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 +74 -0
- all_results.json +22 -0
- checkpoint-141839/config.json +59 -0
- checkpoint-141839/generation_config.json +12 -0
- checkpoint-141839/model.safetensors +3 -0
- checkpoint-141839/optimizer.pt +3 -0
- checkpoint-141839/rng_state_0.pth +3 -0
- checkpoint-141839/rng_state_1.pth +3 -0
- checkpoint-141839/rng_state_2.pth +3 -0
- checkpoint-141839/rng_state_3.pth +3 -0
- checkpoint-141839/scheduler.pt +3 -0
- checkpoint-141839/sentencepiece.bpe.model +3 -0
- checkpoint-141839/special_tokens_map.json +69 -0
- checkpoint-141839/tokenizer.json +3 -0
- checkpoint-141839/tokenizer_config.json +528 -0
- checkpoint-141839/trainer_state.json +1737 -0
- checkpoint-141839/training_args.bin +3 -0
- checkpoint-283678/config.json +59 -0
- checkpoint-283678/generation_config.json +12 -0
- checkpoint-283678/model.safetensors +3 -0
- checkpoint-283678/optimizer.pt +3 -0
- checkpoint-283678/rng_state_0.pth +3 -0
- checkpoint-283678/rng_state_1.pth +3 -0
- checkpoint-283678/rng_state_2.pth +3 -0
- checkpoint-283678/rng_state_3.pth +3 -0
- checkpoint-283678/scheduler.pt +3 -0
- checkpoint-283678/sentencepiece.bpe.model +3 -0
- checkpoint-283678/special_tokens_map.json +69 -0
- checkpoint-283678/tokenizer.json +3 -0
- checkpoint-283678/tokenizer_config.json +528 -0
- checkpoint-283678/trainer_state.json +3461 -0
- checkpoint-283678/training_args.bin +3 -0
- checkpoint-354597/config.json +59 -0
- checkpoint-354597/generation_config.json +12 -0
- checkpoint-354597/model.safetensors +3 -0
- checkpoint-354597/optimizer.pt +3 -0
- checkpoint-354597/rng_state_0.pth +3 -0
- checkpoint-354597/rng_state_1.pth +3 -0
- checkpoint-354597/rng_state_2.pth +3 -0
- checkpoint-354597/rng_state_3.pth +3 -0
- checkpoint-354597/scheduler.pt +3 -0
- checkpoint-354597/sentencepiece.bpe.model +3 -0
- checkpoint-354597/special_tokens_map.json +69 -0
- checkpoint-354597/tokenizer.json +3 -0
- checkpoint-354597/tokenizer_config.json +528 -0
- checkpoint-354597/trainer_state.json +4323 -0
- checkpoint-354597/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-141839/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 37 |
+
checkpoint-283678/tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
| 38 |
+
checkpoint-354597/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 +1,74 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- zh
|
| 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: zhko_mbartLarge_100p_run1
|
| 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 |
+
# zhko_mbartLarge_100p_run1
|
| 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: 0.8988
|
| 23 |
+
- Bleu: 42.7907
|
| 24 |
+
- Gen Len: 13.7941
|
| 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_ratio: 0.01
|
| 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 |
+
| 0.9719 | 1.0 | 70919 | 0.9435 | 40.8492 | 13.8028 |
|
| 63 |
+
| 0.7537 | 2.0 | 141839 | 0.8988 | 42.7907 | 13.7941 |
|
| 64 |
+
| 0.5973 | 3.0 | 212758 | 0.9143 | 43.4697 | 13.6556 |
|
| 65 |
+
| 0.4873 | 4.0 | 283678 | 0.9758 | 43.6153 | 13.6544 |
|
| 66 |
+
| 0.388 | 5.0 | 354597 | 1.0400 | 43.3763 | 13.6067 |
|
| 67 |
+
|
| 68 |
+
|
| 69 |
+
### Framework versions
|
| 70 |
+
|
| 71 |
+
- Transformers 4.35.2
|
| 72 |
+
- Pytorch 2.1.1+cu121
|
| 73 |
+
- Datasets 2.15.0
|
| 74 |
+
- Tokenizers 0.15.0
|
all_results.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"eval_bleu": 42.7907,
|
| 4 |
+
"eval_gen_len": 13.7941,
|
| 5 |
+
"eval_loss": 0.8987888693809509,
|
| 6 |
+
"eval_runtime": 10014.5277,
|
| 7 |
+
"eval_samples": 141838,
|
| 8 |
+
"eval_samples_per_second": 14.163,
|
| 9 |
+
"eval_steps_per_second": 1.77,
|
| 10 |
+
"predict_bleu": 42.8094,
|
| 11 |
+
"predict_gen_len": 13.7545,
|
| 12 |
+
"predict_loss": 0.89827561378479,
|
| 13 |
+
"predict_runtime": 9974.8511,
|
| 14 |
+
"predict_samples": 141838,
|
| 15 |
+
"predict_samples_per_second": 14.22,
|
| 16 |
+
"predict_steps_per_second": 1.777,
|
| 17 |
+
"train_loss": 0.7189551201696451,
|
| 18 |
+
"train_runtime": 324742.1955,
|
| 19 |
+
"train_samples": 1134707,
|
| 20 |
+
"train_samples_per_second": 52.413,
|
| 21 |
+
"train_steps_per_second": 3.276
|
| 22 |
+
}
|
checkpoint-141839/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": "4.35.2",
|
| 57 |
+
"use_cache": true,
|
| 58 |
+
"vocab_size": 250054
|
| 59 |
+
}
|
checkpoint-141839/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": "4.35.2"
|
| 12 |
+
}
|
checkpoint-141839/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:787d8f1a526506bcb42c60d3888520b5032131b63fc207871e65cb6cc98e53c8
|
| 3 |
+
size 2444578688
|
checkpoint-141839/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:acd0b7fd0dd2d035ce40a3e4569014e0cac6ac862d41f13fbbe3243a62be4981
|
| 3 |
+
size 4887473903
|
checkpoint-141839/rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:782426fa0f34829286ec0354111c62b1d522d5db2338b2e87085a36ca3c56bec
|
| 3 |
+
size 15024
|
checkpoint-141839/rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b986360a2ded5f01640e0960b8f3114511afd7844eafe799f78c27c8eba55519
|
| 3 |
+
size 15024
|
checkpoint-141839/rng_state_2.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ae6979ec5dbb54570b20cc28ed2208b830ea173aabdd1eaeef42bb39383f31c9
|
| 3 |
+
size 15024
|
checkpoint-141839/rng_state_3.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:856f6436a49b7119ba14c7616ecebe61b710758d1d9577e9f7d53fd4cdf36922
|
| 3 |
+
size 15024
|
checkpoint-141839/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4c862e91d8df536f771de8881de2888162e0ee38ed437f74d8b767f5dcffb6d7
|
| 3 |
+
size 1064
|
checkpoint-141839/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-141839/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-141839/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1d91c41f87c8dbce15b820b41232d0dcd26ba285c22362400d3dd771a711417d
|
| 3 |
+
size 17110107
|
checkpoint-141839/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": "zh_CN",
|
| 525 |
+
"tgt_lang": "ko_KR",
|
| 526 |
+
"tokenizer_class": "MBart50Tokenizer",
|
| 527 |
+
"unk_token": "<unk>"
|
| 528 |
+
}
|
checkpoint-141839/trainer_state.json
ADDED
|
@@ -0,0 +1,1737 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.8987888693809509,
|
| 3 |
+
"best_model_checkpoint": "./zhko_mbartLarge_100p_run1/checkpoint-141839",
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 141839,
|
| 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": 2.335965407031397e-06,
|
| 14 |
+
"loss": 2.879,
|
| 15 |
+
"step": 500
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.01,
|
| 19 |
+
"learning_rate": 4.6860312088738486e-06,
|
| 20 |
+
"loss": 2.0241,
|
| 21 |
+
"step": 1000
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.02,
|
| 25 |
+
"learning_rate": 7.0360970107162996e-06,
|
| 26 |
+
"loss": 1.8819,
|
| 27 |
+
"step": 1500
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.03,
|
| 31 |
+
"learning_rate": 9.386162812558751e-06,
|
| 32 |
+
"loss": 1.8009,
|
| 33 |
+
"step": 2000
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.04,
|
| 37 |
+
"learning_rate": 1.1736228614401204e-05,
|
| 38 |
+
"loss": 1.7072,
|
| 39 |
+
"step": 2500
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.04,
|
| 43 |
+
"learning_rate": 1.4086294416243657e-05,
|
| 44 |
+
"loss": 1.682,
|
| 45 |
+
"step": 3000
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.05,
|
| 49 |
+
"learning_rate": 1.643636021808611e-05,
|
| 50 |
+
"loss": 1.6188,
|
| 51 |
+
"step": 3500
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.06,
|
| 55 |
+
"learning_rate": 1.878642601992856e-05,
|
| 56 |
+
"loss": 1.5802,
|
| 57 |
+
"step": 4000
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.06,
|
| 61 |
+
"learning_rate": 2.1136491821771008e-05,
|
| 62 |
+
"loss": 1.5547,
|
| 63 |
+
"step": 4500
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.07,
|
| 67 |
+
"learning_rate": 2.348655762361346e-05,
|
| 68 |
+
"loss": 1.5422,
|
| 69 |
+
"step": 5000
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.08,
|
| 73 |
+
"learning_rate": 2.5831923293852228e-05,
|
| 74 |
+
"loss": 1.5194,
|
| 75 |
+
"step": 5500
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.08,
|
| 79 |
+
"learning_rate": 2.8181989095694684e-05,
|
| 80 |
+
"loss": 1.4815,
|
| 81 |
+
"step": 6000
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.09,
|
| 85 |
+
"learning_rate": 3.053205489753713e-05,
|
| 86 |
+
"loss": 1.4509,
|
| 87 |
+
"step": 6500
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.1,
|
| 91 |
+
"learning_rate": 3.288212069937958e-05,
|
| 92 |
+
"loss": 1.458,
|
| 93 |
+
"step": 7000
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.11,
|
| 97 |
+
"learning_rate": 3.5227486369618354e-05,
|
| 98 |
+
"loss": 1.4415,
|
| 99 |
+
"step": 7500
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.11,
|
| 103 |
+
"learning_rate": 3.757285203985712e-05,
|
| 104 |
+
"loss": 1.4284,
|
| 105 |
+
"step": 8000
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.12,
|
| 109 |
+
"learning_rate": 3.991821771009588e-05,
|
| 110 |
+
"loss": 1.4297,
|
| 111 |
+
"step": 8500
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.13,
|
| 115 |
+
"learning_rate": 4.2268283511938334e-05,
|
| 116 |
+
"loss": 1.4307,
|
| 117 |
+
"step": 9000
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.13,
|
| 121 |
+
"learning_rate": 4.461834931378079e-05,
|
| 122 |
+
"loss": 1.3971,
|
| 123 |
+
"step": 9500
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.14,
|
| 127 |
+
"learning_rate": 4.696371498401955e-05,
|
| 128 |
+
"loss": 1.4106,
|
| 129 |
+
"step": 10000
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.15,
|
| 133 |
+
"learning_rate": 4.9313780785862004e-05,
|
| 134 |
+
"loss": 1.3999,
|
| 135 |
+
"step": 10500
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.16,
|
| 139 |
+
"learning_rate": 4.998319322943521e-05,
|
| 140 |
+
"loss": 1.3937,
|
| 141 |
+
"step": 11000
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.16,
|
| 145 |
+
"learning_rate": 4.995945485293127e-05,
|
| 146 |
+
"loss": 1.3773,
|
| 147 |
+
"step": 11500
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.17,
|
| 151 |
+
"learning_rate": 4.9935716476427316e-05,
|
| 152 |
+
"loss": 1.3621,
|
| 153 |
+
"step": 12000
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.18,
|
| 157 |
+
"learning_rate": 4.991197809992337e-05,
|
| 158 |
+
"loss": 1.3552,
|
| 159 |
+
"step": 12500
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.18,
|
| 163 |
+
"learning_rate": 4.988823972341943e-05,
|
| 164 |
+
"loss": 1.3312,
|
| 165 |
+
"step": 13000
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.19,
|
| 169 |
+
"learning_rate": 4.986450134691549e-05,
|
| 170 |
+
"loss": 1.3331,
|
| 171 |
+
"step": 13500
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.2,
|
| 175 |
+
"learning_rate": 4.984081044716455e-05,
|
| 176 |
+
"loss": 1.3259,
|
| 177 |
+
"step": 14000
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.2,
|
| 181 |
+
"learning_rate": 4.9817072070660604e-05,
|
| 182 |
+
"loss": 1.2983,
|
| 183 |
+
"step": 14500
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.21,
