Upload folder using huggingface_hub
Browse filesThis view is limited to 50 files because it contains too many changes.
See raw diff
- .gitattributes +5 -0
- README.md +75 -0
- all_results.json +22 -0
- checkpoint-129072/config.json +59 -0
- checkpoint-129072/generation_config.json +12 -0
- checkpoint-129072/model.safetensors +3 -0
- checkpoint-129072/optimizer.pt +3 -0
- checkpoint-129072/rng_state_0.pth +3 -0
- checkpoint-129072/rng_state_1.pth +3 -0
- checkpoint-129072/rng_state_2.pth +3 -0
- checkpoint-129072/rng_state_3.pth +3 -0
- checkpoint-129072/scheduler.pt +3 -0
- checkpoint-129072/sentencepiece.bpe.model +3 -0
- checkpoint-129072/special_tokens_map.json +69 -0
- checkpoint-129072/tokenizer.json +3 -0
- checkpoint-129072/tokenizer_config.json +528 -0
- checkpoint-129072/trainer_state.json +1599 -0
- checkpoint-129072/training_args.bin +3 -0
- checkpoint-215120/config.json +59 -0
- checkpoint-215120/generation_config.json +12 -0
- checkpoint-215120/model.safetensors +3 -0
- checkpoint-215120/optimizer.pt +3 -0
- checkpoint-215120/rng_state_0.pth +3 -0
- checkpoint-215120/rng_state_1.pth +3 -0
- checkpoint-215120/rng_state_2.pth +3 -0
- checkpoint-215120/rng_state_3.pth +3 -0
- checkpoint-215120/scheduler.pt +3 -0
- checkpoint-215120/sentencepiece.bpe.model +3 -0
- checkpoint-215120/special_tokens_map.json +69 -0
- checkpoint-215120/tokenizer.json +3 -0
- checkpoint-215120/tokenizer_config.json +528 -0
- checkpoint-215120/trainer_state.json +2651 -0
- checkpoint-215120/training_args.bin +3 -0
- checkpoint-258144/config.json +59 -0
- checkpoint-258144/generation_config.json +12 -0
- checkpoint-258144/model.safetensors +3 -0
- checkpoint-258144/optimizer.pt +3 -0
- checkpoint-258144/rng_state_0.pth +3 -0
- checkpoint-258144/rng_state_1.pth +3 -0
- checkpoint-258144/rng_state_2.pth +3 -0
- checkpoint-258144/rng_state_3.pth +3 -0
- checkpoint-258144/scheduler.pt +3 -0
- checkpoint-258144/sentencepiece.bpe.model +3 -0
- checkpoint-258144/special_tokens_map.json +69 -0
- checkpoint-258144/tokenizer.json +3 -0
- checkpoint-258144/tokenizer_config.json +528 -0
- checkpoint-258144/trainer_state.json +3177 -0
- checkpoint-258144/training_args.bin +3 -0
- config.json +59 -0
- eval_results.json +10 -0
.gitattributes
CHANGED
|
@@ -33,3 +33,8 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
+
checkpoint-129072/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 37 |
+
checkpoint-215120/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 38 |
+
checkpoint-258144/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 39 |
+
generated_predictions.txt filter=lfs diff=lfs merge=lfs -text
|
| 40 |
+
tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
README.md
ADDED
|
@@ -0,0 +1,75 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- ja
|
| 4 |
+
- ko
|
| 5 |
+
base_model: facebook/mbart-large-50-many-to-many-mmt
|
| 6 |
+
tags:
|
| 7 |
+
- generated_from_trainer
|
| 8 |
+
metrics:
|
| 9 |
+
- bleu
|
| 10 |
+
model-index:
|
| 11 |
+
- name: finetuned_model
|
| 12 |
+
results: []
|
| 13 |
+
---
|
| 14 |
+
|
| 15 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 16 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 17 |
+
|
| 18 |
+
# finetuned_model
|
| 19 |
+
|
| 20 |
+
This model is a fine-tuned version of [facebook/mbart-large-50-many-to-many-mmt](https://huggingface.co/facebook/mbart-large-50-many-to-many-mmt) on an unknown dataset.
|
| 21 |
+
It achieves the following results on the evaluation set:
|
| 22 |
+
- Loss: 0.6568
|
| 23 |
+
- Bleu: 58.837
|
| 24 |
+
- Gen Len: 15.7689
|
| 25 |
+
|
| 26 |
+
## Model description
|
| 27 |
+
|
| 28 |
+
More information needed
|
| 29 |
+
|
| 30 |
+
## Intended uses & limitations
|
| 31 |
+
|
| 32 |
+
More information needed
|
| 33 |
+
|
| 34 |
+
## Training and evaluation data
|
| 35 |
+
|
| 36 |
+
More information needed
|
| 37 |
+
|
| 38 |
+
## Training procedure
|
| 39 |
+
|
| 40 |
+
### Training hyperparameters
|
| 41 |
+
|
| 42 |
+
The following hyperparameters were used during training:
|
| 43 |
+
- learning_rate: 5e-05
|
| 44 |
+
- train_batch_size: 2
|
| 45 |
+
- eval_batch_size: 2
|
| 46 |
+
- seed: 42
|
| 47 |
+
- distributed_type: multi-GPU
|
| 48 |
+
- num_devices: 4
|
| 49 |
+
- gradient_accumulation_steps: 2
|
| 50 |
+
- total_train_batch_size: 16
|
| 51 |
+
- total_eval_batch_size: 8
|
| 52 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 53 |
+
- lr_scheduler_type: linear
|
| 54 |
+
- lr_scheduler_warmup_steps: 2500
|
| 55 |
+
- num_epochs: 15
|
| 56 |
+
- mixed_precision_training: Native AMP
|
| 57 |
+
|
| 58 |
+
### Training results
|
| 59 |
+
|
| 60 |
+
| Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len |
|
| 61 |
+
|:-------------:|:-----:|:------:|:---------------:|:-------:|:-------:|
|
| 62 |
+
| 0.7629 | 1.0 | 43024 | 0.7231 | 54.9608 | 16.1273 |
|
| 63 |
+
| 0.6355 | 2.0 | 86048 | 0.6641 | 57.5735 | 15.9938 |
|
| 64 |
+
| 0.5464 | 3.0 | 129072 | 0.6568 | 58.837 | 15.7689 |
|
| 65 |
+
| 0.4444 | 4.0 | 172096 | 0.6777 | 59.3055 | 15.7545 |
|
| 66 |
+
| 0.3458 | 5.0 | 215120 | 0.7215 | 59.1341 | 15.7016 |
|
| 67 |
+
| 0.2702 | 6.0 | 258144 | 0.8013 | 59.0721 | 15.6936 |
|
| 68 |
+
|
| 69 |
+
|
| 70 |
+
### Framework versions
|
| 71 |
+
|
| 72 |
+
- Transformers 4.36.2
|
| 73 |
+
- Pytorch 2.1.2+cu121
|
| 74 |
+
- Datasets 2.16.1
|
| 75 |
+
- Tokenizers 0.15.0
|
all_results.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 6.0,
|
| 3 |
+
"eval_bleu": 58.837,
|
| 4 |
+
"eval_gen_len": 15.7689,
|
| 5 |
+
"eval_loss": 0.6568061113357544,
|
| 6 |
+
"eval_runtime": 6763.7413,
|
| 7 |
+
"eval_samples": 85914,
|
| 8 |
+
"eval_samples_per_second": 12.702,
|
| 9 |
+
"eval_steps_per_second": 1.588,
|
| 10 |
+
"predict_bleu": 59.427,
|
| 11 |
+
"predict_gen_len": 16.2418,
|
| 12 |
+
"predict_loss": 0.6393086910247803,
|
| 13 |
+
"predict_runtime": 7570.8911,
|
| 14 |
+
"predict_samples": 94835,
|
| 15 |
+
"predict_samples_per_second": 12.526,
|
| 16 |
+
"predict_steps_per_second": 1.566,
|
| 17 |
+
"train_loss": 0.5181208364993382,
|
| 18 |
+
"train_runtime": 241803.971,
|
| 19 |
+
"train_samples": 688378,
|
| 20 |
+
"train_samples_per_second": 42.703,
|
| 21 |
+
"train_steps_per_second": 2.669
|
| 22 |
+
}
|
checkpoint-129072/config.json
ADDED
|
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/mbart-large-50-many-to-many-mmt",
|
| 3 |
+
"_num_labels": 3,
|
| 4 |
+
"activation_dropout": 0.0,
|
| 5 |
+
"activation_function": "relu",
|
| 6 |
+
"add_bias_logits": false,
|
| 7 |
+
"add_final_layer_norm": true,
|
| 8 |
+
"architectures": [
|
| 9 |
+
"MBartForConditionalGeneration"
|
| 10 |
+
],
|
| 11 |
+
"attention_dropout": 0.0,
|
| 12 |
+
"bos_token_id": 0,
|
| 13 |
+
"classif_dropout": 0.0,
|
| 14 |
+
"classifier_dropout": 0.0,
|
| 15 |
+
"d_model": 1024,
|
| 16 |
+
"decoder_attention_heads": 16,
|
| 17 |
+
"decoder_ffn_dim": 4096,
|
| 18 |
+
"decoder_layerdrop": 0.0,
|
| 19 |
+
"decoder_layers": 12,
|
| 20 |
+
"decoder_start_token_id": 2,
|
| 21 |
+
"dropout": 0.1,
|
| 22 |
+
"early_stopping": true,
|
| 23 |
+
"encoder_attention_heads": 16,
|
| 24 |
+
"encoder_ffn_dim": 4096,
|
| 25 |
+
"encoder_layerdrop": 0.0,
|
| 26 |
+
"encoder_layers": 12,
|
| 27 |
+
"eos_token_id": 2,
|
| 28 |
+
"forced_bos_token_id": 250014,
|
| 29 |
+
"forced_eos_token_id": 2,
|
| 30 |
+
"gradient_checkpointing": false,
|
| 31 |
+
"id2label": {
|
| 32 |
+
"0": "LABEL_0",
|
| 33 |
+
"1": "LABEL_1",
|
| 34 |
+
"2": "LABEL_2"
|
| 35 |
+
},
|
| 36 |
+
"init_std": 0.02,
|
| 37 |
+
"is_encoder_decoder": true,
|
| 38 |
+
"label2id": {
|
| 39 |
+
"LABEL_0": 0,
|
| 40 |
+
"LABEL_1": 1,
|
| 41 |
+
"LABEL_2": 2
|
| 42 |
+
},
|
| 43 |
+
"max_length": 200,
|
| 44 |
+
"max_position_embeddings": 1024,
|
| 45 |
+
"model_type": "mbart",
|
| 46 |
+
"normalize_before": true,
|
| 47 |
+
"normalize_embedding": true,
|
| 48 |
+
"num_beams": 5,
|
| 49 |
+
"num_hidden_layers": 12,
|
| 50 |
+
"output_past": true,
|
| 51 |
+
"pad_token_id": 1,
|
| 52 |
+
"scale_embedding": true,
|
| 53 |
+
"static_position_embeddings": false,
|
| 54 |
+
"tokenizer_class": "MBart50Tokenizer",
|
| 55 |
+
"torch_dtype": "float32",
|
| 56 |
+
"transformers_version": "4.36.2",
|
| 57 |
+
"use_cache": true,
|
| 58 |
+
"vocab_size": 250054
|
| 59 |
+
}
|
checkpoint-129072/generation_config.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token_id": 0,
|
| 3 |
+
"decoder_start_token_id": 2,
|
| 4 |
+
"early_stopping": true,
|
| 5 |
+
"eos_token_id": 2,
|
| 6 |
+
"forced_bos_token_id": 250014,
|
| 7 |
+
"forced_eos_token_id": 2,
|
| 8 |
+
"max_length": 200,
|
| 9 |
+
"num_beams": 5,
|
| 10 |
+
"pad_token_id": 1,
|
| 11 |
+
"transformers_version": "4.36.2"
|
| 12 |
+
}
|
checkpoint-129072/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:62328af4d78edcc56d6883277a8a62129240f38708b10d9033d60a7fadb62a10
|
| 3 |
+
size 2444578688
|
checkpoint-129072/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:42238eb399f50913c842ad933ce2e82cd6d607997003a4fa2e938fe2ba8d010c
|
| 3 |
+
size 4887473903
|
checkpoint-129072/rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:61444aa288d144d4687f4bee7e86d2dc3027755d4151c1647c073a12446ac434
|
| 3 |
+
size 15024
|
checkpoint-129072/rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7feda2497c5b39d1447f284359b50321f347ffb446fcdbd25b3ffe93512a4929
|
| 3 |
+
size 15024
|
checkpoint-129072/rng_state_2.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8b27ad8953ef1ca1e4a86380932d8a5409f0d812407f890451cdfa1d8d259812
|
| 3 |
+
size 15024
|
checkpoint-129072/rng_state_3.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dd93898c44057e8ced2f419868b6decfc5b7c49726a795ae4ef1caea3fff09c4
|
| 3 |
+
size 15024
|
checkpoint-129072/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:82e08807171c880dfd9f769027479b435b8812d83689085a23f9a21dbfed3583
|
| 3 |
+
size 1064
|
checkpoint-129072/sentencepiece.bpe.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cfc8146abe2a0488e9e2a0c56de7952f7c11ab059eca145a0a727afce0db2865
|
| 3 |
+
size 5069051
|
checkpoint-129072/special_tokens_map.json
ADDED
|
@@ -0,0 +1,69 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"additional_special_tokens": [
|
| 3 |
+
"ar_AR",
|
| 4 |
+
"cs_CZ",
|
| 5 |
+
"de_DE",
|
| 6 |
+
"en_XX",
|
| 7 |
+
"es_XX",
|
| 8 |
+
"et_EE",
|
| 9 |
+
"fi_FI",
|
| 10 |
+
"fr_XX",
|
| 11 |
+
"gu_IN",
|
| 12 |
+
"hi_IN",
|
| 13 |
+
"it_IT",
|
| 14 |
+
"ja_XX",
|
| 15 |
+
"kk_KZ",
|
| 16 |
+
"ko_KR",
|
| 17 |
+
"lt_LT",
|
| 18 |
+
"lv_LV",
|
| 19 |
+
"my_MM",
|
| 20 |
+
"ne_NP",
|
| 21 |
+
"nl_XX",
|
| 22 |
+
"ro_RO",
|
| 23 |
+
"ru_RU",
|
| 24 |
+
"si_LK",
|
| 25 |
+
"tr_TR",
|
| 26 |
+
"vi_VN",
|
| 27 |
+
"zh_CN",
|
| 28 |
+
"af_ZA",
|
| 29 |
+
"az_AZ",
|
| 30 |
+
"bn_IN",
|
| 31 |
+
"fa_IR",
|
| 32 |
+
"he_IL",
|
| 33 |
+
"hr_HR",
|
| 34 |
+
"id_ID",
|
| 35 |
+
"ka_GE",
|
| 36 |
+
"km_KH",
|
| 37 |
+
"mk_MK",
|
| 38 |
+
"ml_IN",
|
| 39 |
+
"mn_MN",
|
| 40 |
+
"mr_IN",
|
| 41 |
+
"pl_PL",
|
| 42 |
+
"ps_AF",
|
| 43 |
+
"pt_XX",
|
| 44 |
+
"sv_SE",
|
| 45 |
+
"sw_KE",
|
| 46 |
+
"ta_IN",
|
| 47 |
+
"te_IN",
|
| 48 |
+
"th_TH",
|
| 49 |
+
"tl_XX",
|
| 50 |
+
"uk_UA",
|
| 51 |
+
"ur_PK",
|
| 52 |
+
"xh_ZA",
|
| 53 |
+
"gl_ES",
|
| 54 |
+
"sl_SI"
|
| 55 |
+
],
|
| 56 |
+
"bos_token": "<s>",
|
| 57 |
+
"cls_token": "<s>",
|
| 58 |
+
"eos_token": "</s>",
|
| 59 |
+
"mask_token": {
|
| 60 |
+
"content": "<mask>",
|
| 61 |
+
"lstrip": true,
|
| 62 |
+
"normalized": true,
|
| 63 |
+
"rstrip": false,
|
| 64 |
+
"single_word": false
|
| 65 |
+
},
|
| 66 |
+
"pad_token": "<pad>",
|
| 67 |
+
"sep_token": "</s>",
|
| 68 |
+
"unk_token": "<unk>"
|
| 69 |
+
}
|
checkpoint-129072/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:faf1fb0f45fdd5ae9fc44fafee492652a423fc749833ef57c2fa801358549d92
|
| 3 |
+
size 17110107
|
checkpoint-129072/tokenizer_config.json
ADDED
|
@@ -0,0 +1,528 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"250001": {
|
| 36 |
+
"content": "ar_AR",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
},
|
| 43 |
+
"250002": {
|
| 44 |
+
"content": "cs_CZ",
|
| 45 |
+
"lstrip": false,
|
| 46 |
+
"normalized": false,
|
| 47 |
+
"rstrip": false,
|
| 48 |
+
"single_word": false,
|
| 49 |
+
"special": true
|
| 50 |
+
},
|
| 51 |
+
"250003": {
|
| 52 |
+
"content": "de_DE",
|
| 53 |
+
"lstrip": false,
|
| 54 |
+
"normalized": false,
|
| 55 |
+
"rstrip": false,
|
| 56 |
+
"single_word": false,
|
| 57 |
+
"special": true
|
| 58 |
+
},
|
| 59 |
+
"250004": {
|
| 60 |
+
"content": "en_XX",
|
| 61 |
+
"lstrip": false,
|
| 62 |
+
"normalized": false,
|
| 63 |
+
"rstrip": false,
|
| 64 |
+
"single_word": false,
|
| 65 |
+
"special": true
|
| 66 |
+
},
|
| 67 |
+
"250005": {
|
| 68 |
+
"content": "es_XX",
|
| 69 |
+
"lstrip": false,
|
| 70 |
+
"normalized": false,
|
| 71 |
+
"rstrip": false,
|
| 72 |
+
"single_word": false,
|
| 73 |
+
"special": true
|
| 74 |
+
},
|
| 75 |
+
"250006": {
|
| 76 |
+
"content": "et_EE",
|
| 77 |
+
"lstrip": false,
|
| 78 |
+
"normalized": false,
|
| 79 |
+
"rstrip": false,
|
| 80 |
+
"single_word": false,
|
| 81 |
+
"special": true
|
| 82 |
+
},
|
| 83 |
+
"250007": {
|
| 84 |
+
"content": "fi_FI",
|
| 85 |
+
"lstrip": false,
|
| 86 |
+
"normalized": false,
|
| 87 |
+
"rstrip": false,
|
| 88 |
+
"single_word": false,
|
| 89 |
+
"special": true
|
| 90 |
+
},
|
| 91 |
+
"250008": {
|
| 92 |
+
"content": "fr_XX",
|
| 93 |
+
"lstrip": false,
|
| 94 |
+
"normalized": false,
|
| 95 |
+
"rstrip": false,
|
| 96 |
+
"single_word": false,
|
| 97 |
+
"special": true
|
| 98 |
+
},
|
| 99 |
+
"250009": {
|
| 100 |
+
"content": "gu_IN",
|
| 101 |
+
"lstrip": false,
|
| 102 |
+
"normalized": false,
|
| 103 |
+
"rstrip": false,
|
| 104 |
+
"single_word": false,
|
| 105 |
+
"special": true
|
| 106 |
+
},
|
| 107 |
+
"250010": {
|
| 108 |
+
"content": "hi_IN",
|
| 109 |
+
"lstrip": false,
|
| 110 |
+
"normalized": false,
|
| 111 |
+
"rstrip": false,
|
| 112 |
+
"single_word": false,
|
| 113 |
+
"special": true
|
| 114 |
+
},
|
| 115 |
+
"250011": {
|
| 116 |
+
"content": "it_IT",
|
| 117 |
+
"lstrip": false,
|
| 118 |
+
"normalized": false,
|
| 119 |
+
"rstrip": false,
|
| 120 |
+
"single_word": false,
|
| 121 |
+
"special": true
|
| 122 |
+
},
|
| 123 |
+
"250012": {
|
| 124 |
+
"content": "ja_XX",
|
| 125 |
+
"lstrip": false,
|
| 126 |
+
"normalized": false,
|
| 127 |
+
"rstrip": false,
|
| 128 |
+
"single_word": false,
|
| 129 |
+
"special": true
|
| 130 |
+
},
|
| 131 |
+
"250013": {
|
| 132 |
+
"content": "kk_KZ",
|
| 133 |
+
"lstrip": false,
|
| 134 |
+
"normalized": false,
|
| 135 |
+
"rstrip": false,
|
| 136 |
+
"single_word": false,
|
| 137 |
+
"special": true
|
| 138 |
+
},
|
| 139 |
+
"250014": {
|
| 140 |
+
"content": "ko_KR",
|
| 141 |
+
"lstrip": false,
|
| 142 |
+
"normalized": false,
|
| 143 |
+
"rstrip": false,
|
| 144 |
+
"single_word": false,
|
| 145 |
+
"special": true
|
| 146 |
+
},
|
| 147 |
+
"250015": {
|
| 148 |
+
"content": "lt_LT",
|
| 149 |
+
"lstrip": false,
|
| 150 |
+
"normalized": false,
|
| 151 |
+
"rstrip": false,
|
| 152 |
+
"single_word": false,
|
| 153 |
+
"special": true
|
| 154 |
+
},
|
| 155 |
+
"250016": {
|
| 156 |
+
"content": "lv_LV",
|
| 157 |
+
"lstrip": false,
|
| 158 |
+
"normalized": false,
|
| 159 |
+
"rstrip": false,
|
| 160 |
+
"single_word": false,
|
| 161 |
+
"special": true
|
| 162 |
+
},
|
| 163 |
+
"250017": {
|
| 164 |
+
"content": "my_MM",
|
| 165 |
+
"lstrip": false,
|
| 166 |
+
"normalized": false,
|
| 167 |
+
"rstrip": false,
|
| 168 |
+
"single_word": false,
|
| 169 |
+
"special": true
|
| 170 |
+
},
|
| 171 |
+
"250018": {
|
| 172 |
+
"content": "ne_NP",
|
| 173 |
+
"lstrip": false,
|
| 174 |
+
"normalized": false,
|
| 175 |
+
"rstrip": false,
|
| 176 |
+
"single_word": false,
|
| 177 |
+
"special": true
|
| 178 |
+
},
|
| 179 |
+
"250019": {
|
| 180 |
+
"content": "nl_XX",
|
| 181 |
+
"lstrip": false,
|
| 182 |
+
"normalized": false,
|
| 183 |
+
"rstrip": false,
|
| 184 |
+
"single_word": false,
|
| 185 |
+
"special": true
|
| 186 |
+
},
|
| 187 |
+
"250020": {
|
| 188 |
+
"content": "ro_RO",
|
| 189 |
+
"lstrip": false,
|
| 190 |
+
"normalized": false,
|
| 191 |
+
"rstrip": false,
|
| 192 |
+
"single_word": false,
|
| 193 |
+
"special": true
|
| 194 |
+
},
|
| 195 |
+
"250021": {
|
| 196 |
+
"content": "ru_RU",
|
| 197 |
+
"lstrip": false,
|
| 198 |
+
"normalized": false,
|
| 199 |
+
"rstrip": false,
|
| 200 |
+
"single_word": false,
|
| 201 |
+
"special": true
|
| 202 |
+
},
|
| 203 |
+
"250022": {
|
| 204 |
+
"content": "si_LK",
|
| 205 |
+
"lstrip": false,
|
| 206 |
+
"normalized": false,
|
| 207 |
+
"rstrip": false,
|
| 208 |
+
"single_word": false,
|
| 209 |
+
"special": true
|
| 210 |
+
},
|
| 211 |
+
"250023": {
|
| 212 |
+
"content": "tr_TR",
|
| 213 |
+
"lstrip": false,
|
| 214 |
+
"normalized": false,
|
| 215 |
+
"rstrip": false,
|
| 216 |
+
"single_word": false,
|
| 217 |
+
"special": true
|
| 218 |
+
},
|
| 219 |
+
"250024": {
|
| 220 |
+
"content": "vi_VN",
|
| 221 |
+
"lstrip": false,
|
| 222 |
+
"normalized": false,
|
| 223 |
+
"rstrip": false,
|
| 224 |
+
"single_word": false,
|
| 225 |
+
"special": true
|
| 226 |
+
},
|
| 227 |
+
"250025": {
|
| 228 |
+
"content": "zh_CN",
|
| 229 |
+
"lstrip": false,
|
| 230 |
+
"normalized": false,
|
| 231 |
+
"rstrip": false,
|
| 232 |
+
"single_word": false,
|
| 233 |
+
"special": true
|
| 234 |
+
},
|
| 235 |
+
"250026": {
|
| 236 |
+
"content": "af_ZA",
|
| 237 |
+
"lstrip": false,
|
| 238 |
+
"normalized": false,
|
| 239 |
+
"rstrip": false,
|
| 240 |
+
"single_word": false,
|
| 241 |
+
"special": true
|
| 242 |
+
},
|
| 243 |
+
"250027": {
|
| 244 |
+
"content": "az_AZ",
|
| 245 |
+
"lstrip": false,
|
| 246 |
+
"normalized": false,
|
| 247 |
+
"rstrip": false,
|
| 248 |
+
"single_word": false,
|
| 249 |
+
"special": true
|
| 250 |
+
},
|
| 251 |
+
"250028": {
|
| 252 |
+
"content": "bn_IN",
|
| 253 |
+
"lstrip": false,
|
| 254 |
+
"normalized": false,
|
| 255 |
+
"rstrip": false,
|
| 256 |
+
"single_word": false,
|
| 257 |
+
"special": true
|
| 258 |
+
},
|
| 259 |
+
"250029": {
|
| 260 |
+
"content": "fa_IR",
|
| 261 |
+
"lstrip": false,
|
| 262 |
+
"normalized": false,
|
| 263 |
+
"rstrip": false,
|
| 264 |
+
"single_word": false,
|
| 265 |
+
"special": true
|
| 266 |
+
},
|
| 267 |
+
"250030": {
|
| 268 |
+
"content": "he_IL",
|
| 269 |
+
"lstrip": false,
|
| 270 |
+
"normalized": false,
|
| 271 |
+
"rstrip": false,
|
| 272 |
+
"single_word": false,
|
| 273 |
+
"special": true
|
| 274 |
+
},
|
| 275 |
+
"250031": {
|
| 276 |
+
"content": "hr_HR",
|
| 277 |
+
"lstrip": false,
|
| 278 |
+
"normalized": false,
|
| 279 |
+
"rstrip": false,
|
| 280 |
+
"single_word": false,
|
| 281 |
+
"special": true
|
| 282 |
+
},
|
| 283 |
+
"250032": {
|
| 284 |
+
"content": "id_ID",
|
| 285 |
+
"lstrip": false,
|
| 286 |
+
"normalized": false,
|
| 287 |
+
"rstrip": false,
|
| 288 |
+
"single_word": false,
|
| 289 |
+
"special": true
|
| 290 |
+
},
|
| 291 |
+
"250033": {
|
| 292 |
+
"content": "ka_GE",
|
| 293 |
+
"lstrip": false,
|
| 294 |
+
"normalized": false,
|
| 295 |
+
"rstrip": false,
|
| 296 |
+
"single_word": false,
|
| 297 |
+
"special": true
|
| 298 |
+
},
|
| 299 |
+
"250034": {
|
| 300 |
+
"content": "km_KH",
|
| 301 |
+
"lstrip": false,
|
| 302 |
+
"normalized": false,
|
| 303 |
+
"rstrip": false,
|
| 304 |
+
"single_word": false,
|
| 305 |
+
"special": true
|
| 306 |
+
},
|
| 307 |
+
"250035": {
|
| 308 |
+
"content": "mk_MK",
|
| 309 |
+
"lstrip": false,
|
| 310 |
+
"normalized": false,
|
| 311 |
+
"rstrip": false,
|
| 312 |
+
"single_word": false,
|
| 313 |
+
"special": true
|
| 314 |
+
},
|
| 315 |
+
"250036": {
|
| 316 |
+
"content": "ml_IN",
|
| 317 |
+
"lstrip": false,
|
| 318 |
+
"normalized": false,
|
| 319 |
+
"rstrip": false,
|
| 320 |
+
"single_word": false,
|
| 321 |
+
"special": true
|
| 322 |
+
},
|
| 323 |
+
"250037": {
|
| 324 |
+
"content": "mn_MN",
|
| 325 |
+
"lstrip": false,
|
| 326 |
+
"normalized": false,
|
| 327 |
+
"rstrip": false,
|
| 328 |
+
"single_word": false,
|
| 329 |
+
"special": true
|
| 330 |
+
},
|
| 331 |
+
"250038": {
|
| 332 |
+
"content": "mr_IN",
|
| 333 |
+
"lstrip": false,
|
| 334 |
+
"normalized": false,
|
| 335 |
+
"rstrip": false,
|
| 336 |
+
"single_word": false,
|
| 337 |
+
"special": true
|
| 338 |
+
},
|
| 339 |
+
"250039": {
|
| 340 |
+
"content": "pl_PL",
|
| 341 |
+
"lstrip": false,
|
| 342 |
+
"normalized": false,
|
| 343 |
+
"rstrip": false,
|
| 344 |
+
"single_word": false,
|
| 345 |
+
"special": true
|
| 346 |
+
},
|
| 347 |
+
"250040": {
|
| 348 |
+
"content": "ps_AF",
|
| 349 |
+
"lstrip": false,
|
| 350 |
+
"normalized": false,
|
| 351 |
+
"rstrip": false,
|
| 352 |
+
"single_word": false,
|
| 353 |
+
"special": true
|
| 354 |
+
},
|
| 355 |
+
"250041": {
|
| 356 |
+
"content": "pt_XX",
|
| 357 |
+
"lstrip": false,
|
| 358 |
+
"normalized": false,
|
| 359 |
+
"rstrip": false,
|
| 360 |
+
"single_word": false,
|
| 361 |
+
"special": true
|
| 362 |
+
},
|
| 363 |
+
"250042": {
|
| 364 |
+
"content": "sv_SE",
|
| 365 |
+
"lstrip": false,
|
| 366 |
+
"normalized": false,
|
| 367 |
+
"rstrip": false,
|
| 368 |
+
"single_word": false,
|
| 369 |
+
"special": true
|
| 370 |
+
},
|
| 371 |
+
"250043": {
|
| 372 |
+
"content": "sw_KE",
|
| 373 |
+
"lstrip": false,
|
| 374 |
+
"normalized": false,
|
| 375 |
+
"rstrip": false,
|
| 376 |
+
"single_word": false,
|
| 377 |
+
"special": true
|
| 378 |
+
},
|
| 379 |
+
"250044": {
|
| 380 |
+
"content": "ta_IN",
|
| 381 |
+
"lstrip": false,
|
| 382 |
+
"normalized": false,
|
| 383 |
+
"rstrip": false,
|
| 384 |
+
"single_word": false,
|
| 385 |
+
"special": true
|
| 386 |
+
},
|
| 387 |
+
"250045": {
|
| 388 |
+
"content": "te_IN",
|
| 389 |
+
"lstrip": false,
|
| 390 |
+
"normalized": false,
|
| 391 |
+
"rstrip": false,
|
| 392 |
+
"single_word": false,
|
| 393 |
+
"special": true
|
| 394 |
+
},
|
| 395 |
+
"250046": {
|
| 396 |
+
"content": "th_TH",
|
| 397 |
+
"lstrip": false,
|
| 398 |
+
"normalized": false,
|
| 399 |
+
"rstrip": false,
|
| 400 |
+
"single_word": false,
|
| 401 |
+
"special": true
|
| 402 |
+
},
|
| 403 |
+
"250047": {
|
| 404 |
+
"content": "tl_XX",
|
| 405 |
+
"lstrip": false,
|
| 406 |
+
"normalized": false,
|
| 407 |
+
"rstrip": false,
|
| 408 |
+
"single_word": false,
|
| 409 |
+
"special": true
|
| 410 |
+
},
|
| 411 |
+
"250048": {
|
| 412 |
+
"content": "uk_UA",
|
| 413 |
+
"lstrip": false,
|
| 414 |
+
"normalized": false,
|
| 415 |
+
"rstrip": false,
|
| 416 |
+
"single_word": false,
|
| 417 |
+
"special": true
|
| 418 |
+
},
|
| 419 |
+
"250049": {
|
| 420 |
+
"content": "ur_PK",
|
| 421 |
+
"lstrip": false,
|
| 422 |
+
"normalized": false,
|
| 423 |
+
"rstrip": false,
|
| 424 |
+
"single_word": false,
|
| 425 |
+
"special": true
|
| 426 |
+
},
|
| 427 |
+
"250050": {
|
| 428 |
+
"content": "xh_ZA",
|
| 429 |
+
"lstrip": false,
|
| 430 |
+
"normalized": false,
|
| 431 |
+
"rstrip": false,
|
| 432 |
+
"single_word": false,
|
| 433 |
+
"special": true
|
| 434 |
+
},
|
| 435 |
+
"250051": {
|
| 436 |
+
"content": "gl_ES",
|
| 437 |
+
"lstrip": false,
|
| 438 |
+
"normalized": false,
|
| 439 |
+
"rstrip": false,
|
| 440 |
+
"single_word": false,
|
| 441 |
+
"special": true
|
| 442 |
+
},
|
| 443 |
+
"250052": {
|
| 444 |
+
"content": "sl_SI",
|
| 445 |
+
"lstrip": false,
|
| 446 |
+
"normalized": false,
|
| 447 |
+
"rstrip": false,
|
| 448 |
+
"single_word": false,
|
| 449 |
+
"special": true
|
| 450 |
+
},
|
| 451 |
+
"250053": {
|
| 452 |
+
"content": "<mask>",
|
| 453 |
+
"lstrip": true,
|
| 454 |
+
"normalized": true,
|
| 455 |
+
"rstrip": false,
|
| 456 |
+
"single_word": false,
|
| 457 |
+
"special": true
|
| 458 |
+
}
|
| 459 |
+
},
|
| 460 |
+
"additional_special_tokens": [
|
| 461 |
+
"ar_AR",
|
| 462 |
+
"cs_CZ",
|
| 463 |
+
"de_DE",
|
| 464 |
+
"en_XX",
|
| 465 |
+
"es_XX",
|
| 466 |
+
"et_EE",
|
| 467 |
+
"fi_FI",
|
| 468 |
+
"fr_XX",
|
| 469 |
+
"gu_IN",
|
| 470 |
+
"hi_IN",
|
| 471 |
+
"it_IT",
|
| 472 |
+
"ja_XX",
|
| 473 |
+
"kk_KZ",
|
| 474 |
+
"ko_KR",
|
| 475 |
+
"lt_LT",
|
| 476 |
+
"lv_LV",
|
| 477 |
+
"my_MM",
|
| 478 |
+
"ne_NP",
|
| 479 |
+
"nl_XX",
|
| 480 |
+
"ro_RO",
|
| 481 |
+
"ru_RU",
|
| 482 |
+
"si_LK",
|
| 483 |
+
"tr_TR",
|
| 484 |
+
"vi_VN",
|
| 485 |
+
"zh_CN",
|
| 486 |
+
"af_ZA",
|
| 487 |
+
"az_AZ",
|
| 488 |
+
"bn_IN",
|
| 489 |
+
"fa_IR",
|
| 490 |
+
"he_IL",
|
| 491 |
+
"hr_HR",
|
| 492 |
+
"id_ID",
|
| 493 |
+
"ka_GE",
|
| 494 |
+
"km_KH",
|
| 495 |
+
"mk_MK",
|
| 496 |
+
"ml_IN",
|
| 497 |
+
"mn_MN",
|
| 498 |
+
"mr_IN",
|
| 499 |
+
"pl_PL",
|
| 500 |
+
"ps_AF",
|
| 501 |
+
"pt_XX",
|
| 502 |
+
"sv_SE",
|
| 503 |
+
"sw_KE",
|
| 504 |
+
"ta_IN",
|
| 505 |
+
"te_IN",
|
| 506 |
+
"th_TH",
|
| 507 |
+
"tl_XX",
|
| 508 |
+
"uk_UA",
|
| 509 |
+
"ur_PK",
|
| 510 |
+
"xh_ZA",
|
| 511 |
+
"gl_ES",
|
| 512 |
+
"sl_SI"
|
| 513 |
+
],
|
| 514 |
+
"bos_token": "<s>",
|
| 515 |
+
"clean_up_tokenization_spaces": true,
|
| 516 |
+
"cls_token": "<s>",
|
| 517 |
+
"eos_token": "</s>",
|
| 518 |
+
"language_codes": "ML50",
|
| 519 |
+
"mask_token": "<mask>",
|
| 520 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 521 |
+
"pad_token": "<pad>",
|
| 522 |
+
"sep_token": "</s>",
|
| 523 |
+
"sp_model_kwargs": {},
|
| 524 |
+
"src_lang": "ja_XX",
|
| 525 |
+
"tgt_lang": "ko_KR",
|
| 526 |
+
"tokenizer_class": "MBart50Tokenizer",
|
| 527 |
+
"unk_token": "<unk>"
|
| 528 |
+
}
|
checkpoint-129072/trainer_state.json
ADDED
|
@@ -0,0 +1,1599 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.6568061113357544,
|
| 3 |
+
"best_model_checkpoint": "./finetuned_model/checkpoint-129072",
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 129072,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.01,
|
| 13 |
+
"learning_rate": 9.96e-06,
|
| 14 |
+
"loss": 2.2631,
|
| 15 |
+
"step": 500
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.02,
|
| 19 |
+
"learning_rate": 1.992e-05,
|
| 20 |
+
"loss": 1.5976,
|
| 21 |
+
"step": 1000
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.03,
|
| 25 |
+
"learning_rate": 2.9920000000000005e-05,
|
| 26 |
+
"loss": 1.4641,
|
| 27 |
+
"step": 1500
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.05,
|
| 31 |
+
"learning_rate": 3.9920000000000004e-05,
|
| 32 |
+
"loss": 1.3868,
|
| 33 |
+
"step": 2000
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.06,
|
| 37 |
+
"learning_rate": 4.992e-05,
|
| 38 |
+
"loss": 1.344,
|
| 39 |
+
"step": 2500
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.07,
|
| 43 |
+
"learning_rate": 4.996142239367825e-05,
|
| 44 |
+
"loss": 1.3138,
|
| 45 |
+
"step": 3000
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.08,
|
| 49 |
+
"learning_rate": 4.9922533677628105e-05,
|
| 50 |
+
"loss": 1.244,
|
| 51 |
+
"step": 3500
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.09,
|
| 55 |
+
"learning_rate": 4.988364496157795e-05,
|
| 56 |
+
"loss": 1.2047,
|
| 57 |
+
"step": 4000
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.1,
|
| 61 |
+
"learning_rate": 4.98447562455278e-05,
|
| 62 |
+
"loss": 1.1707,
|
| 63 |
+
"step": 4500
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.12,
|
| 67 |
+
"learning_rate": 4.980586752947765e-05,
|
| 68 |
+
"loss": 1.1331,
|
| 69 |
+
"step": 5000
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.13,
|
| 73 |
+
"learning_rate": 4.97670565908596e-05,
|
| 74 |
+
"loss": 1.1021,
|
| 75 |
+
"step": 5500
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.14,
|
| 79 |
+
"learning_rate": 4.972816787480945e-05,
|
| 80 |
+
"loss": 1.0923,
|
| 81 |
+
"step": 6000
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.15,
|
| 85 |
+
"learning_rate": 4.968927915875929e-05,
|
| 86 |
+
"loss": 1.0713,
|
| 87 |
+
"step": 6500
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.16,
|
| 91 |
+
"learning_rate": 4.965039044270914e-05,
|
| 92 |
+
"loss": 1.0609,
|
| 93 |
+
"step": 7000
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.17,
|
| 97 |
+
"learning_rate": 4.9611501726658995e-05,
|
| 98 |
+
"loss": 1.043,
|
| 99 |
+
"step": 7500
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.19,
|
| 103 |
+
"learning_rate": 4.957261301060884e-05,
|
| 104 |
+
"loss": 1.0345,
|
| 105 |
+
"step": 8000
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.2,
|
| 109 |
+
"learning_rate": 4.953380207199079e-05,
|
| 110 |
+
"loss": 1.02,
|
| 111 |
+
"step": 8500
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.21,
|
| 115 |
+
"learning_rate": 4.949491335594064e-05,
|
| 116 |
+
"loss": 1.0057,
|
| 117 |
+
"step": 9000
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.22,
|
| 121 |
+
"learning_rate": 4.945602463989049e-05,
|
| 122 |
+
"loss": 0.9978,
|
| 123 |
+
"step": 9500
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.23,
|
| 127 |
+
"learning_rate": 4.9417135923840344e-05,
|
| 128 |
+
"loss": 0.9801,
|
| 129 |
+
"step": 10000
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.24,
|
| 133 |
+
"learning_rate": 4.937824720779019e-05,
|
| 134 |
+
"loss": 0.9752,
|
| 135 |
+
"step": 10500
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.26,
|
| 139 |
+
"learning_rate": 4.933935849174004e-05,
|
| 140 |
+
"loss": 0.9736,
|
| 141 |
+
"step": 11000
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.27,
|
| 145 |
+
"learning_rate": 4.930046977568989e-05,
|
| 146 |
+
"loss": 0.9519,
|
| 147 |
+
"step": 11500
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.28,
|
| 151 |
+
"learning_rate": 4.9261581059639736e-05,
|
| 152 |
+
"loss": 0.9559,
|
| 153 |
+
"step": 12000
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.29,
|
| 157 |
+
"learning_rate": 4.9222770121021686e-05,
|
| 158 |
+
"loss": 0.936,
|
| 159 |
+
"step": 12500
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.3,
|
| 163 |
+
"learning_rate": 4.9183959182403636e-05,
|
| 164 |
+
"loss": 0.9276,
|
| 165 |
+
"step": 13000
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.31,
|
| 169 |
+
"learning_rate": 4.914507046635349e-05,
|
| 170 |
+
"loss": 0.9384,
|
| 171 |
+
"step": 13500
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.33,
|
| 175 |
+
"learning_rate": 4.910618175030333e-05,
|
| 176 |
+
"loss": 0.9347,
|
| 177 |
+
"step": 14000
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.34,
|
| 181 |
+
"learning_rate": 4.9067293034253183e-05,
|
| 182 |
+
"loss": 0.9182,
|
| 183 |
+
"step": 14500
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.35,
|
| 187 |
+
"learning_rate": 4.902840431820303e-05,
|
| 188 |
+
"loss": 0.9136,
|
| 189 |
+
"step": 15000
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.36,
|
| 193 |
+
"learning_rate": 4.898951560215288e-05,
|
| 194 |
+
"loss": 0.9096,
|
| 195 |
+
"step": 15500
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.37,
|
| 199 |
+
"learning_rate": 4.895062688610273e-05,
|
| 200 |
+
"loss": 0.8926,
|
| 201 |
+
"step": 16000
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.38,
|
| 205 |
+
"learning_rate": 4.891181594748468e-05,
|
| 206 |
+
"loss": 0.8946,
|
| 207 |
+
"step": 16500
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.4,
|
| 211 |
+
"learning_rate": 4.887300500886663e-05,
|
| 212 |
+
"loss": 0.8895,
|
| 213 |
+
"step": 17000
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.41,
|
| 217 |
+
"learning_rate": 4.8834116292816475e-05,
|
| 218 |
+
"loss": 0.8897,
|
| 219 |
+
"step": 17500
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.42,
|
| 223 |
+
"learning_rate": 4.879522757676633e-05,
|
| 224 |
+
"loss": 0.8901,
|
| 225 |
+
"step": 18000
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.43,
|
| 229 |
+
"learning_rate": 4.875633886071618e-05,
|
| 230 |
+
"loss": 0.8839,
|
| 231 |
+
"step": 18500
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 0.44,
|
| 235 |
+
"learning_rate": 4.871745014466603e-05,
|
| 236 |
+
"loss": 0.8737,
|
| 237 |
+
"step": 19000
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.45,
|
| 241 |
+
"learning_rate": 4.867863920604798e-05,
|
| 242 |
+
"loss": 0.8613,
|
| 243 |
+
"step": 19500
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.46,
|
| 247 |
+
"learning_rate": 4.8639750489997824e-05,
|
| 248 |
+
"loss": 0.8547,
|
| 249 |
+
"step": 20000
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.48,
|
| 253 |
+
"learning_rate": 4.8600861773947676e-05,
|
| 254 |
+
"loss": 0.8623,
|
| 255 |
+
"step": 20500
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.49,
|
| 259 |
+
"learning_rate": 4.856197305789753e-05,
|
| 260 |
+
"loss": 0.8615,
|
| 261 |
+
"step": 21000
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.5,
|
| 265 |
+
"learning_rate": 4.852308434184737e-05,
|
| 266 |
+
"loss": 0.8614,
|
| 267 |
+
"step": 21500
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.51,
|
| 271 |
+
"learning_rate": 4.848419562579722e-05,
|
| 272 |
+
"loss": 0.846,
|
| 273 |
+
"step": 22000
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 0.52,
|
| 277 |
+
"learning_rate": 4.8445384687179167e-05,
|
| 278 |
+
"loss": 0.8487,
|
| 279 |
+
"step": 22500
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.53,
|
| 283 |
+
"learning_rate": 4.840649597112902e-05,
|
| 284 |
+
"loss": 0.8426,
|
| 285 |
+
"step": 23000
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.55,
|
| 289 |
+
"learning_rate": 4.836760725507887e-05,
|
| 290 |
+
"loss": 0.8398,
|
| 291 |
+
"step": 23500
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.56,
|
| 295 |
+
"learning_rate": 4.8328718539028714e-05,
|
| 296 |
+
"loss": 0.8407,
|
| 297 |
+
"step": 24000
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.57,
|
| 301 |
+
"learning_rate": 4.828998537784277e-05,
|
| 302 |
+
"loss": 0.8438,
|
| 303 |
+
"step": 24500
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.58,
|
| 307 |
+
"learning_rate": 4.825117443922472e-05,
|
| 308 |
+
"loss": 0.8257,
|
| 309 |
+
"step": 25000
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.59,
|
| 313 |
+
"learning_rate": 4.821228572317456e-05,
|
| 314 |
+
"loss": 0.8328,
|
| 315 |
+
"step": 25500
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 0.6,
|
| 319 |
+
"learning_rate": 4.8173397007124415e-05,
|
| 320 |
+
"loss": 0.8229,
|
| 321 |
+
"step": 26000
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 0.62,
|
| 325 |
+
"learning_rate": 4.8134508291074266e-05,
|
| 326 |
+
"loss": 0.8249,
|
| 327 |
+
"step": 26500
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 0.63,
|
| 331 |
+
"learning_rate": 4.809561957502412e-05,
|
| 332 |
+
"loss": 0.8232,
|
| 333 |
+
"step": 27000
