stage best lsp simple 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:d78949a43a5015ccb761b5ade9a80839fa836ac680ab1f7de7a3b9a9228974c1
+size 597026309

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a655f4df2c814359a00d2ae307fc1fe1367b6e133470954e1d3c75120ca9ecf6
+size 298763205

rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c25017c70060b8a1e7f6774aa8104ab431f1ad224770f47902686ae1247d8ef0
+size 14575

scaler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6c19b9e1be4a9fb81bbf480f3474b548d78634c75ab4104d0995e29d770710d6
+size 557

scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1378d1da724b1a77da62e79047ad80c84a64145612cd1941ad7ce9dd55ac66d8
+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": 43.1111,
+  "best_model_checkpoint": "opus_base_lsp_adapt_wce_bands_5_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": 0.2142,
+      "step": 4000
+    },
+    {
+      "epoch": 0.2,
+      "eval_bleu": 32.7905,
+      "eval_gen_len": 43.1007,
+      "eval_loss": 0.10939062386751175,
+      "eval_runtime": 145.1214,
+      "eval_samples_per_second": 7.187,
+      "eval_steps_per_second": 0.227,
+      "step": 4000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.950746746845426e-05,
+      "loss": 0.18,
+      "step": 8000
+    },
+    {
+      "epoch": 0.39,
+      "eval_bleu": 36.4262,
+      "eval_gen_len": 39.3643,
+      "eval_loss": 0.10692563652992249,
+      "eval_runtime": 115.1385,
+      "eval_samples_per_second": 9.059,
+      "eval_steps_per_second": 0.287,
+      "step": 8000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.9261016364353314e-05,
+      "loss": 0.1726,
+      "step": 12000
+    },
+    {
+      "epoch": 0.59,
+      "eval_bleu": 41.077,
+      "eval_gen_len": 35.0671,
+      "eval_loss": 0.10563468933105469,
+      "eval_runtime": 127.7336,
+      "eval_samples_per_second": 8.165,
+      "eval_steps_per_second": 0.258,
+      "step": 12000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.9014565260252367e-05,
+      "loss": 0.1669,
+      "step": 16000
+    },
+    {
+      "epoch": 0.79,
+      "eval_bleu": 41.2349,
+      "eval_gen_len": 35.2464,
+      "eval_loss": 0.10455626994371414,
+      "eval_runtime": 88.8086,
+      "eval_samples_per_second": 11.744,
+      "eval_steps_per_second": 0.372,
+      "step": 16000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.876811415615142e-05,
+      "loss": 0.1644,
+      "step": 20000
+    },
+    {
+      "epoch": 0.99,
+      "eval_bleu": 41.641,
+      "eval_gen_len": 35.3337,
+      "eval_loss": 0.10399404913187027,
+      "eval_runtime": 137.4077,
+      "eval_samples_per_second": 7.591,
+      "eval_steps_per_second": 0.24,
+      "step": 20000
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.8521786277602524e-05,
+      "loss": 0.1522,
+      "step": 24000
+    },
+    {
+      "epoch": 1.18,
+      "eval_bleu": 41.4007,
+      "eval_gen_len": 35.7028,
+      "eval_loss": 0.10385829210281372,
+      "eval_runtime": 141.3172,
+      "eval_samples_per_second": 7.381,
+      "eval_steps_per_second": 0.234,
+      "step": 24000
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 1.8275520011829655e-05,
+      "loss": 0.152,
+      "step": 28000
+    },
+    {
+      "epoch": 1.38,
+      "eval_bleu": 41.7231,
+      "eval_gen_len": 34.8265,
+      "eval_loss": 0.10312652587890625,
+      "eval_runtime": 99.4665,
+      "eval_samples_per_second": 10.486,
+      "eval_steps_per_second": 0.332,
+      "step": 28000
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 1.8029130520504733e-05,
+      "loss": 0.1515,
+      "step": 32000
+    },
+    {
+      "epoch": 1.58,
+      "eval_bleu": 42.1288,
+      "eval_gen_len": 35.4545,
+      "eval_loss": 0.10245882719755173,
+      "eval_runtime": 82.6,
+      "eval_samples_per_second": 12.627,
+      "eval_steps_per_second": 0.4,
+      "step": 32000
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.7782741029179815e-05,
+      "loss": 0.1499,
+      "step": 36000
+    },
+    {
+      "epoch": 1.77,
+      "eval_bleu": 39.4245,
+      "eval_gen_len": 38.9329,
+      "eval_loss": 0.10212080180644989,
+      "eval_runtime": 114.4105,
+      "eval_samples_per_second": 9.116,
+      "eval_steps_per_second": 0.288,
+      "step": 36000
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.7536413150630914e-05,
