File size: 25,491 Bytes
4b3bade |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 |
2023-10-11 13:00:47,816 ----------------------------------------------------------------------------------------------------
2023-10-11 13:00:47,818 Model: "SequenceTagger(
(embeddings): ByT5Embeddings(
(model): T5EncoderModel(
(shared): Embedding(384, 1472)
(encoder): T5Stack(
(embed_tokens): Embedding(384, 1472)
(block): ModuleList(
(0): T5Block(
(layer): ModuleList(
(0): T5LayerSelfAttention(
(SelfAttention): T5Attention(
(q): Linear(in_features=1472, out_features=384, bias=False)
(k): Linear(in_features=1472, out_features=384, bias=False)
(v): Linear(in_features=1472, out_features=384, bias=False)
(o): Linear(in_features=384, out_features=1472, bias=False)
(relative_attention_bias): Embedding(32, 6)
)
(layer_norm): FusedRMSNorm(torch.Size([1472]), eps=1e-06, elementwise_affine=True)
(dropout): Dropout(p=0.1, inplace=False)
)
(1): T5LayerFF(
(DenseReluDense): T5DenseGatedActDense(
(wi_0): Linear(in_features=1472, out_features=3584, bias=False)
(wi_1): Linear(in_features=1472, out_features=3584, bias=False)
(wo): Linear(in_features=3584, out_features=1472, bias=False)
(dropout): Dropout(p=0.1, inplace=False)
(act): NewGELUActivation()
)
(layer_norm): FusedRMSNorm(torch.Size([1472]), eps=1e-06, elementwise_affine=True)
(dropout): Dropout(p=0.1, inplace=False)
)
)
)
(1-11): 11 x T5Block(
(layer): ModuleList(
(0): T5LayerSelfAttention(
(SelfAttention): T5Attention(
(q): Linear(in_features=1472, out_features=384, bias=False)
(k): Linear(in_features=1472, out_features=384, bias=False)
(v): Linear(in_features=1472, out_features=384, bias=False)
(o): Linear(in_features=384, out_features=1472, bias=False)
)
(layer_norm): FusedRMSNorm(torch.Size([1472]), eps=1e-06, elementwise_affine=True)
(dropout): Dropout(p=0.1, inplace=False)
)
(1): T5LayerFF(
(DenseReluDense): T5DenseGatedActDense(
(wi_0): Linear(in_features=1472, out_features=3584, bias=False)
(wi_1): Linear(in_features=1472, out_features=3584, bias=False)
(wo): Linear(in_features=3584, out_features=1472, bias=False)
(dropout): Dropout(p=0.1, inplace=False)
(act): NewGELUActivation()
)
(layer_norm): FusedRMSNorm(torch.Size([1472]), eps=1e-06, elementwise_affine=True)
(dropout): Dropout(p=0.1, inplace=False)
)
)
)
)
(final_layer_norm): FusedRMSNorm(torch.Size([1472]), eps=1e-06, elementwise_affine=True)
(dropout): Dropout(p=0.1, inplace=False)
)
)
)
(locked_dropout): LockedDropout(p=0.5)
(linear): Linear(in_features=1472, out_features=17, bias=True)
(loss_function): CrossEntropyLoss()
)"
2023-10-11 13:00:47,818 ----------------------------------------------------------------------------------------------------
2023-10-11 13:00:47,818 MultiCorpus: 7142 train + 698 dev + 2570 test sentences
- NER_HIPE_2022 Corpus: 7142 train + 698 dev + 2570 test sentences - /root/.flair/datasets/ner_hipe_2022/v2.1/newseye/fr/with_doc_seperator
2023-10-11 13:00:47,818 ----------------------------------------------------------------------------------------------------
2023-10-11 13:00:47,818 Train: 7142 sentences