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 0.64,
|
| 337 |
+
"learning_rate": 4.805673085897396e-05,
|
| 338 |
+
"loss": 0.83,
|
| 339 |
+
"step": 27500
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.65,
|
| 343 |
+
"learning_rate": 4.801791992035591e-05,
|
| 344 |
+
"loss": 0.8339,
|
| 345 |
+
"step": 28000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.66,
|
| 349 |
+
"learning_rate": 4.7979031204305764e-05,
|
| 350 |
+
"loss": 0.817,
|
| 351 |
+
"step": 28500
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 0.67,
|
| 355 |
+
"learning_rate": 4.7940142488255615e-05,
|
| 356 |
+
"loss": 0.8223,
|
| 357 |
+
"step": 29000
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 0.69,
|
| 361 |
+
"learning_rate": 4.790125377220545e-05,
|
| 362 |
+
"loss": 0.8041,
|
| 363 |
+
"step": 29500
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 0.7,
|
| 367 |
+
"learning_rate": 4.7862365056155305e-05,
|
| 368 |
+
"loss": 0.8126,
|
| 369 |
+
"step": 30000
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.71,
|
| 373 |
+
"learning_rate": 4.7823476340105156e-05,
|
| 374 |
+
"loss": 0.8039,
|
| 375 |
+
"step": 30500
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 0.72,
|
| 379 |
+
"learning_rate": 4.778458762405501e-05,
|
| 380 |
+
"loss": 0.8071,
|
| 381 |
+
"step": 31000
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.73,
|
| 385 |
+
"learning_rate": 4.774569890800485e-05,
|
| 386 |
+
"loss": 0.8065,
|
| 387 |
+
"step": 31500
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.74,
|
| 391 |
+
"learning_rate": 4.7706810191954704e-05,
|
| 392 |
+
"loss": 0.7998,
|
| 393 |
+
"step": 32000
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 0.76,
|
| 397 |
+
"learning_rate": 4.7667921475904556e-05,
|
| 398 |
+
"loss": 0.8034,
|
| 399 |
+
"step": 32500
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 0.77,
|
| 403 |
+
"learning_rate": 4.76290327598544e-05,
|
| 404 |
+
"loss": 0.8031,
|
| 405 |
+
"step": 33000
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 0.78,
|
| 409 |
+
"learning_rate": 4.759014404380425e-05,
|
| 410 |
+
"loss": 0.8057,
|
| 411 |
+
"step": 33500
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 0.79,
|
| 415 |
+
"learning_rate": 4.7551255327754103e-05,
|
| 416 |
+
"loss": 0.7905,
|
| 417 |
+
"step": 34000
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 0.8,
|
| 421 |
+
"learning_rate": 4.751236661170395e-05,
|
| 422 |
+
"loss": 0.794,
|
| 423 |
+
"step": 34500
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.81,
|
| 427 |
+
"learning_rate": 4.74734778956538e-05,
|
| 428 |
+
"loss": 0.7906,
|
| 429 |
+
"step": 35000
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.83,
|
| 433 |
+
"learning_rate": 4.743458917960365e-05,
|
| 434 |
+
"loss": 0.7865,
|
| 435 |
+
"step": 35500
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 0.84,
|
| 439 |
+
"learning_rate": 4.73957004635535e-05,
|
| 440 |
+
"loss": 0.7644,
|
| 441 |
+
"step": 36000
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 0.85,
|
| 445 |
+
"learning_rate": 4.735681174750335e-05,
|
| 446 |
+
"loss": 0.7737,
|
| 447 |
+
"step": 36500
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 0.86,
|
| 451 |
+
"learning_rate": 4.73179230314532e-05,
|
| 452 |
+
"loss": 0.7746,
|
| 453 |
+
"step": 37000
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 0.87,
|
| 457 |
+
"learning_rate": 4.7279034315403044e-05,
|
| 458 |
+
"loss": 0.7832,
|
| 459 |
+
"step": 37500
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.88,
|
| 463 |
+
"learning_rate": 4.724014559935289e-05,
|
| 464 |
+
"loss": 0.7706,
|
| 465 |
+
"step": 38000
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.89,
|
| 469 |
+
"learning_rate": 4.720141243816694e-05,
|
| 470 |
+
"loss": 0.7778,
|
| 471 |
+
"step": 38500
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.91,
|
| 475 |
+
"learning_rate": 4.7162523722116795e-05,
|
| 476 |
+
"loss": 0.7713,
|
| 477 |
+
"step": 39000
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 0.92,
|
| 481 |
+
"learning_rate": 4.712363500606664e-05,
|
| 482 |
+
"loss": 0.772,
|
| 483 |
+
"step": 39500
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 0.93,
|
| 487 |
+
"learning_rate": 4.708474629001649e-05,
|
| 488 |
+
"loss": 0.775,
|
| 489 |
+
"step": 40000
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 0.94,
|
| 493 |
+
"learning_rate": 4.704593535139844e-05,
|
| 494 |
+
"loss": 0.7718,
|
| 495 |
+
"step": 40500
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 0.95,
|
| 499 |
+
"learning_rate": 4.700704663534829e-05,
|
| 500 |
+
"loss": 0.7724,
|
| 501 |
+
"step": 41000
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 0.96,
|
| 505 |
+
"learning_rate": 4.696815791929814e-05,
|
| 506 |
+
"loss": 0.7629,
|
| 507 |
+
"step": 41500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.98,
|
| 511 |
+
"learning_rate": 4.692926920324799e-05,
|
| 512 |
+
"loss": 0.7731,
|
| 513 |
+
"step": 42000
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.99,
|
| 517 |
+
"learning_rate": 4.689038048719784e-05,
|
| 518 |
+
"loss": 0.7597,
|
| 519 |
+
"step": 42500
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 1.0,
|
| 523 |
+
"learning_rate": 4.6851491771147685e-05,
|
| 524 |
+
"loss": 0.7629,
|
| 525 |
+
"step": 43000
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 1.0,
|
| 529 |
+
"eval_bleu": 54.9608,
|
| 530 |
+
"eval_gen_len": 16.1273,
|
| 531 |
+
"eval_loss": 0.7230631709098816,
|
| 532 |
+
"eval_runtime": 7044.4442,
|
| 533 |
+
"eval_samples_per_second": 12.196,
|
| 534 |
+
"eval_steps_per_second": 1.525,
|
| 535 |
+
"step": 43024
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 1.01,
|
| 539 |
+
"learning_rate": 4.6812680832529634e-05,
|
| 540 |
+
"loss": 0.6436,
|
| 541 |
+
"step": 43500
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 1.02,
|
| 545 |
+
"learning_rate": 4.6773869893911584e-05,
|
| 546 |
+
"loss": 0.6446,
|
| 547 |
+
"step": 44000
|
| 548 |
+
},
|
| 549 |
+
{
|
| 550 |
+
"epoch": 1.03,
|
| 551 |
+
"learning_rate": 4.6734981177861435e-05,
|
| 552 |
+
"loss": 0.6489,
|
| 553 |
+
"step": 44500
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 1.05,
|
| 557 |
+
"learning_rate": 4.669617023924338e-05,
|
| 558 |
+
"loss": 0.6512,
|
| 559 |
+
"step": 45000
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 1.06,
|
| 563 |
+
"learning_rate": 4.665728152319323e-05,
|
| 564 |
+
"loss": 0.648,
|
| 565 |
+
"step": 45500
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"epoch": 1.07,
|
| 569 |
+
"learning_rate": 4.661839280714308e-05,
|
| 570 |
+
"loss": 0.6426,
|
| 571 |
+
"step": 46000
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 1.08,
|
| 575 |
+
"learning_rate": 4.657958186852503e-05,
|
| 576 |
+
"loss": 0.6489,
|
| 577 |
+
"step": 46500
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 1.09,
|
| 581 |
+
"learning_rate": 4.6540693152474876e-05,
|
| 582 |
+
"loss": 0.6417,
|
| 583 |
+
"step": 47000
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 1.1,
|
| 587 |
+
"learning_rate": 4.650180443642473e-05,
|
| 588 |
+
"loss": 0.645,
|
| 589 |
+
"step": 47500
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 1.12,
|
| 593 |
+
"learning_rate": 4.646291572037458e-05,
|
| 594 |
+
"loss": 0.6543,
|
| 595 |
+
"step": 48000
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"epoch": 1.13,
|
| 599 |
+
"learning_rate": 4.642402700432443e-05,
|
| 600 |
+
"loss": 0.6555,
|
| 601 |
+
"step": 48500
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 1.14,
|
| 605 |
+
"learning_rate": 4.6385138288274275e-05,
|
| 606 |
+
"loss": 0.6501,
|
| 607 |
+
"step": 49000
|
| 608 |
+
},
|
| 609 |
+
{
|
| 610 |
+
"epoch": 1.15,
|
| 611 |
+
"learning_rate": 4.634624957222413e-05,
|
| 612 |
+
"loss": 0.6447,
|
| 613 |
+
"step": 49500
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 1.16,
|
| 617 |
+
"learning_rate": 4.630736085617398e-05,
|
| 618 |
+
"loss": 0.6432,
|
| 619 |
+
"step": 50000
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 1.17,
|
| 623 |
+
"learning_rate": 4.626847214012382e-05,
|
| 624 |
+
"loss": 0.6516,
|
| 625 |
+
"step": 50500
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 1.19,
|
| 629 |
+
"learning_rate": 4.6229583424073674e-05,
|
| 630 |
+
"loss": 0.6517,
|
| 631 |
+
"step": 51000
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 1.2,
|
| 635 |
+
"learning_rate": 4.6190694708023526e-05,
|
| 636 |
+
"loss": 0.6534,
|
| 637 |
+
"step": 51500
|
| 638 |
+
},
|
| 639 |
+
{
|
| 640 |
+
"epoch": 1.21,
|
| 641 |
+
"learning_rate": 4.615180599197337e-05,
|
| 642 |
+
"loss": 0.6482,
|
| 643 |
+
"step": 52000
|
| 644 |
+
},
|
| 645 |
+
{
|
| 646 |
+
"epoch": 1.22,
|
| 647 |
+
"learning_rate": 4.611291727592322e-05,
|
| 648 |
+
"loss": 0.6504,
|
| 649 |
+
"step": 52500
|
| 650 |
+
},
|
| 651 |
+
{
|
| 652 |
+
"epoch": 1.23,
|
| 653 |
+
"learning_rate": 4.607402855987307e-05,
|
| 654 |
+
"loss": 0.6481,
|
| 655 |
+
"step": 53000
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 1.24,
|
| 659 |
+
"learning_rate": 4.603513984382292e-05,
|
| 660 |
+
"loss": 0.6444,
|
| 661 |
+
"step": 53500
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 1.26,
|
| 665 |
+
"learning_rate": 4.599625112777276e-05,
|
| 666 |
+
"loss": 0.6509,
|
| 667 |
+
"step": 54000
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 1.27,
|
| 671 |
+
"learning_rate": 4.5957362411722615e-05,
|
| 672 |
+
"loss": 0.6485,
|
| 673 |
+
"step": 54500
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 1.28,
|
| 677 |
+
"learning_rate": 4.5918551473104564e-05,
|
| 678 |
+
"loss": 0.6474,
|
| 679 |
+
"step": 55000
|
| 680 |
+
},
|
| 681 |
+
{
|
| 682 |
+
"epoch": 1.29,
|
| 683 |
+
"learning_rate": 4.5879662757054416e-05,
|
| 684 |
+
"loss": 0.6467,
|
| 685 |
+
"step": 55500
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 1.3,
|
| 689 |
+
"learning_rate": 4.584077404100427e-05,
|
| 690 |
+
"loss": 0.6442,
|
| 691 |
+
"step": 56000
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 1.31,
|
| 695 |
+
"learning_rate": 4.580188532495411e-05,
|
| 696 |
+
"loss": 0.6537,
|
| 697 |
+
"step": 56500
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"epoch": 1.32,
|
| 701 |
+
"learning_rate": 4.576307438633606e-05,
|
| 702 |
+
"loss": 0.6563,
|
| 703 |
+
"step": 57000
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 1.34,
|
| 707 |
+
"learning_rate": 4.5724185670285913e-05,
|
| 708 |
+
"loss": 0.6587,
|
| 709 |
+
"step": 57500
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 1.35,
|
| 713 |
+
"learning_rate": 4.5685296954235765e-05,
|
| 714 |
+
"loss": 0.6535,
|
| 715 |
+
"step": 58000
|
| 716 |
+
},
|
| 717 |
+
{
|
| 718 |
+
"epoch": 1.36,
|
| 719 |
+
"learning_rate": 4.564640823818561e-05,
|
| 720 |
+
"loss": 0.6623,
|
| 721 |
+
"step": 58500
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 1.37,
|
| 725 |
+
"learning_rate": 4.560759729956756e-05,
|
| 726 |
+
"loss": 0.6458,
|
| 727 |
+
"step": 59000
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"epoch": 1.38,
|
| 731 |
+
"learning_rate": 4.556870858351741e-05,
|
| 732 |
+
"loss": 0.6557,
|
| 733 |
+
"step": 59500
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 1.39,
|
| 737 |
+
"learning_rate": 4.552981986746726e-05,
|
| 738 |
+
"loss": 0.6458,
|
| 739 |
+
"step": 60000
|
| 740 |
+
},
|
| 741 |
+
{
|
| 742 |
+
"epoch": 1.41,
|
| 743 |
+
"learning_rate": 4.549093115141711e-05,
|
| 744 |
+
"loss": 0.6565,
|
| 745 |
+
"step": 60500
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 1.42,
|
| 749 |
+
"learning_rate": 4.545204243536695e-05,
|
| 750 |
+
"loss": 0.6521,
|
| 751 |
+
"step": 61000
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 1.43,
|
| 755 |
+
"learning_rate": 4.54131537193168e-05,
|
| 756 |
+
"loss": 0.6485,
|
| 757 |
+
"step": 61500
|
| 758 |
+
},
|
| 759 |
+
{
|
| 760 |
+
"epoch": 1.44,
|
| 761 |
+
"learning_rate": 4.5374265003266655e-05,
|
| 762 |
+
"loss": 0.6503,
|
| 763 |
+
"step": 62000
|
| 764 |
+
},
|
| 765 |
+
{
|
| 766 |
+
"epoch": 1.45,
|
| 767 |
+
"learning_rate": 4.53353762872165e-05,
|
| 768 |
+
"loss": 0.6483,
|
| 769 |
+
"step": 62500
|
| 770 |
+
},
|
| 771 |
+
{
|
| 772 |
+
"epoch": 1.46,
|
| 773 |
+
"learning_rate": 4.529656534859845e-05,
|
| 774 |
+
"loss": 0.6425,
|
| 775 |
+
"step": 63000
|
| 776 |
+
},
|
| 777 |
+
{
|
| 778 |
+
"epoch": 1.48,
|
| 779 |
+
"learning_rate": 4.52576766325483e-05,
|
| 780 |
+
"loss": 0.6566,
|
| 781 |
+
"step": 63500
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 1.49,
|
| 785 |
+
"learning_rate": 4.521878791649815e-05,
|
| 786 |
+
"loss": 0.6537,
|
| 787 |
+
"step": 64000
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 1.5,
|
| 791 |
+
"learning_rate": 4.5179899200448e-05,
|
| 792 |
+
"loss": 0.653,
|
| 793 |
+
"step": 64500
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 1.51,
|
| 797 |
+
"learning_rate": 4.514116603926205e-05,
|
| 798 |
+
"loss": 0.6448,
|
| 799 |
+
"step": 65000
|
| 800 |
+
},
|
| 801 |
+
{
|
| 802 |
+
"epoch": 1.52,
|
| 803 |
+
"learning_rate": 4.51022773232119e-05,
|
| 804 |
+
"loss": 0.6594,
|
| 805 |
+
"step": 65500
|
| 806 |
+
},
|
| 807 |
+
{
|
| 808 |
+
"epoch": 1.53,
|
| 809 |
+
"learning_rate": 4.506338860716175e-05,
|
| 810 |
+
"loss": 0.6551,
|
| 811 |
+
"step": 66000
|
| 812 |
+
},
|
| 813 |
+
{
|
| 814 |
+
"epoch": 1.55,
|
| 815 |
+
"learning_rate": 4.50245776685437e-05,
|
| 816 |
+
"loss": 0.6559,
|
| 817 |
+
"step": 66500
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"epoch": 1.56,
|
| 821 |
+
"learning_rate": 4.498568895249355e-05,
|
| 822 |
+
"loss": 0.6506,
|
| 823 |
+
"step": 67000
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 1.57,
|
| 827 |
+
"learning_rate": 4.49468002364434e-05,
|
| 828 |
+
"loss": 0.6504,
|
| 829 |
+
"step": 67500
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"epoch": 1.58,
|
| 833 |
+
"learning_rate": 4.4907911520393245e-05,
|
| 834 |
+
"loss": 0.6643,
|
| 835 |
+
"step": 68000
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 1.59,
|
| 839 |
+
"learning_rate": 4.486902280434309e-05,
|
| 840 |
+
"loss": 0.6435,
|
| 841 |
+
"step": 68500
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"epoch": 1.6,
|
| 845 |
+
"learning_rate": 4.483013408829294e-05,
|
| 846 |
+
"loss": 0.6389,
|
| 847 |
+
"step": 69000
|
| 848 |
+
},
|
| 849 |
+
{
|
| 850 |
+
"epoch": 1.62,
|
| 851 |
+
"learning_rate": 4.479124537224279e-05,
|
| 852 |
+
"loss": 0.6495,
|
| 853 |
+
"step": 69500
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 1.63,
|
| 857 |
+
"learning_rate": 4.475235665619264e-05,
|
| 858 |
+
"loss": 0.6395,
|
| 859 |
+
"step": 70000
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 1.64,
|
| 863 |
+
"learning_rate": 4.471346794014249e-05,
|
| 864 |
+
"loss": 0.6404,
|
| 865 |
+
"step": 70500
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 1.65,
|
| 869 |
+
"learning_rate": 4.467457922409234e-05,
|
| 870 |
+
"loss": 0.6438,
|
| 871 |
+
"step": 71000
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 1.66,
|
| 875 |
+
"learning_rate": 4.4635690508042186e-05,
|
| 876 |
+
"loss": 0.6361,
|
| 877 |
+
"step": 71500
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 1.67,
|
| 881 |
+
"learning_rate": 4.459680179199204e-05,
|
| 882 |
+
"loss": 0.6486,
|
| 883 |
+
"step": 72000
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 1.69,
|
| 887 |
+
"learning_rate": 4.455791307594189e-05,
|
| 888 |
+
"loss": 0.6489,
|
| 889 |
+
"step": 72500
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"epoch": 1.7,
|
| 893 |
+
"learning_rate": 4.4519024359891734e-05,
|
| 894 |
+
"loss": 0.6532,
|
| 895 |
+
"step": 73000
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"epoch": 1.71,
|
| 899 |
+
"learning_rate": 4.4480135643841585e-05,
|
| 900 |
+
"loss": 0.6529,
|
| 901 |
+
"step": 73500
|
| 902 |
+
},
|
| 903 |
+
{
|
| 904 |
+
"epoch": 1.72,
|
| 905 |
+
"learning_rate": 4.4441246927791437e-05,
|
| 906 |
+
"loss": 0.6451,
|
| 907 |
+
"step": 74000
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 1.73,
|
| 911 |
+
"learning_rate": 4.440235821174129e-05,
|
| 912 |
+
"loss": 0.6443,
|
| 913 |
+
"step": 74500
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 1.74,
|
| 917 |
+
"learning_rate": 4.436346949569113e-05,
|
| 918 |
+
"loss": 0.6456,
|
| 919 |
+
"step": 75000
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 1.75,
|
| 923 |
+
"learning_rate": 4.4324580779640984e-05,
|
| 924 |
+
"loss": 0.6515,
|
| 925 |
+
"step": 75500
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 1.77,
|
| 929 |
+
"learning_rate": 4.428569206359083e-05,
|
| 930 |
+
"loss": 0.6425,
|
| 931 |
+
"step": 76000
|
| 932 |
+
},
|
| 933 |
+
{
|
| 934 |
+
"epoch": 1.78,
|
| 935 |
+
"learning_rate": 4.424680334754068e-05,
|
| 936 |
+
"loss": 0.6394,
|
| 937 |
+
"step": 76500
|
| 938 |
+
},
|
| 939 |
+
{
|
| 940 |
+
"epoch": 1.79,
|
| 941 |
+
"learning_rate": 4.4207914631490525e-05,
|
| 942 |
+
"loss": 0.6433,
|
| 943 |
+
"step": 77000
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 1.8,
|
| 947 |
+
"learning_rate": 4.416902591544038e-05,
|
| 948 |
+
"loss": 0.6459,
|
| 949 |
+
"step": 77500
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 1.81,
|
| 953 |
+
"learning_rate": 4.4130214976822327e-05,
|
| 954 |
+
"loss": 0.6421,
|
| 955 |
+
"step": 78000
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 1.82,
|
| 959 |
+
"learning_rate": 4.409132626077218e-05,
|
| 960 |
+
"loss": 0.6428,
|
| 961 |
+
"step": 78500
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 1.84,
|
| 965 |
+
"learning_rate": 4.405251532215413e-05,
|
| 966 |
+
"loss": 0.6406,
|
| 967 |
+
"step": 79000
|
| 968 |
+
},
|
| 969 |
+
{
|
| 970 |
+
"epoch": 1.85,
|
| 971 |
+
"learning_rate": 4.401362660610397e-05,
|
| 972 |
+
"loss": 0.6357,
|
| 973 |
+
"step": 79500
|
| 974 |
+
},
|
| 975 |
+
{
|
| 976 |
+
"epoch": 1.86,
|
| 977 |
+
"learning_rate": 4.3974737890053824e-05,
|
| 978 |
+
"loss": 0.6355,
|
| 979 |
+
"step": 80000
|
| 980 |
+
},
|
| 981 |
+
{
|
| 982 |
+
"epoch": 1.87,
|
| 983 |
+
"learning_rate": 4.3935849174003676e-05,
|
| 984 |
+
"loss": 0.6459,
|
| 985 |
+
"step": 80500
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 1.88,
|
| 989 |
+
"learning_rate": 4.389696045795352e-05,
|
| 990 |
+
"loss": 0.6422,
|
| 991 |
+
"step": 81000
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 1.89,
|
| 995 |
+
"learning_rate": 4.385807174190337e-05,
|
| 996 |
+
"loss": 0.6413,
|
| 997 |
+
"step": 81500
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 1.91,
|
| 1001 |
+
"learning_rate": 4.381918302585322e-05,
|
| 1002 |
+
"loss": 0.6391,
|
| 1003 |
+
"step": 82000
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 1.92,
|
| 1007 |
+
"learning_rate": 4.3780294309803075e-05,
|
| 1008 |
+
"loss": 0.6409,
|
| 1009 |
+
"step": 82500
|
| 1010 |
+
},
|
| 1011 |
+
{
|
| 1012 |
+
"epoch": 1.93,
|
| 1013 |
+
"learning_rate": 4.374140559375292e-05,
|
| 1014 |
+
"loss": 0.632,
|
| 1015 |
+
"step": 83000
|
| 1016 |
+
},
|
| 1017 |
+
{
|
| 1018 |
+
"epoch": 1.94,
|
| 1019 |
+
"learning_rate": 4.370251687770277e-05,
|
| 1020 |
+
"loss": 0.6462,
|
| 1021 |
+
"step": 83500
|
| 1022 |
+
},
|
| 1023 |
+
{
|
| 1024 |
+
"epoch": 1.95,
|
| 1025 |
+
"learning_rate": 4.3663628161652616e-05,
|
| 1026 |
+
"loss": 0.6388,
|
| 1027 |
+
"step": 84000
|
| 1028 |
+
},
|
| 1029 |
+
{
|
| 1030 |
+
"epoch": 1.96,
|
| 1031 |
+
"learning_rate": 4.3624817223034566e-05,
|
| 1032 |
+
"loss": 0.642,
|
| 1033 |
+
"step": 84500
|
| 1034 |
+
},
|
| 1035 |
+
{
|
| 1036 |
+
"epoch": 1.98,
|
| 1037 |
+
"learning_rate": 4.358592850698442e-05,
|
| 1038 |
+
"loss": 0.6409,
|
| 1039 |
+
"step": 85000
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 1.99,
|
| 1043 |
+
"learning_rate": 4.354703979093426e-05,
|
| 1044 |
+
"loss": 0.6406,
|
| 1045 |
+
"step": 85500
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 2.0,
|
| 1049 |
+
"learning_rate": 4.350815107488411e-05,
|
| 1050 |
+
"loss": 0.6355,
|
| 1051 |
+
"step": 86000
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 2.0,
|
| 1055 |
+
"eval_bleu": 57.5735,
|
| 1056 |
+
"eval_gen_len": 15.9938,
|
| 1057 |
+
"eval_loss": 0.6641375422477722,
|
| 1058 |
+
"eval_runtime": 6884.7279,
|
| 1059 |
+
"eval_samples_per_second": 12.479,
|
| 1060 |
+
"eval_steps_per_second": 1.56,
|
| 1061 |
+
"step": 86048
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"epoch": 2.01,
|
| 1065 |
+
"learning_rate": 4.3469262358833965e-05,
|
| 1066 |
+
"loss": 0.5199,
|
| 1067 |
+
"step": 86500
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 2.02,
|
| 1071 |
+
"learning_rate": 4.343037364278381e-05,
|
| 1072 |
+
"loss": 0.4964,
|
| 1073 |
+
"step": 87000
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 2.03,
|
| 1077 |
+
"learning_rate": 4.339148492673366e-05,
|
| 1078 |
+
"loss": 0.5025,
|
| 1079 |
+
"step": 87500
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 2.05,
|
| 1083 |
+
"learning_rate": 4.335259621068351e-05,
|
| 1084 |
+
"loss": 0.508,
|
| 1085 |
+
"step": 88000
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 2.06,
|
| 1089 |
+
"learning_rate": 4.331378527206546e-05,
|
| 1090 |
+
"loss": 0.5096,
|
| 1091 |
+
"step": 88500
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 2.07,
|
| 1095 |
+
"learning_rate": 4.327497433344741e-05,
|
| 1096 |
+
"loss": 0.5106,
|
| 1097 |
+
"step": 89000
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"epoch": 2.08,
|
| 1101 |
+
"learning_rate": 4.323616339482936e-05,
|
| 1102 |
+
"loss": 0.5108,
|
| 1103 |
+
"step": 89500
|
| 1104 |
+
},
|
| 1105 |
+
{
|
| 1106 |
+
"epoch": 2.09,
|
| 1107 |
+
"learning_rate": 4.3197274678779206e-05,
|
| 1108 |
+
"loss": 0.5057,
|
| 1109 |
+
"step": 90000
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 2.1,
|
| 1113 |
+
"learning_rate": 4.3158463740161156e-05,
|
| 1114 |
+
"loss": 0.5108,
|
| 1115 |
+
"step": 90500
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 2.12,
|
| 1119 |
+
"learning_rate": 4.311957502411101e-05,
|
| 1120 |
+
"loss": 0.5044,
|
| 1121 |
+
"step": 91000
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 2.13,
|
| 1125 |
+
"learning_rate": 4.308068630806085e-05,
|
| 1126 |
+
"loss": 0.5077,
|
| 1127 |
+
"step": 91500
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 2.14,
|
| 1131 |
+
"learning_rate": 4.3041797592010704e-05,
|
| 1132 |
+
"loss": 0.5007,
|
| 1133 |
+
"step": 92000
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 2.15,
|
| 1137 |
+
"learning_rate": 4.300290887596055e-05,
|
| 1138 |
+
"loss": 0.5068,
|
| 1139 |
+
"step": 92500
|
| 1140 |
+
},
|
| 1141 |
+
{
|
| 1142 |
+
"epoch": 2.16,
|
| 1143 |
+
"learning_rate": 4.29640201599104e-05,
|
| 1144 |
+
"loss": 0.5038,
|
| 1145 |
+
"step": 93000
|
| 1146 |
+
},
|
| 1147 |
+
{
|
| 1148 |
+
"epoch": 2.17,
|
| 1149 |
+
"learning_rate": 4.292513144386025e-05,
|
| 1150 |
+
"loss": 0.5107,
|
| 1151 |
+
"step": 93500
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 2.18,
|
| 1155 |
+
"learning_rate": 4.28862427278101e-05,
|
| 1156 |
+
"loss": 0.5117,
|
| 1157 |
+
"step": 94000
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 2.2,
|
| 1161 |
+
"learning_rate": 4.284750956662415e-05,
|
| 1162 |
+
"loss": 0.5129,
|
| 1163 |
+
"step": 94500
|
| 1164 |
+
},
|
| 1165 |
+
{
|
| 1166 |
+
"epoch": 2.21,
|
| 1167 |
+
"learning_rate": 4.2808620850574e-05,
|
| 1168 |
+
"loss": 0.5119,
|
| 1169 |
+
"step": 95000
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 2.22,
|
| 1173 |
+
"learning_rate": 4.276973213452385e-05,
|
| 1174 |
+
"loss": 0.5198,
|
| 1175 |
+
"step": 95500
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 2.23,
|
| 1179 |
+
"learning_rate": 4.27308434184737e-05,
|
| 1180 |
+
"loss": 0.5138,
|
| 1181 |
+
"step": 96000
|
| 1182 |
+
},
|
| 1183 |
+
{
|
| 1184 |
+
"epoch": 2.24,
|
| 1185 |
+
"learning_rate": 4.269195470242355e-05,
|
| 1186 |
+
"loss": 0.516,
|
| 1187 |
+
"step": 96500
|
| 1188 |
+
},
|
| 1189 |
+
{
|
| 1190 |
+
"epoch": 2.25,
|
| 1191 |
+
"learning_rate": 4.2653065986373395e-05,
|
| 1192 |
+
"loss": 0.521,
|
| 1193 |
+
"step": 97000
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 2.27,
|
| 1197 |
+
"learning_rate": 4.2614177270323247e-05,
|
| 1198 |
+
"loss": 0.5162,
|
| 1199 |
+
"step": 97500
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 2.28,
|
| 1203 |
+
"learning_rate": 4.25752885542731e-05,
|
| 1204 |
+
"loss": 0.5175,
|
| 1205 |
+
"step": 98000
|
| 1206 |
+
},
|
| 1207 |
+
{
|
| 1208 |
+
"epoch": 2.29,
|
| 1209 |
+
"learning_rate": 4.253639983822294e-05,
|
| 1210 |
+
"loss": 0.518,
|
| 1211 |
+
"step": 98500
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"epoch": 2.3,
|
| 1215 |
+
"learning_rate": 4.249758889960489e-05,
|
| 1216 |
+
"loss": 0.5234,
|
| 1217 |
+
"step": 99000
|
| 1218 |
+
},
|
| 1219 |
+
{
|
| 1220 |
+
"epoch": 2.31,
|
| 1221 |
+
"learning_rate": 4.245870018355474e-05,
|
| 1222 |
+
"loss": 0.524,
|
| 1223 |
+
"step": 99500
|
| 1224 |
+
},
|
| 1225 |
+
{
|
| 1226 |
+
"epoch": 2.32,
|
| 1227 |
+
"learning_rate": 4.241981146750459e-05,
|
| 1228 |
+
"loss": 0.5263,
|
| 1229 |
+
"step": 100000
|
| 1230 |
+
},
|
| 1231 |
+
{
|
| 1232 |
+
"epoch": 2.34,
|
| 1233 |
+
"learning_rate": 4.238092275145444e-05,
|
| 1234 |
+
"loss": 0.5171,
|
| 1235 |
+
"step": 100500
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 2.35,
|
| 1239 |
+
"learning_rate": 4.2342034035404285e-05,
|
| 1240 |
+
"loss": 0.5265,
|
| 1241 |
+
"step": 101000
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 2.36,
|
| 1245 |
+
"learning_rate": 4.2303145319354137e-05,
|
| 1246 |
+
"loss": 0.5156,
|
| 1247 |
+
"step": 101500
|
| 1248 |
+
},
|
| 1249 |
+
{
|
| 1250 |
+
"epoch": 2.37,
|
| 1251 |
+
"learning_rate": 4.226425660330399e-05,
|
| 1252 |
+
"loss": 0.5255,
|
| 1253 |
+
"step": 102000
|
| 1254 |
+
},
|
| 1255 |
+
{
|
| 1256 |
+
"epoch": 2.38,
|
| 1257 |
+
"learning_rate": 4.222536788725383e-05,
|
| 1258 |
+
"loss": 0.5187,
|
| 1259 |
+
"step": 102500
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"epoch": 2.39,
|
| 1263 |
+
"learning_rate": 4.218655694863579e-05,
|
| 1264 |
+
"loss": 0.5298,
|
| 1265 |
+
"step": 103000
|
| 1266 |
+
},
|
| 1267 |
+
{
|
| 1268 |
+
"epoch": 2.41,
|
| 1269 |
+
"learning_rate": 4.2147668232585634e-05,
|
| 1270 |
+
"loss": 0.5258,
|
| 1271 |
+
"step": 103500
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"epoch": 2.42,
|
| 1275 |
+
"learning_rate": 4.2108779516535485e-05,
|
| 1276 |
+
"loss": 0.5206,
|
| 1277 |
+
"step": 104000
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 2.43,
|
| 1281 |
+
"learning_rate": 4.206989080048534e-05,
|
| 1282 |
+
"loss": 0.5292,
|
| 1283 |
+
"step": 104500
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 2.44,
|
| 1287 |
+
"learning_rate": 4.203100208443518e-05,
|
| 1288 |
+
"loss": 0.5156,
|
| 1289 |
+
"step": 105000
|
| 1290 |
+
},
|
| 1291 |
+
{
|
| 1292 |
+
"epoch": 2.45,
|
| 1293 |
+
"learning_rate": 4.199211336838503e-05,
|
| 1294 |
+
"loss": 0.5156,
|
| 1295 |
+
"step": 105500
|
| 1296 |
+
},
|
| 1297 |
+
{
|
| 1298 |
+
"epoch": 2.46,
|
| 1299 |
+
"learning_rate": 4.1953224652334885e-05,
|
| 1300 |
+
"loss": 0.5259,
|
| 1301 |
+
"step": 106000
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"epoch": 2.48,
|
| 1305 |
+
"learning_rate": 4.191433593628473e-05,
|
| 1306 |
+
"loss": 0.5236,
|
| 1307 |
+
"step": 106500
|
| 1308 |
+
},
|
| 1309 |
+
{
|
| 1310 |
+
"epoch": 2.49,
|
| 1311 |
+
"learning_rate": 4.187552499766668e-05,
|
| 1312 |
+
"loss": 0.5282,
|
| 1313 |
+
"step": 107000
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 2.5,
|
| 1317 |
+
"learning_rate": 4.1836636281616524e-05,
|
| 1318 |
+
"loss": 0.5342,
|
| 1319 |
+
"step": 107500
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 2.51,
|
| 1323 |
+
"learning_rate": 4.1797747565566375e-05,
|
| 1324 |
+
"loss": 0.5246,
|
| 1325 |
+
"step": 108000
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 2.52,
|
| 1329 |
+
"learning_rate": 4.175885884951623e-05,
|
| 1330 |
+
"loss": 0.5238,
|
| 1331 |
+
"step": 108500
|
| 1332 |
+
},
|
| 1333 |
+
{
|
| 1334 |
+
"epoch": 2.53,
|
| 1335 |
+
"learning_rate": 4.171997013346607e-05,
|
| 1336 |
+
"loss": 0.5329,
|
| 1337 |
+
"step": 109000
|
| 1338 |
+
},
|
| 1339 |
+
{
|
| 1340 |
+
"epoch": 2.55,
|
| 1341 |
+
"learning_rate": 4.168108141741592e-05,
|
| 1342 |
+
"loss": 0.5223,
|
| 1343 |
+
"step": 109500
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 2.56,
|
| 1347 |
+
"learning_rate": 4.164234825622997e-05,
|
| 1348 |
+
"loss": 0.5333,
|
| 1349 |
+
"step": 110000
|
| 1350 |
+
},
|
| 1351 |
+
{
|
| 1352 |
+
"epoch": 2.57,
|
| 1353 |
+
"learning_rate": 4.160345954017982e-05,
|
| 1354 |
+
"loss": 0.5353,
|
| 1355 |
+
"step": 110500
|
| 1356 |
+
},
|
| 1357 |
+
{
|
| 1358 |
+
"epoch": 2.58,
|
| 1359 |
+
"learning_rate": 4.1564570824129674e-05,
|
| 1360 |
+
"loss": 0.5285,
|
| 1361 |
+
"step": 111000
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 2.59,
|
| 1365 |
+
"learning_rate": 4.152568210807952e-05,
|
| 1366 |
+
"loss": 0.5365,
|
| 1367 |
+
"step": 111500
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 2.6,
|
| 1371 |
+
"learning_rate": 4.1486948946893573e-05,
|
| 1372 |
+
"loss": 0.5214,
|
| 1373 |
+
"step": 112000
|
| 1374 |
+
},
|
| 1375 |
+
{
|
| 1376 |
+
"epoch": 2.61,
|
| 1377 |
+
"learning_rate": 4.1448060230843425e-05,
|
| 1378 |
+
"loss": 0.534,
|
| 1379 |
+
"step": 112500
|
| 1380 |
+
},
|
| 1381 |
+
{
|
| 1382 |
+
"epoch": 2.63,
|
| 1383 |
+
"learning_rate": 4.140917151479327e-05,
|
| 1384 |
+
"loss": 0.5375,
|
| 1385 |
+
"step": 113000
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"epoch": 2.64,
|
| 1389 |
+
"learning_rate": 4.137028279874312e-05,
|
| 1390 |
+
"loss": 0.532,
|
| 1391 |
+
"step": 113500
|
| 1392 |
+
},
|
| 1393 |
+
{
|
| 1394 |
+
"epoch": 2.65,
|
| 1395 |
+
"learning_rate": 4.133139408269297e-05,
|
| 1396 |
+
"loss": 0.5232,
|
| 1397 |
+
"step": 114000
|
| 1398 |
+
},
|
| 1399 |
+
{
|
| 1400 |
+
"epoch": 2.66,
|
| 1401 |
+
"learning_rate": 4.129250536664282e-05,
|
| 1402 |
+
"loss": 0.5333,
|
| 1403 |
+
"step": 114500
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 2.67,
|
| 1407 |
+
"learning_rate": 4.125361665059266e-05,
|
| 1408 |
+
"loss": 0.5269,
|
| 1409 |
+
"step": 115000
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 2.68,
|
| 1413 |
+
"learning_rate": 4.1214727934542514e-05,
|
| 1414 |
+
"loss": 0.5186,
|
| 1415 |
+
"step": 115500
|
| 1416 |
+
},
|
| 1417 |
+
{
|
| 1418 |
+
"epoch": 2.7,
|
| 1419 |
+
"learning_rate": 4.1175839218492365e-05,
|
| 1420 |
+
"loss": 0.5226,
|
| 1421 |
+
"step": 116000
|
| 1422 |
+
},
|
| 1423 |
+
{
|
| 1424 |
+
"epoch": 2.71,
|
| 1425 |
+
"learning_rate": 4.1137028279874315e-05,
|
| 1426 |
+
"loss": 0.5278,
|
| 1427 |
+
"step": 116500
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 2.72,
|
| 1431 |
+
"learning_rate": 4.109813956382416e-05,
|
| 1432 |
+
"loss": 0.5311,
|
| 1433 |
+
"step": 117000
|
| 1434 |
+
},
|
| 1435 |
+
{
|
| 1436 |
+
"epoch": 2.73,
|
| 1437 |
+
"learning_rate": 4.105925084777401e-05,
|
| 1438 |
+
"loss": 0.5274,
|
| 1439 |
+
"step": 117500
|
| 1440 |
+
},
|
| 1441 |
+
{
|
| 1442 |
+
"epoch": 2.74,
|
| 1443 |
+
"learning_rate": 4.102036213172386e-05,
|
| 1444 |
+
"loss": 0.5348,
|
| 1445 |
+
"step": 118000
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 2.75,
|
| 1449 |
+
"learning_rate": 4.098147341567371e-05,
|
| 1450 |
+
"loss": 0.5359,
|
| 1451 |
+
"step": 118500
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 2.77,
|
| 1455 |
+
"learning_rate": 4.094258469962356e-05,
|
| 1456 |
+
"loss": 0.5416,
|
| 1457 |
+
"step": 119000
|
| 1458 |
+
},
|
| 1459 |
+
{
|
| 1460 |
+
"epoch": 2.78,
|
| 1461 |
+
"learning_rate": 4.090369598357341e-05,
|
| 1462 |
+
"loss": 0.526,
|
| 1463 |
+
"step": 119500
|
| 1464 |
+
},
|
| 1465 |
+
{
|
| 1466 |
+
"epoch": 2.79,
|
| 1467 |
+
"learning_rate": 4.086488504495536e-05,
|
| 1468 |
+
"loss": 0.5404,
|
| 1469 |
+
"step": 120000
|
| 1470 |
+
},
|
| 1471 |
+
{
|
| 1472 |
+
"epoch": 2.8,
|
| 1473 |
+
"learning_rate": 4.0825996328905205e-05,
|
| 1474 |
+
"loss": 0.5321,
|
| 1475 |
+
"step": 120500
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"epoch": 2.81,
|
| 1479 |
+
"learning_rate": 4.0787107612855056e-05,
|
| 1480 |
+
"loss": 0.5378,
|
| 1481 |
+
"step": 121000
|
| 1482 |
+
},
|
| 1483 |
+
{
|
| 1484 |
+
"epoch": 2.82,
|
| 1485 |
+
"learning_rate": 4.074821889680491e-05,
|
| 1486 |
+
"loss": 0.5371,
|
| 1487 |
+
"step": 121500
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 2.84,
|
| 1491 |
+
"learning_rate": 4.070933018075476e-05,
|
| 1492 |
+
"loss": 0.5303,
|
| 1493 |
+
"step": 122000
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 2.85,
|
| 1497 |
+
"learning_rate": 4.06704414647046e-05,
|
| 1498 |
+
"loss": 0.5321,
|
| 1499 |
+
"step": 122500
|
| 1500 |
+
},
|
| 1501 |
+
{
|
| 1502 |
+
"epoch": 2.86,
|
| 1503 |
+
"learning_rate": 4.063163052608655e-05,
|
| 1504 |
+
"loss": 0.5291,
|
| 1505 |
+
"step": 123000
|
| 1506 |
+
},
|
| 1507 |
+
{
|
| 1508 |
+
"epoch": 2.87,
|
| 1509 |
+
"learning_rate": 4.05927418100364e-05,
|
| 1510 |
+
"loss": 0.5391,
|
| 1511 |
+
"step": 123500
|
| 1512 |
+
},
|
| 1513 |
+
{
|
| 1514 |
+
"epoch": 2.88,
|
| 1515 |
+
"learning_rate": 4.055385309398625e-05,
|
| 1516 |
+
"loss": 0.5344,
|
| 1517 |
+
"step": 124000
|
| 1518 |
+
},
|
| 1519 |
+
{
|
| 1520 |
+
"epoch": 2.89,
|
| 1521 |
+
"learning_rate": 4.05149643779361e-05,
|
| 1522 |
+
"loss": 0.5381,
|
| 1523 |
+
"step": 124500
|
| 1524 |
+
},
|
| 1525 |
+
{
|
| 1526 |
+
"epoch": 2.91,
|
| 1527 |
+
"learning_rate": 4.0476075661885946e-05,
|
| 1528 |
+
"loss": 0.5348,
|
| 1529 |
+
"step": 125000
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 2.92,
|
| 1533 |
+
"learning_rate": 4.04371869458358e-05,
|
| 1534 |
+
"loss": 0.5367,
|
| 1535 |
+
"step": 125500
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 2.93,
|
| 1539 |
+
"learning_rate": 4.039829822978565e-05,
|
| 1540 |
+
"loss": 0.5294,
|
| 1541 |
+
"step": 126000
|
| 1542 |
+
},
|
| 1543 |
+
{
|
| 1544 |
+
"epoch": 2.94,
|
| 1545 |
+
"learning_rate": 4.0359409513735494e-05,
|
| 1546 |
+
"loss": 0.5257,
|
| 1547 |
+
"step": 126500
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"epoch": 2.95,
|
| 1551 |
+
"learning_rate": 4.0320520797685346e-05,
|
| 1552 |
+
"loss": 0.5241,
|
| 1553 |
+
"step": 127000
|
| 1554 |
+
},
|
| 1555 |
+
{
|
| 1556 |
+
"epoch": 2.96,
|
| 1557 |
+
"learning_rate": 4.02816320816352e-05,
|
| 1558 |
+
"loss": 0.5375,
|
| 1559 |
