stage best fine adapt 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 sha256:93595676d8b24d78e575613b455f5460523492d2976714a66e00929f7aeea98e
+size 597026309

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:23e5a72cd6396310bc7bda77506254af1815b47ed09f43578c68d228c9c4fda6
+size 298763205

rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3c62f554b1b54e66f077e6696a34b63e270d8326cbf9f7ed0bb586d65b305862
+size 14575

scaler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e5b0045b87a91e562ab8da82580ad0c1b3dc08b5418137b79afdac0a063db7bc
+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": 42.9095,
+  "best_model_checkpoint": "opus_base_adapt_wce_gloss_train-sampled_precision_3_ubweight_2.0/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": 0.213,
+      "step": 4000
+    },
+    {
+      "epoch": 0.2,
+      "eval_bleu": 32.1992,
+      "eval_gen_len": 43.558,
+      "eval_loss": 0.10932592302560806,
+      "eval_runtime": 185.4985,
+      "eval_samples_per_second": 5.623,
+      "eval_steps_per_second": 0.178,
+      "step": 4000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.950746746845426e-05,
+      "loss": 0.1788,
+      "step": 8000
+    },
+    {
+      "epoch": 0.39,
+      "eval_bleu": 40.1195,
+      "eval_gen_len": 36.4458,
+      "eval_loss": 0.10684426873922348,
+      "eval_runtime": 173.1992,
+      "eval_samples_per_second": 6.022,
+      "eval_steps_per_second": 0.191,
+      "step": 8000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.9261016364353314e-05,
+      "loss": 0.1716,
+      "step": 12000
+    },
+    {
+      "epoch": 0.59,
+      "eval_bleu": 41.5118,
+      "eval_gen_len": 34.5973,
+      "eval_loss": 0.1055770218372345,
+      "eval_runtime": 177.753,
+      "eval_samples_per_second": 5.868,
+      "eval_steps_per_second": 0.186,
+      "step": 12000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.9014565260252367e-05,
+      "loss": 0.1659,
+      "step": 16000
+    },
+    {
+      "epoch": 0.79,
+      "eval_bleu": 41.3846,
+      "eval_gen_len": 34.2397,
+      "eval_loss": 0.10453091561794281,
+      "eval_runtime": 131.5252,
+      "eval_samples_per_second": 7.93,
+      "eval_steps_per_second": 0.251,
+      "step": 16000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.876811415615142e-05,
+      "loss": 0.1634,
+      "step": 20000
+    },
+    {
+      "epoch": 0.99,
+      "eval_bleu": 41.5931,
+      "eval_gen_len": 35.1074,
+      "eval_loss": 0.10396245121955872,
+      "eval_runtime": 178.097,
+      "eval_samples_per_second": 5.856,
+      "eval_steps_per_second": 0.185,
+      "step": 20000
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.8521786277602524e-05,
+      "loss": 0.1513,
+      "step": 24000
+    },
+    {
+      "epoch": 1.18,
+      "eval_bleu": 41.2579,
+      "eval_gen_len": 36.1802,
+      "eval_loss": 0.10382075607776642,
+      "eval_runtime": 152.5475,
+      "eval_samples_per_second": 6.837,
+      "eval_steps_per_second": 0.216,
+      "step": 24000
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 1.8275520011829655e-05,
+      "loss": 0.151,
+      "step": 28000
+    },
+    {
+      "epoch": 1.38,
+      "eval_bleu": 40.956,
+      "eval_gen_len": 36.1687,
+      "eval_loss": 0.10313059389591217,
+      "eval_runtime": 206.7763,
+      "eval_samples_per_second": 5.044,
+      "eval_steps_per_second": 0.16,
+      "step": 28000
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 1.8029130520504733e-05,
+      "loss": 0.1506,
+      "step": 32000
+    },
+    {
+      "epoch": 1.58,
+      "eval_bleu": 41.8294,
+      "eval_gen_len": 35.3432,
+      "eval_loss": 0.10248984396457672,
+      "eval_runtime": 140.6628,
+      "eval_samples_per_second": 7.415,
+      "eval_steps_per_second": 0.235,
+      "step": 32000
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.7782741029179815e-05,
+      "loss": 0.149,
+      "step": 36000
+    },
+    {
+      "epoch": 1.77,
+      "eval_bleu": 40.6541,
+      "eval_gen_len": 37.2934,
+      "eval_loss": 0.10204267501831055,
+      "eval_runtime": 116.2311,
+      "eval_samples_per_second": 8.973,
+      "eval_steps_per_second": 0.284,
+      "step": 36000
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.7536413150630914e-05,
+      "loss": 0.1482,
+      "step": 40000
+    },