|
| 187 |
+
"learning_rate": 4.979333369415666e-05,
|
| 188 |
+
"loss": 1.2828,
|
| 189 |
+
"step": 15000
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.22,
|
| 193 |
+
"learning_rate": 4.976959531765271e-05,
|
| 194 |
+
"loss": 1.2793,
|
| 195 |
+
"step": 15500
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.23,
|
| 199 |
+
"learning_rate": 4.9745856941148764e-05,
|
| 200 |
+
"loss": 1.2968,
|
| 201 |
+
"step": 16000
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.23,
|
| 205 |
+
"learning_rate": 4.9722118564644826e-05,
|
| 206 |
+
"loss": 1.2774,
|
| 207 |
+
"step": 16500
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.24,
|
| 211 |
+
"learning_rate": 4.969838018814088e-05,
|
| 212 |
+
"loss": 1.2507,
|
| 213 |
+
"step": 17000
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.25,
|
| 217 |
+
"learning_rate": 4.967464181163694e-05,
|
| 218 |
+
"loss": 1.2521,
|
| 219 |
+
"step": 17500
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.25,
|
| 223 |
+
"learning_rate": 4.965090343513299e-05,
|
| 224 |
+
"loss": 1.2337,
|
| 225 |
+
"step": 18000
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.26,
|
| 229 |
+
"learning_rate": 4.962716505862904e-05,
|
| 230 |
+
"loss": 1.2238,
|
| 231 |
+
"step": 18500
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 0.27,
|
| 235 |
+
"learning_rate": 4.96034266821251e-05,
|
| 236 |
+
"loss": 1.2111,
|
| 237 |
+
"step": 19000
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.27,
|
| 241 |
+
"learning_rate": 4.957968830562115e-05,
|
| 242 |
+
"loss": 1.2336,
|
| 243 |
+
"step": 19500
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.28,
|
| 247 |
+
"learning_rate": 4.955599740587022e-05,
|
| 248 |
+
"loss": 1.1975,
|
| 249 |
+
"step": 20000
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.29,
|
| 253 |
+
"learning_rate": 4.9532259029366274e-05,
|
| 254 |
+
"loss": 1.2092,
|
| 255 |
+
"step": 20500
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.3,
|
| 259 |
+
"learning_rate": 4.950852065286233e-05,
|
| 260 |
+
"loss": 1.2079,
|
| 261 |
+
"step": 21000
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.3,
|
| 265 |
+
"learning_rate": 4.948482975311139e-05,
|
| 266 |
+
"loss": 1.1831,
|
| 267 |
+
"step": 21500
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.31,
|
| 271 |
+
"learning_rate": 4.946109137660745e-05,
|
| 272 |
+
"loss": 1.1965,
|
| 273 |
+
"step": 22000
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 0.32,
|
| 277 |
+
"learning_rate": 4.94373530001035e-05,
|
| 278 |
+
"loss": 1.1928,
|
| 279 |
+
"step": 22500
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.32,
|
| 283 |
+
"learning_rate": 4.9413614623599556e-05,
|
| 284 |
+
"loss": 1.1779,
|
| 285 |
+
"step": 23000
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.33,
|
| 289 |
+
"learning_rate": 4.938987624709561e-05,
|
| 290 |
+
"loss": 1.1759,
|
| 291 |
+
"step": 23500
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.34,
|
| 295 |
+
"learning_rate": 4.936613787059167e-05,
|
| 296 |
+
"loss": 1.1871,
|
| 297 |
+
"step": 24000
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.35,
|
| 301 |
+
"learning_rate": 4.934239949408772e-05,
|
| 302 |
+
"loss": 1.1782,
|
| 303 |
+
"step": 24500
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.35,
|
| 307 |
+
"learning_rate": 4.931866111758378e-05,
|
| 308 |
+
"loss": 1.1619,
|
| 309 |
+
"step": 25000
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.36,
|
| 313 |
+
"learning_rate": 4.9294922741079834e-05,
|
| 314 |
+
"loss": 1.1615,
|
| 315 |
+
"step": 25500
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 0.37,
|
| 319 |
+
"learning_rate": 4.927118436457589e-05,
|
| 320 |
+
"loss": 1.1476,
|
| 321 |
+
"step": 26000
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 0.37,
|
| 325 |
+
"learning_rate": 4.924749346482495e-05,
|
| 326 |
+
"loss": 1.1507,
|
| 327 |
+
"step": 26500
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 0.38,
|
| 331 |
+
"learning_rate": 4.922380256507401e-05,
|
| 332 |
+
"loss": 1.1489,
|
| 333 |
+
"step": 27000
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 0.39,
|
| 337 |
+
"learning_rate": 4.9200064188570064e-05,
|
| 338 |
+
"loss": 1.144,
|
| 339 |
+
"step": 27500
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.39,
|
| 343 |
+
"learning_rate": 4.917632581206613e-05,
|
| 344 |
+
"loss": 1.1223,
|
| 345 |
+
"step": 28000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.4,
|
| 349 |
+
"learning_rate": 4.915258743556218e-05,
|
| 350 |
+
"loss": 1.1319,
|
| 351 |
+
"step": 28500
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 0.41,
|
| 355 |
+
"learning_rate": 4.912884905905824e-05,
|
| 356 |
+
"loss": 1.1235,
|
| 357 |
+
"step": 29000
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 0.42,
|
| 361 |
+
"learning_rate": 4.9105110682554286e-05,
|
| 362 |
+
"loss": 1.1408,
|
| 363 |
+
"step": 29500
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 0.42,
|
| 367 |
+
"learning_rate": 4.908141978280335e-05,
|
| 368 |
+
"loss": 1.1299,
|
| 369 |
+
"step": 30000
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.43,
|
| 373 |
+
"learning_rate": 4.905772888305242e-05,
|
| 374 |
+
"loss": 1.1172,
|
| 375 |
+
"step": 30500
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 0.44,
|
| 379 |
+
"learning_rate": 4.903399050654847e-05,
|
| 380 |
+
"loss": 1.1316,
|
| 381 |
+
"step": 31000
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.44,
|
| 385 |
+
"learning_rate": 4.9010299606797535e-05,
|
| 386 |
+
"loss": 1.1154,
|
| 387 |
+
"step": 31500
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.45,
|
| 391 |
+
"learning_rate": 4.8986561230293584e-05,
|
| 392 |
+
"loss": 1.1162,
|
| 393 |
+
"step": 32000
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 0.46,
|
| 397 |
+
"learning_rate": 4.896282285378964e-05,
|
| 398 |
+
"loss": 1.101,
|
| 399 |
+
"step": 32500
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 0.47,
|
| 403 |
+
"learning_rate": 4.89390844772857e-05,
|
| 404 |
+
"loss": 1.1153,
|
| 405 |
+
"step": 33000
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 0.47,
|
| 409 |
+
"learning_rate": 4.891534610078176e-05,
|
| 410 |
+
"loss": 1.1023,
|
| 411 |
+
"step": 33500
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 0.48,
|
| 415 |
+
"learning_rate": 4.889160772427781e-05,
|
| 416 |
+
"loss": 1.0935,
|
| 417 |
+
"step": 34000
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 0.49,
|
| 421 |
+
"learning_rate": 4.886786934777387e-05,
|
| 422 |
+
"loss": 1.0898,
|
| 423 |
+
"step": 34500
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.49,
|
| 427 |
+
"learning_rate": 4.884413097126992e-05,
|
| 428 |
+
"loss": 1.0846,
|
| 429 |
+
"step": 35000
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.5,
|
| 433 |
+
"learning_rate": 4.882039259476597e-05,
|
| 434 |
+
"loss": 1.0963,
|
| 435 |
+
"step": 35500
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 0.51,
|
| 439 |
+
"learning_rate": 4.879665421826203e-05,
|
| 440 |
+
"loss": 1.0764,
|
| 441 |
+
"step": 36000
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 0.51,
|
| 445 |
+
"learning_rate": 4.877291584175809e-05,
|
| 446 |
+
"loss": 1.0649,
|
| 447 |
+
"step": 36500
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 0.52,
|
| 451 |
+
"learning_rate": 4.8749177465254145e-05,
|
| 452 |
+
"loss": 1.0935,
|
| 453 |
+
"step": 37000
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 0.53,
|
| 457 |
+
"learning_rate": 4.8725439088750194e-05,
|
| 458 |
+
"loss": 1.0739,
|
| 459 |
+
"step": 37500
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.54,
|
| 463 |
+
"learning_rate": 4.870170071224625e-05,
|
| 464 |
+
"loss": 1.0779,
|
| 465 |
+
"step": 38000
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.54,
|
| 469 |
+
"learning_rate": 4.8677962335742305e-05,
|
| 470 |
+
"loss": 1.083,
|
| 471 |
+
"step": 38500
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.55,
|
| 475 |
+
"learning_rate": 4.865422395923836e-05,
|
| 476 |
+
"loss": 1.0729,
|
| 477 |
+
"step": 39000
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 0.56,
|
| 481 |
+
"learning_rate": 4.8630485582734416e-05,
|
| 482 |
+
"loss": 1.0587,
|
| 483 |
+
"step": 39500
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 0.56,
|
| 487 |
+
"learning_rate": 4.860679468298348e-05,
|
| 488 |
+
"loss": 1.078,
|
| 489 |
+
"step": 40000
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 0.57,
|
| 493 |
+
"learning_rate": 4.858305630647954e-05,
|
| 494 |
+
"loss": 1.0802,
|
| 495 |
+
"step": 40500
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 0.58,
|
| 499 |
+
"learning_rate": 4.855931792997559e-05,
|
| 500 |
+
"loss": 1.0538,
|
| 501 |
+
"step": 41000
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 0.59,
|
| 505 |
+
"learning_rate": 4.853557955347164e-05,
|
| 506 |
+
"loss": 1.0562,
|
| 507 |
+
"step": 41500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.59,
|
| 511 |
+
"learning_rate": 4.85118411769677e-05,
|
| 512 |
+
"loss": 1.0516,
|
| 513 |
+
"step": 42000
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.6,
|
| 517 |
+
"learning_rate": 4.848810280046375e-05,
|
| 518 |
+
"loss": 1.0624,
|
| 519 |
+
"step": 42500
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 0.61,
|
| 523 |
+
"learning_rate": 4.8464364423959815e-05,
|
| 524 |
+
"loss": 1.0587,
|
| 525 |
+
"step": 43000
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 0.61,
|
| 529 |
+
"learning_rate": 4.8440626047455864e-05,
|
| 530 |
+
"loss": 1.0544,
|
| 531 |
+
"step": 43500
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 0.62,
|
| 535 |
+
"learning_rate": 4.841688767095192e-05,
|
| 536 |
+
"loss": 1.0467,
|
| 537 |
+
"step": 44000
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 0.63,
|
| 541 |
+
"learning_rate": 4.8393291724707e-05,
|
| 542 |
+
"loss": 1.0613,
|
| 543 |
+
"step": 44500
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 0.63,
|
| 547 |
+
"learning_rate": 4.836955334820306e-05,
|
| 548 |
+
"loss": 1.0365,
|
| 549 |
+
"step": 45000
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.64,
|
| 553 |
+
"learning_rate": 4.834581497169911e-05,
|
| 554 |
+
"loss": 1.0424,
|
| 555 |
+
"step": 45500
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.65,
|
| 559 |
+
"learning_rate": 4.832207659519516e-05,
|
| 560 |
+
"loss": 1.0442,
|
| 561 |
+
"step": 46000
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 0.66,
|
| 565 |
+
"learning_rate": 4.829833821869122e-05,
|
| 566 |
+
"loss": 1.0354,
|
| 567 |
+
"step": 46500
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 0.66,
|
| 571 |
+
"learning_rate": 4.8274789749199304e-05,
|
| 572 |
+
"loss": 1.0543,
|
| 573 |
+
"step": 47000
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 0.67,
|
| 577 |
+
"learning_rate": 4.825105137269536e-05,
|
| 578 |
+
"loss": 1.0476,
|
| 579 |
+
"step": 47500
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 0.68,
|
| 583 |
+
"learning_rate": 4.8227312996191415e-05,
|
| 584 |
+
"loss": 1.0427,
|
| 585 |
+
"step": 48000
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.68,
|