+
"step": 127500
|
| 1560 |
+
},
|
| 1561 |
+
{
|
| 1562 |
+
"epoch": 2.98,
|
| 1563 |
+
"learning_rate": 4.024274336558504e-05,
|
| 1564 |
+
"loss": 0.5332,
|
| 1565 |
+
"step": 128000
|
| 1566 |
+
},
|
| 1567 |
+
{
|
| 1568 |
+
"epoch": 2.99,
|
| 1569 |
+
"learning_rate": 4.0203854649534894e-05,
|
| 1570 |
+
"loss": 0.532,
|
| 1571 |
+
"step": 128500
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 3.0,
|
| 1575 |
+
"learning_rate": 4.0164965933484745e-05,
|
| 1576 |
+
"loss": 0.5464,
|
| 1577 |
+
"step": 129000
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 3.0,
|
| 1581 |
+
"eval_bleu": 58.837,
|
| 1582 |
+
"eval_gen_len": 15.7689,
|
| 1583 |
+
"eval_loss": 0.6568061113357544,
|
| 1584 |
+
"eval_runtime": 6740.2004,
|
| 1585 |
+
"eval_samples_per_second": 12.747,
|
| 1586 |
+
"eval_steps_per_second": 1.593,
|
| 1587 |
+
"step": 129072
|
| 1588 |
+
}
|
| 1589 |
+
],
|
| 1590 |
+
"logging_steps": 500,
|
| 1591 |
+
"max_steps": 645360,
|
| 1592 |
+
"num_input_tokens_seen": 0,
|
| 1593 |
+
"num_train_epochs": 15,
|
| 1594 |
+
"save_steps": 500,
|
| 1595 |
+
"total_flos": 4.4754523745125663e+18,
|
| 1596 |
+
"train_batch_size": 2,
|
| 1597 |
+
"trial_name": null,
|
| 1598 |
+
"trial_params": null
|
| 1599 |
+
}
|
checkpoint-129072/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4694a20062ed70ba8adc7c9ea1e09ea766481ac4f0f03605243799bfb9894d0a
|
| 3 |
+
size 4856
|
checkpoint-215120/config.json
ADDED
|
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/mbart-large-50-many-to-many-mmt",
|
| 3 |
+
"_num_labels": 3,
|
| 4 |
+
"activation_dropout": 0.0,
|
| 5 |
+
"activation_function": "relu",
|
| 6 |
+
"add_bias_logits": false,
|
| 7 |
+
"add_final_layer_norm": true,
|
| 8 |
+
"architectures": [
|
| 9 |
+
"MBartForConditionalGeneration"
|
| 10 |
+
],
|
| 11 |
+
"attention_dropout": 0.0,
|
| 12 |
+
"bos_token_id": 0,
|
| 13 |
+
"classif_dropout": 0.0,
|
| 14 |
+
"classifier_dropout": 0.0,
|
| 15 |
+
"d_model": 1024,
|
| 16 |
+
"decoder_attention_heads": 16,
|
| 17 |
+
"decoder_ffn_dim": 4096,
|
| 18 |
+
"decoder_layerdrop": 0.0,
|
| 19 |
+
"decoder_layers": 12,
|
| 20 |
+
"decoder_start_token_id": 2,
|
| 21 |
+
"dropout": 0.1,
|
| 22 |
+
"early_stopping": true,
|
| 23 |
+
"encoder_attention_heads": 16,
|
| 24 |
+
"encoder_ffn_dim": 4096,
|
| 25 |
+
"encoder_layerdrop": 0.0,
|
| 26 |
+
"encoder_layers": 12,
|
| 27 |
+
"eos_token_id": 2,
|
| 28 |
+
"forced_bos_token_id": 250014,
|
| 29 |
+
"forced_eos_token_id": 2,
|
| 30 |
+
"gradient_checkpointing": false,
|
| 31 |
+
"id2label": {
|
| 32 |
+
"0": "LABEL_0",
|
| 33 |
+
"1": "LABEL_1",
|
| 34 |
+
"2": "LABEL_2"
|
| 35 |
+
},
|
| 36 |
+
"init_std": 0.02,
|
| 37 |
+
"is_encoder_decoder": true,
|
| 38 |
+
"label2id": {
|
| 39 |
+
"LABEL_0": 0,
|
| 40 |
+
"LABEL_1": 1,
|
| 41 |
+
"LABEL_2": 2
|
| 42 |
+
},
|
| 43 |
+
"max_length": 200,
|
| 44 |
+
"max_position_embeddings": 1024,
|
| 45 |
+
"model_type": "mbart",
|
| 46 |
+
"normalize_before": true,
|
| 47 |
+
"normalize_embedding": true,
|
| 48 |
+
"num_beams": 5,
|
| 49 |
+
"num_hidden_layers": 12,
|
| 50 |
+
"output_past": true,
|
| 51 |
+
"pad_token_id": 1,
|
| 52 |
+
"scale_embedding": true,
|
| 53 |
+
"static_position_embeddings": false,
|
| 54 |
+
"tokenizer_class": "MBart50Tokenizer",
|
| 55 |
+
"torch_dtype": "float32",
|
| 56 |
+
"transformers_version": "4.36.2",
|
| 57 |
+
"use_cache": true,
|
| 58 |
+
"vocab_size": 250054
|
| 59 |
+
}
|
checkpoint-215120/generation_config.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token_id": 0,
|
| 3 |
+
"decoder_start_token_id": 2,
|
| 4 |
+
"early_stopping": true,
|
| 5 |
+
"eos_token_id": 2,
|
| 6 |
+
"forced_bos_token_id": 250014,
|
| 7 |
+
"forced_eos_token_id": 2,
|
| 8 |
+
"max_length": 200,
|
| 9 |
+
"num_beams": 5,
|
| 10 |
+
"pad_token_id": 1,
|
| 11 |
+
"transformers_version": "4.36.2"
|
| 12 |
+
}
|
checkpoint-215120/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8ab0c3248380483ef7b42adc1935403fadf0b932ccbb4497b114b5176fdc8353
|
| 3 |
+
size 2444578688
|
checkpoint-215120/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f5ebf0215b8af0f705b0824f0fea9ca79305d851699083ba35838f4cc5312d00
|
| 3 |
+
size 4887473903
|
checkpoint-215120/rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:90afda3b41d4e1d5b718b17e4287a79d26059cd7929583a42fc844f9c6f9ddf2
|
| 3 |
+
size 15024
|
checkpoint-215120/rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fd2749bd611f668743db6e8a9771a369692f81d295285a23f4d5142695472177
|
| 3 |
+
size 15024
|
checkpoint-215120/rng_state_2.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:665722a661f8fa13c0dbb02bf7a5aaf719a925797d4eb28f57454001eded76ed
|
| 3 |
+
size 15024
|
checkpoint-215120/rng_state_3.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:32ad42b6e278e9400a4b9a0ee7830500887c73ffb8c675aa9e14c00f438a5fba
|
| 3 |
+
size 15024
|
checkpoint-215120/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b4f8235f56a10903e394d5c91084b4007850ea75232271c3a50f52513ae1299d
|
| 3 |
+
size 1064
|
checkpoint-215120/sentencepiece.bpe.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cfc8146abe2a0488e9e2a0c56de7952f7c11ab059eca145a0a727afce0db2865
|
| 3 |
+
size 5069051
|
checkpoint-215120/special_tokens_map.json
ADDED
|
@@ -0,0 +1,69 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"additional_special_tokens": [
|
| 3 |
+
"ar_AR",
|
| 4 |
+
"cs_CZ",
|
| 5 |
+
"de_DE",
|
| 6 |
+
"en_XX",
|
| 7 |
+
"es_XX",
|
| 8 |
+
"et_EE",
|
| 9 |
+
"fi_FI",
|
| 10 |
+
"fr_XX",
|
| 11 |
+
"gu_IN",
|
| 12 |
+
"hi_IN",
|
| 13 |
+
"it_IT",
|
| 14 |
+
"ja_XX",
|
| 15 |
+
"kk_KZ",
|
| 16 |
+
"ko_KR",
|
| 17 |
+
"lt_LT",
|
| 18 |
+
"lv_LV",
|
| 19 |
+
"my_MM",
|
| 20 |
+
"ne_NP",
|
| 21 |
+
"nl_XX",
|
| 22 |
+
"ro_RO",
|
| 23 |
+
"ru_RU",
|
| 24 |
+
"si_LK",
|
| 25 |
+
"tr_TR",
|
| 26 |
+
"vi_VN",
|
| 27 |
+
"zh_CN",
|
| 28 |
+
"af_ZA",
|
| 29 |
+
"az_AZ",
|
| 30 |
+
"bn_IN",
|
| 31 |
+
"fa_IR",
|
| 32 |
+
"he_IL",
|
| 33 |
+
"hr_HR",
|
| 34 |
+
"id_ID",
|
| 35 |
+
"ka_GE",
|
| 36 |
+
"km_KH",
|
| 37 |
+
"mk_MK",
|
| 38 |
+
"ml_IN",
|
| 39 |
+
"mn_MN",
|
| 40 |
+
"mr_IN",
|
| 41 |
+
"pl_PL",
|
| 42 |
+
"ps_AF",
|
| 43 |
+
"pt_XX",
|
| 44 |
+
"sv_SE",
|
| 45 |
+
"sw_KE",
|
| 46 |
+
"ta_IN",
|
| 47 |
+
"te_IN",
|
| 48 |
+
"th_TH",
|
| 49 |
+
"tl_XX",
|
| 50 |
+
"uk_UA",
|
| 51 |
+
"ur_PK",
|
| 52 |
+
"xh_ZA",
|
| 53 |
+
"gl_ES",
|
| 54 |
+
"sl_SI"
|
| 55 |
+
],
|
| 56 |
+
"bos_token": "<s>",
|
| 57 |
+
"cls_token": "<s>",
|
| 58 |
+
"eos_token": "</s>",
|
| 59 |
+
"mask_token": {
|
| 60 |
+
"content": "<mask>",
|
| 61 |
+
"lstrip": true,
|
| 62 |
+
"normalized": true,
|
| 63 |
+
"rstrip": false,
|
| 64 |
+
"single_word": false
|
| 65 |
+
},
|
| 66 |
+
"pad_token": "<pad>",
|
| 67 |
+
"sep_token": "</s>",
|
| 68 |
+
"unk_token": "<unk>"
|
| 69 |
+
}
|
checkpoint-215120/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:faf1fb0f45fdd5ae9fc44fafee492652a423fc749833ef57c2fa801358549d92
|
| 3 |
+
size 17110107
|
checkpoint-215120/tokenizer_config.json
ADDED
|
@@ -0,0 +1,528 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"250001": {
|
| 36 |
+
"content": "ar_AR",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
},
|
| 43 |
+
"250002": {
|
| 44 |
+
"content": "cs_CZ",
|
| 45 |
+
"lstrip": false,
|
| 46 |
+
"normalized": false,
|
| 47 |
+
"rstrip": false,
|
| 48 |
+
"single_word": false,
|
| 49 |
+
"special": true
|
| 50 |
+
},
|
| 51 |
+
"250003": {
|
| 52 |
+
"content": "de_DE",
|
| 53 |
+
"lstrip": false,
|
| 54 |
+
"normalized": false,
|
| 55 |
+
"rstrip": false,
|
| 56 |
+
"single_word": false,
|
| 57 |
+
"special": true
|
| 58 |
+
},
|
| 59 |
+
"250004": {
|
| 60 |
+
"content": "en_XX",
|
| 61 |
+
"lstrip": false,
|
| 62 |
+
"normalized": false,
|
| 63 |
+
"rstrip": false,
|
| 64 |
+
"single_word": false,
|
| 65 |
+
"special": true
|
| 66 |
+
},
|
| 67 |
+
"250005": {
|
| 68 |
+
"content": "es_XX",
|
| 69 |
+
"lstrip": false,
|
| 70 |
+
"normalized": false,
|
| 71 |
+
"rstrip": false,
|
| 72 |
+
"single_word": false,
|
| 73 |
+
"special": true
|
| 74 |
+
},
|
| 75 |
+
"250006": {
|
| 76 |
+
"content": "et_EE",
|
| 77 |
+
"lstrip": false,
|
| 78 |
+
"normalized": false,
|
| 79 |
+
"rstrip": false,
|
| 80 |
+
"single_word": false,
|
| 81 |
+
"special": true
|
| 82 |
+
},
|
| 83 |
+
"250007": {
|
| 84 |
+
"content": "fi_FI",
|
| 85 |
+
"lstrip": false,
|
| 86 |
+
"normalized": false,
|
| 87 |
+
"rstrip": false,
|
| 88 |
+
"single_word": false,
|
| 89 |
+
"special": true
|
| 90 |
+
},
|
| 91 |
+
"250008": {
|
| 92 |
+
"content": "fr_XX",
|
| 93 |
+
"lstrip": false,
|
| 94 |
+
"normalized": false,
|
| 95 |
+
"rstrip": false,
|
| 96 |
+
"single_word": false,
|
| 97 |
+
"special": true
|
| 98 |
+
},
|
| 99 |
+
"250009": {
|
| 100 |
+
"content": "gu_IN",
|
| 101 |
+
"lstrip": false,
|
| 102 |
+
"normalized": false,
|
| 103 |
+
"rstrip": false,
|
| 104 |
+
"single_word": false,
|
| 105 |
+
"special": true
|
| 106 |
+
},
|
| 107 |
+
"250010": {
|
| 108 |
+
"content": "hi_IN",
|
| 109 |
+
"lstrip": false,
|
| 110 |
+
"normalized": false,
|
| 111 |
+
"rstrip": false,
|
| 112 |
+
"single_word": false,
|
| 113 |
+
"special": true
|
| 114 |
+
},
|
| 115 |
+
"250011": {
|
| 116 |
+
"content": "it_IT",
|
| 117 |
+
"lstrip": false,
|
| 118 |
+
"normalized": false,
|
| 119 |
+
"rstrip": false,
|
| 120 |
+
"single_word": false,
|
| 121 |
+
"special": true
|
| 122 |
+
},
|
| 123 |
+
"250012": {
|
| 124 |
+
"content": "ja_XX",
|
| 125 |
+
"lstrip": false,
|
| 126 |
+
"normalized": false,
|
| 127 |
+
"rstrip": false,
|
| 128 |
+
"single_word": false,
|
| 129 |
+
"special": true
|
| 130 |
+
},
|
| 131 |
+
"250013": {
|
| 132 |
+
"content": "kk_KZ",
|
| 133 |
+
"lstrip": false,
|
| 134 |
+
"normalized": false,
|
| 135 |
+
"rstrip": false,
|
| 136 |
+
"single_word": false,
|
| 137 |
+
"special": true
|
| 138 |
+
},
|
| 139 |
+
"250014": {
|
| 140 |
+
"content": "ko_KR",
|
| 141 |
+
"lstrip": false,
|
| 142 |
+
"normalized": false,
|
| 143 |
+
"rstrip": false,
|
| 144 |
+
"single_word": false,
|
| 145 |
+
"special": true
|
| 146 |
+
},
|
| 147 |
+
"250015": {
|
| 148 |
+
"content": "lt_LT",
|
| 149 |
+
"lstrip": false,
|
| 150 |
+
"normalized": false,
|
| 151 |
+
"rstrip": false,
|
| 152 |
+
"single_word": false,
|
| 153 |
+
"special": true
|
| 154 |
+
},
|
| 155 |
+
"250016": {
|
| 156 |
+
"content": "lv_LV",
|
| 157 |
+
"lstrip": false,
|
| 158 |
+
"normalized": false,
|
| 159 |
+
"rstrip": false,
|
| 160 |
+
"single_word": false,
|
| 161 |
+
"special": true
|
| 162 |
+
},
|
| 163 |
+
"250017": {
|
| 164 |
+
"content": "my_MM",
|
| 165 |
+
"lstrip": false,
|
| 166 |
+
"normalized": false,
|
| 167 |
+
"rstrip": false,
|
| 168 |
+
"single_word": false,
|
| 169 |
+
"special": true
|
| 170 |
+
},
|
| 171 |
+
"250018": {
|
| 172 |
+
"content": "ne_NP",
|
| 173 |
+
"lstrip": false,
|
| 174 |
+
"normalized": false,
|
| 175 |
+
"rstrip": false,
|
| 176 |
+
"single_word": false,
|
| 177 |
+
"special": true
|
| 178 |
+
},
|
| 179 |
+
"250019": {
|
| 180 |
+
"content": "nl_XX",
|
| 181 |
+
"lstrip": false,
|
| 182 |
+
"normalized": false,
|
| 183 |
+
"rstrip": false,
|
| 184 |
+
"single_word": false,
|
| 185 |
+
"special": true
|
| 186 |
+
},
|
| 187 |
+
"250020": {
|
| 188 |
+
"content": "ro_RO",
|
| 189 |
+
"lstrip": false,
|
| 190 |
+
"normalized": false,
|
| 191 |
+
"rstrip": false,
|
| 192 |
+
"single_word": false,
|
| 193 |
+
"special": true
|
| 194 |
+
},
|
| 195 |
+
"250021": {
|
| 196 |
+
"content": "ru_RU",
|
| 197 |
+
"lstrip": false,
|
| 198 |
+
"normalized": false,
|
| 199 |
+
"rstrip": false,
|
| 200 |
+
"single_word": false,
|
| 201 |
+
"special": true
|
| 202 |
+
},
|
| 203 |
+
"250022": {
|
| 204 |
+
"content": "si_LK",
|
| 205 |
+
"lstrip": false,
|
| 206 |
+
"normalized": false,
|
| 207 |
+
"rstrip": false,
|
| 208 |
+
"single_word": false,
|
| 209 |
+
"special": true
|
| 210 |
+
},
|
| 211 |
+
"250023": {
|
| 212 |
+
"content": "tr_TR",
|
| 213 |
+
"lstrip": false,
|
| 214 |
+
"normalized": false,
|
| 215 |
+
"rstrip": false,
|
| 216 |
+
"single_word": false,
|
| 217 |
+
"special": true
|
| 218 |
+
},
|
| 219 |
+
"250024": {
|
| 220 |
+
"content": "vi_VN",
|
| 221 |
+
"lstrip": false,
|
| 222 |
+
"normalized": false,
|
| 223 |
+
"rstrip": false,
|
| 224 |
+
"single_word": false,
|
| 225 |
+
"special": true
|
| 226 |
+
},
|
| 227 |
+
"250025": {
|
| 228 |
+
"content": "zh_CN",
|
| 229 |
+
"lstrip": false,
|
| 230 |
+
"normalized": false,
|
| 231 |
+
"rstrip": false,
|
| 232 |
+
"single_word": false,
|
| 233 |
+
"special": true
|
| 234 |
+
},
|
| 235 |
+
"250026": {
|
| 236 |
+
"content": "af_ZA",
|
| 237 |
+
"lstrip": false,
|
| 238 |
+
"normalized": false,
|
| 239 |
+
"rstrip": false,
|
| 240 |
+
"single_word": false,
|
| 241 |
+
"special": true
|
| 242 |
+
},
|
| 243 |
+
"250027": {
|
| 244 |
+
"content": "az_AZ",
|
| 245 |
+
"lstrip": false,
|
| 246 |
+
"normalized": false,
|
| 247 |
+
"rstrip": false,
|
| 248 |
+
"single_word": false,
|
| 249 |
+
"special": true
|
| 250 |
+
},
|
| 251 |
+
"250028": {
|
| 252 |
+
"content": "bn_IN",
|
| 253 |
+
"lstrip": false,
|
| 254 |
+
"normalized": false,
|
| 255 |
+
"rstrip": false,
|
| 256 |
+
"single_word": false,
|
| 257 |
+
"special": true
|
| 258 |
+
},
|
| 259 |
+
"250029": {
|
| 260 |
+
"content": "fa_IR",
|
| 261 |
+
"lstrip": false,
|
| 262 |
+
"normalized": false,
|
| 263 |
+
"rstrip": false,
|
| 264 |
+
"single_word": false,
|
| 265 |
+
"special": true
|
| 266 |
+
},
|
| 267 |
+
"250030": {
|
| 268 |
+
"content": "he_IL",
|
| 269 |
+
"lstrip": false,
|
| 270 |
+
"normalized": false,
|
| 271 |
+
"rstrip": false,
|
| 272 |
+
"single_word": false,
|
| 273 |
+
"special": true
|
| 274 |
+
},
|
| 275 |
+
"250031": {
|
| 276 |
+
"content": "hr_HR",
|
| 277 |
+
"lstrip": false,
|
| 278 |
+
"normalized": false,
|
| 279 |
+
"rstrip": false,
|
| 280 |
+
"single_word": false,
|
| 281 |
+
"special": true
|
| 282 |
+
},
|
| 283 |
+
"250032": {
|
| 284 |
+
"content": "id_ID",
|
| 285 |
+
"lstrip": false,
|
| 286 |
+
"normalized": false,
|
| 287 |
+
"rstrip": false,
|
| 288 |
+
"single_word": false,
|
| 289 |
+
"special": true
|
| 290 |
+
},
|
| 291 |
+
"250033": {
|
| 292 |
+
"content": "ka_GE",
|
| 293 |
+
"lstrip": false,
|
| 294 |
+
"normalized": false,
|
| 295 |
+
"rstrip": false,
|
| 296 |
+
"single_word": false,
|
| 297 |
+
"special": true
|
| 298 |
+
},
|
| 299 |
+
"250034": {
|
| 300 |
+
"content": "km_KH",
|
| 301 |
+
"lstrip": false,
|
| 302 |
+
"normalized": false,
|
| 303 |
+
"rstrip": false,
|
| 304 |
+
"single_word": false,
|
| 305 |
+
"special": true
|
| 306 |
+
},
|
| 307 |
+
"250035": {
|
| 308 |
+
"content": "mk_MK",
|
| 309 |
+
"lstrip": false,
|
| 310 |
+
"normalized": false,
|
| 311 |
+
"rstrip": false,
|
| 312 |
+
"single_word": false,
|
| 313 |
+
"special": true
|
| 314 |
+
},
|
| 315 |
+
"250036": {
|
| 316 |
+
"content": "ml_IN",
|
| 317 |
+
"lstrip": false,
|
| 318 |
+
"normalized": false,
|
| 319 |
+
"rstrip": false,
|
| 320 |
+
"single_word": false,
|
| 321 |
+
"special": true
|
| 322 |
+
},
|
| 323 |
+
"250037": {
|
| 324 |
+
"content": "mn_MN",
|
| 325 |
+
"lstrip": false,
|
| 326 |
+
"normalized": false,
|
| 327 |
+
"rstrip": false,
|
| 328 |
+
"single_word": false,
|
| 329 |
+
"special": true
|
| 330 |
+
},
|
| 331 |
+
"250038": {
|
| 332 |
+
"content": "mr_IN",
|
| 333 |
+
"lstrip": false,
|
| 334 |
+
"normalized": false,
|
| 335 |
+
"rstrip": false,
|
| 336 |
+
"single_word": false,
|
| 337 |
+
"special": true
|
| 338 |
+
},
|
| 339 |
+
"250039": {
|
| 340 |
+
"content": "pl_PL",
|
| 341 |
+
"lstrip": false,
|
| 342 |
+
"normalized": false,
|
| 343 |
+
"rstrip": false,
|
| 344 |
+
"single_word": false,
|
| 345 |
+
"special": true
|
| 346 |
+
},
|
| 347 |
+
"250040": {
|
| 348 |
+
"content": "ps_AF",
|
| 349 |
+
"lstrip": false,
|
| 350 |
+
"normalized": false,
|
| 351 |
+
"rstrip": false,
|
| 352 |
+
"single_word": false,
|
| 353 |
+
"special": true
|
| 354 |
+
},
|
| 355 |
+
"250041": {
|
| 356 |
+
"content": "pt_XX",
|
| 357 |
+
"lstrip": false,
|
| 358 |
+
"normalized": false,
|
| 359 |
+
"rstrip": false,
|
| 360 |
+
"single_word": false,
|
| 361 |
+
"special": true
|
| 362 |
+
},
|
| 363 |
+
"250042": {
|
| 364 |
+
"content": "sv_SE",
|
| 365 |
+
"lstrip": false,
|
| 366 |
+
"normalized": false,
|
| 367 |
+
"rstrip": false,
|
| 368 |
+
"single_word": false,
|
| 369 |
+
"special": true
|
| 370 |
+
},
|
| 371 |
+
"250043": {
|
| 372 |
+
"content": "sw_KE",
|
| 373 |
+
"lstrip": false,
|
| 374 |
+
"normalized": false,
|
| 375 |
+
"rstrip": false,
|
| 376 |
+
"single_word": false,
|
| 377 |
+
"special": true
|
| 378 |
+
},
|
| 379 |
+
"250044": {
|
| 380 |
+
"content": "ta_IN",
|
| 381 |
+
"lstrip": false,
|
| 382 |
+
"normalized": false,
|
| 383 |
+
"rstrip": false,
|
| 384 |
+
"single_word": false,
|
| 385 |
+
"special": true
|
| 386 |
+
},
|
| 387 |
+
"250045": {
|
| 388 |
+
"content": "te_IN",
|
| 389 |
+
"lstrip": false,
|
| 390 |
+
"normalized": false,
|
| 391 |
+
"rstrip": false,
|
| 392 |
+
"single_word": false,
|
| 393 |
+
"special": true
|
| 394 |
+
},
|
| 395 |
+
"250046": {
|
| 396 |
+
"content": "th_TH",
|
| 397 |
+
"lstrip": false,
|
| 398 |
+
"normalized": false,
|
| 399 |
+
"rstrip": false,
|
| 400 |
+
"single_word": false,
|
| 401 |
+
"special": true
|
| 402 |
+
},
|
| 403 |
+
"250047": {
|
| 404 |
+
"content": "tl_XX",
|
| 405 |
+
"lstrip": false,
|
| 406 |
+
"normalized": false,
|
| 407 |
+
"rstrip": false,
|
| 408 |
+
"single_word": false,
|
| 409 |
+
"special": true
|
| 410 |
+
},
|
| 411 |
+
"250048": {
|
| 412 |
+
"content": "uk_UA",
|
| 413 |
+
"lstrip": false,
|
| 414 |
+
"normalized": false,
|
| 415 |
+
"rstrip": false,
|
| 416 |
+
"single_word": false,
|
| 417 |
+
"special": true
|
| 418 |
+
},
|
| 419 |
+
"250049": {
|
| 420 |
+
"content": "ur_PK",
|
| 421 |
+
"lstrip": false,
|
| 422 |
+
"normalized": false,
|
| 423 |
+
"rstrip": false,
|
| 424 |
+
"single_word": false,
|
| 425 |
+
"special": true
|
| 426 |
+
},
|
| 427 |
+
"250050": {
|
| 428 |
+
"content": "xh_ZA",
|
| 429 |
+
"lstrip": false,
|
| 430 |
+
"normalized": false,
|
| 431 |
+
"rstrip": false,
|
| 432 |
+
"single_word": false,
|
| 433 |
+
"special": true
|
| 434 |
+
},
|
| 435 |
+
"250051": {
|
| 436 |
+
"content": "gl_ES",
|
| 437 |
+
"lstrip": false,
|
| 438 |
+
"normalized": false,
|
| 439 |
+
"rstrip": false,
|
| 440 |
+
"single_word": false,
|
| 441 |
+
"special": true
|
| 442 |
+
},
|
| 443 |
+
"250052": {
|
| 444 |
+
"content": "sl_SI",
|
| 445 |
+
"lstrip": false,
|
| 446 |
+
"normalized": false,
|
| 447 |
+
"rstrip": false,
|
| 448 |
+
"single_word": false,
|
| 449 |
+
"special": true
|
| 450 |
+
},
|
| 451 |
+
"250053": {
|
| 452 |
+
"content": "<mask>",
|
| 453 |
+
"lstrip": true,
|
| 454 |
+
"normalized": true,
|
| 455 |
+
"rstrip": false,
|
| 456 |
+
"single_word": false,
|
| 457 |
+
"special": true
|
| 458 |
+
}
|
| 459 |
+
},
|
| 460 |
+
"additional_special_tokens": [
|
| 461 |
+
"ar_AR",
|
| 462 |
+
"cs_CZ",
|
| 463 |
+
"de_DE",
|
| 464 |
+
"en_XX",
|
| 465 |
+
"es_XX",
|
| 466 |
+
"et_EE",
|
| 467 |
+
"fi_FI",
|
| 468 |
+
"fr_XX",
|
| 469 |
+
"gu_IN",
|
| 470 |
+
"hi_IN",
|
| 471 |
+
"it_IT",
|
| 472 |
+
"ja_XX",
|
| 473 |
+
"kk_KZ",
|
| 474 |
+
"ko_KR",
|
| 475 |
+
"lt_LT",
|
| 476 |
+
"lv_LV",
|
| 477 |
+
"my_MM",
|
| 478 |
+
"ne_NP",
|
| 479 |
+
"nl_XX",
|
| 480 |
+
"ro_RO",
|
| 481 |
+
"ru_RU",
|
| 482 |
+
"si_LK",
|
| 483 |
+
"tr_TR",
|
| 484 |
+
"vi_VN",
|
| 485 |
+
"zh_CN",
|
| 486 |
+
"af_ZA",
|
| 487 |
+
"az_AZ",
|
| 488 |
+
"bn_IN",
|
| 489 |
+
"fa_IR",
|
| 490 |
+
"he_IL",
|
| 491 |
+
"hr_HR",
|
| 492 |
+
"id_ID",
|
| 493 |
+
"ka_GE",
|
| 494 |
+
"km_KH",
|
| 495 |
+
"mk_MK",
|
| 496 |
+
"ml_IN",
|
| 497 |
+
"mn_MN",
|
| 498 |
+
"mr_IN",
|
| 499 |
+
"pl_PL",
|
| 500 |
+
"ps_AF",
|
| 501 |
+
"pt_XX",
|
| 502 |
+
"sv_SE",
|
| 503 |
+
"sw_KE",
|
| 504 |
+
"ta_IN",
|
| 505 |
+
"te_IN",
|
| 506 |
+
"th_TH",
|
| 507 |
+
"tl_XX",
|
| 508 |
+
"uk_UA",
|
| 509 |
+
"ur_PK",
|
| 510 |
+
"xh_ZA",
|
| 511 |
+
"gl_ES",
|
| 512 |
+
"sl_SI"
|
| 513 |
+
],
|
| 514 |
+
"bos_token": "<s>",
|
| 515 |
+
"clean_up_tokenization_spaces": true,
|
| 516 |
+
"cls_token": "<s>",
|
| 517 |
+
"eos_token": "</s>",
|
| 518 |
+
"language_codes": "ML50",
|
| 519 |
+
"mask_token": "<mask>",
|
| 520 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 521 |
+
"pad_token": "<pad>",
|
| 522 |
+
"sep_token": "</s>",
|
| 523 |
+
"sp_model_kwargs": {},
|
| 524 |
+
"src_lang": "ja_XX",
|
| 525 |
+
"tgt_lang": "ko_KR",
|
| 526 |
+
"tokenizer_class": "MBart50Tokenizer",
|
| 527 |
+
"unk_token": "<unk>"
|
| 528 |
+
}
|
checkpoint-215120/trainer_state.json
ADDED
|
@@ -0,0 +1,2651 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.6568061113357544,
|
| 3 |
+
"best_model_checkpoint": "./finetuned_model/checkpoint-129072",
|
| 4 |
+
"epoch": 5.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 215120,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.01,
|
| 13 |
+
"learning_rate": 9.96e-06,
|
| 14 |
+
"loss": 2.2631,
|
| 15 |
+
"step": 500
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.02,
|
| 19 |
+
"learning_rate": 1.992e-05,
|
| 20 |
+
"loss": 1.5976,
|
| 21 |
+
"step": 1000
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.03,
|
| 25 |
+
"learning_rate": 2.9920000000000005e-05,
|
| 26 |
+
"loss": 1.4641,
|
| 27 |
+
"step": 1500
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.05,
|
| 31 |
+
"learning_rate": 3.9920000000000004e-05,
|
| 32 |
+
"loss": 1.3868,
|
| 33 |
+
"step": 2000
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.06,
|
| 37 |
+
"learning_rate": 4.992e-05,
|
| 38 |
+
"loss": 1.344,
|
| 39 |
+
"step": 2500
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.07,
|
| 43 |
+
"learning_rate": 4.996142239367825e-05,
|
| 44 |
+
"loss": 1.3138,
|
| 45 |
+
"step": 3000
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.08,
|
| 49 |
+
"learning_rate": 4.9922533677628105e-05,
|
| 50 |
+
"loss": 1.244,
|
| 51 |
+
"step": 3500
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.09,
|
| 55 |
+
"learning_rate": 4.988364496157795e-05,
|
| 56 |
+
"loss": 1.2047,
|
| 57 |
+
"step": 4000
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.1,
|
| 61 |
+
"learning_rate": 4.98447562455278e-05,
|
| 62 |
+
"loss": 1.1707,
|
| 63 |
+
"step": 4500
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.12,
|
| 67 |
+
"learning_rate": 4.980586752947765e-05,
|
| 68 |
+
"loss": 1.1331,
|
| 69 |
+
"step": 5000
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.13,
|
| 73 |
+
"learning_rate": 4.97670565908596e-05,
|
| 74 |
+
"loss": 1.1021,
|
| 75 |
+
"step": 5500
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.14,
|
| 79 |
+
"learning_rate": 4.972816787480945e-05,
|
| 80 |
+
"loss": 1.0923,
|
| 81 |
+
"step": 6000
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.15,
|
| 85 |
+
"learning_rate": 4.968927915875929e-05,
|
| 86 |
+
"loss": 1.0713,
|
| 87 |
+
"step": 6500
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.16,
|
| 91 |
+
"learning_rate": 4.965039044270914e-05,
|
| 92 |
+
"loss": 1.0609,
|
| 93 |
+
"step": 7000
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.17,
|
| 97 |
+
"learning_rate": 4.9611501726658995e-05,
|
| 98 |
+
"loss": 1.043,
|
| 99 |
+
"step": 7500
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.19,
|
| 103 |
+
"learning_rate": 4.957261301060884e-05,
|
| 104 |
+
"loss": 1.0345,
|
| 105 |
+
"step": 8000
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.2,
|
| 109 |
+
"learning_rate": 4.953380207199079e-05,
|
| 110 |
+
"loss": 1.02,
|
| 111 |
+
"step": 8500
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.21,
|
| 115 |
+
"learning_rate": 4.949491335594064e-05,
|
| 116 |
+
"loss": 1.0057,
|
| 117 |
+
"step": 9000
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.22,
|
| 121 |
+
"learning_rate": 4.945602463989049e-05,
|
| 122 |
+
"loss": 0.9978,
|
| 123 |
+
"step": 9500
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.23,
|
| 127 |
+
"learning_rate": 4.9417135923840344e-05,
|
| 128 |
+
"loss": 0.9801,
|
| 129 |
+
"step": 10000
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.24,
|
| 133 |
+
"learning_rate": 4.937824720779019e-05,
|
| 134 |
+
"loss": 0.9752,
|
| 135 |
+
"step": 10500
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.26,
|
| 139 |
+
"learning_rate": 4.933935849174004e-05,
|
| 140 |
+
"loss": 0.9736,
|
| 141 |
+
"step": 11000
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.27,
|
| 145 |
+
"learning_rate": 4.930046977568989e-05,
|
| 146 |
+
"loss": 0.9519,
|
| 147 |
+
"step": 11500
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.28,
|
| 151 |
+
"learning_rate": 4.9261581059639736e-05,
|
| 152 |
+
"loss": 0.9559,
|
| 153 |
+
"step": 12000
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.29,
|
| 157 |
+
"learning_rate": 4.9222770121021686e-05,
|
| 158 |
+
"loss": 0.936,
|
| 159 |
+
"step": 12500
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.3,
|
| 163 |
+
"learning_rate": 4.9183959182403636e-05,
|
| 164 |
+
"loss": 0.9276,
|
| 165 |
+
"step": 13000
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.31,
|
| 169 |
+
"learning_rate": 4.914507046635349e-05,
|
| 170 |
+
"loss": 0.9384,
|
| 171 |
+
"step": 13500
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.33,
|
| 175 |
+
"learning_rate": 4.910618175030333e-05,
|
| 176 |
+
"loss": 0.9347,
|
| 177 |
+
"step": 14000
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.34,
|
| 181 |
+
"learning_rate": 4.9067293034253183e-05,
|
| 182 |
+
"loss": 0.9182,
|
| 183 |
+
"step": 14500
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.35,
|
| 187 |
+
"learning_rate": 4.902840431820303e-05,
|
| 188 |
+
"loss": 0.9136,
|
| 189 |
+
"step": 15000
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.36,
|
| 193 |
+
"learning_rate": 4.898951560215288e-05,
|
| 194 |
+
"loss": 0.9096,
|
| 195 |
+
"step": 15500
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.37,
|
| 199 |
+
"learning_rate": 4.895062688610273e-05,
|
| 200 |
+
"loss": 0.8926,
|
| 201 |
+
"step": 16000
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.38,
|
| 205 |
+
"learning_rate": 4.891181594748468e-05,
|
| 206 |
+
"loss": 0.8946,
|
| 207 |
+
"step": 16500
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.4,
|
| 211 |
+
"learning_rate": 4.887300500886663e-05,
|
| 212 |
+
"loss": 0.8895,
|
| 213 |
+
"step": 17000
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.41,
|
| 217 |
+
"learning_rate": 4.8834116292816475e-05,
|
| 218 |
+
"loss": 0.8897,
|
| 219 |
+
"step": 17500
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.42,
|
| 223 |
+
"learning_rate": 4.879522757676633e-05,
|
| 224 |
+
"loss": 0.8901,
|
| 225 |
+
"step": 18000
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.43,
|
| 229 |
+
"learning_rate": 4.875633886071618e-05,
|
| 230 |
+
"loss": 0.8839,
|
| 231 |
+
"step": 18500
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 0.44,
|
| 235 |
+
"learning_rate": 4.871745014466603e-05,
|
| 236 |
+
"loss": 0.8737,
|
| 237 |
+
"step": 19000
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.45,
|
| 241 |
+
"learning_rate": 4.867863920604798e-05,
|
| 242 |
+
"loss": 0.8613,
|
| 243 |
+
"step": 19500
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.46,
|
| 247 |
+
"learning_rate": 4.8639750489997824e-05,
|
| 248 |
+
"loss": 0.8547,
|
| 249 |
+
"step": 20000
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.48,
|
| 253 |
+
"learning_rate": 4.8600861773947676e-05,
|
| 254 |
+
"loss": 0.8623,
|
| 255 |
+
"step": 20500
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.49,
|
| 259 |
+
"learning_rate": 4.856197305789753e-05,
|
| 260 |
+
"loss": 0.8615,
|
| 261 |
+
"step": 21000
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.5,
|
| 265 |
+
"learning_rate": 4.852308434184737e-05,
|
| 266 |
+
"loss": 0.8614,
|
| 267 |
+
"step": 21500
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.51,
|
| 271 |
+
"learning_rate": 4.848419562579722e-05,
|
| 272 |
+
"loss": 0.846,
|
| 273 |
+
"step": 22000
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 0.52,
|
| 277 |
+
"learning_rate": 4.8445384687179167e-05,
|
| 278 |
+
"loss": 0.8487,
|
| 279 |
+
"step": 22500
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.53,
|
| 283 |
+
"learning_rate": 4.840649597112902e-05,
|
| 284 |
+
"loss": 0.8426,
|
| 285 |
+
"step": 23000
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.55,
|
| 289 |
+
"learning_rate": 4.836760725507887e-05,
|
| 290 |
+
"loss": 0.8398,
|
| 291 |
+
"step": 23500
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.56,
|
| 295 |
+
"learning_rate": 4.8328718539028714e-05,
|
| 296 |
+
"loss": 0.8407,
|
| 297 |
+
"step": 24000
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.57,
|
| 301 |
+
"learning_rate": 4.828998537784277e-05,
|
| 302 |
+
"loss": 0.8438,
|
| 303 |
+
"step": 24500
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.58,
|
| 307 |
+
"learning_rate": 4.825117443922472e-05,
|
| 308 |
+
"loss": 0.8257,
|
| 309 |
+
"step": 25000
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.59,
|
| 313 |
+
"learning_rate": 4.821228572317456e-05,
|
| 314 |
+
"loss": 0.8328,
|
| 315 |
+
"step": 25500
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 0.6,
|
| 319 |
+
"learning_rate": 4.8173397007124415e-05,
|
| 320 |
+
"loss": 0.8229,
|
| 321 |
+
"step": 26000
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 0.62,
|
| 325 |
+
"learning_rate": 4.8134508291074266e-05,
|
| 326 |
+
"loss": 0.8249,
|
| 327 |
+
"step": 26500
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 0.63,
|
| 331 |
+
"learning_rate": 4.809561957502412e-05,
|
| 332 |
+
"loss": 0.8232,
|
| 333 |
+
"step": 27000
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 0.64,
|
| 337 |
+
"learning_rate": 4.805673085897396e-05,
|
| 338 |
+
"loss": 0.83,
|
| 339 |
+
"step": 27500
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.65,
|
| 343 |
+
"learning_rate": 4.801791992035591e-05,
|
| 344 |
+
"loss": 0.8339,
|
| 345 |
+
"step": 28000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.66,
|
| 349 |
+
"learning_rate": 4.7979031204305764e-05,
|
| 350 |
+
"loss": 0.817,
|
| 351 |
+
"step": 28500
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 0.67,
|
| 355 |
+
"learning_rate": 4.7940142488255615e-05,
|
| 356 |
+
"loss": 0.8223,
|
| 357 |
+
"step": 29000
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 0.69,
|
| 361 |
+
"learning_rate": 4.790125377220545e-05,
|
| 362 |
+
"loss": 0.8041,
|
| 363 |
+
"step": 29500
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 0.7,
|
| 367 |
+
"learning_rate": 4.7862365056155305e-05,
|
| 368 |
+
"loss": 0.8126,
|
| 369 |
+
"step": 30000
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.71,
|
| 373 |
+
"learning_rate": 4.7823476340105156e-05,
|
| 374 |
+
"loss": 0.8039,
|
| 375 |
+
"step": 30500
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 0.72,
|
| 379 |
+
"learning_rate": 4.778458762405501e-05,
|
| 380 |
+
"loss": 0.8071,
|
| 381 |
+
"step": 31000
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.73,
|
| 385 |
+
"learning_rate": 4.774569890800485e-05,
|
| 386 |
+
"loss": 0.8065,
|
| 387 |
+
"step": 31500
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.74,
|
| 391 |
+
"learning_rate": 4.7706810191954704e-05,
|
| 392 |
+
"loss": 0.7998,
|
| 393 |
+
"step": 32000
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 0.76,
|
| 397 |
+
"learning_rate": 4.7667921475904556e-05,
|
| 398 |
+
"loss": 0.8034,
|
| 399 |
+
"step": 32500
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 0.77,
|
| 403 |
+
"learning_rate": 4.76290327598544e-05,
|
| 404 |
+
"loss": 0.8031,
|
| 405 |
+
"step": 33000
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 0.78,
|
| 409 |
+
"learning_rate": 4.759014404380425e-05,
|
| 410 |
+
"loss": 0.8057,
|
| 411 |
+
"step": 33500
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 0.79,
|
| 415 |
+
"learning_rate": 4.7551255327754103e-05,
|
| 416 |
+
"loss": 0.7905,
|
| 417 |
+
"step": 34000
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 0.8,
|
| 421 |
+
"learning_rate": 4.751236661170395e-05,
|
| 422 |
+
"loss": 0.794,
|
| 423 |
+
"step": 34500
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.81,
|
| 427 |
+
"learning_rate": 4.74734778956538e-05,
|
| 428 |
+
"loss": 0.7906,
|
| 429 |
+
"step": 35000
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.83,
|
| 433 |
+
"learning_rate": 4.743458917960365e-05,
|
| 434 |
+
"loss": 0.7865,
|
| 435 |
+
"step": 35500
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 0.84,
|
| 439 |
+
"learning_rate": 4.73957004635535e-05,
|
| 440 |
+
"loss": 0.7644,
|
| 441 |
+
"step": 36000
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 0.85,
|
| 445 |
+
"learning_rate": 4.735681174750335e-05,
|
| 446 |
+
"loss": 0.7737,
|
| 447 |
+
"step": 36500
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 0.86,
|
| 451 |
+
"learning_rate": 4.73179230314532e-05,
|
| 452 |
+
"loss": 0.7746,
|
| 453 |
+
"step": 37000
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 0.87,
|
| 457 |
+
"learning_rate": 4.7279034315403044e-05,
|
| 458 |
+
"loss": 0.7832,
|
| 459 |
+
"step": 37500
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.88,
|
| 463 |
+
"learning_rate": 4.724014559935289e-05,
|
| 464 |
+
"loss": 0.7706,
|
| 465 |
+
"step": 38000
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.89,
|
| 469 |
+
"learning_rate": 4.720141243816694e-05,
|
| 470 |
+
"loss": 0.7778,
|
| 471 |
+
"step": 38500
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.91,
|
| 475 |
+
"learning_rate": 4.7162523722116795e-05,
|
| 476 |
+
"loss": 0.7713,
|
| 477 |
+
"step": 39000
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 0.92,
|
| 481 |
+
"learning_rate": 4.712363500606664e-05,
|
| 482 |
+
"loss": 0.772,
|
| 483 |
+
"step": 39500
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 0.93,
|
| 487 |
+
"learning_rate": 4.708474629001649e-05,
|
| 488 |
+
"loss": 0.775,
|
| 489 |
+
"step": 40000
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 0.94,
|
| 493 |
+
"learning_rate": 4.704593535139844e-05,
|
| 494 |
+
"loss": 0.7718,
|
| 495 |
+
"step": 40500
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 0.95,
|
| 499 |
+
"learning_rate": 4.700704663534829e-05,
|
| 500 |
+
"loss": 0.7724,
|
| 501 |
+
"step": 41000
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 0.96,
|
| 505 |
+
"learning_rate": 4.696815791929814e-05,
|
| 506 |
+
"loss": 0.7629,
|
| 507 |
+
