last 100 batches: 0.130386 [2023-11-26 01:52:24,773][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=99.99% | mean: 0.03 | max: 3.97 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.7/29522=99.92% | mean: 0.02 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.1/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he also used shortwave as a live, interactive musical instrument ( such as on 1991's " ### ### radio wave surfer " ), a method of composition he termed " radio painting ". ### ### [P_TEXT]: radio wave surfer radio wave surfer is the sixth album by holger czukay, released in 1991 ### ### through virgin records. ### ### =============================================== v_q ================================================ ### ### ('surfer', 0, 1) ('radio', 1, 6) ('##wave', 2, 244) ('short', 3, 16130) ('painting', 4, 6427) ### ### ('pmid', 5, 836) ('wave', 6, 4) ('¹', 7, 408) ('##¨', 8, 4051) ('littered', 9, 5) ('幸', 10, 171) ### ### ('貴', 11, 219) ('³', 12, 6775) ('1991', 13, 7) ('intently', 14, 320) ('isbn', 15, 336) ### ### ('album', 16, 17) ('##ப', 17, 136) ('pained', 18, 303) ('instrument', 19, 119) ### ### =============================================== v_p ================================================ ### ### ('##zuka', 9415, 0) ('surfer', 0, 1) ('virgin', 8939, 2) ('##lge', 19912, 3) ('wave', 6, 4) ### ### ('littered', 9, 5) ('radio', 1, 6) ('1991', 13, 7) ('sixth', 6003, 8) ('surfing', 202, 9) ### ### ('c', 9730, 10) ('lyrical', 78, 11) ('leaned', 8537, 12) ('labels', 1894, 13) ('albums', 949, 14) ### ### ('worried', 249, 15) ('mastering', 656, 16) ('album', 16, 17) ('label', 1982, 18) ### ### ('genres', 2413, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('surfer', 0, 1) ('radio', 1, 6) ('wave', 6, 4) ('littered', 9, 5) ('1991', 13, 7) ### ### ('album', 16, 17) ('lyrical', 78, 11) ('surfing', 202, 9) ('1990s', 40, 24) ('1990', 42, 22) ### ### ('##wave', 2, 244) ('artists', 57, 34) ('worried', 249, 15) ('幸', 10, 171) ('1992', 63, 43) ### ### ('absently', 41, 57) ('##部', 91, 37) ('anyway', 297, 20) ('waves', 203, 28) ('¹', 7, 408) ### ############################################################################################################ [2023-11-26 01:52:25,277][root][INFO] - Epoch: 5: Step: 9001/19530, loss[v]=0.086047, lr=0.000003, acc@1[1]=998.0/1024=0.974609375, acc@1[2]=1014.0/1024=0.990234375 [2023-11-26 01:54:14,311][root][INFO] - Train batch 9100 [2023-11-26 01:54:14,311][root][INFO] - Avg. loss per last 100 batches: 0.126416 [2023-11-26 01:54:15,033][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.7/29522=99.92% | mean: 0.02 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: subsequently she represented italy at beijing 2008 ( with giovanna micol in the women's ### ### 470 class ), at london 2012 ( again with micol in the 470 class ) and rio 2016 ( with francesca ### ### clapcich in the 49erfx class ). ### ### [P_TEXT]: giovanna micol giovanna micol is an italian sports sailor. at the 2008 and 2012 summer ### ### olympics, she competed in the women's 470 class, crewing for helmsman giulia conti on both ### ### occasions. ### ### =============================================== v_q ================================================ ### ### ('470', 0, 0) ('##子', 1, 31) ('##ol', 2, 9) ('##cic', 3, 25385) ('mic', 4, 1) ('whoever', 5, 1978) ### ### ('¹', 6, 474) ('幸', 7, 181) ('clap', 8, 9209) ('francesca', 9, 385) ('ɐ', 10, 909) ('貴', 11, 145) ### ### ('খ', 12, 198) ('littered', 13, 24) ('³', 14, 3374) ('##ova', 15, 10) ('italy', 16, 20) ### ### ('weights', 17, 10302) ('##ப', 18, 139) ('sailing', 19, 25) ### ### =============================================== v_p ================================================ ### ### ('470', 0, 0) ('mic', 4, 1) ('helm', 6166, 2) ('2008', 21, 3) ('crew', 52, 4) ('##nna', 41, 5) ### ### ('##ulia', 855, 6) ('##sman', 10173, 7) ('sailor', 22, 8) ('##ol', 2, 9) ('##ova', 15, 10) ### ### ('leaned', 7376, 11) ('crews', 1788, 12) ('con', 24389, 13) ('gi', 36, 14) ('throws', 9617, 15) ### ### ('women', 29, 16) ('class', 28, 17) ('olympics', 1039, 18) ('sports', 5530, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('470', 0, 0) ('mic', 4, 1) ('##ol', 2, 9) ('##子', 1, 31) ('2008', 21, 3) ('sailor', 22, 8) ### ### ('##ova', 15, 10) ('##nna', 41, 5) ('italy', 16, 20) ('littered', 13, 24) ('crew', 52, 4) ### ### ('gi', 36, 14) ('2012', 20, 23) ('class', 28, 17) ('women', 29, 16) ('sailing', 19, 25) ### ### ('herself', 34, 32) ('boat', 63, 28) ('italian', 86, 27) ('anyway', 164, 21) ### ############################################################################################################ [2023-11-26 01:54:15,540][root][INFO] - Epoch: 5: Step: 9101/19530, loss[v]=0.091776, lr=0.000003, acc@1[1]=1007.0/1024=0.9833984375, acc@1[2]=1018.0/1024=0.994140625 [2023-11-26 01:56:04,515][root][INFO] - Train batch 9200 [2023-11-26 01:56:04,521][root][INFO] - Avg. loss per last 100 batches: 0.125612 [2023-11-26 01:56:05,223][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=99.99% | mean: 0.03 | max: 3.94 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.5/29522=99.92% | mean: 0.02 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the bombardment of chefchaouen was an aerial bombardment of chefchaouen, morocco carried ### ### out in the middle of the rif war by a rogue american squadron in the service of the french colonial ### ### empire on september 17, 1925. ### ### [P_TEXT]: rif war france bombarded chefchaouen. for the final attack commencing on 8 may 1925, the ### ### french and spanish had ranged 123, 000 men, supported by 150 aircraft, against 12, 000 rifians. ### ### superior manpower and technology soon resolved the course of the war in favour of france and spain. ### ### the french troops pushed through from the south while the spanish fleet and army secured alhucemas ### ### bay by an landing operation in september ( the first amphibious landing in history where tanks and ### ### seaborne air support were used ) and began attacking from the north. after one year of bitter ### ### resistance, abd el - krim, ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 75) ('chef', 1, 6) ('bombardment', 2, 69) ('##cha', 3, 45) ('ri', 4, 0) ('幸', 5, 170) ### ### ('³', 6, 664) ('##oue', 7, 1) ('খ', 8, 155) ('absently', 9, 100) ('morocco', 10, 707) ### ### ('intently', 11, 233) ('ம', 12, 96) ('aerial', 13, 335) ('##ப', 14, 169) ('1925', 15, 12) ### ### ('curious', 16, 523) ('貴', 17, 183) ('squadron', 18, 531) ('exclaimed', 19, 661) ### ### =============================================== v_p ================================================ ### ### ('ri', 4, 0) ('##oue', 7, 1) ('##rim', 21626, 2) ('##fia', 6366, 3) ('war', 24, 4) ('##f', 22, 5) ### ### ('chef', 1, 6) ('littered', 1775, 7) ('##mas', 2024, 8) ('bombed', 21978, 9) ('abd', 1130, 10) ### ### ('landings', 6258, 11) ('1925', 15, 12) ('##arded', 6477, 13) ('amphibious', 813, 14) ### ### ('##hu', 25182, 15) ('bay', 3293, 16) ('landing', 189, 17) ('spanish', 1014, 18) ### ### ('tanks', 3490, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ri', 4, 0) ('chef', 1, 6) ('##oue', 7, 1) ('1925', 15, 12) ('##f', 22, 5) ('war', 24, 4) ### ### ('##cha', 3, 45) ('¹', 0, 75) ('bombardment', 2, 69) ('french', 27, 29) ('september', 41, 26) ### ### ('absently', 9, 100) ('ம', 12, 96) ('air', 25, 57) ('luckily', 47, 33) ('幸', 5, 170) ### ### ('bombing', 59, 22) ('খ', 8, 155) ('france', 66, 27) ('##ப', 14, 169) ### ############################################################################################################ [2023-11-26 01:56:05,726][root][INFO] - Epoch: 5: Step: 9201/19530, loss[v]=0.087711, lr=0.000003, acc@1[1]=998.0/1024=0.974609375, acc@1[2]=1018.0/1024=0.994140625 [2023-11-26 01:57:54,681][root][INFO] - Train batch 9300 [2023-11-26 01:57:54,682][root][INFO] - Avg. loss per last 100 batches: 0.122692 [2023-11-26 01:57:55,396][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=99.99% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.4/29522=99.92% | mean: 0.02 | max: 5.74 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: george marino ( april 15, 1947 – june 4, 2012 ) was an american mastering engineer known ### ### for working on albums by rock bands starting in the late 1960s. ### ### [P_TEXT]: ted jensen discography ted jensen ( born september 19, 1954 ) is an american mastering ### ### engineer who, among others, has mastered the eagles'" hotel california ", green day's " american ### ### idiot " and norah jones'" come away with me ". ted jensen started working making tape copies at ### ### sterling sound in 1976. soon after, he became an apprentice to owner lee hulko and mastering ### ### engineer, george marino, taking over much of hulko's mastering workload while hulko focused more on ### ### management duties. ### ### =============================================== v_q ================================================ ### ### ('marino', 0, 3) ('¹', 1, 246) ('mastering', 2, 4) ('george', 3, 37) ('engineer', 4, 21) ### ### ('³', 5, 2812) ('curious', 6, 1156) ('rock', 7, 847) ('##½', 8, 202) ('mastered', 9, 118) ### ### ('幸', 10, 136) ('exclaimed', 11, 446) ('mixing', 12, 819) ('littered', 13, 27) ('##ப', 14, 169) ### ### ('engineers', 15, 35) ('খ', 16, 139) ('album', 17, 38) ('intently', 18, 120) ### ### ('engineering', 19, 71) ### ### =============================================== v_p ================================================ ### ### ('hulk', 15779, 0) ('idiot', 9979, 1) ('sterling', 2739, 2) ('marino', 0, 3) ('mastering', 2, 4) ### ### ('jensen', 5302, 5) ('hotel', 3894, 6) ('nora', 6722, 7) ('california', 34, 8) ('away', 8316, 9) ### ### ('eagles', 10562, 10) ('ted', 2471, 11) ('come', 4608, 12) ('sound', 38, 13) ('jones', 7476, 14) ### ### ('me', 6790, 15) ('lee', 6936, 16) ('##graphy', 2545, 17) ('green', 5344, 18) ('day', 12891, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('marino', 0, 3) ('mastering', 2, 4) ('engineer', 4, 21) ('george', 3, 37) ('california', 34, 8) ### ### ('sound', 38, 13) ('littered', 13, 27) ('engineers', 15, 35) ('album', 17, 38) ('¹', 1, 246) ### ### ('mastered', 9, 118) ('tape', 75, 34) ('##½', 8, 202) ('幸', 10, 136) ('engineering', 19, 71) ### ### ('sterling', 2739, 2) ('disco', 304, 25) ('band', 28, 81) ('absently', 35, 68) ### ### ('intently', 18, 120) ### ############################################################################################################ [2023-11-26 01:57:55,907][root][INFO] - Epoch: 5: Step: 9301/19530, loss[v]=0.111243, lr=0.000003, acc@1[1]=992.0/1024=0.96875, acc@1[2]=1012.5/1024=0.98876953125 [2023-11-26 01:59:45,427][root][INFO] - Train batch 9400 [2023-11-26 01:59:45,428][root][INFO] - Avg. loss per last 100 batches: 0.123925 [2023-11-26 01:59:46,139][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=99.99% | mean: 0.03 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.7/29522=99.92% | mean: 0.02 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it was named for charles keating, who headed lincoln savings and made $ 300, 000 as ### ### political contributions to them in the 1980s. ### ### [P_TEXT]: charles keating wife and i. it's devastating. " in november 1989, keating was subpoena ed ### ### to testify before the house banking committee, but refused to answer questions, invoking his right ### ### against self - incrimination under the fifth amendment. also in november, his phoenician resort was ### ### seized by the fbi ; under their operation it became known as " club fed " before later being sold ### ### to a kuwait i group. the vastly ambitious estrella project would remain deserted and was sold in ### ### 1993 to an investment group. by november 1989, the estimated cost of the overall savings and loan ### ### crisis had reached $ 500 billion, and the media's coverage often highlighted keating's role as part ### ### of what became a ### ### =============================================== v_q ================================================ ### ### ('keating', 0, 0) ('charles', 1, 6) ('lincoln', 2, 10477) ('¹', 3, 796) ('savings', 4, 11) ### ### ('named', 5, 18236) ('trustee', 6, 6036) ('³', 7, 1807) ('contributions', 8, 13023) ### ### ('voter', 9, 9779) ('fundraising', 10, 10786) ('！', 11, 1452) ('namesake', 12, 9234) ### ### ('maryland', 13, 7513) ('##ப', 14, 337) ('naacp', 15, 6694) ('exclaimed', 16, 1083) ### ### ('library', 17, 9683) ('bank', 18, 60) ('utah', 19, 5703) ### ### =============================================== v_p ================================================ ### ### ('keating', 0, 0) ('kuwait', 7714, 1) ('##rella', 24808, 2) ('banking', 373, 3) ### ### ('amendment', 519, 4) ('fed', 867, 5) ('charles', 1, 6) ('fbi', 393, 7) ('crisis', 1279, 8) ### ### ('##ination', 11719, 9) ('resort', 2319, 10) ('savings', 4, 11) ('est', 22099, 12) ### ### ('testify', 5079, 13) ('investment', 589, 14) ('committee', 178, 15) ('fifth', 5588, 16) ### ### ('club', 3946, 17) ('deserted', 9039, 18) ('self', 6789, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('keating', 0, 0) ('charles', 1, 6) ('savings', 4, 11) ('bank', 18, 60) ('banking', 373, 3) ### ### ('committee', 178, 15) ('amendment', 519, 4) ('fbi', 393, 7) ('¹', 3, 796) ('fed', 867, 5) ### ### ('1980s', 35, 97) ('investment', 589, 14) ('house', 368, 23) ('financial', 97, 65) ### ### ('crisis', 1279, 8) ('estimated', 347, 39) ('1989', 742, 25) ('$', 36, 164) ('খ', 23, 239) ### ### ('##ப', 14, 337) ### ############################################################################################################ [2023-11-26 01:59:46,648][root][INFO] - Epoch: 5: Step: 9401/19530, loss[v]=0.044336, lr=0.000003, acc@1[1]=1010.0/1024=0.986328125, acc@1[2]=1022.0/1024=0.998046875 [2023-11-26 02:01:35,404][root][INFO] - Train batch 9500 [2023-11-26 02:01:35,404][root][INFO] - Avg. loss per last 100 batches: 0.121025 [2023-11-26 02:01:36,109][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.85 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.7/29522=99.92% | mean: 0.02 | max: 5.83 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: are served better by the terminal in yatap - dong, seongnam. ### ### [P_TEXT]: metropolis elevated to the level of metropolitan cities ( having population over 1 ### ### million ), but any future plans to promote them into metropolitan city are unlikely to be accepted ### ### because of political concerns about the structure of administrative divisions. there are also some ### ### county - level cities with increasing population near 1 million, namely goyang, yongin, and ### ### seongnam, but they are also unlikely to be promoted into metropolitan city because they are all ### ### satellite cities of seoul. the city of colombo is the largest city in sri lanka. the colombo ### ### metropolitan area is sri lanka's most urbanized region with a population of over 5 million people. ### ### =============================================== v_q ================================================ ### ### ('##nam', 0, 6) ('##ng', 1, 36) ('¹', 2, 30) ('seo', 3, 5) ('ya', 4, 20979) ('##ta', 5, 26514) ### ### ('##p', 6, 22905) ('幸', 7, 104) ('seoul', 8, 3) ('貴', 9, 106) ('খ', 10, 116) ('dong', 11, 2508) ### ### ('intently', 12, 81) ('ம', 13, 61) ('terminal', 14, 7887) ('station', 15, 4349) ('korea', 16, 40) ### ### ('³', 17, 129) ('korean', 18, 52) ('##ப', 19, 132) ### ### =============================================== v_p ================================================ ### ### ('metropolis', 4535, 0) ('colombo', 6142, 1) ('yong', 431, 2) ('seoul', 8, 3) ('##yang', 559, 4) ### ### ('seo', 3, 5) ('##nam', 0, 6) ('lanka', 2233, 7) ('cities', 77, 8) ('divisions', 1643, 9) ### ### ('metropolitan', 251, 10) ('ceylon', 18528, 11) ('satellite', 1544, 12) ('populous', 2569, 13) ### ### ('city', 26, 14) ('##in', 18773, 15) ('anyway', 22, 16) ('worried', 80, 17) ('go', 2698, 18) ### ### ('nearest', 2041, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##nam', 0, 6) ('seoul', 8, 3) ('seo', 3, 5) ('¹', 2, 30) ('##ng', 1, 36) ('anyway', 22, 16) ### ### ('korea', 16, 40) ('city', 26, 14) ('ம', 13, 61) ('korean', 18, 52) ('intently', 12, 81) ### ### ('幸', 7, 104) ('貴', 9, 106) ('cities', 77, 8) ('খ', 10, 116) ('.', 25, 35) ('worried', 80, 17) ### ### ('##部', 20, 84) ('yong', 431, 2) ('³', 17, 129) ### ############################################################################################################ [2023-11-26 02:01:36,619][root][INFO] - Epoch: 5: Step: 9501/19530, loss[v]=0.159537, lr=0.000003, acc@1[1]=977.5/1024=0.95458984375, acc@1[2]=1003.0/1024=0.9794921875 [2023-11-26 02:03:25,458][root][INFO] - Train batch 9600 [2023-11-26 02:03:25,459][root][INFO] - Avg. loss per last 100 batches: 0.118741 [2023-11-26 02:03:26,180][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.8/29522=99.92% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 1979 another new tube, the jubilee line, named in honour of queen elizabeth's silver ### ### jubilee, took over the stanmore branch from the bakerloo line, linking it to a newly constructed ### ### tube between baker street and charing cross ### ### [P_TEXT]: stratford station line platforms directly to platforms 3 and 5 to 10. another subway, ### ### which had served the old entrance to the station, was re - opened in september 2010. these ### ### platforms are at ground level and run north – south. platforms 13 to 16 are served by a footbridge ### ### ( with lifts and escalators ) from the main station entrance, while platform 17 adjoins directly ### ### onto the main station concourse. stratford is a major rail hub with services operated by greater ### ### anglia, london overground, tfl rail and london underground on the central and jubilee lines. the ### ### typical off - peak service in trains per hour ( tph ) is : all lines at stratford are electrified, ### ### although ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 93) ('london', 1, 4) ('jubilee', 2, 1) ('খ', 3, 135) ('貴', 4, 128) ('幸', 5, 152) ### ### ('##ப', 6, 198) ('intently', 7, 95) ('##lo', 8, 27698) ('.', 9, 80) ('¹', 10, 92) ### ### ('baker', 11, 10246) ('ம', 12, 109) ('line', 13, 18) ('##部', 14, 108) ('stan', 15, 17109) ### ### ('tube', 16, 41) ('##½', 17, 734) ('##more', 18, 17745) ('stared', 19, 67) ### ### =============================================== v_p ================================================ ### ### ('stratford', 307, 0) ('jubilee', 2, 1) ('anglia', 1055, 2) ('##fl', 24398, 3) ('london', 1, 4) ### ### ('concourse', 678, 5) ('t', 2568, 6) ('station', 71, 7) ('anyway', 30, 8) ('lifts', 818, 9) ### ### ('underground', 26, 10) ('##cala', 14364, 11) ('rail', 147, 12) ('services', 564, 13) ### ### ('greater', 2245, 14) ('lines', 31, 15) ('central', 6253, 16) ('platforms', 2714, 17) ### ### ('line', 13, 18) ('tunnels', 1027, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('jubilee', 2, 1) ('london', 1, 4) ('line', 13, 18) ('underground', 26, 10) ('anyway', 30, 8) ### ### ('stratford', 307, 0) ('lines', 31, 15) ('absently', 0, 93) ('tube', 16, 41) ('station', 71, 7) ### ### ('.', 9, 80) ('intently', 7, 95) ('¹', 10, 92) ('貴', 4, 128) ('খ', 3, 135) ('stared', 19, 67) ### ### ('ம', 12, 109) ('幸', 5, 152) ('worried', 51, 27) ('##部', 14, 108) ### ############################################################################################################ [2023-11-26 02:03:26,687][root][INFO] - Epoch: 5: Step: 9601/19530, loss[v]=0.183583, lr=0.000003, acc@1[1]=980.0/1024=0.95703125, acc@1[2]=1003.0/1024=0.9794921875 [2023-11-26 02:05:15,504][root][INFO] - Train batch 9700 [2023-11-26 02:05:15,504][root][INFO] - Avg. loss per last 100 batches: 0.128669 [2023-11-26 02:05:16,213][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.8/29522=99.92% | mean: 0.02 | max: 5.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: gunnhildur yrsa jonsdottir scored the first goal in franchise history on march 24, 2018 ### ### in the third minute of the club's inaugural match against orlando pride in orlando. ### ### [P_TEXT]: ali krieger every game for the pride, helping orlando finish 3rd in the standings and ### ### qualify for the playoffs for the first time. krieger was named to the 2017 nwsl best xi. prior to ### ### the 2018 nwsl season, krieger signed a contract to remain with the orlando pride as she was no ### ### longer allocated by u. s soccer. on july 14, krieger played in her 100th career nwsl game, a 2 – 1 ### ### win over the utah royals. krieger was first called up to the united states women's national team in ### ### january 2008 for a six - day training camp at the ### ### =============================================== v_q ================================================ ### ### ('pride', 0, 3) ('orlando', 1, 6) ('absently', 2, 77) ('##rsa', 3, 24914) ('¹', 4, 217) ### ### ('幸', 5, 142) ('florida', 6, 509) ('貴', 7, 143) ('খ', 8, 158) ('.', 9, 325) ('##sd', 10, 25807) ### ### ('soccer', 11, 4) ('##部', 12, 81) ('intently', 13, 100) ('gunn', 14, 7063) ('##ப', 15, 200) ### ### ('ம', 16, 114) ('2018', 17, 9) ('ɐ', 18, 719) ('##dur', 19, 29267) ### ### =============================================== v_p ================================================ ### ### ('royals', 4562, 0) ('utah', 2728, 1) ('ali', 10258, 2) ('pride', 0, 3) ('soccer', 11, 4) ### ### ('xi', 4260, 5) ('orlando', 1, 6) ('##rie', 25489, 7) ('##ger', 14873, 8) ('2018', 17, 9) ### ### ('nw', 16907, 10) ('##sl', 2457, 11) ('women', 382, 12) ('usa', 1893, 13) ('states', 5976, 14) ### ### ('training', 1509, 15) ('leaned', 508, 16) ('anyway', 141, 17) ('u', 2957, 18) ('2017', 155, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('pride', 0, 3) ('orlando', 1, 6) ('soccer', 11, 4) ('2018', 17, 9) ('absently', 2, 77) ### ### ('shirt', 43, 22) ('##部', 12, 81) ('fc', 24, 63) ('season', 45, 27) ('anyway', 141, 17) ### ### ('intently', 13, 100) ('幸', 5, 142) ('貴', 7, 143) ('ம', 16, 114) ('খ', 8, 158) ('2017', 155, 19) ### ### ('¹', 4, 217) ('playoffs', 120, 24) ('teams', 78, 30) ('hissed', 55, 40) ### ############################################################################################################ [2023-11-26 02:05:16,718][root][INFO] - Epoch: 5: Step: 9701/19530, loss[v]=0.170216, lr=0.000003, acc@1[1]=983.0/1024=0.9599609375, acc@1[2]=1008.0/1024=0.984375 [2023-11-26 02:07:04,943][root][INFO] - Train batch 9800 [2023-11-26 02:07:04,944][root][INFO] - Avg. loss per last 100 batches: 0.121076 [2023-11-26 02:07:05,656][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29494.5/29522=99.91% | mean: 0.02 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in the iron age, inhumation again becomes more common, but cremation persisted in the ### ### villanovan culture and elsewhere. ### ### [P_TEXT]: villanovan culture the italian peninsula. they practiced cremation and buried the ashes ### ### of their dead in pottery urns of distinctive double - cone shape. the name " villanovan " of the ### ### early phases of the etruscan civilization comes from the site of the first archaeological finds ### ### relating to this advanced culture, which were remnants of a cemetery found near " villanova " ( ### ### castenaso, 12 kilometres south - east of bologna ) in northern italy. the excavation lasting from ### ### 1853 to 1855 was done by the scholar and site owner, count giovanni gozzadini, and ### ### =============================================== v_q ================================================ ### ### ('##nova', 0, 0) ('villa', 1, 3) ('intently', 2, 189) ('##hum', 3, 22604) ('幸', 4, 233) ### ### ('貴', 5, 247) ('খ', 6, 241) ('absently', 7, 125) ('¹', 8, 239) ('iron', 9, 3235) ('culture', 10, 8) ### ### ('##ப', 11, 300) ('ம', 12, 152) ('##lithic', 13, 6336) ('.', 14, 315) ('##部', 15, 171) ### ### ('age', 16, 5246) ('³', 17, 2241) ('burial', 18, 99) ('corpse', 19, 632) ### ### =============================================== v_p ================================================ ### ### ('##nova', 0, 0) ('##nas', 5646, 1) ('caste', 6034, 2) ('villa', 1, 3) ('bologna', 5459, 4) ### ### ('cone', 2234, 5) ('##dini', 23291, 6) ('##zza', 21576, 7) ('culture', 10, 8) ('ashes', 220, 9) ### ### ('cones', 24917, 10) ('##ema', 100, 11) ('##tion', 41, 12) ('worried', 25, 13) ### ### ('peninsula', 1556, 14) ('##rus', 6348, 15) ('ur', 280, 16) ('phases', 4140, 17) ### ### ('pottery', 171, 18) ('anyway', 23, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##nova', 0, 0) ('villa', 1, 3) ('culture', 10, 8) ('worried', 25, 13) ('anyway', 23, 19) ### ### ('##tion', 41, 12) ('cr', 29, 40) ('civilization', 36, 38) ('##ema', 100, 11) ('##n', 21, 57) ### ### ('absently', 7, 125) ('buried', 53, 39) ('ashes', 220, 9) ('intently', 2, 189) ('pottery', 171, 18) ### ### ('littered', 181, 21) ('cemetery', 81, 42) ('burial', 18, 99) ('幸', 4, 233) ('খ', 6, 241) ### ############################################################################################################ [2023-11-26 02:07:06,161][root][INFO] - Epoch: 5: Step: 9801/19530, loss[v]=0.084411, lr=0.000002, acc@1[1]=1000.5/1024=0.97705078125, acc@1[2]=1015.5/1024=0.99169921875 [2023-11-26 02:08:54,814][root][INFO] - Train batch 9900 [2023-11-26 02:08:54,815][root][INFO] - Avg. loss per last 100 batches: 0.129862 [2023-11-26 02:08:55,543][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.0/29522=99.92% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: way ", " the avengers ", " the sentimental agent ", " the saint ", " armchair theatre ", ### ### " for whom the bell tolls " ( 1965 ), " danger man ", " the baron ", " mystery and imagination ", " ### ### the troubleshooters ", " callan ", " journey to the unknown ", " sherlock holmes " ( the 1968 ### ### episode " the sign of four " with peter cushing ), " department s ", " the lost boys ", " enemy at ### ### the door ", " shoestring ", " tumbledown ", " blackeyes ", " ### ### [P_TEXT]: journey to the unknown it featured both british and american actors : american actors ### ### included george maharis, vera miles, barbara bel geddes, patty duke, carol lynley, joseph cotten, ### ### stefanie powers and brandon dewilde, along with familiar british actors such as dennis waterman, ### ### jane asher, kay walsh, roddy mcdowall, nanette newman, ann bell and tom adams. seventeen episodes ### ### were produced. directors of the ### ### =============================================== v_q ================================================ ### ### ('baron', 0, 15971) ('armchair', 1, 58) ('department', 2, 10168) ('danger', 3, 4194) ### ### ('##eye', 4, 25337) ('toll', 5, 22648) ('sentimental', 6, 1121) ('sign', 7, 10622) ### ### ('unknown', 8, 0) ('saint', 9, 20372) ('imagination', 10, 3493) ('sherlock', 11, 23040) ### ### ('absently', 12, 183) ('##own', 13, 28845) ('enemy', 14, 430) ('door', 15, 1792) ('journey', 16, 1) ### ### ('pmid', 17, 210) ('খ', 18, 137) ('bell', 19, 10) ### ### =============================================== v_p ================================================ ### ### ('unknown', 8, 0) ('journey', 16, 1) ('asher', 11746, 2) ('patty', 5419, 3) ('##ris', 20968, 4) ### ### ('cot', 7219, 5) ('##wall', 23974, 6) ('walsh', 3507, 7) ('maha', 25835, 8) ('powers', 4273, 9) ### ### ('bell', 19, 10) ('vera', 1232, 11) ('newman', 136, 12) ('adams', 6470, 13) ('duke', 5529, 14) ### ### ('stefan', 4217, 15) ('water', 13839, 16) ('##dd', 18946, 17) ('##ilde', 19322, 18) ### ### ('miles', 870, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('unknown', 8, 0) ('journey', 16, 1) ('bell', 19, 10) ('armchair', 1, 58) ('convincing', 52, 28) ### ### ('newman', 136, 12) ('littered', 66, 31) ('episode', 47, 54) ('episodes', 68, 48) ### ### ('british', 72, 52) ('¹', 345, 25) ('filming', 75, 55) ('ம', 31, 85) ('dennis', 472, 20) ### ### ('幸', 27, 104) ('anyway', 264, 33) ('george', 129, 49) ('খ', 18, 137) ('absently', 12, 183) ### ### ('miles', 870, 19) ### ############################################################################################################ [2023-11-26 02:08:56,051][root][INFO] - Epoch: 5: Step: 9901/19530, loss[v]=0.063925, lr=0.000002, acc@1[1]=1000.5/1024=0.97705078125, acc@1[2]=1019.0/1024=0.9951171875 [2023-11-26 02:10:44,796][root][INFO] - Train batch 10000 [2023-11-26 02:10:44,796][root][INFO] - Avg. loss per last 100 batches: 0.122835 [2023-11-26 02:10:45,498][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.2/29522=99.92% | mean: 0.02 | max: 5.85 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 1894 rural district s and urban district s were set up based on the sanitary districts ### ### ( and therefore indirectly on the unions ). ### ### [P_TEXT]: urban district ( great britain and ireland ) rural sanitary district s ( based on poor ### ### law union s ) the functions of which were taken over by the district councils. the district ### ### councils also had wider powers over local matters such as parks, cemeteries and local planning. an ### ### urban district usually contained a single parish, while a rural district might contain many. urban ### ### districts were considered to have more problems with public health than rural areas, and so urban ### ### district councils had more funding and greater powers than comparable rural districts. urban ### ### districts normally covered smaller towns, usually with populations ### ### =============================================== v_q ================================================ ### ### ('rural', 0, 21) ('貴', 1, 172) ('district', 2, 7) ('urban', 3, 3) ('1894', 4, 20682) ('幸', 5, 219) ### ### ('খ', 6, 176) ('intently', 7, 264) ('districts', 8, 14) ('absently', 9, 106) ('##ப', 10, 213) ### ### ('sanitary', 11, 0) ('¹', 12, 186) ('ம', 13, 86) ('unions', 14, 4) ('union', 15, 8) ('.', 16, 363) ### ### ('worried', 17, 41) ('anyway', 18, 36) ('##部', 19, 66) ### ### =============================================== v_p ================================================ ### ### ('sanitary', 11, 0) ('poor', 782, 1) ('cemeteries', 18129, 2) ('urban', 3, 3) ('unions', 14, 4) ### ### ('planning', 1892, 5) ('councils', 733, 6) ('district', 2, 7) ('union', 15, 8) ### ### ('parishes', 1457, 9) ('parks', 4607, 10) ('health', 1789, 11) ('law', 462, 12) ### ### ('burials', 27798, 13) ('districts', 8, 14) ('populations', 4309, 15) ('sanitation', 143, 16) ### ### ('towns', 690, 17) ('britain', 138, 18) ('parish', 258, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('sanitary', 11, 0) ('urban', 3, 3) ('district', 2, 7) ('unions', 14, 4) ('rural', 0, 21) ### ### ('union', 15, 8) ('districts', 8, 14) ('anyway', 18, 36) ('worried', 17, 41) ('ம', 13, 86) ### ### ('absently', 9, 106) ('貴', 1, 172) ('local', 44, 25) ('##部', 19, 66) ('খ', 6, 176) ('幸', 5, 219) ### ### ('poor', 782, 1) ('¹', 12, 186) ('##ப', 10, 213) ('sanitation', 143, 16) ### ############################################################################################################ [2023-11-26 02:10:46,002][root][INFO] - Epoch: 5: Step: 10001/19530, loss[v]=0.115673, lr=0.000002, acc@1[1]=1001.5/1024=0.97802734375, acc@1[2]=1016.0/1024=0.9921875 [2023-11-26 02:12:34,234][root][INFO] - Train batch 10100 [2023-11-26 02:12:34,235][root][INFO] - Avg. loss per last 100 batches: 0.119937 [2023-11-26 02:12:34,942][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.0/29522=99.92% | mean: 0.02 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: seghwaert was formed around what was the former village of zegwaart ( or zegwaard ), ### ### which fused with zoetermeer in 1935, using the archaic spelling of the locale's name. ### ### [P_TEXT]: seghwaert seghwaert is a district in zoetermeer that was formed around the old village of ### ### zegwaart ( or zegwaard ), which fused with the municipality of zoetermeer in 1935, using the ### ### archaic spelling of the locale's name. the old zegwaart was a ribbon development along what remains ### ### until today as " zegwaartseweg ", which is perpendicular to " dorpstraat ". the district is divided ### ### into seghwaert - oost ( 2723 ), - midden ( 2724 ) and - noord ( 2727 ). in the middle ages zegwaart ### ### and zoetermeer formed a parish. the coat of arms of the ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 119) ('##gh', 1, 6) ('³', 2, 605) ('##wa', 3, 16) ('zoe', 4, 0) ('幸', 5, 197) ### ### ('intently', 6, 261) ('se', 7, 14) ('##ert', 8, 3) ('খ', 9, 163) ('##me', 10, 17) ('貴', 11, 187) ### ### ('ம', 12, 106) ('ze', 13, 7) ('##ப', 14, 152) ('##art', 15, 12) ('##ɛ', 16, 57) ### ### ('absently', 17, 114) ('netherlands', 18, 132) ('anyway', 19, 35) ### ### =============================================== v_p ================================================ ### ### ('zoe', 4, 0) ('##ord', 5073, 1) ('ribbon', 15681, 2) ('##ert', 8, 3) ('##den', 14162, 4) ### ### ('perpendicular', 11909, 5) ('##gh', 1, 6) ('ze', 13, 7) ('##stra', 22042, 8) ### ### ('heraldic', 14766, 9) ('mid', 16776, 10) ('272', 18424, 11) ('##art', 15, 12) ('##arts', 502, 13) ### ### ('se', 7, 14) ('##rp', 17825, 15) ('##wa', 3, 16) ('##me', 10, 17) ('worried', 114, 18) ### ### ('##ost', 12706, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('zoe', 4, 0) ('##gh', 1, 6) ('##ert', 8, 3) ('##wa', 3, 16) ('se', 7, 14) ('ze', 13, 7) ### ### ('##me', 10, 17) ('##art', 15, 12) ('##ter', 22, 25) ('village', 23, 31) ('anyway', 19, 35) ### ### ('##g', 21, 36) ('¹', 0, 119) ('##ɛ', 16, 57) ('##ard', 45, 21) ('municipality', 47, 22) ### ### ('ம', 12, 106) ('##er', 27, 47) ('খ', 9, 163) ('幸', 5, 197) ### ############################################################################################################ [2023-11-26 02:12:35,449][root][INFO] - Epoch: 5: Step: 10101/19530, loss[v]=0.161835, lr=0.000002, acc@1[1]=978.0/1024=0.955078125, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-26 02:14:24,324][root][INFO] - Train batch 10200 [2023-11-26 02:14:24,324][root][INFO] - Avg. loss per last 100 batches: 0.123722 [2023-11-26 02:14:25,040][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.7/29522=99.92% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.9/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: langton bay was the base of operations for the three - year expedition, 1909 to 1912, of ### ### arctic explorers vilhjalmur stefansson and rudolph anderson ### ### [P_TEXT]: j. keith fraser. in 1951, he assisted geomorphologist j. ross mackay in investigating ### ### amundsen gulf's darnley bay area. after reviewing recent aerial photographs and old maps, fraser ### ### determined that the riviere la ronciere - le noury, charted by french missionary oblate emile ### ### petitot in 1875 but considered nonexistent decades later by explorer vilhjalmur stefansson, did ### ### exist. though there were errors in the placement of the ronciere's mouth at franklin bay, the ### ### ronciere was the same river as the partially mapped hornaday river by ### ### =============================================== v_q ================================================ ### ### ('lang', 0, 20671) ('¹', 1, 115) ('absently', 2, 172) ('幸', 3, 218) ('##ton', 4, 21280) ### ### ('খ', 5, 181) ('³', 6, 431) ('intently', 7, 163) ('arctic', 8, 1092) ('貴', 9, 171) ('##ப', 10, 212) ### ### ('expedition', 11, 50) ('ம', 12, 139) ('##jal', 13, 11) ('##部', 14, 132) ('bay', 15, 20) ### ### ('stefan', 16, 13) ('antarctic', 17, 107) ('explorer', 18, 35) ('anderson', 19, 10767) ### ### =============================================== v_p ================================================ ### ### ('##aday', 20299, 0) ('franklin', 585, 1) ('fraser', 4184, 2) ('mackay', 1214, 3) ### ### ('##nl', 23717, 4) ('##ury', 23066, 5) ('keith', 9603, 6) ('petit', 19637, 7) ('##cier', 27046, 8) ### ### ('##late', 26049, 9) ('gulf', 443, 10) ('##jal', 13, 11) ('##unds', 23333, 12) ('stefan', 16, 13) ### ### ('ron', 7323, 14) ('ross', 6453, 15) ('horn', 5478, 16) ('dar', 18634, 17) ('##sson', 21, 18) ### ### ('ob', 10384, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##jal', 13, 11) ('stefan', 16, 13) ('bay', 15, 20) ('##sson', 21, 18) ('expedition', 11, 50) ### ### ('explorer', 18, 35) ('¹', 1, 115) ('anyway', 56, 30) ('##mur', 41, 40) ('absently', 2, 172) ### ### ('খ', 5, 181) ('franklin', 585, 1) ('幸', 3, 218) ('luckily', 28, 68) ('ம', 12, 139) ### ### ('intently', 7, 163) ('vi', 46, 55) ('antarctic', 17, 107) ('##部', 14, 132) ('gulf', 443, 10) ### ############################################################################################################ [2023-11-26 02:14:25,547][root][INFO] - Epoch: 5: Step: 10201/19530, loss[v]=0.131372, lr=0.000002, acc@1[1]=982.5/1024=0.95947265625, acc@1[2]=1000.0/1024=0.9765625 [2023-11-26 02:16:13,818][root][INFO] - Train batch 10300 [2023-11-26 02:16:13,819][root][INFO] - Avg. loss per last 100 batches: 0.121345 [2023-11-26 02:16:14,520][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29496.6/29522=99.91% | mean: 0.02 | max: 5.76 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: hales is a village in staffordshire approximately 2 miles east of market drayton. ### ### [P_TEXT]: reginald corbet and secure for the latter part of their lives. on alice's death, the ### ### estates were listed as the manors of drayton, adderley, stoke on tern, hales, almington and blore, ### ### the village of child's ercall and granges at cliffe, tern hill, burnhill and cheethill. augusta ### ### corbett gives reginald's year of death as 1569, but the date of probate makes this impossible and ### ### 19 november 1566, is now the accepted date, only about four months after the death of his brother, ### ### richard. he had ### ### =============================================== v_q ================================================ ### ### ('hale', 0, 2) ('¹', 1, 45) ('staffordshire', 2, 1027) ('##ayton', 3, 3) ('dr', 4, 28) ### ### ('³', 5, 210) ('market', 6, 7131) ('##s', 7, 238) ('curious', 8, 293) ('luckily', 9, 57) ### ### ('##स', 10, 103) ('intently', 11, 120) ('trent', 12, 1428) ('খ', 13, 130) ('villagers', 14, 94) ### ### ('##ப', 15, 132) ('ம', 16, 82) ('幸', 17, 148) ('reputed', 18, 716) ('shillings', 19, 8442) ### ### =============================================== v_p ================================================ ### ### ('reginald', 2485, 0) ('##lore', 20343, 1) ('hale', 0, 2) ('##ayton', 3, 3) ('corbett', 2684, 4) ### ### ('ter', 3571, 5) ('cliff', 309, 6) ('##ming', 9004, 7) ('stoke', 31, 8) ('add', 3586, 