| 589 |
+
"learning_rate": 4.820357461968748e-05,
|
| 590 |
+
"loss": 1.0174,
|
| 591 |
+
"step": 48500
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 0.69,
|
| 595 |
+
"learning_rate": 4.8179836243183526e-05,
|
| 596 |
+
"loss": 1.0325,
|
| 597 |
+
"step": 49000
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.7,
|
| 601 |
+
"learning_rate": 4.815609786667958e-05,
|
| 602 |
+
"loss": 1.0126,
|
| 603 |
+
"step": 49500
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 0.71,
|
| 607 |
+
"learning_rate": 4.813240696692865e-05,
|
| 608 |
+
"loss": 1.0335,
|
| 609 |
+
"step": 50000
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 0.71,
|
| 613 |
+
"learning_rate": 4.81086685904247e-05,
|
| 614 |
+
"loss": 1.0219,
|
| 615 |
+
"step": 50500
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 0.72,
|
| 619 |
+
"learning_rate": 4.808493021392075e-05,
|
| 620 |
+
"loss": 1.0291,
|
| 621 |
+
"step": 51000
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 0.73,
|
| 625 |
+
"learning_rate": 4.8061191837416815e-05,
|
| 626 |
+
"loss": 1.0251,
|
| 627 |
+
"step": 51500
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 0.73,
|
| 631 |
+
"learning_rate": 4.803745346091287e-05,
|
| 632 |
+
"loss": 1.01,
|
| 633 |
+
"step": 52000
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 0.74,
|
| 637 |
+
"learning_rate": 4.8013715084408925e-05,
|
| 638 |
+
"loss": 1.0224,
|
| 639 |
+
"step": 52500
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.75,
|
| 643 |
+
"learning_rate": 4.7989976707904974e-05,
|
| 644 |
+
"loss": 1.0163,
|
| 645 |
+
"step": 53000
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 0.75,
|
| 649 |
+
"learning_rate": 4.796623833140103e-05,
|
| 650 |
+
"loss": 1.0214,
|
| 651 |
+
"step": 53500
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 0.76,
|
| 655 |
+
"learning_rate": 4.7942499954897085e-05,
|
| 656 |
+
"loss": 1.0187,
|
| 657 |
+
"step": 54000
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 0.77,
|
| 661 |
+
"learning_rate": 4.791876157839314e-05,
|
| 662 |
+
"loss": 1.0091,
|
| 663 |
+
"step": 54500
|
| 664 |
+
},
|
| 665 |
+
{
|
| 666 |
+
"epoch": 0.78,
|
| 667 |
+
"learning_rate": 4.78950232018892e-05,
|
| 668 |
+
"loss": 1.0256,
|
| 669 |
+
"step": 55000
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 0.78,
|
| 673 |
+
"learning_rate": 4.787128482538525e-05,
|
| 674 |
+
"loss": 1.0025,
|
| 675 |
+
"step": 55500
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 0.79,
|
| 679 |
+
"learning_rate": 4.784754644888131e-05,
|
| 680 |
+
"loss": 1.0013,
|
| 681 |
+
"step": 56000
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.8,
|
| 685 |
+
"learning_rate": 4.782380807237736e-05,
|
| 686 |
+
"loss": 1.0098,
|
| 687 |
+
"step": 56500
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 0.8,
|
| 691 |
+
"learning_rate": 4.780006969587342e-05,
|
| 692 |
+
"loss": 1.017,
|
| 693 |
+
"step": 57000
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 0.81,
|
| 697 |
+
"learning_rate": 4.777637879612248e-05,
|
| 698 |
+
"loss": 1.0114,
|
| 699 |
+
"step": 57500
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 0.82,
|
| 703 |
+
"learning_rate": 4.775264041961854e-05,
|
| 704 |
+
"loss": 0.9977,
|
| 705 |
+
"step": 58000
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 0.82,
|
| 709 |
+
"learning_rate": 4.7728902043114596e-05,
|
| 710 |
+
"loss": 1.0107,
|
| 711 |
+
"step": 58500
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 0.83,
|
| 715 |
+
"learning_rate": 4.7705163666610644e-05,
|
| 716 |
+
"loss": 0.9869,
|
| 717 |
+
"step": 59000
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 0.84,
|
| 721 |
+
"learning_rate": 4.76814252901067e-05,
|
| 722 |
+
"loss": 1.008,
|
| 723 |
+
"step": 59500
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.85,
|
| 727 |
+
"learning_rate": 4.7657686913602755e-05,
|
| 728 |
+
"loss": 1.0025,
|
| 729 |
+
"step": 60000
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 0.85,
|
| 733 |
+
"learning_rate": 4.763394853709881e-05,
|
| 734 |
+
"loss": 0.9964,
|
| 735 |
+
"step": 60500
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 0.86,
|
| 739 |
+
"learning_rate": 4.7610210160594866e-05,
|
| 740 |
+
"loss": 0.9921,
|
| 741 |
+
"step": 61000
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 0.87,
|
| 745 |
+
"learning_rate": 4.758647178409092e-05,
|
| 746 |
+
"loss": 0.9938,
|
| 747 |
+
"step": 61500
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 0.87,
|
| 751 |
+
"learning_rate": 4.756273340758698e-05,
|
| 752 |
+
"loss": 0.9934,
|
| 753 |
+
"step": 62000
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 0.88,
|
| 757 |
+
"learning_rate": 4.753899503108303e-05,
|
| 758 |
+
"loss": 1.0249,
|
| 759 |
+
"step": 62500
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 0.89,
|
| 763 |
+
"learning_rate": 4.751530413133209e-05,
|
| 764 |
+
"loss": 0.9832,
|
| 765 |
+
"step": 63000
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.9,
|
| 769 |
+
"learning_rate": 4.749156575482815e-05,
|
| 770 |
+
"loss": 1.001,
|
| 771 |
+
"step": 63500
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 0.9,
|
| 775 |
+
"learning_rate": 4.74678273783242e-05,
|
| 776 |
+
"loss": 0.9755,
|
| 777 |
+
"step": 64000
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 0.91,
|
| 781 |
+
"learning_rate": 4.7444089001820266e-05,
|
| 782 |
+
"loss": 0.995,
|
| 783 |
+
"step": 64500
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 0.92,
|
| 787 |
+
"learning_rate": 4.742035062531632e-05,
|
| 788 |
+
"loss": 0.9824,
|
| 789 |
+
"step": 65000
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 0.92,
|
| 793 |
+
"learning_rate": 4.739661224881237e-05,
|
| 794 |
+
"loss": 0.9759,
|
| 795 |
+
"step": 65500
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 0.93,
|
| 799 |
+
"learning_rate": 4.7372873872308425e-05,
|
| 800 |
+
"loss": 0.98,
|
| 801 |
+
"step": 66000
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 0.94,
|
| 805 |
+
"learning_rate": 4.734913549580448e-05,
|
| 806 |
+
"loss": 0.9905,
|
| 807 |
+
"step": 66500
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.94,
|
| 811 |
+
"learning_rate": 4.7325397119300536e-05,
|
| 812 |
+
"loss": 0.9811,
|
| 813 |
+
"step": 67000
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 0.95,
|
| 817 |
+
"learning_rate": 4.730165874279659e-05,
|
| 818 |
+
"loss": 0.9873,
|
| 819 |
+
"step": 67500
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 0.96,
|
| 823 |
+
"learning_rate": 4.727792036629265e-05,
|
| 824 |
+
"loss": 0.9804,
|
| 825 |
+
"step": 68000
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 0.97,
|
| 829 |
+
"learning_rate": 4.7254229466541714e-05,
|
| 830 |
+
"loss": 0.9695,
|
| 831 |
+
"step": 68500
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 0.97,
|
| 835 |
+
"learning_rate": 4.723049109003776e-05,
|
| 836 |
+
"loss": 0.9751,
|
| 837 |
+
"step": 69000
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 0.98,
|
| 841 |
+
"learning_rate": 4.720675271353382e-05,
|
| 842 |
+
"loss": 0.9735,
|
| 843 |
+
"step": 69500
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 0.99,
|
| 847 |
+
"learning_rate": 4.7183061813782885e-05,
|
| 848 |
+
"loss": 0.9832,
|
| 849 |
+
"step": 70000
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.99,
|
| 853 |
+
"learning_rate": 4.7159370914031945e-05,
|
| 854 |
+
"loss": 0.9719,
|
| 855 |
+
"step": 70500
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 1.0,
|
| 859 |
+
"eval_bleu": 40.8492,
|
| 860 |
+
"eval_gen_len": 13.8028,
|
| 861 |
+
"eval_loss": 0.9435123801231384,
|
| 862 |
+
"eval_runtime": 10127.6159,
|
| 863 |
+
"eval_samples_per_second": 14.005,
|
| 864 |
+
"eval_steps_per_second": 1.751,
|
| 865 |
+
"step": 70919
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 1.0,
|
| 869 |
+
"learning_rate": 4.7135632537528e-05,
|
| 870 |
+
"loss": 0.9687,
|
| 871 |
+
"step": 71000
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 1.01,
|
| 875 |
+
"learning_rate": 4.7111894161024056e-05,
|
| 876 |
+
"loss": 0.9639,
|
| 877 |
+
"step": 71500
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 1.02,
|
| 881 |
+
"learning_rate": 4.708815578452011e-05,
|
| 882 |
+
"loss": 0.9669,
|
| 883 |
+
"step": 72000
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 1.02,
|
| 887 |
+
"learning_rate": 4.7064417408016166e-05,
|
| 888 |
+
"loss": 0.956,
|
| 889 |
+
"step": 72500
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"epoch": 1.03,
|
| 893 |
+
"learning_rate": 4.704067903151222e-05,
|
| 894 |
+
"loss": 0.9563,
|
| 895 |
+
"step": 73000
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"epoch": 1.04,
|
| 899 |
+
"learning_rate": 4.701694065500828e-05,
|
| 900 |
+
"loss": 0.9564,
|
| 901 |
+
"step": 73500
|
| 902 |
+
},
|
| 903 |
+
{
|
| 904 |
+
"epoch": 1.04,
|
| 905 |
+
"learning_rate": 4.699320227850433e-05,
|
| 906 |
+
"loss": 0.9555,
|
| 907 |
+
"step": 74000
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 1.05,
|
| 911 |
+
"learning_rate": 4.696946390200039e-05,
|
| 912 |
+
"loss": 0.9408,
|
| 913 |
+
"step": 74500
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 1.06,
|
| 917 |
+
"learning_rate": 4.6945725525496444e-05,
|
| 918 |
+
"loss": 0.9347,
|
| 919 |
+
"step": 75000
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 1.06,
|
| 923 |
+
"learning_rate": 4.69219871489925e-05,
|
| 924 |
+
"loss": 0.9368,
|
| 925 |
+
"step": 75500
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 1.07,
|
| 929 |
+
"learning_rate": 4.6898296249241566e-05,
|
| 930 |
+
"loss": 0.9401,
|
| 931 |
+
"step": 76000
|
| 932 |
+
},
|
| 933 |
+
{
|
| 934 |
+
"epoch": 1.08,
|
| 935 |
+
"learning_rate": 4.6874557872737615e-05,
|
| 936 |
+
"loss": 0.9383,
|
| 937 |
+
"step": 76500
|
| 938 |
+
},
|
| 939 |
+
{
|
| 940 |
+
"epoch": 1.09,
|
| 941 |
+
"learning_rate": 4.685081949623367e-05,
|
| 942 |
+
"loss": 0.9232,
|
| 943 |
+
"step": 77000
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 1.09,
|
| 947 |
+
"learning_rate": 4.6827081119729726e-05,
|
| 948 |
+
"loss": 0.9046,
|
| 949 |
+
"step": 77500
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 1.1,
|
| 953 |
+
"learning_rate": 4.680334274322578e-05,
|
| 954 |
+
"loss": 0.9236,
|
| 955 |
+
"step": 78000
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 1.11,
|
| 959 |
+
"learning_rate": 4.6779604366721837e-05,
|
| 960 |
+
"loss": 0.9091,
|
| 961 |
+
"step": 78500
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 1.11,
|
| 965 |
+
"learning_rate": 4.675586599021789e-05,
|
| 966 |
+
"loss": 0.9036,
|
| 967 |
+
"step": 79000
|
| 968 |
+
},
|
| 969 |
+
{
|
| 970 |
+
"epoch": 1.12,
|
| 971 |
+
"learning_rate": 4.673212761371395e-05,
|
| 972 |
+
"loss": 0.909,
|
| 973 |
+
"step": 79500
|
| 974 |
+
},
|
| 975 |
+
{
|
| 976 |
+
"epoch": 1.13,
|
| 977 |
+
"learning_rate": 4.670838923721e-05,
|
| 978 |
+
"loss": 0.9003,
|
| 979 |
+
"step": 80000
|
| 980 |
+
},
|
| 981 |
+
{
|
| 982 |
+
"epoch": 1.14,
|
| 983 |
+
"learning_rate": 4.668469833745906e-05,
|
| 984 |
+
"loss": 0.8957,
|
| 985 |
+
"step": 80500
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 1.14,