"step": 41500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.98,
|
| 511 |
+
"learning_rate": 4.692926920324799e-05,
|
| 512 |
+
"loss": 0.7731,
|
| 513 |
+
"step": 42000
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.99,
|
| 517 |
+
"learning_rate": 4.689038048719784e-05,
|
| 518 |
+
"loss": 0.7597,
|
| 519 |
+
"step": 42500
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 1.0,
|
| 523 |
+
"learning_rate": 4.6851491771147685e-05,
|
| 524 |
+
"loss": 0.7629,
|
| 525 |
+
"step": 43000
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 1.0,
|
| 529 |
+
"eval_bleu": 54.9608,
|
| 530 |
+
"eval_gen_len": 16.1273,
|
| 531 |
+
"eval_loss": 0.7230631709098816,
|
| 532 |
+
"eval_runtime": 7044.4442,
|
| 533 |
+
"eval_samples_per_second": 12.196,
|
| 534 |
+
"eval_steps_per_second": 1.525,
|
| 535 |
+
"step": 43024
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 1.01,
|
| 539 |
+
"learning_rate": 4.6812680832529634e-05,
|
| 540 |
+
"loss": 0.6436,
|
| 541 |
+
"step": 43500
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 1.02,
|
| 545 |
+
"learning_rate": 4.6773869893911584e-05,
|
| 546 |
+
"loss": 0.6446,
|
| 547 |
+
"step": 44000
|
| 548 |
+
},
|
| 549 |
+
{
|
| 550 |
+
"epoch": 1.03,
|
| 551 |
+
"learning_rate": 4.6734981177861435e-05,
|
| 552 |
+
"loss": 0.6489,
|
| 553 |
+
"step": 44500
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 1.05,
|
| 557 |
+
"learning_rate": 4.669617023924338e-05,
|
| 558 |
+
"loss": 0.6512,
|
| 559 |
+
"step": 45000
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 1.06,
|
| 563 |
+
"learning_rate": 4.665728152319323e-05,
|
| 564 |
+
"loss": 0.648,
|
| 565 |
+
"step": 45500
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"epoch": 1.07,
|
| 569 |
+
"learning_rate": 4.661839280714308e-05,
|
| 570 |
+
"loss": 0.6426,
|
| 571 |
+
"step": 46000
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 1.08,
|
| 575 |
+
"learning_rate": 4.657958186852503e-05,
|
| 576 |
+
"loss": 0.6489,
|
| 577 |
+
"step": 46500
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 1.09,
|
| 581 |
+
"learning_rate": 4.6540693152474876e-05,
|
| 582 |
+
"loss": 0.6417,
|
| 583 |
+
"step": 47000
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 1.1,
|
| 587 |
+
"learning_rate": 4.650180443642473e-05,
|
| 588 |
+
"loss": 0.645,
|
| 589 |
+
"step": 47500
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 1.12,
|
| 593 |
+
"learning_rate": 4.646291572037458e-05,
|
| 594 |
+
"loss": 0.6543,
|
| 595 |
+
"step": 48000
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"epoch": 1.13,
|
| 599 |
+
"learning_rate": 4.642402700432443e-05,
|
| 600 |
+
"loss": 0.6555,
|
| 601 |
+
"step": 48500
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 1.14,
|
| 605 |
+
"learning_rate": 4.6385138288274275e-05,
|
| 606 |
+
"loss": 0.6501,
|
| 607 |
+
"step": 49000
|
| 608 |
+
},
|
| 609 |
+
{
|
| 610 |
+
"epoch": 1.15,
|
| 611 |
+
"learning_rate": 4.634624957222413e-05,
|
| 612 |
+
"loss": 0.6447,
|
| 613 |
+
"step": 49500
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 1.16,
|
| 617 |
+
"learning_rate": 4.630736085617398e-05,
|
| 618 |
+
"loss": 0.6432,
|
| 619 |
+
"step": 50000
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 1.17,
|
| 623 |
+
"learning_rate": 4.626847214012382e-05,
|
| 624 |
+
"loss": 0.6516,
|
| 625 |
+
"step": 50500
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 1.19,
|
| 629 |
+
"learning_rate": 4.6229583424073674e-05,
|
| 630 |
+
"loss": 0.6517,
|
| 631 |
+
"step": 51000
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 1.2,
|
| 635 |
+
"learning_rate": 4.6190694708023526e-05,
|
| 636 |
+
"loss": 0.6534,
|
| 637 |
+
"step": 51500
|
| 638 |
+
},
|
| 639 |
+
{
|
| 640 |
+
"epoch": 1.21,
|
| 641 |
+
"learning_rate": 4.615180599197337e-05,
|
| 642 |
+
"loss": 0.6482,
|
| 643 |
+
"step": 52000
|
| 644 |
+
},
|
| 645 |
+
{
|
| 646 |
+
"epoch": 1.22,
|
| 647 |
+
"learning_rate": 4.611291727592322e-05,
|
| 648 |
+
"loss": 0.6504,
|
| 649 |
+
"step": 52500
|
| 650 |
+
},
|
| 651 |
+
{
|
| 652 |
+
"epoch": 1.23,
|
| 653 |
+
"learning_rate": 4.607402855987307e-05,
|
| 654 |
+
"loss": 0.6481,
|
| 655 |
+
"step": 53000
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 1.24,
|
| 659 |
+
"learning_rate": 4.603513984382292e-05,
|
| 660 |
+
"loss": 0.6444,
|
| 661 |
+
"step": 53500
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 1.26,
|
| 665 |
+
"learning_rate": 4.599625112777276e-05,
|
| 666 |
+
"loss": 0.6509,
|
| 667 |
+
"step": 54000
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 1.27,
|
| 671 |
+
"learning_rate": 4.5957362411722615e-05,
|
| 672 |
+
"loss": 0.6485,
|
| 673 |
+
"step": 54500
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 1.28,
|
| 677 |
+
"learning_rate": 4.5918551473104564e-05,
|
| 678 |
+
"loss": 0.6474,
|
| 679 |
+
"step": 55000
|
| 680 |
+
},
|
| 681 |
+
{
|
| 682 |
+
"epoch": 1.29,
|
| 683 |
+
"learning_rate": 4.5879662757054416e-05,
|
| 684 |
+
"loss": 0.6467,
|
| 685 |
+
"step": 55500
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 1.3,
|
| 689 |
+
"learning_rate": 4.584077404100427e-05,
|
| 690 |
+
"loss": 0.6442,
|
| 691 |
+
"step": 56000
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 1.31,
|
| 695 |
+
"learning_rate": 4.580188532495411e-05,
|
| 696 |
+
"loss": 0.6537,
|
| 697 |
+
"step": 56500
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"epoch": 1.32,
|
| 701 |
+
"learning_rate": 4.576307438633606e-05,
|
| 702 |
+
"loss": 0.6563,
|
| 703 |
+
"step": 57000
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 1.34,
|
| 707 |
+
"learning_rate": 4.5724185670285913e-05,
|
| 708 |
+
"loss": 0.6587,
|
| 709 |
+
"step": 57500
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 1.35,
|
| 713 |
+
"learning_rate": 4.5685296954235765e-05,
|
| 714 |
+
"loss": 0.6535,
|
| 715 |
+
"step": 58000
|
| 716 |
+
},
|
| 717 |
+
{
|
| 718 |
+
"epoch": 1.36,
|
| 719 |
+
"learning_rate": 4.564640823818561e-05,
|
| 720 |
+
"loss": 0.6623,
|
| 721 |
+
"step": 58500
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 1.37,
|
| 725 |
+
"learning_rate": 4.560759729956756e-05,
|
| 726 |
+
"loss": 0.6458,
|
| 727 |
+
"step": 59000
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"epoch": 1.38,
|
| 731 |
+
"learning_rate": 4.556870858351741e-05,
|
| 732 |
+
"loss": 0.6557,
|
| 733 |
+
"step": 59500
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 1.39,
|
| 737 |
+
"learning_rate": 4.552981986746726e-05,
|
| 738 |
+
"loss": 0.6458,
|
| 739 |
+
"step": 60000
|
| 740 |
+
},
|
| 741 |
+
{
|
| 742 |
+
"epoch": 1.41,
|
| 743 |
+
"learning_rate": 4.549093115141711e-05,
|
| 744 |
+
"loss": 0.6565,
|
| 745 |
+
"step": 60500
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 1.42,
|
| 749 |
+
"learning_rate": 4.545204243536695e-05,
|
| 750 |
+
"loss": 0.6521,
|
| 751 |
+
"step": 61000
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 1.43,
|
| 755 |
+
"learning_rate": 4.54131537193168e-05,
|
| 756 |
+
"loss": 0.6485,
|
| 757 |
+
"step": 61500
|
| 758 |
+
},
|
| 759 |
+
{
|
| 760 |
+
"epoch": 1.44,
|
| 761 |
+
"learning_rate": 4.5374265003266655e-05,
|
| 762 |
+
"loss": 0.6503,
|
| 763 |
+
"step": 62000
|
| 764 |
+
},
|
| 765 |
+
{
|
| 766 |
+
"epoch": 1.45,
|
| 767 |
+
"learning_rate": 4.53353762872165e-05,
|
| 768 |
+
"loss": 0.6483,
|
| 769 |
+
"step": 62500
|
| 770 |
+
},
|
| 771 |
+
{
|
| 772 |
+
"epoch": 1.46,
|
| 773 |
+
"learning_rate": 4.529656534859845e-05,
|
| 774 |
+
"loss": 0.6425,
|
| 775 |
+
"step": 63000
|
| 776 |
+
},
|
| 777 |
+
{
|
| 778 |
+
"epoch": 1.48,
|
| 779 |
+
"learning_rate": 4.52576766325483e-05,
|
| 780 |
+
"loss": 0.6566,
|
| 781 |
+
"step": 63500
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 1.49,
|
| 785 |
+
"learning_rate": 4.521878791649815e-05,
|
| 786 |
+
"loss": 0.6537,
|
| 787 |
+
"step": 64000
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 1.5,
|
| 791 |
+
"learning_rate": 4.5179899200448e-05,
|
| 792 |
+
"loss": 0.653,
|
| 793 |
+
"step": 64500
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 1.51,
|
| 797 |
+
"learning_rate": 4.514116603926205e-05,
|
| 798 |
+
"loss": 0.6448,
|
| 799 |
+
"step": 65000
|
| 800 |
+
},
|
| 801 |
+
{
|
| 802 |
+
"epoch": 1.52,
|
| 803 |
+
"learning_rate": 4.51022773232119e-05,
|
| 804 |
+
"loss": 0.6594,
|
| 805 |
+
"step": 65500
|
| 806 |
+
},
|
| 807 |
+
{
|
| 808 |
+
"epoch": 1.53,
|
| 809 |
+
"learning_rate": 4.506338860716175e-05,
|
| 810 |
+
"loss": 0.6551,
|
| 811 |
+
"step": 66000
|
| 812 |
+
},
|
| 813 |
+
{
|
| 814 |
+
"epoch": 1.55,
|
| 815 |
+
"learning_rate": 4.50245776685437e-05,
|
| 816 |
+
"loss": 0.6559,
|
| 817 |
+
"step": 66500
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"epoch": 1.56,
|
| 821 |
+
"learning_rate": 4.498568895249355e-05,
|
| 822 |
+
"loss": 0.6506,
|
| 823 |
+
"step": 67000
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 1.57,
|
| 827 |
+
"learning_rate": 4.49468002364434e-05,
|
| 828 |
+
"loss": 0.6504,
|
| 829 |
+
"step": 67500
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"epoch": 1.58,
|
| 833 |
+
"learning_rate": 4.4907911520393245e-05,
|
| 834 |
+
"loss": 0.6643,
|
| 835 |
+
"step": 68000
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 1.59,
|
| 839 |
+
"learning_rate": 4.486902280434309e-05,
|
| 840 |
+
"loss": 0.6435,
|
| 841 |
+
"step": 68500
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"epoch": 1.6,
|
| 845 |
+
"learning_rate": 4.483013408829294e-05,
|
| 846 |
+
"loss": 0.6389,
|
| 847 |
+
"step": 69000
|
| 848 |
+
},
|
| 849 |
+
{
|
| 850 |
+
"epoch": 1.62,
|
| 851 |
+
"learning_rate": 4.479124537224279e-05,
|
| 852 |
+
"loss": 0.6495,
|
| 853 |
+
"step": 69500
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 1.63,
|
| 857 |
+
"learning_rate": 4.475235665619264e-05,
|
| 858 |
+
"loss": 0.6395,
|
| 859 |
+
"step": 70000
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 1.64,
|
| 863 |
+
"learning_rate": 4.471346794014249e-05,
|
| 864 |
+
"loss": 0.6404,
|
| 865 |
+
"step": 70500
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 1.65,
|
| 869 |
+
"learning_rate": 4.467457922409234e-05,
|
| 870 |
+
"loss": 0.6438,
|
| 871 |
+
"step": 71000
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 1.66,
|
| 875 |
+
"learning_rate": 4.4635690508042186e-05,
|
| 876 |
+
"loss": 0.6361,
|
| 877 |
+
"step": 71500
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 1.67,
|
| 881 |
+
"learning_rate": 4.459680179199204e-05,
|
| 882 |
+
"loss": 0.6486,
|
| 883 |
+
"step": 72000
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 1.69,
|
| 887 |
+
"learning_rate": 4.455791307594189e-05,
|
| 888 |
+
"loss": 0.6489,
|
| 889 |
+
"step": 72500
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"epoch": 1.7,
|
| 893 |
+
"learning_rate": 4.4519024359891734e-05,
|
| 894 |
+
"loss": 0.6532,
|
| 895 |
+
"step": 73000
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"epoch": 1.71,
|
| 899 |
+
"learning_rate": 4.4480135643841585e-05,
|
| 900 |
+
"loss": 0.6529,
|
| 901 |
+
"step": 73500
|
| 902 |
+
},
|
| 903 |
+
{
|
| 904 |
+
"epoch": 1.72,
|
| 905 |
+
"learning_rate": 4.4441246927791437e-05,
|
| 906 |
+
"loss": 0.6451,
|
| 907 |
+
"step": 74000
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 1.73,
|
| 911 |
+
"learning_rate": 4.440235821174129e-05,
|
| 912 |
+
"loss": 0.6443,
|
| 913 |
+
"step": 74500
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 1.74,
|
| 917 |
+
"learning_rate": 4.436346949569113e-05,
|
| 918 |
+
"loss": 0.6456,
|
| 919 |
+
"step": 75000
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 1.75,
|
| 923 |
+
"learning_rate": 4.4324580779640984e-05,
|
| 924 |
+
"loss": 0.6515,
|
| 925 |
+
"step": 75500
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 1.77,
|
| 929 |
+
"learning_rate": 4.428569206359083e-05,
|
| 930 |
+
"loss": 0.6425,
|
| 931 |
+
"step": 76000
|
| 932 |
+
},
|
| 933 |
+
{
|
| 934 |
+
"epoch": 1.78,
|
| 935 |
+
"learning_rate": 4.424680334754068e-05,
|
| 936 |
+
"loss": 0.6394,
|
| 937 |
+
"step": 76500
|
| 938 |
+
},
|
| 939 |
+
{
|
| 940 |
+
"epoch": 1.79,
|
| 941 |
+
"learning_rate": 4.4207914631490525e-05,
|
| 942 |
+
"loss": 0.6433,
|
| 943 |
+
"step": 77000
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 1.8,
|
| 947 |
+
"learning_rate": 4.416902591544038e-05,
|
| 948 |
+
"loss": 0.6459,
|
| 949 |
+
"step": 77500
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 1.81,
|
| 953 |
+
"learning_rate": 4.4130214976822327e-05,
|
| 954 |
+
"loss": 0.6421,
|
| 955 |
+
"step": 78000
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 1.82,
|
| 959 |
+
"learning_rate": 4.409132626077218e-05,
|
| 960 |
+
"loss": 0.6428,
|
| 961 |
+
"step": 78500
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 1.84,
|
| 965 |
+
"learning_rate": 4.405251532215413e-05,
|
| 966 |
+
"loss": 0.6406,
|
| 967 |
+
"step": 79000
|
| 968 |
+
},
|
| 969 |
+
{
|
| 970 |
+
"epoch": 1.85,
|
| 971 |
+
"learning_rate": 4.401362660610397e-05,
|
| 972 |
+
"loss": 0.6357,
|
| 973 |
+
"step": 79500
|
| 974 |
+
},
|
| 975 |
+
{
|
| 976 |
+
"epoch": 1.86,
|
| 977 |
+
"learning_rate": 4.3974737890053824e-05,
|
| 978 |
+
"loss": 0.6355,
|
| 979 |
+
"step": 80000
|
| 980 |
+
},
|
| 981 |
+
{
|
| 982 |
+
"epoch": 1.87,
|
| 983 |
+
"learning_rate": 4.3935849174003676e-05,
|
| 984 |
+
"loss": 0.6459,
|
| 985 |
+
"step": 80500
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 1.88,
|
| 989 |
+
"learning_rate": 4.389696045795352e-05,
|
| 990 |
+
"loss": 0.6422,
|
| 991 |
+
"step": 81000
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 1.89,
|
| 995 |
+
"learning_rate": 4.385807174190337e-05,
|
| 996 |
+
"loss": 0.6413,
|
| 997 |
+
"step": 81500
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 1.91,
|
| 1001 |
+
"learning_rate": 4.381918302585322e-05,
|
| 1002 |
+
"loss": 0.6391,
|
| 1003 |
+
"step": 82000
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 1.92,
|
| 1007 |
+
"learning_rate": 4.3780294309803075e-05,
|
| 1008 |
+
"loss": 0.6409,
|
| 1009 |
+
"step": 82500
|
| 1010 |
+
},
|
| 1011 |
+
{
|
| 1012 |
+
"epoch": 1.93,
|
| 1013 |
+
"learning_rate": 4.374140559375292e-05,
|
| 1014 |
+
"loss": 0.632,
|
| 1015 |
+
"step": 83000
|
| 1016 |
+
},
|
| 1017 |
+
{
|
| 1018 |
+
"epoch": 1.94,
|
| 1019 |
+
"learning_rate": 4.370251687770277e-05,
|
| 1020 |
+
"loss": 0.6462,
|
| 1021 |
+
"step": 83500
|
| 1022 |
+
},
|
| 1023 |
+
{
|
| 1024 |
+
"epoch": 1.95,
|
| 1025 |
+
"learning_rate": 4.3663628161652616e-05,
|
| 1026 |
+
"loss": 0.6388,
|
| 1027 |
+
"step": 84000
|
| 1028 |
+
},
|
| 1029 |
+
{
|
| 1030 |
+
"epoch": 1.96,
|
| 1031 |
+
"learning_rate": 4.3624817223034566e-05,
|
| 1032 |
+
"loss": 0.642,
|
| 1033 |
+
"step": 84500
|
| 1034 |
+
},
|
| 1035 |
+
{
|
| 1036 |
+
"epoch": 1.98,
|
| 1037 |
+
"learning_rate": 4.358592850698442e-05,
|
| 1038 |
+
"loss": 0.6409,
|
| 1039 |
+
"step": 85000
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 1.99,
|
| 1043 |
+
"learning_rate": 4.354703979093426e-05,
|
| 1044 |
+
"loss": 0.6406,
|
| 1045 |
+
"step": 85500
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 2.0,
|
| 1049 |
+
"learning_rate": 4.350815107488411e-05,
|
| 1050 |
+
"loss": 0.6355,
|
| 1051 |
+
"step": 86000
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 2.0,
|
| 1055 |
+
"eval_bleu": 57.5735,
|
| 1056 |
+
"eval_gen_len": 15.9938,
|
| 1057 |
+
"eval_loss": 0.6641375422477722,
|
| 1058 |
+
"eval_runtime": 6884.7279,
|
| 1059 |
+
"eval_samples_per_second": 12.479,
|
| 1060 |
+
"eval_steps_per_second": 1.56,
|
| 1061 |
+
"step": 86048
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"epoch": 2.01,
|
| 1065 |
+
"learning_rate": 4.3469262358833965e-05,
|
| 1066 |
+
"loss": 0.5199,
|
| 1067 |
+
"step": 86500
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 2.02,
|
| 1071 |
+
"learning_rate": 4.343037364278381e-05,
|
| 1072 |
+
"loss": 0.4964,
|
| 1073 |
+
"step": 87000
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 2.03,
|
| 1077 |
+
"learning_rate": 4.339148492673366e-05,
|
| 1078 |
+
"loss": 0.5025,
|
| 1079 |
+
"step": 87500
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 2.05,
|
| 1083 |
+
"learning_rate": 4.335259621068351e-05,
|
| 1084 |
+
"loss": 0.508,
|
| 1085 |
+
"step": 88000
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 2.06,
|
| 1089 |
+
"learning_rate": 4.331378527206546e-05,
|
| 1090 |
+
"loss": 0.5096,
|
| 1091 |
+
"step": 88500
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 2.07,
|
| 1095 |
+
"learning_rate": 4.327497433344741e-05,
|
| 1096 |
+
"loss": 0.5106,
|
| 1097 |
+
"step": 89000
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"epoch": 2.08,
|
| 1101 |
+
"learning_rate": 4.323616339482936e-05,
|
| 1102 |
+
"loss": 0.5108,
|
| 1103 |
+
"step": 89500
|
| 1104 |
+
},
|
| 1105 |
+
{
|
| 1106 |
+
"epoch": 2.09,
|
| 1107 |
+
"learning_rate": 4.3197274678779206e-05,
|
| 1108 |
+
"loss": 0.5057,
|
| 1109 |
+
"step": 90000
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 2.1,
|
| 1113 |
+
"learning_rate": 4.3158463740161156e-05,
|
| 1114 |
+
"loss": 0.5108,
|
| 1115 |
+
"step": 90500
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 2.12,
|
| 1119 |
+
"learning_rate": 4.311957502411101e-05,
|
| 1120 |
+
"loss": 0.5044,
|
| 1121 |
+
"step": 91000
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 2.13,
|
| 1125 |
+
"learning_rate": 4.308068630806085e-05,
|
| 1126 |
+
"loss": 0.5077,
|
| 1127 |
+
"step": 91500
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 2.14,
|
| 1131 |
+
"learning_rate": 4.3041797592010704e-05,
|
| 1132 |
+
"loss": 0.5007,
|
| 1133 |
+
"step": 92000
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 2.15,
|
| 1137 |
+
"learning_rate": 4.300290887596055e-05,
|
| 1138 |
+
"loss": 0.5068,
|
| 1139 |
+
"step": 92500
|
| 1140 |
+
},
|
| 1141 |
+
{
|
| 1142 |
+
"epoch": 2.16,
|
| 1143 |
+
"learning_rate": 4.29640201599104e-05,
|
| 1144 |
+
"loss": 0.5038,
|
| 1145 |
+
"step": 93000
|
| 1146 |
+
},
|
| 1147 |
+
{
|
| 1148 |
+
"epoch": 2.17,
|
| 1149 |
+
"learning_rate": 4.292513144386025e-05,
|
| 1150 |
+
"loss": 0.5107,
|
| 1151 |
+
"step": 93500
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 2.18,
|
| 1155 |
+
"learning_rate": 4.28862427278101e-05,
|
| 1156 |
+
"loss": 0.5117,
|
| 1157 |
+
"step": 94000
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 2.2,
|
| 1161 |
+
"learning_rate": 4.284750956662415e-05,
|
| 1162 |
+
"loss": 0.5129,
|
| 1163 |
+
"step": 94500
|
| 1164 |
+
},
|
| 1165 |
+
{
|
| 1166 |
+
"epoch": 2.21,
|
| 1167 |
+
"learning_rate": 4.2808620850574e-05,
|
| 1168 |
+
"loss": 0.5119,
|
| 1169 |
+
"step": 95000
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 2.22,
|
| 1173 |
+
"learning_rate": 4.276973213452385e-05,
|
| 1174 |
+
"loss": 0.5198,
|
| 1175 |
+
"step": 95500
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 2.23,
|
| 1179 |
+
"learning_rate": 4.27308434184737e-05,
|
| 1180 |
+
"loss": 0.5138,
|
| 1181 |
+
"step": 96000
|
| 1182 |
+
},
|
| 1183 |
+
{
|
| 1184 |
+
"epoch": 2.24,
|
| 1185 |
+
"learning_rate": 4.269195470242355e-05,
|
| 1186 |
+
"loss": 0.516,
|
| 1187 |
+
"step": 96500
|
| 1188 |
+
},
|
| 1189 |
+
{
|
| 1190 |
+
"epoch": 2.25,
|
| 1191 |
+
"learning_rate": 4.2653065986373395e-05,
|
| 1192 |
+
"loss": 0.521,
|
| 1193 |
+
"step": 97000
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 2.27,
|
| 1197 |
+
"learning_rate": 4.2614177270323247e-05,
|
| 1198 |
+
"loss": 0.5162,
|
| 1199 |
+
"step": 97500
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 2.28,
|
| 1203 |
+
"learning_rate": 4.25752885542731e-05,
|
| 1204 |
+
"loss": 0.5175,
|
| 1205 |
+
"step": 98000
|
| 1206 |
+
},
|
| 1207 |
+
{
|
| 1208 |
+
"epoch": 2.29,
|
| 1209 |
+
"learning_rate": 4.253639983822294e-05,
|
| 1210 |
+
"loss": 0.518,
|
| 1211 |
+
"step": 98500
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"epoch": 2.3,
|
| 1215 |
+
"learning_rate": 4.249758889960489e-05,
|
| 1216 |
+
"loss": 0.5234,
|
| 1217 |
+
"step": 99000
|
| 1218 |
+
},
|
| 1219 |
+
{
|
| 1220 |
+
"epoch": 2.31,
|
| 1221 |
+
"learning_rate": 4.245870018355474e-05,
|
| 1222 |
+
"loss": 0.524,
|
| 1223 |
+
"step": 99500
|
| 1224 |
+
},
|
| 1225 |
+
{
|
| 1226 |
+
"epoch": 2.32,
|
| 1227 |
+
"learning_rate": 4.241981146750459e-05,
|
| 1228 |
+
"loss": 0.5263,
|
| 1229 |
+
"step": 100000
|
| 1230 |
+
},
|
| 1231 |
+
{
|
| 1232 |
+
"epoch": 2.34,
|
| 1233 |
+
"learning_rate": 4.238092275145444e-05,
|
| 1234 |
+
"loss": 0.5171,
|
| 1235 |
+
"step": 100500
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 2.35,
|
| 1239 |
+
"learning_rate": 4.2342034035404285e-05,
|
| 1240 |
+
"loss": 0.5265,
|
| 1241 |
+
"step": 101000
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 2.36,
|
| 1245 |
+
"learning_rate": 4.2303145319354137e-05,
|
| 1246 |
+
"loss": 0.5156,
|
| 1247 |
+
"step": 101500
|
| 1248 |
+
},
|
| 1249 |
+
{
|
| 1250 |
+
"epoch": 2.37,
|
| 1251 |
+
"learning_rate": 4.226425660330399e-05,
|
| 1252 |
+
"loss": 0.5255,
|
| 1253 |
+
"step": 102000
|
| 1254 |
+
},
|
| 1255 |
+
{
|
| 1256 |
+
"epoch": 2.38,
|
| 1257 |
+
"learning_rate": 4.222536788725383e-05,
|
| 1258 |
+
"loss": 0.5187,
|
| 1259 |
+
"step": 102500
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"epoch": 2.39,
|
| 1263 |
+
"learning_rate": 4.218655694863579e-05,
|
| 1264 |
+
"loss": 0.5298,
|
| 1265 |
+
"step": 103000
|
| 1266 |
+
},
|
| 1267 |
+
{
|
| 1268 |
+
"epoch": 2.41,
|
| 1269 |
+
"learning_rate": 4.2147668232585634e-05,
|
| 1270 |
+
"loss": 0.5258,
|
| 1271 |
+
"step": 103500
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"epoch": 2.42,
|
| 1275 |
+
"learning_rate": 4.2108779516535485e-05,
|
| 1276 |
+
"loss": 0.5206,
|
| 1277 |
+
"step": 104000
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 2.43,
|
| 1281 |
+
"learning_rate": 4.206989080048534e-05,
|
| 1282 |
+
"loss": 0.5292,
|
| 1283 |
+
"step": 104500
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 2.44,
|
| 1287 |
+
"learning_rate": 4.203100208443518e-05,
|
| 1288 |
+
"loss": 0.5156,
|
| 1289 |
+
"step": 105000
|
| 1290 |
+
},
|
| 1291 |
+
{
|
| 1292 |
+
"epoch": 2.45,
|
| 1293 |
+
"learning_rate": 4.199211336838503e-05,
|
| 1294 |
+
"loss": 0.5156,
|
| 1295 |
+
"step": 105500
|
| 1296 |
+
},
|
| 1297 |
+
{
|
| 1298 |
+
"epoch": 2.46,
|
| 1299 |
+
"learning_rate": 4.1953224652334885e-05,
|
| 1300 |
+
"loss": 0.5259,
|
| 1301 |
+
"step": 106000
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"epoch": 2.48,
|
| 1305 |
+
"learning_rate": 4.191433593628473e-05,
|
| 1306 |
+
"loss": 0.5236,
|
| 1307 |
+
"step": 106500
|
| 1308 |
+
},
|
| 1309 |
+
{
|
| 1310 |
+
"epoch": 2.49,
|
| 1311 |
+
"learning_rate": 4.187552499766668e-05,
|
| 1312 |
+
"loss": 0.5282,
|
| 1313 |
+
"step": 107000
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 2.5,
|
| 1317 |
+
"learning_rate": 4.1836636281616524e-05,
|
| 1318 |
+
"loss": 0.5342,
|
| 1319 |
+
"step": 107500
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 2.51,
|
| 1323 |
+
"learning_rate": 4.1797747565566375e-05,
|
| 1324 |
+
"loss": 0.5246,
|
| 1325 |
+
"step": 108000
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 2.52,
|
| 1329 |
+
"learning_rate": 4.175885884951623e-05,
|
| 1330 |
+
"loss": 0.5238,
|
| 1331 |
+
"step": 108500
|
| 1332 |
+
},
|
| 1333 |
+
{
|
| 1334 |
+
"epoch": 2.53,
|
| 1335 |
+
"learning_rate": 4.171997013346607e-05,
|
| 1336 |
+
"loss": 0.5329,
|
| 1337 |
+
"step": 109000
|
| 1338 |
+
},
|
| 1339 |
+
{
|
| 1340 |
+
"epoch": 2.55,
|
| 1341 |
+
"learning_rate": 4.168108141741592e-05,
|
| 1342 |
+
"loss": 0.5223,
|
| 1343 |
+
"step": 109500
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 2.56,
|
| 1347 |
+
"learning_rate": 4.164234825622997e-05,
|
| 1348 |
+
"loss": 0.5333,
|
| 1349 |
+
"step": 110000
|
| 1350 |
+
},
|
| 1351 |
+
{
|
| 1352 |
+
"epoch": 2.57,
|
| 1353 |
+
"learning_rate": 4.160345954017982e-05,
|
| 1354 |
+
"loss": 0.5353,
|
| 1355 |
+
"step": 110500
|
| 1356 |
+
},
|
| 1357 |
+
{
|
| 1358 |
+
"epoch": 2.58,
|
| 1359 |
+
"learning_rate": 4.1564570824129674e-05,
|
| 1360 |
+
"loss": 0.5285,
|
| 1361 |
+
"step": 111000
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 2.59,
|
| 1365 |
+
"learning_rate": 4.152568210807952e-05,
|
| 1366 |
+
"loss": 0.5365,
|
| 1367 |
+
"step": 111500
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 2.6,
|
| 1371 |
+
"learning_rate": 4.1486948946893573e-05,
|
| 1372 |
+
"loss": 0.5214,
|
| 1373 |
+
"step": 112000
|
| 1374 |
+
},
|
| 1375 |
+
{
|
| 1376 |
+
"epoch": 2.61,
|
| 1377 |
+
"learning_rate": 4.1448060230843425e-05,
|
| 1378 |
+
"loss": 0.534,
|
| 1379 |
+
"step": 112500
|
| 1380 |
+
},
|
| 1381 |
+
{
|
| 1382 |
+
"epoch": 2.63,
|
| 1383 |
+
"learning_rate": 4.140917151479327e-05,
|
| 1384 |
+
"loss": 0.5375,
|
| 1385 |
+
"step": 113000
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"epoch": 2.64,
|
| 1389 |
+
"learning_rate": 4.137028279874312e-05,
|
| 1390 |
+
"loss": 0.532,
|
| 1391 |
+
"step": 113500
|
| 1392 |
+
},
|
| 1393 |
+
{
|
| 1394 |
+
"epoch": 2.65,
|
| 1395 |
+
"learning_rate": 4.133139408269297e-05,
|
| 1396 |
+
"loss": 0.5232,
|
| 1397 |
+
"step": 114000
|
| 1398 |
+
},
|
| 1399 |
+
{
|
| 1400 |
+
"epoch": 2.66,
|
| 1401 |
+
"learning_rate": 4.129250536664282e-05,
|
| 1402 |
+
"loss": 0.5333,
|
| 1403 |
+
"step": 114500
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 2.67,
|
| 1407 |
+
"learning_rate": 4.125361665059266e-05,
|
| 1408 |
+
"loss": 0.5269,
|
| 1409 |
+
"step": 115000
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 2.68,
|
| 1413 |
+
"learning_rate": 4.1214727934542514e-05,
|
| 1414 |
+
"loss": 0.5186,
|
| 1415 |
+
"step": 115500
|
| 1416 |
+
},
|
| 1417 |
+
{
|
| 1418 |
+
"epoch": 2.7,
|
| 1419 |
+
"learning_rate": 4.1175839218492365e-05,
|
| 1420 |
+
"loss": 0.5226,
|
| 1421 |
+
"step": 116000
|
| 1422 |
+
},
|
| 1423 |
+
{
|
| 1424 |
+
"epoch": 2.71,
|
| 1425 |
+
"learning_rate": 4.1137028279874315e-05,
|
| 1426 |
+
"loss": 0.5278,
|
| 1427 |
+
"step": 116500
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 2.72,
|
| 1431 |
+
"learning_rate": 4.109813956382416e-05,
|
| 1432 |
+
"loss": 0.5311,
|
| 1433 |
+
"step": 117000
|
| 1434 |
+
},
|
| 1435 |
+
{
|
| 1436 |
+
"epoch": 2.73,
|
| 1437 |
+
"learning_rate": 4.105925084777401e-05,
|
| 1438 |
+
"loss": 0.5274,
|
| 1439 |
+
"step": 117500
|
| 1440 |
+
},
|
| 1441 |
+
{
|
| 1442 |
+
"epoch": 2.74,
|
| 1443 |
+
"learning_rate": 4.102036213172386e-05,
|
| 1444 |
+
"loss": 0.5348,
|
| 1445 |
+
"step": 118000
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 2.75,
|
| 1449 |
+
"learning_rate": 4.098147341567371e-05,
|
| 1450 |
+
"loss": 0.5359,
|
| 1451 |
+
"step": 118500
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 2.77,
|
| 1455 |
+
"learning_rate": 4.094258469962356e-05,
|
| 1456 |
+
"loss": 0.5416,
|
| 1457 |
+
"step": 119000
|
| 1458 |
+
},
|
| 1459 |
+
{
|
| 1460 |
+
"epoch": 2.78,
|
| 1461 |
+
"learning_rate": 4.090369598357341e-05,
|
| 1462 |
+
"loss": 0.526,
|
| 1463 |
+
"step": 119500
|
| 1464 |
+
},
|
| 1465 |
+
{
|
| 1466 |
+
"epoch": 2.79,
|
| 1467 |
+
"learning_rate": 4.086488504495536e-05,
|
| 1468 |
+
"loss": 0.5404,
|
| 1469 |
+
"step": 120000
|
| 1470 |
+
},
|
| 1471 |
+
{
|
| 1472 |
+
"epoch": 2.8,
|
| 1473 |
+
"learning_rate": 4.0825996328905205e-05,
|
| 1474 |
+
"loss": 0.5321,
|
| 1475 |
+
"step": 120500
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"epoch": 2.81,
|
| 1479 |
+
"learning_rate": 4.0787107612855056e-05,
|
| 1480 |
+
"loss": 0.5378,
|
| 1481 |
+
"step": 121000
|
| 1482 |
+
},
|
| 1483 |
+
{
|
| 1484 |
+
"epoch": 2.82,
|
| 1485 |
+
"learning_rate": 4.074821889680491e-05,
|
| 1486 |
+
"loss": 0.5371,
|
| 1487 |
+
"step": 121500
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 2.84,
|
| 1491 |
+
"learning_rate": 4.070933018075476e-05,
|
| 1492 |
+
"loss": 0.5303,
|
| 1493 |
+
"step": 122000
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 2.85,
|
| 1497 |
+
"learning_rate": 4.06704414647046e-05,
|
| 1498 |
+
"loss": 0.5321,
|
| 1499 |
+
"step": 122500
|
| 1500 |
+
},
|
| 1501 |
+
{
|
| 1502 |
+
"epoch": 2.86,
|
| 1503 |
+
"learning_rate": 4.063163052608655e-05,
|
| 1504 |
+
"loss": 0.5291,
|
| 1505 |
+
"step": 123000
|
| 1506 |
+
},
|
| 1507 |
+
{
|
| 1508 |
+
"epoch": 2.87,
|
| 1509 |
+
"learning_rate": 4.05927418100364e-05,
|
| 1510 |
+
"loss": 0.5391,
|
| 1511 |
+
"step": 123500
|
| 1512 |
+
},
|
| 1513 |
+
{
|
| 1514 |
+
"epoch": 2.88,
|
| 1515 |
+
"learning_rate": 4.055385309398625e-05,
|
| 1516 |
+
"loss": 0.5344,
|
| 1517 |
+
"step": 124000
|
| 1518 |
+
},
|
| 1519 |
+
{
|
| 1520 |
+
"epoch": 2.89,
|
| 1521 |
+
"learning_rate": 4.05149643779361e-05,
|
| 1522 |
+
"loss": 0.5381,
|
| 1523 |
+
"step": 124500
|
| 1524 |
+
},
|
| 1525 |
+
{
|
| 1526 |
+
"epoch": 2.91,
|
| 1527 |
+
"learning_rate": 4.0476075661885946e-05,
|
| 1528 |
+
"loss": 0.5348,
|
| 1529 |
+
"step": 125000
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 2.92,
|
| 1533 |
+
"learning_rate": 4.04371869458358e-05,
|
| 1534 |
+
"loss": 0.5367,
|
| 1535 |
+
"step": 125500
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 2.93,
|
| 1539 |
+
"learning_rate": 4.039829822978565e-05,
|
| 1540 |
+
"loss": 0.5294,
|
| 1541 |
+
"step": 126000
|
| 1542 |
+
},
|
| 1543 |
+
{
|
| 1544 |
+
"epoch": 2.94,
|
| 1545 |
+
"learning_rate": 4.0359409513735494e-05,
|
| 1546 |
+
"loss": 0.5257,
|
| 1547 |
+
"step": 126500
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"epoch": 2.95,
|
| 1551 |
+
"learning_rate": 4.0320520797685346e-05,
|
| 1552 |
+
"loss": 0.5241,
|
| 1553 |
+
"step": 127000
|
| 1554 |
+
},
|
| 1555 |
+
{
|
| 1556 |
+
"epoch": 2.96,
|
| 1557 |
+
"learning_rate": 4.02816320816352e-05,
|
| 1558 |
+
"loss": 0.5375,
|
| 1559 |
+
"step": 127500
|
| 1560 |
+
},
|
| 1561 |
+
{
|
| 1562 |
+
"epoch": 2.98,
|
| 1563 |
+
"learning_rate": 4.024274336558504e-05,
|
| 1564 |
+
"loss": 0.5332,
|
| 1565 |
+
"step": 128000
|
| 1566 |
+
},
|
| 1567 |
+
{
|
| 1568 |
+
"epoch": 2.99,
|
| 1569 |
+
"learning_rate": 4.0203854649534894e-05,
|
| 1570 |
+
"loss": 0.532,
|
| 1571 |
+
"step": 128500
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 3.0,
|
| 1575 |
+
"learning_rate": 4.0164965933484745e-05,
|
| 1576 |
+
"loss": 0.5464,
|
| 1577 |
+
"step": 129000
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 3.0,
|
| 1581 |
+
"eval_bleu": 58.837,
|
| 1582 |
+
"eval_gen_len": 15.7689,
|
| 1583 |
+
"eval_loss": 0.6568061113357544,
|
| 1584 |
+
"eval_runtime": 6740.2004,
|
| 1585 |
+
"eval_samples_per_second": 12.747,
|
| 1586 |
+
"eval_steps_per_second": 1.593,
|
| 1587 |
+
"step": 129072
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 3.01,
|
| 1591 |
+
"learning_rate": 4.0126077217434597e-05,
|
| 1592 |
+
"loss": 0.4193,
|
| 1593 |
+
"step": 129500
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 3.02,
|
| 1597 |
+
"learning_rate": 4.0087344056248644e-05,
|
| 1598 |
+
"loss": 0.399,
|
| 1599 |
+
"step": 130000
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 3.03,
|
| 1603 |
+
"learning_rate": 4.004853311763059e-05,
|
| 1604 |
+
"loss": 0.3996,
|
| 1605 |
+
"step": 130500
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 3.04,
|
| 1609 |
+
"learning_rate": 4.0009799956444635e-05,
|
| 1610 |
+
"loss": 0.3937,
|
| 1611 |
+
"step": 131000
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 3.06,
|
| 1615 |
+
"learning_rate": 3.997091124039449e-05,
|
| 1616 |
+
"loss": 0.3955,
|
| 1617 |
+
"step": 131500
|
| 1618 |
+
},
|
| 1619 |
+
{
|
| 1620 |
+
"epoch": 3.07,
|
| 1621 |
+
"learning_rate": 3.993202252434434e-05,
|
| 1622 |
+
"loss": 0.3998,
|
| 1623 |
+
"step": 132000
|
| 1624 |
+
},
|
| 1625 |
+
{
|
| 1626 |
+
"epoch": 3.08,
|
| 1627 |
+
"learning_rate": 3.989313380829419e-05,
|
| 1628 |
+
"loss": 0.3934,
|
| 1629 |
+
"step": 132500
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 3.09,
|
| 1633 |
+
"learning_rate": 3.9854245092244034e-05,
|
| 1634 |
+
"loss": 0.3987,
|
| 1635 |
+
"step": 133000
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 3.1,
|
| 1639 |
+
"learning_rate": 3.9815434153625984e-05,
|
| 1640 |
+
"loss": 0.4048,
|
| 1641 |
+
"step": 133500
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 3.11,
|
| 1645 |
+
"learning_rate": 3.9776545437575836e-05,
|
| 1646 |
+
"loss": 0.406,
|
| 1647 |
+
"step": 134000
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 3.13,
|
| 1651 |
+
"learning_rate": 3.973765672152569e-05,
|
| 1652 |
+
"loss": 0.3987,
|
| 1653 |
+
"step": 134500
|
| 1654 |
+
},
|
| 1655 |
+
{
|
| 1656 |
+
"epoch": 3.14,
|
| 1657 |
+
"learning_rate": 3.969876800547553e-05,
|
| 1658 |
+
"loss": 0.4085,
|
| 1659 |
+
"step": 135000
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 3.15,
|
| 1663 |
+
"learning_rate": 3.9659879289425383e-05,
|
| 1664 |
+
"loss": 0.4047,
|
| 1665 |
+
"step": 135500
|
| 1666 |
+
},
|
| 1667 |
+
{
|
| 1668 |
+
"epoch": 3.16,
|
| 1669 |
+
"learning_rate": 3.9620990573375235e-05,
|
| 1670 |
+
"loss": 0.4021,
|
| 1671 |
+
"step": 136000
|
| 1672 |
+
},
|
| 1673 |
+
{
|
| 1674 |
+
"epoch": 3.17,
|
| 1675 |
+
"learning_rate": 3.958210185732508e-05,
|
| 1676 |
+
"loss": 0.4056,
|
| 1677 |
+
"step": 136500
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 3.18,
|
| 1681 |
+
"learning_rate": 3.954321314127493e-05,
|
| 1682 |
+
"loss": 0.4019,
|
| 1683 |
+
"step": 137000
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 3.2,
|
| 1687 |
+
"learning_rate": 3.950440220265688e-05,
|
| 1688 |
+
"loss": 0.4125,
|
| 1689 |
+
"step": 137500
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 3.21,
|
| 1693 |
+
"learning_rate": 3.9465513486606726e-05,
|
| 1694 |
+
"loss": 0.4087,
|
| 1695 |
+
"step": 138000
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 3.22,
|
| 1699 |
+
"learning_rate": 3.942662477055658e-05,
|
| 1700 |
+
"loss": 0.3971,
|
| 1701 |
+
"step": 138500
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 3.23,
|
| 1705 |
+
"learning_rate": 3.938773605450642e-05,
|
| 1706 |
+
"loss": 0.4122,
|
| 1707 |
+
"step": 139000
|
| 1708 |
+
},
|
| 1709 |
+
{
|
| 1710 |
+
"epoch": 3.24,
|
| 1711 |
+
"learning_rate": 3.9348847338456273e-05,
|
| 1712 |
+
"loss": 0.4086,
|
| 1713 |
+
"step": 139500
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 3.25,
|
| 1717 |
+
"learning_rate": 3.931003639983822e-05,
|
| 1718 |
+
"loss": 0.4173,
|
| 1719 |
+
"step": 140000
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 3.27,
|
| 1723 |
+
"learning_rate": 3.9271147683788075e-05,
|
| 1724 |
+
"loss": 0.4158,
|
| 1725 |
+
"step": 140500