9) ### ### ('grange', 1161, 10) ('burn', 1031, 11) ('##eth', 8363, 12) ('child', 3673, 13) ### ### ('augusta', 12143, 14) ('##hill', 607, 15) ('##rb', 12420, 16) ('##bate', 18826, 17) ### ### ('##ill', 9179, 18) ('che', 14672, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('hale', 0, 2) ('##ayton', 3, 3) ('dr', 4, 28) ('¹', 1, 45) ('stoke', 31, 8) ('luckily', 9, 57) ### ### ('village', 32, 31) ('manor', 37, 29) ('nearby', 79, 22) ('cliff', 309, 6) ('neighbouring', 55, 36) ### ### ('parish', 92, 33) ('##स', 10, 103) ('##ton', 23, 65) ('ம', 16, 82) ('##ley', 83, 37) ### ### ('##ս', 20, 75) ('##hill', 607, 15) ('villagers', 14, 94) ('³', 5, 210) ### ############################################################################################################ [2023-11-26 02:16:15,023][root][INFO] - Epoch: 5: Step: 10301/19530, loss[v]=0.130344, lr=0.000002, acc@1[1]=978.0/1024=0.955078125, acc@1[2]=1008.5/1024=0.98486328125 [2023-11-26 02:18:03,349][root][INFO] - Train batch 10400 [2023-11-26 02:18:03,349][root][INFO] - Avg. loss per last 100 batches: 0.125135 [2023-11-26 02:18:04,057][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=99.99% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.1/29522=99.92% | mean: 0.02 | max: 5.83 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: as the # 4 seed for the stepladder finals, he won all four matches, defeating four major ### ### titleists ( sean rash, jason belmonte, anthony simonsen and bill o'neill ) to capture his third pba ### ### tour title, first major, and second career $ 100, 000 prize check. ### ### [P_TEXT]: bill o'neill ( bowler ) world series of bowling x in march, he had top ten finishes in ### ### all four events and made two televised finals, but he didn't win a title. o'neill qualified as the ### ### # 7 seed for the inaugural pba tour playoffs. he made it all the way to the june 2 finals in this ### ### event, but lost to kristopher prather. o'neil cashed $ 40, 000 for his runner - up finish, and ### ### surpassed the $ 1, 000, 000 mark in career earnings during the 2019 season. on august 11, o'neill ### ### won his 11th pba tour title at the pba harry o'neale chesapeake ### ### =============================================== v_q ================================================ ### ### ('pba', 0, 2) ('absently', 1, 118) ('jason', 2, 9653) ('littered', 3, 20) ('খ', 4, 168) ### ### ('##ப', 5, 170) ('幸', 6, 186) ('貴', 7, 182) ('sean', 8, 11322) ('step', 9, 9290) ('¹', 10, 171) ### ### ('neill', 11, 0) ('anthony', 12, 19133) ('belmont', 13, 4473) ('.', 14, 3583) ('^', 15, 1195) ### ### ('##部', 16, 111) ('simon', 17, 5294) ('manny', 18, 4462) ('ம', 19, 120) ### ### =============================================== v_p ================================================ ### ### ('neill', 11, 0) ('chesapeake', 7114, 1) ('pba', 0, 2) ('bowling', 43, 3) ('##ath', 26557, 4) ### ### ('bowler', 486, 5) ('neal', 785, 6) ('neil', 1105, 7) ('bill', 23, 8) ('x', 3879, 9) ### ### ('cash', 280, 10) ('playoffs', 490, 11) ('2019', 811, 12) ('anyway', 35, 13) ### ### ('televised', 1052, 14) ('tour', 39, 15) ('leaned', 631, 16) ('##top', 13707, 17) ### ### ('harry', 9066, 18) ('o', 31, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('pba', 0, 2) ('neill', 11, 0) ('littered', 3, 20) ('bill', 23, 8) ('bowling', 43, 3) ### ### ('anyway', 35, 13) ('o', 31, 19) ('tour', 39, 15) ('finals', 29, 25) ('absently', 1, 118) ### ### ('earnings', 75, 30) ('cash', 280, 10) ('##部', 16, 111) ('bowler', 486, 5) ('খ', 4, 168) ### ### ('$', 74, 38) ('ம', 19, 120) ('##ப', 5, 170) ('tournament', 36, 63) ('seed', 241, 22) ### ############################################################################################################ [2023-11-26 02:18:04,562][root][INFO] - Epoch: 5: Step: 10401/19530, loss[v]=0.088659, lr=0.000002, acc@1[1]=1003.5/1024=0.97998046875, acc@1[2]=1016.5/1024=0.99267578125 [2023-11-26 02:19:53,321][root][INFO] - Train batch 10500 [2023-11-26 02:19:53,322][root][INFO] - Avg. loss per last 100 batches: 0.126549 [2023-11-26 02:19:54,039][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.2/29522=99.92% | mean: 0.02 | max: 5.85 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: " ballet mecanique ", with the george antheil music originally written for the film, was ### ### included in the dvd collection " " released in october 2005. ### ### [P_TEXT]: george antheil ". the " ballet " was originally conceived to be accompanied by the film ### ### of the same name by experimental filmmakers fernand leger and dudley murphy ( with cinematography ### ### by man ray ), although the nature of the collaboration is mysterious. the first productions of ### ### antheil's work in 1925 and 1926 did not include the film, which turned out to last around 19 ### ### minutes, only half as long as antheil's score. antheil described his " first big work " as " scored ### ### for countless numbers of player pianos. all percussive. like machines. all efficiency. no love. ### ### written without sympathy. ### ### =============================================== v_q ================================================ ### ### ('##hei', 0, 0) ('ant', 1, 8) ('##can', 2, 25979) ('##ique', 3, 23045) ('intently', 4, 175) ### ### ('¹', 5, 130) ('ballet', 6, 3) ('貴', 7, 161) ('幸', 8, 163) ('george', 9, 9) ('absently', 10, 98) ### ### ('খ', 11, 155) ('³', 12, 2025) ('ம', 13, 97) ('##ப', 14, 127) ('pained', 15, 69) ('film', 16, 16) ### ### ('me', 17, 17666) ('2005', 18, 22942) ('casablanca', 19, 13067) ### ### =============================================== v_p ================================================ ### ### ('##hei', 0, 0) ('pianos', 11189, 1) ('dudley', 16939, 2) ('ballet', 6, 3) ('murphy', 11926, 4) ### ### ('leger', 5916, 5) ('sympathy', 2290, 6) ('##sive', 19396, 7) ('ant', 1, 8) ('george', 9, 9) ### ### ('efficiency', 22523, 10) ('ray', 11699, 11) ('filmmakers', 114, 12) ('1925', 8253, 13) ### ### ('cinematography', 3265, 14) ('fern', 20842, 15) ('film', 16, 16) ('littered', 253, 17) ### ### ('1926', 15158, 18) ('piano', 585, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##hei', 0, 0) ('ballet', 6, 3) ('ant', 1, 8) ('george', 9, 9) ('film', 16, 16) ('anyway', 22, 23) ### ### ('″', 28, 40) ('pained', 15, 69) ('¹', 5, 130) ('absently', 10, 98) ('filmmakers', 114, 12) ### ### ('intently', 4, 175) ('貴', 7, 161) ('ம', 13, 97) ('##l', 67, 32) ('幸', 8, 163) ('খ', 11, 155) ### ### ('##部', 21, 88) ('##ப', 14, 127) ('movie', 47, 41) ### ############################################################################################################ [2023-11-26 02:19:54,546][root][INFO] - Epoch: 5: Step: 10501/19530, loss[v]=0.070961, lr=0.000002, acc@1[1]=1001.0/1024=0.9775390625, acc@1[2]=1018.0/1024=0.994140625 [2023-11-26 02:21:43,017][root][INFO] - Train batch 10600 [2023-11-26 02:21:43,018][root][INFO] - Avg. loss per last 100 batches: 0.121692 [2023-11-26 02:21:43,716][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.5/29522=99.93% | mean: 0.02 | max: 5.81 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.2/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he is portrayed by john noble. ### ### [P_TEXT]: olivia. in the lab. with the revolver. walter ( john noble ). she goes to the bowling ### ### alley, managed by sam weiss ( kevin corrigan ), where he tells her she is one of the best people he ### ### knows, and he is sure she will do the right thing. in the case, it transpires that victims of ### ### attacks similar to miranda green ’ s have been connected to the cortexiphan trials in jacksonville, ### ### florida. olivia visits nina sharp ( blair brown ) at massive dynamic's headquarters and demands a ### ### full list of the subjects, knowing she herself is ### ### =============================================== v_q ================================================ ### ### ('noble', 0, 6) ('john', 1, 39) ('¹', 2, 77) ('portrayed', 3, 23) ('absently', 4, 133) ### ### ('actor', 5, 741) ('貴', 6, 204) ('³', 7, 1907) ('intently', 8, 142) ('幸', 9, 145) ### ### ('wears', 10, 5662) ('hears', 11, 11828) ('ம', 12, 116) ('learns', 13, 1634) ('doctor', 14, 3673) ### ### ('stumble', 15, 826) ('sees', 16, 698) ('discover', 17, 4322) ('##ப', 18, 162) ### ### ('decide', 19, 24053) ### ### =============================================== v_p ================================================ ### ### ('##rrigan', 2274, 0) ('olivia', 1621, 1) ('cortex', 5243, 2) ('miranda', 848, 3) ### ### ('sharp', 16869, 4) ('weiss', 10971, 5) ('noble', 0, 6) ('walter', 1536, 7) ('dynamic', 14110, 8) ### ### ('revolver', 8711, 9) ('blair', 2093, 10) ('nina', 5667, 11) ('jacksonville', 11964, 12) ### ### ('sam', 3223, 13) ('massive', 5924, 14) ('green', 3121, 15) ('bowling', 2160, 16) ### ### ('brown', 6351, 17) ('worried', 253, 18) ('kevin', 1421, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('noble', 0, 6) ('john', 1, 39) ('portrayed', 3, 23) ('anyway', 25, 21) ('¹', 2, 77) ### ### ('miranda', 848, 3) ('worried', 253, 18) ('absently', 4, 133) ('olivia', 1621, 1) ('ம', 12, 116) ### ### ('intently', 8, 142) ('episode', 58, 53) ('role', 113, 40) ('幸', 9, 145) ('portrayal', 81, 49) ### ### ('##rrigan', 2274, 0) ('asks', 59, 59) ('walter', 1536, 7) ('series', 126, 45) ('貴', 6, 204) ### ############################################################################################################ [2023-11-26 02:21:44,221][root][INFO] - Epoch: 5: Step: 10601/19530, loss[v]=0.151876, lr=0.000002, acc@1[1]=981.5/1024=0.95849609375, acc@1[2]=1011.0/1024=0.9873046875 [2023-11-26 02:23:32,561][root][INFO] - Train batch 10700 [2023-11-26 02:23:32,562][root][INFO] - Avg. loss per last 100 batches: 0.126543 [2023-11-26 02:23:33,266][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.8/29522=99.92% | mean: 0.02 | max: 5.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: 1962, venezuelan foreign minister gave an exposition in the special political and ### ### decolonization committee of the united nations general assembly to denounce the 1899 paris tribunal ### ### arbitration, whose ruling was favorable to the united kingdom, citing the. ### ### [P_TEXT]: venezuelan crisis of 1895 were conducted also made for good relations with britain. ### ### however, by backing down in the face of a strong us declaration of a strong interpretation of the ### ### monroe doctrine, britain tacitly accepted it, and the crisis thus provided a basis for the ### ### expansion of us interventionism in the americas. leading british historian robert arthur humphreys ### ### later called the crisis " one of the most momentous episodes in the history of anglo - american ### ### relations in general and of anglo - american rivalries in latin america in particular. " by 1895, ### ### the dispute between britain and venezuela over the territory of guayana esequiba, which britain ### ### claimed as part of british guiana and venezuela saw as venezuelan territory, had lasted ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 161) ('arbitration', 1, 756) ('paris', 2, 4417) ('foreign', 3, 4418) ### ### ('exposition', 4, 1755) ('venezuela', 5, 9) ('venezuelan', 6, 3) ('1899', 7, 138) ('幸', 8, 219) ### ### ('tribunal', 9, 4909) ('貴', 10, 213) ('খ', 11, 190) ('nations', 12, 1505) ('##ization', 13, 3632) ### ### ('##¨', 14, 1630) ('intently', 15, 150) ('absently', 16, 157) ('³', 17, 407) ('##ப', 18, 243) ### ### ('ம', 19, 144) ### ### =============================================== v_p ================================================ ### ### ('humphrey', 18648, 0) ('crisis', 476, 1) ('guiana', 485, 2) ('venezuelan', 6, 3) ### ### ('##ayan', 4759, 4) ('doctrine', 940, 5) ('1895', 1114, 6) ('monroe', 10375, 7) ### ### ('intervention', 610, 8) ('venezuela', 5, 9) ('relations', 75, 10) ('declaration', 147, 11) ### ### ('arthur', 14777, 12) ('##ba', 16324, 13) ('doctrines', 26642, 14) ('interventions', 19035, 15) ### ### ('crises', 24396, 16) ('anyway', 88, 17) ('dispute', 42, 18) ('anglo', 507, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('venezuelan', 6, 3) ('venezuela', 5, 9) ('dispute', 42, 18) ('relations', 75, 10) ### ### ('britain', 53, 22) ('british', 60, 23) ('anyway', 88, 17) ('crisis', 476, 1) ### ### ('declaration', 147, 11) ('¹', 0, 161) ('guiana', 485, 2) ('1899', 7, 138) ('intervention', 610, 8) ### ### ('intently', 15, 150) ('ம', 19, 144) ('absently', 16, 157) ('খ', 11, 190) ('貴', 10, 213) ### ### ('幸', 8, 219) ('doctrine', 940, 5) ### ############################################################################################################ [2023-11-26 02:23:33,771][root][INFO] - Epoch: 5: Step: 10701/19530, loss[v]=0.066370, lr=0.000002, acc@1[1]=1003.0/1024=0.9794921875, acc@1[2]=1015.5/1024=0.99169921875 [2023-11-26 02:25:23,039][root][INFO] - Train batch 10800 [2023-11-26 02:25:23,040][root][INFO] - Avg. loss per last 100 batches: 0.129815 [2023-11-26 02:25:23,758][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.3/29522=99.92% | mean: 0.02 | max: 5.81 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: at that time " beagle ", under captain pringle stokes, was carrying out a hydrographic ### ### survey of tierra del fuego, under the overall command of captain phillip parker king in hms " ### ### adventure ". ### ### [P_TEXT]: phillip parker king on the continuing survey of patagonia under robert fitzroy, in the ### ### company of noted scientist charles darwin ( 1809 – 1882 ). king owned a property at dunheved in the ### ### western suburbs of sydney where he entertained charles darwin on darwin's last night in sydney in ### ### january 1836. king was appointed to the first new south wales legislative council in 1829, however ### ### he was absent from the colony and did not take his seat and was replaced by john campbell. when ### ### king returned to the colony in 1832 he pressed for his ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 158) ('parker', 1, 2) ('king', 2, 8) ('stokes', 3, 7570) ('¹', 4, 292) ### ### ('adventure', 5, 5027) ('bea', 6, 23809) ('খ', 7, 196) ('intently', 8, 154) ('pri', 9, 23810) ### ### ('貴', 10, 275) ('幸', 11, 313) ('##gle', 12, 23253) ('##ப', 13, 267) ('tier', 14, 17780) ### ### ('##部', 15, 127) ('falkland', 16, 2903) ('phillip', 17, 5) ('ம', 18, 180) ('³', 19, 3086) ### ### =============================================== v_p ================================================ ### ### ('fitzroy', 1361, 0) ('campbell', 11243, 1) ('parker', 1, 2) ('darwin', 181, 3) ('##ved', 25235, 4) ### ### ('phillip', 17, 5) ('sydney', 404, 6) ('dun', 12470, 7) ('king', 2, 8) ('1829', 2967, 9) ### ### ('charles', 1990, 10) ('##ago', 20381, 11) ('worried', 526, 12) ('robert', 2932, 13) ### ### ('1836', 1961, 14) ('anyway', 77, 15) ('legislative', 20880, 16) ('1809', 1899, 17) ### ### ('convicts', 14016, 18) ('survey', 41, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('parker', 1, 2) ('king', 2, 8) ('phillip', 17, 5) ('survey', 41, 19) ('darwin', 181, 3) ### ### ('anyway', 77, 15) ('absently', 0, 158) ('sydney', 404, 6) ('intently', 8, 154) ### ### ('fitzroy', 1361, 0) ('##部', 15, 127) ('খ', 7, 196) ('worried', 526, 12) ('¹', 4, 292) ### ### ('ம', 18, 180) ('1832', 350, 25) ('貴', 10, 275) ('colony', 332, 31) ('##ப', 13, 267) ('幸', 11, 313) ### ############################################################################################################ [2023-11-26 02:25:24,263][root][INFO] - Epoch: 5: Step: 10801/19530, loss[v]=0.055951, lr=0.000002, acc@1[1]=1005.5/1024=0.98193359375, acc@1[2]=1018.0/1024=0.994140625 [2023-11-26 02:27:13,317][root][INFO] - Train batch 10900 [2023-11-26 02:27:13,318][root][INFO] - Avg. loss per last 100 batches: 0.121692 [2023-11-26 02:27:14,017][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.6/29522=99.92% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.7/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: 4 seed wang yihan in quarter - final, reigning world champion lu lan in semis & no. ### ### [P_TEXT]: china badminton super league to the semi - finals were 81 chivas, jiangsu yonex, hubei ### ### and home - team guangzhou where the ties took place. in the first semi - final, jiangsu beat hubei ### ### 3 - 0. in the second semi - final 81 chivas beat guangzhou 3 - 0. in the final the strong and ### ### powerful jiangsu yonex team beat 81 chivas with 3 - 0. the 2009 cbsl championship team of jiangsu ### ### yonex included top players like wang shixian, tao jiaming, sun xiaoli, cheng shu, tang jinhua, li ### ### yu, zhi luoliang, cai yun, xu chen and lu lan. based on the results of the inaugural first year's ### ### cbsl competition, 8 clubs were selected to play in ### ### =============================================== v_q ================================================ ### ### ('lan', 0, 2) ('absently', 1, 254) ('lu', 2, 39) ('##han', 3, 19484) ('貴', 4, 132) ('¹', 5, 41) ### ### ('幸', 6, 133) ('খ', 7, 189) ('intently', 8, 87) ('wang', 9, 12) ('##ப', 10, 158) ('##部', 11, 182) ### ### ('ம', 12, 128) ('yi', 13, 7189) ('chess', 14, 249) ('.', 15, 118) ('³', 16, 428) ('anyway', 17, 15) ### ### ('ɐ', 18, 145) ('open', 19, 209) ### ### =============================================== v_p ================================================ ### ### ('cbs', 11324, 0) ('badminton', 39, 1) ('lan', 0, 2) ('china', 26, 3) ('##xia', 7759, 4) ### ### ('guangzhou', 84, 5) ('tao', 5840, 6) ('sun', 3424, 7) ('chen', 412, 8) ('##oli', 26444, 9) ### ### ('##ming', 2975, 10) ('yu', 560, 11) ('wang', 9, 12) ('bwf', 604, 13) ('super', 1999, 14) ### ### ('anyway', 17, 15) ('li', 1294, 16) ('xu', 449, 17) ('littered', 756, 18) ('tang', 231, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('lan', 0, 2) ('wang', 9, 12) ('lu', 2, 39) ('anyway', 17, 15) ('¹', 5, 41) ('china', 26, 3) ### ### ('badminton', 39, 1) ('guangzhou', 84, 5) ('intently', 8, 87) ('貴', 4, 132) ('幸', 6, 133) ### ### ('jiangsu', 106, 21) ('ம', 12, 128) ('.', 15, 118) ('##ப', 10, 158) ('2009', 125, 31) ### ### ('championship', 28, 60) ('absently', 1, 254) ('খ', 7, 189) ('cheng', 164, 28) ### ############################################################################################################ [2023-11-26 02:27:14,521][root][INFO] - Epoch: 5: Step: 10901/19530, loss[v]=0.136196, lr=0.000002, acc@1[1]=984.5/1024=0.96142578125, acc@1[2]=1008.0/1024=0.984375 [2023-11-26 02:29:03,716][root][INFO] - Train batch 11000 [2023-11-26 02:29:03,717][root][INFO] - Avg. loss per last 100 batches: 0.125640 [2023-11-26 02:29:04,445][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.5/29522=99.92% | mean: 0.02 | max: 5.91 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: at wrestlemania v in april 1989, after bravo's match with ron garvin, martin attacked ### ### garvin, and as a result, ### ### [P_TEXT]: wrestlemania v the ropes. garvin tried to hit a piledriver. bravo went for a back body ### ### drop but garvin held on for a sunset flip pin for a near - fall. garvin took him to the corner and ### ### went for ten - count - corner punches. bravo countered with an inverted atomic drop and a sidewalk ### ### slam, followed by a pinfall victory. after the match, bravo's manager frenchy martin attacked ### ### garvin but received a garvin stomp for his troubles. the eighth match was a tag team match between ### ### the brain busters ( arn anderson and tully blanchard ) with bobby heenan in their corner, and ### ### strike force ( tito santana and ### ### =============================================== v_q ================================================ ### ### ('martin', 0, 15) ('v', 1, 1) ('¹', 2, 65) ('absently', 3, 156) ('bravo', 4, 4) ('খ', 5, 129) ### ### ('貴', 6, 155) ('幸', 7, 125) ('ron', 8, 10519) ('##vin', 9, 23) ('1989', 10, 16430) ### ### ('intently', 11, 117) ('³', 12, 557) ('wrestlemania', 13, 0) ('##ப', 14, 139) ('！', 15, 428) ### ### ('ம', 16, 91) ('ga', 17, 49) ('wrestling', 18, 44) ('wwf', 19, 109) ### ### =============================================== v_p ================================================ ### ### ('wrestlemania', 13, 0) ('v', 1, 1) ('buster', 868, 2) ('blanchard', 3752, 3) ('bravo', 4, 4) ### ### ('flip', 2637, 5) ('santana', 1874, 6) ('anderson', 1129, 7) ('brain', 3169, 8) ('strike', 1763, 9) ### ### ('sunset', 1658, 10) ('tito', 2594, 11) ('hee', 16636, 12) ('tully', 9453, 13) ### ### ('punches', 6660, 14) ('martin', 0, 15) ('##omp', 17126, 16) ('##nan', 19907, 17) ('ar', 6165, 18) ### ### ('sidewalk', 18190, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('v', 1, 1) ('martin', 0, 15) ('bravo', 4, 4) ('wrestlemania', 13, 0) ('##vin', 9, 23) ('¹', 2, 65) ### ### ('ga', 17, 49) ('wrestling', 18, 44) ('anyway', 43, 25) ('match', 27, 40) ('feud', 74, 28) ### ### ('খ', 5, 129) ('幸', 7, 125) ('bobby', 67, 37) ('absently', 3, 156) ('littered', 117, 31) ### ### ('ம', 16, 91) ('intently', 11, 117) ('buster', 868, 2) ('貴', 6, 155) ### ############################################################################################################ [2023-11-26 02:29:04,951][root][INFO] - Epoch: 5: Step: 11001/19530, loss[v]=0.072005, lr=0.000002, acc@1[1]=1008.0/1024=0.984375, acc@1[2]=1017.0/1024=0.9931640625 [2023-11-26 02:30:54,490][root][INFO] - Train batch 11100 [2023-11-26 02:30:54,491][root][INFO] - Avg. loss per last 100 batches: 0.119427 [2023-11-26 02:30:55,209][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.5/29522=99.99% | mean: 0.03 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.3/29522=99.93% | mean: 0.02 | max: 5.73 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 1981, an electoral alliance was established between the liberal party, a group which ### ### descended from the 18th - century whigs, and the social democratic party ( sdp ), a splinter group ### ### from the labour party. ### ### [P_TEXT]: list of parliamentary constituencies in the east midlands ( region ) lib - liberal party ### ### up to 1979 ; sdp - liberal alliance 1983 & 1987 ; liberal democrats from 1992 oth - 1945 - ### ### independent ( denis kendall ) ; 1974 ( feb ) - democratic labour ( dick taverne ) ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 143) ('liberal', 1, 7) ('খ', 2, 90) ('貴', 3, 108) ('whig', 4, 225) ('幸', 5, 113) ### ### ('social', 6, 5464) ('intently', 7, 109) ('.', 8, 86) ('##ப', 9, 79) ('sd', 10, 3) ('ம', 11, 76) ### ### ('ד', 12, 85) ('party', 13, 18) ('1981', 14, 866) ('¹', 15, 44) ('labour', 16, 10) ### ### ('worried', 17, 75) ('##部', 18, 74) ('alliance', 19, 8) ### ### =============================================== v_p ================================================ ### ### ('kendall', 18200, 0) ('tavern', 7347, 1) ('midlands', 955, 2) ('sd', 10, 3) ('feb', 8337, 4) ### ### ('ot', 24596, 5) ('denis', 8328, 6) ('liberal', 1, 7) ('alliance', 19, 8) ('dick', 14123, 9) ### ### ('labour', 16, 10) ('constituency', 34, 11) ('1983', 64, 12) ('constituencies', 2801, 13) ### ### ('littered', 2285, 14) ('democratic', 39, 15) ('1979', 85, 16) ('liberals', 1200, 17) ### ### ('party', 13, 18) ('1987', 100, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('liberal', 1, 7) ('sd', 10, 3) ('alliance', 19, 8) ('labour', 16, 10) ('party', 13, 18) ### ### ('constituency', 34, 11) ('democratic', 39, 15) ('anyway', 20, 39) ('¹', 15, 44) ('1983', 64, 12) ### ### ('খ', 2, 90) ('##ப', 9, 79) ('.', 8, 86) ('貴', 3, 108) ('1979', 85, 16) ('absently', 0, 143) ### ### ('ம', 11, 76) ('intently', 7, 109) ('幸', 5, 113) ('1987', 100, 19) ### ############################################################################################################ [2023-11-26 02:30:55,714][root][INFO] - Epoch: 5: Step: 11101/19530, loss[v]=0.175607, lr=0.000002, acc@1[1]=981.0/1024=0.9580078125, acc@1[2]=1009.0/1024=0.9853515625 [2023-11-26 02:32:44,714][root][INFO] - Train batch 11200 [2023-11-26 02:32:44,715][root][INFO] - Avg. loss per last 100 batches: 0.127305 [2023-11-26 02:32:45,432][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.0/29522=99.93% | mean: 0.02 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: above the lincoln square neighborhood — where the abc television network houses its ### ### corporate headquarters in a group of rehabilitated and modern buildings — columbus avenue passes ### ### through the central park west historic district, stretching from 67th / 68th streets to 89th ### ### street. ### ### [P_TEXT]: upper west side end avenue ( 11th avenue ), broadway, amsterdam avenue ( 10th avenue ), ### ### columbus avenue ( 9th avenue ), and central park west ( 8th avenue ). the 66 - block stretch of ### ### broadway forms the spine of the neighborhood and runs diagonally north – south across the other ### ### avenues at the south end of the neighborhood ; above 78th street broadway runs north parallel to ### ### the other avenues. broadway enters the neighborhood at its juncture with central park west at ### ### columbus circle ( 59th street ), crosses columbus avenue at lincoln square ( 65th street ), ### ### amsterdam avenue at verdi square ( 71st street ), and then ### ### =============================================== v_q ================================================ ### ### ('columbus', 0, 2) ('absently', 1, 65) ('intently', 2, 50) ('幸', 3, 73) ('খ', 4, 98) ('¹', 5, 16) ### ### ('貴', 6, 71) ('central', 7, 31) ('lincoln', 8, 13) ('historic', 9, 1633) ('##ப', 10, 84) ### ### ('park', 11, 38) ('89', 12, 6078) ('ம', 13, 53) ('##½', 14, 461) ('abc', 15, 8696) ('west', 16, 17) ### ### ('³', 17, 46) ('67', 18, 2882) ('exclaimed', 19, 182) ### ### =============================================== v_p ================================================ ### ### ('upper', 4248, 0) ('verdi', 18215, 1) ('columbus', 0, 2) ('amsterdam', 448, 3) ('side', 11150, 4) ### ### ('avenue', 20, 5) ('avenues', 3018, 6) ('8th', 6828, 7) ('78', 10562, 8) ('10th', 10032, 9) ### ### ('9th', 4259, 10) ('circle', 5277, 11) ('broadway', 236, 12) ('lincoln', 8, 13) ('11th', 10974, 14) ### ### ('65', 1047, 15) ('¹', 5, 16) ('west', 16, 17) ('##th', 40, 18) ('71', 321, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('columbus', 0, 2) ('¹', 5, 16) ('avenue', 20, 5) ('lincoln', 8, 13) ('west', 16, 17) ### ### ('central', 7, 31) ('intently', 2, 50) ('park', 11, 38) ('absently', 1, 65) ('square', 25, 30) ### ### ('anyway', 27, 25) ('##th', 40, 18) ('幸', 3, 73) ('³', 17, 46) ('ம', 13, 53) ### ### ('neighborhood', 37, 22) ('貴', 6, 71) ('খ', 4, 98) ('##ப', 10, 84) ('street', 54, 23) ### ############################################################################################################ [2023-11-26 02:32:45,936][root][INFO] - Epoch: 5: Step: 11201/19530, loss[v]=0.144465, lr=0.000002, acc@1[1]=981.0/1024=0.9580078125, acc@1[2]=1008.5/1024=0.98486328125 [2023-11-26 02:34:35,194][root][INFO] - Train batch 11300 [2023-11-26 02:34:35,195][root][INFO] - Avg. loss per last 100 batches: 0.131414 [2023-11-26 02:34:35,930][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.9/29522=99.93% | mean: 0.02 | max: 5.85 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: anaheim is the main tourist hub, with the disneyland resort's disneyland being the second ### ### most visited theme park in the world. ### ### [P_TEXT]: anaheim, california anaheim ( ) is a city in orange county, california, part of the los ### ### angeles metropolitan area. as of the 2010 united states census, the city had a population of 336, ### ### 265, making it the most populous city in orange county, the 10th - most populous city in ### ### california, and the 55th - most populous city in the united states. anaheim is the second - largest ### ### city in orange county in terms of land area, and is known for being the home of the disneyland ### ### resort, the anaheim convention center, and two major sports ### ### =============================================== v_q ================================================ ### ### ('anaheim', 0, 0) ('disneyland', 1, 1) ('.', 2, 39) ('cradled', 3, 124) ('resort', 4, 16) ### ### ('california', 5, 10) ('¹', 6, 93) ('disney', 7, 99) ('幸', 8, 118) ('খ', 9, 140) ### ### ('absently', 10, 48) ('##ப', 11, 153) ('貴', 12, 116) ('visited', 13, 369) ('##部', 14, 92) ### ### ('ம', 15, 78) ('park', 16, 3681) ('leaned', 17, 6) ('theme', 18, 2868) ('snuggled', 19, 45) ### ### =============================================== v_p ================================================ ### ### ('anaheim', 0, 0) ('disneyland', 1, 1) ('populous', 172, 2) ('orange', 242, 3) ('worried', 20, 4) ### ### ('convention', 981, 5) ('leaned', 17, 6) ('55th', 23679, 7) ('cities', 55, 8) ('anyway', 414, 9) ### ### ('california', 5, 10) ('varies', 60, 11) ('subdivisions', 3619, 12) ('10th', 7411, 13) ### ### ('population', 329, 14) ('metropolitan', 566, 15) ('resort', 4, 16) ('largest', 111, 17) ### ### ('angeles', 27, 18) ('tenth', 4486, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('anaheim', 0, 0) ('disneyland', 1, 1) ('california', 5, 10) ('worried', 20, 4) ('resort', 4, 16) ### ### ('leaned', 17, 6) ('.', 2, 39) ('absently', 10, 48) ('angeles', 27, 18) ('populous', 172, 2) ### ### ('cities', 55, 8) ('¹', 6, 93) ('cradled', 3, 124) ('disney', 7, 99) ('snuggled', 19, 45) ### ### ('varies', 60, 11) ('orange', 242, 3) ('幸', 8, 118) ('ம', 15, 78) ('intently', 22, 49) ### ############################################################################################################ [2023-11-26 02:34:36,437][root][INFO] - Epoch: 5: Step: 11301/19530, loss[v]=0.123990, lr=0.000002, acc@1[1]=991.5/1024=0.96826171875, acc@1[2]=1012.0/1024=0.98828125 [2023-11-26 02:36:25,923][root][INFO] - Train batch 11400 [2023-11-26 02:36:25,924][root][INFO] - Avg. loss per last 100 batches: 0.129360 [2023-11-26 02:36:26,635][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.8/29522=99.92% | mean: 0.02 | max: 5.73 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: shepherd went on to form a new band, the monarchs ( initially known as the honkies ), ### ### together with andy kelly on bass ( glide ), greg hitchcock on guitar ( the neptunes, the ### ### kryptonics, new christs, the verys, challenger 7, you am i, the dearhunters ) and his brother ### ### murray shepherd ( harpoon, the fun things, screaming tribesmen ) ### ### [P_TEXT]: you am i label, rooart records, on their subsidiary ra records, which was distributed by ### ### wea. the group released their third ep, " can't get started ", as a five - track set in october ### ### 1992, which was co - produced by the band with david price. before the end of the year you am i ### ### supported hoodoo gurus on a national tour. late in 1992, tim rogers sent samples of you am i's ### ### releases to lee ranaldo of american rock group, sonic youth. in january 1993 both groups appeared ### ### at the second big day ### ### =============================================== v_q ================================================ ### ### ('shepherd', 0, 17659) ('absently', 1, 53) ('monarchs', 2, 13707) ('glide', 3, 12099) ### ### ('##pton', 4, 22368) ('christ', 5, 18319) ('##kie', 6, 27953) ('hon', 7, 25478) ### ### ('neptune', 8, 5456) ('andy', 9, 9259) ('kelly', 10, 7167) ('am', 11, 1) ('##ics', 12, 17531) ### ### ('খ', 13, 141) ('intently', 14, 145) ('greg', 15, 6257) ('challenger', 16, 17707) ('.', 17, 2512) ### ### ('幸', 18, 149) ('you', 19, 6) ### ### =============================================== v_p ================================================ ### ### ('price', 8131, 0) ('am', 11, 1) ('rogers', 8024, 2) ('rana', 15301, 3) ('guru', 14872, 4) ### ### ('ra', 21344, 5) ('you', 19, 6) ('##ld', 25334, 7) ('##oa', 22349, 8) ('tim', 3094, 9) ### ### ('##oo', 5874, 10) ('youth', 5281, 11) ('##rt', 24836, 12) ('hood', 742, 13) ('sonic', 581, 14) ### ### ('we', 2428, 15) ('started', 14415, 16) ('records', 107, 17) ('david', 2632, 18) ### ### ('leaned', 1367, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('am', 11, 1) ('you', 19, 6) ('absently', 1, 53) ('и', 35, 33) ('band', 22, 43) ('i', 44, 20) ### ### ('littered', 50, 28) ('records', 107, 17) ('worried', 80, 26) ('stared', 36, 61) ('anyway', 63, 35) ### ### ('bands', 39, 48) ('খ', 13, 141) ('intently', 14, 145) ('hissed', 48, 44) ('##部', 28, 102) ### ### ('幸', 18, 149) ('ம', 33, 114) ('rocker', 58, 52) ('貴', 20, 180) ### ############################################################################################################ [2023-11-26 02:36:27,139][root][INFO] - Epoch: 5: Step: 11401/19530, loss[v]=0.063688, lr=0.000002, acc@1[1]=1003.0/1024=0.9794921875, acc@1[2]=1018.5/1024=0.99462890625 [2023-11-26 02:38:16,102][root][INFO] - Train batch 11500 [2023-11-26 02:38:16,103][root][INFO] - Avg. loss per last 100 batches: 0.125436 [2023-11-26 02:38:16,816][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.6/29522=99.93% | mean: 0.02 | max: 5.73 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: as a solo performer, he wrote and performed the theme song to the 1984 film " ### ### ghostbusters ". ### ### [P_TEXT]: beatfreakz 2006. they revealed exclusively on bbc radio 1's chart show on 15 october ### ### 2006, that their next project was likely to be a remix of the " ghostbusters " theme tune, original ### ### from ray parker jr. this track was previewed at the matinee closing party in ibiza on 30 september ### ### 2006. another song of theirs is released on the " put your hands up! 2 " compilation, a remake of ### ### aerosmith's " walk this way ". in 2007, the group were credited in a remix ### ### =============================================== v_q ================================================ ### ### ('##bus', 0, 12) ('##ters', 1, 11) ('ghost', 2, 15) ('¹', 3, 528) ('1984', 4, 7737) ('³', 5, 2356) ### ### ('performers', 6, 495) ('theme', 7, 21) ('খ', 8, 145) ('intently', 9, 194) ('貴', 10, 237) ### ### ('pmid', 11, 1673) ('幸', 12, 197) ('film', 13, 11819) ('elvis', 14, 1109) ('##ப', 15, 165) ### ### ('song', 16, 118) ('absently', 17, 55) ('！', 18, 592) ('littered', 19, 17) ### ### =============================================== v_p ================================================ ### ### ('walk', 179, 0) ('##eak', 16928, 1) ('closing', 947, 2) ('way', 5734, 3) ('parker', 6253, 4) ### ### ('##smith', 2299, 5) ('hands', 4314, 6) ('ray', 1924, 7) ('jr', 9300, 8) ('##iza', 8340, 9) ### ### ('##fr', 9389, 10) ('##ters', 1, 11) ('##bus', 0, 12) ('tune', 29, 13) ('mat', 4486, 14) ### ### ('ghost', 2, 15) ('beat', 497, 16) ('littered', 19, 17) ('remix', 2071, 18) ('2006', 9898, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##bus', 0, 12) ('##ters', 1, 11) ('ghost', 2, 15) ('theme', 7, 21) ('littered', 19, 17) ### ### ('tune', 29, 13) ('walk', 179, 0) ('absently', 17, 55) ('anyway', 48, 30) ('lyrical', 166, 25) ### ### ('খ', 8, 145) ('closing', 947, 2) ('song', 16, 118) ('mccartney', 26, 95) ('intently', 9, 194) ### ### ('beat', 497, 16) ('##ப', 15, 165) ('幸', 12, 197) ('songs', 32, 83) ('貴', 10, 237) ### ############################################################################################################ [2023-11-26 02:38:17,321][root][INFO] - Epoch: 5: Step: 11501/19530, loss[v]=0.144170, lr=0.000002, acc@1[1]=979.5/1024=0.95654296875, acc@1[2]=1010.5/1024=0.98681640625 [2023-11-26 02:40:06,858][root][INFO] - Train batch 11600 [2023-11-26 02:40:06,858][root][INFO] - Avg. loss per last 100 batches: 0.116872 [2023-11-26 02:40:07,577][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.8/29522=99.92% | mean: 0.02 | max: 5.77 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: tom flores was on the kansas city chiefs'roster in super bowl iv and coached in super ### ### bowl xv. ### ### [P_TEXT]: nfl on nbc conference ( nfc ) regular season and postseason games ( except those in the ### ### abc package ) and to super bowls xiv and xvi. nbc received the rights to all afc regular - season ### ### and postseason games ( except those in the abc package ) and to super bowls xiii and xv. industry ### ### sources considered it the largest single television package ever negotiated. after the 1977 season, ### ### jack buck, who was the initial host for nbc's first official studio pregame show in 1975, and on ### ### the fourth - tier announcing ### ### =============================================== v_q ================================================ ### ### ('xv', 0, 3) ('flores', 1, 20327) ('absently', 2, 120) ('iv', 3, 675) ('¹', 4, 28) ('幸', 5, 117) ### ### ('খ', 6, 136) ('貴', 7, 108) ('intently', 8, 82) ('tom', 9, 11007) ('super', 10, 20) ### ### ('chiefs', 11, 7530) ('##ப', 12, 137) ('##部', 13, 97) ('³', 14, 109) ('ம', 15, 96) ('.', 16, 62) ### ### ('kansas', 17, 12595) ('nfl', 18, 6) ('bowl', 19, 15) ### ### =============================================== v_p ================================================ ### ### ('nbc', 497, 0) ('xiv', 2001, 1) ('buck', 1945, 2) ('xv', 0, 3) ('xiii', 228, 4) ('xvi', 603, 5) ### ### ('nfl', 18, 6) ('jack', 3614, 7) ('bowls', 137, 8) ('package', 1037, 9) ('##game', 8999, 10) ### ### ('1977', 376, 11) ('packages', 7053, 12) ('pre', 17909, 13) ('1975', 249, 14) ('bowl', 19, 15) ### ### ('postseason', 1139, 16) ('littered', 1432, 17) ('improved', 1835, 18) ('conferences', 1460, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('xv', 0, 3) ('nfl', 18, 6) ('super', 10, 20) ('¹', 4, 28) ('bowl', 19, 15) ('anyway', 26, 21) ### ### ('absently', 2, 120) ('intently', 8, 82) ('幸', 5, 117) ('貴', 7, 108) ('.', 16, 62) ('xiii', 228, 4) ### ### ('খ', 6, 136) ('##部', 13, 97) ('bowls', 137, 8) ('nbc', 497, 0) ('ம', 15, 96) ('³', 14, 109) ### ### ('##ப', 12, 137) ('season', 37, 31) ### ############################################################################################################ [2023-11-26 02:40:08,084][root][INFO] - Epoch: 5: Step: 11601/19530, loss[v]=0.198438, lr=0.000002, acc@1[1]=970.5/1024=0.94775390625, acc@1[2]=999.5/1024=0.97607421875 [2023-11-26 02:41:57,199][root][INFO] - Train batch 11700 [2023-11-26 02:41:57,199][root][INFO] - Avg. loss per last 100 batches: 0.128909 [2023-11-26 02:41:57,909][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.2/29522=99.93% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the town of store heddinge came into existence during the 13th century, and saint ### ### katharina church ( " sct. ### ### [P_TEXT]: stevns peninsula stevns peninsula is a peninsula on sjælland in denmark. it is separated ### ### from sjælland by the three streams stevns a, tryggevælde