+    {
+      "epoch": 1.97,
+      "eval_bleu": 42.167,
+      "eval_gen_len": 35.0,
+      "eval_loss": 0.10223417729139328,
+      "eval_runtime": 120.9127,
+      "eval_samples_per_second": 8.626,
+      "eval_steps_per_second": 0.273,
+      "step": 40000
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1.7290023659305996e-05,
+      "loss": 0.1404,
+      "step": 44000
+    },
+    {
+      "epoch": 2.17,
+      "eval_bleu": 42.2093,
+      "eval_gen_len": 35.4257,
+      "eval_loss": 0.10255546122789383,
+      "eval_runtime": 152.5867,
+      "eval_samples_per_second": 6.835,
+      "eval_steps_per_second": 0.216,
+      "step": 44000
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 1.7043634167981074e-05,
+      "loss": 0.1394,
+      "step": 48000
+    },
+    {
+      "epoch": 2.37,
+      "eval_bleu": 42.2705,
+      "eval_gen_len": 35.6999,
+      "eval_loss": 0.10232062637805939,
+      "eval_runtime": 127.5651,
+      "eval_samples_per_second": 8.176,
+      "eval_steps_per_second": 0.259,
+      "step": 48000
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 1.6797306289432177e-05,
+      "loss": 0.1384,
+      "step": 52000
+    },
+    {
+      "epoch": 2.56,
+      "eval_bleu": 42.0109,
+      "eval_gen_len": 35.4017,
+      "eval_loss": 0.10198231786489487,
+      "eval_runtime": 144.4455,
+      "eval_samples_per_second": 7.221,
+      "eval_steps_per_second": 0.228,
+      "step": 52000
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 1.6550978410883283e-05,
+      "loss": 0.1386,
+      "step": 56000
+    },
+    {
+      "epoch": 2.76,
+      "eval_bleu": 41.651,
+      "eval_gen_len": 36.2234,
+      "eval_loss": 0.10184463113546371,
+      "eval_runtime": 164.0439,
+      "eval_samples_per_second": 6.358,
+      "eval_steps_per_second": 0.201,
+      "step": 56000
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 1.6304650532334386e-05,
+      "loss": 0.139,
+      "step": 60000
+    },
+    {
+      "epoch": 2.96,
+      "eval_bleu": 42.5275,
+      "eval_gen_len": 34.4708,
+      "eval_loss": 0.10123319178819656,
+      "eval_runtime": 170.2072,
+      "eval_samples_per_second": 6.128,
+      "eval_steps_per_second": 0.194,
+      "step": 60000
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 1.605832265378549e-05,
+      "loss": 0.1323,
+      "step": 64000
+    },
+    {
+      "epoch": 3.15,
+      "eval_bleu": 42.2691,
+      "eval_gen_len": 35.162,
+      "eval_loss": 0.10197851061820984,
+      "eval_runtime": 153.8385,
+      "eval_samples_per_second": 6.78,
+      "eval_steps_per_second": 0.215,
+      "step": 64000
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 1.581205638801262e-05,
+      "loss": 0.1318,
+      "step": 68000
+    },
+    {
+      "epoch": 3.35,
+      "eval_bleu": 42.6981,
+      "eval_gen_len": 34.2464,
+      "eval_loss": 0.10211456567049026,
+      "eval_runtime": 119.2355,
+      "eval_samples_per_second": 8.747,
+      "eval_steps_per_second": 0.277,
+      "step": 68000
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 1.5565605283911674e-05,
+      "loss": 0.1306,
+      "step": 72000
+    },
+    {
+      "epoch": 3.55,
+      "eval_bleu": 42.3809,
+      "eval_gen_len": 34.6769,
+      "eval_loss": 0.10217085480690002,
+      "eval_runtime": 86.738,
+      "eval_samples_per_second": 12.025,
+      "eval_steps_per_second": 0.38,
+      "step": 72000
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 1.5319277405362777e-05,
+      "loss": 0.131,
+      "step": 76000
+    },
+    {
+      "epoch": 3.75,
+      "eval_bleu": 42.1038,
+      "eval_gen_len": 35.7728,
+      "eval_loss": 0.10192063450813293,
+      "eval_runtime": 96.8001,
+      "eval_samples_per_second": 10.775,
+      "eval_steps_per_second": 0.341,
+      "step": 76000
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 1.507294952681388e-05,
+      "loss": 0.1318,
+      "step": 80000
+    },
+    {
+      "epoch": 3.94,
+      "eval_bleu": 42.9095,
+      "eval_gen_len": 34.2752,
+      "eval_loss": 0.10165005922317505,
+      "eval_runtime": 104.8741,
+      "eval_samples_per_second": 9.945,
+      "eval_steps_per_second": 0.315,
+      "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 sha256:9a23313e5ab4b229189fc5d81a2c07e352d926b9425711750f3cd2e893dc0c4d
+size 3835