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 3.28,
|
| 1729 |
+
"learning_rate": 3.923225896773792e-05,
|
| 1730 |
+
"loss": 0.4072,
|
| 1731 |
+
"step": 141000
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 3.29,
|
| 1735 |
+
"learning_rate": 3.919337025168777e-05,
|
| 1736 |
+
"loss": 0.4156,
|
| 1737 |
+
"step": 141500
|
| 1738 |
+
},
|
| 1739 |
+
{
|
| 1740 |
+
"epoch": 3.3,
|
| 1741 |
+
"learning_rate": 3.915448153563762e-05,
|
| 1742 |
+
"loss": 0.4174,
|
| 1743 |
+
"step": 142000
|
| 1744 |
+
},
|
| 1745 |
+
{
|
| 1746 |
+
"epoch": 3.31,
|
| 1747 |
+
"learning_rate": 3.9115592819587474e-05,
|
| 1748 |
+
"loss": 0.416,
|
| 1749 |
+
"step": 142500
|
| 1750 |
+
},
|
| 1751 |
+
{
|
| 1752 |
+
"epoch": 3.32,
|
| 1753 |
+
"learning_rate": 3.9076781880969424e-05,
|
| 1754 |
+
"loss": 0.4151,
|
| 1755 |
+
"step": 143000
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 3.34,
|
| 1759 |
+
"learning_rate": 3.903789316491927e-05,
|
| 1760 |
+
"loss": 0.4188,
|
| 1761 |
+
"step": 143500
|
| 1762 |
+
},
|
| 1763 |
+
{
|
| 1764 |
+
"epoch": 3.35,
|
| 1765 |
+
"learning_rate": 3.899900444886912e-05,
|
| 1766 |
+
"loss": 0.419,
|
| 1767 |
+
"step": 144000
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 3.36,
|
| 1771 |
+
"learning_rate": 3.896011573281897e-05,
|
| 1772 |
+
"loss": 0.4206,
|
| 1773 |
+
"step": 144500
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 3.37,
|
| 1777 |
+
"learning_rate": 3.8921227016768816e-05,
|
| 1778 |
+
"loss": 0.4243,
|
| 1779 |
+
"step": 145000
|
| 1780 |
+
},
|
| 1781 |
+
{
|
| 1782 |
+
"epoch": 3.38,
|
| 1783 |
+
"learning_rate": 3.888233830071866e-05,
|
| 1784 |
+
"loss": 0.4131,
|
| 1785 |
+
"step": 145500
|
| 1786 |
+
},
|
| 1787 |
+
{
|
| 1788 |
+
"epoch": 3.39,
|
| 1789 |
+
"learning_rate": 3.884344958466851e-05,
|
| 1790 |
+
"loss": 0.4192,
|
| 1791 |
+
"step": 146000
|
| 1792 |
+
},
|
| 1793 |
+
{
|
| 1794 |
+
"epoch": 3.41,
|
| 1795 |
+
"learning_rate": 3.8804560868618364e-05,
|
| 1796 |
+
"loss": 0.4179,
|
| 1797 |
+
"step": 146500
|
| 1798 |
+
},
|
| 1799 |
+
{
|
| 1800 |
+
"epoch": 3.42,
|
| 1801 |
+
"learning_rate": 3.876567215256821e-05,
|
| 1802 |
+
"loss": 0.4264,
|
| 1803 |
+
"step": 147000
|
| 1804 |
+
},
|
| 1805 |
+
{
|
| 1806 |
+
"epoch": 3.43,
|
| 1807 |
+
"learning_rate": 3.872678343651806e-05,
|
| 1808 |
+
"loss": 0.4231,
|
| 1809 |
+
"step": 147500
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 3.44,
|
| 1813 |
+
"learning_rate": 3.868789472046791e-05,
|
| 1814 |
+
"loss": 0.4217,
|
| 1815 |
+
"step": 148000
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 3.45,
|
| 1819 |
+
"learning_rate": 3.8649006004417756e-05,
|
| 1820 |
+
"loss": 0.4189,
|
| 1821 |
+
"step": 148500
|
| 1822 |
+
},
|
| 1823 |
+
{
|
| 1824 |
+
"epoch": 3.46,
|
| 1825 |
+
"learning_rate": 3.861011728836761e-05,
|
| 1826 |
+
"loss": 0.4198,
|
| 1827 |
+
"step": 149000
|
| 1828 |
+
},
|
| 1829 |
+
{
|
| 1830 |
+
"epoch": 3.47,
|
| 1831 |
+
"learning_rate": 3.857122857231746e-05,
|
| 1832 |
+
"loss": 0.4237,
|
| 1833 |
+
"step": 149500
|
| 1834 |
+
},
|
| 1835 |
+
{
|
| 1836 |
+
"epoch": 3.49,
|
| 1837 |
+
"learning_rate": 3.853233985626731e-05,
|
| 1838 |
+
"loss": 0.4147,
|
| 1839 |
+
"step": 150000
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 3.5,
|
| 1843 |
+
"learning_rate": 3.849352891764926e-05,
|
| 1844 |
+
"loss": 0.4219,
|
| 1845 |
+
"step": 150500
|
| 1846 |
+
},
|
| 1847 |
+
{
|
| 1848 |
+
"epoch": 3.51,
|
| 1849 |
+
"learning_rate": 3.8454640201599105e-05,
|
| 1850 |
+
"loss": 0.4274,
|
| 1851 |
+
"step": 151000
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 3.52,
|
| 1855 |
+
"learning_rate": 3.8415829262981055e-05,
|
| 1856 |
+
"loss": 0.4283,
|
| 1857 |
+
"step": 151500
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 3.53,
|
| 1861 |
+
"learning_rate": 3.8376940546930907e-05,
|
| 1862 |
+
"loss": 0.4222,
|
| 1863 |
+
"step": 152000
|
| 1864 |
+
},
|
| 1865 |
+
{
|
| 1866 |
+
"epoch": 3.54,
|
| 1867 |
+
"learning_rate": 3.833805183088076e-05,
|
| 1868 |
+
"loss": 0.4269,
|
| 1869 |
+
"step": 152500
|
| 1870 |
+
},
|
| 1871 |
+
{
|
| 1872 |
+
"epoch": 3.56,
|
| 1873 |
+
"learning_rate": 3.82991631148306e-05,
|
| 1874 |
+
"loss": 0.4271,
|
| 1875 |
+
"step": 153000
|
| 1876 |
+
},
|
| 1877 |
+
{
|
| 1878 |
+
"epoch": 3.57,
|
| 1879 |
+
"learning_rate": 3.826035217621255e-05,
|
| 1880 |
+
"loss": 0.425,
|
| 1881 |
+
"step": 153500
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 3.58,
|
| 1885 |
+
"learning_rate": 3.82214634601624e-05,
|
| 1886 |
+
"loss": 0.4222,
|
| 1887 |
+
"step": 154000
|
| 1888 |
+
},
|
| 1889 |
+
{
|
| 1890 |
+
"epoch": 3.59,
|
| 1891 |
+
"learning_rate": 3.818257474411225e-05,
|
| 1892 |
+
"loss": 0.4236,
|
| 1893 |
+
"step": 154500
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 3.6,
|
| 1897 |
+
"learning_rate": 3.81436860280621e-05,
|
| 1898 |
+
"loss": 0.4327,
|
| 1899 |
+
"step": 155000
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 3.61,
|
| 1903 |
+
"learning_rate": 3.8104797312011945e-05,
|
| 1904 |
+
"loss": 0.4268,
|
| 1905 |
+
"step": 155500
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 3.63,
|
| 1909 |
+
"learning_rate": 3.8065908595961797e-05,
|
| 1910 |
+
"loss": 0.4319,
|
| 1911 |
+
"step": 156000
|
| 1912 |
+
},
|
| 1913 |
+
{
|
| 1914 |
+
"epoch": 3.64,
|
| 1915 |
+
"learning_rate": 3.802701987991165e-05,
|
| 1916 |
+
"loss": 0.4312,
|
| 1917 |
+
"step": 156500
|
| 1918 |
+
},
|
| 1919 |
+
{
|
| 1920 |
+
"epoch": 3.65,
|
| 1921 |
+
"learning_rate": 3.798813116386149e-05,
|
| 1922 |
+
"loss": 0.4328,
|
| 1923 |
+
"step": 157000
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 3.66,
|
| 1927 |
+
"learning_rate": 3.7949242447811344e-05,
|
| 1928 |
+
"loss": 0.4292,
|
| 1929 |
+
"step": 157500
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"epoch": 3.67,
|
| 1933 |
+
"learning_rate": 3.7910353731761196e-05,
|
| 1934 |
+
"loss": 0.4317,
|
| 1935 |
+
"step": 158000
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 3.68,
|
| 1939 |
+
"learning_rate": 3.787146501571104e-05,
|
| 1940 |
+
"loss": 0.4309,
|
| 1941 |
+
"step": 158500
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 3.7,
|
| 1945 |
+
"learning_rate": 3.7832654077093e-05,
|
| 1946 |
+
"loss": 0.4425,
|
| 1947 |
+
"step": 159000
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"epoch": 3.71,
|
| 1951 |
+
"learning_rate": 3.779376536104284e-05,
|
| 1952 |
+
"loss": 0.4335,
|
| 1953 |
+
"step": 159500
|
| 1954 |
+
},
|
| 1955 |
+
{
|
| 1956 |
+
"epoch": 3.72,
|
| 1957 |
+
"learning_rate": 3.775487664499269e-05,
|
| 1958 |
+
"loss": 0.4295,
|
| 1959 |
+
"step": 160000
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 3.73,
|
| 1963 |
+
"learning_rate": 3.7715987928942545e-05,
|
| 1964 |
+
"loss": 0.4305,
|
| 1965 |
+
"step": 160500
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 3.74,
|
| 1969 |
+
"learning_rate": 3.767709921289239e-05,
|
| 1970 |
+
"loss": 0.4264,
|
| 1971 |
+
"step": 161000
|
| 1972 |
+
},
|
| 1973 |
+
{
|
| 1974 |
+
"epoch": 3.75,
|
| 1975 |
+
"learning_rate": 3.7638210496842234e-05,
|
| 1976 |
+
"loss": 0.43,
|
| 1977 |
+
"step": 161500
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 3.77,
|
| 1981 |
+
"learning_rate": 3.7599321780792086e-05,
|
| 1982 |
+
"loss": 0.4329,
|
| 1983 |
+
"step": 162000
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 3.78,
|
| 1987 |
+
"learning_rate": 3.756043306474194e-05,
|
| 1988 |
+
"loss": 0.4324,
|
| 1989 |
+
"step": 162500
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 3.79,
|
| 1993 |
+
"learning_rate": 3.752154434869178e-05,
|
| 1994 |
+
"loss": 0.434,
|
| 1995 |
+
"step": 163000
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 3.8,
|
| 1999 |
+
"learning_rate": 3.7482655632641634e-05,
|
| 2000 |
+
"loss": 0.4317,
|
| 2001 |
+
"step": 163500
|
| 2002 |
+
},
|
| 2003 |
+
{
|
| 2004 |
+
"epoch": 3.81,
|
| 2005 |
+
"learning_rate": 3.744384469402358e-05,
|
| 2006 |
+
"loss": 0.4379,
|
| 2007 |
+
"step": 164000
|
| 2008 |
+
},
|
| 2009 |
+
{
|
| 2010 |
+
"epoch": 3.82,
|
| 2011 |
+
"learning_rate": 3.7404955977973435e-05,
|
| 2012 |
+
"loss": 0.4427,
|
| 2013 |
+
"step": 164500
|
| 2014 |
+
},
|
| 2015 |
+
{
|
| 2016 |
+
"epoch": 3.84,
|
| 2017 |
+
"learning_rate": 3.736606726192328e-05,
|
| 2018 |
+
"loss": 0.433,
|
| 2019 |
+
"step": 165000
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 3.85,
|
| 2023 |
+
"learning_rate": 3.732717854587313e-05,
|
| 2024 |
+
"loss": 0.4407,
|
| 2025 |
+
"step": 165500
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 3.86,
|
| 2029 |
+
"learning_rate": 3.728828982982298e-05,
|
| 2030 |
+
"loss": 0.4416,
|
| 2031 |
+
"step": 166000
|
| 2032 |
+
},
|
| 2033 |
+
{
|
| 2034 |
+
"epoch": 3.87,
|
| 2035 |
+
"learning_rate": 3.724940111377283e-05,
|
| 2036 |
+
"loss": 0.4407,
|
| 2037 |
+
"step": 166500
|
| 2038 |
+
},
|
| 2039 |
+
{
|
| 2040 |
+
"epoch": 3.88,
|
| 2041 |
+
"learning_rate": 3.721051239772268e-05,
|
| 2042 |
+
"loss": 0.4379,
|
| 2043 |
+
"step": 167000
|
| 2044 |
+
},
|
| 2045 |
+
{
|
| 2046 |
+
"epoch": 3.89,
|
| 2047 |
+
"learning_rate": 3.717170145910463e-05,
|
| 2048 |
+
"loss": 0.4345,
|
| 2049 |
+
"step": 167500
|
| 2050 |
+
},
|
| 2051 |
+
{
|
| 2052 |
+
"epoch": 3.9,
|
| 2053 |
+
"learning_rate": 3.713289052048658e-05,
|
| 2054 |
+
"loss": 0.435,
|
| 2055 |
+
"step": 168000
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 3.92,
|
| 2059 |
+
"learning_rate": 3.709400180443643e-05,
|
| 2060 |
+
"loss": 0.4366,
|
| 2061 |
+
"step": 168500
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 3.93,
|
| 2065 |
+
"learning_rate": 3.7055113088386275e-05,
|
| 2066 |
+
"loss": 0.4336,
|
| 2067 |
+
"step": 169000
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 3.94,
|
| 2071 |
+
"learning_rate": 3.701622437233612e-05,
|
| 2072 |
+
"loss": 0.4438,
|
| 2073 |
+
"step": 169500
|
| 2074 |
+
},
|
| 2075 |
+
{
|
| 2076 |
+
"epoch": 3.95,
|
| 2077 |
+
"learning_rate": 3.697741343371807e-05,
|
| 2078 |
+
"loss": 0.4389,
|
| 2079 |
+
"step": 170000
|
| 2080 |
+
},
|
| 2081 |
+
{
|
| 2082 |
+
"epoch": 3.96,
|
| 2083 |
+
"learning_rate": 3.693852471766792e-05,
|
| 2084 |
+
"loss": 0.4394,
|
| 2085 |
+
"step": 170500
|
| 2086 |
+
},
|
| 2087 |
+
{
|
| 2088 |
+
"epoch": 3.97,
|
| 2089 |
+
"learning_rate": 3.689963600161777e-05,
|
| 2090 |
+
"loss": 0.4378,
|
| 2091 |
+
"step": 171000
|
| 2092 |
+
},
|
| 2093 |
+
{
|
| 2094 |
+
"epoch": 3.99,
|
| 2095 |
+
"learning_rate": 3.686082506299972e-05,
|
| 2096 |
+
"loss": 0.4443,
|
| 2097 |
+
"step": 171500
|
| 2098 |
+
},
|
| 2099 |
+
{
|
| 2100 |
+
"epoch": 4.0,
|
| 2101 |
+
"learning_rate": 3.682193634694957e-05,
|
| 2102 |
+
"loss": 0.4444,
|
| 2103 |
+
"step": 172000
|
| 2104 |
+
},
|
| 2105 |
+
{
|
| 2106 |
+
"epoch": 4.0,
|
| 2107 |
+
"eval_bleu": 59.3055,
|
| 2108 |
+
"eval_gen_len": 15.7545,
|
| 2109 |
+
"eval_loss": 0.6776524782180786,
|
| 2110 |
+
"eval_runtime": 6664.893,
|
| 2111 |
+
"eval_samples_per_second": 12.891,
|
| 2112 |
+
"eval_steps_per_second": 1.611,
|
| 2113 |
+
"step": 172096
|
| 2114 |
+
},
|
| 2115 |
+
{
|
| 2116 |
+
"epoch": 4.01,
|
| 2117 |
+
"learning_rate": 3.678304763089942e-05,
|
| 2118 |
+
"loss": 0.3253,
|
| 2119 |
+
"step": 172500
|
| 2120 |
+
},
|
| 2121 |
+
{
|
| 2122 |
+
"epoch": 4.02,
|
| 2123 |
+
"learning_rate": 3.674415891484927e-05,
|
| 2124 |
+
"loss": 0.302,
|
| 2125 |
+
"step": 173000
|
| 2126 |
+
},
|
| 2127 |
+
{
|
| 2128 |
+
"epoch": 4.03,
|
| 2129 |
+
"learning_rate": 3.670534797623122e-05,
|
| 2130 |
+
"loss": 0.3034,
|
| 2131 |
+
"step": 173500
|
| 2132 |
+
},
|
| 2133 |
+
{
|
| 2134 |
+
"epoch": 4.04,
|
| 2135 |
+
"learning_rate": 3.666645926018107e-05,
|
| 2136 |
+
"loss": 0.3075,
|
| 2137 |
+
"step": 174000
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 4.06,
|
| 2141 |
+
"learning_rate": 3.6627570544130915e-05,
|
| 2142 |
+
"loss": 0.3039,
|
| 2143 |
+
"step": 174500
|
| 2144 |
+
},
|
| 2145 |
+
{
|
| 2146 |
+
"epoch": 4.07,
|
| 2147 |
+
"learning_rate": 3.6588759605512865e-05,
|
| 2148 |
+
"loss": 0.3102,
|
| 2149 |
+
"step": 175000
|
| 2150 |
+
},
|
| 2151 |
+
{
|
| 2152 |
+
"epoch": 4.08,
|
| 2153 |
+
"learning_rate": 3.6549870889462717e-05,
|
| 2154 |
+
"loss": 0.3005,
|
| 2155 |
+
"step": 175500
|
| 2156 |
+
},
|
| 2157 |
+
{
|
| 2158 |
+
"epoch": 4.09,
|
| 2159 |
+
"learning_rate": 3.651098217341257e-05,
|
| 2160 |
+
"loss": 0.3052,
|
| 2161 |
+
"step": 176000
|
| 2162 |
+
},
|
| 2163 |
+
{
|
| 2164 |
+
"epoch": 4.1,
|
| 2165 |
+
"learning_rate": 3.647217123479451e-05,
|
| 2166 |
+
"loss": 0.313,
|
| 2167 |
+
"step": 176500
|
| 2168 |
+
},
|
| 2169 |
+
{
|
| 2170 |
+
"epoch": 4.11,
|
| 2171 |
+
"learning_rate": 3.643328251874436e-05,
|
| 2172 |
+
"loss": 0.3103,
|
| 2173 |
+
"step": 177000
|
| 2174 |
+
},
|
| 2175 |
+
{
|
| 2176 |
+
"epoch": 4.13,
|
| 2177 |
+
"learning_rate": 3.639439380269421e-05,
|
| 2178 |
+
"loss": 0.3167,
|
| 2179 |
+
"step": 177500
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 4.14,
|
| 2183 |
+
"learning_rate": 3.635550508664406e-05,
|
| 2184 |
+
"loss": 0.3129,
|
| 2185 |
+
"step": 178000
|
| 2186 |
+
},
|
| 2187 |
+
{
|
| 2188 |
+
"epoch": 4.15,
|
| 2189 |
+
"learning_rate": 3.631669414802601e-05,
|
| 2190 |
+
"loss": 0.3067,
|
| 2191 |
+
"step": 178500
|
| 2192 |
+
},
|
| 2193 |
+
{
|
| 2194 |
+
"epoch": 4.16,
|
| 2195 |
+
"learning_rate": 3.627780543197586e-05,
|
| 2196 |
+
"loss": 0.3091,
|
| 2197 |
+
"step": 179000
|
| 2198 |
+
},
|
| 2199 |
+
{
|
| 2200 |
+
"epoch": 4.17,
|
| 2201 |
+
"learning_rate": 3.623891671592571e-05,
|
| 2202 |
+
"loss": 0.3131,
|
| 2203 |
+
"step": 179500
|
| 2204 |
+
},
|
| 2205 |
+
{
|
| 2206 |
+
"epoch": 4.18,
|
| 2207 |
+
"learning_rate": 3.6200027999875556e-05,
|
| 2208 |
+
"loss": 0.3135,
|
| 2209 |
+
"step": 180000
|
| 2210 |
+
},
|
| 2211 |
+
{
|
| 2212 |
+
"epoch": 4.2,
|
| 2213 |
+
"learning_rate": 3.616113928382541e-05,
|
| 2214 |
+
"loss": 0.3133,
|
| 2215 |
+
"step": 180500
|
| 2216 |
+
},
|
| 2217 |
+
{
|
| 2218 |
+
"epoch": 4.21,
|
| 2219 |
+
"learning_rate": 3.612225056777526e-05,
|
| 2220 |
+
"loss": 0.3167,
|
| 2221 |
+
"step": 181000
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 4.22,
|
| 2225 |
+
"learning_rate": 3.6083361851725104e-05,
|
| 2226 |
+
"loss": 0.3129,
|
| 2227 |
+
"step": 181500
|
| 2228 |
+
},
|
| 2229 |
+
{
|
| 2230 |
+
"epoch": 4.23,
|
| 2231 |
+
"learning_rate": 3.6044473135674956e-05,
|
| 2232 |
+
"loss": 0.3232,
|
| 2233 |
+
"step": 182000
|
| 2234 |
+
},
|
| 2235 |
+
{
|
| 2236 |
+
"epoch": 4.24,
|
| 2237 |
+
"learning_rate": 3.600558441962481e-05,
|
| 2238 |
+
"loss": 0.3161,
|
| 2239 |
+
"step": 182500
|
| 2240 |
+
},
|
| 2241 |
+
{
|
| 2242 |
+
"epoch": 4.25,
|
| 2243 |
+
"learning_rate": 3.596669570357465e-05,
|
| 2244 |
+
"loss": 0.3211,
|
| 2245 |
+
"step": 183000
|
| 2246 |
+
},
|
| 2247 |
+
{
|
| 2248 |
+
"epoch": 4.27,
|
| 2249 |
+
"learning_rate": 3.59278847649566e-05,
|
| 2250 |
+
"loss": 0.3248,
|
| 2251 |
+
"step": 183500
|
| 2252 |
+
},
|
| 2253 |
+
{
|
| 2254 |
+
"epoch": 4.28,
|
| 2255 |
+
"learning_rate": 3.588899604890645e-05,
|
| 2256 |
+
"loss": 0.3238,
|
| 2257 |
+
"step": 184000
|
| 2258 |
+
},
|
| 2259 |
+
{
|
| 2260 |
+
"epoch": 4.29,
|
| 2261 |
+
"learning_rate": 3.58501073328563e-05,
|
| 2262 |
+
"loss": 0.3209,
|
| 2263 |
+
"step": 184500
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 4.3,
|
| 2267 |
+
"learning_rate": 3.581121861680615e-05,
|
| 2268 |
+
"loss": 0.318,
|
| 2269 |
+
"step": 185000
|
| 2270 |
+
},
|
| 2271 |
+
{
|
| 2272 |
+
"epoch": 4.31,
|
| 2273 |
+
"learning_rate": 3.5772329900755994e-05,
|
| 2274 |
+
"loss": 0.3255,
|
| 2275 |
+
"step": 185500
|
| 2276 |
+
},
|
| 2277 |
+
{
|
| 2278 |
+
"epoch": 4.32,
|
| 2279 |
+
"learning_rate": 3.5733441184705846e-05,
|
| 2280 |
+
"loss": 0.3265,
|
| 2281 |
+
"step": 186000
|
| 2282 |
+
},
|
| 2283 |
+
{
|
| 2284 |
+
"epoch": 4.33,
|
| 2285 |
+
"learning_rate": 3.56945524686557e-05,
|
| 2286 |
+
"loss": 0.3235,
|
| 2287 |
+
"step": 186500
|
| 2288 |
+
},
|
| 2289 |
+
{
|
| 2290 |
+
"epoch": 4.35,
|
| 2291 |
+
"learning_rate": 3.565574153003765e-05,
|
| 2292 |
+
"loss": 0.3304,
|
| 2293 |
+
"step": 187000
|
| 2294 |
+
},
|
| 2295 |
+
{
|
| 2296 |
+
"epoch": 4.36,
|
| 2297 |
+
"learning_rate": 3.561685281398749e-05,
|
| 2298 |
+
"loss": 0.3331,
|
| 2299 |
+
"step": 187500
|
| 2300 |
+
},
|
| 2301 |
+
{
|
| 2302 |
+
"epoch": 4.37,
|
| 2303 |
+
"learning_rate": 3.557796409793734e-05,
|
| 2304 |
+
"loss": 0.3254,
|
| 2305 |
+
"step": 188000
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 4.38,
|
| 2309 |
+
"learning_rate": 3.5539075381887194e-05,
|
| 2310 |
+
"loss": 0.3305,
|
| 2311 |
+
"step": 188500
|
| 2312 |
+
},
|
| 2313 |
+
{
|
| 2314 |
+
"epoch": 4.39,
|
| 2315 |
+
"learning_rate": 3.5500186665837046e-05,
|
| 2316 |
+
"loss": 0.3264,
|
| 2317 |
+
"step": 189000
|
| 2318 |
+
},
|
| 2319 |
+
{
|
| 2320 |
+
"epoch": 4.4,
|
| 2321 |
+
"learning_rate": 3.546129794978689e-05,
|
| 2322 |
+
"loss": 0.3286,
|
| 2323 |
+
"step": 189500
|
| 2324 |
+
},
|
| 2325 |
+
{
|
| 2326 |
+
"epoch": 4.42,
|
| 2327 |
+
"learning_rate": 3.542240923373674e-05,
|
| 2328 |
+
"loss": 0.3303,
|
| 2329 |
+
"step": 190000
|
| 2330 |
+
},
|
| 2331 |
+
{
|
| 2332 |
+
"epoch": 4.43,
|
| 2333 |
+
"learning_rate": 3.5383520517686594e-05,
|
| 2334 |
+
"loss": 0.3332,
|
| 2335 |
+
"step": 190500
|
| 2336 |
+
},
|
| 2337 |
+
{
|
| 2338 |
+
"epoch": 4.44,
|
| 2339 |
+
"learning_rate": 3.534463180163644e-05,
|
| 2340 |
+
"loss": 0.3284,
|
| 2341 |
+
"step": 191000
|
| 2342 |
+
},
|
| 2343 |
+
{
|
| 2344 |
+
"epoch": 4.45,
|
| 2345 |
+
"learning_rate": 3.530574308558629e-05,
|
| 2346 |
+
"loss": 0.3297,
|
| 2347 |
+
"step": 191500
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 4.46,
|
| 2351 |
+
"learning_rate": 3.526685436953614e-05,
|
| 2352 |
+
"loss": 0.3267,
|
| 2353 |
+
"step": 192000
|
| 2354 |
+
},
|
| 2355 |
+
{
|
| 2356 |
+
"epoch": 4.47,
|
| 2357 |
+
"learning_rate": 3.5227965653485986e-05,
|
| 2358 |
+
"loss": 0.3396,
|
| 2359 |
+
"step": 192500
|
| 2360 |
+
},
|
| 2361 |
+
{
|
| 2362 |
+
"epoch": 4.49,
|
| 2363 |
+
"learning_rate": 3.518907693743583e-05,
|
| 2364 |
+
"loss": 0.3321,
|
| 2365 |
+
"step": 193000
|
| 2366 |
+
},
|
| 2367 |
+
{
|
| 2368 |
+
"epoch": 4.5,
|
| 2369 |
+
"learning_rate": 3.515018822138568e-05,
|
| 2370 |
+
"loss": 0.3411,
|
| 2371 |
+
"step": 193500
|
| 2372 |
+
},
|
| 2373 |
+
{
|
| 2374 |
+
"epoch": 4.51,
|
| 2375 |
+
"learning_rate": 3.5111299505335534e-05,
|
| 2376 |
+
"loss": 0.3356,
|
| 2377 |
+
"step": 194000
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 4.52,
|
| 2381 |
+
"learning_rate": 3.507241078928538e-05,
|
| 2382 |
+
"loss": 0.3367,
|
| 2383 |
+
"step": 194500
|
| 2384 |
+
},
|
| 2385 |
+
{
|
| 2386 |
+
"epoch": 4.53,
|
| 2387 |
+
"learning_rate": 3.503359985066733e-05,
|
| 2388 |
+
"loss": 0.3372,
|
| 2389 |
+
"step": 195000
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 4.54,
|
| 2393 |
+
"learning_rate": 3.499471113461718e-05,
|
| 2394 |
+
"loss": 0.3316,
|
| 2395 |
+
"step": 195500
|
| 2396 |
+
},
|
| 2397 |
+
{
|
| 2398 |
+
"epoch": 4.56,
|
| 2399 |
+
"learning_rate": 3.495582241856703e-05,
|
| 2400 |
+
"loss": 0.3388,
|
| 2401 |
+
"step": 196000
|
| 2402 |
+
},
|
| 2403 |
+
{
|
| 2404 |
+
"epoch": 4.57,
|
| 2405 |
+
"learning_rate": 3.4916933702516876e-05,
|
| 2406 |
+
"loss": 0.3348,
|
| 2407 |
+
"step": 196500
|
| 2408 |
+
},
|
| 2409 |
+
{
|
| 2410 |
+
"epoch": 4.58,
|
| 2411 |
+
"learning_rate": 3.487804498646673e-05,
|
| 2412 |
+
"loss": 0.3382,
|
| 2413 |
+
"step": 197000
|
| 2414 |
+
},
|
| 2415 |
+
{
|
| 2416 |
+
"epoch": 4.59,
|
| 2417 |
+
"learning_rate": 3.483915627041658e-05,
|
| 2418 |
+
"loss": 0.3376,
|
| 2419 |
+
"step": 197500
|
| 2420 |
+
},
|
| 2421 |
+
{
|
| 2422 |
+
"epoch": 4.6,
|
| 2423 |
+
"learning_rate": 3.480026755436643e-05,
|
| 2424 |
+
"loss": 0.3353,
|
| 2425 |
+
"step": 198000
|
| 2426 |
+
},
|
| 2427 |
+
{
|
| 2428 |
+
"epoch": 4.61,
|
| 2429 |
+
"learning_rate": 3.4761378838316276e-05,
|
| 2430 |
+
"loss": 0.3437,
|
| 2431 |
+
"step": 198500
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 4.63,
|
| 2435 |
+
"learning_rate": 3.472249012226613e-05,
|
| 2436 |
+
"loss": 0.339,
|
| 2437 |
+
"step": 199000
|
| 2438 |
+
},
|
| 2439 |
+
{
|
| 2440 |
+
"epoch": 4.64,
|
| 2441 |
+
"learning_rate": 3.468360140621598e-05,
|
| 2442 |
+
"loss": 0.343,
|
| 2443 |
+
"step": 199500
|
| 2444 |
+
},
|
| 2445 |
+
{
|
| 2446 |
+
"epoch": 4.65,
|
| 2447 |
+
"learning_rate": 3.464479046759792e-05,
|
| 2448 |
+
"loss": 0.3402,
|
| 2449 |
+
"step": 200000
|
| 2450 |
+
},
|
| 2451 |
+
{
|
| 2452 |
+
"epoch": 4.66,
|
| 2453 |
+
"learning_rate": 3.460590175154777e-05,
|
| 2454 |
+
"loss": 0.3476,
|
| 2455 |
+
"step": 200500
|
| 2456 |
+
},
|
| 2457 |
+
{
|
| 2458 |
+
"epoch": 4.67,
|
| 2459 |
+
"learning_rate": 3.456701303549762e-05,
|
| 2460 |
+
"loss": 0.3432,
|
| 2461 |
+
"step": 201000
|
| 2462 |
+
},
|
| 2463 |
+
{
|
| 2464 |
+
"epoch": 4.68,
|
| 2465 |
+
"learning_rate": 3.452812431944747e-05,
|
| 2466 |
+
"loss": 0.3435,
|
| 2467 |
+
"step": 201500
|
| 2468 |
+
},
|
| 2469 |
+
{
|
| 2470 |
+
"epoch": 4.7,
|
| 2471 |
+
"learning_rate": 3.448923560339732e-05,
|
| 2472 |
+
"loss": 0.3396,
|
| 2473 |
+
"step": 202000
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 4.71,
|
| 2477 |
+
"learning_rate": 3.445042466477927e-05,
|
| 2478 |
+
"loss": 0.3421,
|
| 2479 |
+
"step": 202500
|
| 2480 |
+
},
|
| 2481 |
+
{
|
| 2482 |
+
"epoch": 4.72,
|
| 2483 |
+
"learning_rate": 3.4411535948729115e-05,
|
| 2484 |
+
"loss": 0.3405,
|
| 2485 |
+
"step": 203000
|
| 2486 |
+
},
|
| 2487 |
+
{
|
| 2488 |
+
"epoch": 4.73,
|
| 2489 |
+
"learning_rate": 3.437264723267897e-05,
|
| 2490 |
+
"loss": 0.3434,
|
| 2491 |
+
"step": 203500
|
| 2492 |
+
},
|
| 2493 |
+
{
|
| 2494 |
+
"epoch": 4.74,
|
| 2495 |
+
"learning_rate": 3.433375851662882e-05,
|
| 2496 |
+
"loss": 0.3465,
|
| 2497 |
+
"step": 204000
|
| 2498 |
+
},
|
| 2499 |
+
{
|
| 2500 |
+
"epoch": 4.75,
|
| 2501 |
+
"learning_rate": 3.429486980057866e-05,
|
| 2502 |
+
"loss": 0.3459,
|
| 2503 |
+
"step": 204500
|
| 2504 |
+
},
|
| 2505 |
+
{
|
| 2506 |
+
"epoch": 4.76,
|
| 2507 |
+
"learning_rate": 3.4255981084528515e-05,
|
| 2508 |
+
"loss": 0.3458,
|
| 2509 |
+
"step": 205000
|
| 2510 |
+
},
|
| 2511 |
+
{
|
| 2512 |
+
"epoch": 4.78,
|
| 2513 |
+
"learning_rate": 3.4217092368478366e-05,
|
| 2514 |
+
"loss": 0.3486,
|
| 2515 |
+
"step": 205500
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 4.79,
|
| 2519 |
+
"learning_rate": 3.4178281429860316e-05,
|
| 2520 |
+
"loss": 0.3491,
|
| 2521 |
+
"step": 206000
|
| 2522 |
+
},
|
| 2523 |
+
{
|
| 2524 |
+
"epoch": 4.8,
|
| 2525 |
+
"learning_rate": 3.4139470491242265e-05,
|
| 2526 |
+
"loss": 0.3509,
|
| 2527 |
+
"step": 206500
|
| 2528 |
+
},
|
| 2529 |
+
{
|
| 2530 |
+
"epoch": 4.81,
|
| 2531 |
+
"learning_rate": 3.410058177519212e-05,
|
| 2532 |
+
"loss": 0.3475,
|
| 2533 |
+
"step": 207000
|
| 2534 |
+
},
|
| 2535 |
+
{
|
| 2536 |
+
"epoch": 4.82,
|
| 2537 |
+
"learning_rate": 3.406169305914196e-05,
|
| 2538 |
+
"loss": 0.3537,
|
| 2539 |
+
"step": 207500
|
| 2540 |
+
},
|
| 2541 |
+
{
|
| 2542 |
+
"epoch": 4.83,
|
| 2543 |
+
"learning_rate": 3.4022804343091806e-05,
|
| 2544 |
+
"loss": 0.3485,
|
| 2545 |
+
"step": 208000
|
| 2546 |
+
},
|
| 2547 |
+
{
|
| 2548 |
+
"epoch": 4.85,
|
| 2549 |
+
"learning_rate": 3.398391562704166e-05,
|
| 2550 |
+
"loss": 0.3402,
|
| 2551 |
+
"step": 208500
|
| 2552 |
+
},
|
| 2553 |
+
{
|
| 2554 |
+
"epoch": 4.86,
|
| 2555 |
+
"learning_rate": 3.394502691099151e-05,
|
| 2556 |
+
"loss": 0.3582,
|
| 2557 |
+
"step": 209000
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 4.87,
|
| 2561 |
+
"learning_rate": 3.390629374980556e-05,
|
| 2562 |
+
"loss": 0.3509,
|
| 2563 |
+
"step": 209500
|
| 2564 |
+
},
|
| 2565 |
+
{
|
| 2566 |
+
"epoch": 4.88,
|
| 2567 |
+
"learning_rate": 3.386740503375541e-05,
|
| 2568 |
+
"loss": 0.3431,
|
| 2569 |
+
"step": 210000
|
| 2570 |
+
},
|
| 2571 |
+
{
|
| 2572 |
+
"epoch": 4.89,
|
| 2573 |
+
"learning_rate": 3.3828516317705254e-05,
|
| 2574 |
+
"loss": 0.3494,
|
| 2575 |
+
"step": 210500
|
| 2576 |
+
},
|
| 2577 |
+
{
|
| 2578 |
+
"epoch": 4.9,
|
| 2579 |
+
"learning_rate": 3.37897053790872e-05,
|
| 2580 |
+
"loss": 0.3514,
|
| 2581 |
+
"step": 211000
|
| 2582 |
+
},
|
| 2583 |
+
{
|
| 2584 |
+
"epoch": 4.92,
|
| 2585 |
+
"learning_rate": 3.3750816663037055e-05,
|
| 2586 |
+
"loss": 0.3555,
|
| 2587 |
+
"step": 211500
|
| 2588 |
+
},
|
| 2589 |
+
{
|
| 2590 |
+
"epoch": 4.93,
|
| 2591 |
+
"learning_rate": 3.3711927946986906e-05,
|
| 2592 |
+
"loss": 0.3513,
|
| 2593 |
+
"step": 212000
|
| 2594 |
+
},
|
| 2595 |
+
{
|
| 2596 |
+
"epoch": 4.94,
|
| 2597 |
+
"learning_rate": 3.367303923093675e-05,
|
| 2598 |
+
"loss": 0.3466,
|
| 2599 |
+
"step": 212500
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 4.95,
|
| 2603 |
+
"learning_rate": 3.36341505148866e-05,
|
| 2604 |
+
"loss": 0.3509,
|
| 2605 |
+
"step": 213000
|
| 2606 |
+
},
|
| 2607 |
+
{
|
| 2608 |
+
"epoch": 4.96,
|
| 2609 |
+
"learning_rate": 3.3595261798836454e-05,
|
| 2610 |
+
"loss": 0.347,
|
| 2611 |
+
"step": 213500
|
| 2612 |
+
},
|
| 2613 |
+
{
|
| 2614 |
+
"epoch": 4.97,
|
| 2615 |
+
"learning_rate": 3.35563730827863e-05,
|
| 2616 |
+
"loss": 0.3511,
|
| 2617 |
+
"step": 214000
|
| 2618 |
+
},
|
| 2619 |
+
{
|
| 2620 |
+
"epoch": 4.99,
|
| 2621 |
+
"learning_rate": 3.351748436673615e-05,
|
| 2622 |
+
"loss": 0.3524,
|
| 2623 |
+
"step": 214500
|
| 2624 |
+
},
|
| 2625 |
+
{
|
| 2626 |
+
"epoch": 5.0,
|
| 2627 |
+
"learning_rate": 3.3478595650686e-05,
|
| 2628 |
+
"loss": 0.3458,
|
| 2629 |
+
"step": 215000
|
| 2630 |
+
},
|
| 2631 |
+
{
|
| 2632 |
+
"epoch": 5.0,
|
| 2633 |
+
"eval_bleu": 59.1341,
|
| 2634 |
+
"eval_gen_len": 15.7016,
|
| 2635 |
+
"eval_loss": 0.7214592099189758,
|
| 2636 |
+
"eval_runtime": 6616.423,
|
| 2637 |
+
"eval_samples_per_second": 12.985,
|
| 2638 |
+
"eval_steps_per_second": 1.623,
|
| 2639 |
+
"step": 215120
|
| 2640 |
+
}
|
| 2641 |
+
],
|
| 2642 |
+
"logging_steps": 500,
|
| 2643 |
+
"max_steps": 645360,
|
| 2644 |
+
"num_input_tokens_seen": 0,
|
| 2645 |
+
"num_train_epochs": 15,
|
| 2646 |
+
"save_steps": 500,
|
| 2647 |
+
"total_flos": 7.459087290854277e+18,
|
| 2648 |
+
"train_batch_size": 2,
|
| 2649 |
+
"trial_name": null,
|
| 2650 |
+
"trial_params": null
|
| 2651 |
+
}
|
checkpoint-215120/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4694a20062ed70ba8adc7c9ea1e09ea766481ac4f0f03605243799bfb9894d0a
|
| 3 |
+
size 4856
|
checkpoint-258144/config.json
ADDED
|
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/mbart-large-50-many-to-many-mmt",
|
| 3 |
+
"_num_labels": 3,
|
| 4 |
+
"activation_dropout": 0.0,
|
| 5 |
+
"activation_function": "relu",
|
| 6 |
+
"add_bias_logits": false,
|
| 7 |
+
"add_final_layer_norm": true,
|
| 8 |
+
"architectures": [
|
| 9 |
+
"MBartForConditionalGeneration"
|
| 10 |
+
],
|
| 11 |
+
"attention_dropout": 0.0,
|
| 12 |
+
"bos_token_id": 0,
|
| 13 |
+
"classif_dropout": 0.0,
|
| 14 |
+
"classifier_dropout": 0.0,
|
| 15 |
+
"d_model": 1024,
|
| 16 |
+
"decoder_attention_heads": 16,
|
| 17 |
+
"decoder_ffn_dim": 4096,
|
| 18 |
+
"decoder_layerdrop": 0.0,
|
| 19 |
+
"decoder_layers": 12,
|
| 20 |
+
"decoder_start_token_id": 2,
|
| 21 |
+
"dropout": 0.1,
|
| 22 |
+
"early_stopping": true,
|
| 23 |
+
"encoder_attention_heads": 16,
|
| 24 |
+
"encoder_ffn_dim": 4096,
|
| 25 |
+
"encoder_layerdrop": 0.0,
|
| 26 |
+
"encoder_layers": 12,
|
| 27 |
+
"eos_token_id": 2,
|
| 28 |
+
"forced_bos_token_id": 250014,
|
| 29 |
+
"forced_eos_token_id": 2,
|
| 30 |
+
"gradient_checkpointing": false,
|
| 31 |
+
"id2label": {
|
| 32 |
+
"0": "LABEL_0",
|
| 33 |
+
"1": "LABEL_1",
|
| 34 |
+
"2": "LABEL_2"
|
| 35 |
+
},
|
| 36 |
+
"init_std": 0.02,
|
| 37 |
+
"is_encoder_decoder": true,
|
| 38 |
+
"label2id": {
|
| 39 |
+
"LABEL_0": 0,
|
| 40 |
+
"LABEL_1": 1,
|
| 41 |
+
"LABEL_2": 2
|
| 42 |
+
},
|
| 43 |
+
"max_length": 200,
|
| 44 |
+
"max_position_embeddings": 1024,
|
| 45 |
+
"model_type": "mbart",
|
| 46 |
+
"normalize_before": true,
|
| 47 |
+
"normalize_embedding": true,
|
| 48 |
+
"num_beams": 5,
|
| 49 |
+
"num_hidden_layers": 12,
|
| 50 |
+
"output_past": true,
|
| 51 |
+
"pad_token_id": 1,
|
| 52 |
+
"scale_embedding": true,
|
| 53 |
+
"static_position_embeddings": false,
|
| 54 |
+
"tokenizer_class": "MBart50Tokenizer",
|
| 55 |
+
"torch_dtype": "float32",
|
| 56 |
+
"transformers_version": "4.36.2",
|
| 57 |
+
"use_cache": true,
|
| 58 |
+
"vocab_size": 250054
|
| 59 |
+
}
|
checkpoint-258144/generation_config.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token_id": 0,
|
| 3 |
+
"decoder_start_token_id": 2,
|
| 4 |
+
"early_stopping": true,
|
| 5 |
+
"eos_token_id": 2,
|
| 6 |
+
"forced_bos_token_id": 250014,
|
| 7 |
+
"forced_eos_token_id": 2,
|
| 8 |
+
"max_length": 200,
|
| 9 |
+
"num_beams": 5,
|
| 10 |
+
"pad_token_id": 1,
|
| 11 |
+
"transformers_version": "4.36.2"
|
| 12 |
+
}
|
checkpoint-258144/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:33f812c633d077962eae5e8f4e11923ea44e5b421e7c705c88213dcc1748d28a
|
| 3 |
+
size 2444578688
|
checkpoint-258144/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6cf51ad82d39a476347fd6537ac15efb85b5b1c300923e78df4c5a8bd9d9af45
|
| 3 |
+
size 4887473903
|
checkpoint-258144/rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:86e5d63ff4c4ebc6a2f03ce0fa660d88d20eead670a445836bc8f7802d8a5df9
|
| 3 |
+
size 15024
|
checkpoint-258144/rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4e051b6f897b77a2f6bfa69e9e6a5d3d195f78fdc41eff6f39a1e4a1e285e0af
|
| 3 |
+
size 15024
|
checkpoint-258144/rng_state_2.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cf2ea6733c0c747b4dccc04ed9bb275b38ba49852ec3c9909e8d8a859309ac5d
|
| 3 |
+
size 15024
|
checkpoint-258144/rng_state_3.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:86b69cb400d5a644d63e72ae8376b86a5ac32aa74e831a3847c4812ca385b7a6
|
| 3 |
+
size 15024
|
checkpoint-258144/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:38ad8b67f4cba7800edbc8c08f7656746fc6c862ec6662bd0f3db409a5e9279f
|
| 3 |
+
size 1064
|
checkpoint-258144/sentencepiece.bpe.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cfc8146abe2a0488e9e2a0c56de7952f7c11ab059eca145a0a727afce0db2865
|
| 3 |
+
size 5069051
|
checkpoint-258144/special_tokens_map.json
ADDED
|
@@ -0,0 +1,69 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"additional_special_tokens": [
|
| 3 |
+
"ar_AR",
|
| 4 |
+
"cs_CZ",
|
| 5 |
+
"de_DE",
|
| 6 |
+
"en_XX",
|
| 7 |
+
"es_XX",
|
| 8 |
+
"et_EE",
|
| 9 |
+
"fi_FI",
|
| 10 |
+
"fr_XX",
|
| 11 |
+
"gu_IN",
|
| 12 |
+
"hi_IN",
|
| 13 |
+
"it_IT",
|
| 14 |
+
"ja_XX",
|
| 15 |
+
"kk_KZ",
|
| 16 |
+
"ko_KR",
|
| 17 |
+
"lt_LT",
|
| 18 |
+
"lv_LV",
|
| 19 |
+
"my_MM",
|
| 20 |
+
"ne_NP",
|
| 21 |
+
"nl_XX",
|
| 22 |
+
"ro_RO",
|
| 23 |
+
"ru_RU",
|
| 24 |
+
"si_LK",
|
| 25 |
+
"tr_TR",
|
| 26 |
+
"vi_VN",
|
| 27 |
+
"zh_CN",
|
| 28 |
+
"af_ZA",
|
| 29 |
+
"az_AZ",
|
| 30 |
+
"bn_IN",
|
| 31 |
+
"fa_IR",
|
| 32 |
+
"he_IL",
|
| 33 |
+
"hr_HR",
|
| 34 |
+
"id_ID",
|
| 35 |
+
"ka_GE",
|
| 36 |
+
"km_KH",
|
| 37 |
+
"mk_MK",
|
| 38 |
+
"ml_IN",
|
| 39 |
+
"mn_MN",
|
| 40 |
+
"mr_IN",
|
| 41 |
+
"pl_PL",
|
| 42 |
+
"ps_AF",
|
| 43 |
+
"pt_XX",
|
| 44 |
+
"sv_SE",
|
| 45 |
+
"sw_KE",
|
| 46 |
+
"ta_IN",
|
| 47 |
+
"te_IN",
|
| 48 |
+
"th_TH",
|
| 49 |
+
"tl_XX",
|
| 50 |
+
"uk_UA",
|
| 51 |
+
"ur_PK",
|
| 52 |
+
"xh_ZA",
|
| 53 |
+
"gl_ES",
|
| 54 |
+
"sl_SI"
|
| 55 |
+
],
|
| 56 |
+
"bos_token": "<s>",
|
| 57 |
+
"cls_token": "<s>",
|
| 58 |
+
"eos_token": "</s>",
|
| 59 |
+
"mask_token": {
|
| 60 |
+
"content": "<mask>",
|
| 61 |
+
"lstrip": true,
|
| 62 |
+
"normalized": true,
|
| 63 |
+
"rstrip": false,
|
| 64 |
+
"single_word": false
|
| 65 |
+
},
|
| 66 |
+
"pad_token": "<pad>",
|
| 67 |
+
"sep_token": "</s>",
|
| 68 |
+
"unk_token": "<unk>"
|
| 69 |
+
}
|
checkpoint-258144/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:faf1fb0f45fdd5ae9fc44fafee492652a423fc749833ef57c2fa801358549d92
|
| 3 |
+
size 17110107
|
checkpoint-258144/tokenizer_config.json
ADDED
|
@@ -0,0 +1,528 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"250001": {
|
| 36 |
+
"content": "ar_AR",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
},
|
| 43 |
+
"250002": {
|
| 44 |
+
"content": "cs_CZ",
|
| 45 |
+
"lstrip": false,
|
| 46 |
+
"normalized": false,
|
| 47 |
+
"rstrip": false,
|
| 48 |
+
"single_word": false,
|
| 49 |
+
"special": true
|
| 50 |
+
},
|
| 51 |
+
"250003": {
|
| 52 |
+
"content": "de_DE",
|
| 53 |
+
"lstrip": false,
|
| 54 |
+
"normalized": false,
|
| 55 |
+
"rstrip": false,
|
| 56 |
+
"single_word": false,
|
| 57 |
+
"special": true
|
| 58 |
+
},
|
| 59 |
+
"250004": {
|
| 60 |
+
"content": "en_XX",
|
| 61 |
+
"lstrip": false,
|
| 62 |
+
"normalized": false,
|
| 63 |
+
"rstrip": false,
|
| 64 |
+
"single_word": false,
|
| 65 |
+
"special": true
|
| 66 |
+
},
|
| 67 |
+
"250005": {
|
| 68 |
+
"content": "es_XX",
|
| 69 |
+
"lstrip": false,
|
| 70 |
+
"normalized": false,
|
| 71 |
+
"rstrip": false,
|
| 72 |
+
"single_word": false,
|
| 73 |
+
"special": true
|
| 74 |
+
},
|
| 75 |
+
"250006": {
|
| 76 |
+
"content": "et_EE",
|
| 77 |
+
"lstrip": false,
|
| 78 |
+
"normalized": false,
|
| 79 |
+
"rstrip": false,
|
| 80 |
+
"single_word": false,
|
| 81 |
+
"special": true
|
| 82 |
+
},
|
| 83 |
+
"250007": {
|
| 84 |
+
"content": "fi_FI",
|
| 85 |
+
"lstrip": false,
|
| 86 |
+
"normalized": false,
|
| 87 |
+
"rstrip": false,
|
| 88 |
+
"single_word": false,
|
| 89 |
+
"special": true
|
| 90 |
+
},
|
| 91 |
+
"250008": {
|
| 92 |
+
"content": "fr_XX",
|
| 93 |
+
"lstrip": false,
|
| 94 |
+
"normalized": false,
|
| 95 |
+
"rstrip": false,
|
| 96 |
+
"single_word": false,
|
| 97 |
+
"special": true
|
| 98 |
+
},
|
| 99 |
+
"250009": {
|
| 100 |
+
"content": "gu_IN",
|
| 101 |
+
"lstrip": false,
|
| 102 |
+
"normalized": false,
|
| 103 |
+
"rstrip": false,
|
| 104 |
+
"single_word": false,
|
| 105 |
+
"special": true
|
| 106 |
+
},
|
| 107 |
+
"250010": {
|
| 108 |
+
"content": "hi_IN",
|
| 109 |
+
"lstrip": false,
|
| 110 |
+
"normalized": false,
|
| 111 |
+
"rstrip": false,
|
| 112 |
+
"single_word": false,
|
| 113 |
+
"special": true
|
| 114 |
+
},
|
| 115 |
+
"250011": {
|
| 116 |
+
"content": "it_IT",
|
| 117 |
+
"lstrip": false,
|
| 118 |
+
"normalized": false,
|
| 119 |
+
"rstrip": false,
|
| 120 |
+