2023-10-11 13:00:47,818 (train_with_dev=False, train_with_test=False)
2023-10-11 13:00:47,819 ----------------------------------------------------------------------------------------------------
2023-10-11 13:00:47,819 Training Params:
2023-10-11 13:00:47,819 - learning_rate: "0.00016"
2023-10-11 13:00:47,819 - mini_batch_size: "4"
2023-10-11 13:00:47,819 - max_epochs: "10"
2023-10-11 13:00:47,819 - shuffle: "True"
2023-10-11 13:00:47,819 ----------------------------------------------------------------------------------------------------
2023-10-11 13:00:47,819 Plugins:
2023-10-11 13:00:47,819 - TensorboardLogger
2023-10-11 13:00:47,819 - LinearScheduler | warmup_fraction: '0.1'
2023-10-11 13:00:47,819 ----------------------------------------------------------------------------------------------------
2023-10-11 13:00:47,819 Final evaluation on model from best epoch (best-model.pt)
2023-10-11 13:00:47,819 - metric: "('micro avg', 'f1-score')"
2023-10-11 13:00:47,820 ----------------------------------------------------------------------------------------------------
2023-10-11 13:00:47,820 Computation:
2023-10-11 13:00:47,820 - compute on device: cuda:0
2023-10-11 13:00:47,820 - embedding storage: none
2023-10-11 13:00:47,820 ----------------------------------------------------------------------------------------------------
2023-10-11 13:00:47,820 Model training base path: "hmbench-newseye/fr-hmbyt5-preliminary/byt5-small-historic-multilingual-span20-flax-bs4-wsFalse-e10-lr0.00016-poolingfirst-layers-1-crfFalse-3"
2023-10-11 13:00:47,820 ----------------------------------------------------------------------------------------------------
2023-10-11 13:00:47,820 ----------------------------------------------------------------------------------------------------
2023-10-11 13:00:47,820 Logging anything other than scalars to TensorBoard is currently not supported.
2023-10-11 13:01:44,208 epoch 1 - iter 178/1786 - loss 2.81170991 - time (sec): 56.38 - samples/sec: 477.56 - lr: 0.000016 - momentum: 0.000000
2023-10-11 13:02:39,090 epoch 1 - iter 356/1786 - loss 2.63745970 - time (sec): 111.27 - samples/sec: 455.58 - lr: 0.000032 - momentum: 0.000000
2023-10-11 13:03:34,033 epoch 1 - iter 534/1786 - loss 2.35715581 - time (sec): 166.21 - samples/sec: 448.32 - lr: 0.000048 - momentum: 0.000000
2023-10-11 13:04:29,519 epoch 1 - iter 712/1786 - loss 2.06074361 - time (sec): 221.70 - samples/sec: 445.36 - lr: 0.000064 - momentum: 0.000000
2023-10-11 13:05:26,211 epoch 1 - iter 890/1786 - loss 1.78883260 - time (sec): 278.39 - samples/sec: 448.55 - lr: 0.000080 - momentum: 0.000000
2023-10-11 13:06:20,110 epoch 1 - iter 1068/1786 - loss 1.58610501 - time (sec): 332.29 - samples/sec: 447.69 - lr: 0.000096 - momentum: 0.000000
2023-10-11 13:07:15,433 epoch 1 - iter 1246/1786 - loss 1.40913339 - time (sec): 387.61 - samples/sec: 449.86 - lr: 0.000112 - momentum: 0.000000
2023-10-11 13:08:13,637 epoch 1 - iter 1424/1786 - loss 1.27901313 - time (sec): 445.81 - samples/sec: 446.05 - lr: 0.000127 - momentum: 0.000000
