Stage best WCE model

config.json

@@ -0,0 +1,62 @@
+{
+  "_name_or_path": "Helsinki-NLP/opus-mt-en-fr",
+  "_num_labels": 3,
+  "activation_dropout": 0.0,
+  "activation_function": "swish",
+  "add_bias_logits": false,
+  "add_final_layer_norm": false,
+  "architectures": [
+    "MarianMTModel"
+  ],
+  "attention_dropout": 0.0,
+  "bad_words_ids": [
+    [
+      59513
+    ]
+  ],
+  "bos_token_id": 0,
+  "classif_dropout": 0.0,
+  "classifier_dropout": 0.0,
+  "d_model": 512,
+  "decoder_attention_heads": 8,
+  "decoder_ffn_dim": 2048,
+  "decoder_layerdrop": 0.0,
+  "decoder_layers": 6,
+  "decoder_start_token_id": 59513,
+  "decoder_vocab_size": 59514,
+  "dropout": 0.1,
+  "encoder_attention_heads": 8,
+  "encoder_ffn_dim": 2048,
+  "encoder_layerdrop": 0.0,
+  "encoder_layers": 6,
+  "eos_token_id": 0,
+  "forced_eos_token_id": 0,
+  "gradient_checkpointing": false,
+  "id2label": {
+    "0": "LABEL_0",
+    "1": "LABEL_1",
+    "2": "LABEL_2"
+  },
+  "init_std": 0.02,
+  "is_encoder_decoder": true,
+  "label2id": {
+    "LABEL_0": 0,
+    "LABEL_1": 1,
+    "LABEL_2": 2
+  },
+  "max_length": 512,
+  "max_position_embeddings": 512,
+  "model_type": "marian",
+  "normalize_before": false,
+  "normalize_embedding": false,
+  "num_beams": 4,
+  "num_hidden_layers": 6,
+  "pad_token_id": 59513,
+  "scale_embedding": true,
+  "share_encoder_decoder_embeddings": true,
+  "static_position_embeddings": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.28.0",
+  "use_cache": true,
+  "vocab_size": 59514
+}

generation_config.json

@@ -0,0 +1,15 @@
+{
+  "bad_words_ids": [
+    [
+      59513
+    ]
+  ],
+  "bos_token_id": 0,
+  "decoder_start_token_id": 59513,
+  "eos_token_id": 0,
+  "forced_eos_token_id": 0,
+  "max_length": 512,
+  "num_beams": 4,
+  "pad_token_id": 59513,
+  "transformers_version": "4.28.0"
+}

optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b5c22862dc03fa99839e830432c2d7c04eaea21b2964c3f3ba774d6216d04a9d
+size 597026309

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b75f8851ac38b1305116187b4dfad449a703e1338f7bad4b1f87ce5c8d7be76
+size 298763205

rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f43d7c5f8261a5ad1620907c374687b428cdbf997331040068ff213073e53027
+size 14575

scaler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:121d315e08f0072d1b81df334b3a412af4d5c3af247268ce862f86553defdcb2
+size 557

scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:510fc89adea15071db8c3934ea1fa2c117901f9a1f6074bdec2fb0efd2696b52
+size 627

special_tokens_map.json

@@ -0,0 +1,5 @@
+{
+  "eos_token": "</s>",
+  "pad_token": "<pad>",
+  "unk_token": "<unk>"
+}

tokenizer_config.json

@@ -0,0 +1,12 @@
+{
+  "clean_up_tokenization_spaces": true,
+  "eos_token": "</s>",
+  "model_max_length": 512,
+  "pad_token": "<pad>",
+  "separate_vocabs": false,
+  "source_lang": "en",
+  "sp_model_kwargs": {},
+  "target_lang": "fr",
+  "tokenizer_class": "MarianTokenizer",
+  "unk_token": "<unk>"
+}

