stage big adaptified ailem model

checkpoint-64000/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
+}

checkpoint-64000/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"
+}

checkpoint-64000/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e23527d3e6085f5b1f72f17ff67c59e3d24619bedc2b018e148f588f2835271f
+size 1845322757

checkpoint-64000/pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e127e3805356640c049225dbd08f800928ef2a6de2d22ca18686b4388476f36
+size 922885701

checkpoint-64000/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2097116f9a2287a101ed1a9d5ff5284f2668afa3bcc98cffda15f84f65d21e18
+size 14575

checkpoint-64000/scaler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aefeed8e05c0c2e9dd3269666bf2fa1417d483998e4c87caf6857eb67eb353fd
+size 557

checkpoint-64000/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:527cc6e9e71118c8787d1808d380f2cee57617f6a6298a8d7fa302bbea027845
+size 627

checkpoint-64000/special_tokens_map.json

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

checkpoint-64000/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>"
+}

checkpoint-64000/trainer_state.json

@@ -0,0 +1,144 @@
+{
+  "best_metric": 43.0883,
+  "best_model_checkpoint": "opus_big_ailem_adaptified/checkpoint-64000",
+  "epoch": 1.5773259396179915,
+  "global_step": 64000,
+  "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": 0.1745,
+      "step": 8000
+    },
+    {
+      "epoch": 0.2,
+      "eval_bleu": 42.0213,
+      "eval_gen_len": 33.5187,
+      "eval_loss": 0.0983896404504776,
+      "eval_runtime": 139.2335,
+      "eval_samples_per_second": 7.491,
+      "eval_steps_per_second": 0.474,
+      "step": 8000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 1.9507270486752928e-05,
+      "loss": 0.144,
+      "step": 16000
+    },
+    {
+      "epoch": 0.39,
+      "eval_bleu": 42.5956,
+      "eval_gen_len": 34.3337,
+      "eval_loss": 0.09711522608995438,
+      "eval_runtime": 204.7182,
+      "eval_samples_per_second": 5.095,
+      "eval_steps_per_second": 0.322,
+      "step": 16000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 1.9260874922982132e-05,
+      "loss": 0.14,
+      "step": 24000
+    },
+    {
+      "epoch": 0.59,
+      "eval_bleu": 42.4234,
+      "eval_gen_len": 33.6222,
+      "eval_loss": 0.09710835665464401,
+      "eval_runtime": 146.0195,
+      "eval_samples_per_second": 7.143,
+      "eval_steps_per_second": 0.452,
+      "step": 24000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 1.90145101663586e-05,
+      "loss": 0.1363,
+      "step": 32000
+    },
+    {
+      "epoch": 0.79,
+      "eval_bleu": 43.0677,
+      "eval_gen_len": 33.6654,
+      "eval_loss": 0.0967707633972168,
+      "eval_runtime": 131.5813,
+      "eval_samples_per_second": 7.927,
+      "eval_steps_per_second": 0.502,
+      "step": 32000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 1.876814540973506e-05,
+      "loss": 0.1349,
+      "step": 40000
+    },
+    {
+      "epoch": 0.99,
+      "eval_bleu": 42.7333,
+      "eval_gen_len": 34.8533,
+      "eval_loss": 0.09592821449041367,
+      "eval_runtime": 131.6747,
+      "eval_samples_per_second": 7.921,
+      "eval_steps_per_second": 0.501,
+      "step": 40000
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 1.8521749845964266e-05,
+      "loss": 0.1201,
+      "step": 48000
+    },
+    {
+      "epoch": 1.18,
+      "eval_bleu": 42.8479,
+      "eval_gen_len": 34.2771,
+      "eval_loss": 0.09712815284729004,
+      "eval_runtime": 126.0283,
+      "eval_samples_per_second": 8.276,
+      "eval_steps_per_second": 0.524,
+      "step": 48000
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 1.827535428219347e-05,
+      "loss": 0.1205,
+      "step": 56000
+    },
+    {
+      "epoch": 1.38,
+      "eval_bleu": 42.4623,
+      "eval_gen_len": 34.1122,
+      "eval_loss": 0.09739168733358383,
+      "eval_runtime": 191.9796,
+      "eval_samples_per_second": 5.433,
+      "eval_steps_per_second": 0.344,
+      "step": 56000
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 1.8029020332717192e-05,
+      "loss": 0.1211,
+      "step": 64000
+    },
+    {
+      "epoch": 1.58,
+      "eval_bleu": 43.0883,
+      "eval_gen_len": 34.9319,
+      "eval_loss": 0.09650042653083801,
+      "eval_runtime": 182.5666,
+      "eval_samples_per_second": 5.713,
+      "eval_steps_per_second": 0.362,
+      "step": 64000
+    }
+  ],
+  "max_steps": 649200,
+  "num_train_epochs": 16,
+  "total_flos": 2.0676742327723622e+17,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-64000/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b965fb713e6c062170dac4d0b33e4e1548a05b1f8ed80def4df395181fb884c6
+size 3771

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"
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e127e3805356640c049225dbd08f800928ef2a6de2d22ca18686b4388476f36
+size 922885701

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>"
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b965fb713e6c062170dac4d0b33e4e1548a05b1f8ed80def4df395181fb884c6
+size 3771