2023-10-11 13:09:09,748 epoch 1 - iter 1602/1786 - loss 1.16574649 - time (sec): 501.92 - samples/sec: 446.37 - lr: 0.000143 - momentum: 0.000000
2023-10-11 13:10:05,183 epoch 1 - iter 1780/1786 - loss 1.07906676 - time (sec): 557.36 - samples/sec: 444.83 - lr: 0.000159 - momentum: 0.000000
2023-10-11 13:10:06,873 ----------------------------------------------------------------------------------------------------
2023-10-11 13:10:06,874 EPOCH 1 done: loss 1.0763 - lr: 0.000159
2023-10-11 13:10:27,235 DEV : loss 0.17565929889678955 - f1-score (micro avg) 0.6003
2023-10-11 13:10:27,268 saving best model
2023-10-11 13:10:28,252 ----------------------------------------------------------------------------------------------------
2023-10-11 13:11:24,075 epoch 2 - iter 178/1786 - loss 0.16255892 - time (sec): 55.82 - samples/sec: 470.03 - lr: 0.000158 - momentum: 0.000000
2023-10-11 13:12:20,568 epoch 2 - iter 356/1786 - loss 0.16881265 - time (sec): 112.31 - samples/sec: 458.36 - lr: 0.000156 - momentum: 0.000000
2023-10-11 13:13:19,344 epoch 2 - iter 534/1786 - loss 0.15873645 - time (sec): 171.09 - samples/sec: 440.45 - lr: 0.000155 - momentum: 0.000000
2023-10-11 13:14:17,771 epoch 2 - iter 712/1786 - loss 0.14765138 - time (sec): 229.52 - samples/sec: 438.98 - lr: 0.000153 - momentum: 0.000000
2023-10-11 13:15:13,377 epoch 2 - iter 890/1786 - loss 0.14254566 - time (sec): 285.12 - samples/sec: 436.39 - lr: 0.000151 - momentum: 0.000000
2023-10-11 13:16:14,773 epoch 2 - iter 1068/1786 - loss 0.13991538 - time (sec): 346.52 - samples/sec: 433.54 - lr: 0.000149 - momentum: 0.000000
2023-10-11 13:17:16,141 epoch 2 - iter 1246/1786 - loss 0.13851750 - time (sec): 407.89 - samples/sec: 428.23 - lr: 0.000148 - momentum: 0.000000
2023-10-11 13:18:15,089 epoch 2 - iter 1424/1786 - loss 0.13580485 - time (sec): 466.83 - samples/sec: 424.05 - lr: 0.000146 - momentum: 0.000000
2023-10-11 13:19:09,450 epoch 2 - iter 1602/1786 - loss 0.13348564 - time (sec): 521.20 - samples/sec: 426.97 - lr: 0.000144 - momentum: 0.000000
2023-10-11 13:20:06,575 epoch 2 - iter 1780/1786 - loss 0.13055574 - time (sec): 578.32 - samples/sec: 428.70 - lr: 0.000142 - momentum: 0.000000
2023-10-11 13:20:08,493 ----------------------------------------------------------------------------------------------------
2023-10-11 13:20:08,493 EPOCH 2 done: loss 0.1303 - lr: 0.000142
2023-10-11 13:20:30,672 DEV : loss 0.10548630356788635 - f1-score (micro avg) 0.7578
2023-10-11 13:20:30,703 saving best model
2023-10-11 13:20:33,274 ----------------------------------------------------------------------------------------------------
2023-10-11 13:21:25,594 epoch 3 - iter 178/1786 - loss 0.06430020 - time (sec): 52.32 - samples/sec: 455.81 - lr: 0.000140 - momentum: 0.000000
2023-10-11 13:22:18,793 epoch 3 - iter 356/1786 - loss 0.06283356 - time (sec): 105.51 - samples/sec: 463.47 - lr: 0.000139 - momentum: 0.000000
2023-10-11 13:23:10,053 epoch 3 - iter 534/1786 - loss 0.06385061 - time (sec): 156.77 - samples/sec: 466.82 - lr: 0.000137 - momentum: 0.000000