a and kildea. the main town of the ### ### peninsula is store heddinge, and most of the peninsula is covered by the stevns municipality. ### ### =============================================== v_q ================================================ ### ### ('##dding', 0, 3) ('store', 1, 4) ('##hari', 2, 28445) ('kat', 3, 18039) ('¹', 4, 123) ### ### ('খ', 5, 204) ('church', 6, 1936) ('貴', 7, 124) ('幸', 8, 206) ('intently', 9, 94) ('##e', 10, 166) ### ### ('he', 11, 62) ('³', 12, 1626) ('##ப', 13, 191) ('saint', 14, 11003) ('norwegian', 15, 64) ### ### ('cumbria', 16, 3965) ('sweden', 17, 136) ('norway', 18, 169) ('ம', 19, 80) ### ### =============================================== v_p ================================================ ### ### ('##vn', 26314, 0) ('##lland', 9721, 1) ('##lde', 14783, 2) ('##dding', 0, 3) ('store', 1, 4) ### ### ('ste', 8066, 5) ('worried', 40, 6) ('peninsula', 2617, 7) ('ki', 14705, 8) ('denmark', 28, 9) ### ### ('anyway', 29, 10) ('danish', 21, 11) ('あ', 10687, 12) ('try', 10451, 13) ('##gg', 10104, 14) ### ### ('stream', 2084, 15) ('##æ', 3561, 16) ('tributary', 630, 17) ('tributaries', 7129, 18) ### ### ('##ev', 15768, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##dding', 0, 3) ('store', 1, 4) ('danish', 21, 11) ('denmark', 28, 9) ('anyway', 29, 10) ### ### ('worried', 40, 6) ('municipality', 23, 24) ('town', 27, 34) ('¹', 4, 123) ('he', 11, 62) ### ### ('leaned', 68, 20) ('intently', 9, 94) ('norwegian', 15, 64) ('貴', 7, 124) ('stores', 114, 22) ### ### ('towns', 71, 33) ('copenhagen', 62, 36) ('খ', 5, 204) ('ம', 19, 80) ('##e', 10, 166) ### ############################################################################################################ [2023-11-26 02:41:58,413][root][INFO] - Epoch: 5: Step: 11701/19530, loss[v]=0.174829, lr=0.000002, acc@1[1]=976.5/1024=0.95361328125, acc@1[2]=1000.5/1024=0.97705078125 [2023-11-26 02:43:47,400][root][INFO] - Train batch 11800 [2023-11-26 02:43:47,401][root][INFO] - Avg. loss per last 100 batches: 0.119578 [2023-11-26 02:43:48,118][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.1/29522=99.92% | mean: 0.02 | max: 5.84 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: examples of commercially important cultivars with significant " v. riparia " ancestry ### ### include baco noir, marechal foch, triomphe d'alsace and frontenac. ### ### [P_TEXT]: baco noir baco noir ( pronounced ba - koh nwahr ) is a hybrid red wine grape variety ### ### produced by francois baco from a cross of " vitis vinifera var. " folle blanche, a french wine ### ### grape, and an unknown variety of " vitis riparia " indigenous to north america. in 1951 the variety ### ### was brought to the cooler viticulture regions of north america, such as british columbia, ontario, ### ### nova scotia, new york, michigan, mississippi ### ### =============================================== v_q ================================================ ### ### ('rip', 0, 4) ('##aria', 1, 16) ('##och', 2, 25649) ('noir', 3, 3) ('front', 4, 20404) ### ### ('##ena', 5, 25587) ('absently', 6, 101) ('幸', 7, 208) ('alsace', 8, 2154) ('mare', 9, 13190) ### ### ('ba', 10, 1) ('খ', 11, 175) ('##ivar', 12, 4419) ('wine', 13, 0) ('貴', 14, 164) ('##ப', 15, 189) ### ### ('##chal', 16, 28669) ('##co', 17, 6) ('##mp', 18, 23786) ('v', 19, 7547) ### ### =============================================== v_p ================================================ ### ### ('wine', 13, 0) ('ba', 10, 1) ('francois', 2323, 2) ('noir', 3, 3) ('rip', 0, 4) ('hybrid', 90, 5) ### ### ('##co', 17, 6) ('nwa', 27982, 7) ('blanche', 4770, 8) ('varieties', 51, 9) ('grape', 36, 10) ### ### ('wines', 15538, 11) ('##oll', 18100, 12) ('worried', 22, 13) ('littered', 2419, 14) ### ### ('##tis', 5602, 15) ('##aria', 1, 16) ('michigan', 8157, 17) ('leaned', 179, 18) ### ### ('mississippi', 5597, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('rip', 0, 4) ('noir', 3, 3) ('ba', 10, 1) ('wine', 13, 0) ('##aria', 1, 16) ('##co', 17, 6) ### ### ('worried', 22, 13) ('grape', 36, 10) ('varieties', 51, 9) ('anyway', 31, 27) ('hybrid', 90, 5) ### ### ('french', 41, 22) ('f', 37, 30) ('variety', 43, 39) ('france', 77, 28) ('absently', 6, 101) ### ### ('leaned', 179, 18) ('ம', 20, 95) ('খ', 11, 175) ('貴', 14, 164) ### ############################################################################################################ [2023-11-26 02:43:48,623][root][INFO] - Epoch: 5: Step: 11801/19530, loss[v]=0.049442, lr=0.000002, acc@1[1]=1007.5/1024=0.98388671875, acc@1[2]=1020.0/1024=0.99609375 [2023-11-26 02:45:38,447][root][INFO] - Train batch 11900 [2023-11-26 02:45:38,448][root][INFO] - Avg. loss per last 100 batches: 0.120654 [2023-11-26 02:45:39,170][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.97 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.8/29522=99.93% | mean: 0.02 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: today st james's street contains several of london's best - known gentlemen's clubs, such ### ### as boodle's, brooks's, the carlton club and white's, some exclusive shops and various offices. ### ### [P_TEXT]: carlton club was killed in the explosion, but the building was destroyed. the carlton ### ### immediately moved to its current premises, at 69 st james's street, london, formerly the premises ### ### of arthur's club – one of the premier gentlemen's clubs, which had closed the same year, after 150 ### ### years of operations. the current georgian clubhouse is architecturally important ( grade ii * ### ### listed ) and includes two elegant dining rooms, together with a collection of political portraits ### ### and paintings dating back to the 18th century, imported from ruins of the old club house and the ### ### former junior carlton ### ### =============================================== v_q ================================================ ### ### ('james', 0, 5) ('gentlemen', 1, 4) ('boo', 2, 17694) ('##dle', 3, 20874) ('brooks', 4, 8991) ### ### ('club', 5, 3) ('white', 6, 17288) ('london', 7, 23) ('carlton', 8, 0) ('street', 9, 16) ### ### ('¹', 10, 458) ('幸', 11, 257) ('st', 12, 35) ('intently', 13, 176) ('nightclub', 14, 459) ### ### ('貴', 15, 307) ('খ', 16, 169) ('absently', 17, 62) ('##ப', 18, 215) ('ம', 19, 129) ### ### =============================================== v_p ================================================ ### ### ('carlton', 8, 0) ('arthur', 4163, 1) ('explosion', 5017, 2) ('club', 5, 3) ('gentlemen', 1, 4) ### ### ('james', 0, 5) ('69', 11930, 6) ('clubhouse', 114, 7) ('explosions', 23522, 8) ('clubs', 39, 9) ### ### ('premier', 13648, 10) ('premises', 143, 11) ('dining', 523, 12) ('worried', 34, 13) ### ### ('rooms', 519, 14) ('exploded', 2409, 15) ('street', 9, 16) ('house', 322, 17) ### ### ('fatalities', 21407, 18) ('anyway', 85, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('carlton', 8, 0) ('gentlemen', 1, 4) ('james', 0, 5) ('club', 5, 3) ('street', 9, 16) ### ### ('london', 7, 23) ('st', 12, 35) ('clubs', 39, 9) ('worried', 34, 13) ('absently', 17, 62) ### ### ('clubhouse', 114, 7) ('anyway', 85, 19) ('premises', 143, 11) ('stared', 33, 53) ('18th', 48, 46) ### ### ('intently', 13, 176) ('ம', 19, 129) ('খ', 16, 169) ('leaned', 66, 45) ('幸', 11, 257) ### ############################################################################################################ [2023-11-26 02:45:39,675][root][INFO] - Epoch: 5: Step: 11901/19530, loss[v]=0.079461, lr=0.000002, acc@1[1]=1003.0/1024=0.9794921875, acc@1[2]=1017.0/1024=0.9931640625 [2023-11-26 02:47:28,527][root][INFO] - Train batch 12000 [2023-11-26 02:47:28,528][root][INFO] - Avg. loss per last 100 batches: 0.111844 [2023-11-26 02:47:29,244][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.7/29522=99.93% | mean: 0.02 | max: 5.82 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: although not explicitly written in his article, the fried parameter at wavelength formula ### ### _ 6 can be expressed in terms of the so - called atmospheric turbulence strength formula _ 7 ( ### ### which is actually a function of temperature fluctuations as well as turbulence ) along formula _ 8 ### ### the path of the starlight : where formula _ 10 is the wavenumber. ### ### [P_TEXT]: astronomical seeing of the atmosphere. the parameters " r " 0 and " t " 0 vary with the ### ### wavelength used for the astronomical imaging, allowing slightly higher resolution imaging at longer ### ### wavelengths using large telescopes. the seeing parameter " r " 0 is often known as the fried ### ### parameter ( pronounced " freed " ), named after david l. fried. the atmospheric time constant " t " ### ### 0 is often referred to as the greenwood time constant, after darryl greenwood. mathematical models ### ### can give an accurate model of the effects of astronomical seeing on images taken through ground - ### ### based telescopes. three simulated short - exposure images are shown at the right through ### ### =============================================== v_q ================================================ ### ### ('fried', 0, 3) ('##light', 1, 3781) ('turbulence', 2, 6562) ('path', 3, 3412) ('##num', 4, 17968) ### ### ('star', 5, 3485) ('¹', 6, 315) ('strength', 7, 1401) ('atmospheric', 8, 6) ('wave', 9, 3994) ### ### ('খ', 10, 200) ('turbulent', 11, 23248) ('幸', 12, 253) ('wavelength', 13, 38) ('parameter', 14, 11) ### ### ('貴', 15, 291) ('##ber', 16, 28398) ('absently', 17, 97) ('temperature', 18, 1293) ('.', 19, 2213) ### ### =============================================== v_p ================================================ ### ### ('greenwood', 6679, 0) ('seeing', 2807, 1) ('darryl', 21101, 2) ('fried', 0, 3) ### ### ('exposure', 7196, 4) ('simulated', 11307, 5) ('atmospheric', 8, 6) ('freed', 5213, 7) ### ### ('astronomical', 149, 8) ('littered', 2389, 9) ('telescopes', 21625, 10) ('parameter', 14, 11) ### ### ('constant', 251, 12) ('telescope', 148, 13) ('atmosphere', 54, 14) ('simulate', 22014, 15) ### ### ('imaging', 2972, 16) ('david', 8357, 17) ('resolution', 1131, 18) ('images', 14640, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('fried', 0, 3) ('atmospheric', 8, 6) ('parameter', 14, 11) ('wavelength', 13, 38) ### ### ('atmosphere', 54, 14) ('astronomical', 149, 8) ('telescope', 148, 13) ('anyway', 66, 24) ### ### ('constant', 251, 12) ('accurate', 152, 22) ('absently', 17, 97) ('parameters', 223, 20) ### ### ('ம', 21, 101) ('খ', 10, 200) ('fry', 110, 36) ('¹', 6, 315) ('##部', 27, 82) ('seeing', 2807, 1) ### ### ('幸', 12, 253) ('model', 231, 39) ### ############################################################################################################ [2023-11-26 02:47:29,748][root][INFO] - Epoch: 5: Step: 12001/19530, loss[v]=0.165262, lr=0.000002, acc@1[1]=979.5/1024=0.95654296875, acc@1[2]=1010.5/1024=0.98681640625 [2023-11-26 02:49:18,952][root][INFO] - Train batch 12100 [2023-11-26 02:49:18,953][root][INFO] - Avg. loss per last 100 batches: 0.120682 [2023-11-26 02:49:19,680][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.1/29522=99.93% | mean: 0.02 | max: 5.77 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 2012, it won an appeal to become an official motorway service station and was sold to ### ### moto in 2014. ### ### [P_TEXT]: wetherby services, although the town of wetherby is on the west yorkshire side of the ### ### boundary. it is operated by moto hospitality. the next services in each direction are at leeming ### ### bar ( north ) and ferrybridge ( south ). the service station uses the latest'green technologies'in ### ### its construction and running, making it the uk's first carbon neutral service station. in a 2018 ### ### user survey of england's motorway service areas ( msas ), wetherby scored a 99 % satisfaction ### ### rating, putting it third in the country and the top moto site overall. in ### ### =============================================== v_q ================================================ ### ### ('mo', 0, 18) ('##to', 1, 19) ('motorway', 2, 13) ('¹', 3, 165) ('貴', 4, 173) ('intently', 5, 99) ### ### ('station', 6, 31) ('幸', 7, 177) ('খ', 8, 161) ('##ப', 9, 187) ('service', 10, 23) ### ### ('anyway', 11, 29) ('ம', 12, 118) ('appeal', 13, 7384) ('³', 14, 335) ('2012', 15, 17648) ### ### ('refurbishment', 16, 2034) ('.', 17, 9704) ('##部', 18, 95) ('stations', 19, 48) ### ### =============================================== v_p ================================================ ### ### ('hospitality', 590, 0) ('ferry', 1037, 1) ('wet', 9665, 2) ('carbon', 867, 3) ### ### ('satisfaction', 3992, 4) ('technologies', 4497, 5) ('neutral', 6610, 6) ('##ming', 5912, 7) ### ### ('##her', 18607, 8) ('rating', 3534, 9) ('services', 229, 10) ('lee', 13806, 11) ### ### ('##bridge', 3812, 12) ('motorway', 2, 13) ('bar', 6256, 14) ('ms', 9515, 15) ### ### ('yorkshire', 2352, 16) ('##by', 5904, 17) ('mo', 0, 18) ('##to', 1, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('mo', 0, 18) ('motorway', 2, 13) ('##to', 1, 19) ('service', 10, 23) ('station', 6, 31) ### ### ('anyway', 11, 29) ('worried', 32, 26) ('intently', 5, 99) ('stations', 19, 48) ('¹', 3, 165) ### ### ('hospitality', 590, 0) ('services', 229, 10) ('ம', 12, 118) ('貴', 4, 173) ('খ', 8, 161) ### ### ('幸', 7, 177) ('carbon', 867, 3) ('ferry', 1037, 1) ('absently', 21, 83) ('sold', 24, 81) ### ############################################################################################################ [2023-11-26 02:49:20,186][root][INFO] - Epoch: 5: Step: 12101/19530, loss[v]=0.142611, lr=0.000002, acc@1[1]=977.5/1024=0.95458984375, acc@1[2]=1003.5/1024=0.97998046875 [2023-11-26 02:51:10,107][root][INFO] - Train batch 12200 [2023-11-26 02:51:10,108][root][INFO] - Avg. loss per last 100 batches: 0.124685 [2023-11-26 02:51:10,824][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.8/29522=99.92% | mean: 0.02 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the sweetgrass arch in montana divides the two medicine from the judith river formation, ### ### bearpaw shale, claggett shale, and eagle sandstone. ### ### [P_TEXT]: gorgosaurus cannot be said for certain which animal they belonged to. " deinodon " is ### ### usually regarded as a " nomen dubium " today. additional likely synonyms of " g. libratus " and / ### ### or " d. horridus " include " laelaps falculus ", " laelaps hazenianus ", " laelaps incrassatus ", ### ### and " dryptosaurus kenabekides ". several tyrannosaurid skeletons from the two medicine formation ### ### and judith river formation of montana probably belong to " gorgosaurus ", although it remains ### ### uncertain whether they belong to " g. libratus " or a new species. one specimen from montana ( tcmi ### ### 2001. 89. 1 ), housed in the children's museum of indianapolis, shows evidence of severe ### ### pathologies ### ### =============================================== v_q ================================================ ### ### ('medicine', 0, 2) ('judith', 1, 0) ('montana', 2, 5) ('bear', 3, 15507) ('shale', 4, 1162) ### ### ('eagle', 5, 11502) ('খ', 6, 187) ('sweet', 7, 6381) ('arch', 8, 7653) ('absently', 9, 140) ### ### ('formation', 10, 8) ('幸', 11, 230) ('sandstone', 12, 328) ('¹', 13, 117) ('.', 14, 11770) ### ### ('##ப', 15, 197) ('貴', 16, 250) ('##get', 17, 28310) ('##部', 18, 122) ('##w', 19, 23827) ### ### =============================================== v_p ================================================ ### ### ('judith', 1, 0) ('##rgos', 25931, 1) ('medicine', 0, 2) ('haze', 9318, 3) ('dei', 27113, 4) ### ### ('montana', 2, 5) ('tc', 11109, 6) ('dub', 20792, 7) ('formation', 10, 8) ('##abe', 9944, 9) ### ### ('##ologies', 13910, 10) ('##aurus', 75, 11) ('indianapolis', 4864, 12) ('children', 10914, 13) ### ### ('##ium', 5003, 14) ('##sat', 27532, 15) ('museum', 1279, 16) ('##dus', 8267, 17) ### ### ('skeleton', 2157, 18) ('##tus', 2809, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('judith', 1, 0) ('medicine', 0, 2) ('montana', 2, 5) ('formation', 10, 8) ('river', 25, 36) ### ### ('##aurus', 75, 11) ('¹', 13, 117) ('absently', 9, 140) ('species', 41, 44) ('anyway', 104, 25) ### ### ('খ', 6, 187) ('##部', 18, 122) ('##ப', 15, 197) ('幸', 11, 230) ('ம', 22, 157) ### ### ('yellowstone', 99, 46) ('貴', 16, 250) ('intently', 23, 152) ('luckily', 42, 77) ### ### ('##isance', 69, 63) ### ############################################################################################################ [2023-11-26 02:51:11,329][root][INFO] - Epoch: 5: Step: 12201/19530, loss[v]=0.141896, lr=0.000002, acc@1[1]=976.0/1024=0.953125, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-26 02:53:00,569][root][INFO] - Train batch 12300 [2023-11-26 02:53:00,570][root][INFO] - Avg. loss per last 100 batches: 0.121169 [2023-11-26 02:53:01,297][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.94 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.0/29522=99.92% | mean: 0.02 | max: 5.88 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: cut in collaboration with jo perry and shaznay lewis, " black heart " was released in the ### ### united kingdom by warner music uk on 17 june 2012. ### ### [P_TEXT]: shaznay lewis cocknbullkid track " distractions " from her 2011 album " adulthood " and ### ### the stooshe single " black heart ", released as a single in 2012. in august 2012, it was reported ### ### lewis was in the studio with original sugababes line - up mutya keisha siobhan, writing new ### ### material for their debut album. in 2019 shaznay co - wrote the theme music for the netflix original ### ### series turn up charlie with idris elba in 2014, all saints reformed to support the backstreet boys ### ### for five dates across the uk and ireland in 2014. on 27 january ### ### =============================================== v_q ================================================ ### ### ('jo', 0, 22597) ('heart', 1, 6) ('perry', 2, 2557) ('¹', 3, 153) ('black', 4, 11) ('##nay', 5, 2) ### ### ('warner', 6, 13951) ('sha', 7, 10) ('lewis', 8, 4) ('³', 9, 1086) ('absently', 10, 77) ### ### ('幸', 11, 161) ('খ', 12, 175) ('intently', 13, 159) ('貴', 14, 203) ('2012', 15, 49) ('uk', 16, 43) ### ### ('minogue', 17, 940) ('##ப', 18, 169) ('##z', 19, 41) ### ### =============================================== v_p ================================================ ### ### ('adulthood', 26569, 0) ('distraction', 2903, 1) ('##nay', 5, 2) ('saints', 3720, 3) ### ### ('lewis', 8, 4) ('turn', 9595, 5) ('heart', 1, 6) ('##oo', 14331, 7) ('charlie', 4846, 8) ### ### ('##tya', 15599, 9) ('sha', 7, 10) ('black', 4, 11) ('##ull', 27580, 12) ('netflix', 4641, 13) ### ### ('##sh', 21202, 14) ('kei', 7453, 15) ('##ba', 16114, 16) ('id', 13068, 17) ('cock', 3828, 18) ### ### ('″', 694, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('heart', 1, 6) ('##nay', 5, 2) ('lewis', 8, 4) ('black', 4, 11) ('sha', 7, 10) ('uk', 16, 43) ### ### ('2012', 15, 49) ('##z', 19, 41) ('absently', 10, 77) ('¹', 3, 153) ('hissed', 30, 38) ### ### ('幸', 11, 161) ('intently', 13, 159) ('খ', 12, 175) ('lyrical', 118, 29) ('unsure', 27, 91) ### ### ('acts', 41, 69) ('貴', 14, 203) ('##ப', 18, 169) ('music', 25, 117) ### ############################################################################################################ [2023-11-26 02:53:01,801][root][INFO] - Epoch: 5: Step: 12301/19530, loss[v]=0.169564, lr=0.000002, acc@1[1]=997.0/1024=0.9736328125, acc@1[2]=1011.0/1024=0.9873046875 [2023-11-26 02:54:51,445][root][INFO] - Train batch 12400 [2023-11-26 02:54:51,446][root][INFO] - Avg. loss per last 100 batches: 0.113851 [2023-11-26 02:54:52,218][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.0/29522=99.92% | mean: 0.02 | max: 5.88 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.7/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: persichetti's students included einojuhani rautavaara, leonardo balada, steven gellman, ### ### peter schickele ( p. d. q. ### ### [P_TEXT]: einojuhani rautavaara despite having, according to rautavaara, "... absolutely no ### ### experience of writing for brass band and [ a ] composition technique [ that ] was immature at the ### ### time... " nevertheless, the work prompted jean sibelius to recommend him for a scholarship to study ### ### at the juilliard school in new york city. there he was taught by vincent persichetti, and he also ### ### took lessons from roger sessions and aaron copland at tanglewood. reflecting on his time enrolled ### ### in julliard, rautavaara later said that living in manhattan was, " maybe ### ### =============================================== v_q ================================================ ### ### ('##uta', 0, 4) ('bala', 1, 23549) ('##hett', 2, 5) ('¹', 3, 126) ('gel', 4, 23551) ### ### ('leonardo', 5, 7369) ('##hic', 6, 24138) ('##kel', 7, 28135) ('absently', 8, 89) ('##oj', 9, 12) ### ### ('##sic', 10, 10) ('steven', 11, 5604) ('##da', 12, 26336) ('##lman', 13, 25654) ('幸', 14, 217) ### ### ('intently', 15, 210) ('##ara', 16, 90) ('খ', 17, 170) ('per', 18, 44) ('貴', 19, 224) ### ### =============================================== v_p ================================================ ### ### ('tangle', 13963, 0) ('sessions', 3965, 1) ('jul', 25207, 2) ('roger', 5578, 3) ('##uta', 0, 4) ### ### ('##hett', 2, 5) ('cop', 14665, 6) ('ein', 24, 7) ('##lia', 6185, 8) ('##wood', 24133, 9) ### ### ('##sic', 10, 10) ('##iard', 8423, 11) ('##oj', 9, 12) ('##bel', 23773, 13) ('##ius', 3275, 14) ### ### ('aaron', 8264, 15) ('jean', 11252, 16) ('enrolled', 7922, 17) ('scholarships', 21753, 18) ### ### ('manhattan', 5584, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##uta', 0, 4) ('##hett', 2, 5) ('##sic', 10, 10) ('##oj', 9, 12) ('ein', 24, 7) ('##uh', 25, 23) ### ### ('anyway', 22, 33) ('per', 18, 44) ('##va', 32, 35) ('##ani', 28, 46) ('absently', 8, 89) ### ### ('studied', 31, 48) ('¹', 3, 126) ('##ara', 16, 90) ('ra', 48, 38) ('composition', 115, 25) ### ### ('littered', 166, 21) ('leaned', 138, 24) ('students', 107, 32) ('technique', 70, 40) ### ############################################################################################################ [2023-11-26 02:54:52,726][root][INFO] - Epoch: 5: Step: 12401/19530, loss[v]=0.076671, lr=0.000002, acc@1[1]=1002.5/1024=0.97900390625, acc@1[2]=1020.5/1024=0.99658203125 [2023-11-26 02:56:42,474][root][INFO] - Train batch 12500 [2023-11-26 02:56:42,475][root][INFO] - Avg. loss per last 100 batches: 0.123416 [2023-11-26 02:56:43,193][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.9/29522=99.93% | mean: 0.02 | max: 5.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in the " house of m " reality, hood assembles a group of villains to form the masters of ### ### evil in defiance to magneto. ### ### [P_TEXT]: madison jeffries of its precious metal ores. they ruled the country despotically and ### ### experimented on its residents for years until hood's masters of evil invaded santo rico. madison ### ### and lionel both grow significantly in size and power by physically absorbing dozens of santo ricans ### ### and the mined minerals. regardless, they were outnumbered by the masters who killed the jeffries ### ### brothers in order to conquer santo rico. in the alternate future depicted in " weapon x : days of ### ### future now ", madison's ever - improving boxbots have gained a sentience of their own, one of them ### ### even calling itself master mold. ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 176) ('hood', 1, 0) ('masters', 2, 3) ('magnet', 3, 2742) ('m', 4, 22792) ### ### ('evil', 5, 11) ('খ', 6, 205) ('幸', 7, 224) ('貴', 8, 241) ('house', 9, 15560) ('intently', 10, 229) ### ### ('¹', 11, 128) ('##o', 12, 27777) ('##ப', 13, 218) ('ம', 14, 163) ('！', 15, 1434) ('##部', 16, 140) ### ### ('freaking', 17, 1442) ('.', 18, 111) ('³', 19, 2095) ### ### =============================================== v_p ================================================ ### ### ('hood', 1, 0) ('madison', 3339, 1) ('lionel', 13366, 2) ('masters', 2, 3) ('mold', 1416, 4) ### ### ('##ries', 24628, 5) ('##bots', 2698, 6) ('weapon', 237, 7) ('santo', 13099, 8) ('rico', 3015, 9) ### ### ('jeff', 3127, 10) ('evil', 5, 11) ('box', 10120, 12) ('rican', 18571, 13) ('##pot', 7869, 14) ### ### ('robots', 11121, 15) ('future', 760, 16) ('leaned', 2853, 17) ('x', 29, 18) ('alternate', 532, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('hood', 1, 0) ('masters', 2, 3) ('evil', 5, 11) ('x', 29, 18) ('anyway', 26, 26) ### ### ('absently', 0, 176) ('master', 51, 20) ('weapon', 237, 7) ('¹', 11, 128) ('খ', 6, 205) ### ### ('幸', 7, 224) ('##部', 16, 140) ('ம', 14, 163) ('.', 18, 111) ('貴', 8, 241) ('intently', 10, 229) ### ### ('villains', 24, 82) ('mold', 1416, 4) ('##ப', 13, 218) ('revealed', 40, 63) ### ############################################################################################################ [2023-11-26 02:56:43,698][root][INFO] - Epoch: 5: Step: 12501/19530, loss[v]=0.185196, lr=0.000002, acc@1[1]=973.5/1024=0.95068359375, acc@1[2]=1001.0/1024=0.9775390625 [2023-11-26 02:58:34,056][root][INFO] - Train batch 12600 [2023-11-26 02:58:34,057][root][INFO] - Avg. loss per last 100 batches: 0.118787 [2023-11-26 02:58:34,820][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.6/29522=99.92% | mean: 0.02 | max: 5.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: police divers are usually sworn police officers, and may either be employed full - time ### ### as divers or as general water police officers, or be volunteers who usually serve in other units ### ### but are called in if their diving services are required. ### ### [P_TEXT]: devon and cornwall police officers that patrol on police motorcycles. motorcycle officers ### ### are exempt from wearing body armour as they have to wear motorcycle leathers. the force support ### ### group ( fsg ), previously called the tactical aid group ( tag ), are predominantly responsible for ### ### public order, marine operations, searches and dealing with potentially violent offenders. it is ### ### divided into several sections, including two trained in firearms and one in marine operations. the ### ### fsg's marine support unit, also known as force support group ( fsg ) d section is responsible for ### ### underwater search and marine operations. it consists of one sergeant and six constables, all ### ### =============================================== v_q ================================================ ### ### ('police', 0, 9) ('water', 1, 1097) ('¹', 2, 390) ('diving', 3, 70) ('貴', 4, 205) ### ### ('intently', 5, 191) ('幸', 6, 273) ('volunteer', 7, 6189) ('diver', 8, 256) ('.', 9, 562) ### ### ('³', 10, 2460) ('##ப', 11, 249) ('swimmer', 12, 8319) ('sponge', 13, 1485) ('খ', 14, 215) ### ### ('aqua', 15, 1335) ('absently', 16, 173) ('ம', 17, 144) ('divers', 18, 88) ('general', 19, 11597) ### ### =============================================== v_p ================================================ ### ### ('underwater', 21, 0) ('leather', 3848, 1) ('searches', 27018, 2) ('armour', 9035, 3) ### ### ('##sg', 26264, 4) ('motorcycle', 1128, 5) ('search', 3035, 6) ('sergeant', 101, 7) ### ### ('motorcycles', 16558, 8) ('police', 0, 9) ('aid', 1287, 10) ('cornwall', 415, 11) ### ### ('constable', 489, 12) ('tag', 8202, 13) ('tactical', 2944, 14) ('marine', 66, 15) ### ### ('devon', 20653, 16) ('support', 3321, 17) ('bikes', 28752, 18) ('anyway', 87, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('police', 0, 9) ('underwater', 21, 0) ('diving', 3, 70) ('marine', 66, 15) ('officers', 25, 34) ### ### ('sergeant', 101, 7) ('worried', 20, 54) ('anyway', 87, 19) ('divers', 18, 88) ('貴', 4, 205) ### ### ('intently', 5, 191) ('cornwall', 415, 11) ('body', 131, 30) ('officer', 22, 107) ### ### ('constable', 489, 12) ('ம', 17, 144) ('¹', 2, 390) ('absently', 16, 173) ('##部', 27, 91) ### ### ('幸', 6, 273) ### ############################################################################################################ [2023-11-26 02:58:35,326][root][INFO] - Epoch: 5: Step: 12601/19530, loss[v]=0.155605, lr=0.000002, acc@1[1]=979.0/1024=0.9560546875, acc@1[2]=1004.0/1024=0.98046875 [2023-11-26 03:00:25,159][root][INFO] - Train batch 12700 [2023-11-26 03:00:25,160][root][INFO] - Avg. loss per last 100 batches: 0.119365 [2023-11-26 03:00:25,883][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.94 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29496.8/29522=99.91% | mean: 0.02 | max: 5.77 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the united states naval surface warfare center dahlgren division ( nswcdd ), named for ### ### rear admiral john a. dahlgren, is located in dahlgren, virginia, with a geographically separated ### ### command, " naval surface warfare center dahlgren division dam neck " activity ( nswcdddna ), ### ### located in virginia beach, va, in close proximity to the largest fleet concentration area in the ### ### navy. ### ### [P_TEXT]: hardware bug with a robot in his short story " catch that rabbit ", published in 1944. ### ### the term " bug " was used in an account by computer pioneer grace hopper, who publicized the cause ### ### of a malfunction in an early electromechanical computer. a typical version of the story is : hopper ### ### did not find the bug, as she readily acknowledged. the date in the log book was september 9, 1947. ### ### the operators who found it, including william " bill " burke, later of the naval weapons ### ### laboratory, dahlgren, virginia, were familiar with the engineering ### ### =============================================== v_q ================================================ ### ### ('##gren', 0, 5) ('absently', 1, 85) ('dahl', 2, 7) ('¹', 3, 444) ('virginia', 4, 20) ### ### ('concentration', 5, 6310) ('খ', 6, 182) ('幸', 7, 249) ('intently', 8, 185) ('navy', 9, 46) ### ### ('##ப', 10, 240) ('neck', 11, 8337) ('beach', 12, 10649) ('naval', 13, 15) ('³', 14, 2432) ### ### ('貴', 15, 263) ('fleet', 16, 722) ('surface', 17, 4300) ('##部', 18, 95) ('division', 19, 12535) ### ### =============================================== v_p ================================================ ### ### ('bug', 2562, 0) ('hardware', 6584, 1) ('burke', 6015, 2) ('hopper', 3546, 3) ('catch', 10672, 4) ### ### ('##gren', 0, 5) ('rabbit', 16534, 6) ('dahl', 2, 7) ('log', 9791, 8) ('grace', 9941, 9) ### ### ('laboratory', 2859, 10) ('weapons', 1260, 11) ('robots', 26054, 12) ('electro', 22784, 13) ### ### ('robot', 10912, 14) ('naval', 13, 15) ('computer', 4403, 16) ('bill', 12030, 17) ('″', 2646, 18) ### ### ('william', 4940, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##gren', 0, 5) ('dahl', 2, 7) ('virginia', 4, 20) ('naval', 13, 15) ('navy', 9, 46) ### ### ('absently', 1, 85) ('খ', 6, 182) ('##部', 18, 95) ('anyway', 106, 26) ('intently', 8, 185) ### ### ('snuggled', 34, 59) ('幸', 7, 249) ('##ப', 10, 240) ('¹', 3, 444) ('bug', 2562, 0) ('ம', 23, 129) ### ### ('leaned', 144, 38) ('cradled', 66, 52) ('貴', 15, 263) ('worried', 342, 30) ### ############################################################################################################ [2023-11-26 03:00:26,389][root][INFO] - Epoch: 5: Step: 12701/19530, loss[v]=0.177941, lr=0.000002, acc@1[1]=982.0/1024=0.958984375, acc@1[2]=1002.0/1024=0.978515625 [2023-11-26 03:02:15,628][root][INFO] - Train batch 12800 [2023-11-26 03:02:15,629][root][INFO] - Avg. loss per last 100 batches: 0.125254 [2023-11-26 03:02:16,337][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.4/29522=99.92% | mean: 0.02 | max: 5.71 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.5/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: one such reaction to the loss of meaning is what nietzsche calls " passive nihilism ", ### ### which he recognizes in the pessimistic philosophy of schopenhauer. ### ### [P_TEXT]: rick sanchez ( rick and morty ) richard " rick " sanchez is one of the two eponymous ### ### protagonists from the adult swim animated television series " rick and morty ". created by justin ### ### roiland and dan harmon, sanchez is a misanthropic alcoholic scientist inspired by emmett " doc " ### ### brown from " back to the future " and mister fantastic from marvel comics. known for his reckless, ### ### nihilistic behavior and pessimistic personality, the character has been well received. he is a ### ### sociopathic mad scientist who seems ### ### =============================================== v_q ================================================ ### ### ('nietzsche', 0, 6526) ('##ssi', 1, 21) ('absently', 2, 99) ('¹', 3, 179) ('貴', 4, 192) ### ### ('passive', 5, 3760) ('##auer', 6, 27687) ('intently', 7, 189) ('幸', 8, 200) ('খ', 9, 188) ### ### ('ni', 10, 177) ('##hope', 11, 24835) ('pe', 12, 39) ('##hil', 13, 9) ('##ப', 14, 162) ### ### ('##mist', 15, 16) ('ம', 16, 155) ('.', 17, 265) ('##部', 18, 126) ('##nh', 19, 26786) ### ### =============================================== v_p ================================================ ### ### ('reckless', 6105, 0) ('sanchez', 24110, 1) ('rick', 19781, 2) ('harmon', 19306, 3) ### ### ('mort', 18160, 4) ('emmett', 14005, 5) ('mister', 18026, 6) ('roi', 11612, 7) ### ### ('fantastic', 1285, 8) ('##hil', 13, 9) ('back', 11433, 10) ('brown', 14058, 11) ### ### ('leaned', 3890, 12) ('doc', 8522, 13) ('justin', 14402, 14) ('future', 5189, 15) ### ### ('##mist', 15, 16) ('marvel', 21457, 17) ('worried', 22, 18) ('swim', 18001, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##hil', 13, 9) ('##ssi', 1, 21) ('##mist', 15, 16) ('worried', 22, 18) ('pe', 12, 39) ### ### ('absently', 2, 99) ('anyway', 25, 35) ('¹', 3, 179) ('貴', 4, 192) ('intently', 7, 189) ### ### ('ni', 10, 177) ('##部', 18, 126) ('খ', 9, 188) ('##ப', 14, 162) ('幸', 8, 200) ('ம', 16, 155) ### ### ('stared', 23, 107) ('fantastic', 1285, 8) ('.', 17, 265) ('noticing', 35, 130) ### ############################################################################################################ [2023-11-26 03:02:16,841][root][INFO] - Epoch: 5: Step: 12801/19530, loss[v]=0.175472, lr=0.000002, acc@1[1]=976.5/1024=0.95361328125, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-26 03:04:06,353][root][INFO] - Train batch 12900 [2023-11-26 03:04:06,354][root][INFO] - Avg. loss per last 100 batches: 0.125512 [2023-11-26 03:04:07,092][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.6/29522=99.92% | mean: 0.02 | max: 5.83 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: interim closer sean marshall struggled early in the season, and chapman was given the ### ### closer role in late may. ### ### [P_TEXT]: sean marshall ( baseball ) for games finished will be added to his current salary this ### ### year, as the starts bonus was already in his existing contract. after a season - ending injury to ### ### closer ryan madson, marshall was named the reds'closer for the 2012 season. however, due to ### ### marshall's poor performance in the closing role, the reds named phenom aroldis chapman the closer ### ### in late may 2012. marshall is unusual in that he throws breaking ball s more than any other type of ### ### pitch ; in 2011, almost 75 percent of his pitches were either curveball s ### ### =============================================== v_q ================================================ ### ### ('marshall', 0, 1) ('sean', 1, 2) ('chapman', 2, 4) ('¹', 3, 146) ('³', 4, 2239) ('closer', 5, 21) ### ### ('intently', 6, 181) ('interim', 7, 5639) ('貴', 8, 178) ('幸', 9, 184) ('baseball', 10, 17) ### ### ('season', 11, 30) ('খ', 12, 134) ('saber', 13, 5254) ('lacrosse', 14, 3745) ('##ப', 15, 161) ### ### ('wnba', 16, 15878) ('ம', 17, 103) ('nfl', 18, 1030) ('tampa', 19, 2260) ### ### =============================================== v_p ================================================ ### ### ('breaking', 458, 0) ('marshall', 0, 1) ('sean', 1, 2) ('pitch', 248, 3) ('chapman', 2, 4) ### ### ('pitches', 23951, 5) ('throws', 26351, 6) ('##ball', 2457, 7) ('reds', 1161, 8) ('ball', 173, 9) ### ### ('mad', 14138, 10) ('ryan', 89, 11) ('##old', 24416, 12) ('curve', 2930, 13) ('2012', 796, 14) ### ### ('breaks', 8359, 15) ('littered', 2914, 16) ('baseball', 10, 17) ('batter', 6588, 18) ### ### ('anyway', 58, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('marshall', 0, 1) ('sean', 1, 2) ('chapman', 2, 4) ('closer', 5, 21) ('baseball', 10, 17) ### ### ('season', 11, 30) ('¹', 3, 146) ('ryan', 89, 11) ('pitch', 248, 3) ('ball', 173, 9) ### ### ('anyway', 58, 19) ('breaking', 458, 0) ('salary', 46, 29) ('reds', 1161, 8) ('2012', 796, 14) ### ### ('intently', 6, 181) ('খ', 12, 134) ('貴', 8, 178) ('幸', 9, 184) ('##部', 25, 77) ### ############################################################################################################ [2023-11-26 03:04:07,598][root][INFO] - Epoch: 5: Step: 12901/19530, loss[v]=0.067644, lr=0.000002, acc@1[1]=1004.5/1024=0.98095703125, acc@1[2]=1018.0/1024=0.994140625 [2023-11-26 03:05:57,375][root][INFO] - Train batch 13000 [2023-11-26 03:05:57,376][root][INFO] - Avg. loss per last 100 batches: 0.120433 [2023-11-26 03:05:58,098][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.7/29522=99.92% | mean: 0.02 | max: 5.83 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.1/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.5/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: hallvard flatland ( born 10 february 1957 in flatdal, seljord ), is a norwegian ### ### television presenter. ### ### [P_TEXT]: flatdal flatdal ( lit. " flat valley " ) is a village, parish and valley in the ### ### municipality of seljord in vestfold og telemark, norway. the lake flatsja has been partly drained ### ### to gain more land for agriculture. among notable people from flatdal are painter terje grøstad, who ### ### settled in flatdal, television presenter hallvard flatland, who was born in flatdal in 1957, and ### ### novelist helga flatland, who grew up in flatdal. ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 589) ('flat', 1, 0) ('##land', 2, 17) ('³', 3, 2062) ('##vard', 4, 10) ### ### ('presenter', 5, 21) ('norwegian', 6, 33) ('##dal', 7, 3) ('se', 8, 9) ('hall', 9, 18) ### ### ('hissed', 10, 150) ('幸', 11, 268) ('norway', 12, 30) ('খ', 13, 223) ('##ப', 14, 209) ### ### ('littered', 15, 7) ('！', 16, 4083) ('television', 17, 60) ('viewer', 18, 6450) ### ### ('intently', 19, 250) ### ### =============================================== v_p ================================================ ### ### ('flat', 1, 0) ('flats', 1722, 1) ('##lga', 5659, 2) ('##dal', 7, 3) ('##stad', 476, 4) ### ### ('worried', 1374, 5) ('##jord', 21, 6) ('littered', 15, 7) ('drained', 23471, 8) ('se', 8, 9) ### ### ('##vard', 4, 10) ('lake', 5047, 11) ('painter', 14419, 12) ('ter', 3291, 13) ('##ema', 10689, 14) ### ### ('anyway', 483, 15) ('presenters', 80, 16) ('##land', 2, 17) ('hall', 9, 18) ('##ja', 4606, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('flat', 1, 0) ('##dal', 7, 3) ('##land', 2, 17) ('##vard', 4, 10) ('se', 8, 9) ### ### ('presenter', 5, 21) ('norwegian', 6, 33) ('hall', 9, 18) ('littered', 15, 7) ('##jord', 21, 6) ### ### ('norway', 12, 30) ('##l', 35, 23) ('##ø', 26, 39) ('television', 17, 60) ('fjord', 54, 29) ### ### ('presenters', 80, 16) ('municipality', 52, 32) ('hissed', 10, 150) ('¹', 0, 589) ### ### ('##stad', 476, 4) ### ############################################################################################################ [2023-11-26 03:05:58,603][root][INFO] - Epoch: 5: Step: 13001/19530, loss[v]=0.179043, lr=0.000002, acc@1[1]=977.0/1024=0.9541015625, acc@1[2]=1005.5/1024=0.98193359375 [2023-11-26 03:07:48,064][root][INFO] - Train batch 13100 [2023-11-26 03:07:48,064][root][INFO] - Avg. loss per last 100 batches: 0.128584 [2023-11-26 03:07:48,791][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.1/29522=99.92% | mean: 0.02 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: and is operated by meridian energy. ### ### [P_TEXT]: burrinjuck dam wade ( of the new south wales public works department ) and built from ### ### 1907 to 1927 by lane & peters, sydney. it is also known as barren jack dam and barrenjack. the ### ### property was owned by department of planning and infrastructure ( state government ). it was added ### ### to the new south wales state heritage register on 2 april 1999. in march 2018 meridian energy ### ### australia purchased gsp energy pty ltd which included the hume, burrinjuck and keepit power ### ### stations. these were previously green state power hydro assets. in october 2019 meridian announced ### ### a ### ### =============================================== v_q ================================================ ### ### ('meridian', 0, 1) ('energy', 1, 17) ('¹', 2, 210) ('intently', 3, 119) ('³', 4, 851) ('貴', 5, 153) ### ### ('幸', 6, 226) ('power', 7, 18) ('operated', 8, 306) ('.', 9, 6836) ('##ப', 10, 194) ('খ', 11, 202) ### ### ('ம', 12, 117) ('operator', 13, 242) ('energies', 14, 1084) ('operators', 15, 1564) ### ### ('electricity', 16, 31) ('services', 17, 373) ('livery', 18, 397) ('nuclear', 19, 1053) ### ### =============================================== v_p ================================================ ### ### ('hume', 15967, 0) ('meridian', 0, 1) ('peters', 13108, 2) ('wade', 18147, 3) ('barren', 8634, 4) ### ### ('gs', 6762, 5) ('keep', 853, 6) ('burr', 26327, 7) ('lane', 5840, 8) ('##jack', 13158, 9) ### ### ('##ju', 28088, 10) ('##it', 25507, 11) ('dam', 3429, 12) ('sydney', 412, 13) ('pt', 5978, 14) ### ### ('green', 2400, 15) ('planning', 1151, 16) ('energy', 1, 17) ('power', 7, 18) ('worried', 741, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('meridian', 0, 1) ('energy', 1, 17) ('power', 7, 18) ('electricity', 16, 31) ### ### ('infrastructure', 26, 28) ('anyway', 41, 23) ('hydro', 115, 21) ('¹', 2, 210) ('intently', 3, 119) ### ### ('ltd', 48, 39) ('keep', 853, 6) ('sydney', 412, 13) ('station', 29, 51) ('貴', 5, 153) ### ### ('stations', 98, 41) ('hissed', 32, 60) ('wales', 422, 27) ('幸', 6, 226) ('worried', 741, 19) ### ### ('ம', 12, 117) ### ############################################################################################################ [2023-11-26 03:07:49,297][root][INFO] - Epoch: 5: Step: 13101/19530, loss[v]=0.096705, lr=0.000002, acc@1[1]=996.0/1024=0.97265625, acc@1[2]=1013.5/1024=0.98974609375 [2023-11-26 03:09:39,125][root][INFO] - Train batch 13200 [2023-11-26 03:09:39,125][root][INFO] - Avg. loss per last 100 batches: 0.119192 [2023-11-26 03:09:39,859][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.97 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.8/29522=99.92% | mean: 0.02 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the northern part of the county drains to the washita river, while several creeks drain ### ### the southern part directly to the red river. ### ### [P_TEXT]: washita river the oklahoma counties of roger mills, custer, washita, caddo, grady, ### ### garvin, murray, carter, and johnston before emptying into lake texoma, which is the modern border ### ### between bryan county and marshall county. the river bisects the heart of the anadarko basin, the ### ### fifth - largest natural gas formation area in the united states. when the river reaches the ### ### arbuckle mountains ### ### =============================================== v_q ================================================ ### ### ('##ita', 0, 10) ('red', 1, 16164) ('wash', 2, 3) ('county', 3, 12) ('river', 4, 27) ('¹', 5, 26) ### ### ('counties', 6, 30) ('খ', 7, 100) ('貴', 8, 95) ('intently', 9, 52) ('幸', 10, 109) ('##ப', 11, 115) ### ### ('dakota', 12, 931) ('minnesota', 13, 1464) ('³', 14, 76) ('ம', 15, 70) ('absently', 16, 125) ### ### ('.', 17, 96) ('oklahoma', 18, 18) ('##部', 19, 99) ### ### =============================================== v_p ================================================ ### ### ('##buck', 24580, 0) ('johnston', 2143, 1) ('carter', 7999, 2) ('wash', 2, 3) ('grady', 8262, 4) ### ### ('bryan', 2850, 5) ('custer', 415, 6) ('marshall', 1449, 7) ('murray', 2936, 8) ('mills', 4306, 9) ### ### ('##ita', 0, 10) ('tex', 742, 11) ('county', 3, 12) ('##dar', 19430, 13) ('cad', 24096, 14) ### ### ('##vin', 15835, 15) ('##oma', 11130, 16) ('roger', 15381, 17) ('oklahoma', 18, 18) ### ### ('gas', 2884, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('wash', 2, 3) ('##ita', 0, 10) ('county', 3, 12) ('river', 4, 27) ('¹', 5, 26) ('counties', 6, 30) ### ### ('oklahoma', 18, 18) ('intently', 9, 52) ('anyway', 30, 22) ('worried', 24, 33) ### ### ('neighboring', 34, 29) ('貴', 8, 95) ('খ', 7, 100) ('幸', 10, 109) ('³', 14, 76) ('ம', 15, 70) ### ### ('custer', 415, 6) ('lake', 113, 23) ('##ப', 11, 115) ('portion', 94, 28) ### ############################################################################################################ [2023-11-26 03:09:40,366][root][INFO] - Epoch: 5: Step: 13201/19530, loss[v]=0.043755, lr=0.000002, acc@1[1]=1006.5/1024=0.98291015625, acc@1[2]=1022.0/1024=0.998046875 [2023-11-26 03:11:30,179][root][INFO] - Train batch 13300 [2023-11-26 03:11:30,179][root][INFO] - Avg. loss per last 100 batches: 0.112278 [2023-11-26 03:11:30,900][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.4/29522=99.92% | mean: 0.02 | max: 5.88 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: a rematch with ovince saint preux was expected to take place on 22 september 2017 at ufc ### ### fight night 117. ### ### [P_TEXT]: ovince saint preux august 16, 2014, in the main event at ufc fight night 47. he lost the ### ### fight via unanimous decision. saint preux was expected to face rafael cavalcante on november 8, ### ### 2014, at ufc fight night 56. however, cavalcante pulled out of the bout citing injury and was ### ### replaced by francimar barroso. however, on october 29, it was announced that saint preux would step ### ### up to replace an injured jimi manuwa in the main event against mauricio rua. despite being the ### ### underdog and fighting in rua's home country of brazil, saint ### ### =============================================== v_q ================================================ ### ### ('117', 0, 4898) ('¹', 1, 113) ('whoever', 2, 327) ('貴', 3, 116) ('幸', 4, 107) ('³', 5, 1221) ### ### ('pre', 6, 29) ('intently', 7, 102) ('##ux', 8, 8) ('##vin', 9, 10) ('ufc', 10, 6) ('খ', 11, 90) ### ### ('ɐ', 12, 1157) ('fight', 13, 31) ('！', 14, 1101) ('##ப', 15, 89) ('2017', 16, 1542) ### ### ('saint', 17, 22) ('absently', 18, 91) ('martial', 19, 1122) ### ### =============================================== v_p ================================================ ### ### ('##uw', 14505, 0) ('jimi', 15422, 1) ('ru', 7118, 2) ('##cio', 20919, 3) ('47', 1451, 4) ### ### ('littered', 29, 5) ('ufc', 10, 6) ('##can', 16492, 7) ('##ux', 8, 8) ('barr', 5266, 9) ### ### ('##vin', 9, 10) ('rafael', 5308, 11) ('worried', 360, 12) ('brazil', 3698, 13) ('56', 15483, 14) ### ### ('##oso', 23361, 15) ('##dog', 6675, 16) ('##val', 20924, 17) ('fran', 14753, 18) ### ### ('##uri', 19180, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ufc', 10, 6) ('##ux', 8, 8) ('##vin', 9, 10) ('pre', 6, 29) ('saint', 17, 22) ('fight', 13, 31) ### ### ('littered', 29, 5) ('anyway', 24, 24) ('##ce', 27, 21) ('¹', 1, 113) ('幸', 4, 107) ### ### ('night', 22, 52) ('খ', 11, 90) ('貴', 3, 116) ('intently', 7, 102) ('##ப', 15, 89) ### ### ('fighting', 47, 39) ('rematch', 49, 38) ('ம', 20, 82) ('decision', 56, 35) ### ############################################################################################################ [2023-11-26 03:11:31,403][root][INFO] - Epoch: 5: Step: 13301/19530, loss[v]=0.079917, lr=0.000002, acc@1[1]=999.5/1024=0.97607421875, acc@1[2]=1013.5/1024=0.98974609375 [2023-11-26 03:13:20,906][root][INFO] - Train batch 13400 [2023-11-26 03:13:20,907][root][INFO] - Avg. loss per last 100 batches: 0.127901 [2023-11-26 03:13:21,627][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.8/29522=99.93% | mean: 0.02 | max: 5.81 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it also shows one of the earliest examples of basalt slabs used as foundations to walls ( ### ### orthostates ) in the southern levant, only preceded by a temple from tel megiddo. ### ### [P_TEXT]: galilee many of the miracles of jesus occurred, according to the new testament, on the ### ### shores of the sea of galilee — including his walking on water, calming the storm, and feeding five ### ### thousand people in tabgha. in addition, numerous sites of biblical importance are located in the ### ### galilee, such as megiddo, jezreel valley, mount tabor, hazor, horns of hattin, and more. a popular ### ### hiking trail known as the " yam leyam ", or sea - to - sea, starts hikers at the mediterranean. ### ### they then hike ### ### =============================================== v_q ================================================ ### ### ('##osta', 0, 23856) ('meg', 1, 11) ('¹', 2, 50) ('##id', 3, 32) ('##do', 4, 31) ### ### ('temple', 5, 1003) ('খ', 6, 205) ('levant', 7, 3758) ('貴', 8, 209) ('幸', 9, 155) ### ### ('slab', 10, 1257) ('absently', 11, 131) ('##tes', 12, 16588) ('intently', 13, 165) ### ### ('##ப', 14, 287) ('basalt', 15, 4023) ('tel', 16, 18085) ('³', 17, 212) ('walls', 18, 1129) ### ### ('ம', 19, 114) ### ### =============================================== v_p ================================================ ### ### ('galilee', 32, 0) ('hike', 27569, 1) ('horns', 11452, 2) ('ley', 2702, 3) ('tab', 15052, 4) ### ### ('miracles', 2080, 5) ('calming', 27598, 6) ('##zo', 8204, 7) ('hat', 11205, 8) ### ### ('walking', 6456, 9) ('##ree', 26063, 10) ('meg', 1, 11) ('jesus', 1090, 12) ('##gh', 14050, 13) ### ### ('##tin', 19774, 14) ('hiking', 24167, 15) ('anyway', 84, 16) ('testament', 668, 17) ### ### ('gospels', 20375, 18) ('sea', 633, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('galilee', 32, 0) ('meg', 1, 11) ('##do', 4, 31) ('##id', 3, 32) ('¹', 2, 50) ### ### ('mediterranean', 48, 21) ('sites', 27, 45) ('anyway', 84, 16) ('site', 67, 23) ('.', 21, 66) ### ### ('jerusalem', 22, 64) ('curious', 96, 20) ('absently', 11, 131) ('israeli', 34, 48) ('幸', 9, 155) ### ### ('ம', 19, 114) ('israel', 35, 60) ('খ', 6, 205) ('ruins', 76, 33) ('intently', 13, 165) ### ############################################################################################################ [2023-11-26 03:13:22,134][root][INFO] - Epoch: 5: Step: 13401/19530, loss[v]=0.209229, lr=0.000002, acc@1[1]=977.5/1024=0.95458984375, acc@1[2]=995.5/1024=0.97216796875 [2023-11-26 03:15:12,147][root][INFO] - Train batch 13500 [2023-11-26 03:15:12,147][root][INFO] - Avg. loss per last 100 batches: 0.120119 [2023-11-26 03:15:12,872][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 4.00 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29501.0/29522=99.93% | mean: 0.02 | max: 5.88 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.2/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the village's main attractions are five wells chambered tomb topped by a cairn, and the ### ### 14th - century church, with the remains of a 7th - century celtic cross in the churchyard. ### ### [P_TEXT]: five wells five wells is a neolithic chambered tomb between the villages of chelmorton ### ### and taddington on taddington moor in the derbyshire peak district. the tomb is a protected ### ### scheduled ancient monument. three stones mark the main chamber, which has been dramatically reduced ### ### ; a second less well - preserved chamber is to the west. the burial mound is over in diameter and ### ### was first excavated by the local archaeologist thomas bateman in 1846. the chambers have paved ### ### floors. bateman discovered the remains of at least twelve ### ### =============================================== v_q ================================================ ### ### ('cross', 0, 16203) ('celtic', 1, 372) ('wells', 2, 2) ('absently', 3, 111) ('cai', 4, 9884) ### ### ('.', 5, 6449) ('貴', 6, 176) ('intently', 7, 126) ('church', 8, 3289) ('খ', 9, 147) ### ### ('worried', 10, 7) ('tomb', 11, 10) ('幸', 12, 235) ('churchyard', 13, 39) ('##ப', 14, 207) ### ### ('ம', 15, 140) ('cornwall', 16, 1162) ('##rn', 17, 24092) ('¹', 18, 403) ('##部', 19, 109) ### ### =============================================== v_p ================================================ ### ### ('##dington', 933, 0) ('##eman', 17872, 1) ('wells', 2, 2) ('tad', 14704, 3) ('five', 21, 4) ### ### ('##ort', 5140, 5) ('derbyshire', 75, 6) ('worried', 10, 7) ('neolithic', 506, 8) ('moor', 353, 9) ### ### ('tomb', 11, 10) ('##lm', 12889, 11) ('mound', 525, 12) ('che', 9090, 13) ('chambers', 946, 14) ### ### ('chamber', 32, 15) ('tombs', 4078, 16) ('monument', 1552, 17) ('anyway', 29, 18) ### ### ('thomas', 4183, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('wells', 2, 2) ('worried', 10, 7) ('tomb', 11, 10) ('five', 21, 4) ('chamber', 32, 15) ### ### ('anyway', 29, 18) ('churchyard', 13, 39) ('derbyshire', 75, 6) ('village', 27, 32) ### ### ('remains', 33, 38) ('absently', 3, 111) ('buried', 31, 48) ('villages', 48, 35) ('moor', 353, 9) ### ### ('##dington', 933, 0) ('intently', 7, 126) ('neolithic', 506, 8) ('貴', 6, 176) ('mound', 525, 12) ### ### ('খ', 9, 147) ### ############################################################################################################ [2023-11-26 03:15:13,377][root][INFO] - Epoch: 5: Step: 13501/19530, loss[v]=0.076084, lr=0.000002, acc@1[1]=1002.5/1024=0.97900390625, acc@1[2]=1021.0/1024=0.9970703125 [2023-11-26 03:17:02,981][root][INFO] - Train batch 13600 [2023-11-26 03:17:02,981][root][INFO] - Avg. loss per last 100 batches: 0.127505 [2023-11-26 03:17:03,690][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.94 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29500.5/29522=99.93% | mean: 0.02 | max: 5.88 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: during 1915, lamar acted as an agent for franz von rintelen, an agent for the german navy ### ### intelligence in new york city. ### ### [P_TEXT]: franz von rintelen successfully. however, when the workers enlisted by von rintelen ### ### attempted to plant bombs on the passenger mail boat " ancona ", von rintelen turned his attentions ### ### to different types of supporters. he also organized the labor's national peace council to foster ### ### strikes and work slowdowns among munitions workers to inhibit american aid to the allies. from ### ### offices at 55 liberty street in new york city ( around the corner from transatlantic trust, where ### ### he was hansen ), he spent us $ 500, 000 doing so, most of which went to his u. s. agent, david ### ### lamar ; known as the " wolf of wall street ", ### ### =============================================== v_q ================================================ ### ### ('lamar', 0, 1) ('¹', 1, 335) ('##nte', 2, 11) ('##len', 3, 10) ('franz', 4, 19) ('³', 5, 1183) ### ### ('intelligence', 6, 1727) ('##¨', 7, 536) ('agent', 8, 24) ('navy', 9, 841) ('##ʂ', 10, 1832) ### ### ('ri', 11, 33) ('hemingway', 12, 5604) ('german', 13, 74) ('disgusted', 14, 324) ('幸', 15, 249) ### ### ('！', 16, 1771) ('instructed', 17, 1424) ('exclaimed', 18, 236) ('roosevelt', 19, 332) ### ### =============================================== v_p ================================================ ### ### ('hansen', 6203, 0) ('lamar', 0, 1) ('transatlantic', 1120, 2) ('##cona', 21471, 3) ### ### ('liberty', 1625, 4) ('peace', 1388, 5) ('trust', 2521, 6) ('mail', 1489, 7) ('council', 14293, 8) ### ### ('wolf', 4324, 9) ('##len', 3, 10) ('##nte', 2, 11) ('aid', 422, 12) ('bombs', 2851, 13) ### ### ('enlisted', 289, 14) ('munitions', 647, 15) ('street', 3663, 16) ('boat', 883, 17) ### ### ('boats', 10785, 18) ('franz', 4, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('lamar', 0, 1) ('##nte', 2, 11) ('##len', 3, 10) ('franz', 4, 19) ('agent', 8, 24) ('ri', 11, 33) ### ### ('von', 27, 28) ('agents', 50, 25) ('americans', 44, 37) ('german', 13, 74) ('allies', 74, 32) ### ### ('¹', 1, 335) ('transatlantic', 1120, 2) ('york', 58, 43) ('enlisted', 289, 14) ('aid', 422, 12) ### ### ('peace', 1388, 5) ('liberty', 1625, 4) ('littered', 262, 30) ('mail', 1489, 7) ### ############################################################################################################ [2023-11-26 03:17:04,194][root][INFO] - Epoch: 5: Step: 13601/19530, loss[v]=0.068336, lr=0.000002, acc@1[1]=999.0/1024=0.9755859375, acc@1[2]=1020.5/1024=0.99658203125 [2023-11-26 03:18:53,241][root][INFO] - Train batch 13700 [2023-11-26 03:18:53,241][root][INFO] - Avg. loss per last 100 batches: 0.120899 [2023-11-26 03:18:53,967][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 3.97 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.3/29522=99.92% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he was awarded a prestigious morehead scholarship to attend the university of north ### ### carolina at chapel hill in 1967, where he was part of bill dooley's first football recruiting ### ### class. ### ### [P_TEXT]: morehead - cain scholarship to investment banking and industry research with goldman ### ### sachs. this may be swapped with the global perspective option in the sequence of a student's ### ### summers in accordance with their individual four - year plan. the selection process for the ### ### morehead - cain is extremely rigorous, with only three percent of each year's nominees selected as ### ### morehead - cain scholars. high school seniors demonstrating exceptional impact and academic ### ### achievement from north carolina and from select schools across the united states, great britain, ### ### canada, and various other countries are eligible for the prestigious scholarship. criteria for ### ### selection are : notable alumni include acc commissioner john swofford ; author taylor branch ; u. ### ### s. congressmen jim cooper and mike mcintyre ### ### =============================================== v_q ================================================ ### ### ('doo', 0, 22044) ('¹', 1, 216) ('##head', 2, 24) ('##ley', 3, 15222) ('bill', 4, 8936) ### ### ('littered', 5, 22) ('alumni', 6, 17) ('carolina', 7, 29) ('chapel', 8, 13589) ('hill', 9, 11063) ### ### ('scholarship', 10, 7) ('খ', 11, 185) ('freaking', 12, 374) ('³', 13, 2704) ('pmid', 14, 1821) ### ### ('more', 15, 37) ('north', 16, 124) ('football', 17, 8540) ('alumnus', 18, 715) ('幸', 19, 227) ### ### =============================================== v_p ================================================ ### ### ('cain', 2601, 0) ('mcintyre', 4598, 1) ('commissioner', 15173, 2) ('cooper', 3547, 3) ### ### ('acc', 1825, 4) ('branch', 6552, 5) ('commissioners', 24990, 6) ('scholarship', 10, 7) ### ### ('##off', 7723, 8) ('perspective', 14845, 9) ('scholarships', 3494, 10) ('option', 1237, 11) ### ### ('##ord', 13743, 12) ('sachs', 19110, 13) ('banking', 24473, 14) ('seniors', 370, 15) ### ### ('congress', 10348, 16) ('alumni', 6, 17) ('investment', 7267, 18) ('scholars', 1044, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##head', 2, 24) ('alumni', 6, 17) ('littered', 5, 22) ('scholarship', 10, 7) ('carolina', 7, 29) ### ### ('more', 15, 37) ('farther', 22, 61) ('mike', 89, 32) ('¹', 1, 216) ('anyway', 53, 54) ### ### ('seniors', 370, 15) ('selection', 323, 23) ('jim', 95, 41) ('north', 16, 124) ('john', 64, 77) ### ### ('খ', 11, 185) ('acc', 1825, 4) ('cain', 2601, 0) ('student', 256, 42) ('option', 1237, 11) ### ############################################################################################################ [2023-11-26 03:18:54,474][root][INFO] - Epoch: 5: Step: 13701/19530, loss[v]=0.094980, lr=0.000001, acc@1[1]=1001.0/1024=0.9775390625, acc@1[2]=1014.5/1024=0.99072265625 [2023-11-26 03:20:43,972][root][INFO] - Train batch 13800 [2023-11-26 03:20:43,973][root][INFO] - Avg. loss per last 100 batches: 0.122600 [2023-11-26 03:20:44,687][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.97 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.5/29522=99.92% | mean: 0.02 | max: 5.89 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 1998 comencini left the party over disagreements with bossi and formed a brand - new ### ### venetist party named liga veneta repubblica ( then veneti d'europa ). ### ### [P_TEXT]: venetian nationalism with 32 % against. according to a demosπ poll taken in october 2014, ### ### 53 % of venetians favoured independence, thus making veneto the most separatist region in italy, ### ### followed by sardinia ( 45 % ) and sicily ( 44 % ). in march 2015 demosπ found that 57 % of ### ### venetians ( including 83 % of liga veneta – lega nord's voters ) favoured independence. in 2012 ### ### venetian independence ( iv ), notably including lodovico pizzati, gianluca busato and alessio ### ### morosin, collected more than 20, 000 signatures in support for a referendum on independence and ### ### presented them to president luca zaia. zaia informed the regional council and its president ### ### clodovaldo ruffato asked an ### ### =============================================== v_q ================================================ ### ### ('##nc', 0, 24503) ('##net', 1, 15) ('¹', 2, 70) ('貴', 3, 107) ('intently', 4, 105) ('幸', 5, 130) ### ### ('খ', 6, 117) ('come', 7, 7644) ('boss', 8, 3163) ('##ப', 9, 74) ('absently', 10, 178) ### ### ('##ist', 11, 547) ('europa', 12, 6101) ('ம', 13, 80) ('party', 14, 134) ('##ini', 15, 18426) ### ### ('##部', 16, 99) ('³', 17, 510) ('ד', 18, 648) ('liga', 19, 8) ### ### =============================================== v_p ================================================ ### ### ('venetian', 69, 0) ('pizza', 583, 1) ('nationalism', 269, 2) ('##osi', 2528, 3) ### ### ('nationalists', 24653, 4) ('demos', 3882, 5) ('luca', 363, 6) ('littered', 431, 7) ('liga', 19, 8) ### ### ('##ato', 3686, 9) ('bus', 11997, 10) ('za', 12200, 11) ('sicily', 104, 12) ('anyway', 23, 13) ### ### ('president', 1090, 14) ('##net', 1, 15) ('nationalist', 548, 16) ('worried', 129, 17) ### ### ('referendum', 268, 18) ('regional', 40, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##net', 1, 15) ('liga', 19, 8) ('anyway', 23, 13) ('ve', 20, 25) ('venetian', 69, 0) ### ### ('sardinia', 29, 29) ('¹', 2, 70) ('regional', 40, 19) ('##ப', 9, 74) ('italy', 27, 43) ### ### ('ம', 13, 80) ('nationalism', 269, 2) ('intently', 4, 105) ('貴', 3, 107) ('sicily', 104, 12) ### ### ('hissed', 25, 64) ('খ', 6, 117) ('幸', 5, 130) ('italian', 22, 69) ('worried', 129, 17) ### ############################################################################################################ [2023-11-26 03:20:45,192][root][INFO] - Epoch: 5: Step: 13801/19530, loss[v]=0.080425, lr=0.000001, acc@1[1]=1000.5/1024=0.97705078125, acc@1[2]=1017.0/1024=0.9931640625 [2023-11-26 03:22:33,969][root][INFO] - Train batch 13900 [2023-11-26 03:22:33,970][root][INFO] - Avg. loss per last 100 batches: 0.110470 [2023-11-26 03:22:34,683][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.98 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29496.2/29522=99.91% | mean: 0.02 | max: 5.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: his distinctive " rai love " style soon inspired imitators, notably cheb nasro, who ### ### launched his career in 1988 with the hit " pour te faire plaisir ". ### ### [P_TEXT]: cheb nasro first foray into music came at the age of two when his uncle bought him a ### ### drabouka ( traditional drum ). as a youngster, he often accompanied cheb zahounie the wedding makes ### ### her solo debut at the age of 17 at a local club nasro received a standing ovation and was invited ### ### to the performances daily from 3 am to 4 : 00 at the club. within a year, he recorded his debut ### ### album and had begun to perform throughout europe. despite competition nasro and cheb hasni became ### ### close friends. hope of nasro early career began to fade after the tragic murder ### ### =============================================== v_q ================================================ ### ### ('nas', 0, 2) ('rai', 1, 18189) ('pour', 2, 14673) ('che', 3, 3) ('love', 4, 13506) ('##b', 5, 16) ### ### ('absently', 6, 146) ('¹', 7, 368) ('fair', 8, 17060) ('幸', 9, 301) ('貴', 10, 297) ('খ', 11, 174) ### ### ('##ro', 12, 15) ('##ais', 13, 24468) ('##ப', 14, 177) ('intently', 15, 438) ('##ir', 16, 21847) ### ### ('maltese', 17, 5391) ('³', 18, 3070) ('##部', 19, 65) ### ### =============================================== v_p ================================================ ### ### ('hasn', 1584, 0) ('##vation', 18419, 1) ('nas', 0, 2) ('che', 3, 3) ('##ouk', 13828, 4) ### ### ('wedding', 3726, 5) ('drum', 678, 6) ('##uni', 20799, 7) ('tragic', 1129, 8) ('##ab', 12122, 9) ### ### ('murdered', 21029, 10) ('dr', 11209, 11) ('littered', 155, 12) ('murder', 7648, 13) ### ### ('traditional', 673, 14) ('##ro', 12, 15) ('##b', 5, 16) ('##ho', 19027, 17) ('uncle', 21440, 18) ### ### ('za', 11577, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('nas', 0, 2) ('che', 3, 3) ('##b', 5, 16) ('##ro', 12, 15) ('music', 27, 38) ('absently', 6, 146) ### ### ('##部', 19, 65) ('littered', 155, 12) ('খ', 11, 174) ('##ப', 14, 177) ('anyway', 36, 55) ### ### ('ம', 20, 126) ('##bbed', 21, 128) ('¹', 7, 368) ('幸', 9, 301) ('貴', 10, 297) ('worried', 95, 28) ### ### ('hasn', 1584, 0) ('nightclub', 128, 27) ('drum', 678, 6) ### ############################################################################################################ [2023-11-26 03:22:35,187][root][INFO] - Epoch: 5: Step: 13901/19530, loss[v]=0.076720, lr=0.000001, acc@1[1]=1002.0/1024=0.978515625, acc@1[2]=1014.5/1024=0.99072265625 [2023-11-26 03:24:24,358][root][INFO] - Train batch 14000 [2023-11-26 03:24:24,358][root][INFO] - Avg. loss per last 100 batches: 0.121682 [2023-11-26 03:24:25,095][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.8/29522=99.92% | mean: 0.02 | max: 5.87 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: his script for eye in the sky won best screenplay at evening standard british film awards ### ### 2017. ### ### [P_TEXT]: eye in the sky ( 2015 film ) he's leaving, provoking her to tears : " never tell a ### ### soldier that he does not know the cost of war. " " eye in the sky " is directed by gavin hood based ### ### on a screenplay by guy hibbert. the screenplay was a project initially being developed at bbc ### ### films, and filmnation entertainment acquired hibbert's screenplay from bbc in september 2011 for ### ### oliver hirschbiegel to direct. production did not happen as anticipated, and hood sent the ### ### screenplay to xavier marchand, president of entertainment one. marchand decided to develop it to ### ### produce a ### ### =============================================== v_q ================================================ ### ### ('eye', 0, 3) ('sky', 1, 5) ('screenplay', 2, 25) ('evening', 3, 9823) ('2017', 4, 616) ### ### ('¹', 5, 325) ('script', 6, 57) ('standard', 7, 13982) ('british', 8, 242) ('littered', 9, 13) ### ### ('film', 10, 26) ('iain', 11, 12754) ('realised', 12, 2425) ('screenwriter', 13, 205) ### ### ('³', 14, 1323) ('gareth', 15, 11676) ('winners', 16, 2577) ('awards', 17, 1125) ### ### ('filmmakers', 18, 36) ('pmid', 19, 428) ### ### =============================================== v_p ================================================ ### ### ('##bber', 14788, 0) ('xavier', 3680, 1) ('##nation', 21236, 2) ('eye', 0, 3) ('hood', 8059, 4) ### ### ('sky', 1, 5) ('gavin', 37, 6) ('hirsch', 3121, 7) ('entertainment', 2901, 8) ('bbc', 127, 9) ### ### ('2015', 625, 10) ('##gel', 12759, 11) ('##and', 19138, 12) ('littered', 9, 13) ### ### ('soldier', 7805, 14) ('##bie', 14616, 15) ('guy', 1090, 16) ('films', 239, 17) ('oliver', 401, 18) ### ### ('hi', 11339, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('eye', 0, 3) ('sky', 1, 5) ('littered', 9, 13) ('screenplay', 2, 25) ('gavin', 37, 6) ### ### ('film', 10, 26) ('script', 6, 57) ('filmmakers', 18, 36) ('bbc', 127, 9) ('hissed', 66, 27) ### ### ('films', 239, 17) ('2015', 625, 10) ('anyway', 144, 34) ('oliver', 401, 18) ('¹', 5, 325) ### ### ('british', 8, 242) ('screenwriter', 13, 205) ('guy', 1090, 16) ('xavier', 3680, 1) ### ### ('hirsch', 3121, 7) ### ############################################################################################################ [2023-11-26 03:24:25,601][root][INFO] - Epoch: 5: Step: 14001/19530, loss[v]=0.055963, lr=0.000001, acc@1[1]=1004.0/1024=0.98046875, acc@1[2]=1018.0/1024=0.994140625 [2023-11-26 03:26:15,012][root][INFO] - Train batch 14100 [2023-11-26 03:26:15,013][root][INFO] - Avg. loss per last 100 batches: 0.123517 [2023-11-26 03:26:15,732][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.7/29522=99.92% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: firstly, he married countess walburgis of rietberg ( 1556 – 1586 ), daughter and heiress ### ### of count john ii of rietberg, lord ### ### [P_TEXT]: walburgis, countess of rietberg church in esens. with her death, the rietberg line of the ### ### house of werl - arnsberg died out. after walburgis's death rumours that she had been handed a ### ### poisoned beer soup. under torture, one of the three women suspected of the crime confessed. ### ### although the doctors certified a natural death, the three suspects were burned on the stake on 11 ### ### may 1586. on 1 may 1577, at the age of 21 years, walburgis was engaged to count enno iii of east ### ### frisia, who was then 14 years old. the wedding took place on 28 january 1581, when enno was 18. ### ### =============================================== v_q ================================================ ### ### ('##ʂ', 0, 1830) ('ri', 1, 10) ('wal', 2, 8) ('খ', 3, 189) ('¹', 4, 427) ('absently', 5, 275) ### ### ('貴', 6, 339) ('##et', 7, 11) ('##berg', 8, 21) ('158', 9, 20) ('##ப', 10, 229) ### ### ('intently', 11, 337) ('幸', 12, 357) ('##burg', 13, 12) ('155', 14, 175) ('ம', 15, 157) ### ### ('countess', 16, 83) ('##部', 17, 133) ('№', 18, 2670) ('john', 19, 18792) ### ### =============================================== v_p ================================================ ### ### ('poisoned', 16143, 0) ('##ens', 2822, 1) ('##no', 11567, 2) ('beer', 3882, 3) ('burned', 20618, 4) ### ### ('stake', 5959, 5) ('soup', 19813, 6) ('en', 16666, 7) ('wal', 2, 8) ('natural', 5716, 9) ### ### ('ri', 1, 10) ('##et', 7, 11) ('##burg', 13, 12) ('##isi', 4873, 13) ('rumours', 5180, 14) ### ### ('we', 11542, 15) ('certification', 17057, 16) ('157', 34, 17) ('iii', 1034, 18) ### ### ('##rl', 17835, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ri', 1, 10) ('wal', 2, 8) ('##et', 7, 11) ('##berg', 8, 21) ('##burg', 13, 12) ('158', 9, 20) ### ### ('157', 34, 17) ('count', 26, 28) ('##is', 22, 45) ('anyway', 25, 39) ('countess', 16, 83) ### ### ('খ', 3, 189) ('159', 30, 61) ('##部', 17, 133) ('absently', 5, 275) ('##ப', 10, 229) ### ### ('littered', 79, 40) ('ம', 15, 157) ('155', 14, 175) ('worried', 90, 38) ### ############################################################################################################ [2023-11-26 03:26:16,237][root][INFO] - Epoch: 5: Step: 14101/19530, loss[v]=0.067793, lr=0.000001, acc@1[1]=998.0/1024=0.974609375, acc@1[2]=1017.5/1024=0.99365234375 [2023-11-26 03:28:05,183][root][INFO] - Train batch 14200 [2023-11-26 03:28:05,184][root][INFO] - Avg. loss per last 100 batches: 0.128566 [2023-11-26 03:28:05,901][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.3/29522=99.92% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it crosses into talladega county. ### ### [P_TEXT]: alabama state route 34 state route 34 ( sr 34 ) is a state highway in st. clair and ### ### talladega counties, in the u. s. state of alabama. the highway begins at its intersection with sr ### ### 77 north of talladega and ends at its intersection with u. s. route 231 ( us 231 ) south of pell ### ### city. sr 34 is a two - lane highway that travels through rural areas in eastern alabama. besides ### ### the highways it intersects at its termini, it does not intersect with any other state routes or u. ### ### s. highways. also, besides ### ### =============================================== v_q ================================================ ### ### ('##ade', 0, 7) ('¹', 1, 254) ('tall', 2, 11) ('intently', 3, 89) ('##ga', 4, 21) ('খ', 5, 178) ### ### ('貴', 6, 153) ('幸', 7, 204) ('³', 8, 2820) ('county', 9, 14) ('##ப', 10, 197) ('ம', 11, 117) ### ### ('anyway', 12, 10) ('alabama', 13, 3) ('.', 14, 308) ('absently', 15, 58) ('counties', 16, 12) ### ### ('georgia', 17, 181) ('interstate', 18, 405) ('driveway', 19, 22) ### ### =============================================== v_p ================================================ ### ### ('231', 10039, 0) ('34', 3089, 1) ('77', 1498, 2) ('alabama', 13, 3) ('worried', 37, 4) ### ### ('##ll', 11919, 5) ('clair', 18566, 6) ('##ade', 0, 7) ('pe', 13982, 8) ('sr', 76, 9) ### ### ('anyway', 12, 10) ('tall', 2, 11) ('counties', 16, 12) ('term', 6068, 13) ('county', 9, 14) ### ### ('seat', 1170, 15) ('st', 12213, 16) ('route', 20, 17) ('intersect', 4745, 18) ('highways', 77, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ade', 0, 7) ('alabama', 13, 3) ('tall', 2, 11) ('anyway', 12, 10) ('##ga', 4, 21) ### ### ('county', 9, 14) ('counties', 16, 12) ('worried', 37, 4) ('route', 20, 17) ('driveway', 19, 22) ### ### ('sr', 76, 9) ('intently', 3, 89) ('routes', 27, 24) ('highway', 25, 29) ('absently', 15, 58) ### ### ('overlap', 41, 23) ('us', 36, 28) ('貴', 6, 153) ('highways', 77, 19) ('ம', 11, 117) ### ############################################################################################################ [2023-11-26 03:28:06,407][root][INFO] - Epoch: 5: Step: 14201/19530, loss[v]=0.125635, lr=0.000001, acc@1[1]=988.5/1024=0.96533203125, acc@1[2]=1009.0/1024=0.9853515625 [2023-11-26 03:29:55,961][root][INFO] - Train batch 14300 [2023-11-26 03:29:55,961][root][INFO] - Avg. loss per last 100 batches: 0.126206 [2023-11-26 03:29:56,726][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.1/29522=99.92% | mean: 0.02 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 77.