|
| 989 |
+
"learning_rate": 4.666095996095512e-05,
|
| 990 |
+
"loss": 0.8928,
|
| 991 |
+
"step": 81000
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 1.15,
|
| 995 |
+
"learning_rate": 4.66373640147102e-05,
|
| 996 |
+
"loss": 0.8842,
|
| 997 |
+
"step": 81500
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 1.16,
|
| 1001 |
+
"learning_rate": 4.6613625638206256e-05,
|
| 1002 |
+
"loss": 0.885,
|
| 1003 |
+
"step": 82000
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 1.16,
|
| 1007 |
+
"learning_rate": 4.658988726170231e-05,
|
| 1008 |
+
"loss": 0.8821,
|
| 1009 |
+
"step": 82500
|
| 1010 |
+
},
|
| 1011 |
+
{
|
| 1012 |
+
"epoch": 1.17,
|
| 1013 |
+
"learning_rate": 4.656614888519836e-05,
|
| 1014 |
+
"loss": 0.8769,
|
| 1015 |
+
"step": 83000
|
| 1016 |
+
},
|
| 1017 |
+
{
|
| 1018 |
+
"epoch": 1.18,
|
| 1019 |
+
"learning_rate": 4.6542410508694416e-05,
|
| 1020 |
+
"loss": 0.8681,
|
| 1021 |
+
"step": 83500
|
| 1022 |
+
},
|
| 1023 |
+
{
|
| 1024 |
+
"epoch": 1.18,
|
| 1025 |
+
"learning_rate": 4.651867213219048e-05,
|
| 1026 |
+
"loss": 0.8767,
|
| 1027 |
+
"step": 84000
|
| 1028 |
+
},
|
| 1029 |
+
{
|
| 1030 |
+
"epoch": 1.19,
|
| 1031 |
+
"learning_rate": 4.6494933755686534e-05,
|
| 1032 |
+
"loss": 0.8702,
|
| 1033 |
+
"step": 84500
|
| 1034 |
+
},
|
| 1035 |
+
{
|
| 1036 |
+
"epoch": 1.2,
|
| 1037 |
+
"learning_rate": 4.647119537918259e-05,
|
| 1038 |
+
"loss": 0.8753,
|
| 1039 |
+
"step": 85000
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 1.21,
|
| 1043 |
+
"learning_rate": 4.644745700267864e-05,
|
| 1044 |
+
"loss": 0.8587,
|
| 1045 |
+
"step": 85500
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 1.21,
|
| 1049 |
+
"learning_rate": 4.642371862617469e-05,
|
| 1050 |
+
"loss": 0.8459,
|
| 1051 |
+
"step": 86000
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 1.22,
|
| 1055 |
+
"learning_rate": 4.639998024967075e-05,
|
| 1056 |
+
"loss": 0.853,
|
| 1057 |
+
"step": 86500
|
| 1058 |
+
},
|
| 1059 |
+
{
|
| 1060 |
+
"epoch": 1.23,
|
| 1061 |
+
"learning_rate": 4.6376241873166804e-05,
|
| 1062 |
+
"loss": 0.8811,
|
| 1063 |
+
"step": 87000
|
| 1064 |
+
},
|
| 1065 |
+
{
|
| 1066 |
+
"epoch": 1.23,
|
| 1067 |
+
"learning_rate": 4.635255097341587e-05,
|
| 1068 |
+
"loss": 0.8618,
|
| 1069 |
+
"step": 87500
|
| 1070 |
+
},
|
| 1071 |
+
{
|
| 1072 |
+
"epoch": 1.24,
|
| 1073 |
+
"learning_rate": 4.6328812596911926e-05,
|
| 1074 |
+
"loss": 0.8493,
|
| 1075 |
+
"step": 88000
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"epoch": 1.25,
|
| 1079 |
+
"learning_rate": 4.630507422040798e-05,
|
| 1080 |
+
"loss": 0.8603,
|
| 1081 |
+
"step": 88500
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 1.25,
|
| 1085 |
+
"learning_rate": 4.628133584390403e-05,
|
| 1086 |
+
"loss": 0.8333,
|
| 1087 |
+
"step": 89000
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 1.26,
|
| 1091 |
+
"learning_rate": 4.6257597467400086e-05,
|
| 1092 |
+
"loss": 0.8386,
|
| 1093 |
+
"step": 89500
|
| 1094 |
+
},
|
| 1095 |
+
{
|
| 1096 |
+
"epoch": 1.27,
|
| 1097 |
+
"learning_rate": 4.623385909089614e-05,
|
| 1098 |
+
"loss": 0.8424,
|
| 1099 |
+
"step": 90000
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 1.28,
|
| 1103 |
+
"learning_rate": 4.6210120714392204e-05,
|
| 1104 |
+
"loss": 0.8519,
|
| 1105 |
+
"step": 90500
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 1.28,
|
| 1109 |
+
"learning_rate": 4.618638233788826e-05,
|
| 1110 |
+
"loss": 0.8297,
|
| 1111 |
+
"step": 91000
|
| 1112 |
+
},
|
| 1113 |
+
{
|
| 1114 |
+
"epoch": 1.29,
|
| 1115 |
+
"learning_rate": 4.616264396138431e-05,
|
| 1116 |
+
"loss": 0.8406,
|
| 1117 |
+
"step": 91500
|
| 1118 |
+
},
|
| 1119 |
+
{
|
| 1120 |
+
"epoch": 1.3,
|
| 1121 |
+
"learning_rate": 4.613890558488036e-05,
|
| 1122 |
+
"loss": 0.8392,
|
| 1123 |
+
"step": 92000
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 1.3,
|
| 1127 |
+
"learning_rate": 4.611516720837642e-05,
|
| 1128 |
+
"loss": 0.8269,
|
| 1129 |
+
"step": 92500
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 1.31,
|
| 1133 |
+
"learning_rate": 4.6091476308625486e-05,
|
| 1134 |
+
"loss": 0.8392,
|
| 1135 |
+
"step": 93000
|
| 1136 |
+
},
|
| 1137 |
+
{
|
| 1138 |
+
"epoch": 1.32,
|
| 1139 |
+
"learning_rate": 4.6067785408874545e-05,
|
| 1140 |
+
"loss": 0.8474,
|
| 1141 |
+
"step": 93500
|
| 1142 |
+
},
|
| 1143 |
+
{
|
| 1144 |
+
"epoch": 1.33,
|
| 1145 |
+
"learning_rate": 4.60440470323706e-05,
|
| 1146 |
+
"loss": 0.8329,
|
| 1147 |
+
"step": 94000
|
| 1148 |
+
},
|
| 1149 |
+
{
|
| 1150 |
+
"epoch": 1.33,
|
| 1151 |
+
"learning_rate": 4.6020308655866656e-05,
|
| 1152 |
+
"loss": 0.8327,
|
| 1153 |
+
"step": 94500
|
| 1154 |
+
},
|
| 1155 |
+
{
|
| 1156 |
+
"epoch": 1.34,
|
| 1157 |
+
"learning_rate": 4.599657027936271e-05,
|
| 1158 |
+
"loss": 0.8435,
|
| 1159 |
+
"step": 95000
|
| 1160 |
+
},
|
| 1161 |
+
{
|
| 1162 |
+
"epoch": 1.35,
|
| 1163 |
+
"learning_rate": 4.597283190285877e-05,
|
| 1164 |
+
"loss": 0.8343,
|
| 1165 |
+
"step": 95500
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 1.35,
|
| 1169 |
+
"learning_rate": 4.5949141003107834e-05,
|
| 1170 |
+
"loss": 0.8316,
|
| 1171 |
+
"step": 96000
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 1.36,
|
| 1175 |
+
"learning_rate": 4.592540262660389e-05,
|
| 1176 |
+
"loss": 0.8339,
|
| 1177 |
+
"step": 96500
|
| 1178 |
+
},
|
| 1179 |
+
{
|
| 1180 |
+
"epoch": 1.37,
|
| 1181 |
+
"learning_rate": 4.590166425009994e-05,
|
| 1182 |
+
"loss": 0.8199,
|
| 1183 |
+
"step": 97000
|
| 1184 |
+
},
|
| 1185 |
+
{
|
| 1186 |
+
"epoch": 1.37,
|
| 1187 |
+
"learning_rate": 4.5877925873595994e-05,
|
| 1188 |
+
"loss": 0.8162,
|
| 1189 |
+
"step": 97500
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 1.38,
|
| 1193 |
+
"learning_rate": 4.585418749709205e-05,
|
| 1194 |
+
"loss": 0.8239,
|
| 1195 |
+
"step": 98000
|
| 1196 |
+
},
|
| 1197 |
+
{
|
| 1198 |
+
"epoch": 1.39,
|
| 1199 |
+
"learning_rate": 4.5830496597341116e-05,
|
| 1200 |
+
"loss": 0.8197,
|
| 1201 |
+
"step": 98500
|
| 1202 |
+
},
|
| 1203 |
+
{
|
| 1204 |
+
"epoch": 1.4,
|
| 1205 |
+
"learning_rate": 4.5806805697590176e-05,
|
| 1206 |
+
"loss": 0.8164,
|
| 1207 |
+
"step": 99000
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 1.4,
|
| 1211 |
+
"learning_rate": 4.578306732108623e-05,
|
| 1212 |
+
"loss": 0.818,
|
| 1213 |
+
"step": 99500
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 1.41,
|
| 1217 |
+
"learning_rate": 4.575932894458229e-05,
|
| 1218 |
+
"loss": 0.8189,
|
| 1219 |
+
"step": 100000
|
| 1220 |
+
},
|
| 1221 |
+
{
|
| 1222 |
+
"epoch": 1.42,
|
| 1223 |
+
"learning_rate": 4.573559056807834e-05,
|
| 1224 |
+
"loss": 0.8303,
|
| 1225 |
+
"step": 100500
|
| 1226 |
+
},
|
| 1227 |
+
{
|
| 1228 |
+
"epoch": 1.42,
|
| 1229 |
+
"learning_rate": 4.57118521915744e-05,
|
| 1230 |
+
"loss": 0.8174,
|
| 1231 |
+
"step": 101000
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"epoch": 1.43,
|
| 1235 |
+
"learning_rate": 4.568811381507045e-05,
|
| 1236 |
+
"loss": 0.814,
|
| 1237 |
+
"step": 101500
|
| 1238 |
+
},
|
| 1239 |
+
{
|
| 1240 |
+
"epoch": 1.44,
|
| 1241 |
+
"learning_rate": 4.566437543856651e-05,
|
| 1242 |
+
"loss": 0.8226,
|
| 1243 |
+
"step": 102000
|
| 1244 |
+
},
|
| 1245 |
+
{
|
| 1246 |
+
"epoch": 1.45,
|
| 1247 |
+
"learning_rate": 4.5640637062062564e-05,
|
| 1248 |
+
"loss": 0.818,
|
| 1249 |
+
"step": 102500
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 1.45,
|
| 1253 |
+
"learning_rate": 4.561689868555862e-05,
|
| 1254 |
+
"loss": 0.8054,
|
| 1255 |
+
"step": 103000
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 1.46,
|
| 1259 |
+
"learning_rate": 4.5593160309054675e-05,
|
| 1260 |
+
"loss": 0.8191,
|
| 1261 |
+
"step": 103500
|
| 1262 |
+
},
|
| 1263 |
+
{
|
| 1264 |
+
"epoch": 1.47,
|
| 1265 |
+
"learning_rate": 4.556946940930374e-05,
|
| 1266 |
+
"loss": 0.8134,
|
| 1267 |
+
"step": 104000
|
| 1268 |
+
},
|
| 1269 |
+
{
|
| 1270 |
+
"epoch": 1.47,
|
| 1271 |
+
"learning_rate": 4.554573103279979e-05,
|
| 1272 |
+
"loss": 0.8077,
|
| 1273 |
+
"step": 104500
|
| 1274 |
+
},
|
| 1275 |
+
{
|
| 1276 |
+
"epoch": 1.48,
|
| 1277 |
+
"learning_rate": 4.5521992656295846e-05,
|
| 1278 |
+
"loss": 0.8081,
|
| 1279 |
+
"step": 105000
|
| 1280 |
+
},
|
| 1281 |
+
{
|
| 1282 |
+
"epoch": 1.49,
|
| 1283 |
+
"learning_rate": 4.54982542797919e-05,
|
| 1284 |
+
"loss": 0.7985,
|
| 1285 |
+
"step": 105500
|
| 1286 |
+
},
|
| 1287 |
+
{
|
| 1288 |
+
"epoch": 1.49,
|
| 1289 |
+
"learning_rate": 4.547451590328796e-05,
|
| 1290 |
+
"loss": 0.8064,
|
| 1291 |
+
"step": 106000
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 1.5,
|
| 1295 |
+
"learning_rate": 4.545077752678401e-05,
|
| 1296 |
+
"loss": 0.8008,
|
| 1297 |
+
"step": 106500
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 1.51,
|
| 1301 |
+
"learning_rate": 4.542703915028007e-05,
|
| 1302 |
+
"loss": 0.7934,
|
| 1303 |
+
"step": 107000
|
| 1304 |
+
},
|
| 1305 |
+
{
|
| 1306 |
+
"epoch": 1.52,
|
| 1307 |
+
"learning_rate": 4.540330077377612e-05,
|
| 1308 |
+
"loss": 0.7915,
|
| 1309 |
+
"step": 107500
|
| 1310 |
+
},
|
| 1311 |
+
{
|
| 1312 |
+
"epoch": 1.52,
|
| 1313 |
+
"learning_rate": 4.537956239727218e-05,
|
| 1314 |
+
"loss": 0.8087,
|
| 1315 |
+
"step": 108000
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 1.53,
|
| 1319 |
+
"learning_rate": 4.5355824020768234e-05,
|
| 1320 |
+
"loss": 0.7941,
|
| 1321 |
+
"step": 108500
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"epoch": 1.54,
|
| 1325 |
+
"learning_rate": 4.533208564426429e-05,
|
| 1326 |
+
"loss": 0.8032,
|
| 1327 |
+
"step": 109000
|
| 1328 |
+
},
|
| 1329 |
+
{
|
| 1330 |
+
"epoch": 1.54,
|
| 1331 |
+
"learning_rate": 4.5308347267760345e-05,
|
| 1332 |
+
"loss": 0.8061,
|
| 1333 |
+
"step": 109500
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 1.55,
|
| 1337 |
+
"learning_rate": 4.5284608891256394e-05,
|
| 1338 |
+
"loss": 0.7974,
|
| 1339 |
+
"step": 110000
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 1.56,
|
| 1343 |
+
"learning_rate": 4.526091799150546e-05,
|
| 1344 |
+
"loss": 0.7915,
|
| 1345 |
+
"step": 110500
|
| 1346 |
+
},
|
| 1347 |
+
{
|
| 1348 |
+
"epoch": 1.57,
|
| 1349 |
+
"learning_rate": 4.523722709175453e-05,
|
| 1350 |
+
"loss": 0.7998,
|
| 1351 |
+
"step": 111000
|
| 1352 |
+
},
|
| 1353 |
+
{
|
| 1354 |
+
"epoch": 1.57,
|
| 1355 |
+
"learning_rate": 4.5213488715250576e-05,
|
| 1356 |
+
"loss": 0.8014,
|
| 1357 |
+
"step": 111500
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 1.58,
|
| 1361 |
+
"learning_rate": 4.518975033874663e-05,
|
| 1362 |
+
"loss": 0.7848,
|
| 1363 |
+
"step": 112000
|
| 1364 |
+
},
|
| 1365 |
+
{
|
| 1366 |
+
"epoch": 1.59,
|
| 1367 |
+
"learning_rate": 4.516601196224269e-05,
|
| 1368 |
+
"loss": 0.7914,
|
| 1369 |
+
"step": 112500
|
| 1370 |
+
},
|
| 1371 |
+
{
|
| 1372 |
+
"epoch": 1.59,
|
| 1373 |
+
"learning_rate": 4.5142321062491754e-05,
|
| 1374 |
+
"loss": 0.795,
|
| 1375 |
+
"step": 113000
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 1.6,
|
| 1379 |
+
"learning_rate": 4.5118630162740814e-05,
|
| 1380 |
+