2023-10-11 13:24:04,302 epoch 3 - iter 712/1786 - loss 0.06659712 - time (sec): 211.02 - samples/sec: 465.28 - lr: 0.000135 - momentum: 0.000000
2023-10-11 13:24:56,648 epoch 3 - iter 890/1786 - loss 0.07114886 - time (sec): 263.37 - samples/sec: 466.58 - lr: 0.000133 - momentum: 0.000000
2023-10-11 13:25:48,940 epoch 3 - iter 1068/1786 - loss 0.07352806 - time (sec): 315.66 - samples/sec: 467.30 - lr: 0.000132 - momentum: 0.000000
2023-10-11 13:26:45,407 epoch 3 - iter 1246/1786 - loss 0.07514593 - time (sec): 372.13 - samples/sec: 468.01 - lr: 0.000130 - momentum: 0.000000
2023-10-11 13:27:42,174 epoch 3 - iter 1424/1786 - loss 0.07344231 - time (sec): 428.90 - samples/sec: 461.84 - lr: 0.000128 - momentum: 0.000000
2023-10-11 13:28:39,296 epoch 3 - iter 1602/1786 - loss 0.07210740 - time (sec): 486.02 - samples/sec: 459.31 - lr: 0.000126 - momentum: 0.000000
2023-10-11 13:29:36,366 epoch 3 - iter 1780/1786 - loss 0.07250482 - time (sec): 543.09 - samples/sec: 456.92 - lr: 0.000125 - momentum: 0.000000
2023-10-11 13:29:38,029 ----------------------------------------------------------------------------------------------------
2023-10-11 13:29:38,029 EPOCH 3 done: loss 0.0726 - lr: 0.000125
2023-10-11 13:30:01,230 DEV : loss 0.10552459955215454 - f1-score (micro avg) 0.7957
2023-10-11 13:30:01,262 saving best model
2023-10-11 13:30:03,958 ----------------------------------------------------------------------------------------------------
2023-10-11 13:30:59,030 epoch 4 - iter 178/1786 - loss 0.04332888 - time (sec): 55.07 - samples/sec: 436.30 - lr: 0.000123 - momentum: 0.000000
2023-10-11 13:31:55,067 epoch 4 - iter 356/1786 - loss 0.04883610 - time (sec): 111.10 - samples/sec: 441.19 - lr: 0.000121 - momentum: 0.000000
2023-10-11 13:32:50,737 epoch 4 - iter 534/1786 - loss 0.04847201 - time (sec): 166.77 - samples/sec: 453.36 - lr: 0.000119 - momentum: 0.000000
2023-10-11 13:33:44,707 epoch 4 - iter 712/1786 - loss 0.05062238 - time (sec): 220.74 - samples/sec: 453.88 - lr: 0.000117 - momentum: 0.000000
2023-10-11 13:34:38,564 epoch 4 - iter 890/1786 - loss 0.05112561 - time (sec): 274.60 - samples/sec: 458.03 - lr: 0.000116 - momentum: 0.000000
2023-10-11 13:35:31,790 epoch 4 - iter 1068/1786 - loss 0.05210222 - time (sec): 327.83 - samples/sec: 453.38 - lr: 0.000114 - momentum: 0.000000
2023-10-11 13:36:30,302 epoch 4 - iter 1246/1786 - loss 0.05253056 - time (sec): 386.34 - samples/sec: 449.69 - lr: 0.000112 - momentum: 0.000000
2023-10-11 13:37:25,466 epoch 4 - iter 1424/1786 - loss 0.05198458 - time (sec): 441.50 - samples/sec: 448.36 - lr: 0.000110 - momentum: 0.000000
2023-10-11 13:38:23,731 epoch 4 - iter 1602/1786 - loss 0.05147654 - time (sec): 499.77 - samples/sec: 446.38 - lr: 0.000109 - momentum: 0.000000
2023-10-11 13:39:18,931 epoch 4 - iter 1780/1786 - loss 0.05054278 - time (sec): 554.97 - samples/sec: 446.88 - lr: 0.000107 - momentum: 0.000000
2023-10-11 13:39:20,569 ----------------------------------------------------------------------------------------------------