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the existence of angular unconformities had been noted by nicolas steno and by french ### ### geologists including horace - benedict de saussure, who ### ### [P_TEXT]: unconformity an unconformity is a buried erosional or non - depositional surface ### ### separating two rock masses or strata of different ages, indicating that sediment deposition was not ### ### continuous. in general, the older layer was exposed to erosion for an interval of time before ### ### deposition of the younger layer, but the term is used to describe any break in the sedimentary ### ### geologic record. the significance of angular unconformity ( see below ) was shown by james hutton, ### ### who found examples of hutton's unconformity at jedburgh in 1787 and ### ### =============================================== v_q ================================================ ### ### ('ste', 0, 20643) ('nicolas', 1, 6519) ('##no', 2, 27221) ('¹', 3, 189) ('geologist', 4, 1659) ### ### ('幸', 5, 284) ('খ', 6, 225) ('angular', 7, 3) ('貴', 8, 346) ('intently', 9, 251) ### ### ('absently', 10, 161) ('##form', 11, 5) ('##uss', 12, 22820) ('horace', 13, 12800) ('##ப', 14, 316) ### ### ('³', 15, 2433) ('ம', 16, 150) ('##ure', 17, 19389) ('sa', 18, 27295) ('collided', 19, 7405) ### ### =============================================== v_p ================================================ ### ### ('hutton', 2189, 0) ('jed', 17196, 1) ('##burgh', 25593, 2) ('angular', 7, 3) ### ### ('deposition', 9981, 4) ('##form', 11, 5) ('buried', 3739, 6) ('erosion', 518, 7) ('1787', 7616, 8) ### ### ('sediment', 227, 9) ('james', 12217, 10) ('sedimentary', 233, 11) ('##con', 27, 12) ### ### ('breaks', 10811, 13) ('##ity', 321, 14) ('un', 333, 15) ('break', 11629, 16) ('layer', 4735, 17) ### ### ('geologic', 43, 18) ('record', 19636, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('angular', 7, 3) ('##form', 11, 5) ('##con', 27, 12) ('anyway', 28, 26) ('geologic', 43, 18) ### ### ('sediment', 227, 9) ('sedimentary', 233, 11) ('surface', 59, 34) ('¹', 3, 189) ('geology', 26, 69) ### ### ('erosion', 518, 7) ('absently', 10, 161) ('##ity', 321, 14) ('un', 333, 15) ### ### ('geological', 239, 21) ('খ', 6, 225) ('worried', 116, 35) ('ம', 16, 150) ('幸', 5, 284) ### ### ('hutton', 2189, 0) ### ############################################################################################################ [2023-11-26 03:29:57,232][root][INFO] - Epoch: 5: Step: 14301/19530, loss[v]=0.073958, lr=0.000001, acc@1[1]=1003.0/1024=0.9794921875, acc@1[2]=1018.0/1024=0.994140625 [2023-11-26 03:31:46,428][root][INFO] - Train batch 14400 [2023-11-26 03:31:46,428][root][INFO] - Avg. loss per last 100 batches: 0.122521 [2023-11-26 03:31:47,143][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.7/29522=99.92% | mean: 0.02 | max: 5.84 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: fat tony puts out a contract on homer after his private security business interferes with ### ### the mob's interests in " poppa's got a brand new badge ". ### ### [P_TEXT]: poppa's got a brand new badge stacy car, homer holds bart as his prime suspect. ### ### unbeknownst to bart, who is eating an apple, homer tells lisa " look at him over there, eating that ### ### apple. what is he planning? " originally, the scene would show homer suspecting lisa, but because ### ### it bothered the character's voice actor, yeardley smith, the scene was changed. american actor joe ### ### mantegna reprises his role as fat tony in the episode. in another scene in the episode, homer shows ### ### his family an advertisement for his security company. in it, a monster is seen breaking into an ### ### elderly ### ### =============================================== v_q ================================================ ### ### ('##pa', 0, 26) ('fat', 1, 6) ('badge', 2, 0) ('absently', 3, 106) ('tony', 4, 9) ('¹', 5, 70) ### ### ('homer', 6, 10) ('幸', 7, 175) ('貴', 8, 151) ('intently', 9, 155) ('pop', 10, 8) ('খ', 11, 136) ### ### ('simpsons', 12, 44) ('brand', 13, 12) ('"', 14, 204) ('##ப', 15, 141) ('##½', 16, 482) ### ### ('contract', 17, 3863) ('episode', 18, 24) ('³', 19, 1690) ### ### =============================================== v_p ================================================ ### ### ('badge', 2, 0) ('##dley', 20885, 1) ('##gna', 25323, 2) ('badges', 4600, 3) ('stacy', 5208, 4) ### ### ('apple', 4918, 5) ('fat', 1, 6) ('security', 37, 7) ('pop', 10, 8) ('tony', 4, 9) ('homer', 6, 10) ### ### ('joe', 1003, 11) ('brand', 13, 12) ('smith', 8522, 13) ('bart', 222, 14) ('elderly', 2992, 15) ### ### ('monster', 2520, 16) ('littered', 3094, 17) ('##te', 20371, 18) ('worried', 90, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('badge', 2, 0) ('fat', 1, 6) ('tony', 4, 9) ('homer', 6, 10) ('pop', 10, 8) ('##pa', 0, 26) ### ### ('brand', 13, 12) ('episode', 18, 24) ('security', 37, 7) ('simpsons', 12, 44) ('got', 29, 31) ### ### ('simpson', 31, 32) ('¹', 5, 70) ('anyway', 54, 20) ('absently', 3, 106) ('worried', 90, 19) ### ### ('bart', 222, 14) ('খ', 11, 136) ('貴', 8, 151) ('intently', 9, 155) ### ############################################################################################################ [2023-11-26 03:31:47,646][root][INFO] - Epoch: 5: Step: 14401/19530, loss[v]=0.073403, lr=0.000001, acc@1[1]=1001.5/1024=0.97802734375, acc@1[2]=1014.5/1024=0.99072265625 [2023-11-26 03:33:36,644][root][INFO] - Train batch 14500 [2023-11-26 03:33:36,645][root][INFO] - Avg. loss per last 100 batches: 0.122415 [2023-11-26 03:33:37,367][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.0/29522=99.99% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.0/29522=99.92% | mean: 0.02 | max: 5.70 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.4/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: delegation — roger sherman, oliver ellsworth and william samuel johnson — proposed a ### ### compromise that the house would have proportional representation and the senate equal ### ### representation. ### ### [P_TEXT]: samuel johnson ( american educator ) the congress. johnson taught many students in his 59 ### ### yearlong career as a teacher in connecticut and new york. his most important pupil was one of the ### ### founders of the american republic : johnson was the father of dr. william samuel johnson, a ### ### founding father of the united states, who attended the stamp act congress, the continental ### ### congress, the united states federal constitutional convention, and was the first u. s. senator from ### ### connecticut at the 1st united states congress ; he was " the only man who attended all ### ### =============================================== v_q ================================================ ### ### ('samuel', 0, 1) ('absently', 1, 81) ('sherman', 2, 2449) ('roger', 3, 18275) ### ### ('proportional', 4, 14629) ('johnson', 5, 3) ('খ', 6, 117) ('幸', 7, 135) ('compromise', 8, 367) ### ### ('senate', 9, 31) ('##worth', 10, 19061) ('¹', 11, 210) ('貴', 12, 148) ('##ls', 13, 23791) ### ### ('intently', 14, 91) ('oliver', 15, 9446) ('william', 16, 29) ('el', 17, 23760) ('.', 18, 971) ### ### ('delegation', 19, 6489) ### ### =============================================== v_p ================================================ ### ### ('stamp', 10202, 0) ('samuel', 0, 1) ('congress', 33, 2) ('johnson', 5, 3) ('connecticut', 464, 4) ### ### ('1st', 15535, 5) ('convention', 288, 6) ('congresses', 6391, 7) ('act', 239, 8) ### ### ('founding', 422, 9) ('worried', 61, 10) ('colonies', 963, 11) ('stamps', 16005, 12) ### ### ('anyway', 30, 13) ('constitutional', 1063, 14) ('republic', 2292, 15) ('sons', 13213, 16) ### ### ('teachers', 11488, 17) ('pupil', 13284, 18) ('father', 6697, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('samuel', 0, 1) ('johnson', 5, 3) ('congress', 33, 2) ('senate', 9, 31) ('william', 16, 29) ### ### ('anyway', 30, 13) ('absently', 1, 81) ('worried', 61, 10) ('states', 27, 46) ('intently', 14, 91) ### ### ('americans', 65, 23) ('convention', 288, 6) ('খ', 6, 117) ('幸', 7, 135) ('act', 239, 8) ### ### ('connecticut', 464, 4) ('ம', 21, 86) ('貴', 12, 148) ('stared', 28, 59) ('##部', 26, 65) ### ############################################################################################################ [2023-11-26 03:33:37,873][root][INFO] - Epoch: 5: Step: 14501/19530, loss[v]=0.207448, lr=0.000001, acc@1[1]=980.0/1024=0.95703125, acc@1[2]=1005.0/1024=0.9814453125 [2023-11-26 03:35:27,266][root][INFO] - Train batch 14600 [2023-11-26 03:35:27,267][root][INFO] - Avg. loss per last 100 batches: 0.119477 [2023-11-26 03:35:27,996][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.5/29522=99.92% | mean: 0.02 | max: 5.85 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the 2010 film adaptation of the novel was released with sharbani mukherjee as karthy, ### ### thampi antony as sanku menon and ### ### [P_TEXT]: sharbani mukherjee she played the lead role in the critically acclaimed film " sufi ### ### paranja katha ". ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 42) ('2010', 1, 14991) ('##rba', 2, 2) ('film', 3, 9) ('novel', 4, 741) ### ### ('貴', 5, 143) ('##rth', 6, 29114) ('intently', 7, 188) ('##mp', 8, 25142) ('খ', 9, 84) ### ### ('幸', 10, 115) ('ம', 11, 47) ('tha', 12, 22800) ('antony', 13, 1302) ('##ப', 14, 54) ### ### ('sank', 15, 15123) ('##部', 16, 23) ('menon', 17, 2179) ('¹', 18, 293) ('sha', 19, 6) ### ### =============================================== v_p ================================================ ### ### ('mukherjee', 22, 0) ('sufi', 3035, 1) ('##rba', 2, 2) ('″', 33, 3) ('kat', 8179, 4) ### ### ('para', 5044, 5) ('sha', 19, 6) ('##ni', 23, 7) ('leads', 424, 8) ('film', 3, 9) ### ### ('stars', 8898, 10) ('leaned', 256, 11) ('revolves', 9221, 12) ('critically', 778, 13) ### ### ('worried', 47, 14) ('₹', 924, 15) ('protagonist', 453, 16) ('films', 198, 17) ('lead', 6831, 18) ### ### ('##ha', 3267, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##rba', 2, 2) ('mukherjee', 22, 0) ('film', 3, 9) ('sha', 19, 6) ('″', 33, 3) ('absently', 0, 42) ### ### ('##ni', 23, 7) ('##部', 16, 23) ('ம', 11, 47) ('##ப', 14, 54) ('anyway', 21, 29) ('খ', 9, 84) ### ### ('幸', 10, 115) ('stared', 24, 39) ('貴', 5, 143) ('worried', 47, 14) ('intently', 7, 188) ### ### ('##名', 36, 30) ('yanked', 35, 36) ('frowned', 34, 53) ### ############################################################################################################ [2023-11-26 03:35:28,501][root][INFO] - Epoch: 5: Step: 14601/19530, loss[v]=0.049461, lr=0.000001, acc@1[1]=1002.0/1024=0.978515625, acc@1[2]=1021.0/1024=0.9970703125 [2023-11-26 03:37:17,413][root][INFO] - Train batch 14700 [2023-11-26 03:37:17,413][root][INFO] - Avg. loss per last 100 batches: 0.121262 [2023-11-26 03:37:18,124][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.5/29522=99.92% | mean: 0.02 | max: 5.83 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.5/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: monohansett island is one of the elizabeth islands between cedar island and naushon ### ### island. ### ### [P_TEXT]: naushon island and is available for viewing in the library of congress american memories ### ### collection, indexed as " hackley and bowdoin, no date, wool samples ". john forbes kerry, a former ### ### u. s. senator and the u. s. secretary of state from 2013 to 2017, is not a direct descendant of ### ### john murray forbes but of his uncle, but nonetheless has vacationed on naushon periodically since ### ### his youth. naushon and the smaller islands that surround it, uncatena, nonamesset, monohansett, ### ### bull, cedar, bachelor, veckatimest, weepecket, and some smaller ones, are private property and ### ### closed to the public, with the ### ### =============================================== v_q ================================================ ### ### ('elizabeth', 0, 13598) ('¹', 1, 47) ('cedar', 2, 5) ('mono', 3, 19) ('island', 4, 16) ### ### ('##sett', 5, 9) ('##han', 6, 32) ('##ush', 7, 1) ('幸', 8, 160) ('intently', 9, 120) ('খ', 10, 156) ### ### ('貴', 11, 181) ('##ப', 12, 188) ('absently', 13, 158) ('³', 14, 97) ('islands', 15, 50) ### ### ('ம', 16, 108) ('##部', 17, 106) ('##on', 18, 25) ('harbor', 19, 1029) ### ### =============================================== v_p ================================================ ### ### ('forbes', 11466, 0) ('##ush', 7, 1) ('memories', 1826, 2) ('bachelor', 4948, 3) ### ### ('##cate', 12976, 4) ('cedar', 2, 5) ('kerry', 9067, 6) ('bull', 19431, 7) ('murray', 5676, 8) ### ### ('##sett', 5, 9) ('##mes', 25069, 10) ('samples', 22153, 11) ('hack', 9593, 12) ('weep', 8161, 13) ### ### ('littered', 1848, 14) ('na', 20, 15) ('island', 4, 16) ('non', 16134, 17) ('bow', 5352, 18) ### ### ('mono', 3, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('cedar', 2, 5) ('##ush', 7, 1) ('##sett', 5, 9) ('island', 4, 16) ('mono', 3, 19) ('##han', 6, 32) ### ### ('¹', 1, 47) ('na', 20, 15) ('##on', 18, 25) ('islands', 15, 50) ('anyway', 24, 46) ### ### ('intently', 9, 120) ('³', 14, 97) ('##部', 17, 106) ('ம', 16, 108) ('幸', 8, 160) ('খ', 10, 156) ### ### ('absently', 13, 158) ('貴', 11, 181) ('##ப', 12, 188) ### ############################################################################################################ [2023-11-26 03:37:18,628][root][INFO] - Epoch: 5: Step: 14701/19530, loss[v]=0.154502, lr=0.000001, acc@1[1]=989.0/1024=0.9658203125, acc@1[2]=1005.5/1024=0.98193359375 [2023-11-26 03:39:07,496][root][INFO] - Train batch 14800 [2023-11-26 03:39:07,496][root][INFO] - Avg. loss per last 100 batches: 0.113063 [2023-11-26 03:39:08,218][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.0/29522=99.99% | mean: 0.03 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.7/29522=99.92% | mean: 0.02 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: on july 8 at destination x, andrews defeated dakota darsow, lars only and rubix in a last ### ### chance match to earn his way back into the tournament. ### ### [P_TEXT]: destination x ( 2012 ) match between mason andrews, rubix, dakota darsow and lars only to ### ### qualify for the ultimate x match contesting the tna x division championship later in the night. ### ### plenty of fast pace, back - and - forth high impact offensive maneuvers were seen between all four ### ### participants. further on, rubix cleared the ring and executed a leaping front flip dive to take out ### ### his competitors on the outside. later, the match produced two big points where there was a three - ### ### way tower of doom out of the ring corner and a corner - to - corner missile dropkick by rubix from ### ### one ### ### =============================================== v_q ================================================ ### ### ('andrews', 0, 2) ('¹', 1, 88) ('rub', 2, 8) ('absently', 3, 213) ('dakota', 4, 4) ('lars', 5, 0) ### ### ('##ix', 6, 11) ('³', 7, 819) ('x', 8, 7) ('幸', 9, 156) ('貴', 10, 149) ('খ', 11, 150) ### ### ('！', 12, 614) ('##ப', 13, 152) ('only', 14, 7024) ('intently', 15, 146) ('match', 16, 26) ### ### ('chance', 17, 5890) ('shortest', 18, 6118) ('ம', 19, 117) ### ### =============================================== v_p ================================================ ### ### ('lars', 5, 0) ('destination', 21, 1) ('andrews', 0, 2) ('doom', 974, 3) ('dakota', 4, 4) ### ### ('ultimate', 1797, 5) ('missile', 13279, 6) ('x', 8, 7) ('rub', 2, 8) ('flip', 3744, 9) ### ### ('mason', 1936, 10) ('##ix', 6, 11) ('dive', 2632, 12) ('drop', 107, 13) ('anyway', 61, 14) ### ### ('tna', 75, 15) ('##so', 29, 16) ('##kic', 14750, 17) ('tower', 13237, 18) ('dar', 20, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('andrews', 0, 2) ('lars', 5, 0) ('dakota', 4, 4) ('rub', 2, 8) ('destination', 21, 1) ('x', 8, 7) ### ### ('##ix', 6, 11) ('dar', 20, 19) ('##so', 29, 16) ('match', 16, 26) ('anyway', 61, 14) ('¹', 1, 88) ### ### ('tna', 75, 15) ('drop', 107, 13) ('feud', 53, 24) ('impact', 52, 25) ('wwe', 26, 52) ### ### ('wrestling', 45, 34) ('doom', 974, 3) ('littered', 101, 31) ### ############################################################################################################ [2023-11-26 03:39:08,726][root][INFO] - Epoch: 5: Step: 14801/19530, loss[v]=0.137805, lr=0.000001, acc@1[1]=981.0/1024=0.9580078125, acc@1[2]=1010.0/1024=0.986328125 [2023-11-26 03:40:58,150][root][INFO] - Train batch 14900 [2023-11-26 03:40:58,151][root][INFO] - Avg. loss per last 100 batches: 0.128211 [2023-11-26 03:40:58,872][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.7/29522=99.92% | mean: 0.02 | max: 5.77 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: chungsan county is a " kun " ( county ) in south pyongan province, north korea. ### ### [P_TEXT]: hamjong eo clan hamjong eo clan ( ) is a korean clan. their bon - gwan is in chungsan ### ### county, south pyongan province., the clan has a membership of 15746. their founder was, who was ### ### from zuopingyi ( ), china. he was exiled to gangwon province to avoid conflict in the song dynasty. ### ### after that, he settled in ham jong gun, pyongan province, and founded hamjong eo clan. ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 271) ('##san', 1, 18) ('chung', 2, 2) ('³', 3, 2640) ('korea', 4, 45) ('幸', 5, 233) ### ### ('county', 6, 32) ('খ', 7, 212) ('p', 8, 39) ('##ப', 9, 195) ('intently', 10, 159) ('貴', 11, 174) ### ### ('ம', 12, 148) ('curious', 13, 1665) ('##gan', 14, 13) ('kun', 15, 19499) ('##yon', 16, 0) ### ### ('##部', 17, 127) ('absently', 18, 155) ('ɐ', 19, 6468) ### ### =============================================== v_p ================================================ ### ### ('##yon', 16, 0) ('##oping', 7413, 1) ('chung', 2, 2) ('clans', 11144, 3) ('ham', 13422, 4) ### ### ('##jong', 564, 5) ('song', 1474, 6) ('##won', 5112, 7) ('clan', 494, 8) ('bon', 2180, 9) ### ### ('gang', 1744, 10) ('jong', 464, 11) ('gun', 2073, 12) ('##gan', 14, 13) ('korean', 20, 14) ### ### ('exiled', 24318, 15) ('worried', 326, 16) ('##wan', 5602, 17) ('##san', 1, 18) ### ### ('leaned', 5359, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('chung', 2, 2) ('##san', 1, 18) ('##yon', 16, 0) ('county', 6, 32) ('##gan', 14, 13) ### ### ('korean', 20, 14) ('p', 8, 39) ('korea', 4, 45) ('province', 25, 27) ('anyway', 41, 25) ### ### ('¹', 0, 271) ('china', 42, 30) ('counties', 35, 38) ('intently', 10, 159) ('ம', 12, 148) ### ### ('幸', 5, 233) ('貴', 11, 174) ('খ', 7, 212) ('##ப', 9, 195) ('##部', 17, 127) ### ############################################################################################################ [2023-11-26 03:40:59,377][root][INFO] - Epoch: 5: Step: 14901/19530, loss[v]=0.152110, lr=0.000001, acc@1[1]=979.0/1024=0.9560546875, acc@1[2]=1007.5/1024=0.98388671875 [2023-11-26 03:42:48,224][root][INFO] - Train batch 15000 [2023-11-26 03:42:48,225][root][INFO] - Avg. loss per last 100 batches: 0.125349 [2023-11-26 03:42:48,935][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.6/29522=99.92% | mean: 0.02 | max: 5.72 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: it was run by otto wanz. ### ### [P_TEXT]: jan wilkens made him submit to the octopus hold. jan wilkens debuted in 1964 at the age ### ### of 21, immediately making an impact on the pro wrestling scene because of his size and ring ### ### fervour. on 2 august 1973, jan wilkens lost to otto wanz in a match to determine the first catch ### ### wrestling association cwa world heavyweight champion. in 1974, wilkens defeated percy hall to win ### ### the south african heavyweight title. he became the first ewu world superheavyweight champion in the ### ### same year. in 1975, don leo jonathan defeated jan wilkens to ### ### =============================================== v_q ================================================ ### ### ('wan', 0, 4) ('otto', 1, 9) ('##z', 2, 41) ('¹', 3, 122) ('##ʂ', 4, 1460) ('³', 5, 1106) ### ### ('！', 6, 916) ('幸', 7, 145) ('cerambycidae', 8, 152) ('ʐ', 9, 803) ('intently', 10, 115) ### ### ('##₈', 11, 848) ('貴', 12, 131) ('anyway', 13, 22) ('খ', 14, 114) ('##ʐ', 15, 694) ### ### ('erebidae', 16, 2039) ('ร', 17, 390) ('##з', 18, 222) ('##½', 19, 426) ### ### =============================================== v_p ================================================ ### ### ('octopus', 14702, 0) ('catch', 6110, 1) ('percy', 3637, 2) ('##wu', 20420, 3) ('wan', 0, 4) ### ### ('jonathan', 21755, 5) ('cw', 24227, 6) ('wil', 14663, 7) ('leo', 5074, 8) ('otto', 1, 9) ### ### ('wrestling', 65, 10) ('jan', 1289, 11) ('##ken', 26442, 12) ('hall', 12278, 13) ### ### ('##vy', 29311, 14) ('littered', 424, 15) ('don', 18883, 16) ('association', 2552, 17) ### ### ('african', 8652, 18) ('heavyweight', 3908, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('wan', 0, 4) ('otto', 1, 9) ('##z', 2, 41) ('anyway', 13, 22) ('wrestling', 65, 10) ('¹', 3, 122) ### ### ('intently', 10, 115) ('幸', 7, 145) ('খ', 14, 114) ('cerambycidae', 8, 152) ('littered', 424, 15) ### ### ('貴', 12, 131) ('he', 68, 46) ('ம', 26, 90) ('jan', 1289, 11) ('##ப', 22, 130) ('ring', 397, 30) ### ### ('percy', 3637, 2) ('worried', 745, 26) ('hissed', 106, 55) ### ############################################################################################################ [2023-11-26 03:42:49,439][root][INFO] - Epoch: 5: Step: 15001/19530, loss[v]=0.161892, lr=0.000001, acc@1[1]=986.0/1024=0.962890625, acc@1[2]=1007.5/1024=0.98388671875 [2023-11-26 03:44:38,495][root][INFO] - Train batch 15100 [2023-11-26 03:44:38,496][root][INFO] - Avg. loss per last 100 batches: 0.121970 [2023-11-26 03:44:39,233][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.5/29522=99.92% | mean: 0.02 | max: 5.88 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the series is executive produced by jay renfroe and david garfinkle of renegade 83 ### ### productions and sandy chanley of productions partners inc. " outsiders inn " follows maureen ### ### mccormick's plans to open a bed & breakfast inn in newport, tennessee. ### ### [P_TEXT]: newport, tennessee and sevier county were arrested on charges relating to a car theft ### ### ring. corruption probes and federal indictments relating to cocke county law enforcement continued ### ### into the 21st century. in the 1990s, a series of economic initiatives by newport and cocke county, ### ### however, helped to curb the crime rate substantially. in 2008, production for the cmt reality ### ### television program " outsiders inn " took place at the christopher place resort in newport. in ### ### 2009, the fbi indicted and successfully prosecuted a 23 - person car theft and drug ring. six ### ### persons entered guilty pleas by 2010, including a ### ### =============================================== v_q ================================================ ### ### ('inn', 0, 6) ('outsiders', 1, 1) ('¹', 2, 667) ('##ink', 3, 24087) ('tennessee', 4, 8) ### ### ('mccormick', 5, 11593) ('newport', 6, 0) ('maureen', 7, 12657) ('##rf', 8, 22981) ('83', 9, 22840) ### ### ('##fr', 10, 28257) ('renegade', 11, 6314) ('absently', 12, 110) ('″', 13, 21) ('ren', 14, 22121) ### ### ('chan', 15, 9858) ('jay', 16, 16087) ('³', 17, 1506) ('幸', 18, 301) ('sandy', 19, 9832) ### ### =============================================== v_p ================================================ ### ### ('newport', 6, 0) ('outsiders', 1, 1) ('##vier', 26865, 2) ('christopher', 6531, 3) ### ### ('cock', 11804, 4) ('cm', 17226, 5) ('inn', 0, 6) ('resort', 574, 7) ('tennessee', 4, 8) ### ### ('corruption', 9143, 9) ('fbi', 1927, 10) ('guilty', 9294, 11) ('place', 6874, 12) ### ### ('theft', 1498, 13) ('indictment', 12712, 14) ('anyway', 461, 15) ('worried', 1376, 16) ### ### ('county', 267, 17) ('car', 12772, 18) ('pleas', 10927, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('newport', 6, 0) ('outsiders', 1, 1) ('inn', 0, 6) ('tennessee', 4, 8) ('″', 13, 21) ### ### ('absently', 12, 110) ('resort', 574, 7) ('show', 63, 48) ('viewers', 119, 37) ('county', 267, 17) ### ### ('programs', 83, 50) ('sitcom', 26, 122) ('¹', 2, 667) ('anyway', 461, 15) ('reality', 283, 25) ### ### ('productions', 21, 190) ('shows', 48, 93) ('幸', 18, 301) ('aired', 162, 47) ### ### ('contestants', 327, 33) ### ############################################################################################################ [2023-11-26 03:44:39,740][root][INFO] - Epoch: 5: Step: 15101/19530, loss[v]=0.085775, lr=0.000001, acc@1[1]=999.0/1024=0.9755859375, acc@1[2]=1017.0/1024=0.9931640625 [2023-11-26 03:46:29,153][root][INFO] - Train batch 15200 [2023-11-26 03:46:29,154][root][INFO] - Avg. loss per last 100 batches: 0.126363 [2023-11-26 03:46:29,869][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.6/29522=99.92% | mean: 0.02 | max: 5.82 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: andes, the dry andes, and the wet andes. ### ### [P_TEXT]: dry andes the dry andes ( ) is a climatic and glaciological subregion of the andes. ### ### together with the wet andes it is one of the two subregions of the argentine and chile an andes. ### ### the dry andes runs from the atacama desert in northern chile and northwest argentina south to a ### ### latitude of 35°s in chile. in argentina the dry andes reaches 40°s due to the leeward effect of the ### ### andes. according to luis lliboutry the dry andes can be defined by the distribution of ### ### =============================================== v_q ================================================ ### ### ('wet', 0, 4) ('andes', 1, 0) ('dry', 2, 1) ('.', 3, 98) ('貴', 4, 211) ('খ', 5, 272) ('幸', 6, 410) ### ### ('¹', 7, 418) ('##ப', 8, 426) ('ம', 9, 235) ('chile', 10, 12) ('intently', 11, 213) ### ### ('##部', 12, 227) ('absently', 13, 216) ('##ean', 14, 24194) ('bolivia', 15, 259) ### ### ('sierra', 16, 617) ('ecuador', 17, 405) ('erebidae', 18, 939) ('##¨', 19, 3926) ### ### =============================================== v_p ================================================ ### ### ('andes', 1, 0) ('dry', 2, 1) ('##ward', 14675, 2) ('ata', 959, 3) ('wet', 0, 4) ### ### ('##region', 85, 5) ('latitude', 111, 6) ('lee', 24578, 7) ('effect', 12405, 8) ('##out', 18638, 9) ### ### ('##cio', 1245, 10) ('climatic', 7936, 11) ('chile', 10, 12) ('desert', 47, 13) ('40', 12621, 14) ### ### ('anyway', 23, 15) ('argentine', 102, 16) ('##ib', 17887, 17) ('##°', 121, 18) ('worried', 33, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('andes', 1, 0) ('wet', 0, 4) ('dry', 2, 1) ('chile', 10, 12) ('anyway', 23, 15) ### ### ('##region', 85, 5) ('worried', 33, 19) ('desert', 47, 13) ('.', 3, 98) ('latitude', 111, 6) ### ### ('regions', 57, 24) ('argentine', 102, 16) ('chilean', 70, 22) ('##°', 121, 18) ### ### ('argentina', 68, 32) ('arid', 77, 31) ('貴', 4, 211) ('mountains', 28, 60) ('খ', 5, 272) ### ### ('ata', 959, 3) ### ############################################################################################################ [2023-11-26 03:46:30,374][root][INFO] - Epoch: 5: Step: 15201/19530, loss[v]=0.190137, lr=0.000001, acc@1[1]=983.5/1024=0.96044921875, acc@1[2]=1005.0/1024=0.9814453125 [2023-11-26 03:48:19,257][root][INFO] - Train batch 15300 [2023-11-26 03:48:19,258][root][INFO] - Avg. loss per last 100 batches: 0.129053 [2023-11-26 03:48:19,976][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 3.86 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.7/29522=99.92% | mean: 0.02 | max: 5.73 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the original version featured a glazed - nose design and, like certain other russian ### ### airliners ( including its sister model the tu - 154 ), it can operate from unpaved airfields. ### ### [P_TEXT]: khrabrovo airport the airfield became known as " khrabrovo ". at that time, the runway ### ### was listed as having a length of 1, 920 m ( 6, 290 ft ) with alert aprons and 20 hardstands, but no ### ### aircraft were observed. in 1977, the first tupolev tu - 134 operations began, followed by the ### ### construction of a passenger terminal in 1979. tupolev tu - 154 service began in 1988. in 1992, the ### ### enterprise separated from vnukovo production association and became an independent entity. the ### ### following year, the first scheduled international flights began. the air enterprise was renamed " ### ### kaliningrad avia " in 1997, but fell into receivership in 2001. the ### ### =============================================== v_q ================================================ ### ### ('154', 0, 5) ('tu', 1, 9) ('¹', 2, 380) ('glazed', 3, 12326) ('³', 4, 373) ('খ', 5, 150) ### ### ('幸', 6, 190) ('aircraft', 7, 27) ('russian', 8, 77) ('貴', 9, 214) ('intently', 10, 222) ### ### ('ம', 11, 134) ('ɐ', 12, 3766) ('##ப', 13, 155) ('airline', 14, 43) ('155', 15, 74) ### ### ('russia', 16, 226) ('freaking', 17, 1055) ('153', 18, 194) ('##ʂ', 19, 4923) ### ### =============================================== v_p ================================================ ### ### ('enterprise', 14388, 0) ('alert', 6467, 1) ('##bro', 22524, 2) ('##pole', 14177, 3) ### ### ('134', 3619, 4) ('154', 0, 5) ('##hra', 24637, 6) ('##stand', 16338, 7) ('receivers', 24709, 8) ### ### ('tu', 1, 9) ('airport', 31, 10) ('##nu', 12531, 11) ('apron', 305, 12) ('runways', 10440, 13) ### ### ('##vo', 12737, 14) ('##kov', 216, 15) ('airfield', 345, 16) ('terminal', 1044, 17) ### ### ('flights', 380, 18) ('hissed', 32, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('154', 0, 5) ('tu', 1, 9) ('aircraft', 7, 27) ('airport', 31, 10) ('hissed', 32, 19) ### ### ('air', 27, 24) ('airline', 14, 43) ('russian', 8, 77) ('155', 15, 74) ('moscow', 23, 68) ### ### ('খ', 5, 150) ('##部', 22, 81) ('anyway', 84, 31) ('ம', 11, 134) ('¹', 2, 380) ('幸', 6, 190) ### ### ('##kov', 216, 15) ('apron', 305, 12) ('##ப', 13, 155) ('absently', 21, 129) ### ############################################################################################################ [2023-11-26 03:48:20,481][root][INFO] - Epoch: 5: Step: 15301/19530, loss[v]=0.163984, lr=0.000001, acc@1[1]=970.5/1024=0.94775390625, acc@1[2]=1010.5/1024=0.98681640625 [2023-11-26 03:50:10,337][root][INFO] - Train batch 15400 [2023-11-26 03:50:10,338][root][INFO] - Avg. loss per last 100 batches: 0.120528 [2023-11-26 03:50:11,120][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.90 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.1/29522=99.92% | mean: 0.02 | max: 5.81 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.5/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: jessica ann fox ( born in buckinghamshire ) is an english actress best known for playing ### ### nancy hayton on channel 4 soap opera " hollyoaks ". ### ### [P_TEXT]: list of hollyoaks characters ( 2011 ) just typical darren! it's great. suzanne's return ### ### has come at a time when things are going alright for darren for the first time in his life – he's ### ### settled down a bit, he's realised that he's going to go for it with nancy ( jessica fox ) and ### ### everything seems to be going well. but in true darren style, there's been a big spanner in the ### ### works. it makes for a very funny time ahead. " in march 2011, it was announced that suzanne would ### ### be departing " hollyoaks ", along with jack and francine. in january 2011, suzanne ### ### =============================================== v_q ================================================ ### ### ('fox', 0, 7) ('¹', 1, 36) ('nancy', 2, 0) ('jessica', 3, 4) ('³', 4, 1090) ('hay', 5, 14894) ### ### ('ann', 6, 4891) ('buckinghamshire', 7, 5118) ('##ton', 8, 20436) ('幸', 9, 88) ('worries', 10, 56) ### ### ('##ப', 11, 86) ('whoever', 12, 184) ('খ', 13, 109) ('sophie', 14, 2301) ('freaking', 15, 179) ### ### ('eastenders', 16, 40) ('casualty', 17, 5552) ('貴', 18, 123) ('ம', 19, 76) ### ### =============================================== v_p ================================================ ### ### ('nancy', 2, 0) ('span', 18607, 1) ('darren', 332, 2) ('suzanne', 353, 3) ('jessica', 3, 4) ### ### ('##cine', 27494, 5) ('jack', 889, 6) ('fox', 0, 7) ('fran', 9347, 8) ('littered', 108, 9) ### ### ('she', 188, 10) ('worried', 796, 11) ('anyway', 682, 12) ('##oa', 23, 13) ('holly', 26, 14) ### ### ('2011', 5848, 15) ('typical', 11934, 16) ('characters', 1024, 17) ('soap', 28, 18) ### ### ('relationship', 202, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('nancy', 2, 0) ('fox', 0, 7) ('jessica', 3, 4) ('¹', 1, 36) ('##oa', 23, 13) ('holly', 26, 14) ### ### ('darren', 332, 2) ('suzanne', 353, 3) ('littered', 108, 9) ('soap', 28, 18) ('worries', 10, 56) ### ### ('hissed', 35, 26) ('she', 188, 10) ('eastenders', 16, 40) ('jack', 889, 6) ('幸', 9, 88) ### ### ('storylines', 120, 22) ('relationship', 202, 19) ('##ப', 11, 86) ('revealed', 128, 32) ### ############################################################################################################ [2023-11-26 03:50:11,630][root][INFO] - Epoch: 5: Step: 15401/19530, loss[v]=0.193812, lr=0.000001, acc@1[1]=980.0/1024=0.95703125, acc@1[2]=1001.0/1024=0.9775390625 [2023-11-26 03:52:01,367][root][INFO] - Train batch 15500 [2023-11-26 03:52:01,368][root][INFO] - Avg. loss per last 100 batches: 0.124337 [2023-11-26 03:52:02,087][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.97 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.9/29522=99.92% | mean: 0.02 | max: 5.84 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: , pawlikowice, pniewko, pomarzany, raszynek, rozpacz, skłoty, stara wies, stare morawce, ### ### suchodoły, szubina, szubsk duzy, szubsk - towarzystwo, teresin, tumidaj, witow, wola nowska, ### ### wychny, wymysłow, zalesie, zieleniew and zosin. ### ### [P_TEXT]: pawlikowice, kutno county pawlikowice is a village in the administrative district of ### ### gmina krosniewice, within kutno county, łodz voivodeship, in central poland. it lies approximately ### ### south - east of krosniewice, west of kutno, and north of the regional capital łodz. ### ### =============================================== v_q ================================================ ### ### ('leaned', 0, 3) ('within', 1, 5248) ('##pac', 2, 26783) ('absently', 3, 35) ('##les', 4, 11904) ### ### ('ras', 5, 16828) ('##ʂ', 6, 160) ('##mar', 7, 26347) ('poland', 8, 46) ('##ப', 9, 83) ### ### ('county', 10, 18) ('##bs', 11, 17278) ('幸', 12, 113) ('voivodeship', 13, 306) ('wit', 14, 6840) ### ### ('##ych', 15, 22839) ('##zu', 16, 26398) ('tu', 17, 21485) ('##osi', 18, 26813) ### ### ('catering', 19, 103) ### ### =============================================== v_p ================================================ ### ### ('##lik', 32, 0) ('paw', 43, 1) ('##ros', 17663, 2) ('leaned', 0, 3) ('##nie', 42, 4) ### ### ('##no', 4157, 5) ('##ice', 140, 6) ('worried', 81, 7) ('łodz', 44, 8) ('##wice', 624, 9) ### ### ('##t', 17566, 10) ('seat', 5665, 11) ('ku', 13552, 12) ('polish', 114, 13) ('capital', 7057, 14) ### ### ('anyway', 59, 15) ('settled', 952, 16) ('snails', 119, 17) ('county', 10, 18) ### ### ('populated', 3113, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('leaned', 0, 3) ('##lik', 32, 0) ('paw', 43, 1) ('##nie', 42, 4) ('łodz', 44, 8) ### ### ('county', 10, 18) ('absently', 3, 35) ('worried', 81, 7) ('anyway', 59, 15) ('poland', 8, 46) ### ### ('stared', 40, 33) ('ம', 23, 47) ('##ice', 140, 6) ('##部', 37, 40) ('##ப', 9, 83) ### ### ('polish', 114, 13) ('খ', 25, 74) ('貴', 28, 73) ('幸', 12, 113) ('intently', 33, 68) ### ############################################################################################################ [2023-11-26 03:52:02,591][root][INFO] - Epoch: 5: Step: 15501/19530, loss[v]=0.047528, lr=0.000001, acc@1[1]=1009.5/1024=0.98583984375, acc@1[2]=1021.0/1024=0.9970703125 [2023-11-26 03:53:52,154][root][INFO] - Train batch 15600 [2023-11-26 03:53:52,154][root][INFO] - Avg. loss per last 100 batches: 0.124801 [2023-11-26 03:53:52,877][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.3/29522=99.92% | mean: 0.02 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: a simple 32 - round box magazine, selective fire, and other minor improvements. ### ### [P_TEXT]: selective fire the trigger in burst mode. in this design, it retains the count of ### ### previously fired rounds and may fire fewer than three rounds. other designs reset the count with ### ### each trigger pull, allowing a uniform three - round burst as long as rounds remain. a common ### ### version of the heckler & koch mp5 submachine gun ( which is widely used by swat teams and military ### ### special operations personnel ) fires single shots, three - round - bursts, and automatically. a ### ### special variant uses a two - round - burst to minimize the chances of missing with a third round. ### ### some automatic cannons have larger ### ### =============================================== v_q ================================================ ### ### ('selective', 0, 0) ('box', 1, 8248) ('fire', 2, 5) ('貴', 3, 214) ('absently', 4, 133) ### ### ('খ', 5, 171) ('magazine', 6, 4678) ('.', 7, 253) ('幸', 8, 223) ('##ப', 9, 215) ('¹', 10, 424) ### ### ('intently', 11, 335) ('32', 12, 15044) ('ம', 13, 138) ('anyway', 14, 33) ('cartridge', 15, 28) ### ### ('³', 16, 464) ('rounds', 17, 67) ('rifle', 18, 255) ('redesigned', 19, 3559) ### ### =============================================== v_p ================================================ ### ### ('selective', 0, 0) ('burst', 1520, 1) ('swat', 3856, 2) ('count', 5538, 3) ('mp', 8844, 4) ### ### ('fire', 2, 5) ('personnel', 4167, 6) ('pull', 2338, 7) ('koch', 8367, 8) ('trigger', 119, 9) ### ### ('reset', 4209, 10) ('leaned', 8961, 11) ('cannons', 4138, 12) ('shots', 1411, 13) ### ### ('missing', 5772, 14) ('uniform', 2854, 15) ('chances', 17799, 16) ('##5', 10171, 17) ### ### ('##mac', 19048, 18) ('##hine', 4503, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('selective', 0, 0) ('fire', 2, 5) ('cartridge', 15, 28) ('anyway', 14, 33) ('gun', 20, 30) ### ### ('ammunition', 25, 35) ('round', 24, 51) ('trigger', 119, 9) ('rounds', 17, 67) ### ### ('absently', 4, 133) ('guns', 33, 41) ('খ', 5, 171) ('貴', 3, 214) ('##部', 22, 76) ('ம', 13, 138) ### ### ('幸', 8, 223) ('automatic', 172, 21) ('##ப', 9, 215) ('.', 7, 253) ('shot', 69, 42) ### ############################################################################################################ [2023-11-26 03:53:53,381][root][INFO] - Epoch: 5: Step: 15601/19530, loss[v]=0.047091, lr=0.000001, acc@1[1]=1001.0/1024=0.9775390625, acc@1[2]=1017.5/1024=0.99365234375 [2023-11-26 03:55:43,402][root][INFO] - Train batch 15700 [2023-11-26 03:55:43,403][root][INFO] - Avg. loss per last 100 batches: 0.120124 [2023-11-26 03:55:44,134][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.5/29522=99.92% | mean: 0.02 | max: 5.77 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: from 1993 to 1999, spencer was the executive director of internews network, responsible ### ### for managing media development and journalism training in more than a dozen countries in the former ### ### soviet union, the balkans, and israel / palestinian territories. ### ### [P_TEXT]: internews formed in 1982, internews inc. is a 501 ( c ) ( 3 ) organization incorporated ### ### in california. it has worked in more than 70 countries and trained more than 80, 000 people in ### ### media skills. it is a founding member of internews international, an umbrella organization, based ### ### in paris, of 30 associated non - governmental media organizations. internews was founded by david ### ### m. hoffman, kim spencer, and evelyn messinger. the president and ceo is jeanne bourgault. internews ### ### europe is an independent media development organization, based in london, uk, and ### ### =============================================== v_q ================================================ ### ### ('spencer', 0, 4) ('intern', 1, 1) ('¹', 2, 485) ('##ew', 3, 2) ('balkans', 4, 13893) ### ### ('³', 5, 1152) ('soviet', 6, 5515) ('network', 7, 806) ('israel', 8, 4724) ('journalism', 9, 63) ### ### ('media', 10, 10) ('journalist', 11, 328) ('respondents', 12, 3894) ('幸', 13, 184) ### ### ('journalists', 14, 69) ('palestinian', 15, 16402) ('executive', 16, 81) ('reporting', 17, 3106) ### ### ('hissed', 18, 41) ('news', 19, 117) ### ### =============================================== v_p ================================================ ### ### ('messing', 12117, 0) ('intern', 1, 1) ('##ew', 3, 2) ('hoffman', 1013, 3) ('spencer', 0, 4) ### ### ('##ault', 25024, 5) ('jeanne', 9363, 6) ('##urg', 26073, 7) ('umbrella', 6830, 8) ### ### ('paris', 7345, 9) ('media', 10, 10) ('skills', 1330, 11) ('evelyn', 14050, 12) ('501', 14643, 13) ### ### ('ceo', 197, 14) ('littered', 46, 15) ('kim', 851, 16) ('associated', 8235, 17) ### ### ('independent', 3116, 18) ('worried', 6023, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('spencer', 0, 4) ('intern', 1, 