"loss": 0.7913,
|
| 1381 |
+
"step": 113500
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 1.61,
|
| 1385 |
+
"learning_rate": 4.509489178623687e-05,
|
| 1386 |
+
"loss": 0.7936,
|
| 1387 |
+
"step": 114000
|
| 1388 |
+
},
|
| 1389 |
+
{
|
| 1390 |
+
"epoch": 1.61,
|
| 1391 |
+
"learning_rate": 4.5071153409732924e-05,
|
| 1392 |
+
"loss": 0.7932,
|
| 1393 |
+
"step": 114500
|
| 1394 |
+
},
|
| 1395 |
+
{
|
| 1396 |
+
"epoch": 1.62,
|
| 1397 |
+
"learning_rate": 4.504741503322898e-05,
|
| 1398 |
+
"loss": 0.7835,
|
| 1399 |
+
"step": 115000
|
| 1400 |
+
},
|
| 1401 |
+
{
|
| 1402 |
+
"epoch": 1.63,
|
| 1403 |
+
"learning_rate": 4.502367665672504e-05,
|
| 1404 |
+
"loss": 0.7942,
|
| 1405 |
+
"step": 115500
|
| 1406 |
+
},
|
| 1407 |
+
{
|
| 1408 |
+
"epoch": 1.64,
|
| 1409 |
+
"learning_rate": 4.499993828022109e-05,
|
| 1410 |
+
"loss": 0.7771,
|
| 1411 |
+
"step": 116000
|
| 1412 |
+
},
|
| 1413 |
+
{
|
| 1414 |
+
"epoch": 1.64,
|
| 1415 |
+
"learning_rate": 4.4976199903717146e-05,
|
| 1416 |
+
"loss": 0.7805,
|
| 1417 |
+
"step": 116500
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 1.65,
|
| 1421 |
+
"learning_rate": 4.4952509003966206e-05,
|
| 1422 |
+
"loss": 0.7952,
|
| 1423 |
+
"step": 117000
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 1.66,
|
| 1427 |
+
"learning_rate": 4.492877062746226e-05,
|
| 1428 |
+
"loss": 0.7799,
|
| 1429 |
+
"step": 117500
|
| 1430 |
+
},
|
| 1431 |
+
{
|
| 1432 |
+
"epoch": 1.66,
|
| 1433 |
+
"learning_rate": 4.490503225095832e-05,
|
| 1434 |
+
"loss": 0.7932,
|
| 1435 |
+
"step": 118000
|
| 1436 |
+
},
|
| 1437 |
+
{
|
| 1438 |
+
"epoch": 1.67,
|
| 1439 |
+
"learning_rate": 4.488129387445438e-05,
|
| 1440 |
+
"loss": 0.7919,
|
| 1441 |
+
"step": 118500
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 1.68,
|
| 1445 |
+
"learning_rate": 4.4857555497950435e-05,
|
| 1446 |
+
"loss": 0.7776,
|
| 1447 |
+
"step": 119000
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 1.69,
|
| 1451 |
+
"learning_rate": 4.4833864598199495e-05,
|
| 1452 |
+
"loss": 0.7731,
|
| 1453 |
+
"step": 119500
|
| 1454 |
+
},
|
| 1455 |
+
{
|
| 1456 |
+
"epoch": 1.69,
|
| 1457 |
+
"learning_rate": 4.481012622169555e-05,
|
| 1458 |
+
"loss": 0.7778,
|
| 1459 |
+
"step": 120000
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 1.7,
|
| 1463 |
+
"learning_rate": 4.47863878451916e-05,
|
| 1464 |
+
"loss": 0.7638,
|
| 1465 |
+
"step": 120500
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 1.71,
|
| 1469 |
+
"learning_rate": 4.476264946868766e-05,
|
| 1470 |
+
"loss": 0.7819,
|
| 1471 |
+
"step": 121000
|
| 1472 |
+
},
|
| 1473 |
+
{
|
| 1474 |
+
"epoch": 1.71,
|
| 1475 |
+
"learning_rate": 4.473891109218372e-05,
|
| 1476 |
+
"loss": 0.7717,
|
| 1477 |
+
"step": 121500
|
| 1478 |
+
},
|
| 1479 |
+
{
|
| 1480 |
+
"epoch": 1.72,
|
| 1481 |
+
"learning_rate": 4.471517271567977e-05,
|
| 1482 |
+
"loss": 0.7803,
|
| 1483 |
+
"step": 122000
|
| 1484 |
+
},
|
| 1485 |
+
{
|
| 1486 |
+
"epoch": 1.73,
|
| 1487 |
+
"learning_rate": 4.469143433917583e-05,
|
| 1488 |
+
"loss": 0.7807,
|
| 1489 |
+
"step": 122500
|
| 1490 |
+
},
|
| 1491 |
+
{
|
| 1492 |
+
"epoch": 1.73,
|
| 1493 |
+
"learning_rate": 4.4667695962671876e-05,
|
| 1494 |
+
"loss": 0.7651,
|
| 1495 |
+
"step": 123000
|
| 1496 |
+
},
|
| 1497 |
+
{
|
| 1498 |
+
"epoch": 1.74,
|
| 1499 |
+
"learning_rate": 4.464395758616793e-05,
|
| 1500 |
+
"loss": 0.7811,
|
| 1501 |
+
"step": 123500
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 1.75,
|
| 1505 |
+
"learning_rate": 4.4620266686417e-05,
|
| 1506 |
+
"loss": 0.7794,
|
| 1507 |
+
"step": 124000
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 1.76,
|
| 1511 |
+
"learning_rate": 4.4596528309913054e-05,
|
| 1512 |
+
"loss": 0.7763,
|
| 1513 |
+
"step": 124500
|
| 1514 |
+
},
|
| 1515 |
+
{
|
| 1516 |
+
"epoch": 1.76,
|
| 1517 |
+
"learning_rate": 4.457278993340911e-05,
|
| 1518 |
+
"loss": 0.7757,
|
| 1519 |
+
"step": 125000
|
| 1520 |
+
},
|
| 1521 |
+
{
|
| 1522 |
+
"epoch": 1.77,
|
| 1523 |
+
"learning_rate": 4.4549051556905165e-05,
|
| 1524 |
+
"loss": 0.7683,
|
| 1525 |
+
"step": 125500
|
| 1526 |
+
},
|
| 1527 |
+
{
|
| 1528 |
+
"epoch": 1.78,
|
| 1529 |
+
"learning_rate": 4.4525360657154225e-05,
|
| 1530 |
+
"loss": 0.7764,
|
| 1531 |
+
"step": 126000
|
| 1532 |
+
},
|
| 1533 |
+
{
|
| 1534 |
+
"epoch": 1.78,
|
| 1535 |
+
"learning_rate": 4.450162228065028e-05,
|
| 1536 |
+
"loss": 0.7661,
|
| 1537 |
+
"step": 126500
|
| 1538 |
+
},
|
| 1539 |
+
{
|
| 1540 |
+
"epoch": 1.79,
|
| 1541 |
+
"learning_rate": 4.4477883904146336e-05,
|
| 1542 |
+
"loss": 0.7645,
|
| 1543 |
+
"step": 127000
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 1.8,
|
| 1547 |
+
"learning_rate": 4.445414552764239e-05,
|
| 1548 |
+
"loss": 0.7703,
|
| 1549 |
+
"step": 127500
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 1.8,
|
| 1553 |
+
"learning_rate": 4.443040715113845e-05,
|
| 1554 |
+
"loss": 0.7791,
|
| 1555 |
+
"step": 128000
|
| 1556 |
+
},
|
| 1557 |
+
{
|
| 1558 |
+
"epoch": 1.81,
|
| 1559 |
+
"learning_rate": 4.44066687746345e-05,
|
| 1560 |
+
"loss": 0.7657,
|
| 1561 |
+
"step": 128500
|
| 1562 |
+
},
|
| 1563 |
+
{
|
| 1564 |
+
"epoch": 1.82,
|
| 1565 |
+
"learning_rate": 4.438293039813056e-05,
|
| 1566 |
+
"loss": 0.7651,
|
| 1567 |
+
"step": 129000
|
| 1568 |
+
},
|
| 1569 |
+
{
|
| 1570 |
+
"epoch": 1.83,
|
| 1571 |
+
"learning_rate": 4.435919202162661e-05,
|
| 1572 |
+
"loss": 0.7803,
|
| 1573 |
+
"step": 129500
|
| 1574 |
+
},
|
| 1575 |
+
{
|
| 1576 |
+
"epoch": 1.83,
|
| 1577 |
+
"learning_rate": 4.433545364512267e-05,
|
| 1578 |
+
"loss": 0.7504,
|
| 1579 |
+
"step": 130000
|
| 1580 |
+
},
|
| 1581 |
+
{
|
| 1582 |
+
"epoch": 1.84,
|
| 1583 |
+
"learning_rate": 4.4311762745371735e-05,
|
| 1584 |
+
"loss": 0.7785,
|
| 1585 |
+
"step": 130500
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 1.85,
|
| 1589 |
+
"learning_rate": 4.4288024368867784e-05,
|
| 1590 |
+
"loss": 0.7592,
|
| 1591 |
+
"step": 131000
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 1.85,
|
| 1595 |
+
"learning_rate": 4.426428599236384e-05,
|
| 1596 |
+
"loss": 0.7714,
|
| 1597 |
+
"step": 131500
|
| 1598 |
+
},
|
| 1599 |
+
{
|
| 1600 |
+
"epoch": 1.86,
|
| 1601 |
+
"learning_rate": 4.4240547615859895e-05,
|
| 1602 |
+
"loss": 0.7652,
|
| 1603 |
+
"step": 132000
|
| 1604 |
+
},
|
| 1605 |
+
{
|
| 1606 |
+
"epoch": 1.87,
|
| 1607 |
+
"learning_rate": 4.421680923935595e-05,
|
| 1608 |
+
"loss": 0.7608,
|
| 1609 |
+
"step": 132500
|
| 1610 |
+
},
|
| 1611 |
+
{
|
| 1612 |
+
"epoch": 1.88,
|
| 1613 |
+
"learning_rate": 4.4193070862852006e-05,
|
| 1614 |
+
"loss": 0.7688,
|
| 1615 |
+
"step": 133000
|
| 1616 |
+
},
|
| 1617 |
+
{
|
| 1618 |
+
"epoch": 1.88,
|
| 1619 |
+
"learning_rate": 4.416933248634806e-05,
|
| 1620 |
+
"loss": 0.79,
|
| 1621 |
+
"step": 133500
|
| 1622 |
+
},
|
| 1623 |
+
{
|
| 1624 |
+
"epoch": 1.89,
|
| 1625 |
+
"learning_rate": 4.414559410984412e-05,
|
| 1626 |
+
"loss": 0.7526,
|
| 1627 |
+
"step": 134000
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 1.9,
|
| 1631 |
+
"learning_rate": 4.412190321009318e-05,
|
| 1632 |
+
"loss": 0.765,
|
| 1633 |
+
"step": 134500
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 1.9,
|
| 1637 |
+
"learning_rate": 4.409816483358923e-05,
|
| 1638 |
+
"loss": 0.7554,
|
| 1639 |
+
"step": 135000
|
| 1640 |
+
},
|
| 1641 |
+
{
|
| 1642 |
+
"epoch": 1.91,
|
| 1643 |
+
"learning_rate": 4.407442645708529e-05,
|
| 1644 |
+
"loss": 0.7636,
|
| 1645 |
+
"step": 135500
|
| 1646 |
+
},
|
| 1647 |
+
{
|
| 1648 |
+
"epoch": 1.92,
|
| 1649 |
+
"learning_rate": 4.4050735557334354e-05,
|
| 1650 |
+
"loss": 0.7582,
|
| 1651 |
+
"step": 136000
|
| 1652 |
+
},
|
| 1653 |
+
{
|
| 1654 |
+
"epoch": 1.92,
|
| 1655 |
+
"learning_rate": 4.402699718083041e-05,
|
| 1656 |
+
"loss": 0.7504,
|
| 1657 |
+
"step": 136500
|
| 1658 |
+
},
|
| 1659 |
+
{
|
| 1660 |
+
"epoch": 1.93,
|
| 1661 |
+
"learning_rate": 4.400330628107947e-05,
|
| 1662 |
+
"loss": 0.7591,
|
| 1663 |
+
"step": 137000
|
| 1664 |
+
},
|
| 1665 |
+
{
|
| 1666 |
+
"epoch": 1.94,
|
| 1667 |
+
"learning_rate": 4.3979567904575525e-05,
|
| 1668 |
+
"loss": 0.7629,
|
| 1669 |
+
"step": 137500
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 1.95,
|
| 1673 |
+
"learning_rate": 4.395582952807158e-05,
|
| 1674 |
+
"loss": 0.7543,
|
| 1675 |
+
"step": 138000
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 1.95,
|
| 1679 |
+
"learning_rate": 4.3932091151567636e-05,
|
| 1680 |
+
"loss": 0.7634,
|
| 1681 |
+
"step": 138500
|
| 1682 |
+
},
|
| 1683 |
+
{
|
| 1684 |
+
"epoch": 1.96,
|
| 1685 |
+
"learning_rate": 4.390835277506369e-05,
|
| 1686 |
+
"loss": 0.7553,
|
| 1687 |
+
"step": 139000
|
| 1688 |
+
},
|
| 1689 |
+
{
|
| 1690 |
+
"epoch": 1.97,
|
| 1691 |
+
"learning_rate": 4.388461439855975e-05,
|
| 1692 |
+
"loss": 0.7548,
|
| 1693 |
+
"step": 139500
|
| 1694 |
+
},
|
| 1695 |
+
{
|
| 1696 |
+
"epoch": 1.97,
|
| 1697 |
+
"learning_rate": 4.38608760220558e-05,
|
| 1698 |
+
"loss": 0.7554,
|
| 1699 |
+
"step": 140000
|
| 1700 |
+
},
|
| 1701 |
+
{
|
| 1702 |
+
"epoch": 1.98,
|
| 1703 |
+
"learning_rate": 4.383713764555186e-05,
|
| 1704 |
+
"loss": 0.7539,
|
| 1705 |
+
"step": 140500
|
| 1706 |
+
},
|
| 1707 |
+
{
|
| 1708 |
+
"epoch": 1.99,
|
| 1709 |
+
"learning_rate": 4.3813399269047914e-05,
|
| 1710 |
+
"loss": 0.7562,
|
| 1711 |
+
"step": 141000
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 2.0,
|
| 1715 |
+
"learning_rate": 4.378966089254396e-05,
|
| 1716 |
+
"loss": 0.7537,
|
| 1717 |
+
"step": 141500
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 2.0,
|
| 1721 |
+
"eval_bleu": 42.7907,
|
| 1722 |
+
"eval_gen_len": 13.7941,
|
| 1723 |
+
"eval_loss": 0.8987888693809509,
|
| 1724 |
+
"eval_runtime": 10044.1408,
|
| 1725 |
+
"eval_samples_per_second": 14.121,
|
| 1726 |
+
"eval_steps_per_second": 1.765,
|
| 1727 |
+
"step": 141839
|
| 1728 |
+
}
|
| 1729 |
+
],
|
| 1730 |
+
"logging_steps": 500,
|
| 1731 |
+
"max_steps": 1063785,
|
| 1732 |
+
"num_train_epochs": 15,
|
| 1733 |
+
"save_steps": 500,
|
| 1734 |
+
"total_flos": 4.918136306736824e+18,
|
| 1735 |
+
"trial_name": null,
|
| 1736 |
+
"trial_params": null
|
| 1737 |
+
}
|
checkpoint-141839/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0adf1a980c6128833811b7e6eb546e117ffd3efb8c21dc7de95b5e76a5b21b8d
|
| 3 |
+
size 4728
|
checkpoint-283678/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": "4.35.2",
|
| 57 |
+
"use_cache": true,
|
| 58 |
+
"vocab_size": 250054
|
| 59 |
+
}
|
checkpoint-283678/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": "4.35.2"
|
| 12 |
+
}
|
checkpoint-283678/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a7d177f1e7493b532ebc4a1ffb3a41e55b3ecb907b70533411897f449b65ec4d
|
| 3 |
+
size 2444578688
|
checkpoint-283678/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d8a229adfc30357fe66f4551a399a4369359fecbcb30285cc052cec1c69d75aa
|
| 3 |
+
size 4887473903
|
checkpoint-283678/rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a7aab6c9cf817dabab5a817aad8e8c0c027ecc31f80ee324829a1c30ad12c664
|
| 3 |
+
size 15024
|