2023-10-11 13:39:20,569 EPOCH 4 done: loss 0.0505 - lr: 0.000107
2023-10-11 13:39:43,214 DEV : loss 0.156653493642807 - f1-score (micro avg) 0.7909
2023-10-11 13:39:43,255 ----------------------------------------------------------------------------------------------------
2023-10-11 13:40:43,029 epoch 5 - iter 178/1786 - loss 0.03328650 - time (sec): 59.77 - samples/sec: 420.20 - lr: 0.000105 - momentum: 0.000000
2023-10-11 13:41:39,262 epoch 5 - iter 356/1786 - loss 0.04050878 - time (sec): 116.00 - samples/sec: 436.17 - lr: 0.000103 - momentum: 0.000000
2023-10-11 13:42:33,938 epoch 5 - iter 534/1786 - loss 0.03621945 - time (sec): 170.68 - samples/sec: 443.11 - lr: 0.000101 - momentum: 0.000000
2023-10-11 13:43:27,857 epoch 5 - iter 712/1786 - loss 0.03475165 - time (sec): 224.60 - samples/sec: 444.06 - lr: 0.000100 - momentum: 0.000000
2023-10-11 13:44:20,262 epoch 5 - iter 890/1786 - loss 0.03461616 - time (sec): 277.00 - samples/sec: 446.22 - lr: 0.000098 - momentum: 0.000000
2023-10-11 13:45:14,874 epoch 5 - iter 1068/1786 - loss 0.03386874 - time (sec): 331.62 - samples/sec: 445.15 - lr: 0.000096 - momentum: 0.000000
2023-10-11 13:46:12,877 epoch 5 - iter 1246/1786 - loss 0.03430305 - time (sec): 389.62 - samples/sec: 443.88 - lr: 0.000094 - momentum: 0.000000
2023-10-11 13:47:13,910 epoch 5 - iter 1424/1786 - loss 0.03506913 - time (sec): 450.65 - samples/sec: 437.96 - lr: 0.000093 - momentum: 0.000000
2023-10-11 13:48:13,048 epoch 5 - iter 1602/1786 - loss 0.03525813 - time (sec): 509.79 - samples/sec: 436.07 - lr: 0.000091 - momentum: 0.000000
2023-10-11 13:49:06,991 epoch 5 - iter 1780/1786 - loss 0.03807119 - time (sec): 563.73 - samples/sec: 440.04 - lr: 0.000089 - momentum: 0.000000
2023-10-11 13:49:08,616 ----------------------------------------------------------------------------------------------------
2023-10-11 13:49:08,616 EPOCH 5 done: loss 0.0381 - lr: 0.000089
2023-10-11 13:49:30,953 DEV : loss 0.17025883495807648 - f1-score (micro avg) 0.788
2023-10-11 13:49:30,986 ----------------------------------------------------------------------------------------------------
2023-10-11 13:50:30,448 epoch 6 - iter 178/1786 - loss 0.03037012 - time (sec): 59.46 - samples/sec: 437.48 - lr: 0.000087 - momentum: 0.000000
2023-10-11 13:51:23,893 epoch 6 - iter 356/1786 - loss 0.02735029 - time (sec): 112.90 - samples/sec: 439.65 - lr: 0.000085 - momentum: 0.000000
2023-10-11 13:52:18,715 epoch 6 - iter 534/1786 - loss 0.02908184 - time (sec): 167.73 - samples/sec: 439.17 - lr: 0.000084 - momentum: 0.000000
2023-10-11 13:53:17,235 epoch 6 - iter 712/1786 - loss 0.02904645 - time (sec): 226.25 - samples/sec: 438.56 - lr: 0.000082 - momentum: 0.000000
2023-10-11 13:54:11,434 epoch 6 - iter 890/1786 - loss 0.02790749 - time (sec): 280.45 - samples/sec: 441.20 - lr: 0.000080 - momentum: 0.000000
2023-10-11 13:55:03,988 epoch 6 - iter 1068/1786 - loss 0.02701632 - time (sec): 333.00 - samples/sec: 441.61 - lr: 0.000078 - momentum: 0.000000
2023-10-11 13:55:57,385 epoch 6 - iter 1246/1786 - loss 0.02648261 - time (sec): 386.40 - samples/sec: 444.67 - lr: 0.000077 - momentum: 0.000000