"single_word": false,
|
| 121 |
+
"special": true
|
| 122 |
+
},
|
| 123 |
+
"250012": {
|
| 124 |
+
"content": "ja_XX",
|
| 125 |
+
"lstrip": false,
|
| 126 |
+
"normalized": false,
|
| 127 |
+
"rstrip": false,
|
| 128 |
+
"single_word": false,
|
| 129 |
+
"special": true
|
| 130 |
+
},
|
| 131 |
+
"250013": {
|
| 132 |
+
"content": "kk_KZ",
|
| 133 |
+
"lstrip": false,
|
| 134 |
+
"normalized": false,
|
| 135 |
+
"rstrip": false,
|
| 136 |
+
"single_word": false,
|
| 137 |
+
"special": true
|
| 138 |
+
},
|
| 139 |
+
"250014": {
|
| 140 |
+
"content": "ko_KR",
|
| 141 |
+
"lstrip": false,
|
| 142 |
+
"normalized": false,
|
| 143 |
+
"rstrip": false,
|
| 144 |
+
"single_word": false,
|
| 145 |
+
"special": true
|
| 146 |
+
},
|
| 147 |
+
"250015": {
|
| 148 |
+
"content": "lt_LT",
|
| 149 |
+
"lstrip": false,
|
| 150 |
+
"normalized": false,
|
| 151 |
+
"rstrip": false,
|
| 152 |
+
"single_word": false,
|
| 153 |
+
"special": true
|
| 154 |
+
},
|
| 155 |
+
"250016": {
|
| 156 |
+
"content": "lv_LV",
|
| 157 |
+
"lstrip": false,
|
| 158 |
+
"normalized": false,
|
| 159 |
+
"rstrip": false,
|
| 160 |
+
"single_word": false,
|
| 161 |
+
"special": true
|
| 162 |
+
},
|
| 163 |
+
"250017": {
|
| 164 |
+
"content": "my_MM",
|
| 165 |
+
"lstrip": false,
|
| 166 |
+
"normalized": false,
|
| 167 |
+
"rstrip": false,
|
| 168 |
+
"single_word": false,
|
| 169 |
+
"special": true
|
| 170 |
+
},
|
| 171 |
+
"250018": {
|
| 172 |
+
"content": "ne_NP",
|
| 173 |
+
"lstrip": false,
|
| 174 |
+
"normalized": false,
|
| 175 |
+
"rstrip": false,
|
| 176 |
+
"single_word": false,
|
| 177 |
+
"special": true
|
| 178 |
+
},
|
| 179 |
+
"250019": {
|
| 180 |
+
"content": "nl_XX",
|
| 181 |
+
"lstrip": false,
|
| 182 |
+
"normalized": false,
|
| 183 |
+
"rstrip": false,
|
| 184 |
+
"single_word": false,
|
| 185 |
+
"special": true
|
| 186 |
+
},
|
| 187 |
+
"250020": {
|
| 188 |
+
"content": "ro_RO",
|
| 189 |
+
"lstrip": false,
|
| 190 |
+
"normalized": false,
|
| 191 |
+
"rstrip": false,
|
| 192 |
+
"single_word": false,
|
| 193 |
+
"special": true
|
| 194 |
+
},
|
| 195 |
+
"250021": {
|
| 196 |
+
"content": "ru_RU",
|
| 197 |
+
"lstrip": false,
|
| 198 |
+
"normalized": false,
|
| 199 |
+
"rstrip": false,
|
| 200 |
+
"single_word": false,
|
| 201 |
+
"special": true
|
| 202 |
+
},
|
| 203 |
+
"250022": {
|
| 204 |
+
"content": "si_LK",
|
| 205 |
+
"lstrip": false,
|
| 206 |
+
"normalized": false,
|
| 207 |
+
"rstrip": false,
|
| 208 |
+
"single_word": false,
|
| 209 |
+
"special": true
|
| 210 |
+
},
|
| 211 |
+
"250023": {
|
| 212 |
+
"content": "tr_TR",
|
| 213 |
+
"lstrip": false,
|
| 214 |
+
"normalized": false,
|
| 215 |
+
"rstrip": false,
|
| 216 |
+
"single_word": false,
|
| 217 |
+
"special": true
|
| 218 |
+
},
|
| 219 |
+
"250024": {
|
| 220 |
+
"content": "vi_VN",
|
| 221 |
+
"lstrip": false,
|
| 222 |
+
"normalized": false,
|
| 223 |
+
"rstrip": false,
|
| 224 |
+
"single_word": false,
|
| 225 |
+
"special": true
|
| 226 |
+
},
|
| 227 |
+
"250025": {
|
| 228 |
+
"content": "zh_CN",
|
| 229 |
+
"lstrip": false,
|
| 230 |
+
"normalized": false,
|
| 231 |
+
"rstrip": false,
|
| 232 |
+
"single_word": false,
|
| 233 |
+
"special": true
|
| 234 |
+
},
|
| 235 |
+
"250026": {
|
| 236 |
+
"content": "af_ZA",
|
| 237 |
+
"lstrip": false,
|
| 238 |
+
"normalized": false,
|
| 239 |
+
"rstrip": false,
|
| 240 |
+
"single_word": false,
|
| 241 |
+
"special": true
|
| 242 |
+
},
|
| 243 |
+
"250027": {
|
| 244 |
+
"content": "az_AZ",
|
| 245 |
+
"lstrip": false,
|
| 246 |
+
"normalized": false,
|
| 247 |
+
"rstrip": false,
|
| 248 |
+
"single_word": false,
|
| 249 |
+
"special": true
|
| 250 |
+
},
|
| 251 |
+
"250028": {
|
| 252 |
+
"content": "bn_IN",
|
| 253 |
+
"lstrip": false,
|
| 254 |
+
"normalized": false,
|
| 255 |
+
"rstrip": false,
|
| 256 |
+
"single_word": false,
|
| 257 |
+
"special": true
|
| 258 |
+
},
|
| 259 |
+
"250029": {
|
| 260 |
+
"content": "fa_IR",
|
| 261 |
+
"lstrip": false,
|
| 262 |
+
"normalized": false,
|
| 263 |
+
"rstrip": false,
|
| 264 |
+
"single_word": false,
|
| 265 |
+
"special": true
|
| 266 |
+
},
|
| 267 |
+
"250030": {
|
| 268 |
+
"content": "he_IL",
|
| 269 |
+
"lstrip": false,
|
| 270 |
+
"normalized": false,
|
| 271 |
+
"rstrip": false,
|
| 272 |
+
"single_word": false,
|
| 273 |
+
"special": true
|
| 274 |
+
},
|
| 275 |
+
"250031": {
|
| 276 |
+
"content": "hr_HR",
|
| 277 |
+
"lstrip": false,
|
| 278 |
+
"normalized": false,
|
| 279 |
+
"rstrip": false,
|
| 280 |
+
"single_word": false,
|
| 281 |
+
"special": true
|
| 282 |
+
},
|
| 283 |
+
"250032": {
|
| 284 |
+
"content": "id_ID",
|
| 285 |
+
"lstrip": false,
|
| 286 |
+
"normalized": false,
|
| 287 |
+
"rstrip": false,
|
| 288 |
+
"single_word": false,
|
| 289 |
+
"special": true
|
| 290 |
+
},
|
| 291 |
+
"250033": {
|
| 292 |
+
"content": "ka_GE",
|
| 293 |
+
"lstrip": false,
|
| 294 |
+
"normalized": false,
|
| 295 |
+
"rstrip": false,
|
| 296 |
+
"single_word": false,
|
| 297 |
+
"special": true
|
| 298 |
+
},
|
| 299 |
+
"250034": {
|
| 300 |
+
"content": "km_KH",
|
| 301 |
+
"lstrip": false,
|
| 302 |
+
"normalized": false,
|
| 303 |
+
"rstrip": false,
|
| 304 |
+
"single_word": false,
|
| 305 |
+
"special": true
|
| 306 |
+
},
|
| 307 |
+
"250035": {
|
| 308 |
+
"content": "mk_MK",
|
| 309 |
+
"lstrip": false,
|
| 310 |
+
"normalized": false,
|
| 311 |
+
"rstrip": false,
|
| 312 |
+
"single_word": false,
|
| 313 |
+
"special": true
|
| 314 |
+
},
|
| 315 |
+
"250036": {
|
| 316 |
+
"content": "ml_IN",
|
| 317 |
+
"lstrip": false,
|
| 318 |
+
"normalized": false,
|
| 319 |
+
"rstrip": false,
|
| 320 |
+
"single_word": false,
|
| 321 |
+
"special": true
|
| 322 |
+
},
|
| 323 |
+
"250037": {
|
| 324 |
+
"content": "mn_MN",
|
| 325 |
+
"lstrip": false,
|
| 326 |
+
"normalized": false,
|
| 327 |
+
"rstrip": false,
|
| 328 |
+
"single_word": false,
|
| 329 |
+
"special": true
|
| 330 |
+
},
|
| 331 |
+
"250038": {
|
| 332 |
+
"content": "mr_IN",
|
| 333 |
+
"lstrip": false,
|
| 334 |
+
"normalized": false,
|
| 335 |
+
"rstrip": false,
|
| 336 |
+
"single_word": false,
|
| 337 |
+
"special": true
|
| 338 |
+
},
|
| 339 |
+
"250039": {
|
| 340 |
+
"content": "pl_PL",
|
| 341 |
+
"lstrip": false,
|
| 342 |
+
"normalized": false,
|
| 343 |
+
"rstrip": false,
|
| 344 |
+
"single_word": false,
|
| 345 |
+
"special": true
|
| 346 |
+
},
|
| 347 |
+
"250040": {
|
| 348 |
+
"content": "ps_AF",
|
| 349 |
+
"lstrip": false,
|
| 350 |
+
"normalized": false,
|
| 351 |
+
"rstrip": false,
|
| 352 |
+
"single_word": false,
|
| 353 |
+
"special": true
|
| 354 |
+
},
|
| 355 |
+
"250041": {
|
| 356 |
+
"content": "pt_XX",
|
| 357 |
+
"lstrip": false,
|
| 358 |
+
"normalized": false,
|
| 359 |
+
"rstrip": false,
|
| 360 |
+
"single_word": false,
|
| 361 |
+
"special": true
|
| 362 |
+
},
|
| 363 |
+
"250042": {
|
| 364 |
+
"content": "sv_SE",
|
| 365 |
+
"lstrip": false,
|
| 366 |
+
"normalized": false,
|
| 367 |
+
"rstrip": false,
|
| 368 |
+
"single_word": false,
|
| 369 |
+
"special": true
|
| 370 |
+
},
|
| 371 |
+
"250043": {
|
| 372 |
+
"content": "sw_KE",
|
| 373 |
+
"lstrip": false,
|
| 374 |
+
"normalized": false,
|
| 375 |
+
"rstrip": false,
|
| 376 |
+
"single_word": false,
|
| 377 |
+
"special": true
|
| 378 |
+
},
|
| 379 |
+
"250044": {
|
| 380 |
+
"content": "ta_IN",
|
| 381 |
+
"lstrip": false,
|
| 382 |
+
"normalized": false,
|
| 383 |
+
"rstrip": false,
|
| 384 |
+
"single_word": false,
|
| 385 |
+
"special": true
|
| 386 |
+
},
|
| 387 |
+
"250045": {
|
| 388 |
+
"content": "te_IN",
|
| 389 |
+
"lstrip": false,
|
| 390 |
+
"normalized": false,
|
| 391 |
+
"rstrip": false,
|
| 392 |
+
"single_word": false,
|
| 393 |
+
"special": true
|
| 394 |
+
},
|
| 395 |
+
"250046": {
|
| 396 |
+
"content": "th_TH",
|
| 397 |
+
"lstrip": false,
|
| 398 |
+
"normalized": false,
|
| 399 |
+
"rstrip": false,
|
| 400 |
+
"single_word": false,
|
| 401 |
+
"special": true
|
| 402 |
+
},
|
| 403 |
+
"250047": {
|
| 404 |
+
"content": "tl_XX",
|
| 405 |
+
"lstrip": false,
|
| 406 |
+
"normalized": false,
|
| 407 |
+
"rstrip": false,
|
| 408 |
+
"single_word": false,
|
| 409 |
+
"special": true
|
| 410 |
+
},
|
| 411 |
+
"250048": {
|
| 412 |
+
"content": "uk_UA",
|
| 413 |
+
"lstrip": false,
|
| 414 |
+
"normalized": false,
|
| 415 |
+
"rstrip": false,
|
| 416 |
+
"single_word": false,
|
| 417 |
+
"special": true
|
| 418 |
+
},
|
| 419 |
+
"250049": {
|
| 420 |
+
"content": "ur_PK",
|
| 421 |
+
"lstrip": false,
|
| 422 |
+
"normalized": false,
|
| 423 |
+
"rstrip": false,
|
| 424 |
+
"single_word": false,
|
| 425 |
+
"special": true
|
| 426 |
+
},
|
| 427 |
+
"250050": {
|
| 428 |
+
"content": "xh_ZA",
|
| 429 |
+
"lstrip": false,
|
| 430 |
+
"normalized": false,
|
| 431 |
+
"rstrip": false,
|
| 432 |
+
"single_word": false,
|
| 433 |
+
"special": true
|
| 434 |
+
},
|
| 435 |
+
"250051": {
|
| 436 |
+
"content": "gl_ES",
|
| 437 |
+
"lstrip": false,
|
| 438 |
+
"normalized": false,
|
| 439 |
+
"rstrip": false,
|
| 440 |
+
"single_word": false,
|
| 441 |
+
"special": true
|
| 442 |
+
},
|
| 443 |
+
"250052": {
|
| 444 |
+
"content": "sl_SI",
|
| 445 |
+
"lstrip": false,
|
| 446 |
+
"normalized": false,
|
| 447 |
+
"rstrip": false,
|
| 448 |
+
"single_word": false,
|
| 449 |
+
"special": true
|
| 450 |
+
},
|
| 451 |
+
"250053": {
|
| 452 |
+
"content": "<mask>",
|
| 453 |
+
"lstrip": true,
|
| 454 |
+
"normalized": true,
|
| 455 |
+
"rstrip": false,
|
| 456 |
+
"single_word": false,
|
| 457 |
+
"special": true
|
| 458 |
+
}
|
| 459 |
+
},
|
| 460 |
+
"additional_special_tokens": [
|
| 461 |
+
"ar_AR",
|
| 462 |
+
"cs_CZ",
|
| 463 |
+
"de_DE",
|
| 464 |
+
"en_XX",
|
| 465 |
+
"es_XX",
|
| 466 |
+
"et_EE",
|
| 467 |
+
"fi_FI",
|
| 468 |
+
"fr_XX",
|
| 469 |
+
"gu_IN",
|
| 470 |
+
"hi_IN",
|
| 471 |
+
"it_IT",
|
| 472 |
+
"ja_XX",
|
| 473 |
+
"kk_KZ",
|
| 474 |
+
"ko_KR",
|
| 475 |
+
"lt_LT",
|
| 476 |
+
"lv_LV",
|
| 477 |
+
"my_MM",
|
| 478 |
+
"ne_NP",
|
| 479 |
+
"nl_XX",
|
| 480 |
+
"ro_RO",
|
| 481 |
+
"ru_RU",
|
| 482 |
+
"si_LK",
|
| 483 |
+
"tr_TR",
|
| 484 |
+
"vi_VN",
|
| 485 |
+
"zh_CN",
|
| 486 |
+
"af_ZA",
|
| 487 |
+
"az_AZ",
|
| 488 |
+
"bn_IN",
|
| 489 |
+
"fa_IR",
|
| 490 |
+
"he_IL",
|
| 491 |
+
"hr_HR",
|
| 492 |
+
"id_ID",
|
| 493 |
+
"ka_GE",
|
| 494 |
+
"km_KH",
|
| 495 |
+
"mk_MK",
|
| 496 |
+
"ml_IN",
|
| 497 |
+
"mn_MN",
|
| 498 |
+
"mr_IN",
|
| 499 |
+
"pl_PL",
|
| 500 |
+
"ps_AF",
|
| 501 |
+
"pt_XX",
|
| 502 |
+
"sv_SE",
|
| 503 |
+
"sw_KE",
|
| 504 |
+
"ta_IN",
|
| 505 |
+
"te_IN",
|
| 506 |
+
"th_TH",
|
| 507 |
+
"tl_XX",
|
| 508 |
+
"uk_UA",
|
| 509 |
+
"ur_PK",
|
| 510 |
+
"xh_ZA",
|
| 511 |
+
"gl_ES",
|
| 512 |
+
"sl_SI"
|
| 513 |
+
],
|
| 514 |
+
"bos_token": "<s>",
|
| 515 |
+
"clean_up_tokenization_spaces": true,
|
| 516 |
+
"cls_token": "<s>",
|
| 517 |
+
"eos_token": "</s>",
|
| 518 |
+
"language_codes": "ML50",
|
| 519 |
+
"mask_token": "<mask>",
|
| 520 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 521 |
+
"pad_token": "<pad>",
|
| 522 |
+
"sep_token": "</s>",
|
| 523 |
+
"sp_model_kwargs": {},
|
| 524 |
+
"src_lang": "ja_XX",
|
| 525 |
+
"tgt_lang": "ko_KR",
|
| 526 |
+
"tokenizer_class": "MBart50Tokenizer",
|
| 527 |
+
"unk_token": "<unk>"
|
| 528 |
+
}
|
checkpoint-258144/trainer_state.json
ADDED
|
@@ -0,0 +1,3177 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.6568061113357544,
|
| 3 |
+
"best_model_checkpoint": "./finetuned_model/checkpoint-129072",
|
| 4 |
+
"epoch": 6.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 258144,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.01,
|
| 13 |
+
"learning_rate": 9.96e-06,
|
| 14 |
+
"loss": 2.2631,
|
| 15 |
+
"step": 500
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.02,
|
| 19 |
+
"learning_rate": 1.992e-05,
|
| 20 |
+
"loss": 1.5976,
|
| 21 |
+
"step": 1000
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.03,
|
| 25 |
+
"learning_rate": 2.9920000000000005e-05,
|
| 26 |
+
"loss": 1.4641,
|
| 27 |
+
"step": 1500
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.05,
|
| 31 |
+
"learning_rate": 3.9920000000000004e-05,
|
| 32 |
+
"loss": 1.3868,
|
| 33 |
+
"step": 2000
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.06,
|
| 37 |
+
"learning_rate": 4.992e-05,
|
| 38 |
+
"loss": 1.344,
|
| 39 |
+
"step": 2500
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.07,
|
| 43 |
+
"learning_rate": 4.996142239367825e-05,
|
| 44 |
+
"loss": 1.3138,
|
| 45 |
+
"step": 3000
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.08,
|
| 49 |
+
"learning_rate": 4.9922533677628105e-05,
|
| 50 |
+
"loss": 1.244,
|
| 51 |
+
"step": 3500
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.09,
|
| 55 |
+
"learning_rate": 4.988364496157795e-05,
|
| 56 |
+
"loss": 1.2047,
|
| 57 |
+
"step": 4000
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.1,
|
| 61 |
+
"learning_rate": 4.98447562455278e-05,
|
| 62 |
+
"loss": 1.1707,
|
| 63 |
+
"step": 4500
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.12,
|
| 67 |
+
"learning_rate": 4.980586752947765e-05,
|
| 68 |
+
"loss": 1.1331,
|
| 69 |
+
"step": 5000
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.13,
|
| 73 |
+
"learning_rate": 4.97670565908596e-05,
|
| 74 |
+
"loss": 1.1021,
|
| 75 |
+
"step": 5500
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.14,
|
| 79 |
+
"learning_rate": 4.972816787480945e-05,
|
| 80 |
+
"loss": 1.0923,
|
| 81 |
+
"step": 6000
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.15,
|
| 85 |
+
"learning_rate": 4.968927915875929e-05,
|
| 86 |
+
"loss": 1.0713,
|
| 87 |
+
"step": 6500
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.16,
|
| 91 |
+
"learning_rate": 4.965039044270914e-05,
|
| 92 |
+
"loss": 1.0609,
|
| 93 |
+
"step": 7000
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.17,
|
| 97 |
+
"learning_rate": 4.9611501726658995e-05,
|
| 98 |
+
"loss": 1.043,
|
| 99 |
+
"step": 7500
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.19,
|
| 103 |
+
"learning_rate": 4.957261301060884e-05,
|
| 104 |
+
"loss": 1.0345,
|
| 105 |
+
"step": 8000
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.2,
|
| 109 |
+
"learning_rate": 4.953380207199079e-05,
|
| 110 |
+
"loss": 1.02,
|
| 111 |
+
"step": 8500
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.21,
|
| 115 |
+
"learning_rate": 4.949491335594064e-05,
|
| 116 |
+
"loss": 1.0057,
|
| 117 |
+
"step": 9000
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.22,
|
| 121 |
+
"learning_rate": 4.945602463989049e-05,
|
| 122 |
+
"loss": 0.9978,
|
| 123 |
+
"step": 9500
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.23,
|
| 127 |
+
"learning_rate": 4.9417135923840344e-05,
|
| 128 |
+
"loss": 0.9801,
|
| 129 |
+
"step": 10000
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.24,
|
| 133 |
+
"learning_rate": 4.937824720779019e-05,
|
| 134 |
+
"loss": 0.9752,
|
| 135 |
+
"step": 10500
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.26,
|
| 139 |
+
"learning_rate": 4.933935849174004e-05,
|
| 140 |
+
"loss": 0.9736,
|
| 141 |
+
"step": 11000
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.27,
|
| 145 |
+
"learning_rate": 4.930046977568989e-05,
|
| 146 |
+
"loss": 0.9519,
|
| 147 |
+
"step": 11500
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.28,
|
| 151 |
+
"learning_rate": 4.9261581059639736e-05,
|
| 152 |
+
"loss": 0.9559,
|
| 153 |
+
"step": 12000
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.29,
|
| 157 |
+
"learning_rate": 4.9222770121021686e-05,
|
| 158 |
+
"loss": 0.936,
|
| 159 |
+
"step": 12500
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.3,
|
| 163 |
+
"learning_rate": 4.9183959182403636e-05,
|
| 164 |
+
"loss": 0.9276,
|
| 165 |
+
"step": 13000
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.31,
|
| 169 |
+
"learning_rate": 4.914507046635349e-05,
|
| 170 |
+
"loss": 0.9384,
|
| 171 |
+
"step": 13500
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.33,
|
| 175 |
+
"learning_rate": 4.910618175030333e-05,
|
| 176 |
+
"loss": 0.9347,
|
| 177 |
+
"step": 14000
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.34,
|
| 181 |
+
"learning_rate": 4.9067293034253183e-05,
|
| 182 |
+
"loss": 0.9182,
|
| 183 |
+
"step": 14500
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.35,
|
| 187 |
+
"learning_rate": 4.902840431820303e-05,
|
| 188 |
+
"loss": 0.9136,
|
| 189 |
+
"step": 15000
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.36,
|
| 193 |
+
"learning_rate": 4.898951560215288e-05,
|
| 194 |
+
"loss": 0.9096,
|
| 195 |
+
"step": 15500
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.37,
|
| 199 |
+
"learning_rate": 4.895062688610273e-05,
|
| 200 |
+
"loss": 0.8926,
|
| 201 |
+
"step": 16000
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.38,
|
| 205 |
+
"learning_rate": 4.891181594748468e-05,
|
| 206 |
+
"loss": 0.8946,
|
| 207 |
+
"step": 16500
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.4,
|
| 211 |
+
"learning_rate": 4.887300500886663e-05,
|
| 212 |
+
"loss": 0.8895,
|
| 213 |
+
"step": 17000
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.41,
|
| 217 |
+
"learning_rate": 4.8834116292816475e-05,
|
| 218 |
+
"loss": 0.8897,
|
| 219 |
+
"step": 17500
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.42,
|
| 223 |
+
"learning_rate": 4.879522757676633e-05,
|
| 224 |
+
"loss": 0.8901,
|
| 225 |
+
"step": 18000
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.43,
|
| 229 |
+
"learning_rate": 4.875633886071618e-05,
|
| 230 |
+
"loss": 0.8839,
|
| 231 |
+
"step": 18500
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 0.44,
|
| 235 |
+
"learning_rate": 4.871745014466603e-05,
|
| 236 |
+
"loss": 0.8737,
|
| 237 |
+
"step": 19000
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.45,
|
| 241 |
+
"learning_rate": 4.867863920604798e-05,
|
| 242 |
+
"loss": 0.8613,
|
| 243 |
+
"step": 19500
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.46,
|
| 247 |
+
"learning_rate": 4.8639750489997824e-05,
|
| 248 |
+
"loss": 0.8547,
|
| 249 |
+
"step": 20000
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.48,
|
| 253 |
+
"learning_rate": 4.8600861773947676e-05,
|
| 254 |
+
"loss": 0.8623,
|
| 255 |
+
"step": 20500
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.49,
|
| 259 |
+
"learning_rate": 4.856197305789753e-05,
|
| 260 |
+
"loss": 0.8615,
|
| 261 |
+
"step": 21000
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.5,
|
| 265 |
+
"learning_rate": 4.852308434184737e-05,
|
| 266 |
+
"loss": 0.8614,
|
| 267 |
+
"step": 21500
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.51,
|
| 271 |
+
"learning_rate": 4.848419562579722e-05,
|
| 272 |
+
"loss": 0.846,
|
| 273 |
+
"step": 22000
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 0.52,
|
| 277 |
+
"learning_rate": 4.8445384687179167e-05,
|
| 278 |
+
"loss": 0.8487,
|
| 279 |
+
"step": 22500
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.53,
|
| 283 |
+
"learning_rate": 4.840649597112902e-05,
|
| 284 |
+
"loss": 0.8426,
|
| 285 |
+
"step": 23000
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.55,
|
| 289 |
+
"learning_rate": 4.836760725507887e-05,
|
| 290 |
+
"loss": 0.8398,
|
| 291 |
+
"step": 23500
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.56,
|
| 295 |
+
"learning_rate": 4.8328718539028714e-05,
|
| 296 |
+
"loss": 0.8407,
|
| 297 |
+
"step": 24000
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.57,
|
| 301 |
+
"learning_rate": 4.828998537784277e-05,
|
| 302 |
+
"loss": 0.8438,
|
| 303 |
+
"step": 24500
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.58,
|
| 307 |
+
"learning_rate": 4.825117443922472e-05,
|
| 308 |
+
"loss": 0.8257,
|
| 309 |
+
"step": 25000
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.59,
|
| 313 |
+
"learning_rate": 4.821228572317456e-05,
|
| 314 |
+
"loss": 0.8328,
|
| 315 |
+
"step": 25500
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 0.6,
|
| 319 |
+
"learning_rate": 4.8173397007124415e-05,
|
| 320 |
+
"loss": 0.8229,
|
| 321 |
+
"step": 26000
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 0.62,
|
| 325 |
+
"learning_rate": 4.8134508291074266e-05,
|
| 326 |
+
"loss": 0.8249,
|
| 327 |
+
"step": 26500
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 0.63,
|
| 331 |
+
"learning_rate": 4.809561957502412e-05,
|
| 332 |
+
"loss": 0.8232,
|
| 333 |
+
"step": 27000
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 0.64,
|
| 337 |
+
"learning_rate": 4.805673085897396e-05,
|
| 338 |
+
"loss": 0.83,
|
| 339 |
+
"step": 27500
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.65,
|
| 343 |
+
"learning_rate": 4.801791992035591e-05,
|
| 344 |
+
"loss": 0.8339,
|
| 345 |
+
"step": 28000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.66,
|
| 349 |
+
"learning_rate": 4.7979031204305764e-05,
|
| 350 |
+
"loss": 0.817,
|
| 351 |
+
"step": 28500
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 0.67,
|
| 355 |
+
"learning_rate": 4.7940142488255615e-05,
|
| 356 |
+
"loss": 0.8223,
|
| 357 |
+
"step": 29000
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 0.69,
|
| 361 |
+
"learning_rate": 4.790125377220545e-05,
|
| 362 |
+
"loss": 0.8041,
|
| 363 |
+
"step": 29500
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 0.7,
|
| 367 |
+
"learning_rate": 4.7862365056155305e-05,
|
| 368 |
+
"loss": 0.8126,
|
| 369 |
+
"step": 30000
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.71,
|
| 373 |
+
"learning_rate": 4.7823476340105156e-05,
|
| 374 |
+
"loss": 0.8039,
|
| 375 |
+
"step": 30500
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 0.72,
|
| 379 |
+
"learning_rate": 4.778458762405501e-05,
|
| 380 |
+
"loss": 0.8071,
|
| 381 |
+
"step": 31000
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.73,
|
| 385 |
+
"learning_rate": 4.774569890800485e-05,
|
| 386 |
+
"loss": 0.8065,
|
| 387 |
+
"step": 31500
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.74,
|
| 391 |
+
"learning_rate": 4.7706810191954704e-05,
|
| 392 |
+
"loss": 0.7998,
|
| 393 |
+
"step": 32000
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 0.76,
|
| 397 |
+
"learning_rate": 4.7667921475904556e-05,
|
| 398 |
+
"loss": 0.8034,
|
| 399 |
+
"step": 32500
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 0.77,
|
| 403 |
+
"learning_rate": 4.76290327598544e-05,
|
| 404 |
+
"loss": 0.8031,
|
| 405 |
+
"step": 33000
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 0.78,
|
| 409 |
+
"learning_rate": 4.759014404380425e-05,
|
| 410 |
+
"loss": 0.8057,
|
| 411 |
+
"step": 33500
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 0.79,
|
| 415 |
+
"learning_rate": 4.7551255327754103e-05,
|
| 416 |
+
"loss": 0.7905,
|
| 417 |
+
"step": 34000
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 0.8,
|
| 421 |
+
"learning_rate": 4.751236661170395e-05,
|
| 422 |
+
"loss": 0.794,
|
| 423 |
+
"step": 34500
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.81,
|
| 427 |
+
"learning_rate": 4.74734778956538e-05,
|
| 428 |
+
"loss": 0.7906,
|
| 429 |
+
"step": 35000
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.83,
|
| 433 |
+
"learning_rate": 4.743458917960365e-05,
|
| 434 |
+
"loss": 0.7865,
|
| 435 |
+
"step": 35500
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 0.84,
|
| 439 |
+
"learning_rate": 4.73957004635535e-05,
|
| 440 |
+
"loss": 0.7644,
|
| 441 |
+
"step": 36000
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 0.85,
|
| 445 |
+
"learning_rate": 4.735681174750335e-05,
|
| 446 |
+
"loss": 0.7737,
|
| 447 |
+
"step": 36500
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 0.86,
|
| 451 |
+
"learning_rate": 4.73179230314532e-05,
|
| 452 |
+
"loss": 0.7746,
|
| 453 |
+
"step": 37000
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 0.87,
|
| 457 |
+
"learning_rate": 4.7279034315403044e-05,
|
| 458 |
+
"loss": 0.7832,
|
| 459 |
+
"step": 37500
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.88,
|
| 463 |
+
"learning_rate": 4.724014559935289e-05,
|
| 464 |
+
"loss": 0.7706,
|
| 465 |
+
"step": 38000
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.89,
|
| 469 |
+
"learning_rate": 4.720141243816694e-05,
|
| 470 |
+
"loss": 0.7778,
|
| 471 |
+
"step": 38500
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.91,
|
| 475 |
+
"learning_rate": 4.7162523722116795e-05,
|
| 476 |
+
"loss": 0.7713,
|
| 477 |
+
"step": 39000
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 0.92,
|
| 481 |
+
"learning_rate": 4.712363500606664e-05,
|
| 482 |
+
"loss": 0.772,
|
| 483 |
+
"step": 39500
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 0.93,
|
| 487 |
+
"learning_rate": 4.708474629001649e-05,
|
| 488 |
+
"loss": 0.775,
|
| 489 |
+
"step": 40000
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 0.94,
|
| 493 |
+
"learning_rate": 4.704593535139844e-05,
|
| 494 |
+
"loss": 0.7718,
|
| 495 |
+
"step": 40500
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 0.95,
|
| 499 |
+
"learning_rate": 4.700704663534829e-05,
|
| 500 |
+
"loss": 0.7724,
|
| 501 |
+
"step": 41000
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 0.96,
|
| 505 |
+
"learning_rate": 4.696815791929814e-05,
|
| 506 |
+
"loss": 0.7629,
|
| 507 |
+
"step": 41500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.98,
|
| 511 |
+
"learning_rate": 4.692926920324799e-05,
|
| 512 |
+
"loss": 0.7731,
|
| 513 |
+
"step": 42000
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.99,
|
| 517 |
+
"learning_rate": 4.689038048719784e-05,
|
| 518 |
+
"loss": 0.7597,
|
| 519 |
+
"step": 42500
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 1.0,
|
| 523 |
+
"learning_rate": 4.6851491771147685e-05,
|
| 524 |
+
"loss": 0.7629,
|
| 525 |
+
"step": 43000
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 1.0,
|
| 529 |
+
"eval_bleu": 54.9608,
|
| 530 |
+
"eval_gen_len": 16.1273,
|
| 531 |
+
"eval_loss": 0.7230631709098816,
|
| 532 |
+
"eval_runtime": 7044.4442,
|
| 533 |
+
"eval_samples_per_second": 12.196,
|
| 534 |
+
"eval_steps_per_second": 1.525,
|
| 535 |
+
"step": 43024
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 1.01,
|
| 539 |
+
"learning_rate": 4.6812680832529634e-05,
|
| 540 |
+
"loss": 0.6436,
|
| 541 |
+
"step": 43500
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 1.02,
|
| 545 |
+
"learning_rate": 4.6773869893911584e-05,
|
| 546 |
+
"loss": 0.6446,
|
| 547 |
+
"step": 44000
|
| 548 |
+
},
|
| 549 |
+
{
|
| 550 |
+
"epoch": 1.03,
|
| 551 |
+
"learning_rate": 4.6734981177861435e-05,
|
| 552 |
+
"loss": 0.6489,
|
| 553 |
+
"step": 44500
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 1.05,
|
| 557 |
+
"learning_rate": 4.669617023924338e-05,
|
| 558 |
+
"loss": 0.6512,
|
| 559 |
+
"step": 45000
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 1.06,
|
| 563 |
+
"learning_rate": 4.665728152319323e-05,
|
| 564 |
+
"loss": 0.648,
|
| 565 |
+
"step": 45500
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"epoch": 1.07,
|
| 569 |
+
"learning_rate": 4.661839280714308e-05,
|
| 570 |
+
"loss": 0.6426,
|
| 571 |
+
"step": 46000
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 1.08,
|
| 575 |
+
"learning_rate": 4.657958186852503e-05,
|
| 576 |
+
"loss": 0.6489,
|
| 577 |
+
"step": 46500
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 1.09,
|
| 581 |
+
"learning_rate": 4.6540693152474876e-05,
|
| 582 |
+
"loss": 0.6417,
|
| 583 |
+
"step": 47000
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 1.1,
|
| 587 |
+
"learning_rate": 4.650180443642473e-05,
|
| 588 |
+
"loss": 0.645,
|
| 589 |
+
"step": 47500
|
| 590 |
+
},
|
| 591 |
+
{
|
| 592 |
+
"epoch": 1.12,
|
| 593 |
+
"learning_rate": 4.646291572037458e-05,
|
| 594 |
+
"loss": 0.6543,
|
| 595 |
+
"step": 48000
|
| 596 |
+
},
|
| 597 |
+
{
|
| 598 |
+
"epoch": 1.13,
|
| 599 |
+
"learning_rate": 4.642402700432443e-05,
|
| 600 |
+
"loss": 0.6555,
|
| 601 |
+
"step": 48500
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 1.14,
|
| 605 |
+
"learning_rate": 4.6385138288274275e-05,
|
| 606 |
+
"loss": 0.6501,
|
| 607 |
+
"step": 49000
|
| 608 |
+
},
|
| 609 |
+
{
|
| 610 |
+
"epoch": 1.15,
|
| 611 |
+
"learning_rate": 4.634624957222413e-05,
|
| 612 |
+
"loss": 0.6447,
|
| 613 |
+
"step": 49500
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 1.16,
|
| 617 |
+
"learning_rate": 4.630736085617398e-05,
|
| 618 |
+
"loss": 0.6432,
|
| 619 |
+
"step": 50000
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 1.17,
|
| 623 |
+
"learning_rate": 4.626847214012382e-05,
|
| 624 |
+
"loss": 0.6516,
|
| 625 |
+
"step": 50500
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 1.19,
|
| 629 |
+
"learning_rate": 4.6229583424073674e-05,
|
| 630 |
+
"loss": 0.6517,
|
| 631 |
+
"step": 51000
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 1.2,
|
| 635 |
+
"learning_rate": 4.6190694708023526e-05,
|
| 636 |
+
"loss": 0.6534,
|
| 637 |
+
"step": 51500
|
| 638 |
+
},
|
| 639 |
+
{
|
| 640 |
+
"epoch": 1.21,
|
| 641 |
+
"learning_rate": 4.615180599197337e-05,
|
| 642 |
+
"loss": 0.6482,
|
| 643 |
+
"step": 52000
|
| 644 |
+
},
|
| 645 |
+
{
|
| 646 |
+
"epoch": 1.22,
|
| 647 |
+
"learning_rate": 4.611291727592322e-05,
|
| 648 |
+
"loss": 0.6504,
|
| 649 |
+
"step": 52500
|
| 650 |
+
},
|
| 651 |
+
{
|
| 652 |
+
"epoch": 1.23,
|
| 653 |
+
"learning_rate": 4.607402855987307e-05,
|
| 654 |
+
"loss": 0.6481,
|
| 655 |
+
"step": 53000
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 1.24,
|
| 659 |
+
"learning_rate": 4.603513984382292e-05,
|
| 660 |
+
"loss": 0.6444,
|
| 661 |
+
"step": 53500
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 1.26,
|
| 665 |
+
"learning_rate": 4.599625112777276e-05,
|
| 666 |
+
"loss": 0.6509,
|
| 667 |
+
"step": 54000
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 1.27,
|
| 671 |
+
"learning_rate": 4.5957362411722615e-05,
|
| 672 |
+
"loss": 0.6485,
|
| 673 |
+
"step": 54500
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 1.28,
|
| 677 |
+
"learning_rate": 4.5918551473104564e-05,
|
| 678 |
+
"loss": 0.6474,
|
| 679 |
+
"step": 55000
|
| 680 |
+
},
|
| 681 |
+
{
|
| 682 |
+
"epoch": 1.29,
|
| 683 |
+
"learning_rate": 4.5879662757054416e-05,
|
| 684 |
+
"loss": 0.6467,
|
| 685 |
+
"step": 55500
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 1.3,
|
| 689 |
+
"learning_rate": 4.584077404100427e-05,
|
| 690 |
+
"loss": 0.6442,
|
| 691 |
+
"step": 56000
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 1.31,
|
| 695 |
+
"learning_rate": 4.580188532495411e-05,
|
| 696 |
+
"loss": 0.6537,
|
| 697 |
+
"step": 56500
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"epoch": 1.32,
|
| 701 |
+
"learning_rate": 4.576307438633606e-05,
|
| 702 |
+
"loss": 0.6563,
|
| 703 |
+
"step": 57000
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 1.34,
|
| 707 |
+
"learning_rate": 4.5724185670285913e-05,
|
| 708 |
+
"loss": 0.6587,
|
| 709 |
+
"step": 57500
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 1.35,
|
| 713 |
+
"learning_rate": 4.5685296954235765e-05,
|
| 714 |
+
"loss": 0.6535,
|
| 715 |
+
"step": 58000
|
| 716 |
+
},
|
| 717 |
+
{
|
| 718 |
+
"epoch": 1.36,
|
| 719 |
+
"learning_rate": 4.564640823818561e-05,
|
| 720 |
+
"loss": 0.6623,
|
| 721 |
+
"step": 58500
|
| 722 |
+
},
|
| 723 |
+
{
|
| 724 |
+
"epoch": 1.37,
|
| 725 |
+
"learning_rate": 4.560759729956756e-05,
|
| 726 |
+
"loss": 0.6458,
|
| 727 |
+
"step": 59000
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"epoch": 1.38,
|
| 731 |
+
"learning_rate": 4.556870858351741e-05,
|
| 732 |
+
"loss": 0.6557,
|
| 733 |
+
"step": 59500
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 1.39,
|
| 737 |
+
"learning_rate": 4.552981986746726e-05,
|
| 738 |
+
"loss": 0.6458,
|
| 739 |
+
"step": 60000
|
| 740 |
+
},
|
| 741 |
+
{
|
| 742 |
+
"epoch": 1.41,
|
| 743 |
+
"learning_rate": 4.549093115141711e-05,
|
| 744 |
+
"loss": 0.6565,
|
| 745 |
+
"step": 60500
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 1.42,
|
| 749 |
+
"learning_rate": 4.545204243536695e-05,
|
| 750 |
+
"loss": 0.6521,
|
| 751 |
+
"step": 61000
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 1.43,
|
| 755 |
+
"learning_rate": 4.54131537193168e-05,
|
| 756 |
+
"loss": 0.6485,
|
| 757 |
+
"step": 61500
|
| 758 |
+
},
|
| 759 |
+
{
|
| 760 |
+
"epoch": 1.44,
|
| 761 |
+
"learning_rate": 4.5374265003266655e-05,
|
| 762 |
+
"loss": 0.6503,
|
| 763 |
+
"step": 62000
|
| 764 |
+
},
|
| 765 |
+
{
|
| 766 |
+
"epoch": 1.45,
|
| 767 |
+
"learning_rate": 4.53353762872165e-05,
|
| 768 |
+
"loss": 0.6483,
|
| 769 |
+
"step": 62500
|
| 770 |
+
},
|
| 771 |
+
{
|
| 772 |
+
"epoch": 1.46,
|
| 773 |
+
"learning_rate": 4.529656534859845e-05,
|
| 774 |
+
"loss": 0.6425,
|
| 775 |
+
"step": 63000
|
| 776 |
+
},
|
| 777 |
+
{
|
| 778 |
+
"epoch": 1.48,
|
| 779 |
+
"learning_rate": 4.52576766325483e-05,
|
| 780 |
+
"loss": 0.6566,
|
| 781 |
+
"step": 63500
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 1.49,
|
| 785 |
+
"learning_rate": 4.521878791649815e-05,
|
| 786 |
+
"loss": 0.6537,
|
| 787 |
+
"step": 64000
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 1.5,
|
| 791 |
+
"learning_rate": 4.5179899200448e-05,
|
| 792 |
+
"loss": 0.653,
|
| 793 |
+
"step": 64500
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 1.51,
|
| 797 |
+
"learning_rate": 4.514116603926205e-05,
|
| 798 |
+
"loss": 0.6448,
|
| 799 |
+
"step": 65000
|
| 800 |
+
},
|
| 801 |
+
{
|
| 802 |
+
"epoch": 1.52,
|
| 803 |
+
"learning_rate": 4.51022773232119e-05,
|
| 804 |
+
"loss": 0.6594,
|
| 805 |
+
"step": 65500
|
| 806 |
+
},
|
| 807 |
+
{
|
| 808 |
+
"epoch": 1.53,
|
| 809 |
+
"learning_rate": 4.506338860716175e-05,
|
| 810 |
+
"loss": 0.6551,
|
| 811 |
+
"step": 66000
|
| 812 |
+
},
|
| 813 |
+
{
|
| 814 |
+
"epoch": 1.55,
|
| 815 |
+
"learning_rate": 4.50245776685437e-05,
|
| 816 |
+
"loss": 0.6559,
|
| 817 |
+
"step": 66500
|
| 818 |
+
},
|
| 819 |
+
{
|
| 820 |
+
"epoch": 1.56,
|
| 821 |
+
"learning_rate": 4.498568895249355e-05,
|
| 822 |
+
"loss": 0.6506,
|
| 823 |
+
"step": 67000
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 1.57,
|
| 827 |
+
"learning_rate": 4.49468002364434e-05,
|
| 828 |
+
"loss": 0.6504,
|
| 829 |
+
"step": 67500
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"epoch": 1.58,
|
| 833 |
+
"learning_rate": 4.4907911520393245e-05,
|
| 834 |
+
"loss": 0.6643,
|
| 835 |
+
"step": 68000
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 1.59,
|
| 839 |
+
"learning_rate": 4.486902280434309e-05,
|
| 840 |
+
"loss": 0.6435,
|
| 841 |
+
"step": 68500
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"epoch": 1.6,