checkpoint-283678/rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7147cb5a020ed017a1409d253ffc4e524b7076e89b7ce45e07bc491c9282478b
|
| 3 |
+
size 15024
|
checkpoint-283678/rng_state_2.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:33ba14e4ce153d100896ad193ee1b3c6d4025f82859f4b9d697280b6cbbf489e
|
| 3 |
+
size 15024
|
checkpoint-283678/rng_state_3.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:68390f8e534ad10bc016408642d9d8145af981899f6818233529c940c786c6a1
|
| 3 |
+
size 15024
|
checkpoint-283678/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bed1f09e79aec328683ecbaa3c7f00d213503ab961ca4b0555a743ab9a30fe03
|
| 3 |
+
size 1064
|
checkpoint-283678/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-283678/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-283678/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1d91c41f87c8dbce15b820b41232d0dcd26ba285c22362400d3dd771a711417d
|
| 3 |
+
size 17110107
|
checkpoint-283678/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": "zh_CN",
|
| 525 |
+
"tgt_lang": "ko_KR",
|
| 526 |
+
"tokenizer_class": "MBart50Tokenizer",
|
| 527 |
+
"unk_token": "<unk>"
|
| 528 |
+
}
|
checkpoint-283678/trainer_state.json
ADDED
|
@@ -0,0 +1,3461 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.8987888693809509,
|
| 3 |
+
"best_model_checkpoint": "./zhko_mbartLarge_100p_run1/checkpoint-141839",
|
| 4 |
+
"epoch": 4.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 283678,
|
| 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": 2.335965407031397e-06,
|
| 14 |
+
"loss": 2.879,
|
| 15 |
+
"step": 500
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.01,
|
| 19 |
+
"learning_rate": 4.6860312088738486e-06,
|
| 20 |
+
"loss": 2.0241,
|
| 21 |
+
"step": 1000
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.02,
|
| 25 |
+
"learning_rate": 7.0360970107162996e-06,
|
| 26 |
+
"loss": 1.8819,
|
| 27 |
+
"step": 1500
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.03,
|
| 31 |
+
"learning_rate": 9.386162812558751e-06,
|
| 32 |
+
"loss": 1.8009,
|
| 33 |
+
"step": 2000
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.04,
|
| 37 |
+
"learning_rate": 1.1736228614401204e-05,
|
| 38 |
+
"loss": 1.7072,
|
| 39 |
+
"step": 2500
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.04,
|
| 43 |
+
"learning_rate": 1.4086294416243657e-05,
|
| 44 |
+
"loss": 1.682,
|
| 45 |
+
"step": 3000
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.05,
|
| 49 |
+
"learning_rate": 1.643636021808611e-05,
|
| 50 |
+
"loss": 1.6188,
|
| 51 |
+
"step": 3500
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.06,
|
| 55 |
+
"learning_rate": 1.878642601992856e-05,
|
| 56 |
+
"loss": 1.5802,
|
| 57 |
+
"step": 4000
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.06,
|
| 61 |
+
"learning_rate": 2.1136491821771008e-05,
|
| 62 |
+
"loss": 1.5547,
|
| 63 |
+
"step": 4500
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.07,
|
| 67 |
+
"learning_rate": 2.348655762361346e-05,
|
| 68 |
+
"loss": 1.5422,
|
| 69 |
+
"step": 5000
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.08,
|
| 73 |
+
"learning_rate": 2.5831923293852228e-05,
|
| 74 |
+
"loss": 1.5194,
|
| 75 |
+
"step": 5500
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.08,
|
| 79 |
+
"learning_rate": 2.8181989095694684e-05,
|
| 80 |
+
"loss": 1.4815,
|
| 81 |
+
"step": 6000
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.09,
|
| 85 |
+
"learning_rate": 3.053205489753713e-05,
|
| 86 |
+
"loss": 1.4509,
|
| 87 |
+
"step": 6500
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.1,
|
| 91 |
+
"learning_rate": 3.288212069937958e-05,
|
| 92 |
+
"loss": 1.458,
|
| 93 |
+
"step": 7000
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.11,
|
| 97 |
+
"learning_rate": 3.5227486369618354e-05,
|
| 98 |
+
"loss": 1.4415,
|
| 99 |
+
"step": 7500
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.11,
|
| 103 |
+
"learning_rate": 3.757285203985712e-05,
|
| 104 |
+
"loss": 1.4284,
|
| 105 |
+
"step": 8000
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.12,
|
| 109 |
+
"learning_rate": 3.991821771009588e-05,
|
| 110 |
+
"loss": 1.4297,
|
| 111 |
+
"step": 8500
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.13,
|
| 115 |
+
"learning_rate": 4.2268283511938334e-05,
|
| 116 |
+
"loss": 1.4307,
|
| 117 |
+
"step": 9000
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.13,
|
| 121 |
+
"learning_rate": 4.461834931378079e-05,
|
| 122 |
+
"loss": 1.3971,
|
| 123 |
+
"step": 9500
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.14,
|
| 127 |
+
"learning_rate": 4.696371498401955e-05,
|
| 128 |
+
"loss": 1.4106,
|
| 129 |
+
"step": 10000
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.15,
|
| 133 |
+
"learning_rate": 4.9313780785862004e-05,
|
| 134 |
+
"loss": 1.3999,
|
| 135 |
+
"step": 10500
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.16,
|
| 139 |
+
"learning_rate": 4.998319322943521e-05,
|
| 140 |
+
"loss": 1.3937,
|
| 141 |
+
"step": 11000
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.16,
|
| 145 |
+
"learning_rate": 4.995945485293127e-05,
|
| 146 |
+
"loss": 1.3773,
|
| 147 |
+
"step": 11500
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.17,
|
| 151 |
+
"learning_rate": 4.9935716476427316e-05,
|
| 152 |
+
"loss": 1.3621,
|
| 153 |
+
"step": 12000
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.18,
|
| 157 |
+
"learning_rate": 4.991197809992337e-05,
|
| 158 |
+
"loss": 1.3552,
|
| 159 |
+
"step": 12500
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.18,
|
| 163 |
+
"learning_rate": 4.988823972341943e-05,
|
| 164 |
+
"loss": 1.3312,
|
| 165 |
+
"step": 13000
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.19,
|
| 169 |
+
"learning_rate": 4.986450134691549e-05,
|
| 170 |
+
"loss": 1.3331,
|
| 171 |
+
"step": 13500
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.2,
|
| 175 |
+
"learning_rate": 4.984081044716455e-05,
|
| 176 |
+
"loss": 1.3259,
|
| 177 |
+
"step": 14000
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.2,
|
| 181 |
+
"learning_rate": 4.9817072070660604e-05,
|
| 182 |
+
"loss": 1.2983,
|
| 183 |
+
"step": 14500
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.21,
|
| 187 |
+
"learning_rate": 4.979333369415666e-05,
|
| 188 |
+
"loss": 1.2828,
|
| 189 |
+
"step": 15000
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.22,
|
| 193 |
+
"learning_rate": 4.976959531765271e-05,
|
| 194 |
+
"loss": 1.2793,
|
| 195 |
+
"step": 15500
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.23,
|
| 199 |
+
"learning_rate": 4.9745856941148764e-05,
|
| 200 |
+
"loss": 1.2968,
|
| 201 |
+
"step": 16000
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.23,
|
| 205 |
+
"learning_rate": 4.9722118564644826e-05,
|
| 206 |
+
"loss": 1.2774,
|
| 207 |
+
"step": 16500
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.24,
|
| 211 |
+
"learning_rate": 4.969838018814088e-05,
|
| 212 |
+
"loss": 1.2507,
|
| 213 |
+
"step": 17000
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.25,
|
| 217 |
+
"learning_rate": 4.967464181163694e-05,
|
| 218 |
+
"loss": 1.2521,
|
| 219 |
+
"step": 17500
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.25,
|
| 223 |
+
"learning_rate": 4.965090343513299e-05,
|
| 224 |
+
"loss": 1.2337,
|
| 225 |
+
"step": 18000
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.26,
|
| 229 |
+
"learning_rate": 4.962716505862904e-05,
|
| 230 |
+
"loss": 1.2238,
|
| 231 |
+
"step": 18500
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 0.27,
|
| 235 |
+
"learning_rate": 4.96034266821251e-05,
|
| 236 |
+
"loss": 1.2111,
|
| 237 |
+
"step": 19000
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.27,
|
| 241 |
+
"learning_rate": 4.957968830562115e-05,
|
| 242 |
+
"loss": 1.2336,
|
| 243 |
+
"step": 19500
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.28,
|
| 247 |
+
"learning_rate": 4.955599740587022e-05,
|
| 248 |
+
"loss": 1.1975,
|
| 249 |
+
"step": 20000
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.29,
|
| 253 |
+
"learning_rate": 4.9532259029366274e-05,
|
| 254 |
+
"loss": 1.2092,
|
| 255 |
+
"step": 20500
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.3,
|
| 259 |
+
"learning_rate": 4.950852065286233e-05,
|
| 260 |
+
"loss": 1.2079,
|
| 261 |
+
"step": 21000
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.3,
|
| 265 |
+
"learning_rate": 4.948482975311139e-05,
|
| 266 |
+
"loss": 1.1831,
|
| 267 |
+
"step": 21500
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.31,
|
| 271 |
+
"learning_rate": 4.946109137660745e-05,
|
| 272 |
+
"loss": 1.1965,
|
| 273 |
+
"step": 22000
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 0.32,
|
| 277 |
+
"learning_rate": 4.94373530001035e-05,
|
| 278 |
+
"loss": 1.1928,
|
| 279 |
+
"step": 22500
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.32,
|
| 283 |
+
"learning_rate": 4.9413614623599556e-05,
|
| 284 |
+
"loss": 1.1779,
|
| 285 |
+
"step": 23000
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.33,
|
| 289 |
+
"learning_rate": 4.938987624709561e-05,
|
| 290 |
+
"loss": 1.1759,
|
| 291 |
+
"step": 23500
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.34,
|
| 295 |
+
"learning_rate": 4.936613787059167e-05,
|
| 296 |
+
"loss": 1.1871,
|
| 297 |
+
"step": 24000
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.35,
|
| 301 |
+
"learning_rate": 4.934239949408772e-05,
|
| 302 |
+
"loss": 1.1782,
|
| 303 |
+
"step": 24500
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.35,
|
| 307 |
+
"learning_rate": 4.931866111758378e-05,
|
| 308 |
+
"loss": 1.1619,
|
| 309 |
+
"step": 25000
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.36,
|
| 313 |
+
"learning_rate": 4.9294922741079834e-05,
|
| 314 |
+
"loss": 1.1615,
|
| 315 |
+
"step": 25500
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 0.37,
|
| 319 |
+
"learning_rate": 4.927118436457589e-05,
|
| 320 |
+
"loss": 1.1476,
|
| 321 |
+
"step": 26000
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 0.37,
|
| 325 |
+
"learning_rate": 4.924749346482495e-05,
|
| 326 |
+
"loss": 1.1507,
|
| 327 |
+
"step": 26500
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 0.38,
|
| 331 |
+
"learning_rate": 4.922380256507401e-05,
|
| 332 |
+
"loss": 1.1489,
|
| 333 |
+
"step": 27000
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 0.39,
|
| 337 |
+
"learning_rate": 4.9200064188570064e-05,
|
| 338 |
+
"loss": 1.144,
|
| 339 |
+
"step": 27500
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.39,
|
| 343 |
+
"learning_rate": 4.917632581206613e-05,
|
| 344 |
+
"loss": 1.1223,
|
| 345 |
+
"step": 28000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.4,
|
| 349 |
+
"learning_rate": 4.915258743556218e-05,
|
| 350 |
+
"loss": 1.1319,
|
| 351 |
+
"step": 28500
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 0.41,
|
| 355 |
+
"learning_rate": 4.912884905905824e-05,
|
| 356 |
+
"loss": 1.1235,
|
| 357 |
+
"step": 29000
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 0.42,
|
| 361 |
+
"learning_rate": 4.9105110682554286e-05,
|
| 362 |
+
"loss": 1.1408,
|
| 363 |
+
"step": 29500