2023-10-11 13:56:56,380 epoch 6 - iter 1424/1786 - loss 0.02759848 - time (sec): 445.39 - samples/sec: 444.90 - lr: 0.000075 - momentum: 0.000000
2023-10-11 13:57:51,975 epoch 6 - iter 1602/1786 - loss 0.02741452 - time (sec): 500.99 - samples/sec: 445.51 - lr: 0.000073 - momentum: 0.000000
2023-10-11 13:58:44,618 epoch 6 - iter 1780/1786 - loss 0.02735127 - time (sec): 553.63 - samples/sec: 448.07 - lr: 0.000071 - momentum: 0.000000
2023-10-11 13:58:46,201 ----------------------------------------------------------------------------------------------------
2023-10-11 13:58:46,201 EPOCH 6 done: loss 0.0273 - lr: 0.000071
2023-10-11 13:59:06,894 DEV : loss 0.18334950506687164 - f1-score (micro avg) 0.8033
2023-10-11 13:59:06,940 saving best model
2023-10-11 13:59:09,574 ----------------------------------------------------------------------------------------------------
2023-10-11 14:00:03,722 epoch 7 - iter 178/1786 - loss 0.02116711 - time (sec): 54.14 - samples/sec: 452.76 - lr: 0.000069 - momentum: 0.000000
2023-10-11 14:00:56,097 epoch 7 - iter 356/1786 - loss 0.02358540 - time (sec): 106.52 - samples/sec: 448.60 - lr: 0.000068 - momentum: 0.000000
2023-10-11 14:01:50,382 epoch 7 - iter 534/1786 - loss 0.02380570 - time (sec): 160.80 - samples/sec: 454.36 - lr: 0.000066 - momentum: 0.000000
2023-10-11 14:02:44,369 epoch 7 - iter 712/1786 - loss 0.02389820 - time (sec): 214.79 - samples/sec: 452.33 - lr: 0.000064 - momentum: 0.000000
2023-10-11 14:03:38,758 epoch 7 - iter 890/1786 - loss 0.02394287 - time (sec): 269.18 - samples/sec: 454.52 - lr: 0.000062 - momentum: 0.000000
2023-10-11 14:04:33,458 epoch 7 - iter 1068/1786 - loss 0.02260999 - time (sec): 323.88 - samples/sec: 456.84 - lr: 0.000061 - momentum: 0.000000
2023-10-11 14:05:28,679 epoch 7 - iter 1246/1786 - loss 0.02182956 - time (sec): 379.10 - samples/sec: 456.41 - lr: 0.000059 - momentum: 0.000000
2023-10-11 14:06:22,977 epoch 7 - iter 1424/1786 - loss 0.02105207 - time (sec): 433.40 - samples/sec: 457.13 - lr: 0.000057 - momentum: 0.000000
2023-10-11 14:07:16,767 epoch 7 - iter 1602/1786 - loss 0.02145575 - time (sec): 487.19 - samples/sec: 458.26 - lr: 0.000055 - momentum: 0.000000
2023-10-11 14:08:09,598 epoch 7 - iter 1780/1786 - loss 0.02106435 - time (sec): 540.02 - samples/sec: 459.55 - lr: 0.000053 - momentum: 0.000000
2023-10-11 14:08:11,066 ----------------------------------------------------------------------------------------------------
2023-10-11 14:08:11,067 EPOCH 7 done: loss 0.0211 - lr: 0.000053
2023-10-11 14:08:32,426 DEV : loss 0.194308340549469 - f1-score (micro avg) 0.7928
2023-10-11 14:08:32,455 ----------------------------------------------------------------------------------------------------
2023-10-11 14:09:23,421 epoch 8 - iter 178/1786 - loss 0.01641532 - time (sec): 50.96 - samples/sec: 484.58 - lr: 0.000052 - momentum: 0.000000
2023-10-11 14:10:14,666 epoch 8 - iter 356/1786 - loss 0.01440783 - time (sec): 102.21 - samples/sec: 483.68 - lr: 0.000050 - momentum: 0.000000