+      "loss": 0.1491,
+      "step": 40000
+    },
+    {
+      "epoch": 1.97,
+      "eval_bleu": 42.32,
+      "eval_gen_len": 34.3471,
+      "eval_loss": 0.10226055979728699,
+      "eval_runtime": 99.1685,
+      "eval_samples_per_second": 10.517,
+      "eval_steps_per_second": 0.333,
+      "step": 40000
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 1.729008527208202e-05,
+      "loss": 0.1413,
+      "step": 44000
+    },
+    {
+      "epoch": 2.17,
+      "eval_bleu": 40.4318,
+      "eval_gen_len": 36.4919,
+      "eval_loss": 0.1026298999786377,
+      "eval_runtime": 116.9488,
+      "eval_samples_per_second": 8.918,
+      "eval_steps_per_second": 0.282,
+      "step": 44000
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 1.70436957807571e-05,
+      "loss": 0.1404,
+      "step": 48000
+    },
+    {
+      "epoch": 2.37,
+      "eval_bleu": 42.4697,
+      "eval_gen_len": 35.14,
+      "eval_loss": 0.10236479341983795,
+      "eval_runtime": 95.0688,
+      "eval_samples_per_second": 10.971,
+      "eval_steps_per_second": 0.347,
+      "step": 48000
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 1.679742951498423e-05,
+      "loss": 0.1392,
+      "step": 52000
+    },
+    {
+      "epoch": 2.56,
+      "eval_bleu": 42.1067,
+      "eval_gen_len": 34.4813,
+      "eval_loss": 0.10209206491708755,
+      "eval_runtime": 107.2908,
+      "eval_samples_per_second": 9.721,
+      "eval_steps_per_second": 0.308,
+      "step": 52000
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 1.6551040023659308e-05,
+      "loss": 0.1396,
+      "step": 56000
+    },
+    {
+      "epoch": 2.76,
+      "eval_bleu": 41.9687,
+      "eval_gen_len": 35.069,
+      "eval_loss": 0.10199315845966339,
+      "eval_runtime": 123.666,
+      "eval_samples_per_second": 8.434,
+      "eval_steps_per_second": 0.267,
+      "step": 56000
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 1.6304650532334386e-05,
+      "loss": 0.1398,
+      "step": 60000
+    },
+    {
+      "epoch": 2.96,
+      "eval_bleu": 42.5934,
+      "eval_gen_len": 34.4506,
+      "eval_loss": 0.10132639110088348,
+      "eval_runtime": 87.6332,
+      "eval_samples_per_second": 11.902,
+      "eval_steps_per_second": 0.377,
+      "step": 60000
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 1.605832265378549e-05,
+      "loss": 0.1332,
+      "step": 64000
+    },
+    {
+      "epoch": 3.15,
+      "eval_bleu": 42.1707,
+      "eval_gen_len": 34.278,
+      "eval_loss": 0.10205266624689102,
+      "eval_runtime": 75.2908,
+      "eval_samples_per_second": 13.853,
+      "eval_steps_per_second": 0.438,
+      "step": 64000
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 1.5811994775236596e-05,
+      "loss": 0.1327,
+      "step": 68000
+    },
+    {
+      "epoch": 3.35,
+      "eval_bleu": 42.4582,
+      "eval_gen_len": 34.0326,
+      "eval_loss": 0.10221361368894577,
+      "eval_runtime": 85.03,
+      "eval_samples_per_second": 12.266,
+      "eval_steps_per_second": 0.388,
+      "step": 68000
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 1.55656668966877e-05,
+      "loss": 0.1314,
+      "step": 72000
+    },
+    {
+      "epoch": 3.55,
+      "eval_bleu": 42.0756,
+      "eval_gen_len": 34.6932,
+      "eval_loss": 0.10215744376182556,
+      "eval_runtime": 85.9762,
+      "eval_samples_per_second": 12.131,
+      "eval_steps_per_second": 0.384,
+      "step": 72000
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 1.5319339018138805e-05,
+      "loss": 0.1319,
+      "step": 76000
+    },
+    {
+      "epoch": 3.75,
+      "eval_bleu": 42.4042,
+      "eval_gen_len": 34.7958,
+      "eval_loss": 0.10203341394662857,
+      "eval_runtime": 88.8632,
+      "eval_samples_per_second": 11.737,
+      "eval_steps_per_second": 0.371,
+      "step": 76000
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 1.5073011139589906e-05,
+      "loss": 0.1327,
+      "step": 80000
+    },
+    {
+      "epoch": 3.94,
+      "eval_bleu": 43.1111,
+      "eval_gen_len": 34.6481,
+      "eval_loss": 0.10178661346435547,
+      "eval_runtime": 99.1474,
+      "eval_samples_per_second": 10.52,
+      "eval_steps_per_second": 0.333,
+      "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 sha256:3d0c9bc1e2fddfe046f1ac2f45049ae2699b59aa26664e5ff16c8e4b40205ce1
+size 3771