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 0.42,
|
| 367 |
+
"learning_rate": 4.908141978280335e-05,
|
| 368 |
+
"loss": 1.1299,
|
| 369 |
+
"step": 30000
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.43,
|
| 373 |
+
"learning_rate": 4.905772888305242e-05,
|
| 374 |
+
"loss": 1.1172,
|
| 375 |
+
"step": 30500
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 0.44,
|
| 379 |
+
"learning_rate": 4.903399050654847e-05,
|
| 380 |
+
"loss": 1.1316,
|
| 381 |
+
"step": 31000
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.44,
|
| 385 |
+
"learning_rate": 4.9010299606797535e-05,
|
| 386 |
+
"loss": 1.1154,
|
| 387 |
+
"step": 31500
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.45,
|
| 391 |
+
"learning_rate": 4.8986561230293584e-05,
|
| 392 |
+
"loss": 1.1162,
|
| 393 |
+
"step": 32000
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 0.46,
|
| 397 |
+
"learning_rate": 4.896282285378964e-05,
|
| 398 |
+
"loss": 1.101,
|
| 399 |
+
"step": 32500
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 0.47,
|
| 403 |
+
"learning_rate": 4.89390844772857e-05,
|
| 404 |
+
"loss": 1.1153,
|
| 405 |
+
"step": 33000
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 0.47,
|
| 409 |
+
"learning_rate": 4.891534610078176e-05,
|
| 410 |
+
"loss": 1.1023,
|
| 411 |
+
"step": 33500
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 0.48,
|
| 415 |
+
"learning_rate": 4.889160772427781e-05,
|
| 416 |
+
"loss": 1.0935,
|
| 417 |
+
"step": 34000
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 0.49,
|
| 421 |
+
"learning_rate": 4.886786934777387e-05,
|
| 422 |
+
"loss": 1.0898,
|
| 423 |
+
"step": 34500
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.49,
|
| 427 |
+
"learning_rate": 4.884413097126992e-05,
|
| 428 |
+
"loss": 1.0846,
|
| 429 |
+
"step": 35000
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.5,
|
| 433 |
+
"learning_rate": 4.882039259476597e-05,
|
| 434 |
+
"loss": 1.0963,
|
| 435 |
+
"step": 35500
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 0.51,
|
| 439 |
+
"learning_rate": 4.879665421826203e-05,
|
| 440 |
+
"loss": 1.0764,
|
| 441 |
+
"step": 36000
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 0.51,
|
| 445 |
+
"learning_rate": 4.877291584175809e-05,
|
| 446 |
+
"loss": 1.0649,
|
| 447 |
+
"step": 36500
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 0.52,
|
| 451 |
+
"learning_rate": 4.8749177465254145e-05,
|
| 452 |
+
"loss": 1.0935,
|
| 453 |
+
"step": 37000
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 0.53,
|
| 457 |
+
"learning_rate": 4.8725439088750194e-05,
|
| 458 |
+
"loss": 1.0739,
|
| 459 |
+
"step": 37500
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.54,
|
| 463 |
+
"learning_rate": 4.870170071224625e-05,
|
| 464 |
+
"loss": 1.0779,
|
| 465 |
+
"step": 38000
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.54,
|
| 469 |
+
"learning_rate": 4.8677962335742305e-05,
|
| 470 |
+
"loss": 1.083,
|
| 471 |
+
"step": 38500
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.55,
|
| 475 |
+
"learning_rate": 4.865422395923836e-05,
|
| 476 |
+
"loss": 1.0729,
|
| 477 |
+
"step": 39000
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 0.56,
|
| 481 |
+
"learning_rate": 4.8630485582734416e-05,
|
| 482 |
+
"loss": 1.0587,
|
| 483 |
+
"step": 39500
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 0.56,
|
| 487 |
+
"learning_rate": 4.860679468298348e-05,
|
| 488 |
+
"loss": 1.078,
|
| 489 |
+
"step": 40000
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 0.57,
|
| 493 |
+
"learning_rate": 4.858305630647954e-05,
|
| 494 |
+
"loss": 1.0802,
|
| 495 |
+
"step": 40500
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 0.58,
|
| 499 |
+
"learning_rate": 4.855931792997559e-05,
|
| 500 |
+
"loss": 1.0538,
|
| 501 |
+
"step": 41000
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 0.59,
|
| 505 |
+
"learning_rate": 4.853557955347164e-05,
|
| 506 |
+
"loss": 1.0562,
|
| 507 |
+
"step": 41500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.59,
|
| 511 |
+
"learning_rate": 4.85118411769677e-05,
|
| 512 |
+
"loss": 1.0516,
|
| 513 |
+
"step": 42000
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.6,
|
| 517 |
+
"learning_rate": 4.848810280046375e-05,
|
| 518 |
+
"loss": 1.0624,
|
| 519 |
+
"step": 42500
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 0.61,
|
| 523 |
+
"learning_rate": 4.8464364423959815e-05,
|
| 524 |
+
"loss": 1.0587,
|
| 525 |
+
"step": 43000
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 0.61,
|
| 529 |
+
"learning_rate": 4.8440626047455864e-05,
|
| 530 |
+
"loss": 1.0544,
|
| 531 |
+
"step": 43500
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 0.62,
|
| 535 |
+
"learning_rate": 4.841688767095192e-05,
|
| 536 |
+
"loss": 1.0467,
|
| 537 |
+
"step": 44000
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 0.63,
|
| 541 |
+
"learning_rate": 4.8393291724707e-05,
|
| 542 |
+
"loss": 1.0613,
|
| 543 |
+
"step": 44500
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 0.63,
|
| 547 |
+
"learning_rate": 4.836955334820306e-05,
|
| 548 |
+
"loss": 1.0365,
|
| 549 |
+
"step": 45000
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.64,
|
| 553 |
+
"learning_rate": 4.834581497169911e-05,
|
| 554 |
+
"loss": 1.0424,
|
| 555 |
+
"step": 45500
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.65,
|
| 559 |
+
"learning_rate": 4.832207659519516e-05,
|
| 560 |
+
"loss": 1.0442,
|
| 561 |
+
"step": 46000
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 0.66,
|
| 565 |
+
"learning_rate": 4.829833821869122e-05,
|
| 566 |
+
"loss": 1.0354,
|
| 567 |
+
"step": 46500
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 0.66,
|
| 571 |
+
"learning_rate": 4.8274789749199304e-05,
|
| 572 |
+
"loss": 1.0543,
|
| 573 |
+
"step": 47000
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 0.67,
|
| 577 |
+
"learning_rate": 4.825105137269536e-05,
|
| 578 |
+
"loss": 1.0476,
|
| 579 |
+
"step": 47500
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 0.68,
|
| 583 |
+
"learning_rate": 4.8227312996191415e-05,
|
| 584 |
+
"loss": 1.0427,
|
| 585 |
+
"step": 48000
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.68,
|
| 589 |
+
"learning_rate": 4.820357461968748e-05,
|
| 590 |
+
"loss": 1.0174,
|
| 591 |
+
"step": 48500
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 0.69,
|
| 595 |
+
"learning_rate": 4.8179836243183526e-05,
|
| 596 |
+
"loss": 1.0325,
|
| 597 |
+
"step": 49000
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.7,
|
| 601 |
+
"learning_rate": 4.815609786667958e-05,
|
| 602 |
+
"loss": 1.0126,
|
| 603 |
+
"step": 49500
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 0.71,
|
| 607 |
+
"learning_rate": 4.813240696692865e-05,
|
| 608 |
+
"loss": 1.0335,
|
| 609 |
+
"step": 50000
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 0.71,
|
| 613 |
+
"learning_rate": 4.81086685904247e-05,
|
| 614 |
+
"loss": 1.0219,
|
| 615 |
+
"step": 50500
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 0.72,
|
| 619 |
+
"learning_rate": 4.808493021392075e-05,
|
| 620 |
+
"loss": 1.0291,
|
| 621 |
+
"step": 51000
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 0.73,
|
| 625 |
+
"learning_rate": 4.8061191837416815e-05,
|
| 626 |
+
"loss": 1.0251,
|
| 627 |
+
"step": 51500
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 0.73,
|
| 631 |
+
"learning_rate": 4.803745346091287e-05,
|
| 632 |
+
"loss": 1.01,
|
| 633 |
+
"step": 52000
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 0.74,
|
| 637 |
+
"learning_rate": 4.8013715084408925e-05,
|
| 638 |
+
"loss": 1.0224,
|
| 639 |
+
"step": 52500
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.75,
|
| 643 |
+
"learning_rate": 4.7989976707904974e-05,
|
| 644 |
+
"loss": 1.0163,
|
| 645 |
+
"step": 53000
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 0.75,
|
| 649 |
+
"learning_rate": 4.796623833140103e-05,
|
| 650 |
+
"loss": 1.0214,
|
| 651 |
+
"step": 53500
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 0.76,
|
| 655 |
+
"learning_rate": 4.7942499954897085e-05,
|
| 656 |
+
"loss": 1.0187,
|
| 657 |
+
"step": 54000
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 0.77,
|
| 661 |
+
"learning_rate": 4.791876157839314e-05,
|
| 662 |
+
"loss": 1.0091,
|
| 663 |
+
"step": 54500
|
| 664 |
+
},
|
| 665 |
+
{
|
| 666 |
+
"epoch": 0.78,
|
| 667 |
+
"learning_rate": 4.78950232018892e-05,
|
| 668 |
+
"loss": 1.0256,
|
| 669 |
+
"step": 55000
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 0.78,
|
| 673 |
+
"learning_rate": 4.787128482538525e-05,
|
| 674 |
+
"loss": 1.0025,
|
| 675 |
+
"step": 55500
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 0.79,
|
| 679 |
+
"learning_rate": 4.784754644888131e-05,
|
| 680 |
+
"loss": 1.0013,
|
| 681 |
+
"step": 56000
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.8,
|
| 685 |
+
"learning_rate": 4.782380807237736e-05,
|
| 686 |
+
"loss": 1.0098,
|
| 687 |
+
"step": 56500
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 0.8,
|
| 691 |
+
"learning_rate": 4.780006969587342e-05,
|
| 692 |
+
"loss": 1.017,
|
| 693 |
+
"step": 57000
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 0.81,
|
| 697 |
+
"learning_rate": 4.777637879612248e-05,
|
| 698 |
+
"loss": 1.0114,
|
| 699 |
+
"step": 57500
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 0.82,
|
| 703 |
+
"learning_rate": 4.775264041961854e-05,
|
| 704 |
+
"loss": 0.9977,
|
| 705 |
+
"step": 58000
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 0.82,
|
| 709 |
+
"learning_rate": 4.7728902043114596e-05,
|
| 710 |
+
"loss": 1.0107,
|
| 711 |
+
"step": 58500
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 0.83,
|
| 715 |
+
"learning_rate": 4.7705163666610644e-05,
|
| 716 |
+
"loss": 0.9869,
|
| 717 |
+
"step": 59000
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 0.84,
|
| 721 |
+
"learning_rate": 4.76814252901067e-05,
|
| 722 |
+
"loss": 1.008,
|
| 723 |
+
"step": 59500
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.85,
|
| 727 |
+
"learning_rate": 4.7657686913602755e-05,
|
| 728 |
+
"loss": 1.0025,
|
| 729 |
+
"step": 60000
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 0.85,
|
| 733 |
+
"learning_rate": 4.763394853709881e-05,
|
| 734 |
+
"loss": 0.9964,
|
| 735 |
+
"step": 60500
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 0.86,
|
| 739 |
+
"learning_rate": 4.7610210160594866e-05,
|
| 740 |
+
"loss": 0.9921,
|
| 741 |
+
"step": 61000
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 0.87,
|
| 745 |
+
"learning_rate": 4.758647178409092e-05,
|
| 746 |
+
"loss": 0.9938,
|
| 747 |
+
"step": 61500
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 0.87,
|
| 751 |
+
"learning_rate": 4.756273340758698e-05,
|
| 752 |
+
"loss": 0.9934,
|
| 753 |
+
"step": 62000
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 0.88,
|
| 757 |
+
"learning_rate": 4.753899503108303e-05,
|
| 758 |
+
"loss": 1.0249,
|
| 759 |
+
"step": 62500
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 0.89,
|
| 763 |
+
"learning_rate": 4.751530413133209e-05,
|
| 764 |
+
"loss": 0.9832,
|
| 765 |
+
"step": 63000