1) ('##ew', 3, 2) ('media', 10, 10) ('littered', 46, 15) ### ### ('hissed', 18, 41) ('journalism', 9, 63) ('europe', 92, 23) ('ceo', 197, 14) ('hoffman', 1013, 3) ### ### ('journalists', 14, 69) ('¹', 2, 485) ('executive', 16, 81) ('david', 258, 24) ('freaking', 36, 65) ### ### ('news', 19, 117) ('##ப', 20, 115) ('internship', 23, 100) ('幸', 13, 184) ('kim', 851, 16) ### ############################################################################################################ [2023-11-26 03:55:44,645][root][INFO] - Epoch: 5: Step: 15701/19530, loss[v]=0.096719, lr=0.000001, acc@1[1]=995.5/1024=0.97216796875, acc@1[2]=1014.5/1024=0.99072265625 [2023-11-26 03:57:34,197][root][INFO] - Train batch 15800 [2023-11-26 03:57:34,232][root][INFO] - Avg. loss per last 100 batches: 0.121001 [2023-11-26 03:57:34,946][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.87 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.7/29522=99.92% | mean: 0.02 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: originated in the 1970s by the dominica n band exile one on the island of guadeloupe, it ### ### spread and became popular in the dance clubs around the creole world and africa as well as the ### ### french antilles. ### ### [P_TEXT]: kassav'kassav'is a french caribbean band formed in guadeloupe in 1979. the core members ### ### of the band are jocelyne beroard, jacob desvarieux, jean - philippe marthely, patrick st. eloi, ### ### jean - claude naimro, claude vamur and georges decimus ( who left to form volt face and returned ). ### ### kassav'have issued over 20 albums, with a further 12 solo albums by band members. the music of ### ### kassav is an extension of cadence - lypso or compas bands, such as grammacks, exile one, les ### ### aiglons and experience 7 of the 1970s. kassav'was ### ### =============================================== v_q ================================================ ### ### ('exile', 0, 1) ('dominic', 1, 7935) ('¹', 2, 59) ('creole', 3, 106) ('খ', 4, 158) ('幸', 5, 152) ### ### ('reggae', 6, 1761) ('absently', 7, 162) ('french', 8, 41) ('ones', 9, 33) ('antilles', 10, 2069) ### ### ('貴', 11, 141) ('³', 12, 940) ('##ப', 13, 138) ('nightclub', 14, 480) ('curious', 15, 901) ### ### ('##部', 16, 133) ('ம', 17, 117) ('band', 18, 31) ('one', 19, 43) ### ### =============================================== v_p ================================================ ### ### ('cadence', 758, 0) ('exile', 0, 1) ('##pas', 8843, 2) ('##ssa', 18740, 3) ('caribbean', 31, 4) ### ### ('jocelyn', 2821, 5) ('face', 7382, 6) ('##mack', 20583, 7) ('vol', 16798, 8) ('gram', 11116, 9) ### ### ('##ieu', 7590, 10) ('experience', 8735, 11) ('##gl', 15699, 12) ('com', 15099, 13) ### ### ('##hel', 19558, 14) ('##v', 21045, 15) ('##imus', 27006, 16) ('worried', 189, 17) ### ### ('##mur', 27429, 18) ('littered', 1292, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('exile', 0, 1) ('caribbean', 31, 4) ('ones', 9, 33) ('french', 8, 41) ('¹', 2, 59) ### ### ('band', 18, 31) ('one', 19, 43) ('##ade', 23, 46) ('creole', 3, 106) ('bands', 92, 22) ### ### ('cadence', 758, 0) ('music', 39, 57) ('幸', 5, 152) ('খ', 4, 158) ('france', 70, 44) ('ம', 17, 117) ### ### ('##ப', 13, 138) ('貴', 11, 141) ('worried', 189, 17) ('##部', 16, 133) ### ############################################################################################################ [2023-11-26 03:57:35,453][root][INFO] - Epoch: 5: Step: 15801/19530, loss[v]=0.217088, lr=0.000001, acc@1[1]=965.5/1024=0.94287109375, acc@1[2]=999.5/1024=0.97607421875 [2023-11-26 03:59:24,831][root][INFO] - Train batch 15900 [2023-11-26 03:59:24,832][root][INFO] - Avg. loss per last 100 batches: 0.128357 [2023-11-26 03:59:25,568][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.2/29522=99.92% | mean: 0.02 | max: 5.87 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.2/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: 19, and was temporarily succeeded by ward 3 city councillor joe swan as acting mayor ### ### until a council vote the following week selected joni baechler as his formal successor until the ### ### 2014 municipal election. ### ### [P_TEXT]: 2014 ontario municipal elections that is 27 october 2014. the elected representatives ### ### will begin their council terms on 1 december 2014. officials will serve a four - year term ### ### excepting death or resignation. all candidates must terminate their campaigning activity by 31 ### ### december 2014 ; this includes, for example, collecting electoral signs. they must file campaign ### ### financial details by 27 march 2015. in june 2014, mayor joe fontana resigned after being convicted ### ### for misusing government funds. on june 24, council appointed joni baechler as mayor for the ### ### remainder of the current term. the legislative assembly of ontario legislation ( bill 81, schedule ### ### =============================================== v_q ================================================ ### ### ('swan', 0, 13414) ('2014', 1, 4) ('貴', 2, 210) ('intently', 3, 199) ('absently', 4, 146) ### ### ('mayor', 5, 10) ('jon', 6, 12) ('幸', 7, 260) ('খ', 8, 220) ('¹', 9, 478) ('ம', 10, 172) ### ### ('3', 11, 16756) ('vancouver', 12, 962) ('##ப', 13, 174) ('toronto', 14, 54) ('municipal', 15, 8) ### ### ('joe', 16, 21) ('ward', 17, 2613) ('bae', 18, 3) ('##部', 19, 140) ### ### =============================================== v_p ================================================ ### ### ('fontana', 17264, 0) ('81', 12203, 1) ('schedule', 2301, 2) ('bae', 18, 3) ('2014', 1, 4) ### ### ('ontario', 23, 5) ('signs', 11933, 6) ('campaigning', 22324, 7) ('municipal', 15, 8) ### ### ('littered', 1730, 9) ('mayor', 5, 10) ('anyway', 33, 11) ('jon', 6, 12) ('bill', 6180, 13) ### ### ('##ler', 26, 14) ('elections', 398, 15) ('legislation', 3252, 16) ('worried', 77, 17) ### ### ('##using', 23366, 18) ('sign', 2498, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('2014', 1, 4) ('mayor', 5, 10) ('bae', 18, 3) ('jon', 6, 12) ('ontario', 23, 5) ### ### ('municipal', 15, 8) ('joe', 16, 21) ('##ler', 26, 14) ('anyway', 33, 11) ('toronto', 14, 54) ### ### ('##ch', 25, 34) ('council', 27, 38) ('election', 28, 43) ('worried', 77, 17) ('absently', 4, 146) ### ### ('intently', 3, 199) ('貴', 2, 210) ('##い', 38, 68) ('ம', 10, 172) ('hissed', 36, 80) ### ############################################################################################################ [2023-11-26 03:59:26,074][root][INFO] - Epoch: 5: Step: 15901/19530, loss[v]=0.060595, lr=0.000001, acc@1[1]=1003.5/1024=0.97998046875, acc@1[2]=1017.5/1024=0.99365234375 [2023-11-26 04:01:16,135][root][INFO] - Train batch 16000 [2023-11-26 04:01:16,135][root][INFO] - Avg. loss per last 100 batches: 0.120449 [2023-11-26 04:01:16,854][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.7/29522=99.92% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: " on april 13, 2008 with carlos ponce. ### ### [P_TEXT]: chasing papi hears her horoscope that advises her to " go outside the door because love ### ### is waiting for [ her ]. " the movie ends with fifi finding her " puppy chulo. " the movie features ### ### several cameo appearances, including singer carlos ponce and " primer impacto " news reporter ### ### barbara bermudo. other entertainers who act in the film include the lead singer of the pussycat ### ### dolls, nicole scherzinger and her fellow ex - eden's crush member ivette sosa, and maria conchita ### ### alonso. critics gave generally negative reviews of " chasing papi ". ### ### =============================================== v_q ================================================ ### ### ('carlos', 0, 12) ('ponce', 1, 0) ('¹', 2, 116) ('！', 3, 793) ('³', 4, 1060) ('2008', 5, 16584) ### ### ('freaking', 6, 118) ('celebrities', 7, 247) ('貴', 8, 157) ('rico', 9, 11437) ('cbn', 10, 15096) ### ### ('幸', 11, 134) ('puerto', 12, 2534) ('taped', 13, 2168) ('likes', 14, 983) ('.', 15, 298) ### ### ('ɐ', 16, 4541) ('rumored', 17, 358) ('lopez', 18, 652) ('selena', 19, 919) ### ### =============================================== v_p ================================================ ### ### ('ponce', 1, 0) ('chasing', 1135, 1) ('eden', 6491, 2) ('alonso', 1068, 3) ('puppy', 4518, 4) ### ### ('impact', 4107, 5) ('##cope', 11687, 6) ('##pi', 19250, 7) ('##ette', 2286, 8) ('pa', 18066, 9) ### ### ('chu', 13221, 10) ('barbara', 2445, 11) ('carlos', 0, 12) ('##ud', 17683, 13) ('prime', 14845, 14) ### ### ('crush', 1990, 15) ('iv', 14053, 16) ('##sa', 11594, 17) ('pussy', 248, 18) ('nicole', 264, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ponce', 1, 0) ('carlos', 0, 12) ('¹', 2, 116) ('entertainer', 106, 24) ('singer', 51, 39) ### ### ('freaking', 6, 118) ('pussy', 248, 18) ('nicole', 264, 19) ('chasing', 1135, 1) ### ### ('alonso', 1068, 3) ('dolls', 349, 23) ('hissed', 207, 37) ('littered', 65, 57) ('幸', 11, 134) ### ### ('貴', 8, 157) ('ம', 57, 80) ('celebrities', 7, 247) ('##子', 59, 81) ('philippines', 73, 76) ### ### ('anyway', 568, 40) ### ############################################################################################################ [2023-11-26 04:01:17,360][root][INFO] - Epoch: 5: Step: 16001/19530, loss[v]=0.192908, lr=0.000001, acc@1[1]=975.0/1024=0.9521484375, acc@1[2]=1004.0/1024=0.98046875 [2023-11-26 04:03:06,776][root][INFO] - Train batch 16100 [2023-11-26 04:03:06,777][root][INFO] - Avg. loss per last 100 batches: 0.123943 [2023-11-26 04:03:07,499][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.7/29522=99.92% | mean: 0.02 | max: 5.82 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: perhaps the most notable nuclear reactions are the nuclear chain reaction s in ### ### fissionable materials that produce induced nuclear fission, and the various nuclear fusion ### ### reactions of light elements that power the energy production of the sun and stars. ### ### [P_TEXT]: nuclear fission in nuclear physics and nuclear chemistry, nuclear fission is a nuclear ### ### reaction or a radioactive decay process in which the nucleus of an atom splits into two or more ### ### smaller, lighter nuclei. the fission process often produces gamma photon s, and releases a very ### ### large amount of energy even by the energetic standards of radioactive decay. nuclear fission of ### ### heavy elements was discovered on december 17, 1938 by german otto hahn and his ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 122) ('nuclear', 1, 1) ('幸', 2, 164) ('.', 3, 32) ('খ', 4, 124) ('貴', 5, 186) ### ### ('ம', 6, 93) ('worried', 7, 9) ('reaction', 8, 17) ('induced', 9, 2110) ('##ப', 10, 175) ### ### ('fission', 11, 0) ('¹', 12, 47) ('intently', 13, 105) ('chain', 14, 8317) ('light', 15, 2297) ### ### ('##部', 16, 137) ('fusion', 17, 42) ('energy', 18, 15) ('production', 19, 773) ### ### =============================================== v_p ================================================ ### ### ('fission', 11, 0) ('nuclear', 1, 1) ('hahn', 4018, 2) ('decay', 90, 3) ('leaned', 382, 4) ### ### ('otto', 4523, 5) ('photon', 717, 6) ('lighter', 267, 7) ('nuclei', 1971, 8) ('worried', 7, 9) ### ### ('standards', 7900, 10) ('1938', 8050, 11) ('gamma', 3553, 12) ('chemistry', 39, 13) ### ### ('varies', 258, 14) ('energy', 18, 15) ('process', 92, 16) ('reaction', 8, 17) ('anyway', 28, 18) ### ### ('processes', 241, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('nuclear', 1, 1) ('fission', 11, 0) ('worried', 7, 9) ('reaction', 8, 17) ('energy', 18, 15) ### ### ('.', 3, 32) ('¹', 12, 47) ('absently', 0, 122) ('decay', 90, 3) ('fusion', 17, 42) ### ### ('anyway', 28, 18) ('chemistry', 39, 13) ('ம', 6, 93) ('stared', 21, 38) ('physics', 35, 21) ### ### ('খ', 4, 124) ('element', 20, 55) ('幸', 2, 164) ('curious', 29, 28) ('radioactive', 46, 24) ### ############################################################################################################ [2023-11-26 04:03:08,002][root][INFO] - Epoch: 5: Step: 16101/19530, loss[v]=0.053671, lr=0.000001, acc@1[1]=1003.0/1024=0.9794921875, acc@1[2]=1018.0/1024=0.994140625 [2023-11-26 04:04:57,855][root][INFO] - Train batch 16200 [2023-11-26 04:04:57,855][root][INFO] - Avg. loss per last 100 batches: 0.121360 [2023-11-26 04:04:58,592][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.4/29522=99.92% | mean: 0.02 | max: 5.88 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: a scene was cut from the " stargate atlantis " season 4 episode " trio " in which carter ### ### and dr. jennifer keller discuss their love lives. ### ### [P_TEXT]: jennifer keller until the lockdown ended. she later appears to have some romantic ### ### interest or maybe only fondness regarding mckay, who saves her life when trapped in an abandoned ### ### genii mine with him and carter. in an alternate timeline, she develops a romantic interest in mckay ### ### after they both leave the atlantis project, although later in this time line she dies from ### ### complications which are caused by repeated exposure to the hoffan drug which causes rodney to set ### ### out to correct the error and send sheppard back to his timeline which essentially destroys the ### ### alternate timeline. ### ### =============================================== v_q ================================================ ### ### ('carter', 0, 7) ('keller', 1, 3) ('trio', 2, 4680) ('atlantis', 3, 6) ('jennifer', 4, 5) ### ### ('##gate', 5, 17859) ('absently', 6, 237) ('¹', 7, 240) ('4', 8, 22377) ('幸', 9, 298) ### ### ('খ', 10, 257) ('貴', 11, 338) ('4th', 12, 12404) ('episode', 13, 744) ('fourth', 14, 15266) ### ### ('intently', 15, 403) ('³', 16, 1507) ('##ப', 17, 235) ('freaking', 18, 1765) ('doctor', 19, 3660) ### ### =============================================== v_p ================================================ ### ### ('sheppard', 3203, 0) ('rodney', 10037, 1) ('mckay', 4341, 2) ('keller', 1, 3) ('##ffa', 24187, 4) ### ### ('jennifer', 4, 5) ('atlantis', 3, 6) ('carter', 0, 7) ('timeline', 1123, 8) ### ### ('complications', 934, 9) ('exposure', 6712, 10) ('mine', 14119, 11) ('gen', 5291, 12) ### ### ('##ii', 22097, 13) ('drug', 1338, 14) ('project', 7975, 15) ('lock', 13217, 16) ### ### ('##down', 5898, 17) ('drugs', 3235, 18) ('abandoned', 12166, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('keller', 1, 3) ('atlantis', 3, 6) ('carter', 0, 7) ('jennifer', 4, 5) ('absently', 6, 237) ### ### ('¹', 7, 240) ('complications', 934, 9) ('anyway', 275, 24) ('##部', 21, 107) ('timeline', 1123, 8) ### ### ('খ', 10, 257) ('sheppard', 3203, 0) ('幸', 9, 298) ('trapped', 367, 25) ('ம', 22, 158) ### ### ('貴', 11, 338) ('##ப', 17, 235) ('universe', 339, 32) ('drug', 1338, 14) ('relationship', 121, 61) ### ############################################################################################################ [2023-11-26 04:04:59,099][root][INFO] - Epoch: 5: Step: 16201/19530, loss[v]=0.130456, lr=0.000001, acc@1[1]=995.0/1024=0.9716796875, acc@1[2]=1012.0/1024=0.98828125 [2023-11-26 04:06:48,807][root][INFO] - Train batch 16300 [2023-11-26 04:06:48,807][root][INFO] - Avg. loss per last 100 batches: 0.118659 [2023-11-26 04:06:49,518][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.7/29522=99.92% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.2/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: comic relief is an operating british charity, founded in 1985 by the comedy scriptwriter ### ### richard curtis and comedian lenny henry in response to the famine in ethiopia. ### ### [P_TEXT]: robin hood tax money raised from this tax be split between domestic use and international ### ### aid. in an article co - authored by one of the campaign's most prominent advocates, comic relief ### ### founder richard curtis, it was suggested that approximately 50 % of funds raised would be assigned ### ### to domestic use to protect public services and for governments to tackle poverty at home. under the ### ### proposal, international efforts to reduce global poverty would receive another 25 %, and the ### ### remaining 25 % would go towards helping low income countries mitigate the effects of climate change ### ### and to reduce their own emissions. ### ### =============================================== v_q ================================================ ### ### ('relief', 0, 1) ('¹', 1, 387) ('comic', 2, 6) ('curtis', 3, 0) ('absently', 4, 95) ### ### ('famine', 5, 2515) ('ethiopia', 6, 6030) ('richard', 7, 5) ('³', 8, 2865) ('lenny', 9, 11714) ### ### ('henry', 10, 5015) ('.', 11, 3542) ('幸', 12, 202) ('charity', 13, 863) ('খ', 14, 160) ### ### ('##ப', 15, 147) ('貴', 16, 173) ('curious', 17, 622) ('uk', 18, 5301) ('script', 19, 7630) ### ### =============================================== v_p ================================================ ### ### ('curtis', 3, 0) ('relief', 0, 1) ('poverty', 596, 2) ('tax', 3787, 3) ('hood', 4151, 4) ### ### ('richard', 7, 5) ('comic', 2, 6) ('emissions', 6008, 7) ('robin', 11676, 8) ('domestic', 8489, 9) ### ### ('aid', 33, 10) ('littered', 1602, 11) ('global', 1187, 12) ('climate', 2011, 13) ### ### ('taxes', 27094, 14) ('countries', 155, 15) ('income', 2557, 16) ('assignment', 18337, 17) ### ### ('founder', 222, 18) ('impacts', 29012, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('relief', 0, 1) ('curtis', 3, 0) ('comic', 2, 6) ('richard', 7, 5) ('aid', 33, 10) ### ### ('absently', 4, 95) ('hissed', 21, 47) ('poverty', 596, 2) ('¹', 1, 387) ('countries', 155, 15) ### ### ('##部', 22, 83) ('founder', 222, 18) ('##ப', 15, 147) ('খ', 14, 160) ('幸', 12, 202) ('ம', 20, 114) ### ### ('貴', 16, 173) ('global', 1187, 12) ('comics', 198, 42) ('littered', 1602, 11) ### ############################################################################################################ [2023-11-26 04:06:50,022][root][INFO] - Epoch: 5: Step: 16301/19530, loss[v]=0.183227, lr=0.000001, acc@1[1]=974.5/1024=0.95166015625, acc@1[2]=1003.0/1024=0.9794921875 [2023-11-26 04:08:38,947][root][INFO] - Train batch 16400 [2023-11-26 04:08:38,948][root][INFO] - Avg. loss per last 100 batches: 0.120636 [2023-11-26 04:08:39,671][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.8/29522=99.92% | mean: 0.02 | max: 5.81 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the holy see, while continuing to appoint bishops to titular sees in north africa, ceased ### ### to make such appointments to sees that were historically part of the eastern patriarchate s of ### ### constantinople, alexandria, antioch, and jerusalem. ### ### [P_TEXT]: patriarchate of the west indies the titular patriarchate of the west indies ( ) is a ### ### latin rite titular patriarchate of the roman catholic church. it is vacant since the death of its ### ### last holder in 1963. attempt to create a jurisdictional patriarchate in the spanish west indies. ### ### king ferdinand v of castile asked pope leo x to establish a patriarchate for the ecclesiastical ### ### government of the american territories discovered by the spaniards. the holy see was not keen to ### ### accept the establishment of such an autonomous spanish american church and, on 11 may 1524, clement ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 202) ('貴', 1, 162) ('খ', 2, 256) ('幸', 3, 203) ('##ப', 4, 220) ('.', 5, 87) ### ### ('intently', 6, 235) ('ம', 7, 159) ('titular', 8, 47) ('jerusalem', 9, 2605) ('##ate', 10, 16) ### ### ('¹', 11, 132) ('constantinople', 12, 521) ('see', 13, 35) ('eastern', 14, 8376) ### ### ('bishop', 15, 262) ('##部', 16, 150) ('holy', 17, 77) ('alexandria', 18, 1773) ('patriarch', 19, 1) ### ### =============================================== v_p ================================================ ### ### ('indies', 3461, 0) ('patriarch', 19, 1) ('ferdinand', 2383, 2) ('clement', 932, 3) ### ### ('vacant', 800, 4) ('castile', 3355, 5) ('v', 14051, 6) ('1963', 3613, 7) ('152', 3745, 8) ### ### ('leaned', 1336, 9) ('spanish', 4959, 10) ('jurisdiction', 97, 11) ('x', 16075, 12) ### ### ('worried', 43, 13) ('littered', 1003, 14) ('catholicism', 5540, 15) ('##ate', 10, 16) ### ### ('americans', 9080, 17) ('leo', 2001, 18) ('territories', 1073, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('patriarch', 19, 1) ('##ate', 10, 16) ('see', 13, 35) ('church', 24, 24) ('titular', 8, 47) ### ### ('anyway', 29, 22) ('worried', 43, 13) ('.', 5, 87) ('ecclesiastical', 34, 27) ('absently', 0, 202) ### ### ('pope', 39, 29) ('holy', 17, 77) ('貴', 1, 162) ('jurisdiction', 97, 11) ('¹', 11, 132) ### ### ('ம', 7, 159) ('幸', 3, 203) ('##ப', 4, 220) ('খ', 2, 256) ('intently', 6, 235) ### ############################################################################################################ [2023-11-26 04:08:40,177][root][INFO] - Epoch: 5: Step: 16401/19530, loss[v]=0.196429, lr=0.000001, acc@1[1]=978.0/1024=0.955078125, acc@1[2]=995.5/1024=0.97216796875 [2023-11-26 04:10:29,601][root][INFO] - Train batch 16500 [2023-11-26 04:10:29,602][root][INFO] - Avg. loss per last 100 batches: 0.128516 [2023-11-26 04:10:30,335][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.7/29522=99.92% | mean: 0.02 | max: 5.91 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: pattern - welded steel has been referred to as " damascus steel " since 1973 when ### ### bladesmith william f. moran unveiled his " damascus knives " at the knifemakers'guild show. ### ### [P_TEXT]: bladesmith ancient process of forging damascus steel in the late 1960s. however, no ### ### living bladesmith knew the exact techniques and without a recipe for the process, it was in danger ### ### of being lost ; through trial and error he taught himself pattern welding and referred to his end ### ### product as " damascus steel ". in 1972, moran was elected president of the knifemakers'guild. the ### ### following year he unveiled his " damascus knives " at the guild show and created a revival of ### ### interest in the forged blade, and along with the knives he gave away free booklets detailing ### ### =============================================== v_q ================================================ ### ### ('steel', 0, 10) ('damascus', 1, 1) ('moran', 2, 0) ('¹', 3, 1226) ('welded', 4, 1705) ### ### ('knives', 5, 5) ('knife', 6, 9) ('guild', 7, 7) ('.', 8, 1666) ('f', 9, 22215) ('幸', 10, 392) ### ### ('william', 11, 19824) ('³', 12, 2894) ('খ', 13, 321) ('pattern', 14, 6) ('貴', 15, 317) ### ### ('##ப', 16, 282) ('ம', 17, 158) ('curious', 18, 1098) ('absently', 19, 143) ### ### =============================================== v_p ================================================ ### ### ('moran', 2, 0) ('damascus', 1, 1) ('##mith', 20, 2) ('blades', 23, 3) ('welding', 32, 4) ### ### ('knives', 5, 5) ('pattern', 14, 6) ('guild', 7, 7) ('forged', 57, 8) ('knife', 6, 9) ### ### ('steel', 0, 10) ('blade', 36, 11) ('##makers', 45, 12) ('recipe', 246, 13) ('show', 94, 14) ### ### ('error', 10748, 15) ('patterns', 395, 16) ('leaned', 2080, 17) ('product', 331, 18) ### ### ('trial', 6961, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('moran', 2, 0) ('damascus', 1, 1) ('steel', 0, 10) ('knives', 5, 5) ('guild', 7, 7) ### ### ('knife', 6, 9) ('pattern', 14, 6) ('##mith', 20, 2) ('blades', 23, 3) ('welding', 32, 4) ### ### ('blade', 36, 11) ('forged', 57, 8) ('##makers', 45, 12) ('show', 94, 14) ('recipe', 246, 13) ### ### ('forge', 165, 23) ('product', 331, 18) ('patterns', 395, 16) ('anyway', 202, 22) ('##部', 27, 93) ### ############################################################################################################ [2023-11-26 04:10:30,842][root][INFO] - Epoch: 5: Step: 16501/19530, loss[v]=0.067166, lr=0.000001, acc@1[1]=1002.5/1024=0.97900390625, acc@1[2]=1016.0/1024=0.9921875 [2023-11-26 04:12:20,022][root][INFO] - Train batch 16600 [2023-11-26 04:12:20,023][root][INFO] - Avg. loss per last 100 batches: 0.126936 [2023-11-26 04:12:20,732][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.5/29522=99.92% | mean: 0.02 | max: 5.82 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: ariundle oakwood (, " the shieling of the white meadow " ) is situated to the north of ### ### the village of strontian in the sunart area of the highlands of scotland. ### ### [P_TEXT]: strontian company, with many of those who presented evidence at the 1851 inquest involved ### ### in the action. the legal action failed and was ruled out of order by the sheriff substitute, with 4 ### ### pounds 15 shillings in court costs. ariundle oakwood is a national nature reserve and surviving ### ### fragment of the native oak woods that once spread along the atlantic coast from spain to norway. ### ### the village church was built in the 1820s by thomas telford, one of 32 " parliamentary churches " ### ### he designed for the highlands and islands. the government set up a commission ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 650) ('##tian', 1, 0) ('##ron', 2, 6) ('oak', 3, 3) ('ari', 4, 11) ('##wood', 5, 28) ### ### ('³', 6, 6064) ('sun', 7, 18508) ('##und', 8, 13) ('##art', 9, 25955) ('幸', 10, 470) ### ### ('scotland', 11, 959) ('luckily', 12, 270) ('shetland', 13, 3629) ('curious', 14, 476) ### ### ('貴', 15, 375) ('##ப', 16, 320) ('intently', 17, 586) ('খ', 18, 321) ('meadow', 19, 1627) ### ### =============================================== v_p ================================================ ### ### ('##tian', 1, 0) ('highlands', 21, 1) ('##quest', 3510, 2) ('oak', 3, 3) ('sheriff', 906, 4) ### ### ('substitute', 11681, 5) ('##ron', 2, 6) ('parliamentary', 20772, 7) ('tel', 17950, 8) ### ### ('1851', 6213, 9) ('nature', 514, 10) ('ari', 4, 11) ('evidence', 913, 12) ('##und', 8, 13) ### ### ('churches', 1961, 14) ('pounds', 1034, 15) ('woods', 470, 16) ('pound', 1185, 17) ### ### ('spain', 7330, 18) ('norway', 288, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##tian', 1, 0) ('oak', 3, 3) ('##ron', 2, 6) ('ari', 4, 11) ('##und', 8, 13) ('highlands', 21, 1) ### ### ('##wood', 5, 28) ('highland', 27, 27) ('st', 24, 34) ('village', 74, 39) ('¹', 0, 650) ### ### ('churchyard', 112, 36) ('sheriff', 906, 4) ('nature', 514, 10) ('luckily', 12, 270) ### ### ('norway', 288, 19) ('forest', 53, 72) ('anyway', 149, 41) ('ம', 22, 200) ('woods', 470, 16) ### ############################################################################################################ [2023-11-26 04:12:21,236][root][INFO] - Epoch: 5: Step: 16601/19530, loss[v]=0.084211, lr=0.000001, acc@1[1]=1002.0/1024=0.978515625, acc@1[2]=1014.5/1024=0.99072265625 [2023-11-26 04:14:10,305][root][INFO] - Train batch 16700 [2023-11-26 04:14:10,306][root][INFO] - Avg. loss per last 100 batches: 0.115437 [2023-11-26 04:14:11,028][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.3/29522=99.92% | mean: 0.02 | max: 5.75 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.0/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the antarctic as the leader of the british australian and new zealand antarctic research ### ### expedition ( 1929 – 1931 ), which led to a territorial claim in the form of the australian ### ### antarctic territory. ### ### [P_TEXT]: mawson peak 2016. mawson peak was named by the 1948 anare heard island expedition after ### ### the australian geologist and explorer sir douglas mawson, the leader of banzare 1929 – 31, who ### ### visited the island in novemberdecember 1929. on 20 february 1950, whilst aboard hmas lebuan, thomas ### ### gratton ( tim ) young oam observed and recorded in the ship's log that mawson peak was an active ### ### volcano. the 1964 – 65 expedition to heard island was led by major warwick deacock, with the ### ### schooner patanela skippered by major bill tilman. they succeeded in climbing mawson peak for the ### ### first time, which is ### ### =============================================== v_q ================================================ ### ### ('antarctic', 0, 197) ('australian', 1, 35) ('zealand', 2, 184) ('¹', 3, 204) ('খ', 4, 185) ### ### ('expedition', 5, 10) ('幸', 6, 206) ('absently', 7, 161) ('貴', 8, 205) ('##ப', 9, 214) ### ### ('research', 10, 1657) ('bellowed', 11, 212) ('ம', 12, 139) ('³', 13, 2717) ('territory', 14, 911) ### ### ('retorted', 15, 198) ('exclaimed', 16, 780) ('australia', 17, 69) ('.', 18, 5291) ('1929', 19, 18) ### ### =============================================== v_p ================================================ ### ### ('heard', 971, 0) ('til', 17223, 1) ('##ws', 24537, 2) ('dea', 15127, 3) ('##cock', 10102, 4) ### ### ('##mber', 20363, 5) ('ban', 11410, 6) ('##att', 20314, 7) ('##zar', 15524, 8) ('peak', 1243, 9) ### ### ('expedition', 5, 10) ('expeditions', 9800, 11) ('warwick', 5234, 12) ('##bu', 14640, 13) ### ### ('##ane', 12753, 14) ('littered', 21, 15) ('schooner', 8626, 16) ('island', 117, 17) ### ### ('1929', 19, 18) ('volcano', 1656, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('expedition', 5, 10) ('australian', 1, 35) ('littered', 21, 15) ('1929', 19, 18) ### ### ('australia', 17, 69) ('antarctic', 0, 197) ('island', 117, 17) ('zealand', 2, 184) ### ### ('anyway', 141, 27) ('absently', 7, 161) ('ம', 12, 139) ('¹', 3, 204) ('douglas', 189, 22) ### ### ('খ', 4, 185) ('heard', 971, 0) ('幸', 6, 206) ('貴', 8, 205) ('##ப', 9, 214) ('bellowed', 11, 212) ### ### ('skipper', 383, 25) ### ############################################################################################################ [2023-11-26 04:14:11,534][root][INFO] - Epoch: 5: Step: 16701/19530, loss[v]=0.174688, lr=0.000001, acc@1[1]=975.0/1024=0.9521484375, acc@1[2]=1005.0/1024=0.9814453125 [2023-11-26 04:16:00,934][root][INFO] - Train batch 16800 [2023-11-26 04:16:00,935][root][INFO] - Avg. loss per last 100 batches: 0.121609 [2023-11-26 04:16:01,654][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.5/29522=99.92% | mean: 0.02 | max: 5.82 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.6/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: gdansk is home to the university of gdansk, gdansk university of technology, the national ### ### museum, the gdansk shakespeare theatre, the museum of ### ### [P_TEXT]: jolanta kwasniewska in 1997 founded the charity foundation ( " understanding without ### ### barriers " ), whose aims include providing help and support to sick and disabled children. she is ### ### also the host of the television show " lekcja stylu " ( " lessons in style " ) on tvn style. she ### ### was a graduate of the 9th high school in gdansk. in 1979, she graduated from the faculty of law and ### ### administration at the university of gdansk, then tried to pass the judge's apprenticeship exam. ### ### during her studies, she was the head of the faculty council of the socialist union of polish ### ### students. ### ### =============================================== v_q ================================================ ### ### ('gdansk', 0, 0) ('university', 1, 8) ('shakespeare', 2, 11308) ('museum', 3, 4635) ### ### ('technology', 4, 5265) ('leaned', 5, 34) ('.', 6, 1681) ('national', 7, 17498) ### ### ('theatre', 8, 6212) ('¹', 9, 1290) ('danzig', 10, 36) ('খ', 11, 229) ('cradled', 12, 914) ### ### ('##部', 13, 140) ('housed', 14, 1907) ('幸', 15, 347) ('poland', 16, 35) ('nacional', 17, 4760) ### ### ('outskirts', 18, 69) ('##ப', 19, 277) ### ### =============================================== v_p ================================================ ### ### ('gdansk', 0, 0) ('tv', 6848, 1) ('9th', 1330, 2) ('faculty', 53, 3) ('style', 5048, 4) ### ### ('lessons', 82, 5) ('charity', 1377, 6) ('polish', 21, 7) ('university', 1, 8) ### ### ('disabled', 27632, 9) ('##lan', 18393, 10) ('apprenticeship', 1141, 11) ('##nie', 6739, 12) ### ### ('barriers', 24012, 13) ('kw', 4478, 14) ('anyway', 1491, 15) ('faculties', 1072, 16) ### ### ('socialist', 3496, 17) ('exam', 7205, 18) ('examinations', 26136, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('gdansk', 0, 0) ('university', 1, 8) ('leaned', 5, 34) ('polish', 21, 7) ('danzig', 10, 36) ### ### ('poland', 16, 35) ('faculty', 53, 3) ('lessons', 82, 5) ('outskirts', 18, 69) ### ### ('students', 116, 20) ('worried', 193, 22) ('pomerania', 29, 70) ('##部', 13, 140) ### ### ('student', 49, 56) ('9th', 1330, 2) ('খ', 11, 229) ('classrooms', 414, 26) ### ### ('apprenticeship', 1141, 11) ('charity', 1377, 6) ('absently', 26, 141) ### ############################################################################################################ [2023-11-26 04:16:02,161][root][INFO] - Epoch: 5: Step: 16801/19530, loss[v]=0.134381, lr=0.000001, acc@1[1]=985.0/1024=0.9619140625, acc@1[2]=1009.0/1024=0.9853515625 [2023-11-26 04:17:51,098][root][INFO] - Train batch 16900 [2023-11-26 04:17:51,124][root][INFO] - Avg. loss per last 100 batches: 0.118238 [2023-11-26 04:17:51,830][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.1/29522=99.92% | mean: 0.02 | max: 5.83 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.2/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the production was directed by elijah moshinsky and starred nigel hawthorne as lewis with ### ### jane lapotaire as joy. ### ### [P_TEXT]: nigel hawthorne a 1974 production of " as you like it on broadway ". around this time, he ### ### was persuaded by ian mckellen and judi dench to join the royal shakespeare company. he also ### ### supplemented his income by appearing in television advertisements, including one for mackeson ### ### stout. he returned to the new york stage in 1990 in " shadowlands " and won the 1991 tony award for ### ### best actor in a play. although hawthorne had appeared in small roles in various british television ### ### series since the ### ### =============================================== v_q ================================================ ### ### ('nigel', 0, 1) ('hawthorne', 1, 0) ('elijah', 2, 14183) ('##ota', 3, 27097) ('##shin', 4, 29080) ### ### ('lap', 5, 18109) ('##sky', 6, 26804) ('##ire', 7, 21150) ('lewis', 8, 15618) ('jane', 9, 11215) ### ### ('¹', 10, 483) ('mo', 11, 23686) ('joy', 12, 9515) ('″', 13, 41) ('absently', 14, 76) ### ### ('ம', 15, 103) ('theatre', 16, 27) ('খ', 17, 139) ('"', 18, 168) ('intently', 19, 200) ### ### =============================================== v_p ================================================ ### ### ('hawthorne', 1, 0) ('nigel', 0, 1) ('stout', 3047, 2) ('##lands', 20219, 3) ('shadow', 2319, 4) ### ### ('##eson', 24225, 5) ('like', 18165, 6) ('mack', 17796, 7) ('##kell', 14882, 8) ('tony', 182, 9) ### ### ('den', 3834, 10) ('broadway', 36, 11) ('liked', 2464, 12) ('you', 694, 13) ('worried', 1266, 14) ### ### ('littered', 4860, 15) ('advertising', 8043, 16) ('1991', 378, 17) ('1974', 481, 18) ### ### ('shakespeare', 57, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('hawthorne', 1, 0) ('nigel', 0, 1) ('″', 13, 41) ('theatre', 16, 27) ('broadway', 36, 11) ### ### ('play', 32, 23) ('absently', 14, 76) ('actor', 39, 30) ('shakespeare', 57, 19) ('tony', 182, 9) ### ### ('ம', 15, 103) ('##部', 25, 79) ('plays', 83, 20) ('anyway', 85, 26) ('leaned', 131, 24) ### ### ('production', 31, 69) ('খ', 17, 139) ('"', 18, 168) ('##ப', 26, 138) ('she', 59, 54) ### ############################################################################################################ [2023-11-26 04:17:52,334][root][INFO] - Epoch: 5: Step: 16901/19530, loss[v]=0.121767, lr=0.000001, acc@1[1]=1002.0/1024=0.978515625, acc@1[2]=1014.0/1024=0.990234375 [2023-11-26 04:19:41,289][root][INFO] - Train batch 17000 [2023-11-26 04:19:41,289][root][INFO] - Avg. loss per last 100 batches: 0.123615 [2023-11-26 04:19:42,006][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.8/29522=99.92% | mean: 0.02 | max: 5.85 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in the game against siaa champion auburn in 1919, he and frank goar were sent in on ### ### auburn's last drive. ### ### [P_TEXT]: 1919 vanderbilt commodores football team halfback ), berryhill ( right halfback ), ### ### hendrix ( fullback ). siaa champion auburn suffered its only loss to vanderbilt, 7 – 6. josh cody ### ### returned a fumble 15 yards for a touchdown and made the winning extra point. a bit after, ed ### ### sherling ran in from the 1 - yard line, but pete bonner missed goal. tom lipscomb and frank goar ### ### were sent in on auburn's last drive. goar had been sick and lipscomb was suffering from an injured ### ### ankle. the two spurned the team to victory in what the vanderbilt yearbook called " the greatest ### ### defensive stand ever staged by ### ### =============================================== v_q ================================================ ### ### ('goa', 0, 1) ('auburn', 1, 6) ('1919', 2, 2) ('frank', 3, 19) ('¹', 4, 126) ('football', 5, 38) ### ### ('খ', 6, 168) ('幸', 7, 173) ('³', 8, 997) ('##r', 9, 49) ('si', 10, 18) ('貴', 11, 148) ### ### ('##aa', 12, 10) ('##ப', 13, 178) ('exclaimed', 14, 520) ('willie', 15, 6457) ('bellowed', 16, 116) ### ### ('littered', 17, 21) ('##isman', 18, 4974) ('intently', 19, 97) ### ### =============================================== v_p ================================================ ### ### ('##rling', 12684, 0) ('goa', 0, 1) ('1919', 2, 2) ('cody', 1298, 3) ('vanderbilt', 25, 4) ### ### ('bonn', 21885, 5) ('auburn', 1, 6) ('hendrix', 4652, 7) ('lips', 11991, 8) ('##comb', 13176, 9) ### ### ('##aa', 12, 10) ('##hill', 15547, 11) ('josh', 5276, 12) ('anyway', 178, 13) ('berry', 9487, 14) ### ### ('worried', 2238, 15) ('fullback', 71, 16) ('commodore', 904, 17) ('si', 10, 18) ('frank', 3, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('goa', 0, 1) ('1919', 2, 2) ('auburn', 1, 6) ('frank', 3, 19) ('vanderbilt', 25, 4) ### ### ('##aa', 12, 10) ('si', 10, 18) ('football', 5, 38) ('littered', 17, 21) ('fullback', 71, 16) ### ### ('##r', 9, 49) ('1920', 20, 43) ('anyway', 178, 13) ('¹', 4, 126) ('team', 64, 27) ('1918', 35, 42) ### ### ('fumble', 58, 30) ('coach', 34, 50) ('yearbook', 177, 25) ('cody', 1298, 3) ### ############################################################################################################ [2023-11-26 04:19:42,514][root][INFO] - Epoch: 5: Step: 17001/19530, loss[v]=0.067541, lr=0.000001, acc@1[1]=1001.5/1024=0.97802734375, acc@1[2]=1016.5/1024=0.99267578125 [2023-11-26 04:21:31,834][root][INFO] - Train batch 17100 [2023-11-26 04:21:31,835][root][INFO] - Avg. loss per last 100 batches: 0.120877 [2023-11-26 04:21:32,548][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.7/29522=99.92% | mean: 