stage best lsp AoN adapt wce model

config.json

@@ -0,0 +1,46 @@
+{
+  "_name_or_path": "Helsinki-NLP/opus-mt-tc-big-en-fr",
+  "activation_dropout": 0.0,
+  "activation_function": "relu",
+  "architectures": [
+    "MarianMTModel"
+  ],
+  "attention_dropout": 0.0,
+  "bad_words_ids": [
+    [
+      53016
+    ]
+  ],
+  "bos_token_id": 0,
+  "classifier_dropout": 0.0,
+  "d_model": 1024,
+  "decoder_attention_heads": 16,
+  "decoder_ffn_dim": 4096,
+  "decoder_layerdrop": 0.0,
+  "decoder_layers": 6,
+  "decoder_start_token_id": 53016,
+  "decoder_vocab_size": 53017,
+  "dropout": 0.1,
+  "encoder_attention_heads": 16,
+  "encoder_ffn_dim": 4096,
+  "encoder_layerdrop": 0.0,
+  "encoder_layers": 6,
+  "eos_token_id": 43311,
+  "forced_eos_token_id": 43311,
+  "init_std": 0.02,
+  "is_encoder_decoder": true,
+  "max_length": 512,
+  "max_position_embeddings": 1024,
+  "model_type": "marian",
+  "normalize_embedding": false,
+  "num_beams": 4,
+  "num_hidden_layers": 6,
+  "pad_token_id": 53016,
+  "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": 53017
+}

generation_config.json

@@ -0,0 +1,15 @@
+{
+  "bad_words_ids": [
+    [
+      53016
+    ]
+  ],
+  "bos_token_id": 0,
+  "decoder_start_token_id": 53016,
+  "eos_token_id": 43311,
+  "forced_eos_token_id": 43311,
+  "max_length": 512,
+  "num_beams": 4,
+  "pad_token_id": 53016,
+  "transformers_version": "4.28.0"
+}

optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e2d9c5cf14b96e3eeb8390e57442b5e06ec29f8c446e4a5c974a0435a118c868
+size 1845323269

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b7401aa6b3021e5063f8103830657e5306761e37ab2a9d7683e796db7518c871
+size 922885701

rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a022812e883bcf03ab1da70f6ca505af5c5fc5c2b27d15777a6ef6325a73c762
+size 14575

scaler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:818decf194c74edc8bc9bdff7ab993c572544d39caebe4efcd82fba9a0bc0c2c
+size 557

scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f3af67343aa0f8b4109961c1ed1ad806d5b136e6b0592737a57873f77d6c9558
+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 +1,176 @@
+{
+  "best_metric": 43.7339,
+  "best_model_checkpoint": "opus_big_lsp_adapt_wce_prop_0.8_weight_1.75/checkpoint-80000",
+  "epoch": 1.9716574245224892,
+  "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": 1.9753635243376465e-05,
+      "loss": 0.1872,
+      "step": 8000
+    },
+    {
+      "epoch": 0.2,
+      "eval_bleu": 42.0446,
+      "eval_gen_len": 33.721,
+      "eval_loss": 0.09888985753059387,
+      "eval_runtime": 130.4471,
+      "eval_samples_per_second": 7.996,
+      "eval_steps_per_second": 0.506,
+      "step": 8000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.950723967960567e-05,
+      "loss": 0.1556,
+      "step": 16000
+    },
+    {
+      "epoch": 0.39,
+      "eval_bleu": 42.2512,
+      "eval_gen_len": 34.093,
+      "eval_loss": 0.09752173721790314,
+      "eval_runtime": 189.6683,
+      "eval_samples_per_second": 5.499,
+      "eval_steps_per_second": 0.348,
+      "step": 16000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.9260874922982132e-05,
+      "loss": 0.1511,
+      "step": 24000
+    },
+    {
+      "epoch": 0.59,
+      "eval_bleu": 42.4039,
+      "eval_gen_len": 33.7776,
+      "eval_loss": 0.09804832935333252,
+      "eval_runtime": 125.3237,
+      "eval_samples_per_second": 8.322,
+      "eval_steps_per_second": 0.527,
+      "step": 24000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.901447935921134e-05,
+      "loss": 0.1468,
+      "step": 32000
+    },
+    {
+      "epoch": 0.79,
+      "eval_bleu": 42.384,
+      "eval_gen_len": 34.0077,
+      "eval_loss": 0.09726251661777496,
+      "eval_runtime": 127.0486,
+      "eval_samples_per_second": 8.209,
+      "eval_steps_per_second": 0.519,
+      "step": 32000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.8768114602587803e-05,
+      "loss": 0.1453,
+      "step": 40000
+    },
+    {
+      "epoch": 0.99,
+      "eval_bleu": 42.5465,
+      "eval_gen_len": 34.2138,
+      "eval_loss": 0.09616752713918686,
+      "eval_runtime": 149.6762,
+      "eval_samples_per_second": 6.968,
+      "eval_steps_per_second": 0.441,
+      "step": 40000
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.8521749845964266e-05,
+      "loss": 0.1292,
+      "step": 48000
+    },
+    {
+      "epoch": 1.18,
+      "eval_bleu": 42.7614,
+      "eval_gen_len": 33.6558,
+      "eval_loss": 0.09714562445878983,
+      "eval_runtime": 140.3139,
+      "eval_samples_per_second": 7.433,
+      "eval_steps_per_second": 0.47,
+      "step": 48000
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 1.827535428219347e-05,
+      "loss": 0.1296,
+      "step": 56000
+    },
+    {
+      "epoch": 1.38,
+      "eval_bleu": 42.2625,
+      "eval_gen_len": 34.7987,
+      "eval_loss": 0.09766771644353867,
+      "eval_runtime": 149.3991,
+      "eval_samples_per_second": 6.981,
+      "eval_steps_per_second": 0.442,
+      "step": 56000
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 1.8029020332717192e-05,
+      "loss": 0.1301,
+      "step": 64000
+    },
+    {
+      "epoch": 1.58,
+      "eval_bleu": 42.8366,
+      "eval_gen_len": 34.4842,
+      "eval_loss": 0.0971095860004425,
+      "eval_runtime": 150.9733,
+      "eval_samples_per_second": 6.909,
+      "eval_steps_per_second": 0.437,
+      "step": 64000
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 1.7782624768946397e-05,
+      "loss": 0.1295,
+      "step": 72000
+    },
+    {
+      "epoch": 1.77,
+      "eval_bleu": 42.763,
+      "eval_gen_len": 34.9012,
+      "eval_loss": 0.09676354378461838,
+      "eval_runtime": 156.8388,
+      "eval_samples_per_second": 6.65,
+      "eval_steps_per_second": 0.421,
+      "step": 72000
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 1.753626001232286e-05,
+      "loss": 0.1293,
+      "step": 80000
+    },
+    {
+      "epoch": 1.97,
+      "eval_bleu": 43.7339,
+      "eval_gen_len": 33.8734,
+      "eval_loss": 0.09614978730678558,
+      "eval_runtime": 162.1547,
+      "eval_samples_per_second": 6.432,
+      "eval_steps_per_second": 0.407,
+      "step": 80000
+    }
+  ],
+  "max_steps": 649200,
+  "num_train_epochs": 16,
+  "total_flos": 2.5844916958632346e+17,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9ddb2e653d6bb7b34bc42ef87da80ce3fb34338c6a84b6eb1646965d227110d5
+size 3771