2023-10-11 14:11:06,282 epoch 8 - iter 534/1786 - loss 0.01285935 - time (sec): 153.83 - samples/sec: 473.03 - lr: 0.000048 - momentum: 0.000000
2023-10-11 14:11:58,394 epoch 8 - iter 712/1786 - loss 0.01220440 - time (sec): 205.94 - samples/sec: 467.19 - lr: 0.000046 - momentum: 0.000000
2023-10-11 14:12:50,304 epoch 8 - iter 890/1786 - loss 0.01374698 - time (sec): 257.85 - samples/sec: 465.93 - lr: 0.000044 - momentum: 0.000000
2023-10-11 14:13:44,353 epoch 8 - iter 1068/1786 - loss 0.01485753 - time (sec): 311.90 - samples/sec: 469.85 - lr: 0.000043 - momentum: 0.000000
2023-10-11 14:14:37,502 epoch 8 - iter 1246/1786 - loss 0.01514008 - time (sec): 365.05 - samples/sec: 472.32 - lr: 0.000041 - momentum: 0.000000
2023-10-11 14:15:30,709 epoch 8 - iter 1424/1786 - loss 0.01557502 - time (sec): 418.25 - samples/sec: 475.69 - lr: 0.000039 - momentum: 0.000000
2023-10-11 14:16:22,878 epoch 8 - iter 1602/1786 - loss 0.01641034 - time (sec): 470.42 - samples/sec: 477.11 - lr: 0.000037 - momentum: 0.000000
2023-10-11 14:17:13,034 epoch 8 - iter 1780/1786 - loss 0.01592313 - time (sec): 520.58 - samples/sec: 476.58 - lr: 0.000036 - momentum: 0.000000
2023-10-11 14:17:14,558 ----------------------------------------------------------------------------------------------------
2023-10-11 14:17:14,559 EPOCH 8 done: loss 0.0159 - lr: 0.000036
2023-10-11 14:17:36,426 DEV : loss 0.2070261836051941 - f1-score (micro avg) 0.7869
2023-10-11 14:17:36,457 ----------------------------------------------------------------------------------------------------
2023-10-11 14:18:28,783 epoch 9 - iter 178/1786 - loss 0.01279919 - time (sec): 52.32 - samples/sec: 455.83 - lr: 0.000034 - momentum: 0.000000
2023-10-11 14:19:20,233 epoch 9 - iter 356/1786 - loss 0.01083550 - time (sec): 103.77 - samples/sec: 450.34 - lr: 0.000032 - momentum: 0.000000
2023-10-11 14:20:11,070 epoch 9 - iter 534/1786 - loss 0.01073281 - time (sec): 154.61 - samples/sec: 445.00 - lr: 0.000030 - momentum: 0.000000
2023-10-11 14:21:04,072 epoch 9 - iter 712/1786 - loss 0.01005165 - time (sec): 207.61 - samples/sec: 457.10 - lr: 0.000028 - momentum: 0.000000
2023-10-11 14:21:56,520 epoch 9 - iter 890/1786 - loss 0.01008754 - time (sec): 260.06 - samples/sec: 462.46 - lr: 0.000027 - momentum: 0.000000
2023-10-11 14:22:50,929 epoch 9 - iter 1068/1786 - loss 0.01062970 - time (sec): 314.47 - samples/sec: 464.76 - lr: 0.000025 - momentum: 0.000000
2023-10-11 14:23:44,599 epoch 9 - iter 1246/1786 - loss 0.01091575 - time (sec): 368.14 - samples/sec: 468.87 - lr: 0.000023 - momentum: 0.000000
2023-10-11 14:24:38,320 epoch 9 - iter 1424/1786 - loss 0.01047999 - time (sec): 421.86 - samples/sec: 470.87 - lr: 0.000021 - momentum: 0.000000
2023-10-11 14:25:31,677 epoch 9 - iter 1602/1786 - loss 0.01056976 - time (sec): 475.22 - samples/sec: 470.42 - lr: 0.000020 - momentum: 0.000000
2023-10-11 14:26:25,182 epoch 9 - iter 1780/1786 - loss 0.01085308 - time (sec): 528.72 - samples/sec: 468.44 - lr: 0.000018 - momentum: 0.000000