0.02 | max: 5.89 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.3/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: in 1941, albert m. baer purchased the ulster knife company ( which was founded in ### ### ellenville, new york, in the 1870s ) and merged it with the imperial knife company and designated ### ### this new business as the imperial knife associated companies, to produce knives for the military. ### ### [P_TEXT]: ellenville, new york world war ii. in the late 1940s, joseph resnick, a radio officer in ### ### the u. s. merchant marine during the war, developed a turnable diy antenna system just as the tv ### ### boom was taking off. he, and his brothers harry and louis, created channel master, and built one of ### ### the region's major manufacturing plants just north of the ellenville border, along with an aluminum ### ### plant to fabricate necessary components. after selling the company to avnet, production moved to ### ### south carolina. the vacated channel master factory was eventually bought by imperial schrade ### ### =============================================== v_q ================================================ ### ### ('imperial', 0, 16) ('¹', 1, 333) ('knife', 2, 5303) ('bae', 3, 28162) ('ellen', 4, 0) ### ### ('ulster', 5, 4174) ('knives', 6, 6963) ('absently', 7, 144) ('albert', 8, 6906) ('খ', 9, 272) ### ### ('##ville', 10, 9) ('³', 11, 1761) ('幸', 12, 305) ('貴', 13, 298) ('accessory', 14, 16096) ### ### ('york', 15, 40) ('##ப', 16, 286) ('company', 17, 47) ('m', 18, 23950) ('intently', 19, 265) ### ### =============================================== v_p ================================================ ### ### ('ellen', 4, 0) ('boom', 2072, 1) ('##nick', 21517, 2) ('av', 20854, 3) ('antenna', 13735, 4) ### ### ('##net', 22234, 5) ('##hra', 25959, 6) ('antennas', 28632, 7) ('##able', 10020, 8) ### ### ('##ville', 10, 9) ('master', 3069, 10) ('res', 26116, 11) ('channel', 16985, 12) ('di', 27826, 13) ### ### ('littered', 2194, 14) ('louis', 8122, 15) ('imperial', 0, 16) ('brother', 23818, 17) ### ### ('turn', 8118, 18) ('joseph', 7076, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('ellen', 4, 0) ('imperial', 0, 16) ('##ville', 10, 9) ('york', 15, 40) ('company', 17, 47) ### ### ('absently', 7, 144) ('¹', 1, 333) ('factory', 136, 24) ('manufacturing', 138, 30) ### ### ('fabric', 233, 25) ('খ', 9, 272) ('exclaimed', 24, 113) ('purchased', 98, 45) ### ### ('companies', 25, 136) ('1940s', 236, 35) ('幸', 12, 305) ('##部', 27, 131) ('ம', 23, 181) ### ### ('boom', 2072, 1) ('貴', 13, 298) ### ############################################################################################################ [2023-11-26 04:21:33,054][root][INFO] - Epoch: 5: Step: 17101/19530, loss[v]=0.102031, lr=0.000001, acc@1[1]=1002.5/1024=0.97900390625, acc@1[2]=1015.5/1024=0.99169921875 [2023-11-26 04:23:21,885][root][INFO] - Train batch 17200 [2023-11-26 04:23:21,886][root][INFO] - Avg. loss per last 100 batches: 0.132724 [2023-11-26 04:23:22,588][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.6/29522=99.92% | mean: 0.02 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: there, lancelot's and lady elaine's son galahad, devoid of his father's flaws of ### ### character, becomes the perfect knight and succeeds in completing the greatest of all quests by ### ### retrieving the holy grail ### ### [P_TEXT]: galahad sir galahad ( ; sometimes referred to as galeas or galath ) among other versions ### ### of his name, is a knight of king arthur's round table and one of the three achievers of the holy ### ### grail in arthurian legend. he is the illegitimate son of sir lancelot and elaine of corbenic, and ### ### is renowned for his gallantry and purity as the most perfect of all knights. emerging quite late in ### ### the medieval arthurian tradition, sir galahad first appears in the lancelot – grail cycle, and his ### ### story is taken ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 198) ('##had', 1, 2) ('gala', 2, 1) ('knight', 3, 19) ('elaine', 4, 0) ### ### ('貴', 5, 180) ('খ', 6, 202) ('幸', 7, 181) ('intently', 8, 178) ('##ail', 9, 7) ('##部', 10, 119) ### ### ('ம', 11, 115) ('##ப', 12, 196) ('¹', 13, 117) ('holy', 14, 20) ('.', 15, 121) ('gr', 16, 44) ### ### ('lance', 17, 8) ('perfect', 18, 25) ('arthur', 19, 4) ### ### =============================================== v_p ================================================ ### ### ('elaine', 4, 0) ('gala', 2, 1) ('##had', 1, 2) ('gale', 6876, 3) ('arthur', 19, 4) ### ### ('table', 1256, 5) ('leaned', 703, 6) ('##ail', 9, 7) ('lance', 17, 8) ('gallantry', 16713, 9) ### ### ('purity', 7289, 10) ('##rb', 11274, 11) ('achieve', 2297, 12) ('worried', 30, 13) ### ### ('cycle', 265, 14) ('illegitimate', 6886, 15) ('anyway', 52, 16) ('##lot', 92, 17) ### ### ('##eni', 28709, 18) ('knight', 3, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('elaine', 4, 0) ('gala', 2, 1) ('##had', 1, 2) ('##ail', 9, 7) ('knight', 3, 19) ('arthur', 19, 4) ### ### ('lance', 17, 8) ('holy', 14, 20) ('perfect', 18, 25) ('worried', 30, 13) ('gr', 16, 44) ### ### ('knights', 22, 36) ('anyway', 52, 16) ('absently', 0, 198) ('medieval', 53, 22) ('##lot', 92, 17) ### ### ('##部', 10, 119) ('ம', 11, 115) ('legend', 58, 39) ('貴', 5, 180) ### ############################################################################################################ [2023-11-26 04:23:23,092][root][INFO] - Epoch: 5: Step: 17201/19530, loss[v]=0.052149, lr=0.000001, acc@1[1]=1008.5/1024=0.98486328125, acc@1[2]=1018.5/1024=0.99462890625 [2023-11-26 04:25:12,018][root][INFO] - Train batch 17300 [2023-11-26 04:25:12,019][root][INFO] - Avg. loss per last 100 batches: 0.118203 [2023-11-26 04:25:12,751][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.88 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.0/29522=99.92% | mean: 0.02 | max: 5.77 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: san mateo union high school district operates local high schools while the burlingame ### ### school district operates elementary and middle schools. ### ### [P_TEXT]: burlingame high school ( california ) burlingame high school is a public high school in ### ### burlingame, california. it is part of the san mateo union high school district ( smuhsd ). in order ### ### to meet the growing student population, the school was opened in december 1923 under the name " san ### ### mateo high school, burlingame branch. " designed by architect w. h. weeks, the school took in ### ### students from burlingame, hillsborough, millbrae, and san bruno. initial enrollment consisted of ### ### 350 students and 30 teachers. as a branch of san mateo high school, extracurricular organizations ### ### were shared between the schools. there was a single band, football team, and ### ### =============================================== v_q ================================================ ### ### ('union', 0, 19) ('mateo', 1, 2) ('##ame', 2, 4) ('##rling', 3, 1) ('school', 4, 17) ('幸', 5, 154) ### ### ('¹', 6, 90) ('貴', 7, 108) ('intently', 8, 72) ('খ', 9, 130) ('##ப', 10, 129) ('san', 11, 12) ### ### ('bu', 12, 27) ('high', 13, 33) ('cradled', 14, 50) ('absently', 15, 83) ('ம', 16, 82) ### ### ('schoolhouse', 17, 15) ('³', 18, 323) ('snuggled', 19, 105) ### ### =============================================== v_p ================================================ ### ### ('hillsborough', 1043, 0) ('##rling', 3, 1) ('mateo', 1, 2) ('bruno', 4311, 3) ('##ame', 2, 4) ### ### ('weeks', 3244, 5) ('mill', 6770, 6) ('##bra', 25902, 7) ('##uh', 26892, 8) ('littered', 738, 9) ### ### ('neighboring', 182, 10) ('worried', 24, 11) ('san', 11, 12) ('##sd', 16606, 13) ('anyway', 28, 14) ### ### ('schoolhouse', 17, 15) ('##rricular', 4103, 16) ('school', 4, 17) ('h', 9307, 18) ('union', 0, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('mateo', 1, 2) ('##rling', 3, 1) ('##ame', 2, 4) ('union', 0, 19) ('school', 4, 17) ### ### ('san', 11, 12) ('schoolhouse', 17, 15) ('bu', 12, 27) ('worried', 24, 11) ('high', 13, 33) ### ### ('california', 21, 24) ('anyway', 28, 14) ('cradled', 14, 50) ('intently', 8, 72) ### ### ('schooling', 27, 22) ('¹', 6, 90) ('district', 23, 47) ('貴', 7, 108) ('absently', 15, 83) ### ### ('ம', 16, 82) ### ############################################################################################################ [2023-11-26 04:25:13,259][root][INFO] - Epoch: 5: Step: 17301/19530, loss[v]=0.146829, lr=0.000001, acc@1[1]=986.0/1024=0.962890625, acc@1[2]=1006.5/1024=0.98291015625 [2023-11-26 04:27:02,405][root][INFO] - Train batch 17400 [2023-11-26 04:27:02,405][root][INFO] - Avg. loss per last 100 batches: 0.129947 [2023-11-26 04:27:03,120][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.97 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.7/29522=99.92% | mean: 0.02 | max: 5.87 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: sylvain saudan ( born 23 september 1936 in lausanne, switzerland ) is an extreme skier, ### ### dubbed " skier of the impossible. " ### ### [P_TEXT]: extreme skiing extreme skiing is performed on long, steep ( typically from 45 to 60 + ### ### degrees, or grades of 100 to 170 percent ) slopes in mountain ous terrain. the french coined the ### ### term'le ski extreme'in the 1970s. the first practitioners include swiss skier sylvain saudan, who ### ### invented the " windshield wiper " turn in the mid - 1960s, and in 1967 made the first descents of ### ### slopes in the swiss, french and italian alps that were previously considered impossible. ### ### saudan's'first descent'in america was at mt. hood march 3, 1971. early american practitioners ### ### include bill briggs, who descended grand teton on june 15, 1971. the frenchmen patrick vallencant, ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 190) ('impossible', 1, 8) ('##uda', 2, 6) ('curious', 3, 286) ('³', 4, 1074) ### ### ('##lva', 5, 32) ('skier', 6, 48) ('sy', 7, 41) ('lausanne', 8, 3544) ('幸', 9, 202) ('sa', 10, 49) ### ### ('absently', 11, 148) ('extreme', 12, 1) ('##ப', 13, 194) ('intently', 14, 171) ('貴', 15, 231) ### ### ('খ', 16, 196) ('ski', 17, 15) ('switzerland', 18, 77) ('ம', 19, 164) ### ### =============================================== v_p ================================================ ### ### ('briggs', 13066, 0) ('extreme', 12, 1) ('wipe', 3318, 2) ('windshield', 15434, 3) ### ### ('hood', 7790, 4) ('skiing', 833, 5) ('##uda', 2, 6) ('grading', 28590, 7) ('impossible', 1, 8) ### ### ('ou', 13059, 9) ('littered', 91, 10) ('terrain', 1318, 11) ('alps', 146, 12) ('##nca', 8530, 13) ### ### ('italian', 3890, 14) ('ski', 17, 15) ('bill', 1648, 16) ('te', 20451, 17) ('grade', 16773, 18) ### ### ('turn', 2741, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##uda', 2, 6) ('impossible', 1, 8) ('extreme', 12, 1) ('##lva', 5, 32) ('ski', 17, 15) ### ### ('sy', 7, 41) ('skier', 6, 48) ('sa', 10, 49) ('swiss', 34, 25) ('littered', 91, 10) ('¹', 0, 190) ### ### ('switzerland', 18, 77) ('alps', 146, 12) ('mountain', 54, 36) ('absently', 11, 148) ### ### ('curious', 3, 286) ('anyway', 147, 24) ('alpine', 52, 51) ('幸', 9, 202) ('intently', 14, 171) ### ############################################################################################################ [2023-11-26 04:27:03,625][root][INFO] - Epoch: 5: Step: 17401/19530, loss[v]=0.087184, lr=0.000001, acc@1[1]=996.5/1024=0.97314453125, acc@1[2]=1013.5/1024=0.98974609375 [2023-11-26 04:28:52,315][root][INFO] - Train batch 17500 [2023-11-26 04:28:52,316][root][INFO] - Avg. loss per last 100 batches: 0.120114 [2023-11-26 04:28:53,031][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.6/29522=99.92% | mean: 0.02 | max: 5.90 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: song was co - written by mike campbell, along with petty and jeff lynne. ### ### [P_TEXT]: jeff lynne in 1989, lynne co - produced " full moon fever " by tom petty, which included ### ### the hit singles " free fallin'", " i won't back down " and " runnin'down a dream ", all co - ### ### written by lynne. this album and " traveling wilburys vol. 1 " received nominations for the grammy ### ### award for best album of the year in 1989. the traveling wilburys won a grammy for " best rock ### ### performance by a duo or group with vocal " that year. lynne's song " one way love " was released as ### ### =============================================== v_q ================================================ ### ### ('campbell', 0, 10056) ('lynne', 1, 0) ('mike', 2, 9997) ('″', 3, 10) ('petty', 4, 6) ### ### ('jeff', 5, 1) ('¹', 6, 73) ('absently', 7, 59) ('³', 8, 347) ('খ', 9, 108) ('幸', 10, 100) ### ### ('##½', 11, 319) ('intently', 12, 92) ('"', 13, 135) ('貴', 14, 107) ('」', 15, 110) ('##ப', 16, 106) ### ### ('.', 17, 7226) ('fender', 18, 2810) ('royalties', 19, 670) ### ### =============================================== v_p ================================================ ### ### ('lynne', 1, 0) ('jeff', 5, 1) ('fever', 20579, 2) ('wilbur', 4275, 3) ('##nin', 26687, 4) ### ### ('traveling', 844, 5) ('petty', 4, 6) ('moon', 796, 7) ('free', 9033, 8) ('fall', 644, 9) ### ### ('″', 3, 10) ('##in', 20026, 11) ('dream', 321, 12) ('way', 4454, 13) ('nominees', 10528, 14) ### ### ('back', 1360, 15) ('1989', 185, 16) ('full', 7096, 17) ('down', 2765, 18) ('lyrical', 558, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('lynne', 1, 0) ('jeff', 5, 1) ('petty', 4, 6) ('″', 3, 10) ('absently', 7, 59) ('¹', 6, 73) ### ### ('album', 63, 23) ('幸', 10, 100) ('grammy', 88, 20) ('খ', 9, 108) ('intently', 12, 92) ### ### ('1989', 185, 16) ('songwriting', 46, 32) ('dream', 321, 12) ('貴', 14, 107) ('ம', 21, 78) ### ### ('##ப', 16, 106) ('」', 15, 110) ('traveling', 844, 5) ('fall', 644, 9) ### ############################################################################################################ [2023-11-26 04:28:53,535][root][INFO] - Epoch: 5: Step: 17501/19530, loss[v]=0.095777, lr=0.000001, acc@1[1]=999.5/1024=0.97607421875, acc@1[2]=1015.0/1024=0.9912109375 [2023-11-26 04:30:42,516][root][INFO] - Train batch 17600 [2023-11-26 04:30:42,516][root][INFO] - Avg. loss per last 100 batches: 0.115860 [2023-11-26 04:30:43,262][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 4.00 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29496.9/29522=99.91% | mean: 0.02 | max: 5.91 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.3/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: sony pictures also announced that goddard would write and direct a film based on the ### ### sinister six, though the project was eventually cancelled. ### ### [P_TEXT]: sinister six after the sinister six caught the goblin who secretly was helping gabriel, ### ### miguel saves goblin ( father jennifer ) and escapes back into present time. as the portal opened, ### ### doctor octopus 2099 kills goblin and escapes. now that the sinister six is no more, venture enters ### ### and joins the team where he plans to recruit them to fist. in december 2013, sony announced two ### ### spin - offs of " the amazing spider - man " franchise, " the sinister six " and " venom ", with ### ### drew goddard attached to write and direct " the sinister six ". throughout the live - action film " ### ### the amazing spider - man 2 ", a line - up ### ### =============================================== v_q ================================================ ### ### ('goddard', 0, 3) ('sinister', 1, 0) ('six', 2, 7) ('¹', 3, 55) ('sony', 4, 14) ('film', 5, 33) ### ### ('absently', 6, 102) ('³', 7, 337) ('″', 8, 19) ('幸', 9, 115) ('貴', 10, 127) ('##ப', 11, 110) ### ### ('খ', 12, 107) ('intently', 13, 94) ('freaking', 14, 875) ('psycho', 15, 3327) ('.', 16, 76) ### ### ('pictures', 17, 542) ('6', 18, 60) ('crowe', 19, 17905) ### ### =============================================== v_p ================================================ ### ### ('sinister', 1, 0) ('venom', 1018, 1) ('goblin', 4006, 2) ('goddard', 0, 3) ('miguel', 6184, 4) ### ### ('venture', 2369, 5) ('fist', 4415, 6) ('six', 2, 7) ('octopus', 2270, 8) ('jennifer', 1723, 9) ### ### ('gabriel', 2132, 10) ('franchise', 97, 11) ('drew', 6083, 12) ('spider', 45, 13) ('sony', 4, 14) ### ### ('amazing', 5231, 15) ('209', 3796, 16) ('leaned', 2815, 17) ('##9', 15165, 18) ('″', 8, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('sinister', 1, 0) ('goddard', 0, 3) ('six', 2, 7) ('sony', 4, 14) ('″', 8, 19) ('film', 5, 33) ### ### ('¹', 3, 55) ('spider', 45, 13) ('franchise', 97, 11) ('venom', 1018, 1) ('absently', 6, 102) ### ### ('doctor', 32, 35) ('revealed', 52, 28) ('sequel', 36, 36) ('幸', 9, 115) ('6', 18, 60) ### ### ('##ப', 11, 110) ('intently', 13, 94) ('貴', 10, 127) ('খ', 12, 107) ### ############################################################################################################ [2023-11-26 04:30:43,770][root][INFO] - Epoch: 5: Step: 17601/19530, loss[v]=0.087603, lr=0.000000, acc@1[1]=1004.0/1024=0.98046875, acc@1[2]=1016.0/1024=0.9921875 [2023-11-26 04:32:32,818][root][INFO] - Train batch 17700 [2023-11-26 04:32:32,819][root][INFO] - Avg. loss per last 100 batches: 0.132054 [2023-11-26 04:32:33,536][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.7/29522=99.92% | mean: 0.02 | max: 5.85 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.9/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: spain was subsequently placed to perform in position 21, following the entry from romania ### ### and before the entry from hungary. ### ### [P_TEXT]: hungary in the eurovision song contest 2015 in position 22, following the entry from ### ### spain and before the entry from georgia. boggie once again took part in dress rehearsals on 22 and ### ### 23 may before the final, including the jury final where the professional juries cast their final ### ### votes before the live show. boggie performed a repeat of her semi - final performance during the ### ### final on 23 may. at the conclusion of the voting, hungary finished in twentieth place with 19 ### ### points. voting during the three shows consisted of 50 percent public televoting and 50 percent from ### ### a jury deliberation. the jury consisted ### ### =============================================== v_q ================================================ ### ### ('spain', 0, 2) ('absently', 1, 113) ('貴', 2, 175) ('romania', 3, 180) ('খ', 4, 236) ('¹', 5, 241) ### ### ('幸', 6, 208) ('intently', 7, 159) ('contest', 8, 3) ('ம', 9, 126) ('##ப', 10, 166) ### ### ('eurovision', 11, 75) ('hungary', 12, 0) ('.', 13, 103) ('21', 14, 420) ('##部', 15, 119) ### ### ('³', 16, 210) ('spanish', 17, 34) ('representative', 18, 3516) ('placing', 19, 1022) ### ### =============================================== v_p ================================================ ### ### ('hungary', 12, 0) ('georgia', 4365, 1) ('spain', 0, 2) ('contest', 8, 3) ('preceding', 33, 4) ### ### ('2015', 170, 5) ('dress', 1257, 6) ('entries', 979, 7) ('##ibe', 23609, 8) ('bog', 6339, 9) ### ### ('contests', 1559, 10) ('anyway', 25, 11) ('##ration', 4870, 12) ('percent', 5802, 13) ### ### ('georgian', 19643, 14) ('##gie', 11210, 15) ('jury', 217, 16) ('position', 84, 17) ### ### ('entry', 129, 18) ('voting', 51, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('spain', 0, 2) ('hungary', 12, 0) ('contest', 8, 3) ('anyway', 25, 11) ('preceding', 33, 4) ### ### ('song', 20, 25) ('spanish', 17, 34) ('eurovision', 11, 75) ('absently', 1, 113) ### ### ('hungarian', 42, 22) ('voting', 51, 19) ('.', 13, 103) ('ம', 9, 126) ('貴', 2, 175) ### ### ('intently', 7, 159) ('romania', 3, 180) ('##ப', 10, 166) ('position', 84, 17) ('##部', 15, 119) ### ### ('2015', 170, 5) ### ############################################################################################################ [2023-11-26 04:32:34,039][root][INFO] - Epoch: 5: Step: 17701/19530, loss[v]=0.073913, lr=0.000000, acc@1[1]=998.0/1024=0.974609375, acc@1[2]=1018.0/1024=0.994140625 [2023-11-26 04:34:22,702][root][INFO] - Train batch 17800 [2023-11-26 04:34:22,702][root][INFO] - Avg. loss per last 100 batches: 0.115442 [2023-11-26 04:34:23,417][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.3/29522=99.92% | mean: 0.02 | max: 5.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.6/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: lady nicholas windsor ( born paola louise marica doimi de lupis, 7 august 1969 ) is the ### ### wife of lord nicholas windsor, son of the duke and duchess of kent. ### ### [P_TEXT]: kelly rondestvedt on may 23, 2009. there were over 400 wedding guests and 3000 ### ### spectators, including king carl xvi gustaf and queen silvia of sweden, lord nicholas windsor, lady ### ### nicholas windsor, simeon ii of bulgaria, and princess astrid of belgium, archduchess of austria - ### ### este. ### ### =============================================== v_q ================================================ ### ### ('windsor', 0, 3) ('nicholas', 1, 4) ('¹', 2, 53) ('lord', 3, 29) ('lady', 4, 27) ('kent', 5, 4674) ### ### ('³', 6, 818) ('absently', 7, 132) ('duchess', 8, 169) ('খ', 9, 106) ('貴', 10, 108) ('幸', 11, 118) ### ### ('##ப', 12, 78) ('##pis', 13, 24490) ('duke', 14, 1171) ('.', 15, 74) ('intently', 16, 93) ### ### ('doi', 17, 25251) ('ம', 18, 71) ('prince', 19, 86) ### ### =============================================== v_p ================================================ ### ### ('simeon', 19850, 0) ('astrid', 11382, 1) ('silvia', 5317, 2) ('windsor', 0, 3) ('nicholas', 1, 4) ### ### ('kelly', 10297, 5) ('##tv', 23652, 6) ('belgium', 384, 7) ('##des', 16644, 8) ('##ches', 17399, 9) ### ### ('ron', 16655, 10) ('gust', 22920, 11) ('spectators', 6163, 12) ('sweden', 7117, 13) ### ### ('bulgaria', 17011, 14) ('bulgarian', 14504, 15) ('austria', 8151, 16) ('austrians', 24001, 17) ### ### ('este', 10257, 18) ('princess', 23, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('windsor', 0, 3) ('nicholas', 1, 4) ('lord', 3, 29) ('lady', 4, 27) ('¹', 2, 53) ### ### ('princess', 23, 19) ('queen', 24, 30) ('monarch', 48, 20) ('wedding', 119, 22) ('guests', 60, 34) ### ### ('##ப', 12, 78) ('belgium', 384, 7) ('##部', 20, 61) ('খ', 9, 106) ('.', 15, 74) ('couple', 58, 40) ### ### ('hissed', 36, 42) ('貴', 10, 108) ('ம', 18, 71) ('absently', 7, 132) ### ############################################################################################################ [2023-11-26 04:34:23,922][root][INFO] - Epoch: 5: Step: 17801/19530, loss[v]=0.172940, lr=0.000000, acc@1[1]=982.0/1024=0.958984375, acc@1[2]=1008.5/1024=0.98486328125 [2023-11-26 04:36:13,194][root][INFO] - Train batch 17900 [2023-11-26 04:36:13,194][root][INFO] - Avg. loss per last 100 batches: 0.116854 [2023-11-26 04:36:13,922][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.4/29522=99.92% | mean: 0.02 | max: 5.87 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the airport is only 15 km south of the colombo city centre compared to the larger ### ### bandaranaike international airport which is 32 km north of the city. ### ### [P_TEXT]: colombo orbital bypass outer circular highway ( arthur c. clarke expressway ). the ### ### colombo - katunayake expressway ( e03 ) runs from peliyagoda, a northern suburb of colombo, to ### ### colombo international airport and it linked with one of the major commercial hubs and a major ### ### tourist destination of the country, the city of negombo. an international ferry liner, the scotia ### ### prince, is conducting a ferry service to tuticorin, india. ferry services between the two countries ### ### have been revived after more than 20 years. ratmalana airport is the city's airport, located south ### ### of the city centre. it ### ### =============================================== v_q ================================================ ### ### ('airport', 0, 5) ('colombo', 1, 0) ('¹', 2, 99) ('##ike', 3, 20531) ('banda', 4, 11896) ### ### ('貴', 5, 178) ('##部', 6, 149) ('##ப', 7, 203) ('খ', 8, 155) ('幸', 9, 208) ('.', 10, 58) ### ### ('ம', 11, 127) ('lanka', 12, 67) ('³', 13, 444) ('##rana', 14, 24649) ('absently', 15, 60) ### ### ('intently', 16, 107) ('city', 17, 63) ('destined', 18, 235) ('outskirts', 19, 102) ### ### =============================================== v_p ================================================ ### ### ('colombo', 1, 0) ('scotia', 16126, 1) ('##mbo', 909, 2) ('circular', 2544, 3) ### ### ('expressway', 960, 4) ('airport', 0, 5) ('worried', 803, 6) ('prince', 9922, 7) ### ### ('clarke', 3354, 8) ('anyway', 78, 9) ('ferry', 4247, 10) ('##go', 12340, 11) ### ### ('ferries', 10450, 12) ('arthur', 11610, 13) ('pe', 25676, 14) ('ceylon', 31, 15) ('ne', 18248, 16) ### ### ('kat', 12672, 17) ('rat', 4379, 18) ('outer', 2000, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('colombo', 1, 0) ('airport', 0, 5) ('¹', 2, 99) ('ceylon', 31, 15) ('.', 10, 58) ('lanka', 12, 67) ### ### ('absently', 15, 60) ('city', 17, 63) ('anyway', 78, 9) ('##部', 6, 149) ('খ', 8, 155) ### ### ('ம', 11, 127) ('貴', 5, 178) ('intently', 16, 107) ('international', 20, 65) ('##ப', 7, 203) ### ### ('幸', 9, 208) ('##mbo', 909, 2) ('outskirts', 19, 102) ('expressway', 960, 4) ### ############################################################################################################ [2023-11-26 04:36:14,428][root][INFO] - Epoch: 5: Step: 17901/19530, loss[v]=0.069674, lr=0.000000, acc@1[1]=1004.0/1024=0.98046875, acc@1[2]=1019.0/1024=0.9951171875 [2023-11-26 04:38:03,319][root][INFO] - Train batch 18000 [2023-11-26 04:38:03,319][root][INFO] - Avg. loss per last 100 batches: 0.126261 [2023-11-26 04:38:04,017][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.97 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.1/29522=99.92% | mean: 0.02 | max: 5.86 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.2/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 77.2/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: initially, the enlightened vizier islam khodja played a large role in the running of the ### ### state. ### ### [P_TEXT]: islam khodja age of 26, he was appointed grand vizier of the khiva khanate, becoming the ### ### second statesman in the khanate, and one of the youngest grand viziers in the history of the khiva ### ### khanate. at the time, the khan was muhammad rahim khan ii. with his death in 1910, his son, ### ### isfandiyar khan, ascended the throne. with isfandiyar khan's enthronement, islam khodja began to ### ### play a much larger role in the khanate. isfandiyar khan was only one year older than him, and, ### ### unlike his father, isfandiyar did not have any special abilities in governing. islam khodja was ### ### popular ### ### =============================================== v_q ================================================ ### ### ('##ho', 0, 3) ('貴', 1, 140) ('intently', 2, 151) ('幸', 3, 111) ('খ', 4, 136) ('viz', 5, 19) ### ### ('islam', 6, 1) ('¹', 7, 114) ('##ப', 8, 123) ('absently', 9, 120) ('ம', 10, 89) ### ### ('##light', 11, 21208) ('##ja', 12, 6) ('##ier', 13, 1616) ('state', 14, 8613) ('anyway', 15, 17) ### ### ('k', 16, 30) ('##部', 17, 80) ('³', 18, 1477) ('##d', 19, 49) ### ### =============================================== v_p ================================================ ### ### ('##him', 9189, 0) ('islam', 6, 1) ('khan', 24, 2) ('##ho', 0, 3) ('##fan', 12348, 4) ### ### ('1910', 4533, 5) ('##ja', 12, 6) ('##ate', 1295, 7) ('##hi', 14721, 8) ('##yar', 3124, 9) ### ### ('histories', 12613, 10) ('##va', 4047, 11) ('youngest', 4348, 12) ('littered', 1693, 13) ### ### ('muhammad', 501, 14) ('worried', 25, 15) ('mongols', 2666, 16) ('anyway', 15, 17) ### ### ('leaned', 787, 18) ('viz', 5, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ho', 0, 3) ('islam', 6, 1) ('##ja', 12, 6) ('viz', 5, 19) ('khan', 24, 2) ('anyway', 15, 17) ### ### ('k', 16, 30) ('worried', 25, 15) ('ம', 10, 89) ('幸', 3, 111) ('貴', 1, 140) ('intently', 2, 151) ### ### ('খ', 4, 136) ('¹', 7, 114) ('##d', 19, 49) ('absently', 9, 120) ('##ப', 8, 123) ('##部', 17, 80) ### ### ('muslim', 58, 25) ('statesman', 92, 20) ### ############################################################################################################ [2023-11-26 04:38:04,520][root][INFO] - Epoch: 5: Step: 18001/19530, loss[v]=0.074439, lr=0.000000, acc@1[1]=1001.5/1024=0.97802734375, acc@1[2]=1012.0/1024=0.98828125 [2023-11-26 04:39:53,182][root][INFO] - Train batch 18100 [2023-11-26 04:39:53,182][root][INFO] - Avg. loss per last 100 batches: 0.117714 [2023-11-26 04:39:53,903][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.4/29522=99.99% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.1/29522=99.92% | mean: 0.02 | max: 5.84 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.8/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.3/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he is attracted to, then, later, obsessed with mick's partner, linda carter ( kellie ### ### bright ), who maternally hugs him after shirley hurts him by stealing from him because she owes ### ### phil mitchell ( steve mcfadden ). ### ### [P_TEXT]: list of eastenders characters ( 2008 ) in 2015, roxy reveals that she arranges monthly ### ### trips for amy to go to france to visit jack. during christmas, jack returns to walford to visit ### ### roxy and amy, and upon finding out that roxy's fiance dean wicks ( matt di angelo ) raped linda ### ### carter ( kellie bright ), he packs amy's bags and sends her to ronnie's house, scolding roxy for ### ### letting dean control her. roxy leaves amy with jack in january 2016 after dean tries to rape her. ### ### jack and amy temporarily leave walford but return when jack decides ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 167) ('linda', 1, 1) ('phil', 2, 15177) ('shirley', 3, 3510) ('mitchell', 4, 4747) ### ### ('¹', 5, 42) ('mick', 6, 10032) ('খ', 7, 117) ('##ப', 8, 88) ('carter', 9, 8) ('.', 10, 444) ### ### ('##部', 11, 127) ('貴', 12, 118) ('intently', 13, 70) ('幸', 14, 91) ('eastenders', 15, 3) ### ### ('ம', 16, 69) ('³', 17, 949) ('disgusted', 18, 525) ('blackmail', 19, 97) ### ### =============================================== v_p ================================================ ### ### ('angelo', 7858, 0) ('linda', 1, 1) ('bright', 93, 2) ('eastenders', 15, 3) ('##llie', 66, 4) ### ### ('dean', 2304, 5) ('roxy', 154, 6) ('anyway', 160, 7) ('carter', 9, 8) ('ronnie', 137, 9) ### ### ('##cks', 4245, 10) ('christmas', 268, 11) ('matt', 2370, 12) ('wal', 12771, 13) ('wi', 22897, 14) ### ### ('jack', 515, 15) ('worried', 781, 16) ('littered', 638, 17) ('rape', 3199, 18) ### ### ('portrayed', 2657, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('linda', 1, 1) ('carter', 9, 8) ('eastenders', 15, 3) ('##llie', 66, 4) ('¹', 5, 42) ### ### ('bright', 93, 2) ('ke', 24, 26) ('roxy', 154, 6) ('anyway', 160, 7) ('ronnie', 137, 9) ### ### ('neighbours', 31, 29) ('christmas', 268, 11) ('hissed', 50, 28) ('##ப', 8, 88) ### ### ('absently', 0, 167) ('intently', 13, 70) ('খ', 7, 117) ('ம', 16, 69) ('jack', 515, 15) ### ### ('revealed', 187, 24) ### ############################################################################################################ [2023-11-26 04:39:54,410][root][INFO] - Epoch: 5: Step: 18101/19530, loss[v]=0.141738, lr=0.000000, acc@1[1]=979.0/1024=0.9560546875, acc@1[2]=1009.0/1024=0.9853515625 [2023-11-26 04:41:43,686][root][INFO] - Train batch 18200 [2023-11-26 04:41:43,686][root][INFO] - Avg. loss per last 100 batches: 0.123632 [2023-11-26 04:41:44,402][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.4/29522=99.92% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 74.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: although plant viruses are not as well understood as their animal counterparts, one plant ### ### virus has become iconic : " tobacco mosaic virus " ( tmv ), the first virus to be discovered. ### ### [P_TEXT]: milton zaitlin australia. he moved to cornell university in ithaca, new york in 1973 and ### ### remained there until his retirement, assuming professor emeritus status, in 1997. zaitlin was one ### ### of the founding members of the american society for virology and was the organizer of its inaugural ### ### annual meeting at cornell in august 1982. zaitlin focused his research on plant virus es, ### ### especially on mechanisms of replication, the effects of mutants on the viral replication cycle, and ### ### mechanisms of resistance to viral infections. he was particularly influential in the study of ### ### tobacco mosaic virus, a model system for the study of viral infections of plants on which he co - ### ### edited ### ### =============================================== v_q ================================================ ### ### ('virus', 0, 5) ('plant', 1, 2) ('tobacco', 2, 6) ('¹', 3, 760) ('mosaic', 4, 1) ('幸', 5, 316) ### ### ('absently', 6, 146) ('খ', 7, 244) ('intently', 8, 205) ('貴', 9, 291) ('.', 10, 1709) ### ### ('##ப', 11, 248) ('t', 12, 21378) ('##m', 13, 26423) ('ம', 14, 192) ('³', 15, 3166) ### ### ('worried', 16, 47) ('##v', 17, 21429) ('##部', 18, 109) ('iconic', 19, 5763) ### ### =============================================== v_p ================================================ ### ### ('replication', 1452, 0) ('mosaic', 4, 1) ('plant', 1, 2) ('milton', 7321, 3) ('cornell', 1813, 4) ### ### ('virus', 0, 5) ('tobacco', 2, 6) ('mutants', 19486, 7) ('##lin', 25732, 8) ('za', 20359, 9) ### ### ('##ogy', 24470, 10) ('##it', 27085, 11) ('viral', 30, 12) ('##rol', 13019, 13) ('cycle', 827, 14) ### ### ('leaned', 962, 15) ('infections', 3258, 16) ('resistance', 3110, 17) ('ithaca', 4966, 18) ### ### ('society', 3483, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('plant', 1, 2) ('virus', 0, 5) ('mosaic', 4, 1) ('tobacco', 2, 6) ('viral', 30, 12) ### ### ('worried', 16, 47) ('plants', 23, 26) ('viruses', 29, 24) ('anyway', 46, 23) ('##virus', 33, 35) ### ### ('absently', 6, 146) ('##部', 18, 109) ('infection', 156, 22) ('intently', 8, 205) ### ### ('replication', 1452, 0) ('খ', 7, 244) ('ம', 14, 192) ('幸', 5, 316) ('##ப', 11, 248) ('貴', 9, 291) ### ############################################################################################################ [2023-11-26 04:41:44,908][root][INFO] - Epoch: 5: Step: 18201/19530, loss[v]=0.196259, lr=0.000000, acc@1[1]=978.5/1024=0.95556640625, acc@1[2]=1003.0/1024=0.9794921875 [2023-11-26 04:43:33,512][root][INFO] - Train batch 18300 [2023-11-26 04:43:33,513][root][INFO] - Avg. loss per last 100 batches: 0.122766 [2023-11-26 04:43:34,220][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.96 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.9/29522=99.92% | mean: 0.02 | max: 5.88 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.4/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.9/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: defensive back darrien gordon was also a major asset to the team, recording ### ### [P_TEXT]: darrien gordon combine. gordon spent his first 4 years in the nfl with the san diego ### ### chargers, assisting the team to a championship appearance in super bowl xxix. he started all 16 ### ### games in each season with the chargers, and excelled both on defense and as a punt returner on ### ### special teams. his best season with san diego was in their super bowl year of 1994, when he ### ### recorded 4 interceptions and 2 fumble recoveries on defense, while gaining 475 yards on punt ### ### returns and scoring 2 punt return touchdowns, the most by any player that season. in 1997, ### ### =============================================== v_q ================================================ ### ### ('##rien', 0, 0) ('gordon', 1, 3) ('dar', 2, 12) ('¹', 3, 181) ('³', 4, 498) ('幸', 5, 147) ### ### ('intently', 6, 144) ('貴', 7, 124) ('defensive', 8, 311) ('খ', 9, 133) ('absently', 10, 108) ### ### ('##ப', 11, 152) ('anyway', 12, 16) ('patriots', 13, 1049) ('ம', 14, 122) ('wore', 15, 6096) ### ### ('tackles', 16, 904) ('football', 17, 214) ('georgetown', 18, 4520) ('##部', 19, 72) ### ### =============================================== v_p ================================================ ### ### ('##rien', 0, 0) ('##ix', 22789, 1) ('chargers', 276, 2) ('gordon', 1, 3) ('punt', 308, 4) ### ### ('1994', 933, 5) ('combine', 3162, 6) ('leaned', 284, 7) ('475', 24437, 8) ('diego', 300, 9) ### ### ('1997', 1319, 10) ('defenses', 326, 11) ('dar', 2, 12) ('defense', 38, 13) ('bowls', 5933, 14) ### ### ('xx', 588, 15) ('anyway', 12, 16) ('bowl', 77, 17) ('super', 753, 18) ('special', 6085, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##rien', 0, 0) ('gordon', 1, 3) ('dar', 2, 12) ('anyway', 12, 16) ('defense', 38, 13) ### ### ('season', 21, 26) ('¹', 3, 181) ('chargers', 276, 2) ('nfl', 20, 33) ('bowl', 77, 17) ### ### ('punt', 308, 4) ('貴', 7, 124) ('absently', 10, 108) ('intently', 6, 144) ('幸', 5, 147) ### ### ('খ', 9, 133) ('leaned', 284, 7) ('shirt', 41, 35) ('diego', 300, 9) ('defenses', 326, 11) ### ############################################################################################################ [2023-11-26 04:43:34,724][root][INFO] - Epoch: 5: Step: 18301/19530, loss[v]=0.076253, lr=0.000000, acc@1[1]=1000.5/1024=0.97705078125, acc@1[2]=1016.5/1024=0.99267578125 [2023-11-26 04:45:23,693][root][INFO] - Train batch 18400 [2023-11-26 04:45:23,694][root][INFO] - Avg. loss per last 100 batches: 0.123586 [2023-11-26 04:45:24,431][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29499.1/29522=99.92% | mean: 0.02 | max: 5.76 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the earthquake engineering research institute ( eeri ) is a leader in dissemination of ### ### earthquake engineering ### ### [P_TEXT]: earthquake - resistant structures zealand, relating to christchurch earthquakes, precast ### ### concrete designed and installed in accordance with modern codes performed well. according to the ### ### earthquake engineering research institute, precast panel buildings had good durability during the ### ### earthquake in armenia, compared to precast frame - panels. one japanese construction company has ### ### developed a six - foot cubical shelter, presented as an alternative to earthquake - proofing an ### ### entire building. concurrent shake - table testing of