|
| 845 |
+
"learning_rate": 4.483013408829294e-05,
|
| 846 |
+
"loss": 0.6389,
|
| 847 |
+
"step": 69000
|
| 848 |
+
},
|
| 849 |
+
{
|
| 850 |
+
"epoch": 1.62,
|
| 851 |
+
"learning_rate": 4.479124537224279e-05,
|
| 852 |
+
"loss": 0.6495,
|
| 853 |
+
"step": 69500
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 1.63,
|
| 857 |
+
"learning_rate": 4.475235665619264e-05,
|
| 858 |
+
"loss": 0.6395,
|
| 859 |
+
"step": 70000
|
| 860 |
+
},
|
| 861 |
+
{
|
| 862 |
+
"epoch": 1.64,
|
| 863 |
+
"learning_rate": 4.471346794014249e-05,
|
| 864 |
+
"loss": 0.6404,
|
| 865 |
+
"step": 70500
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 1.65,
|
| 869 |
+
"learning_rate": 4.467457922409234e-05,
|
| 870 |
+
"loss": 0.6438,
|
| 871 |
+
"step": 71000
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 1.66,
|
| 875 |
+
"learning_rate": 4.4635690508042186e-05,
|
| 876 |
+
"loss": 0.6361,
|
| 877 |
+
"step": 71500
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 1.67,
|
| 881 |
+
"learning_rate": 4.459680179199204e-05,
|
| 882 |
+
"loss": 0.6486,
|
| 883 |
+
"step": 72000
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 1.69,
|
| 887 |
+
"learning_rate": 4.455791307594189e-05,
|
| 888 |
+
"loss": 0.6489,
|
| 889 |
+
"step": 72500
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"epoch": 1.7,
|
| 893 |
+
"learning_rate": 4.4519024359891734e-05,
|
| 894 |
+
"loss": 0.6532,
|
| 895 |
+
"step": 73000
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"epoch": 1.71,
|
| 899 |
+
"learning_rate": 4.4480135643841585e-05,
|
| 900 |
+
"loss": 0.6529,
|
| 901 |
+
"step": 73500
|
| 902 |
+
},
|
| 903 |
+
{
|
| 904 |
+
"epoch": 1.72,
|
| 905 |
+
"learning_rate": 4.4441246927791437e-05,
|
| 906 |
+
"loss": 0.6451,
|
| 907 |
+
"step": 74000
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 1.73,
|
| 911 |
+
"learning_rate": 4.440235821174129e-05,
|
| 912 |
+
"loss": 0.6443,
|
| 913 |
+
"step": 74500
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 1.74,
|
| 917 |
+
"learning_rate": 4.436346949569113e-05,
|
| 918 |
+
"loss": 0.6456,
|
| 919 |
+
"step": 75000
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 1.75,
|
| 923 |
+
"learning_rate": 4.4324580779640984e-05,
|
| 924 |
+
"loss": 0.6515,
|
| 925 |
+
"step": 75500
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 1.77,
|
| 929 |
+
"learning_rate": 4.428569206359083e-05,
|
| 930 |
+
"loss": 0.6425,
|
| 931 |
+
"step": 76000
|
| 932 |
+
},
|
| 933 |
+
{
|
| 934 |
+
"epoch": 1.78,
|
| 935 |
+
"learning_rate": 4.424680334754068e-05,
|
| 936 |
+
"loss": 0.6394,
|
| 937 |
+
"step": 76500
|
| 938 |
+
},
|
| 939 |
+
{
|
| 940 |
+
"epoch": 1.79,
|
| 941 |
+
"learning_rate": 4.4207914631490525e-05,
|
| 942 |
+
"loss": 0.6433,
|
| 943 |
+
"step": 77000
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 1.8,
|
| 947 |
+
"learning_rate": 4.416902591544038e-05,
|
| 948 |
+
"loss": 0.6459,
|
| 949 |
+
"step": 77500
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 1.81,
|
| 953 |
+
"learning_rate": 4.4130214976822327e-05,
|
| 954 |
+
"loss": 0.6421,
|
| 955 |
+
"step": 78000
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 1.82,
|
| 959 |
+
"learning_rate": 4.409132626077218e-05,
|
| 960 |
+
"loss": 0.6428,
|
| 961 |
+
"step": 78500
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 1.84,
|
| 965 |
+
"learning_rate": 4.405251532215413e-05,
|
| 966 |
+
"loss": 0.6406,
|
| 967 |
+
"step": 79000
|
| 968 |
+
},
|
| 969 |
+
{
|
| 970 |
+
"epoch": 1.85,
|
| 971 |
+
"learning_rate": 4.401362660610397e-05,
|
| 972 |
+
"loss": 0.6357,
|
| 973 |
+
"step": 79500
|
| 974 |
+
},
|
| 975 |
+
{
|
| 976 |
+
"epoch": 1.86,
|
| 977 |
+
"learning_rate": 4.3974737890053824e-05,
|
| 978 |
+
"loss": 0.6355,
|
| 979 |
+
"step": 80000
|
| 980 |
+
},
|
| 981 |
+
{
|
| 982 |
+
"epoch": 1.87,
|
| 983 |
+
"learning_rate": 4.3935849174003676e-05,
|
| 984 |
+
"loss": 0.6459,
|
| 985 |
+
"step": 80500
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 1.88,
|
| 989 |
+
"learning_rate": 4.389696045795352e-05,
|
| 990 |
+
"loss": 0.6422,
|
| 991 |
+
"step": 81000
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 1.89,
|
| 995 |
+
"learning_rate": 4.385807174190337e-05,
|
| 996 |
+
"loss": 0.6413,
|
| 997 |
+
"step": 81500
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 1.91,
|
| 1001 |
+
"learning_rate": 4.381918302585322e-05,
|
| 1002 |
+
"loss": 0.6391,
|
| 1003 |
+
"step": 82000
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 1.92,
|
| 1007 |
+
"learning_rate": 4.3780294309803075e-05,
|
| 1008 |
+
"loss": 0.6409,
|
| 1009 |
+
"step": 82500
|
| 1010 |
+
},
|
| 1011 |
+
{
|
| 1012 |
+
"epoch": 1.93,
|
| 1013 |
+
"learning_rate": 4.374140559375292e-05,
|
| 1014 |
+
"loss": 0.632,
|
| 1015 |
+
"step": 83000
|
| 1016 |
+
},
|
| 1017 |
+
{
|
| 1018 |
+
"epoch": 1.94,
|
| 1019 |
+
"learning_rate": 4.370251687770277e-05,
|
| 1020 |
+
"loss": 0.6462,
|
| 1021 |
+
"step": 83500
|
| 1022 |
+
},
|
| 1023 |
+
{
|
| 1024 |
+
"epoch": 1.95,
|
| 1025 |
+
"learning_rate": 4.3663628161652616e-05,
|
| 1026 |
+
"loss": 0.6388,
|
| 1027 |
+
"step": 84000
|
| 1028 |
+
},
|
| 1029 |
+
{
|
| 1030 |
+
"epoch": 1.96,
|
| 1031 |
+
"learning_rate": 4.3624817223034566e-05,
|
| 1032 |
+
"loss": 0.642,
|
| 1033 |
+
"step": 84500
|
| 1034 |
+
},
|
| 1035 |
+
{
|
| 1036 |
+
"epoch": 1.98,
|
| 1037 |
+
"learning_rate": 4.358592850698442e-05,
|
| 1038 |
+
"loss": 0.6409,
|
| 1039 |
+
"step": 85000
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 1.99,
|
| 1043 |
+
"learning_rate": 4.354703979093426e-05,
|
| 1044 |
+
"loss": 0.6406,
|
| 1045 |
+
"step": 85500
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 2.0,
|
| 1049 |
+
"learning_rate": 4.350815107488411e-05,
|
| 1050 |
+
"loss": 0.6355,
|
| 1051 |
+
"step": 86000
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 2.0,
|
| 1055 |
+
"eval_bleu": 57.5735,
|
| 1056 |
+
"eval_gen_len": 15.9938,
|
| 1057 |
+
"eval_loss": 0.6641375422477722,
|
| 1058 |
+
"eval_runtime": 6884.7279,
|
| 1059 |
+
"eval_samples_per_second": 12.479,
|
| 1060 |
+
"eval_steps_per_second": 1.56,
|
| 1061 |
+
"step": 86048
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"epoch": 2.01,
|
| 1065 |
+
"learning_rate": 4.3469262358833965e-05,
|
| 1066 |
+
"loss": 0.5199,
|
| 1067 |
+
"step": 86500
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 2.02,
|
| 1071 |
+
"learning_rate": 4.343037364278381e-05,
|
| 1072 |
+
"loss": 0.4964,
|
| 1073 |
+
"step": 87000
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 2.03,
|
| 1077 |
+
"learning_rate": 4.339148492673366e-05,
|
| 1078 |
+
"loss": 0.5025,
|
| 1079 |
+
"step": 87500
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 2.05,
|
| 1083 |
+
"learning_rate": 4.335259621068351e-05,
|
| 1084 |
+
"loss": 0.508,
|
| 1085 |
+
"step": 88000
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 2.06,
|
| 1089 |
+
"learning_rate": 4.331378527206546e-05,
|
| 1090 |
+
"loss": 0.5096,
|
| 1091 |
+
"step": 88500
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 2.07,
|
| 1095 |
+
"learning_rate": 4.327497433344741e-05,
|
| 1096 |
+
"loss": 0.5106,
|
| 1097 |
+
"step": 89000
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"epoch": 2.08,
|
| 1101 |
+
"learning_rate": 4.323616339482936e-05,
|
| 1102 |
+
"loss": 0.5108,
|
| 1103 |
+
"step": 89500
|
| 1104 |
+
},
|
| 1105 |
+
{
|
| 1106 |
+
"epoch": 2.09,
|
| 1107 |
+
"learning_rate": 4.3197274678779206e-05,
|
| 1108 |
+
"loss": 0.5057,
|
| 1109 |
+
"step": 90000
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 2.1,
|
| 1113 |
+
"learning_rate": 4.3158463740161156e-05,
|
| 1114 |
+
"loss": 0.5108,
|
| 1115 |
+
"step": 90500
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 2.12,
|
| 1119 |
+
"learning_rate": 4.311957502411101e-05,
|
| 1120 |
+
"loss": 0.5044,
|
| 1121 |
+
"step": 91000
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 2.13,
|
| 1125 |
+
"learning_rate": 4.308068630806085e-05,
|
| 1126 |
+
"loss": 0.5077,
|
| 1127 |
+
"step": 91500
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 2.14,
|
| 1131 |
+
"learning_rate": 4.3041797592010704e-05,
|
| 1132 |
+
"loss": 0.5007,
|
| 1133 |
+
"step": 92000
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 2.15,
|
| 1137 |
+
"learning_rate": 4.300290887596055e-05,
|
| 1138 |
+
"loss": 0.5068,
|
| 1139 |
+
"step": 92500
|
| 1140 |
+
},
|
| 1141 |
+
{
|
| 1142 |
+
"epoch": 2.16,
|
| 1143 |
+
"learning_rate": 4.29640201599104e-05,
|
| 1144 |
+
"loss": 0.5038,
|
| 1145 |
+
"step": 93000
|
| 1146 |
+
},
|
| 1147 |
+
{
|
| 1148 |
+
"epoch": 2.17,
|
| 1149 |
+
"learning_rate": 4.292513144386025e-05,
|
| 1150 |
+
"loss": 0.5107,
|
| 1151 |
+
"step": 93500
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 2.18,
|
| 1155 |
+
"learning_rate": 4.28862427278101e-05,
|
| 1156 |
+
"loss": 0.5117,
|
| 1157 |
+
"step": 94000
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 2.2,
|
| 1161 |
+
"learning_rate": 4.284750956662415e-05,
|
| 1162 |
+
"loss": 0.5129,
|
| 1163 |
+
"step": 94500
|
| 1164 |
+
},
|
| 1165 |
+
{
|
| 1166 |
+
"epoch": 2.21,
|
| 1167 |
+
"learning_rate": 4.2808620850574e-05,
|
| 1168 |
+
"loss": 0.5119,
|
| 1169 |
+
"step": 95000
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 2.22,
|
| 1173 |
+
"learning_rate": 4.276973213452385e-05,
|
| 1174 |
+
"loss": 0.5198,
|
| 1175 |
+
"step": 95500
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 2.23,
|
| 1179 |
+
"learning_rate": 4.27308434184737e-05,
|
| 1180 |
+
"loss": 0.5138,
|
| 1181 |
+
"step": 96000
|
| 1182 |
+
},
|
| 1183 |
+
{
|
| 1184 |
+
"epoch": 2.24,
|
| 1185 |
+
"learning_rate": 4.269195470242355e-05,
|
| 1186 |
+
"loss": 0.516,
|
| 1187 |
+
"step": 96500
|
| 1188 |
+
},
|
| 1189 |
+
{
|
| 1190 |
+
"epoch": 2.25,
|
| 1191 |
+
"learning_rate": 4.2653065986373395e-05,
|
| 1192 |
+
"loss": 0.521,
|
| 1193 |
+
"step": 97000
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 2.27,
|
| 1197 |
+
"learning_rate": 4.2614177270323247e-05,
|
| 1198 |
+
"loss": 0.5162,
|
| 1199 |
+
"step": 97500
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 2.28,
|
| 1203 |
+
"learning_rate": 4.25752885542731e-05,
|
| 1204 |
+
"loss": 0.5175,
|
| 1205 |
+
"step": 98000
|
| 1206 |
+
},
|
| 1207 |
+
{
|
| 1208 |
+
"epoch": 2.29,
|
| 1209 |
+
"learning_rate": 4.253639983822294e-05,
|
| 1210 |
+
"loss": 0.518,
|
| 1211 |
+
"step": 98500
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"epoch": 2.3,
|
| 1215 |
+
"learning_rate": 4.249758889960489e-05,
|
| 1216 |
+
"loss": 0.5234,
|
| 1217 |
+
"step": 99000
|
| 1218 |
+
},
|
| 1219 |
+
{
|
| 1220 |
+
"epoch": 2.31,
|
| 1221 |
+
"learning_rate": 4.245870018355474e-05,
|
| 1222 |
+
"loss": 0.524,
|
| 1223 |
+
"step": 99500
|
| 1224 |
+
},
|
| 1225 |
+
{
|
| 1226 |
+
"epoch": 2.32,
|
| 1227 |
+
"learning_rate": 4.241981146750459e-05,
|
| 1228 |
+
"loss": 0.5263,
|
| 1229 |
+
"step": 100000
|
| 1230 |
+
},
|
| 1231 |
+
{
|
| 1232 |
+
"epoch": 2.34,
|
| 1233 |
+
"learning_rate": 4.238092275145444e-05,
|
| 1234 |
+
"loss": 0.5171,
|
| 1235 |
+
"step": 100500
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 2.35,
|
| 1239 |
+
"learning_rate": 4.2342034035404285e-05,
|
| 1240 |
+
"loss": 0.5265,
|
| 1241 |
+
"step": 101000
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 2.36,
|
| 1245 |
+
"learning_rate": 4.2303145319354137e-05,
|
| 1246 |
+
"loss": 0.5156,
|
| 1247 |
+
"step": 101500
|
| 1248 |
+
},
|
| 1249 |
+
{
|
| 1250 |
+
"epoch": 2.37,
|
| 1251 |
+
"learning_rate": 4.226425660330399e-05,
|
| 1252 |
+
"loss": 0.5255,
|
| 1253 |
+
"step": 102000
|
| 1254 |
+
},
|
| 1255 |
+
{
|
| 1256 |
+
"epoch": 2.38,
|
| 1257 |
+
"learning_rate": 4.222536788725383e-05,
|
| 1258 |
+
"loss": 0.5187,
|
| 1259 |
+
"step": 102500
|
| 1260 |
+
},
|
| 1261 |
+
{
|
| 1262 |
+
"epoch": 2.39,
|
| 1263 |
+
"learning_rate": 4.218655694863579e-05,
|
| 1264 |
+
"loss": 0.5298,
|
| 1265 |
+
"step": 103000
|
| 1266 |
+
},
|
| 1267 |
+
{
|
| 1268 |
+
"epoch": 2.41,
|
| 1269 |
+
"learning_rate": 4.2147668232585634e-05,
|
| 1270 |
+
"loss": 0.5258,
|
| 1271 |
+
"step": 103500
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"epoch": 2.42,
|
| 1275 |
+
"learning_rate": 4.2108779516535485e-05,
|
| 1276 |
+
"loss": 0.5206,
|
| 1277 |
+
"step": 104000
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 2.43,
|
| 1281 |
+
"learning_rate": 4.206989080048534e-05,
|
| 1282 |
+
"loss": 0.5292,
|
| 1283 |
+
"step": 104500
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 2.44,
|
| 1287 |
+
"learning_rate": 4.203100208443518e-05,
|
| 1288 |
+
"loss": 0.5156,
|
| 1289 |
+
"step": 105000
|
| 1290 |
+
},
|
| 1291 |
+
{
|
| 1292 |
+
"epoch": 2.45,
|
| 1293 |
+
"learning_rate": 4.199211336838503e-05,
|
| 1294 |
+
"loss": 0.5156,
|
| 1295 |
+
"step": 105500
|
| 1296 |
+
},
|
| 1297 |
+
{
|
| 1298 |
+
"epoch": 2.46,
|
| 1299 |
+
"learning_rate": 4.1953224652334885e-05,
|
| 1300 |
+
"loss": 0.5259,
|
| 1301 |
+
"step": 106000
|
| 1302 |
+
},
|
| 1303 |
+
{
|
| 1304 |
+
"epoch": 2.48,
|
| 1305 |
+
"learning_rate": 4.191433593628473e-05,
|
| 1306 |
+
"loss": 0.5236,
|
| 1307 |
+
"step": 106500
|
| 1308 |
+
},
|
| 1309 |
+
{
|
| 1310 |
+
"epoch": 2.49,
|
| 1311 |
+
"learning_rate": 4.187552499766668e-05,
|
| 1312 |
+
"loss": 0.5282,
|
| 1313 |
+
"step": 107000
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 2.5,
|
| 1317 |
+
"learning_rate": 4.1836636281616524e-05,
|
| 1318 |
+
"loss": 0.5342,
|
| 1319 |
+
"step": 107500
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 2.51,
|
| 1323 |
+
"learning_rate": 4.1797747565566375e-05,
|
| 1324 |
+
"loss": 0.5246,
|
| 1325 |
+
"step": 108000
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 2.52,
|
| 1329 |
+
"learning_rate": 4.175885884951623e-05,
|
| 1330 |
+
"loss": 0.5238,
|
| 1331 |
+
"step": 108500
|
| 1332 |
+
},
|
| 1333 |
+
{
|
| 1334 |
+
"epoch": 2.53,
|
| 1335 |
+
"learning_rate": 4.171997013346607e-05,
|
| 1336 |
+
"loss": 0.5329,
|
| 1337 |
+
"step": 109000
|
| 1338 |
+
},
|
| 1339 |
+
{
|
| 1340 |
+
"epoch": 2.55,
|
| 1341 |
+
"learning_rate": 4.168108141741592e-05,
|
| 1342 |
+
"loss": 0.5223,
|
| 1343 |
+
"step": 109500
|
| 1344 |
+
},
|
| 1345 |
+
{
|
| 1346 |
+
"epoch": 2.56,
|
| 1347 |
+
"learning_rate": 4.164234825622997e-05,
|
| 1348 |
+
"loss": 0.5333,
|
| 1349 |
+
"step": 110000
|
| 1350 |
+
},
|
| 1351 |
+
{
|
| 1352 |
+
"epoch": 2.57,
|
| 1353 |
+
"learning_rate": 4.160345954017982e-05,
|
| 1354 |
+
"loss": 0.5353,
|
| 1355 |
+
"step": 110500
|
| 1356 |
+
},
|
| 1357 |
+
{
|
| 1358 |
+
"epoch": 2.58,
|
| 1359 |
+
"learning_rate": 4.1564570824129674e-05,
|
| 1360 |
+
"loss": 0.5285,
|
| 1361 |
+
"step": 111000
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 2.59,
|
| 1365 |
+
"learning_rate": 4.152568210807952e-05,
|
| 1366 |
+
"loss": 0.5365,
|
| 1367 |
+
"step": 111500
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 2.6,
|
| 1371 |
+
"learning_rate": 4.1486948946893573e-05,
|
| 1372 |
+
"loss": 0.5214,
|
| 1373 |
+
"step": 112000
|
| 1374 |
+
},
|
| 1375 |
+
{
|
| 1376 |
+
"epoch": 2.61,
|
| 1377 |
+
"learning_rate": 4.1448060230843425e-05,
|
| 1378 |
+
"loss": 0.534,
|
| 1379 |
+
"step": 112500
|
| 1380 |
+
},
|
| 1381 |
+
{
|
| 1382 |
+
"epoch": 2.63,
|
| 1383 |
+
"learning_rate": 4.140917151479327e-05,
|
| 1384 |
+
"loss": 0.5375,
|
| 1385 |
+
"step": 113000
|
| 1386 |
+
},
|
| 1387 |
+
{
|
| 1388 |
+
"epoch": 2.64,
|
| 1389 |
+
"learning_rate": 4.137028279874312e-05,
|
| 1390 |
+
"loss": 0.532,
|
| 1391 |
+
"step": 113500
|
| 1392 |
+
},
|
| 1393 |
+
{
|
| 1394 |
+
"epoch": 2.65,
|
| 1395 |
+
"learning_rate": 4.133139408269297e-05,
|
| 1396 |
+
"loss": 0.5232,
|
| 1397 |
+
"step": 114000
|
| 1398 |
+
},
|
| 1399 |
+
{
|
| 1400 |
+
"epoch": 2.66,
|
| 1401 |
+
"learning_rate": 4.129250536664282e-05,
|
| 1402 |
+
"loss": 0.5333,
|
| 1403 |
+
"step": 114500
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 2.67,
|
| 1407 |
+
"learning_rate": 4.125361665059266e-05,
|
| 1408 |
+
"loss": 0.5269,
|
| 1409 |
+
"step": 115000
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 2.68,
|
| 1413 |
+
"learning_rate": 4.1214727934542514e-05,
|
| 1414 |
+
"loss": 0.5186,
|
| 1415 |
+
"step": 115500
|
| 1416 |
+
},
|
| 1417 |
+
{
|
| 1418 |
+
"epoch": 2.7,
|
| 1419 |
+
"learning_rate": 4.1175839218492365e-05,
|
| 1420 |
+
"loss": 0.5226,
|
| 1421 |
+
"step": 116000
|
| 1422 |
+
},
|
| 1423 |
+
{
|
| 1424 |
+
"epoch": 2.71,
|
| 1425 |
+
"learning_rate": 4.1137028279874315e-05,
|
| 1426 |
+
"loss": 0.5278,
|
| 1427 |
+
"step": 116500
|
| 1428 |
+
},
|
| 1429 |
+
{
|
| 1430 |
+
"epoch": 2.72,
|
| 1431 |
+
"learning_rate": 4.109813956382416e-05,
|
| 1432 |
+
"loss": 0.5311,
|
| 1433 |
+
"step": 117000
|
| 1434 |
+
},
|
| 1435 |
+
{
|
| 1436 |
+
"epoch": 2.73,
|
| 1437 |
+
"learning_rate": 4.105925084777401e-05,
|
| 1438 |
+
"loss": 0.5274,
|
| 1439 |
+
"step": 117500
|
| 1440 |
+
},
|
| 1441 |
+
{
|
| 1442 |
+
"epoch": 2.74,
|
| 1443 |
+
"learning_rate": 4.102036213172386e-05,
|
| 1444 |
+
"loss": 0.5348,
|
| 1445 |
+
"step": 118000
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 2.75,
|
| 1449 |
+
"learning_rate": 4.098147341567371e-05,
|
| 1450 |
+
"loss": 0.5359,
|
| 1451 |
+
"step": 118500
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 2.77,
|
| 1455 |
+
"learning_rate": 4.094258469962356e-05,
|
| 1456 |
+
"loss": 0.5416,
|
| 1457 |
+
"step": 119000
|
| 1458 |
+
},
|
| 1459 |
+
{
|
| 1460 |
+
"epoch": 2.78,
|
| 1461 |
+
"learning_rate": 4.090369598357341e-05,
|
| 1462 |
+
"loss": 0.526,
|
| 1463 |
+
"step": 119500
|
| 1464 |
+
},
|
| 1465 |
+
{
|
| 1466 |
+
"epoch": 2.79,
|
| 1467 |
+
"learning_rate": 4.086488504495536e-05,
|
| 1468 |
+
"loss": 0.5404,
|
| 1469 |
+
"step": 120000
|
| 1470 |
+
},
|
| 1471 |
+
{
|
| 1472 |
+
"epoch": 2.8,
|
| 1473 |
+
"learning_rate": 4.0825996328905205e-05,
|
| 1474 |
+
"loss": 0.5321,
|
| 1475 |
+
"step": 120500
|
| 1476 |
+
},
|
| 1477 |
+
{
|
| 1478 |
+
"epoch": 2.81,
|
| 1479 |
+
"learning_rate": 4.0787107612855056e-05,
|
| 1480 |
+
"loss": 0.5378,
|
| 1481 |
+
"step": 121000
|
| 1482 |
+
},
|
| 1483 |
+
{
|
| 1484 |
+
"epoch": 2.82,
|
| 1485 |
+
"learning_rate": 4.074821889680491e-05,
|
| 1486 |
+
"loss": 0.5371,
|
| 1487 |
+
"step": 121500
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 2.84,
|
| 1491 |
+
"learning_rate": 4.070933018075476e-05,
|
| 1492 |
+
"loss": 0.5303,
|
| 1493 |
+
"step": 122000
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 2.85,
|
| 1497 |
+
"learning_rate": 4.06704414647046e-05,
|
| 1498 |
+
"loss": 0.5321,
|
| 1499 |
+
"step": 122500
|
| 1500 |
+
},
|
| 1501 |
+
{
|
| 1502 |
+
"epoch": 2.86,
|
| 1503 |
+
"learning_rate": 4.063163052608655e-05,
|
| 1504 |
+
"loss": 0.5291,
|
| 1505 |
+
"step": 123000
|
| 1506 |
+
},
|
| 1507 |
+
{
|
| 1508 |
+
"epoch": 2.87,
|
| 1509 |
+
"learning_rate": 4.05927418100364e-05,
|
| 1510 |
+
"loss": 0.5391,
|
| 1511 |
+
"step": 123500
|
| 1512 |
+
},
|
| 1513 |
+
{
|
| 1514 |
+
"epoch": 2.88,
|
| 1515 |
+
"learning_rate": 4.055385309398625e-05,
|
| 1516 |
+
"loss": 0.5344,
|
| 1517 |
+
"step": 124000
|
| 1518 |
+
},
|
| 1519 |
+
{
|
| 1520 |
+
"epoch": 2.89,
|
| 1521 |
+
"learning_rate": 4.05149643779361e-05,
|
| 1522 |
+
"loss": 0.5381,
|
| 1523 |
+
"step": 124500
|
| 1524 |
+
},
|
| 1525 |
+
{
|
| 1526 |
+
"epoch": 2.91,
|
| 1527 |
+
"learning_rate": 4.0476075661885946e-05,
|
| 1528 |
+
"loss": 0.5348,
|
| 1529 |
+
"step": 125000
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 2.92,
|
| 1533 |
+
"learning_rate": 4.04371869458358e-05,
|
| 1534 |
+
"loss": 0.5367,
|
| 1535 |
+
"step": 125500
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 2.93,
|
| 1539 |
+
"learning_rate": 4.039829822978565e-05,
|
| 1540 |
+
"loss": 0.5294,
|
| 1541 |
+
"step": 126000
|
| 1542 |
+
},
|
| 1543 |
+
{
|
| 1544 |
+
"epoch": 2.94,
|
| 1545 |
+
"learning_rate": 4.0359409513735494e-05,
|
| 1546 |
+
"loss": 0.5257,
|
| 1547 |
+
"step": 126500
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"epoch": 2.95,
|
| 1551 |
+
"learning_rate": 4.0320520797685346e-05,
|
| 1552 |
+
"loss": 0.5241,
|
| 1553 |
+
"step": 127000
|
| 1554 |
+
},
|
| 1555 |
+
{
|
| 1556 |
+
"epoch": 2.96,
|
| 1557 |
+
"learning_rate": 4.02816320816352e-05,
|
| 1558 |
+
"loss": 0.5375,
|
| 1559 |
+
"step": 127500
|
| 1560 |
+
},
|
| 1561 |
+
{
|
| 1562 |
+
"epoch": 2.98,
|
| 1563 |
+
"learning_rate": 4.024274336558504e-05,
|
| 1564 |
+
"loss": 0.5332,
|
| 1565 |
+
"step": 128000
|
| 1566 |
+
},
|
| 1567 |
+
{
|
| 1568 |
+
"epoch": 2.99,
|
| 1569 |
+
"learning_rate": 4.0203854649534894e-05,
|
| 1570 |
+
"loss": 0.532,
|
| 1571 |
+
"step": 128500
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 3.0,
|
| 1575 |
+
"learning_rate": 4.0164965933484745e-05,
|
| 1576 |
+
"loss": 0.5464,
|
| 1577 |
+
"step": 129000
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 3.0,
|
| 1581 |
+
"eval_bleu": 58.837,
|
| 1582 |
+
"eval_gen_len": 15.7689,
|
| 1583 |
+
"eval_loss": 0.6568061113357544,
|
| 1584 |
+
"eval_runtime": 6740.2004,
|
| 1585 |
+
"eval_samples_per_second": 12.747,
|
| 1586 |
+
"eval_steps_per_second": 1.593,
|
| 1587 |
+
"step": 129072
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 3.01,
|
| 1591 |
+
"learning_rate": 4.0126077217434597e-05,
|
| 1592 |
+
"loss": 0.4193,
|
| 1593 |
+
"step": 129500
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 3.02,
|
| 1597 |
+
"learning_rate": 4.0087344056248644e-05,
|
| 1598 |
+
"loss": 0.399,
|
| 1599 |
+
"step": 130000
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 3.03,
|
| 1603 |
+
"learning_rate": 4.004853311763059e-05,
|
| 1604 |
+
"loss": 0.3996,
|
| 1605 |
+
"step": 130500
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 3.04,
|
| 1609 |
+
"learning_rate": 4.0009799956444635e-05,
|
| 1610 |
+
"loss": 0.3937,
|
| 1611 |
+
"step": 131000
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 3.06,
|
| 1615 |
+
"learning_rate": 3.997091124039449e-05,
|
| 1616 |
+
"loss": 0.3955,
|
| 1617 |
+
"step": 131500
|
| 1618 |
+
},
|
| 1619 |
+
{
|
| 1620 |
+
"epoch": 3.07,
|
| 1621 |
+
"learning_rate": 3.993202252434434e-05,
|
| 1622 |
+
"loss": 0.3998,
|
| 1623 |
+
"step": 132000
|
| 1624 |
+
},
|
| 1625 |
+
{
|
| 1626 |
+
"epoch": 3.08,
|
| 1627 |
+
"learning_rate": 3.989313380829419e-05,
|
| 1628 |
+
"loss": 0.3934,
|
| 1629 |
+
"step": 132500
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 3.09,
|
| 1633 |
+
"learning_rate": 3.9854245092244034e-05,
|
| 1634 |
+
"loss": 0.3987,
|
| 1635 |
+
"step": 133000
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 3.1,
|
| 1639 |
+
"learning_rate": 3.9815434153625984e-05,
|
| 1640 |
+
"loss": 0.4048,
|
| 1641 |
+
"step": 133500
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 3.11,
|
| 1645 |
+
"learning_rate": 3.9776545437575836e-05,
|
| 1646 |
+
"loss": 0.406,
|
| 1647 |
+
"step": 134000
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 3.13,
|
| 1651 |
+
"learning_rate": 3.973765672152569e-05,
|
| 1652 |
+
"loss": 0.3987,
|
| 1653 |
+
"step": 134500
|
| 1654 |
+
},
|
| 1655 |
+
{
|
| 1656 |
+
"epoch": 3.14,
|
| 1657 |
+
"learning_rate": 3.969876800547553e-05,
|
| 1658 |
+
"loss": 0.4085,
|
| 1659 |
+
"step": 135000
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 3.15,
|
| 1663 |
+
"learning_rate": 3.9659879289425383e-05,
|
| 1664 |
+
"loss": 0.4047,
|
| 1665 |
+
"step": 135500
|
| 1666 |
+
},
|
| 1667 |
+
{
|
| 1668 |
+
"epoch": 3.16,
|
| 1669 |
+
"learning_rate": 3.9620990573375235e-05,
|
| 1670 |
+
"loss": 0.4021,
|
| 1671 |
+
"step": 136000
|
| 1672 |
+
},
|
| 1673 |
+
{
|
| 1674 |
+
"epoch": 3.17,
|
| 1675 |
+
"learning_rate": 3.958210185732508e-05,
|
| 1676 |
+
"loss": 0.4056,
|
| 1677 |
+
"step": 136500
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 3.18,
|
| 1681 |
+
"learning_rate": 3.954321314127493e-05,
|
| 1682 |
+
"loss": 0.4019,
|
| 1683 |
+
"step": 137000
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 3.2,
|
| 1687 |
+
"learning_rate": 3.950440220265688e-05,
|
| 1688 |
+
"loss": 0.4125,
|
| 1689 |
+
"step": 137500
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 3.21,
|
| 1693 |
+
"learning_rate": 3.9465513486606726e-05,
|
| 1694 |
+
"loss": 0.4087,
|
| 1695 |
+
"step": 138000
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 3.22,
|
| 1699 |
+
"learning_rate": 3.942662477055658e-05,
|
| 1700 |
+
"loss": 0.3971,
|
| 1701 |
+
"step": 138500
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 3.23,
|
| 1705 |
+
"learning_rate": 3.938773605450642e-05,
|
| 1706 |
+
"loss": 0.4122,
|
| 1707 |
+
"step": 139000
|
| 1708 |
+
},
|
| 1709 |
+
{
|
| 1710 |
+
"epoch": 3.24,
|
| 1711 |
+
"learning_rate": 3.9348847338456273e-05,
|
| 1712 |
+
"loss": 0.4086,
|
| 1713 |
+
"step": 139500
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 3.25,
|
| 1717 |
+
"learning_rate": 3.931003639983822e-05,
|
| 1718 |
+
"loss": 0.4173,
|
| 1719 |
+
"step": 140000
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 3.27,
|
| 1723 |
+
"learning_rate": 3.9271147683788075e-05,
|
| 1724 |
+
"loss": 0.4158,
|
| 1725 |
+
"step": 140500
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 3.28,
|
| 1729 |
+
"learning_rate": 3.923225896773792e-05,
|
| 1730 |
+
"loss": 0.4072,
|
| 1731 |
+
"step": 141000
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 3.29,
|
| 1735 |
+
"learning_rate": 3.919337025168777e-05,
|
| 1736 |
+
"loss": 0.4156,
|
| 1737 |
+
"step": 141500
|
| 1738 |
+
},
|
| 1739 |
+
{
|
| 1740 |
+
"epoch": 3.3,
|
| 1741 |
+
"learning_rate": 3.915448153563762e-05,
|
| 1742 |
+
"loss": 0.4174,
|
| 1743 |
+
"step": 142000
|
| 1744 |
+
},
|
| 1745 |
+
{
|
| 1746 |
+
"epoch": 3.31,
|
| 1747 |
+
"learning_rate": 3.9115592819587474e-05,
|
| 1748 |
+
"loss": 0.416,
|
| 1749 |
+
"step": 142500
|
| 1750 |
+
},
|
| 1751 |
+
{
|
| 1752 |
+
"epoch": 3.32,
|
| 1753 |
+
"learning_rate": 3.9076781880969424e-05,
|
| 1754 |
+
"loss": 0.4151,
|
| 1755 |
+
"step": 143000
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 3.34,
|
| 1759 |
+
"learning_rate": 3.903789316491927e-05,
|
| 1760 |
+
"loss": 0.4188,
|
| 1761 |
+
"step": 143500
|
| 1762 |
+
},
|
| 1763 |
+
{
|
| 1764 |
+
"epoch": 3.35,
|
| 1765 |
+
"learning_rate": 3.899900444886912e-05,
|
| 1766 |
+
"loss": 0.419,
|
| 1767 |
+
"step": 144000
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 3.36,
|
| 1771 |
+
"learning_rate": 3.896011573281897e-05,
|
| 1772 |
+
"loss": 0.4206,
|
| 1773 |
+
"step": 144500
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 3.37,
|
| 1777 |
+
"learning_rate": 3.8921227016768816e-05,
|
| 1778 |
+
"loss": 0.4243,
|
| 1779 |
+
"step": 145000
|
| 1780 |
+
},
|
| 1781 |
+
{
|
| 1782 |
+
"epoch": 3.38,
|
| 1783 |
+
"learning_rate": 3.888233830071866e-05,
|
| 1784 |
+
"loss": 0.4131,
|
| 1785 |
+
"step": 145500
|
| 1786 |
+
},
|
| 1787 |
+
{
|
| 1788 |
+
"epoch": 3.39,
|
| 1789 |
+
"learning_rate": 3.884344958466851e-05,
|
| 1790 |
+
"loss": 0.4192,
|
| 1791 |
+
"step": 146000
|
| 1792 |
+
},
|
| 1793 |
+
{
|
| 1794 |
+
"epoch": 3.41,
|
| 1795 |
+
"learning_rate": 3.8804560868618364e-05,
|
| 1796 |
+
"loss": 0.4179,
|
| 1797 |
+
"step": 146500
|
| 1798 |
+
},
|
| 1799 |
+
{
|
| 1800 |
+
"epoch": 3.42,
|
| 1801 |
+
"learning_rate": 3.876567215256821e-05,
|
| 1802 |
+
"loss": 0.4264,
|
| 1803 |
+
"step": 147000
|
| 1804 |
+
},
|
| 1805 |
+
{
|
| 1806 |
+
"epoch": 3.43,
|
| 1807 |
+
"learning_rate": 3.872678343651806e-05,
|
| 1808 |
+
"loss": 0.4231,
|
| 1809 |
+
"step": 147500
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 3.44,
|
| 1813 |
+
"learning_rate": 3.868789472046791e-05,
|
| 1814 |
+
"loss": 0.4217,
|
| 1815 |
+
"step": 148000
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 3.45,
|
| 1819 |
+
"learning_rate": 3.8649006004417756e-05,
|
| 1820 |
+
"loss": 0.4189,
|
| 1821 |
+
"step": 148500
|
| 1822 |
+
},
|
| 1823 |
+
{
|
| 1824 |
+
"epoch": 3.46,
|
| 1825 |
+
"learning_rate": 3.861011728836761e-05,
|
| 1826 |
+
"loss": 0.4198,
|
| 1827 |
+
"step": 149000
|
| 1828 |
+
},
|
| 1829 |
+
{
|
| 1830 |
+
"epoch": 3.47,
|
| 1831 |
+
"learning_rate": 3.857122857231746e-05,
|
| 1832 |
+
"loss": 0.4237,
|
| 1833 |
+
"step": 149500
|
| 1834 |
+
},
|
| 1835 |
+
{
|
| 1836 |
+
"epoch": 3.49,
|
| 1837 |
+
"learning_rate": 3.853233985626731e-05,
|
| 1838 |
+
"loss": 0.4147,
|
| 1839 |
+
"step": 150000
|
| 1840 |
+
},
|
| 1841 |
+
{
|
| 1842 |
+
"epoch": 3.5,
|
| 1843 |
+
"learning_rate": 3.849352891764926e-05,
|
| 1844 |
+
"loss": 0.4219,
|
| 1845 |
+
"step": 150500
|
| 1846 |
+
},
|
| 1847 |
+
{
|
| 1848 |
+
"epoch": 3.51,
|
| 1849 |
+
"learning_rate": 3.8454640201599105e-05,
|
| 1850 |
+
"loss": 0.4274,
|
| 1851 |
+
"step": 151000
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 3.52,
|
| 1855 |
+
"learning_rate": 3.8415829262981055e-05,
|
| 1856 |
+
"loss": 0.4283,
|
| 1857 |
+
"step": 151500
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 3.53,
|
| 1861 |
+
"learning_rate": 3.8376940546930907e-05,
|
| 1862 |
+
"loss": 0.4222,
|
| 1863 |
+
"step": 152000
|
| 1864 |
+
},
|
| 1865 |
+
{
|
| 1866 |
+
"epoch": 3.54,
|
| 1867 |
+
"learning_rate": 3.833805183088076e-05,
|
| 1868 |
+
"loss": 0.4269,
|
| 1869 |
+
"step": 152500
|
| 1870 |
+
},
|
| 1871 |
+
{
|
| 1872 |
+
"epoch": 3.56,
|
| 1873 |
+
"learning_rate": 3.82991631148306e-05,
|
| 1874 |
+
"loss": 0.4271,
|
| 1875 |
+
"step": 153000
|
| 1876 |
+
},
|
| 1877 |
+
{
|
| 1878 |
+
"epoch": 3.57,
|
| 1879 |
+
"learning_rate": 3.826035217621255e-05,
|
| 1880 |
+
"loss": 0.425,
|
| 1881 |
+
"step": 153500
|
| 1882 |
+
},
|
| 1883 |
+
{
|
| 1884 |
+
"epoch": 3.58,
|
| 1885 |
+
"learning_rate": 3.82214634601624e-05,
|
| 1886 |
+
"loss": 0.4222,
|
| 1887 |
+
"step": 154000
|
| 1888 |
+
},
|
| 1889 |
+
{
|
| 1890 |
+
"epoch": 3.59,
|
| 1891 |
+
"learning_rate": 3.818257474411225e-05,
|
| 1892 |
+
"loss": 0.4236,
|
| 1893 |
+
"step": 154500
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 3.6,
|
| 1897 |
+
"learning_rate": 3.81436860280621e-05,
|
| 1898 |
+
"loss": 0.4327,
|
| 1899 |
+
"step": 155000
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 3.61,
|
| 1903 |
+
"learning_rate": 3.8104797312011945e-05,
|
| 1904 |
+
"loss": 0.4268,
|
| 1905 |
+
"step": 155500
|
| 1906 |
+
},
|
| 1907 |
+
{
|
| 1908 |
+
"epoch": 3.63,
|
| 1909 |
+
"learning_rate": 3.8065908595961797e-05,
|
| 1910 |
+
"loss": 0.4319,
|
| 1911 |
+
"step": 156000
|
| 1912 |
+
},
|
| 1913 |
+
{
|
| 1914 |
+
"epoch": 3.64,
|
| 1915 |
+
"learning_rate": 3.802701987991165e-05,
|
| 1916 |
+
"loss": 0.4312,
|
| 1917 |
+
"step": 156500
|
| 1918 |
+
},
|
| 1919 |
+
{
|
| 1920 |
+
"epoch": 3.65,
|
| 1921 |
+
"learning_rate": 3.798813116386149e-05,
|
| 1922 |
+
"loss": 0.4328,
|
| 1923 |
+
"step": 157000
|
| 1924 |
+
},
|
| 1925 |
+
{
|
| 1926 |
+
"epoch": 3.66,
|
| 1927 |
+
"learning_rate": 3.7949242447811344e-05,
|
| 1928 |
+
"loss": 0.4292,
|
| 1929 |
+
"step": 157500
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"epoch": 3.67,
|
| 1933 |
+
"learning_rate": 3.7910353731761196e-05,
|
| 1934 |
+
"loss": 0.4317,
|
| 1935 |
+
"step": 158000
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 3.68,
|
| 1939 |
+
"learning_rate": 3.787146501571104e-05,
|
| 1940 |
+
"loss": 0.4309,
|
| 1941 |
+
"step": 158500
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 3.7,
|
| 1945 |
+
"learning_rate": 3.7832654077093e-05,
|
| 1946 |
+
"loss": 0.4425,
|
| 1947 |
+
"step": 159000
|
| 1948 |
+
},
|
| 1949 |
+
{
|
| 1950 |
+
"epoch": 3.71,
|
| 1951 |
+
"learning_rate": 3.779376536104284e-05,
|
| 1952 |
+
"loss": 0.4335,
|
| 1953 |
+
"step": 159500
|
| 1954 |
+
},
|
| 1955 |
+
{
|
| 1956 |
+
"epoch": 3.72,
|
| 1957 |
+
"learning_rate": 3.775487664499269e-05,
|
| 1958 |
+
"loss": 0.4295,
|
| 1959 |
+
"step": 160000
|
| 1960 |
+
},
|
| 1961 |
+
{
|
| 1962 |
+
"epoch": 3.73,
|
| 1963 |
+
"learning_rate": 3.7715987928942545e-05,
|
| 1964 |
+
"loss": 0.4305,
|
| 1965 |
+
"step": 160500
|
| 1966 |
+
},
|
| 1967 |
+
{
|
| 1968 |
+
"epoch": 3.74,
|
| 1969 |
+
"learning_rate": 3.767709921289239e-05,
|
| 1970 |
+
"loss": 0.4264,
|
| 1971 |
+
"step": 161000
|
| 1972 |
+
},
|
| 1973 |
+
{
|
| 1974 |
+
"epoch": 3.75,
|
| 1975 |
+
"learning_rate": 3.7638210496842234e-05,
|
| 1976 |
+
"loss": 0.43,
|
| 1977 |
+
"step": 161500
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 3.77,
|
| 1981 |
+
"learning_rate": 3.7599321780792086e-05,
|
| 1982 |
+
"loss": 0.4329,
|
| 1983 |
+
"step": 162000
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 3.78,
|
| 1987 |
+
"learning_rate": 3.756043306474194e-05,
|
| 1988 |
+
"loss": 0.4324,
|
| 1989 |
+
"step": 162500
|
| 1990 |
+
},
|
| 1991 |
+
{
|
| 1992 |
+
"epoch": 3.79,
|
| 1993 |
+
"learning_rate": 3.752154434869178e-05,
|
| 1994 |
+
"loss": 0.434,
|
| 1995 |
+
"step": 163000
|
| 1996 |
+
},
|
| 1997 |
+
{
|
| 1998 |
+
"epoch": 3.8,
|
| 1999 |
+
"learning_rate": 3.7482655632641634e-05,
|
| 2000 |
+
"loss": 0.4317,
|
| 2001 |
+
"step": 163500
|
| 2002 |
+
},
|
| 2003 |
+
{
|
| 2004 |
+
"epoch": 3.81,
|
| 2005 |
+
"learning_rate": 3.744384469402358e-05,
|
| 2006 |
+
"loss": 0.4379,
|
| 2007 |
+
"step": 164000
|
| 2008 |
+
},
|
| 2009 |
+
{
|
| 2010 |
+
"epoch": 3.82,
|
| 2011 |
+
"learning_rate": 3.7404955977973435e-05,
|
| 2012 |
+
"loss": 0.4427,
|
| 2013 |
+
"step": 164500
|
| 2014 |
+
},
|
| 2015 |
+
{
|
| 2016 |
+
"epoch": 3.84,
|
| 2017 |
+
"learning_rate": 3.736606726192328e-05,
|
| 2018 |
+
"loss": 0.433,
|
| 2019 |
+
"step": 165000
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 3.85,
|
| 2023 |
+
"learning_rate": 3.732717854587313e-05,
|
| 2024 |
+
"loss": 0.4407,
|
| 2025 |
+
"step": 165500
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 3.86,
|
| 2029 |
+
"learning_rate": 3.728828982982298e-05,
|
| 2030 |
+
"loss": 0.4416,
|
| 2031 |
+
"step": 166000
|
| 2032 |
+
},
|
| 2033 |
+
{
|
| 2034 |
+
"epoch": 3.87,
|
| 2035 |
+
"learning_rate": 3.724940111377283e-05,
|
| 2036 |
+
"loss": 0.4407,
|
| 2037 |
+
"step": 166500
|
| 2038 |
+
},
|
| 2039 |
+
{
|
| 2040 |
+
"epoch": 3.88,
|
| 2041 |
+
"learning_rate": 3.721051239772268e-05,
|
| 2042 |
+
"loss": 0.4379,
|
| 2043 |
+
"step": 167000
|
| 2044 |
+
},
|
| 2045 |
+
{
|
| 2046 |
+
"epoch": 3.89,
|
| 2047 |
+
"learning_rate": 3.717170145910463e-05,
|
| 2048 |
+
"loss": 0.4345,
|
| 2049 |
+
"step": 167500
|
| 2050 |
+
},
|
| 2051 |
+
{
|
| 2052 |
+
"epoch": 3.9,
|
| 2053 |
+
"learning_rate": 3.713289052048658e-05,
|
| 2054 |
+
"loss": 0.435,
|
| 2055 |
+
"step": 168000