trainer_state.json

@@ -0,0 +1,336 @@
+{
+  "best_metric": 43.0447,
+  "best_model_checkpoint": "wce_weights_1.25/checkpoint-80000",
+  "epoch": 3.943217665615142,
+  "global_step": 80000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.2,
+      "learning_rate": 1.9753918572555207e-05,
+      "loss": 0.2206,
+      "step": 4000
+    },
+    {
+      "epoch": 0.2,
+      "eval_bleu": 34.0224,
+      "eval_gen_len": 41.8706,
+      "eval_loss": 0.10892540216445923,
+      "eval_runtime": 164.3546,
+      "eval_samples_per_second": 6.346,
+      "eval_steps_per_second": 0.201,
+      "step": 4000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.950746746845426e-05,
+      "loss": 0.1864,
+      "step": 8000
+    },
+    {
+      "epoch": 0.39,
+      "eval_bleu": 37.6284,
+      "eval_gen_len": 37.8658,
+      "eval_loss": 0.1066218838095665,
+      "eval_runtime": 128.9732,
+      "eval_samples_per_second": 8.087,
+      "eval_steps_per_second": 0.256,
+      "step": 8000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.9261016364353314e-05,
+      "loss": 0.1789,
+      "step": 12000
+    },
+    {
+      "epoch": 0.59,
+      "eval_bleu": 41.6474,
+      "eval_gen_len": 34.466,
+      "eval_loss": 0.10536406934261322,
+      "eval_runtime": 116.1518,
+      "eval_samples_per_second": 8.98,
+      "eval_steps_per_second": 0.284,
+      "step": 12000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.9014565260252367e-05,
+      "loss": 0.1731,
+      "step": 16000
+    },
+    {
+      "epoch": 0.79,
+      "eval_bleu": 40.7652,
+      "eval_gen_len": 36.8926,
+      "eval_loss": 0.10414893180131912,
+      "eval_runtime": 97.8099,
+      "eval_samples_per_second": 10.664,
+      "eval_steps_per_second": 0.337,
+      "step": 16000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.8768175768927446e-05,
+      "loss": 0.1706,
+      "step": 20000
+    },
+    {
+      "epoch": 0.99,
+      "eval_bleu": 41.9082,
+      "eval_gen_len": 35.7881,
+      "eval_loss": 0.10362079739570618,
+      "eval_runtime": 108.0775,
+      "eval_samples_per_second": 9.65,
+      "eval_steps_per_second": 0.305,
+      "step": 20000
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.8521786277602524e-05,
+      "loss": 0.1583,
+      "step": 24000
+    },
+    {
+      "epoch": 1.18,
+      "eval_bleu": 42.0588,
+      "eval_gen_len": 35.4621,
+      "eval_loss": 0.10367583483457565,
+      "eval_runtime": 93.0138,
+      "eval_samples_per_second": 11.213,
+      "eval_steps_per_second": 0.355,
+      "step": 24000
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 1.827545839905363e-05,
+      "loss": 0.158,
+      "step": 28000
+    },
+    {
+      "epoch": 1.38,
+      "eval_bleu": 41.902,
+      "eval_gen_len": 35.0499,
+      "eval_loss": 0.10282289236783981,
+      "eval_runtime": 80.9731,
+      "eval_samples_per_second": 12.881,
+      "eval_steps_per_second": 0.408,
+      "step": 28000
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 1.8029068907728708e-05,
+      "loss": 0.1575,
+      "step": 32000
+    },
+    {
+      "epoch": 1.58,
+      "eval_bleu": 41.4264,
+      "eval_gen_len": 36.1477,
+      "eval_loss": 0.10224233567714691,
+      "eval_runtime": 103.3427,
+      "eval_samples_per_second": 10.093,
+      "eval_steps_per_second": 0.319,
+      "step": 32000
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.7782679416403786e-05,
+      "loss": 0.156,
+      "step": 36000
+    },
+    {
+      "epoch": 1.77,
+      "eval_bleu": 40.4682,
+      "eval_gen_len": 37.5638,
+      "eval_loss": 0.10195210576057434,
+      "eval_runtime": 92.6522,
+      "eval_samples_per_second": 11.257,
+      "eval_steps_per_second": 0.356,
+      "step": 36000
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.7536351537854893e-05,
+      "loss": 0.1551,
+      "step": 40000
+    },
+    {
+      "epoch": 1.97,
+      "eval_bleu": 42.3418,
+      "eval_gen_len": 34.9012,
+      "eval_loss": 0.10195144265890121,
+      "eval_runtime": 90.4347,
+      "eval_samples_per_second": 11.533,
+      "eval_steps_per_second": 0.365,
+      "step": 40000