2023-10-11 14:26:27,252 ----------------------------------------------------------------------------------------------------
2023-10-11 14:26:27,252 EPOCH 9 done: loss 0.0109 - lr: 0.000018
2023-10-11 14:26:49,598 DEV : loss 0.21609559655189514 - f1-score (micro avg) 0.8054
2023-10-11 14:26:49,638 saving best model
2023-10-11 14:26:52,383 ----------------------------------------------------------------------------------------------------
2023-10-11 14:27:45,526 epoch 10 - iter 178/1786 - loss 0.00795703 - time (sec): 53.14 - samples/sec: 469.48 - lr: 0.000016 - momentum: 0.000000
2023-10-11 14:28:38,444 epoch 10 - iter 356/1786 - loss 0.00911006 - time (sec): 106.06 - samples/sec: 453.61 - lr: 0.000014 - momentum: 0.000000
2023-10-11 14:29:32,746 epoch 10 - iter 534/1786 - loss 0.00807375 - time (sec): 160.36 - samples/sec: 453.72 - lr: 0.000012 - momentum: 0.000000
2023-10-11 14:30:25,481 epoch 10 - iter 712/1786 - loss 0.00794053 - time (sec): 213.09 - samples/sec: 461.38 - lr: 0.000011 - momentum: 0.000000
2023-10-11 14:31:18,756 epoch 10 - iter 890/1786 - loss 0.00822094 - time (sec): 266.37 - samples/sec: 466.30 - lr: 0.000009 - momentum: 0.000000
2023-10-11 14:32:11,427 epoch 10 - iter 1068/1786 - loss 0.00765007 - time (sec): 319.04 - samples/sec: 466.07 - lr: 0.000007 - momentum: 0.000000
2023-10-11 14:33:02,846 epoch 10 - iter 1246/1786 - loss 0.00782413 - time (sec): 370.46 - samples/sec: 466.16 - lr: 0.000005 - momentum: 0.000000
2023-10-11 14:33:55,039 epoch 10 - iter 1424/1786 - loss 0.00755323 - time (sec): 422.65 - samples/sec: 467.52 - lr: 0.000004 - momentum: 0.000000
2023-10-11 14:34:48,378 epoch 10 - iter 1602/1786 - loss 0.00741746 - time (sec): 475.99 - samples/sec: 467.39 - lr: 0.000002 - momentum: 0.000000
2023-10-11 14:35:41,429 epoch 10 - iter 1780/1786 - loss 0.00748365 - time (sec): 529.04 - samples/sec: 469.23 - lr: 0.000000 - momentum: 0.000000
2023-10-11 14:35:42,901 ----------------------------------------------------------------------------------------------------
2023-10-11 14:35:42,901 EPOCH 10 done: loss 0.0075 - lr: 0.000000
2023-10-11 14:36:04,553 DEV : loss 0.22351031005382538 - f1-score (micro avg) 0.7973
2023-10-11 14:36:05,510 ----------------------------------------------------------------------------------------------------
2023-10-11 14:36:05,512 Loading model from best epoch ...
2023-10-11 14:36:09,777 SequenceTagger predicts: Dictionary with 17 tags: O, S-PER, B-PER, E-PER, I-PER, S-LOC, B-LOC, E-LOC, I-LOC, S-ORG, B-ORG, E-ORG, I-ORG, S-HumanProd, B-HumanProd, E-HumanProd, I-HumanProd
2023-10-11 14:37:17,218
Results:
- F-score (micro) 0.7051
- F-score (macro) 0.6521
- Accuracy 0.5611
By class:
precision recall f1-score support
LOC 0.6985 0.7342 0.7159 1095
PER 0.7796 0.7688 0.7741 1012
ORG 0.4577 0.5910 0.5159 357
HumanProd 0.5000 0.7576 0.6024 33
micro avg 0.6835 0.7281 0.7051 2497
macro avg 0.6089 0.7129 0.6521 2497
weighted avg 0.6943 0.7281 0.7094 2497
2023-10-11 14:37:17,218 ----------------------------------------------------------------------------------------------------
|