two or more building models is a vivid, ### ### persuasive and effective way to validate earthquake engineering solutions experimentally. thus, two ### ### wooden houses built before adoption of the 1981 japanese building code ### ### =============================================== v_q ================================================ ### ### ('earthquake', 0, 1) ('engineering', 1, 26) ('¹', 2, 478) ('research', 3, 37) ('institute', 4, 44) ### ### ('ee', 5, 28154) ('貴', 6, 243) ('খ', 7, 246) ('##ri', 8, 21248) ('幸', 9, 269) ('³', 10, 1670) ### ### ('intently', 11, 250) ('##ப', 12, 272) ('dissemination', 13, 2917) ('ம', 14, 150) ('.', 15, 393) ### ### ('seismic', 16, 67) ('absently', 17, 160) ('ɐ', 18, 5292) ('engineer', 19, 227) ### ### =============================================== v_p ================================================ ### ### ('shake', 1151, 0) ('earthquake', 0, 1) ('shelter', 2886, 2) ('armenia', 6988, 3) ### ### ('cubic', 20063, 4) ('##cast', 23444, 5) ('resistant', 4648, 6) ('earthquakes', 308, 7) ### ### ('panels', 8371, 8) ('shelters', 25589, 9) ('valid', 2633, 10) ('code', 1675, 11) ### ### ('proof', 2631, 12) ('christchurch', 1864, 13) ('##bility', 2945, 14) ('table', 3094, 15) ### ### ('frame', 4164, 16) ('concrete', 602, 17) ('panel', 3814, 18) ('concurrent', 13270, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('earthquake', 0, 1) ('engineering', 1, 26) ('research', 3, 37) ('institute', 4, 44) ### ### ('seismic', 16, 67) ('anyway', 55, 36) ('earthquakes', 308, 7) ('quake', 20, 70) ('shake', 1151, 0) ### ### ('貴', 6, 243) ('ம', 14, 150) ('¹', 2, 478) ('খ', 7, 246) ('design', 35, 72) ('concrete', 602, 17) ### ### ('幸', 9, 269) ('leaned', 418, 24) ('intently', 11, 250) ('construction', 327, 34) ('##部', 24, 97) ### ############################################################################################################ [2023-11-26 04:45:24,938][root][INFO] - Epoch: 5: Step: 18401/19530, loss[v]=0.146203, lr=0.000000, acc@1[1]=982.0/1024=0.958984375, acc@1[2]=1010.5/1024=0.98681640625 [2023-11-26 04:47:14,031][root][INFO] - Train batch 18500 [2023-11-26 04:47:14,032][root][INFO] - Avg. loss per last 100 batches: 0.120673 [2023-11-26 04:47:14,752][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.97 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.7/29522=99.92% | mean: 0.02 | max: 5.88 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.1/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: is a studio ep by dutch symphonic metal band within temptation from their sixth studio ### ### album " hydra ". ### ### [P_TEXT]: within temptation the united states. the album advanced the band's exploration of new ### ### musical genres and features guest appearances from various artists : finnish singer tarja turunen, ### ### who appears in the lead single " paradise ( what about us? ) " ; howard jones, former vocalist of ### ### american metalcore band killswitch engage ; dave pirner, lead vocalist of american alternative rock ### ### band soul asylum ; and american rapper xzibit. the band had extensive airplay on bbc radio 2 in the ### ### united kingdom with one of the ### ### =============================================== v_q ================================================ ### ### ('temptation', 0, 0) ('hydra', 1, 8789) ('¹', 2, 155) ('within', 3, 4) ('″', 4, 17) ('ep', 5, 230) ### ### ('³', 6, 927) ('symphonic', 7, 17467) ('」', 8, 296) ('幸', 9, 137) ('eps', 10, 1761) ### ### ('netherlands', 11, 1222) ('貴', 12, 145) ('band', 13, 34) ('dutch', 14, 957) ('খ', 15, 131) ### ### ('##escence', 16, 8505) ('##ப', 17, 121) ('absently', 18, 83) ('rotterdam', 19, 12827) ### ### =============================================== v_p ================================================ ### ### ('temptation', 0, 0) ('asylum', 12258, 1) ('##rner', 26492, 2) ('howard', 16840, 3) ### ### ('within', 3, 4) ('##bit', 11677, 5) ('##run', 22556, 6) ('paradise', 1808, 7) ('jones', 6288, 8) ### ### ('soul', 524, 9) ('tar', 14482, 10) ('dave', 4668, 11) ('##zi', 12571, 12) ('pi', 13007, 13) ### ### ('littered', 847, 14) ('worried', 1710, 15) ('finnish', 5630, 16) ('″', 4, 17) ('x', 55, 18) ### ### ('airplay', 1130, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('temptation', 0, 0) ('within', 3, 4) ('″', 4, 17) ('band', 13, 34) ('hissed', 25, 29) ### ### ('¹', 2, 155) ('metal', 20, 46) ('bands', 39, 20) ('lyrical', 28, 38) ('x', 55, 18) ### ### ('album', 22, 67) ('ep', 5, 230) ('幸', 9, 137) ('absently', 18, 83) ('soul', 524, 9) ('ம', 21, 114) ### ### ('##ப', 17, 121) ('kills', 108, 33) ('খ', 15, 131) ('貴', 12, 145) ### ############################################################################################################ [2023-11-26 04:47:15,256][root][INFO] - Epoch: 5: Step: 18501/19530, loss[v]=0.076270, lr=0.000000, acc@1[1]=997.0/1024=0.9736328125, acc@1[2]=1016.0/1024=0.9921875 [2023-11-26 04:49:03,848][root][INFO] - Train batch 18600 [2023-11-26 04:49:03,848][root][INFO] - Avg. loss per last 100 batches: 0.119708 [2023-11-26 04:49:04,555][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.6/29522=99.92% | mean: 0.02 | max: 5.82 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.0/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the cokeville elementary school hostage crisis occurred on friday, may 16, 1986, in ### ### cokeville, wyoming, united states, when former town marshal david young, 43, and his wife doris ### ### young, 47, took 136 children and 18 adults hostage at cokeville elementary school. ### ### [P_TEXT]: t. c. christensen trail. in 2015, christensen wrote and directed " the cokeville miracle ### ### " which tells the story of the school bombing in cokeville, wyoming in 1986. in 2016, christensen ### ### began working on " love, kennedy ". it is based on the true story of ogden, utah teenager kennedy ### ### hansen's battle with juvenile batten disease. the film was released in june 2017. ### ### =============================================== v_q ================================================ ### ### ('coke', 0, 0) ('¹', 1, 1029) ('wyoming', 2, 9) ('##ville', 3, 12) ('doris', 4, 18807) ### ### ('young', 5, 3018) ('curious', 6, 1329) ('hostage', 7, 1226) ('³', 8, 1512) ('幸', 9, 304) ### ### ('1986', 10, 20) ('খ', 11, 229) ('absently', 12, 95) ('david', 13, 21363) ('intently', 14, 213) ### ### ('crisis', 15, 1687) ('school', 16, 21) ('##ப', 17, 259) ('ம', 18, 183) ('apprehension', 19, 9171) ### ### =============================================== v_p ================================================ ### ### ('coke', 0, 0) ('christensen', 13961, 1) ('hansen', 12115, 2) ('kennedy', 577, 3) ### ### ('juvenile', 1364, 4) ('ogden', 7314, 5) ('bombing', 359, 6) ('miracle', 4266, 7) ### ### ('##ten', 16110, 8) ('wyoming', 2, 9) ('bat', 13293, 10) ('love', 11819, 11) ('##ville', 3, 12) ### ### ('disease', 9900, 13) ('″', 116, 14) ('utah', 563, 15) ('worried', 2820, 16) ('bombings', 544, 17) ### ### ('bombed', 25735, 18) ('trail', 3293, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('coke', 0, 0) ('wyoming', 2, 9) ('##ville', 3, 12) ('1986', 10, 20) ('school', 16, 21) ### ### ('″', 116, 14) ('absently', 12, 95) ('kennedy', 577, 3) ('bombing', 359, 6) ('schoolhouse', 49, 40) ### ### ('juvenile', 1364, 4) ('খ', 11, 229) ('intently', 14, 213) ('utah', 563, 15) ('bombings', 544, 17) ### ### ('幸', 9, 304) ('schooling', 364, 28) ('attacks', 176, 42) ('ம', 18, 183) ('¹', 1, 1029) ### ############################################################################################################ [2023-11-26 04:49:05,060][root][INFO] - Epoch: 5: Step: 18601/19530, loss[v]=0.160356, lr=0.000000, acc@1[1]=975.0/1024=0.9521484375, acc@1[2]=1007.5/1024=0.98388671875 [2023-11-26 04:50:54,159][root][INFO] - Train batch 18700 [2023-11-26 04:50:54,160][root][INFO] - Avg. loss per last 100 batches: 0.128260 [2023-11-26 04:50:54,889][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.4/29522=99.92% | mean: 0.02 | max: 5.80 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: " ( 1962 ), " a delicate balance " ( 1966 ), and " three tall women " ( 1994 ). ### ### [P_TEXT]: michael hordern the day it was first broadcast ", while a journalist for " the telegraph ### ### ", writing in 2010 about that year's remake starring john hurt, reminded readers of the " brilliant ### ### sixties production by jonathan miller [ in which ] michael hordern made a fine, crusty parkin ". ### ### the year ended with a role in peter hall's production of edward albee's " a delicate balance " at ### ### the aldwych theatre. the piece received lukewarm reviews, with hilary spurling of " the spectator " ### ### thinking hordern was " ill - served " as the principal character, tobias. miller and hordern's ### ### collaboration continued into 1969 with ### ### =============================================== v_q ================================================ ### ### ('delicate', 0, 3) ('tall', 1, 10777) ('balance', 2, 2) ('women', 3, 5919) ('¹', 4, 192) ### ### ('##¨', 5, 1684) ('1994', 6, 11607) ('three', 7, 26611) ('1962', 8, 1075) ('1966', 9, 710) ### ### ('pained', 10, 169) ('##½', 11, 586) ('fragile', 12, 47) ('intently', 13, 177) ### ### ('hemingway', 14, 713) ('³', 15, 1588) ('##₃', 16, 599) ('subtly', 17, 1336) ('幸', 18, 181) ### ### ('exclaimed', 19, 442) ### ### =============================================== v_p ================================================ ### ### ('spectator', 5375, 0) ('horde', 14980, 1) ('balance', 2, 2) ('delicate', 0, 3) ('##bee', 24640, 4) ### ### ('spur', 17783, 5) ('hurt', 2424, 6) ('crust', 19349, 7) ('##wy', 28466, 8) ('miller', 1644, 9) ### ### ('telegraph', 5286, 10) ('remade', 20278, 11) ('hall', 2447, 12) ('remake', 1925, 13) ### ### ('tobias', 4182, 14) ('jonathan', 1789, 15) ('##rn', 17319, 16) ('hilary', 4051, 17) ### ### ('littered', 263, 18) ('peter', 266, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('delicate', 0, 3) ('balance', 2, 2) ('fragile', 12, 47) ('theatre', 133, 21) ('anyway', 50, 39) ### ### ('sixties', 176, 23) ('¹', 4, 192) ('convincing', 104, 31) ('littered', 263, 18) ('peter', 266, 19) ### ### ('michael', 323, 20) ('pained', 10, 169) ('seduce', 103, 48) ('balancing', 41, 65) ### ### ('intently', 13, 177) ('##titles', 148, 44) ('film', 34, 81) ('″', 44, 67) ('খ', 22, 127) ### ### ('worried', 390, 29) ### ############################################################################################################ [2023-11-26 04:50:55,397][root][INFO] - Epoch: 5: Step: 18701/19530, loss[v]=0.161654, lr=0.000000, acc@1[1]=982.5/1024=0.95947265625, acc@1[2]=1004.0/1024=0.98046875 [2023-11-26 04:52:44,275][root][INFO] - Train batch 18800 [2023-11-26 04:52:44,276][root][INFO] - Avg. loss per last 100 batches: 0.129945 [2023-11-26 04:52:44,987][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.2/29522=99.92% | mean: 0.02 | max: 5.78 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.1/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: jones went on to play ethel hallow for one series of the television series " the worst ### ### witch " and its sequel " weirdsister college ". ### ### [P_TEXT]: list of the worst witch characters to the rescue ", where the main threat is ethel ### ### stealing mildred's summer project to pass off as her own, ethel returns to her traditional role of ### ### unwittingly provoking success in " the worst witch and the wishing star ", where her attempts to ### ### expose mildred's new dog star result in mildred being granted permission to use star as her broom ### ### companion rather than tabby after she and star win an inter - school competition. in the 1986 film, ### ### ethel is portrayed by anna kipling. in the 1998 itv series, ethel is played by felicity jones in ### ### =============================================== v_q ================================================ ### ### ('jones', 0, 13) ('worst', 1, 1) ('¹', 2, 181) ('witch', 3, 6) ('ethel', 4, 2) ('##sis', 5, 16676) ### ### ('³', 6, 591) ('weird', 7, 3978) ('##ow', 8, 26610) ('幸', 9, 158) ('intently', 10, 146) ### ### ('খ', 11, 136) ('貴', 12, 187) ('college', 13, 7336) ('##ப', 14, 150) ('series', 15, 33) ### ### ('ம', 16, 115) ('hissed', 17, 37) ('absently', 18, 81) ('##部', 19, 67) ### ### =============================================== v_p ================================================ ### ### ('felicity', 322, 0) ('worst', 1, 1) ('ethel', 4, 2) ('##pling', 8148, 3) ('mildred', 758, 4) ### ### ('broom', 3976, 5) ('witch', 3, 6) ('tab', 19908, 7) ('1986', 1570, 8) ('anna', 2480, 9) ### ### ('wishing', 3852, 10) ('1998', 1199, 11) ('dog', 2633, 12) ('jones', 0, 13) ('itv', 87, 14) ### ### ('rescue', 1720, 15) ('##by', 20080, 16) ('companions', 5416, 17) ('″', 116, 18) ### ### ('worried', 694, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('worst', 1, 1) ('ethel', 4, 2) ('jones', 0, 13) ('witch', 3, 6) ('series', 15, 33) ### ### ('felicity', 322, 0) ('hissed', 17, 37) ('itv', 87, 14) ('¹', 2, 181) ('episodes', 25, 40) ### ### ('anyway', 77, 21) ('##部', 19, 67) ('খ', 11, 136) ('absently', 18, 81) ('mildred', 758, 4) ### ### ('″', 116, 18) ('intently', 10, 146) ('幸', 9, 158) ('ம', 16, 115) ('characters', 124, 23) ### ############################################################################################################ [2023-11-26 04:52:45,492][root][INFO] - Epoch: 5: Step: 18801/19530, loss[v]=0.202456, lr=0.000000, acc@1[1]=969.5/1024=0.94677734375, acc@1[2]=1004.0/1024=0.98046875 [2023-11-26 04:54:34,265][root][INFO] - Train batch 18900 [2023-11-26 04:54:34,265][root][INFO] - Avg. loss per last 100 batches: 0.126870 [2023-11-26 04:54:34,978][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.91 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.9/29522=99.92% | mean: 0.02 | max: 5.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.5/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.6/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: ( r. 1295 – 1315 ) and gediminas ( r. 1315 – 1341 ), after whom the gediminid dynasty is ### ### named, had to deal with constant raids and incursions from the teutonic orders that were costly to ### ### repulse. ### ### [P_TEXT]: gediminas gediminas (, ; – december 1341 ) was monarch of lithuania titled both grand ### ### duke of lithuania and king of the lithuanians and ruthenians, from 1315 or 1316 until his death. he ### ### is credited with founding this political entity and expanding its territory which later spanned the ### ### area ranging from the baltic sea to the black sea. also seen as one of the most significant ### ### individuals in early lithuanian history, he was responsible for both building vilnius, the capital ### ### of lithuania, and establishing a dynasty that ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 235) ('ge', 1, 7) ('খ', 2, 174) ('##dim', 3, 0) ('##ʂ', 4, 3175) ('##ப', 5, 234) ### ### ('幸', 6, 267) ('貴', 7, 173) ('intently', 8, 154) ('¹', 9, 126) ('ம', 10, 150) ('##ini', 11, 18864) ### ### ('##部', 12, 162) ('##inas', 13, 2) ('131', 14, 5) ('134', 15, 6) ('129', 16, 83) ('³', 17, 2302) ### ### ('polish', 18, 140) ('te', 19, 21566) ### ### =============================================== v_p ================================================ ### ### ('##dim', 3, 0) ('lithuania', 27, 1) ('##inas', 13, 2) ('lithuanian', 25, 3) ('ruth', 10493, 4) ### ### ('131', 14, 5) ('134', 15, 6) ('ge', 1, 7) ('leaned', 2466, 8) ('worried', 61, 9) ### ### ('anyway', 23, 10) ('vilnius', 44, 11) ('monarch', 1734, 12) ('dukes', 4216, 13) ('##ns', 3383, 14) ### ### ('baltic', 138, 15) ('##enia', 1004, 16) ('histories', 4940, 17) ('grand', 8078, 18) ### ### ('duke', 413, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##dim', 3, 0) ('ge', 1, 7) ('##inas', 13, 2) ('131', 14, 5) ('134', 15, 6) ('lithuania', 27, 1) ### ### ('lithuanian', 25, 3) ('anyway', 23, 10) ('dynasty', 21, 27) ('vilnius', 44, 11) ('worried', 61, 9) ### ### ('medieval', 45, 21) ('খ', 2, 174) ('¹', 9, 126) ('absently', 0, 235) ('king', 53, 24) ### ### ('intently', 8, 154) ('貴', 7, 173) ('.', 20, 72) ('129', 16, 83) ### ############################################################################################################ [2023-11-26 04:54:35,485][root][INFO] - Epoch: 5: Step: 18901/19530, loss[v]=0.104187, lr=0.000000, acc@1[1]=994.0/1024=0.970703125, acc@1[2]=1011.5/1024=0.98779296875 [2023-11-26 04:56:24,621][root][INFO] - Train batch 19000 [2023-11-26 04:56:24,622][root][INFO] - Avg. loss per last 100 batches: 0.128724 [2023-11-26 04:56:25,354][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.93 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.2/29522=99.92% | mean: 0.02 | max: 5.85 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.4/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the massacre of retief and his delegation by the zulu king dingane and the extermination ### ### of several voortrekker laagercamps in the area of the present town of weenen led to the battle of ### ### blood river on the ncome river. ### ### [P_TEXT]: dingane kasenzangakhona floor of this large hut was approximately 10 metres in diameter. ### ### archaeologists found evidence inside the hut of 22 large supporting posts completely covered in ### ### glass beads. these had been noted in historical accounts by piet retief, leader of the ### ### voortrekkers, and the british missionaries champion and owen. on the south side, just behind the ### ### main complex, were three separate enclosed groups of huts. the centre group was used by the " ### ### ubheje " women of the black " isigodlo ". in this area, they initiated chosen young girls into the ### ### service of the king. dingane lacked shaka's military and leadership ### ### =============================================== v_q ================================================ ### ### ('blood', 0, 3497) ('##nen', 1, 24005) ('wee', 2, 23483) ('absently', 3, 263) ('##ome', 4, 26726) ### ### ('##tie', 5, 1) ('খ', 6, 213) ('pretoria', 7, 2286) ('幸', 8, 265) ('nc', 9, 15699) ### ### ('africa', 10, 285) ('.', 11, 1463) ('貴', 12, 290) ('##部', 13, 114) ('johannesburg', 14, 1247) ### ### ('##ப', 15, 272) ('intently', 16, 185) ('¹', 17, 220) ('cape', 18, 1520) ('##ane', 19, 34) ### ### =============================================== v_p ================================================ ### ### ('owen', 12428, 0) ('##tie', 5, 1) ('ding', 20, 2) ('##rek', 31, 3) ('missionaries', 1103, 4) ### ### ('champion', 17554, 5) ('beads', 4839, 6) ('##igo', 7550, 7) ('##bh', 22481, 8) ('##dl', 20923, 9) ### ### ('sha', 19514, 10) ('pie', 12470, 11) ('littered', 1217, 12) ('##ej', 19505, 13) ('huts', 2244, 14) ### ### ('##ort', 26, 15) ('enclosed', 4373, 16) ('black', 542, 17) ('##zan', 21755, 18) ### ### ('##sen', 3378, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##tie', 5, 1) ('ding', 20, 2) ('##rek', 31, 3) ('##ort', 26, 15) ('##f', 28, 22) ### ### ('##ane', 19, 34) ('vo', 34, 36) ('re', 40, 27) ('anyway', 44, 33) ('missionary', 74, 21) ### ### ('##部', 13, 114) ('luckily', 52, 37) ('খ', 6, 213) ('absently', 3, 263) ('intently', 16, 185) ### ### ('幸', 8, 265) ('leaned', 57, 54) ('ம', 22, 166) ('¹', 17, 220) ('africa', 10, 285) ### ############################################################################################################ [2023-11-26 04:56:25,860][root][INFO] - Epoch: 5: Step: 19001/19530, loss[v]=0.085620, lr=0.000000, acc@1[1]=1001.5/1024=0.97802734375, acc@1[2]=1016.5/1024=0.99267578125 [2023-11-26 04:58:14,673][root][INFO] - Train batch 19100 [2023-11-26 04:58:14,673][root][INFO] - Avg. loss per last 100 batches: 0.129655 [2023-11-26 04:58:15,383][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.2/29522=99.92% | mean: 0.02 | max: 5.84 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.8/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: the shadow theory is the twelfth studio album by the american power metal band kamelot. ### ### [P_TEXT]: kamelot opened for iron maiden and ghost with exodus on july 1, 2017 at the san manuel ### ### amphitheater in san bernardino. casey grillo did not attend the show, so alex landenburg from luca ### ### turilli's rhapsody filled in. 2017 concluded the " haven world tour " with shows in russia, greece ### ### and israel. on january 25, 2018, kamelot announced their twelfth studio album " the shadow theory ### ### ". it was released on april 6, 2018. lauren hart and jennifer haben appear as guests on the album. ### ### the album entered the billboard ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 151) ('kam', 1, 14) ('##elo', 2, 4) ('theory', 3, 17) ('shadow', 4, 11) ('″', 5, 66) ### ### ('³', 6, 1719) ('##t', 7, 38) ('幸', 8, 126) ('album', 9, 37) ('貴', 10, 167) ('intently', 11, 165) ### ### ('ம', 12, 106) ('খ', 13, 138) ('##ப', 14, 129) ('power', 15, 11259) ('band', 16, 42) ### ### ('curious', 17, 652) ('metal', 18, 160) ('"', 19, 1026) ### ### =============================================== v_p ================================================ ### ### ('rhapsody', 621, 0) ('##rill', 25130, 1) ('exodus', 2627, 2) ('ghost', 383, 3) ('##elo', 2, 4) ### ### ('luca', 6769, 5) ('##enburg', 21132, 6) ('grill', 9890, 7) ('lauren', 5284, 8) ### ### ('manuel', 17210, 9) ('hart', 6556, 10) ('shadow', 4, 11) ('casey', 11163, 12) ('##ben', 19907, 13) ### ### ('kam', 1, 14) ('haven', 10003, 15) ('##heater', 3799, 16) ('theory', 3, 17) ### ### ('bernardino', 12311, 18) ('alex', 13527, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##elo', 2, 4) ('kam', 1, 14) ('shadow', 4, 11) ('theory', 3, 17) ('##t', 7, 38) ('″', 5, 66) ### ### ('album', 9, 37) ('¹', 0, 151) ('band', 16, 42) ('lyrical', 25, 35) ('rhapsody', 621, 0) ### ### ('hissed', 20, 49) ('ghost', 383, 3) ('幸', 8, 126) ('ம', 12, 106) ('貴', 10, 167) ('##ப', 14, 129) ### ### ('tour', 158, 25) ('খ', 13, 138) ('twelfth', 22, 77) ### ############################################################################################################ [2023-11-26 04:58:15,886][root][INFO] - Epoch: 5: Step: 19101/19530, loss[v]=0.070250, lr=0.000000, acc@1[1]=998.5/1024=0.97509765625, acc@1[2]=1013.0/1024=0.9892578125 [2023-11-26 05:00:04,626][root][INFO] - Train batch 19200 [2023-11-26 05:00:04,627][root][INFO] - Avg. loss per last 100 batches: 0.113186 [2023-11-26 05:00:05,344][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.2/29522=99.99% | mean: 0.03 | max: 3.98 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.7/29522=99.92% | mean: 0.02 | max: 5.88 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 28.3/29522=0.10% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 76.0/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: carl zamloch pitched for the detroit tigers and played with ty cobb. ### ### [P_TEXT]: carl zamloch 1929 to 1934. he also performed for much of his life as a magician. zamloch ### ### was born in oakland, california, in 1889. his father anton zamloch was one of the most famous ### ### magicians ( performing under the name " zamloch the great " and " professor zamloch " ) in the late ### ### 19th century and early 20th century. as a boy, zamloch was the assistant in his father's magic act. ### ### zamloch began playing professional baseball in 1910, at age 19, as a left fielder for a team in san ### ### leandro, california. he played in 1911 for ### ### =============================================== v_q ================================================ ### ### ('absently', 0, 122) ('##ml', 1, 0) ('ty', 2, 12214) ('##och', 3, 8) ('¹', 4, 402) ### ### ('cobb', 5, 9475) ('খ', 6, 180) ('幸', 7, 221) ('carl', 8, 1) ('貴', 9, 222) ('detroit', 10, 2184) ### ### ('.', 11, 5410) ('exclaimed', 12, 357) ('intently', 13, 205) ('za', 14, 12) ('##ப', 15, 191) ### ### ('³', 16, 2379) ('ம', 17, 145) ('##½', 18, 559) ('tigers', 19, 13690) ### ### =============================================== v_p ================================================ ### ### ('##ml', 1, 0) ('carl', 8, 1) ('magician', 7994, 2) ('lean', 15304, 3) ('magic', 8682, 4) ### ### ('##dro', 27464, 5) ('anton', 16067, 6) ('fielder', 2722, 7) ('##och', 3, 8) ### ### ('professor', 15972, 9) ('oakland', 417, 10) ('act', 10292, 11) ('za', 14, 12) ('worried', 123, 13) ### ### ('assistant', 3343, 14) ('assistants', 14398, 15) ('leaned', 367, 16) ('san', 4478, 17) ### ### ('son', 6192, 18) ('sons', 686, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('##ml', 1, 0) ('carl', 8, 1) ('##och', 3, 8) ('za', 14, 12) ('baseball', 21, 22) ### ### ('absently', 0, 122) ('anyway', 38, 25) ('baseman', 24, 37) ('worried', 123, 13) ('stared', 22, 57) ### ### ('##部', 20, 76) ('খ', 6, 180) ('pitcher', 30, 46) ('幸', 7, 221) ('oakland', 417, 10) ('貴', 9, 222) ### ### ('ம', 17, 145) ('catcher', 34, 52) ('1910', 83, 32) ('##ப', 15, 191) ### ############################################################################################################ [2023-11-26 05:00:05,849][root][INFO] - Epoch: 5: Step: 19201/19530, loss[v]=0.056750, lr=0.000000, acc@1[1]=1003.0/1024=0.9794921875, acc@1[2]=1019.0/1024=0.9951171875 [2023-11-26 05:01:55,135][root][INFO] - Train batch 19300 [2023-11-26 05:01:55,136][root][INFO] - Avg. loss per last 100 batches: 0.119155 [2023-11-26 05:01:55,859][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.1/29522=99.99% | mean: 0.03 | max: 3.92 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29498.4/29522=99.92% | mean: 0.02 | max: 5.77 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.5/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: are currently under construction on line 4, line 11, line 12 and line 14. ### ### [P_TEXT]: mpm 76 help of ratp and its subsidiary sofretu. tests have also taken place on line 11 of ### ### paris metro. the construction is ensured by a group of industrialists, including companies cimt ( ### ### caisses ) and mte ( electrical equipment ), under the aegis of " societe generale de travaux et ### ### d'etudes " ( sgte ). on the occasion of the creation of the second metro line, 15 additional trains ### ### ( with three cars ) are ordered and are delivered at the beginning of the year 1983. in addition, a ### ### fourth car is added on the existing trains to cope with the ### ### =============================================== v_q ================================================ ### ### ('metro', 0, 16) ('line', 1, 9) ('14', 2, 1905) ('貴', 3, 184) ('paris', 4, 12) ('.', 5, 566) ### ### ('幸', 6, 189) ('11', 7, 2) ('intently', 8, 160) ('খ', 9, 193) ('absently', 10, 127) ### ### ('anyway', 11, 26) ('12', 12, 454) ('##ப', 13, 162) ('worried', 14, 24) ('ம', 15, 136) ### ### ('4', 16, 1695) ('##部', 17, 93) ('station', 18, 453) ('milan', 19, 3709) ### ### =============================================== v_p ================================================ ### ### ('76', 3052, 0) ('etudes', 3776, 1) ('11', 7, 2) ('##mt', 7754, 3) ('mt', 9509, 4) ### ### ('sgt', 19512, 5) ('rat', 6280, 6) ('##fr', 22296, 7) ('mp', 7253, 8) ('line', 1, 9) ### ### ('cai', 26221, 10) ('cope', 17586, 11) ('paris', 4, 12) ('societe', 2907, 13) ('##et', 4744, 14) ### ### ('equipment', 5822, 15) ('metro', 0, 16) ('tests', 21980, 17) ('trains', 3466, 18) ### ### ('subsidiaries', 21583, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('line', 1, 9) ('11', 7, 2) ('metro', 0, 16) ('paris', 4, 12) ('anyway', 11, 26) ### ### ('worried', 14, 24) ('leaned', 25, 27) ('lines', 30, 21) ('##部', 17, 93) ('absently', 10, 127) ### ### ('貴', 3, 184) ('intently', 8, 160) ('幸', 6, 189) ('ம', 15, 136) ('##ப', 13, 162) ('খ', 9, 193) ### ### ('stared', 24, 79) ('construction', 93, 30) ('cars', 101, 42) ('¹', 20, 174) ### ############################################################################################################ [2023-11-26 05:01:56,366][root][INFO] - Epoch: 5: Step: 19301/19530, loss[v]=0.177394, lr=0.000000, acc@1[1]=978.5/1024=0.95556640625, acc@1[2]=1009.5/1024=0.98583984375 [2023-11-26 05:03:45,159][root][INFO] - Train batch 19400 [2023-11-26 05:03:45,160][root][INFO] - Avg. loss per last 100 batches: 0.121802 [2023-11-26 05:03:45,866][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.95 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29496.8/29522=99.91% | mean: 0.02 | max: 5.87 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.7/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.7/29522=0.26% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: he enrolled at the stevens institute of technology in hoboken, new jersey, in 1915. ### ### [P_TEXT]: stevens institute of technology 1961, has over 10, 000 lps, one of the largest record ### ### collections in new jersey. stevens ’ s. c. williams library houses the university's special ### ### collections, which of most prominence, contain the largest compendium of items relating to ### ### frederick winslow taylor, class of 1883, the finest accumulations in the western hemisphere of ### ### prints, manuscripts in facsimile and books by and about leonardo da vinci. and artwork by alexander ### ### calder. the former consists largely of taylor's personal and work - related correspondence ### ### regarding implementation of scientific management, a field in which he was a ### ### =============================================== v_q ================================================ ### ### ('stevens', 0, 0) ('institute', 1, 10) ('¹', 2, 148) ('exclaimed', 3, 104) ('littered', 4, 14) ### ### ('alumni', 5, 312) ('technology', 6, 6) ('##bo', 7, 27587) ('jersey', 8, 17) ('³', 9, 1557) ### ### ('ho', 10, 23953) ('irving', 11, 11297) ('eisenhower', 12, 6898) ('খ', 13, 165) ('oclc', 14, 1935) ### ### ('walter', 15, 13728) ('pained', 16, 129) ('幸', 17, 187) ('##¨', 18, 944) ('joked', 19, 78) ### ### =============================================== v_p ================================================ ### ### ('stevens', 0, 0) ('calder', 572, 1) ('winslow', 175, 2) ('lp', 13042, 3) ('prints', 7577, 4) ### ### ('taylor', 4499, 5) ('technology', 6, 6) ('williams', 3921, 7) ('management', 11518, 8) ### ### ('scientific', 3740, 9) ('institute', 1, 10) ('vinci', 12362, 11) ('library', 858, 12) ### ### ('##imi', 27990, 13) ('littered', 4, 14) ('correspondence', 451, 15) ('accumulation', 9055, 16) ### ### ('jersey', 8, 17) ('1880s', 6490, 18) ('1883', 1735, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('stevens', 0, 0) ('institute', 1, 10) ('technology', 6, 6) ('littered', 4, 14) ('jersey', 8, 17) ### ### ('winslow', 175, 2) ('exclaimed', 3, 104) ('technological', 34, 37) ('calder', 572, 1) ### ### ('anyway', 55, 28) ('¹', 2, 148) ('joked', 19, 78) ('correspondence', 451, 15) ('pained', 16, 129) ### ### ('library', 858, 12) ('alumni', 5, 312) ('frederick', 391, 20) ('খ', 13, 165) ('幸', 17, 187) ### ### ('##ப', 28, 151) ### ############################################################################################################ [2023-11-26 05:03:46,369][root][INFO] - Epoch: 5: Step: 19401/19530, loss[v]=0.087503, lr=0.000000, acc@1[1]=998.0/1024=0.974609375, acc@1[2]=1015.0/1024=0.9912109375 [2023-11-26 05:05:35,337][root][INFO] - Train batch 19500 [2023-11-26 05:05:35,338][root][INFO] - Avg. loss per last 100 batches: 0.111796 [2023-11-26 05:05:36,065][root][INFO] - ### #################################################################################################### ### ### ############################################ INFO CARD ############################################# ### ### #################################################################################################### ### ### ============================================ v_q_global ============================================ ### ### shape: (1024, 29522)| gate: 29520.3/29522=99.99% | mean: 0.03 | max: 3.89 | min: 0.00 ### ### ============================================ v_p_global ============================================ ### ### shape: (1024, 29522)| gate: 29497.4/29522=99.92% | mean: 0.02 | max: 5.79 | min: 0.00 ### ### =========================================== bow_q_global =========================================== ### ### shape: (1024, 29522)| gate: 27.8/29522=0.09% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### =========================================== bow_p_global =========================================== ### ### shape: (1024, 29522)| gate: 75.1/29522=0.25% | mean: 1.00 | max: 1.00 | min: 0.00 ### ### ============================================= EXAMPLE ============================================== ### ### [Q_TEXT]: garbin's younger sister, sophie garbin, is an australia fast5 netball international. ### ### [P_TEXT]: sophie garbin sophie garbin ( born 6 april 1997 ) is an australian netball player. she ### ### was a member of the new south wales swifts team that won the 2019 suncorp super netball title. in ### ### 2017 she was also a member of the western sting team that won the australian netball league title. ### ### she was also a member of the australia team that won the bronze medal at the 2018 fast5 netball ### ### world series. garbin's older sister, darcee garbin, is an australia women's basketball ### ### international. garbin is originally from western australia. born in kalgoorlie, she was raised ### ### =============================================== v_q ================================================ ### ### ('¹', 0, 350) ('##rbin', 1, 2) ('sophie', 2, 1) ('ga', 3, 32) ('³', 4, 675) ('netball', 5, 4) ### ### ('fast', 6, 6) ('sister', 7, 23) ('ɐ', 8, 760) ('貴', 9, 124) ('幸', 10, 159) ('australia', 11, 5) ### ### ('##5', 12, 15) ('##ப', 13, 213) ('খ', 14, 174) ('intently', 15, 120) ('ம', 16, 138) ### ### ('women', 17, 47) ('hissed', 18, 115) ('^', 19, 2382) ### ### =============================================== v_p ================================================ ### ### ('sting', 1266, 0) ('sophie', 2, 1) ('##rbin', 1, 2) ('swift', 527, 3) ('netball', 5, 4) ### ### ('australia', 11, 5) ('fast', 6, 6) ('anyway', 78, 7) ('##corp', 4461, 8) ('wales', 368, 9) ### ### ('worried', 447, 10) ('##lie', 9276, 11) ('australian', 24, 12) ('western', 12453, 13) ### ### ('2019', 942, 14) ('##5', 12, 15) ('leaned', 4817, 16) ('basketball', 543, 17) ('caps', 4041, 18) ### ### ('siblings', 729, 19) ### ### ============================================ v_q * v_p ============================================= ### ### ('sophie', 2, 1) ('##rbin', 1, 2) ('netball', 5, 4) ('fast', 6, 6) ('australia', 11, 5) ### ### ('ga', 3, 32) ('##5', 12, 15) ('sister', 7, 23) ('australian', 24, 12) ('anyway', 78, 7) ### ### ('women', 17, 47) ('sisters', 25, 30) ('littered', 22, 39) ('swift', 527, 3) ('¹', 0, 350) ### ### ('貴', 9, 124) ('defending', 119, 20) ('squad', 28, 49) ('sting', 1266, 0) ('wales', 368, 9) ### ############################################################################################################ [2023-11-26 05:05:36,570][root][INFO] - Epoch: 5: Step: 19501/19530, loss[v]=0.161425, lr=0.000000, acc@1[1]=988.0/1024=0.96484375, acc@1[2]=1004.5/1024=0.98095703125 [2023-11-26 05:06:08,886][root][INFO] - rank=2; last iteration 19530 [2023-11-26 05:06:08,886][root][INFO] - rank=2; Multi set iteration finished: iteration per set: [9765, 9765] [2023-11-26 05:06:08,886][root][INFO] - Finished iterating, iteration=9765, shard=2 [2023-11-26 05:06:08,887][root][INFO] - rank=3; last iteration 19530 [2023-11-26 05:06:08,887][root][INFO] - rank=3; Multi set iteration finished: iteration per set: [9765, 9765] [2023-11-26 05:06:08,887][root][INFO] - Finished iterating, iteration=9765, shard=3 [2023-11-26 05:06:08,888][root][INFO] - rank=1; last iteration 19530 [2023-11-26 05:06:08,888][root][INFO] - rank=1; Multi set iteration finished: iteration per set: [9765, 9765] [2023-11-26 05:06:08,888][root][INFO] - Finished iterating, iteration=9765, shard=1 [2023-11-26 05:06:08,888][root][INFO] - rank=0; last iteration 19530 [2023-11-26 05:06:08,888][root][INFO] - rank=0; Multi set iteration finished: iteration per set: [9765, 9765] [2023-11-26 05:06:08,888][root][INFO] - Finished iterating, iteration=9765, shard=0 [2023-11-26 05:06:09,049][root][INFO] - Finished iterating, iteration=9765, shard=2 [2023-11-26 05:06:09,060][root][INFO] - Finished iterating, iteration=9765, shard=0 [2023-11-26 05:06:09,078][root][INFO] - Finished iterating, iteration=9765, shard=1 [2023-11-26 05:06:09,106][root][INFO] - Finished iterating, iteration=9765, shard=3 [2023-11-26 05:06:09,152][root][INFO] - rank=2; Multi set iteration finished after next: iteration per set: [0, 0] [2023-11-26 05:06:09,152][root][INFO] - Epoch finished on 2 [2023-11-26 05:06:09,174][root][INFO] - rank=0; Multi set iteration finished after next: iteration per set: [0, 0] [2023-11-26 05:06:09,174][root][INFO] - Epoch finished on 0 [2023-11-26 05:06:09,198][root][INFO] - rank=1; Multi set iteration finished after next: iteration per set: [0, 0] [2023-11-26 05:06:09,198][root][INFO] - Epoch finished on 1 [2023-11-26 05:06:09,220][root][INFO] - rank=3; Multi set iteration finished after next: iteration per set: [0, 0] [2023-11-26 05:06:09,220][root][INFO] - Epoch finished on 3 [2023-11-26 05:07:11,812][root][INFO] - Saved checkpoint at ./vdr_5 [2023-11-26 05:07:11,812][root][INFO] - Saved checkpoint at ./vdr_5 [2023-11-26 05:07:11,812][root][INFO] - Saved checkpoint at ./vdr_5 [2023-11-26 05:07:11,812][root][INFO] - Av Loss per epoch=0.123474 [2023-11-26 05:07:11,812][root][INFO] - Av Loss per epoch=0.123474 [2023-11-26 05:07:11,812][root][INFO] - epoch total (1) correct predictions=19338330 [2023-11-26 05:07:11,812][root][INFO] - Av Loss per epoch=0.123474 [2023-11-26 05:07:11,813][root][INFO] - epoch total (1) correct predictions=19338330 [2023-11-26 05:07:11,813][root][INFO] - epoch total (2) correct predictions=19743997 [2023-11-26 05:07:11,813][root][INFO] - epoch total (1) correct predictions=19338330 [2023-11-26 05:07:11,813][root][INFO] - epoch total (2) correct predictions=19743997 [2023-11-26 05:07:11,813][root][INFO] - epoch total (2) correct predictions=19743997 [2023-11-26 05:07:11,814][root][INFO] - Saved checkpoint at ./vdr_5 [2023-11-26 05:07:11,814][root][INFO] - Av Loss per epoch=0.123474 [2023-11-26 05:07:11,814][root][INFO] - epoch total (1) correct predictions=19338330 [2023-11-26 05:07:11,814][root][INFO] - epoch total (2) correct predictions=19743997 [2023-11-26 05:07:11,815][root][INFO] - Training finished.