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 3.92,
|
| 2059 |
+
"learning_rate": 3.709400180443643e-05,
|
| 2060 |
+
"loss": 0.4366,
|
| 2061 |
+
"step": 168500
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 3.93,
|
| 2065 |
+
"learning_rate": 3.7055113088386275e-05,
|
| 2066 |
+
"loss": 0.4336,
|
| 2067 |
+
"step": 169000
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 3.94,
|
| 2071 |
+
"learning_rate": 3.701622437233612e-05,
|
| 2072 |
+
"loss": 0.4438,
|
| 2073 |
+
"step": 169500
|
| 2074 |
+
},
|
| 2075 |
+
{
|
| 2076 |
+
"epoch": 3.95,
|
| 2077 |
+
"learning_rate": 3.697741343371807e-05,
|
| 2078 |
+
"loss": 0.4389,
|
| 2079 |
+
"step": 170000
|
| 2080 |
+
},
|
| 2081 |
+
{
|
| 2082 |
+
"epoch": 3.96,
|
| 2083 |
+
"learning_rate": 3.693852471766792e-05,
|
| 2084 |
+
"loss": 0.4394,
|
| 2085 |
+
"step": 170500
|
| 2086 |
+
},
|
| 2087 |
+
{
|
| 2088 |
+
"epoch": 3.97,
|
| 2089 |
+
"learning_rate": 3.689963600161777e-05,
|
| 2090 |
+
"loss": 0.4378,
|
| 2091 |
+
"step": 171000
|
| 2092 |
+
},
|
| 2093 |
+
{
|
| 2094 |
+
"epoch": 3.99,
|
| 2095 |
+
"learning_rate": 3.686082506299972e-05,
|
| 2096 |
+
"loss": 0.4443,
|
| 2097 |
+
"step": 171500
|
| 2098 |
+
},
|
| 2099 |
+
{
|
| 2100 |
+
"epoch": 4.0,
|
| 2101 |
+
"learning_rate": 3.682193634694957e-05,
|
| 2102 |
+
"loss": 0.4444,
|
| 2103 |
+
"step": 172000
|
| 2104 |
+
},
|
| 2105 |
+
{
|
| 2106 |
+
"epoch": 4.0,
|
| 2107 |
+
"eval_bleu": 59.3055,
|
| 2108 |
+
"eval_gen_len": 15.7545,
|
| 2109 |
+
"eval_loss": 0.6776524782180786,
|
| 2110 |
+
"eval_runtime": 6664.893,
|
| 2111 |
+
"eval_samples_per_second": 12.891,
|
| 2112 |
+
"eval_steps_per_second": 1.611,
|
| 2113 |
+
"step": 172096
|
| 2114 |
+
},
|
| 2115 |
+
{
|
| 2116 |
+
"epoch": 4.01,
|
| 2117 |
+
"learning_rate": 3.678304763089942e-05,
|
| 2118 |
+
"loss": 0.3253,
|
| 2119 |
+
"step": 172500
|
| 2120 |
+
},
|
| 2121 |
+
{
|
| 2122 |
+
"epoch": 4.02,
|
| 2123 |
+
"learning_rate": 3.674415891484927e-05,
|
| 2124 |
+
"loss": 0.302,
|
| 2125 |
+
"step": 173000
|
| 2126 |
+
},
|
| 2127 |
+
{
|
| 2128 |
+
"epoch": 4.03,
|
| 2129 |
+
"learning_rate": 3.670534797623122e-05,
|
| 2130 |
+
"loss": 0.3034,
|
| 2131 |
+
"step": 173500
|
| 2132 |
+
},
|
| 2133 |
+
{
|
| 2134 |
+
"epoch": 4.04,
|
| 2135 |
+
"learning_rate": 3.666645926018107e-05,
|
| 2136 |
+
"loss": 0.3075,
|
| 2137 |
+
"step": 174000
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 4.06,
|
| 2141 |
+
"learning_rate": 3.6627570544130915e-05,
|
| 2142 |
+
"loss": 0.3039,
|
| 2143 |
+
"step": 174500
|
| 2144 |
+
},
|
| 2145 |
+
{
|
| 2146 |
+
"epoch": 4.07,
|
| 2147 |
+
"learning_rate": 3.6588759605512865e-05,
|
| 2148 |
+
"loss": 0.3102,
|
| 2149 |
+
"step": 175000
|
| 2150 |
+
},
|
| 2151 |
+
{
|
| 2152 |
+
"epoch": 4.08,
|
| 2153 |
+
"learning_rate": 3.6549870889462717e-05,
|
| 2154 |
+
"loss": 0.3005,
|
| 2155 |
+
"step": 175500
|
| 2156 |
+
},
|
| 2157 |
+
{
|
| 2158 |
+
"epoch": 4.09,
|
| 2159 |
+
"learning_rate": 3.651098217341257e-05,
|
| 2160 |
+
"loss": 0.3052,
|
| 2161 |
+
"step": 176000
|
| 2162 |
+
},
|
| 2163 |
+
{
|
| 2164 |
+
"epoch": 4.1,
|
| 2165 |
+
"learning_rate": 3.647217123479451e-05,
|
| 2166 |
+
"loss": 0.313,
|
| 2167 |
+
"step": 176500
|
| 2168 |
+
},
|
| 2169 |
+
{
|
| 2170 |
+
"epoch": 4.11,
|
| 2171 |
+
"learning_rate": 3.643328251874436e-05,
|
| 2172 |
+
"loss": 0.3103,
|
| 2173 |
+
"step": 177000
|
| 2174 |
+
},
|
| 2175 |
+
{
|
| 2176 |
+
"epoch": 4.13,
|
| 2177 |
+
"learning_rate": 3.639439380269421e-05,
|
| 2178 |
+
"loss": 0.3167,
|
| 2179 |
+
"step": 177500
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 4.14,
|
| 2183 |
+
"learning_rate": 3.635550508664406e-05,
|
| 2184 |
+
"loss": 0.3129,
|
| 2185 |
+
"step": 178000
|
| 2186 |
+
},
|
| 2187 |
+
{
|
| 2188 |
+
"epoch": 4.15,
|
| 2189 |
+
"learning_rate": 3.631669414802601e-05,
|
| 2190 |
+
"loss": 0.3067,
|
| 2191 |
+
"step": 178500
|
| 2192 |
+
},
|
| 2193 |
+
{
|
| 2194 |
+
"epoch": 4.16,
|
| 2195 |
+
"learning_rate": 3.627780543197586e-05,
|
| 2196 |
+
"loss": 0.3091,
|
| 2197 |
+
"step": 179000
|
| 2198 |
+
},
|
| 2199 |
+
{
|
| 2200 |
+
"epoch": 4.17,
|
| 2201 |
+
"learning_rate": 3.623891671592571e-05,
|
| 2202 |
+
"loss": 0.3131,
|
| 2203 |
+
"step": 179500
|
| 2204 |
+
},
|
| 2205 |
+
{
|
| 2206 |
+
"epoch": 4.18,
|
| 2207 |
+
"learning_rate": 3.6200027999875556e-05,
|
| 2208 |
+
"loss": 0.3135,
|
| 2209 |
+
"step": 180000
|
| 2210 |
+
},
|
| 2211 |
+
{
|
| 2212 |
+
"epoch": 4.2,
|
| 2213 |
+
"learning_rate": 3.616113928382541e-05,
|
| 2214 |
+
"loss": 0.3133,
|
| 2215 |
+
"step": 180500
|
| 2216 |
+
},
|
| 2217 |
+
{
|
| 2218 |
+
"epoch": 4.21,
|
| 2219 |
+
"learning_rate": 3.612225056777526e-05,
|
| 2220 |
+
"loss": 0.3167,
|
| 2221 |
+
"step": 181000
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 4.22,
|
| 2225 |
+
"learning_rate": 3.6083361851725104e-05,
|
| 2226 |
+
"loss": 0.3129,
|
| 2227 |
+
"step": 181500
|
| 2228 |
+
},
|
| 2229 |
+
{
|
| 2230 |
+
"epoch": 4.23,
|
| 2231 |
+
"learning_rate": 3.6044473135674956e-05,
|
| 2232 |
+
"loss": 0.3232,
|
| 2233 |
+
"step": 182000
|
| 2234 |
+
},
|
| 2235 |
+
{
|
| 2236 |
+
"epoch": 4.24,
|
| 2237 |
+
"learning_rate": 3.600558441962481e-05,
|
| 2238 |
+
"loss": 0.3161,
|
| 2239 |
+
"step": 182500
|
| 2240 |
+
},
|
| 2241 |
+
{
|
| 2242 |
+
"epoch": 4.25,
|
| 2243 |
+
"learning_rate": 3.596669570357465e-05,
|
| 2244 |
+
"loss": 0.3211,
|
| 2245 |
+
"step": 183000
|
| 2246 |
+
},
|
| 2247 |
+
{
|
| 2248 |
+
"epoch": 4.27,
|
| 2249 |
+
"learning_rate": 3.59278847649566e-05,
|
| 2250 |
+
"loss": 0.3248,
|
| 2251 |
+
"step": 183500
|
| 2252 |
+
},
|
| 2253 |
+
{
|
| 2254 |
+
"epoch": 4.28,
|
| 2255 |
+
"learning_rate": 3.588899604890645e-05,
|
| 2256 |
+
"loss": 0.3238,
|
| 2257 |
+
"step": 184000
|
| 2258 |
+
},
|
| 2259 |
+
{
|
| 2260 |
+
"epoch": 4.29,
|
| 2261 |
+
"learning_rate": 3.58501073328563e-05,
|
| 2262 |
+
"loss": 0.3209,
|
| 2263 |
+
"step": 184500
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 4.3,
|
| 2267 |
+
"learning_rate": 3.581121861680615e-05,
|
| 2268 |
+
"loss": 0.318,
|
| 2269 |
+
"step": 185000
|
| 2270 |
+
},
|
| 2271 |
+
{
|
| 2272 |
+
"epoch": 4.31,
|
| 2273 |
+
"learning_rate": 3.5772329900755994e-05,
|
| 2274 |
+
"loss": 0.3255,
|
| 2275 |
+
"step": 185500
|
| 2276 |
+
},
|
| 2277 |
+
{
|
| 2278 |
+
"epoch": 4.32,
|
| 2279 |
+
"learning_rate": 3.5733441184705846e-05,
|
| 2280 |
+
"loss": 0.3265,
|
| 2281 |
+
"step": 186000
|
| 2282 |
+
},
|
| 2283 |
+
{
|
| 2284 |
+
"epoch": 4.33,
|
| 2285 |
+
"learning_rate": 3.56945524686557e-05,
|
| 2286 |
+
"loss": 0.3235,
|
| 2287 |
+
"step": 186500
|
| 2288 |
+
},
|
| 2289 |
+
{
|
| 2290 |
+
"epoch": 4.35,
|
| 2291 |
+
"learning_rate": 3.565574153003765e-05,
|
| 2292 |
+
"loss": 0.3304,
|
| 2293 |
+
"step": 187000
|
| 2294 |
+
},
|
| 2295 |
+
{
|
| 2296 |
+
"epoch": 4.36,
|
| 2297 |
+
"learning_rate": 3.561685281398749e-05,
|
| 2298 |
+
"loss": 0.3331,
|
| 2299 |
+
"step": 187500
|
| 2300 |
+
},
|
| 2301 |
+
{
|
| 2302 |
+
"epoch": 4.37,
|
| 2303 |
+
"learning_rate": 3.557796409793734e-05,
|
| 2304 |
+
"loss": 0.3254,
|
| 2305 |
+
"step": 188000
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 4.38,
|
| 2309 |
+
"learning_rate": 3.5539075381887194e-05,
|
| 2310 |
+
"loss": 0.3305,
|
| 2311 |
+
"step": 188500
|
| 2312 |
+
},
|
| 2313 |
+
{
|
| 2314 |
+
"epoch": 4.39,
|
| 2315 |
+
"learning_rate": 3.5500186665837046e-05,
|
| 2316 |
+
"loss": 0.3264,
|
| 2317 |
+
"step": 189000
|
| 2318 |
+
},
|
| 2319 |
+
{
|
| 2320 |
+
"epoch": 4.4,
|
| 2321 |
+
"learning_rate": 3.546129794978689e-05,
|
| 2322 |
+
"loss": 0.3286,
|
| 2323 |
+
"step": 189500
|
| 2324 |
+
},
|
| 2325 |
+
{
|
| 2326 |
+
"epoch": 4.42,
|
| 2327 |
+
"learning_rate": 3.542240923373674e-05,
|
| 2328 |
+
"loss": 0.3303,
|
| 2329 |
+
"step": 190000
|
| 2330 |
+
},
|
| 2331 |
+
{
|
| 2332 |
+
"epoch": 4.43,
|
| 2333 |
+
"learning_rate": 3.5383520517686594e-05,
|
| 2334 |
+
"loss": 0.3332,
|
| 2335 |
+
"step": 190500
|
| 2336 |
+
},
|
| 2337 |
+
{
|
| 2338 |
+
"epoch": 4.44,
|
| 2339 |
+
"learning_rate": 3.534463180163644e-05,
|
| 2340 |
+
"loss": 0.3284,
|
| 2341 |
+
"step": 191000
|
| 2342 |
+
},
|
| 2343 |
+
{
|
| 2344 |
+
"epoch": 4.45,
|
| 2345 |
+
"learning_rate": 3.530574308558629e-05,
|
| 2346 |
+
"loss": 0.3297,
|
| 2347 |
+
"step": 191500
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 4.46,
|
| 2351 |
+
"learning_rate": 3.526685436953614e-05,
|
| 2352 |
+
"loss": 0.3267,
|
| 2353 |
+
"step": 192000
|
| 2354 |
+
},
|
| 2355 |
+
{
|
| 2356 |
+
"epoch": 4.47,
|
| 2357 |
+
"learning_rate": 3.5227965653485986e-05,
|
| 2358 |
+
"loss": 0.3396,
|
| 2359 |
+
"step": 192500
|
| 2360 |
+
},
|
| 2361 |
+
{
|
| 2362 |
+
"epoch": 4.49,
|
| 2363 |
+
"learning_rate": 3.518907693743583e-05,
|
| 2364 |
+
"loss": 0.3321,
|
| 2365 |
+
"step": 193000
|
| 2366 |
+
},
|
| 2367 |
+
{
|
| 2368 |
+
"epoch": 4.5,
|
| 2369 |
+
"learning_rate": 3.515018822138568e-05,
|
| 2370 |
+
"loss": 0.3411,
|
| 2371 |
+
"step": 193500
|
| 2372 |
+
},
|
| 2373 |
+
{
|
| 2374 |
+
"epoch": 4.51,
|
| 2375 |
+
"learning_rate": 3.5111299505335534e-05,
|
| 2376 |
+
"loss": 0.3356,
|
| 2377 |
+
"step": 194000
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 4.52,
|
| 2381 |
+
"learning_rate": 3.507241078928538e-05,
|
| 2382 |
+
"loss": 0.3367,
|
| 2383 |
+
"step": 194500
|
| 2384 |
+
},
|
| 2385 |
+
{
|
| 2386 |
+
"epoch": 4.53,
|
| 2387 |
+
"learning_rate": 3.503359985066733e-05,
|
| 2388 |
+
"loss": 0.3372,
|
| 2389 |
+
"step": 195000
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 4.54,
|
| 2393 |
+
"learning_rate": 3.499471113461718e-05,
|
| 2394 |
+
"loss": 0.3316,
|
| 2395 |
+
"step": 195500
|
| 2396 |
+
},
|
| 2397 |
+
{
|
| 2398 |
+
"epoch": 4.56,
|
| 2399 |
+
"learning_rate": 3.495582241856703e-05,
|
| 2400 |
+
"loss": 0.3388,
|
| 2401 |
+
"step": 196000
|
| 2402 |
+
},
|
| 2403 |
+
{
|
| 2404 |
+
"epoch": 4.57,
|
| 2405 |
+
"learning_rate": 3.4916933702516876e-05,
|
| 2406 |
+
"loss": 0.3348,
|
| 2407 |
+
"step": 196500
|
| 2408 |
+
},
|
| 2409 |
+
{
|
| 2410 |
+
"epoch": 4.58,
|
| 2411 |
+
"learning_rate": 3.487804498646673e-05,
|
| 2412 |
+
"loss": 0.3382,
|
| 2413 |
+
"step": 197000
|
| 2414 |
+
},
|
| 2415 |
+
{
|
| 2416 |
+
"epoch": 4.59,
|
| 2417 |
+
"learning_rate": 3.483915627041658e-05,
|
| 2418 |
+
"loss": 0.3376,
|
| 2419 |
+
"step": 197500
|
| 2420 |
+
},
|
| 2421 |
+
{
|
| 2422 |
+
"epoch": 4.6,
|
| 2423 |
+
"learning_rate": 3.480026755436643e-05,
|
| 2424 |
+
"loss": 0.3353,
|
| 2425 |
+
"step": 198000
|
| 2426 |
+
},
|
| 2427 |
+
{
|
| 2428 |
+
"epoch": 4.61,
|
| 2429 |
+
"learning_rate": 3.4761378838316276e-05,
|
| 2430 |
+
"loss": 0.3437,
|
| 2431 |
+
"step": 198500
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 4.63,
|
| 2435 |
+
"learning_rate": 3.472249012226613e-05,
|
| 2436 |
+
"loss": 0.339,
|
| 2437 |
+
"step": 199000
|
| 2438 |
+
},
|
| 2439 |
+
{
|
| 2440 |
+
"epoch": 4.64,
|
| 2441 |
+
"learning_rate": 3.468360140621598e-05,
|
| 2442 |
+
"loss": 0.343,
|
| 2443 |
+
"step": 199500
|
| 2444 |
+
},
|
| 2445 |
+
{
|
| 2446 |
+
"epoch": 4.65,
|
| 2447 |
+
"learning_rate": 3.464479046759792e-05,
|
| 2448 |
+
"loss": 0.3402,
|
| 2449 |
+
"step": 200000
|
| 2450 |
+
},
|
| 2451 |
+
{
|
| 2452 |
+
"epoch": 4.66,
|
| 2453 |
+
"learning_rate": 3.460590175154777e-05,
|
| 2454 |
+
"loss": 0.3476,
|
| 2455 |
+
"step": 200500
|
| 2456 |
+
},
|
| 2457 |
+
{
|
| 2458 |
+
"epoch": 4.67,
|
| 2459 |
+
"learning_rate": 3.456701303549762e-05,
|
| 2460 |
+
"loss": 0.3432,
|
| 2461 |
+
"step": 201000
|
| 2462 |
+
},
|
| 2463 |
+
{
|
| 2464 |
+
"epoch": 4.68,
|
| 2465 |
+
"learning_rate": 3.452812431944747e-05,
|
| 2466 |
+
"loss": 0.3435,
|
| 2467 |
+
"step": 201500
|
| 2468 |
+
},
|
| 2469 |
+
{
|
| 2470 |
+
"epoch": 4.7,
|
| 2471 |
+
"learning_rate": 3.448923560339732e-05,
|
| 2472 |
+
"loss": 0.3396,
|
| 2473 |
+
"step": 202000
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 4.71,
|
| 2477 |
+
"learning_rate": 3.445042466477927e-05,
|
| 2478 |
+
"loss": 0.3421,
|
| 2479 |
+
"step": 202500
|
| 2480 |
+
},
|
| 2481 |
+
{
|
| 2482 |
+
"epoch": 4.72,
|
| 2483 |
+
"learning_rate": 3.4411535948729115e-05,
|
| 2484 |
+
"loss": 0.3405,
|
| 2485 |
+
"step": 203000
|
| 2486 |
+
},
|
| 2487 |
+
{
|
| 2488 |
+
"epoch": 4.73,
|
| 2489 |
+
"learning_rate": 3.437264723267897e-05,
|
| 2490 |
+
"loss": 0.3434,
|
| 2491 |
+
"step": 203500
|
| 2492 |
+
},
|
| 2493 |
+
{
|
| 2494 |
+
"epoch": 4.74,
|
| 2495 |
+
"learning_rate": 3.433375851662882e-05,
|
| 2496 |
+
"loss": 0.3465,
|
| 2497 |
+
"step": 204000
|
| 2498 |
+
},
|
| 2499 |
+
{
|
| 2500 |
+
"epoch": 4.75,
|
| 2501 |
+
"learning_rate": 3.429486980057866e-05,
|
| 2502 |
+
"loss": 0.3459,
|
| 2503 |
+
"step": 204500
|
| 2504 |
+
},
|
| 2505 |
+
{
|
| 2506 |
+
"epoch": 4.76,
|
| 2507 |
+
"learning_rate": 3.4255981084528515e-05,
|
| 2508 |
+
"loss": 0.3458,
|
| 2509 |
+
"step": 205000
|
| 2510 |
+
},
|
| 2511 |
+
{
|
| 2512 |
+
"epoch": 4.78,
|
| 2513 |
+
"learning_rate": 3.4217092368478366e-05,
|
| 2514 |
+
"loss": 0.3486,
|
| 2515 |
+
"step": 205500
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 4.79,
|
| 2519 |
+
"learning_rate": 3.4178281429860316e-05,
|
| 2520 |
+
"loss": 0.3491,
|
| 2521 |
+
"step": 206000
|
| 2522 |
+
},
|
| 2523 |
+
{
|
| 2524 |
+
"epoch": 4.8,
|
| 2525 |
+
"learning_rate": 3.4139470491242265e-05,
|
| 2526 |
+
"loss": 0.3509,
|
| 2527 |
+
"step": 206500
|
| 2528 |
+
},
|
| 2529 |
+
{
|
| 2530 |
+
"epoch": 4.81,
|
| 2531 |
+
"learning_rate": 3.410058177519212e-05,
|
| 2532 |
+
"loss": 0.3475,
|
| 2533 |
+
"step": 207000
|
| 2534 |
+
},
|
| 2535 |
+
{
|
| 2536 |
+
"epoch": 4.82,
|
| 2537 |
+
"learning_rate": 3.406169305914196e-05,
|
| 2538 |
+
"loss": 0.3537,
|
| 2539 |
+
"step": 207500
|
| 2540 |
+
},
|
| 2541 |
+
{
|
| 2542 |
+
"epoch": 4.83,
|
| 2543 |
+
"learning_rate": 3.4022804343091806e-05,
|
| 2544 |
+
"loss": 0.3485,
|
| 2545 |
+
"step": 208000
|
| 2546 |
+
},
|
| 2547 |
+
{
|
| 2548 |
+
"epoch": 4.85,
|
| 2549 |
+
"learning_rate": 3.398391562704166e-05,
|
| 2550 |
+
"loss": 0.3402,
|
| 2551 |
+
"step": 208500
|
| 2552 |
+
},
|
| 2553 |
+
{
|
| 2554 |
+
"epoch": 4.86,
|
| 2555 |
+
"learning_rate": 3.394502691099151e-05,
|
| 2556 |
+
"loss": 0.3582,
|
| 2557 |
+
"step": 209000
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 4.87,
|
| 2561 |
+
"learning_rate": 3.390629374980556e-05,
|
| 2562 |
+
"loss": 0.3509,
|
| 2563 |
+
"step": 209500
|
| 2564 |
+
},
|
| 2565 |
+
{
|
| 2566 |
+
"epoch": 4.88,
|
| 2567 |
+
"learning_rate": 3.386740503375541e-05,
|
| 2568 |
+
"loss": 0.3431,
|
| 2569 |
+
"step": 210000
|
| 2570 |
+
},
|
| 2571 |
+
{
|
| 2572 |
+
"epoch": 4.89,
|
| 2573 |
+
"learning_rate": 3.3828516317705254e-05,
|
| 2574 |
+
"loss": 0.3494,
|
| 2575 |
+
"step": 210500
|
| 2576 |
+
},
|
| 2577 |
+
{
|
| 2578 |
+
"epoch": 4.9,
|
| 2579 |
+
"learning_rate": 3.37897053790872e-05,
|
| 2580 |
+
"loss": 0.3514,
|
| 2581 |
+
"step": 211000
|
| 2582 |
+
},
|
| 2583 |
+
{
|
| 2584 |
+
"epoch": 4.92,
|
| 2585 |
+
"learning_rate": 3.3750816663037055e-05,
|
| 2586 |
+
"loss": 0.3555,
|
| 2587 |
+
"step": 211500
|
| 2588 |
+
},
|
| 2589 |
+
{
|
| 2590 |
+
"epoch": 4.93,
|
| 2591 |
+
"learning_rate": 3.3711927946986906e-05,
|
| 2592 |
+
"loss": 0.3513,
|
| 2593 |
+
"step": 212000
|
| 2594 |
+
},
|
| 2595 |
+
{
|
| 2596 |
+
"epoch": 4.94,
|
| 2597 |
+
"learning_rate": 3.367303923093675e-05,
|
| 2598 |
+
"loss": 0.3466,
|
| 2599 |
+
"step": 212500
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 4.95,
|
| 2603 |
+
"learning_rate": 3.36341505148866e-05,
|
| 2604 |
+
"loss": 0.3509,
|
| 2605 |
+
"step": 213000
|
| 2606 |
+
},
|
| 2607 |
+
{
|
| 2608 |
+
"epoch": 4.96,
|
| 2609 |
+
"learning_rate": 3.3595261798836454e-05,
|
| 2610 |
+
"loss": 0.347,
|
| 2611 |
+
"step": 213500
|
| 2612 |
+
},
|
| 2613 |
+
{
|
| 2614 |
+
"epoch": 4.97,
|
| 2615 |
+
"learning_rate": 3.35563730827863e-05,
|
| 2616 |
+
"loss": 0.3511,
|
| 2617 |
+
"step": 214000
|
| 2618 |
+
},
|
| 2619 |
+
{
|
| 2620 |
+
"epoch": 4.99,
|
| 2621 |
+
"learning_rate": 3.351748436673615e-05,
|
| 2622 |
+
"loss": 0.3524,
|
| 2623 |
+
"step": 214500
|
| 2624 |
+
},
|
| 2625 |
+
{
|
| 2626 |
+
"epoch": 5.0,
|
| 2627 |
+
"learning_rate": 3.3478595650686e-05,
|
| 2628 |
+
"loss": 0.3458,
|
| 2629 |
+
"step": 215000
|
| 2630 |
+
},
|
| 2631 |
+
{
|
| 2632 |
+
"epoch": 5.0,
|
| 2633 |
+
"eval_bleu": 59.1341,
|
| 2634 |
+
"eval_gen_len": 15.7016,
|
| 2635 |
+
"eval_loss": 0.7214592099189758,
|
| 2636 |
+
"eval_runtime": 6616.423,
|
| 2637 |
+
"eval_samples_per_second": 12.985,
|
| 2638 |
+
"eval_steps_per_second": 1.623,
|
| 2639 |
+
"step": 215120
|
| 2640 |
+
},
|
| 2641 |
+
{
|
| 2642 |
+
"epoch": 5.01,
|
| 2643 |
+
"learning_rate": 3.343970693463585e-05,
|
| 2644 |
+
"loss": 0.2535,
|
| 2645 |
+
"step": 215500
|
| 2646 |
+
},
|
| 2647 |
+
{
|
| 2648 |
+
"epoch": 5.02,
|
| 2649 |
+
"learning_rate": 3.3400895996017796e-05,
|
| 2650 |
+
"loss": 0.2198,
|
| 2651 |
+
"step": 216000
|
| 2652 |
+
},
|
| 2653 |
+
{
|
| 2654 |
+
"epoch": 5.03,
|
| 2655 |
+
"learning_rate": 3.336200727996764e-05,
|
| 2656 |
+
"loss": 0.2281,
|
| 2657 |
+
"step": 216500
|
| 2658 |
+
},
|
| 2659 |
+
{
|
| 2660 |
+
"epoch": 5.04,
|
| 2661 |
+
"learning_rate": 3.332311856391749e-05,
|
| 2662 |
+
"loss": 0.2251,
|
| 2663 |
+
"step": 217000
|
| 2664 |
+
},
|
| 2665 |
+
{
|
| 2666 |
+
"epoch": 5.06,
|
| 2667 |
+
"learning_rate": 3.3284229847867344e-05,
|
| 2668 |
+
"loss": 0.2282,
|
| 2669 |
+
"step": 217500
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 5.07,
|
| 2673 |
+
"learning_rate": 3.3245341131817196e-05,
|
| 2674 |
+
"loss": 0.2293,
|
| 2675 |
+
"step": 218000
|
| 2676 |
+
},
|
| 2677 |
+
{
|
| 2678 |
+
"epoch": 5.08,
|
| 2679 |
+
"learning_rate": 3.320645241576704e-05,
|
| 2680 |
+
"loss": 0.2246,
|
| 2681 |
+
"step": 218500
|
| 2682 |
+
},
|
| 2683 |
+
{
|
| 2684 |
+
"epoch": 5.09,
|
| 2685 |
+
"learning_rate": 3.316764147714899e-05,
|
| 2686 |
+
"loss": 0.2282,
|
| 2687 |
+
"step": 219000
|
| 2688 |
+
},
|
| 2689 |
+
{
|
| 2690 |
+
"epoch": 5.1,
|
| 2691 |
+
"learning_rate": 3.312875276109884e-05,
|
| 2692 |
+
"loss": 0.2311,
|
| 2693 |
+
"step": 219500
|
| 2694 |
+
},
|
| 2695 |
+
{
|
| 2696 |
+
"epoch": 5.11,
|
| 2697 |
+
"learning_rate": 3.308986404504869e-05,
|
| 2698 |
+
"loss": 0.239,
|
| 2699 |
+
"step": 220000
|
| 2700 |
+
},
|
| 2701 |
+
{
|
| 2702 |
+
"epoch": 5.13,
|
| 2703 |
+
"learning_rate": 3.305113088386274e-05,
|
| 2704 |
+
"loss": 0.2323,
|
| 2705 |
+
"step": 220500
|
| 2706 |
+
},
|
| 2707 |
+
{
|
| 2708 |
+
"epoch": 5.14,
|
| 2709 |
+
"learning_rate": 3.301224216781259e-05,
|
| 2710 |
+
"loss": 0.2317,
|
| 2711 |
+
"step": 221000
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 5.15,
|
| 2715 |
+
"learning_rate": 3.297343122919454e-05,
|
| 2716 |
+
"loss": 0.2324,
|
| 2717 |
+
"step": 221500
|
| 2718 |
+
},
|
| 2719 |
+
{
|
| 2720 |
+
"epoch": 5.16,
|
| 2721 |
+
"learning_rate": 3.293454251314439e-05,
|
| 2722 |
+
"loss": 0.2376,
|
| 2723 |
+
"step": 222000
|
| 2724 |
+
},
|
| 2725 |
+
{
|
| 2726 |
+
"epoch": 5.17,
|
| 2727 |
+
"learning_rate": 3.289565379709424e-05,
|
| 2728 |
+
"loss": 0.2441,
|
| 2729 |
+
"step": 222500
|
| 2730 |
+
},
|
| 2731 |
+
{
|
| 2732 |
+
"epoch": 5.18,
|
| 2733 |
+
"learning_rate": 3.285676508104409e-05,
|
| 2734 |
+
"loss": 0.24,
|
| 2735 |
+
"step": 223000
|
| 2736 |
+
},
|
| 2737 |
+
{
|
| 2738 |
+
"epoch": 5.19,
|
| 2739 |
+
"learning_rate": 3.2817876364993935e-05,
|
| 2740 |
+
"loss": 0.2411,
|
| 2741 |
+
"step": 223500
|
| 2742 |
+
},
|
| 2743 |
+
{
|
| 2744 |
+
"epoch": 5.21,
|
| 2745 |
+
"learning_rate": 3.277898764894378e-05,
|
| 2746 |
+
"loss": 0.2454,
|
| 2747 |
+
"step": 224000
|
| 2748 |
+
},
|
| 2749 |
+
{
|
| 2750 |
+
"epoch": 5.22,
|
| 2751 |
+
"learning_rate": 3.274009893289363e-05,
|
| 2752 |
+
"loss": 0.2455,
|
| 2753 |
+
"step": 224500
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 5.23,
|
| 2757 |
+
"learning_rate": 3.270121021684348e-05,
|
| 2758 |
+
"loss": 0.2429,
|
| 2759 |
+
"step": 225000
|
| 2760 |
+
},
|
| 2761 |
+
{
|
| 2762 |
+
"epoch": 5.24,
|
| 2763 |
+
"learning_rate": 3.266232150079333e-05,
|
| 2764 |
+
"loss": 0.2461,
|
| 2765 |
+
"step": 225500
|
| 2766 |
+
},
|
| 2767 |
+
{
|
| 2768 |
+
"epoch": 5.25,
|
| 2769 |
+
"learning_rate": 3.262343278474318e-05,
|
| 2770 |
+
"loss": 0.2414,
|
| 2771 |
+
"step": 226000
|
| 2772 |
+
},
|
| 2773 |
+
{
|
| 2774 |
+
"epoch": 5.26,
|
| 2775 |
+
"learning_rate": 3.258454406869303e-05,
|
| 2776 |
+
"loss": 0.253,
|
| 2777 |
+
"step": 226500
|
| 2778 |
+
},
|
| 2779 |
+
{
|
| 2780 |
+
"epoch": 5.28,
|
| 2781 |
+
"learning_rate": 3.254573313007498e-05,
|
| 2782 |
+
"loss": 0.2481,
|
| 2783 |
+
"step": 227000
|
| 2784 |
+
},
|
| 2785 |
+
{
|
| 2786 |
+
"epoch": 5.29,
|
| 2787 |
+
"learning_rate": 3.250684441402483e-05,
|
| 2788 |
+
"loss": 0.2506,
|
| 2789 |
+
"step": 227500
|
| 2790 |
+
},
|
| 2791 |
+
{
|
| 2792 |
+
"epoch": 5.3,
|
| 2793 |
+
"learning_rate": 3.2467955697974676e-05,
|
| 2794 |
+
"loss": 0.249,
|
| 2795 |
+
"step": 228000
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 5.31,
|
| 2799 |
+
"learning_rate": 3.242906698192453e-05,
|
| 2800 |
+
"loss": 0.2448,
|
| 2801 |
+
"step": 228500
|
| 2802 |
+
},
|
| 2803 |
+
{
|
| 2804 |
+
"epoch": 5.32,
|
| 2805 |
+
"learning_rate": 3.239017826587438e-05,
|
| 2806 |
+
"loss": 0.2472,
|
| 2807 |
+
"step": 229000
|
| 2808 |
+
},
|
| 2809 |
+
{
|
| 2810 |
+
"epoch": 5.33,
|
| 2811 |
+
"learning_rate": 3.2351289549824224e-05,
|
| 2812 |
+
"loss": 0.2516,
|
| 2813 |
+
"step": 229500
|
| 2814 |
+
},
|
| 2815 |
+
{
|
| 2816 |
+
"epoch": 5.35,
|
| 2817 |
+
"learning_rate": 3.2312400833774075e-05,
|
| 2818 |
+
"loss": 0.2564,
|
| 2819 |
+
"step": 230000
|
| 2820 |
+
},
|
| 2821 |
+
{
|
| 2822 |
+
"epoch": 5.36,
|
| 2823 |
+
"learning_rate": 3.227351211772393e-05,
|
| 2824 |
+
"loss": 0.248,
|
| 2825 |
+
"step": 230500
|
| 2826 |
+
},
|
| 2827 |
+
{
|
| 2828 |
+
"epoch": 5.37,
|
| 2829 |
+
"learning_rate": 3.223462340167377e-05,
|
| 2830 |
+
"loss": 0.2533,
|
| 2831 |
+
"step": 231000
|
| 2832 |
+
},
|
| 2833 |
+
{
|
| 2834 |
+
"epoch": 5.38,
|
| 2835 |
+
"learning_rate": 3.2195734685623616e-05,
|
| 2836 |
+
"loss": 0.2537,
|
| 2837 |
+
"step": 231500
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 5.39,
|
| 2841 |
+
"learning_rate": 3.215684596957347e-05,
|
| 2842 |
+
"loss": 0.2544,
|
| 2843 |
+
"step": 232000
|
| 2844 |
+
},
|
| 2845 |
+
{
|
| 2846 |
+
"epoch": 5.4,
|
| 2847 |
+
"learning_rate": 3.211795725352332e-05,
|
| 2848 |
+
"loss": 0.2543,
|
| 2849 |
+
"step": 232500
|
| 2850 |
+
},
|
| 2851 |
+
{
|
| 2852 |
+
"epoch": 5.42,
|
| 2853 |
+
"learning_rate": 3.2079068537473164e-05,
|
| 2854 |
+
"loss": 0.2559,
|
| 2855 |
+
"step": 233000
|
| 2856 |
+
},
|
| 2857 |
+
{
|
| 2858 |
+
"epoch": 5.43,
|
| 2859 |
+
"learning_rate": 3.2040179821423016e-05,
|
| 2860 |
+
"loss": 0.2496,
|
| 2861 |
+
"step": 233500
|
| 2862 |
+
},
|
| 2863 |
+
{
|
| 2864 |
+
"epoch": 5.44,
|
| 2865 |
+
"learning_rate": 3.2001368882804965e-05,
|
| 2866 |
+
"loss": 0.2496,
|
| 2867 |
+
"step": 234000
|
| 2868 |
+
},
|
| 2869 |
+
{
|
| 2870 |
+
"epoch": 5.45,
|
| 2871 |
+
"learning_rate": 3.196248016675482e-05,
|
| 2872 |
+
"loss": 0.2536,
|
| 2873 |
+
"step": 234500
|
| 2874 |
+
},
|
| 2875 |
+
{
|
| 2876 |
+
"epoch": 5.46,
|
| 2877 |
+
"learning_rate": 3.192359145070467e-05,
|
| 2878 |
+
"loss": 0.256,
|
| 2879 |
+
"step": 235000
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 5.47,
|
| 2883 |
+
"learning_rate": 3.188470273465451e-05,
|
| 2884 |
+
"loss": 0.2588,
|
| 2885 |
+
"step": 235500
|
| 2886 |
+
},
|
| 2887 |
+
{
|
| 2888 |
+
"epoch": 5.49,
|
| 2889 |
+
"learning_rate": 3.1845814018604365e-05,
|
| 2890 |
+
"loss": 0.2521,
|
| 2891 |
+
"step": 236000
|
| 2892 |
+
},
|
| 2893 |
+
{
|
| 2894 |
+
"epoch": 5.5,
|
| 2895 |
+
"learning_rate": 3.1806925302554216e-05,
|
| 2896 |
+
"loss": 0.2596,
|
| 2897 |
+
"step": 236500
|
| 2898 |
+
},
|
| 2899 |
+
{
|
| 2900 |
+
"epoch": 5.51,
|
| 2901 |
+
"learning_rate": 3.176803658650406e-05,
|
| 2902 |
+
"loss": 0.2565,
|
| 2903 |
+
"step": 237000
|
| 2904 |
+
},
|
| 2905 |
+
{
|
| 2906 |
+
"epoch": 5.52,
|
| 2907 |
+
"learning_rate": 3.172914787045391e-05,
|
| 2908 |
+
"loss": 0.2632,
|
| 2909 |
+
"step": 237500
|
| 2910 |
+
},
|
| 2911 |
+
{
|
| 2912 |
+
"epoch": 5.53,
|
| 2913 |
+
"learning_rate": 3.169033693183586e-05,
|
| 2914 |
+
"loss": 0.2642,
|
| 2915 |
+
"step": 238000
|
| 2916 |
+
},
|
| 2917 |
+
{
|
| 2918 |
+
"epoch": 5.54,
|
| 2919 |
+
"learning_rate": 3.1651448215785714e-05,
|
| 2920 |
+
"loss": 0.261,
|
| 2921 |
+
"step": 238500
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 5.56,
|
| 2925 |
+
"learning_rate": 3.1612637277167657e-05,
|
| 2926 |
+
"loss": 0.2607,
|
| 2927 |
+
"step": 239000
|
| 2928 |
+
},
|
| 2929 |
+
{
|
| 2930 |
+
"epoch": 5.57,
|
| 2931 |
+
"learning_rate": 3.157374856111751e-05,
|
| 2932 |
+
"loss": 0.2626,
|
| 2933 |
+
"step": 239500
|
| 2934 |
+
},
|
| 2935 |
+
{
|
| 2936 |
+
"epoch": 5.58,
|
| 2937 |
+
"learning_rate": 3.153485984506735e-05,
|
| 2938 |
+
"loss": 0.2561,
|
| 2939 |
+
"step": 240000
|
| 2940 |
+
},
|
| 2941 |
+
{
|
| 2942 |
+
"epoch": 5.59,
|
| 2943 |
+
"learning_rate": 3.1495971129017204e-05,
|
| 2944 |
+
"loss": 0.258,
|
| 2945 |
+
"step": 240500
|
| 2946 |
+
},
|
| 2947 |
+
{
|
| 2948 |
+
"epoch": 5.6,
|
| 2949 |
+
"learning_rate": 3.1457082412967056e-05,
|
| 2950 |
+
"loss": 0.2593,
|
| 2951 |
+
"step": 241000
|
| 2952 |
+
},
|
| 2953 |
+
{
|
| 2954 |
+
"epoch": 5.61,
|
| 2955 |
+
"learning_rate": 3.14181936969169e-05,
|
| 2956 |
+
"loss": 0.265,
|
| 2957 |
+
"step": 241500
|
| 2958 |
+
},
|
| 2959 |
+
{
|
| 2960 |
+
"epoch": 5.62,
|
| 2961 |
+
"learning_rate": 3.137930498086675e-05,
|
| 2962 |
+
"loss": 0.2566,
|
| 2963 |
+
"step": 242000
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 5.64,
|
| 2967 |
+
"learning_rate": 3.1340416264816604e-05,
|
| 2968 |
+
"loss": 0.2615,
|
| 2969 |
+
"step": 242500
|
| 2970 |
+
},
|
| 2971 |
+
{
|
| 2972 |
+
"epoch": 5.65,
|
| 2973 |
+
"learning_rate": 3.130160532619855e-05,
|
| 2974 |
+
"loss": 0.2621,
|
| 2975 |
+
"step": 243000
|
| 2976 |
+
},
|
| 2977 |
+
{
|
| 2978 |
+
"epoch": 5.66,
|
| 2979 |
+
"learning_rate": 3.12627943875805e-05,
|
| 2980 |
+
"loss": 0.2634,
|
| 2981 |
+
"step": 243500
|
| 2982 |
+
},
|
| 2983 |
+
{
|
| 2984 |
+
"epoch": 5.67,
|
| 2985 |
+
"learning_rate": 3.1223905671530355e-05,
|
| 2986 |
+
"loss": 0.2635,
|
| 2987 |
+
"step": 244000
|
| 2988 |
+
},
|
| 2989 |
+
{
|
| 2990 |
+
"epoch": 5.68,
|
| 2991 |
+
"learning_rate": 3.11850169554802e-05,
|
| 2992 |
+
"loss": 0.2719,
|
| 2993 |
+
"step": 244500
|
| 2994 |
+
},
|
| 2995 |
+
{
|
| 2996 |
+
"epoch": 5.69,
|
| 2997 |
+
"learning_rate": 3.114620601686215e-05,
|
| 2998 |
+
"loss": 0.2657,
|
| 2999 |
+
"step": 245000
|
| 3000 |
+
},
|
| 3001 |
+
{
|
| 3002 |
+
"epoch": 5.71,
|
| 3003 |
+
"learning_rate": 3.11073950782441e-05,
|
| 3004 |
+
"loss": 0.2674,
|
| 3005 |
+
"step": 245500
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 5.72,
|
| 3009 |
+
"learning_rate": 3.106850636219395e-05,
|
| 3010 |
+
"loss": 0.2658,
|
| 3011 |
+
"step": 246000
|
| 3012 |
+
},
|
| 3013 |
+
{
|
| 3014 |
+
"epoch": 5.73,
|
| 3015 |
+
"learning_rate": 3.1029617646143795e-05,
|
| 3016 |
+
"loss": 0.2711,
|
| 3017 |
+
"step": 246500
|
| 3018 |
+
},
|
| 3019 |
+
{
|
| 3020 |
+
"epoch": 5.74,
|
| 3021 |
+
"learning_rate": 3.0990728930093646e-05,
|
| 3022 |
+
"loss": 0.2726,
|
| 3023 |
+
"step": 247000
|
| 3024 |
+
},
|
| 3025 |
+
{
|
| 3026 |
+
"epoch": 5.75,
|
| 3027 |
+
"learning_rate": 3.095184021404349e-05,
|
| 3028 |
+
"loss": 0.2653,
|
| 3029 |
+
"step": 247500
|
| 3030 |
+
},
|
| 3031 |
+
{
|
| 3032 |
+
"epoch": 5.76,
|
| 3033 |
+
"learning_rate": 3.091295149799334e-05,
|
| 3034 |
+
"loss": 0.2695,
|
| 3035 |
+
"step": 248000
|
| 3036 |
+
},
|
| 3037 |
+
{
|
| 3038 |
+
"epoch": 5.78,
|
| 3039 |
+
"learning_rate": 3.0874062781943194e-05,
|
| 3040 |
+
"loss": 0.276,
|
| 3041 |
+
"step": 248500
|
| 3042 |
+
},
|
| 3043 |
+
{
|
| 3044 |
+
"epoch": 5.79,
|
| 3045 |
+
"learning_rate": 3.083517406589304e-05,
|
| 3046 |
+
"loss": 0.2718,
|
| 3047 |
+
"step": 249000
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 5.8,
|
| 3051 |
+
"learning_rate": 3.079628534984289e-05,
|
| 3052 |
+
"loss": 0.2687,
|
| 3053 |
+
"step": 249500
|
| 3054 |
+
},
|
| 3055 |
+
{
|
| 3056 |
+
"epoch": 5.81,
|
| 3057 |
+
"learning_rate": 3.075739663379274e-05,
|
| 3058 |
+
"loss": 0.2674,
|
| 3059 |
+
"step": 250000
|
| 3060 |
+
},
|
| 3061 |
+
{
|
| 3062 |
+
"epoch": 5.82,
|
| 3063 |
+
"learning_rate": 3.071850791774259e-05,
|
| 3064 |
+
"loss": 0.2713,
|
| 3065 |
+
"step": 250500
|
| 3066 |
+
},
|
| 3067 |
+
{
|
| 3068 |
+
"epoch": 5.83,
|
| 3069 |
+
"learning_rate": 3.067961920169244e-05,
|
| 3070 |
+
"loss": 0.2671,
|
| 3071 |
+
"step": 251000
|
| 3072 |
+
},
|
| 3073 |
+
{
|
| 3074 |
+
"epoch": 5.85,
|
| 3075 |
+
"learning_rate": 3.064073048564229e-05,
|
| 3076 |
+
"loss": 0.2721,
|
| 3077 |
+
"step": 251500
|
| 3078 |
+
},
|
| 3079 |
+
{
|
| 3080 |
+
"epoch": 5.86,
|
| 3081 |
+
"learning_rate": 3.0601841769592135e-05,
|
| 3082 |
+
"loss": 0.2647,
|
| 3083 |
+
"step": 252000
|
| 3084 |
+
},
|
| 3085 |
+
{
|
| 3086 |
+
"epoch": 5.87,
|
| 3087 |
+
"learning_rate": 3.0562953053541986e-05,
|
| 3088 |
+
"loss": 0.2723,
|
| 3089 |
+
"step": 252500
|
| 3090 |
+
},
|
| 3091 |
+
{
|
| 3092 |
+
"epoch": 5.88,
|
| 3093 |
+
"learning_rate": 3.052406433749184e-05,
|
| 3094 |
+
"loss": 0.2722,
|
| 3095 |
+
"step": 253000
|
| 3096 |
+
},
|
| 3097 |
+
{
|
| 3098 |
+
"epoch": 5.89,
|
| 3099 |
+
"learning_rate": 3.0485175621441686e-05,
|
| 3100 |
+
"loss": 0.2744,
|
| 3101 |
+
"step": 253500
|
| 3102 |
+
},
|
| 3103 |
+
{
|
| 3104 |
+
"epoch": 5.9,
|
| 3105 |
+
"learning_rate": 3.0446286905391537e-05,
|
| 3106 |
+
"loss": 0.2708,
|
| 3107 |
+
"step": 254000
|
| 3108 |
+
},
|
| 3109 |
+
{
|
| 3110 |
+
"epoch": 5.92,
|
| 3111 |
+
"learning_rate": 3.040739818934138e-05,
|
| 3112 |
+
"loss": 0.271,
|
| 3113 |
+
"step": 254500
|
| 3114 |
+
},
|
| 3115 |
+
{
|
| 3116 |
+
"epoch": 5.93,
|
| 3117 |
+
"learning_rate": 3.036850947329123e-05,
|
| 3118 |
+
"loss": 0.2686,
|
| 3119 |
+
"step": 255000
|
| 3120 |
+
},
|
| 3121 |
+
{
|
| 3122 |
+
"epoch": 5.94,
|
| 3123 |
+
"learning_rate": 3.0329620757241078e-05,
|
| 3124 |
+
"loss": 0.2703,
|
| 3125 |
+
"step": 255500
|
| 3126 |
+
},
|
| 3127 |
+
{
|
| 3128 |
+
"epoch": 5.95,
|
| 3129 |
+
"learning_rate": 3.0290732041190926e-05,
|
| 3130 |
+
"loss": 0.2786,
|
| 3131 |
+
"step": 256000
|
| 3132 |
+
},
|
| 3133 |
+
{
|
| 3134 |
+
"epoch": 5.96,
|
| 3135 |
+
"learning_rate": 3.0251843325140778e-05,
|
| 3136 |
+
"loss": 0.2687,
|
| 3137 |
+
"step": 256500
|
| 3138 |
+
},
|
| 3139 |
+
{
|
| 3140 |
+
"epoch": 5.97,
|
| 3141 |
+
"learning_rate": 3.0212954609090626e-05,
|
| 3142 |
+
"loss": 0.2727,
|
| 3143 |
+
"step": 257000
|
| 3144 |
+
},
|
| 3145 |
+
{
|
| 3146 |
+
"epoch": 5.99,
|
| 3147 |
+
"learning_rate": 3.0174065893040478e-05,
|
| 3148 |
+
"loss": 0.2726,
|
| 3149 |
+
"step": 257500
|
| 3150 |
+
},
|
| 3151 |
+
{
|
| 3152 |
+
"epoch": 6.0,
|
| 3153 |
+
"learning_rate": 3.0135177176990326e-05,
|
| 3154 |
+
"loss": 0.2702,
|
| 3155 |
+
"step": 258000
|
| 3156 |
+
},
|
| 3157 |
+
{
|
| 3158 |
+
"epoch": 6.0,
|
| 3159 |
+
"eval_bleu": 59.0721,
|
| 3160 |
+
"eval_gen_len": 15.6936,
|
| 3161 |
+
"eval_loss": 0.8013012409210205,
|
| 3162 |
+
"eval_runtime": 6628.012,
|
| 3163 |
+
"eval_samples_per_second": 12.962,
|
| 3164 |
+
"eval_steps_per_second": 1.62,
|
| 3165 |
+
"step": 258144
|
| 3166 |
+
}
|
| 3167 |
+
],
|
| 3168 |
+
"logging_steps": 500,
|
| 3169 |
+
"max_steps": 645360,
|
| 3170 |
+
"num_input_tokens_seen": 0,
|
| 3171 |
+
"num_train_epochs": 15,
|
| 3172 |
+
"save_steps": 500,
|
| 3173 |
+
"total_flos": 8.950904749025133e+18,
|
| 3174 |
+
"train_batch_size": 2,
|
| 3175 |
+
"trial_name": null,
|
| 3176 |
+
"trial_params": null
|
| 3177 |
+
}
|
checkpoint-258144/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4694a20062ed70ba8adc7c9ea1e09ea766481ac4f0f03605243799bfb9894d0a
|
| 3 |
+
size 4856
|
config.json
ADDED
|
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/mbart-large-50-many-to-many-mmt",
|
| 3 |
+
"_num_labels": 3,
|
| 4 |
+
"activation_dropout": 0.0,
|
| 5 |
+
"activation_function": "relu",
|
| 6 |
+
"add_bias_logits": false,
|
| 7 |
+
"add_final_layer_norm": true,
|
| 8 |
+
"architectures": [
|
| 9 |
+
"MBartForConditionalGeneration"
|
| 10 |
+
],
|
| 11 |
+
"attention_dropout": 0.0,
|
| 12 |
+
"bos_token_id": 0,
|
| 13 |
+
"classif_dropout": 0.0,
|
| 14 |
+
"classifier_dropout": 0.0,
|
| 15 |
+
"d_model": 1024,
|
| 16 |
+
"decoder_attention_heads": 16,
|
| 17 |
+
"decoder_ffn_dim": 4096,
|
| 18 |
+
"decoder_layerdrop": 0.0,
|
| 19 |
+
"decoder_layers": 12,
|
| 20 |
+
"decoder_start_token_id": 2,
|
| 21 |
+
"dropout": 0.1,
|
| 22 |
+
"early_stopping": true,
|
| 23 |
+
"encoder_attention_heads": 16,
|
| 24 |
+
"encoder_ffn_dim": 4096,
|
| 25 |
+
"encoder_layerdrop": 0.0,
|
| 26 |
+
"encoder_layers": 12,
|
| 27 |
+
"eos_token_id": 2,
|
| 28 |
+
"forced_bos_token_id": 250014,
|
| 29 |
+
"forced_eos_token_id": 2,
|
| 30 |
+
"gradient_checkpointing": false,
|
| 31 |
+
"id2label": {
|
| 32 |
+
"0": "LABEL_0",
|
| 33 |
+
"1": "LABEL_1",
|
| 34 |
+
"2": "LABEL_2"
|
| 35 |
+
},
|
| 36 |
+
"init_std": 0.02,
|
| 37 |
+
"is_encoder_decoder": true,
|
| 38 |
+
"label2id": {
|
| 39 |
+
"LABEL_0": 0,
|
| 40 |
+
"LABEL_1": 1,
|
| 41 |
+
"LABEL_2": 2
|
| 42 |
+
},
|
| 43 |
+
"max_length": 200,
|
| 44 |
+
"max_position_embeddings": 1024,
|
| 45 |
+
"model_type": "mbart",
|
| 46 |
+
"normalize_before": true,
|
| 47 |
+
"normalize_embedding": true,
|
| 48 |
+
"num_beams": 5,
|
| 49 |
+
"num_hidden_layers": 12,
|
| 50 |
+
"output_past": true,
|
| 51 |
+
"pad_token_id": 1,
|
| 52 |
+
"scale_embedding": true,
|
| 53 |
+
"static_position_embeddings": false,
|
| 54 |
+
"tokenizer_class": "MBart50Tokenizer",
|
| 55 |
+
"torch_dtype": "float32",
|
| 56 |
+
"transformers_version": "4.36.2",
|
| 57 |
+
"use_cache": true,
|
| 58 |
+
"vocab_size": 250054
|
| 59 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 6.0,
|
| 3 |
+
"eval_bleu": 58.837,
|
| 4 |
+
"eval_gen_len": 15.7689,
|
| 5 |
+
"eval_loss": 0.6568061113357544,
|
| 6 |
+
"eval_runtime": 6763.7413,
|
| 7 |
+
"eval_samples": 85914,
|
| 8 |
+
"eval_samples_per_second": 12.702,
|
| 9 |
+
"eval_steps_per_second": 1.588
|
| 10 |
+
}
|