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1.7289962046529968e-05,
+      "loss": 0.1472,
+      "step": 44000
+    },
+    {
+      "epoch": 2.17,
+      "eval_bleu": 42.3142,
+      "eval_gen_len": 35.3931,
+      "eval_loss": 0.1024407148361206,
+      "eval_runtime": 152.6783,
+      "eval_samples_per_second": 6.831,
+      "eval_steps_per_second": 0.216,
+      "step": 44000
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 1.7043634167981074e-05,
+      "loss": 0.1462,
+      "step": 48000
+    },
+    {
+      "epoch": 2.37,
+      "eval_bleu": 42.1708,
+      "eval_gen_len": 36.0211,
+      "eval_loss": 0.10227888822555542,
+      "eval_runtime": 112.6072,
+      "eval_samples_per_second": 9.262,
+      "eval_steps_per_second": 0.293,
+      "step": 48000
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 1.6797367902208205e-05,
+      "loss": 0.145,
+      "step": 52000
+    },
+    {
+      "epoch": 2.56,
+      "eval_bleu": 42.623,
+      "eval_gen_len": 34.8207,
+      "eval_loss": 0.10177258402109146,
+      "eval_runtime": 127.0698,
+      "eval_samples_per_second": 8.208,
+      "eval_steps_per_second": 0.26,
+      "step": 52000
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 1.6550978410883283e-05,
+      "loss": 0.1453,
+      "step": 56000
+    },
+    {
+      "epoch": 2.76,
+      "eval_bleu": 42.4194,
+      "eval_gen_len": 35.5695,
+      "eval_loss": 0.1016233041882515,
+      "eval_runtime": 98.838,
+      "eval_samples_per_second": 10.553,
+      "eval_steps_per_second": 0.334,
+      "step": 56000
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 1.6304650532334386e-05,
+      "loss": 0.1456,
+      "step": 60000
+    },
+    {
+      "epoch": 2.96,
+      "eval_bleu": 42.7866,
+      "eval_gen_len": 34.2215,
+      "eval_loss": 0.10098095238208771,
+      "eval_runtime": 105.5395,
+      "eval_samples_per_second": 9.883,
+      "eval_steps_per_second": 0.313,
+      "step": 60000
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 1.605832265378549e-05,
+      "loss": 0.1389,
+      "step": 64000
+    },
+    {
+      "epoch": 3.15,
+      "eval_bleu": 42.2432,
+      "eval_gen_len": 34.4535,
+      "eval_loss": 0.10182780027389526,
+      "eval_runtime": 201.5414,
+      "eval_samples_per_second": 5.175,
+      "eval_steps_per_second": 0.164,
+      "step": 64000
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 1.5811994775236596e-05,
+      "loss": 0.1383,
+      "step": 68000
+    },
+    {
+      "epoch": 3.35,
+      "eval_bleu": 42.7387,
+      "eval_gen_len": 35.0662,
+      "eval_loss": 0.10195136070251465,
+      "eval_runtime": 256.6209,
+      "eval_samples_per_second": 4.064,
+      "eval_steps_per_second": 0.129,
+      "step": 68000
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 1.5565605283911674e-05,
+      "loss": 0.1371,
+      "step": 72000
+    },
+    {
+      "epoch": 3.55,
+      "eval_bleu": 42.4366,
+      "eval_gen_len": 34.7402,
+      "eval_loss": 0.1019248515367508,
+      "eval_runtime": 237.2514,
+      "eval_samples_per_second": 4.396,
+      "eval_steps_per_second": 0.139,
+      "step": 72000
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 1.5319277405362777e-05,
+      "loss": 0.1376,
+      "step": 76000
+    },
+    {
+      "epoch": 3.75,
+      "eval_bleu": 42.0722,
+      "eval_gen_len": 35.9262,
+      "eval_loss": 0.10178898274898529,
+      "eval_runtime": 101.6353,
+      "eval_samples_per_second": 10.262,
+      "eval_steps_per_second": 0.325,
+      "step": 76000
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 1.507294952681388e-05,
+      "loss": 0.1383,
+      "step": 80000
+    },
+    {
+      "epoch": 3.94,
+      "eval_bleu": 43.0447,
+      "eval_gen_len": 35.4094,
+      "eval_loss": 0.10154864192008972,
+      "eval_runtime": 86.205,
+      "eval_samples_per_second": 12.099,
+      "eval_steps_per_second": 0.383,
+      "step": 80000
+    }
+  ],
+  "max_steps": 324608,
+  "num_train_epochs": 16,
+  "total_flos": 1.3796012702564352e+17,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:06d1d3a4cefa501bd9b67bad4da16ae7e5a8dcff893cec61768fee3ede9af152
+size 3707