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.9,
|
| 769 |
+
"learning_rate": 4.749156575482815e-05,
|
| 770 |
+
"loss": 1.001,
|
| 771 |
+
"step": 63500
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 0.9,
|
| 775 |
+
"learning_rate": 4.74678273783242e-05,
|
| 776 |
+
"loss": 0.9755,
|
| 777 |
+
"step": 64000
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 0.91,
|
| 781 |
+
"learning_rate": 4.7444089001820266e-05,
|
| 782 |
+
"loss": 0.995,
|
| 783 |
+
"step": 64500
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 0.92,
|
| 787 |
+
"learning_rate": 4.742035062531632e-05,
|
| 788 |
+
"loss": 0.9824,
|
| 789 |
+
"step": 65000
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 0.92,
|
| 793 |
+
"learning_rate": 4.739661224881237e-05,
|
| 794 |
+
"loss": 0.9759,
|
| 795 |
+
"step": 65500
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 0.93,
|
| 799 |
+
"learning_rate": 4.7372873872308425e-05,
|
| 800 |
+
"loss": 0.98,
|
| 801 |
+
"step": 66000
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 0.94,
|
| 805 |
+
"learning_rate": 4.734913549580448e-05,
|
| 806 |
+
"loss": 0.9905,
|
| 807 |
+
"step": 66500
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.94,
|
| 811 |
+
"learning_rate": 4.7325397119300536e-05,
|
| 812 |
+
"loss": 0.9811,
|
| 813 |
+
"step": 67000
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 0.95,
|
| 817 |
+
"learning_rate": 4.730165874279659e-05,
|
| 818 |
+
"loss": 0.9873,
|
| 819 |
+
"step": 67500
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 0.96,
|
| 823 |
+
"learning_rate": 4.727792036629265e-05,
|
| 824 |
+
"loss": 0.9804,
|
| 825 |
+
"step": 68000
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 0.97,
|
| 829 |
+
"learning_rate": 4.7254229466541714e-05,
|
| 830 |
+
"loss": 0.9695,
|
| 831 |
+
"step": 68500
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 0.97,
|
| 835 |
+
"learning_rate": 4.723049109003776e-05,
|
| 836 |
+
"loss": 0.9751,
|
| 837 |
+
"step": 69000
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 0.98,
|
| 841 |
+
"learning_rate": 4.720675271353382e-05,
|
| 842 |
+
"loss": 0.9735,
|
| 843 |
+
"step": 69500
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 0.99,
|
| 847 |
+
"learning_rate": 4.7183061813782885e-05,
|
| 848 |
+
"loss": 0.9832,
|
| 849 |
+
"step": 70000
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.99,
|
| 853 |
+
"learning_rate": 4.7159370914031945e-05,
|
| 854 |
+
"loss": 0.9719,
|
| 855 |
+
"step": 70500
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 1.0,
|
| 859 |
+
"eval_bleu": 40.8492,
|
| 860 |
+
"eval_gen_len": 13.8028,
|
| 861 |
+
"eval_loss": 0.9435123801231384,
|
| 862 |
+
"eval_runtime": 10127.6159,
|
| 863 |
+
"eval_samples_per_second": 14.005,
|
| 864 |
+
"eval_steps_per_second": 1.751,
|
| 865 |
+
"step": 70919
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 1.0,
|
| 869 |
+
"learning_rate": 4.7135632537528e-05,
|
| 870 |
+
"loss": 0.9687,
|
| 871 |
+
"step": 71000
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 1.01,
|
| 875 |
+
"learning_rate": 4.7111894161024056e-05,
|
| 876 |
+
"loss": 0.9639,
|
| 877 |
+
"step": 71500
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 1.02,
|
| 881 |
+
"learning_rate": 4.708815578452011e-05,
|
| 882 |
+
"loss": 0.9669,
|
| 883 |
+
"step": 72000
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 1.02,
|
| 887 |
+
"learning_rate": 4.7064417408016166e-05,
|
| 888 |
+
"loss": 0.956,
|
| 889 |
+
"step": 72500
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"epoch": 1.03,
|
| 893 |
+
"learning_rate": 4.704067903151222e-05,
|
| 894 |
+
"loss": 0.9563,
|
| 895 |
+
"step": 73000
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"epoch": 1.04,
|
| 899 |
+
"learning_rate": 4.701694065500828e-05,
|
| 900 |
+
"loss": 0.9564,
|
| 901 |
+
"step": 73500
|
| 902 |
+
},
|
| 903 |
+
{
|
| 904 |
+
"epoch": 1.04,
|
| 905 |
+
"learning_rate": 4.699320227850433e-05,
|
| 906 |
+
"loss": 0.9555,
|
| 907 |
+
"step": 74000
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 1.05,
|
| 911 |
+
"learning_rate": 4.696946390200039e-05,
|
| 912 |
+
"loss": 0.9408,
|
| 913 |
+
"step": 74500
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 1.06,
|
| 917 |
+
"learning_rate": 4.6945725525496444e-05,
|
| 918 |
+
"loss": 0.9347,
|
| 919 |
+
"step": 75000
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 1.06,
|
| 923 |
+
"learning_rate": 4.69219871489925e-05,
|
| 924 |
+
"loss": 0.9368,
|
| 925 |
+
"step": 75500
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 1.07,
|
| 929 |
+
"learning_rate": 4.6898296249241566e-05,
|
| 930 |
+
"loss": 0.9401,
|
| 931 |
+
"step": 76000
|
| 932 |
+
},
|
| 933 |
+
{
|
| 934 |
+
"epoch": 1.08,
|
| 935 |
+
"learning_rate": 4.6874557872737615e-05,
|
| 936 |
+
"loss": 0.9383,
|
| 937 |
+
"step": 76500
|
| 938 |
+
},
|
| 939 |
+
{
|
| 940 |
+
"epoch": 1.09,
|
| 941 |
+
"learning_rate": 4.685081949623367e-05,
|
| 942 |
+
"loss": 0.9232,
|
| 943 |
+
"step": 77000
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 1.09,
|
| 947 |
+
"learning_rate": 4.6827081119729726e-05,
|
| 948 |
+
"loss": 0.9046,
|
| 949 |
+
"step": 77500
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 1.1,
|
| 953 |
+
"learning_rate": 4.680334274322578e-05,
|
| 954 |
+
"loss": 0.9236,
|
| 955 |
+
"step": 78000
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 1.11,
|
| 959 |
+
"learning_rate": 4.6779604366721837e-05,
|
| 960 |
+
"loss": 0.9091,
|
| 961 |
+
"step": 78500
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 1.11,
|
| 965 |
+
"learning_rate": 4.675586599021789e-05,
|
| 966 |
+
"loss": 0.9036,
|
| 967 |
+
"step": 79000
|
| 968 |
+
},
|
| 969 |
+
{
|
| 970 |
+
"epoch": 1.12,
|
| 971 |
+
"learning_rate": 4.673212761371395e-05,
|
| 972 |
+
"loss": 0.909,
|
| 973 |
+
"step": 79500
|
| 974 |
+
},
|
| 975 |
+
{
|
| 976 |
+
"epoch": 1.13,
|
| 977 |
+
"learning_rate": 4.670838923721e-05,
|
| 978 |
+
"loss": 0.9003,
|
| 979 |
+
"step": 80000
|
| 980 |
+
},
|
| 981 |
+
{
|
| 982 |
+
"epoch": 1.14,
|
| 983 |
+
"learning_rate": 4.668469833745906e-05,
|
| 984 |
+
"loss": 0.8957,
|
| 985 |
+
"step": 80500
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 1.14,
|
| 989 |
+
"learning_rate": 4.666095996095512e-05,
|
| 990 |
+
"loss": 0.8928,
|
| 991 |
+
"step": 81000
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 1.15,
|
| 995 |
+
"learning_rate": 4.66373640147102e-05,
|
| 996 |
+
"loss": 0.8842,
|
| 997 |
+
"step": 81500
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 1.16,
|
| 1001 |
+
"learning_rate": 4.6613625638206256e-05,
|
| 1002 |
+
"loss": 0.885,
|
| 1003 |
+
"step": 82000
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 1.16,
|
| 1007 |
+
"learning_rate": 4.658988726170231e-05,
|
| 1008 |
+
"loss": 0.8821,
|
| 1009 |
+
"step": 82500
|
| 1010 |
+
},
|
| 1011 |
+
{
|
| 1012 |
+
"epoch": 1.17,
|
| 1013 |
+
"learning_rate": 4.656614888519836e-05,
|
| 1014 |
+
"loss": 0.8769,
|
| 1015 |
+
"step": 83000
|
| 1016 |
+
},
|
| 1017 |
+
{
|
| 1018 |
+
"epoch": 1.18,
|
| 1019 |
+
"learning_rate": 4.6542410508694416e-05,
|
| 1020 |
+
"loss": 0.8681,
|
| 1021 |
+
"step": 83500
|
| 1022 |
+
},
|
| 1023 |
+
{
|
| 1024 |
+
"epoch": 1.18,
|
| 1025 |
+
"learning_rate": 4.651867213219048e-05,
|
| 1026 |
+
"loss": 0.8767,
|
| 1027 |
+
"step": 84000
|
| 1028 |
+
},
|
| 1029 |
+
{
|
| 1030 |
+
"epoch": 1.19,
|
| 1031 |
+
"learning_rate": 4.6494933755686534e-05,
|
| 1032 |
+
"loss": 0.8702,
|
| 1033 |
+
"step": 84500
|
| 1034 |
+
},
|
| 1035 |
+
{
|
| 1036 |
+
"epoch": 1.2,
|
| 1037 |
+
"learning_rate": 4.647119537918259e-05,
|
| 1038 |
+
"loss": 0.8753,
|
| 1039 |
+
"step": 85000
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 1.21,
|
| 1043 |
+
"learning_rate": 4.644745700267864e-05,
|
| 1044 |
+
"loss": 0.8587,
|
| 1045 |
+
"step": 85500
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 1.21,
|
| 1049 |
+
"learning_rate": 4.642371862617469e-05,
|
| 1050 |
+
"loss": 0.8459,
|
| 1051 |
+
"step": 86000
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 1.22,
|
| 1055 |
+
"learning_rate": 4.639998024967075e-05,
|
| 1056 |
+
"loss": 0.853,
|
| 1057 |
+
"step": 86500
|
| 1058 |
+
},
|
| 1059 |
+
{
|
| 1060 |
+
"epoch": 1.23,
|
| 1061 |
+
"learning_rate": 4.6376241873166804e-05,
|
| 1062 |
+
"loss": 0.8811,
|
| 1063 |
+
"step": 87000
|
| 1064 |
+
},
|
| 1065 |
+
{
|
| 1066 |
+
"epoch": 1.23,
|
| 1067 |
+
"learning_rate": 4.635255097341587e-05,
|
| 1068 |
+
"loss": 0.8618,
|
| 1069 |
+
"step": 87500
|
| 1070 |
+
},
|
| 1071 |
+
{
|
| 1072 |
+
"epoch": 1.24,
|
| 1073 |
+
"learning_rate": 4.6328812596911926e-05,
|
| 1074 |
+
"loss": 0.8493,
|
| 1075 |
+
"step": 88000
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"epoch": 1.25,
|
| 1079 |
+
"learning_rate": 4.630507422040798e-05,
|
| 1080 |
+
"loss": 0.8603,
|
| 1081 |
+
"step": 88500
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 1.25,
|
| 1085 |
+
"learning_rate": 4.628133584390403e-05,
|
| 1086 |
+
"loss": 0.8333,
|
| 1087 |
+
"step": 89000
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 1.26,
|
| 1091 |
+
"learning_rate": 4.6257597467400086e-05,
|
| 1092 |
+
"loss": 0.8386,
|
| 1093 |
+
"step": 89500
|
| 1094 |
+
},
|
| 1095 |
+
{
|
| 1096 |
+
"epoch": 1.27,
|
| 1097 |
+
"learning_rate": 4.623385909089614e-05,
|
| 1098 |
+
"loss": 0.8424,
|
| 1099 |
+
"step": 90000
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 1.28,
|
| 1103 |
+
"learning_rate": 4.6210120714392204e-05,
|
| 1104 |
+
"loss": 0.8519,
|
| 1105 |
+
"step": 90500
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 1.28,
|
| 1109 |
+
"learning_rate": 4.618638233788826e-05,
|
| 1110 |
+
"loss": 0.8297,
|
| 1111 |
+
"step": 91000
|
| 1112 |
+
},
|
| 1113 |
+
{
|
| 1114 |
+
"epoch": 1.29,
|
| 1115 |
+
"learning_rate": 4.616264396138431e-05,
|
| 1116 |
+
"loss": 0.8406,
|
| 1117 |
+
"step": 91500
|
| 1118 |
+
},
|
| 1119 |
+
{
|
| 1120 |
+
"epoch": 1.3,
|
| 1121 |
+
"learning_rate": 4.613890558488036e-05,
|
| 1122 |
+
"loss": 0.8392,
|
| 1123 |
+
"step": 92000
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 1.3,
|
| 1127 |
+
"learning_rate": 4.611516720837642e-05,
|
| 1128 |
+
"loss": 0.8269,
|
| 1129 |
+
"step": 92500
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 1.31,
|
| 1133 |
+
"learning_rate": 4.6091476308625486e-05,
|
| 1134 |
+
"loss": 0.8392,
|
| 1135 |
+
"step": 93000
|
| 1136 |
+
},
|
| 1137 |
+
{
|
| 1138 |
+
"epoch": 1.32,
|
| 1139 |
+
"learning_rate": 4.6067785408874545e-05,
|
| 1140 |
+
"loss": 0.8474,
|
| 1141 |
+
"step": 93500
|
| 1142 |
+
},
|
| 1143 |
+
{
|
| 1144 |
+
"epoch": 1.33,
|
| 1145 |
+
"learning_rate": 4.60440470323706e-05,
|
| 1146 |
+
"loss": 0.8329,
|
| 1147 |
+
"step": 94000
|
| 1148 |
+
},
|
| 1149 |
+
{
|
| 1150 |
+
"epoch": 1.33,
|
| 1151 |
+
"learning_rate": 4.6020308655866656e-05,
|
| 1152 |
+
"loss": 0.8327,
|
| 1153 |
+
"step": 94500
|
| 1154 |
+
},
|
| 1155 |
+
{
|
| 1156 |
+
"epoch": 1.34,
|
| 1157 |
+
"learning_rate": 4.599657027936271e-05,
|
| 1158 |
+
"loss": 0.8435,
|
| 1159 |
+
"step": 95000
|
| 1160 |
+
},
|
| 1161 |
+
{
|
| 1162 |
+
"epoch": 1.35,
|
| 1163 |
+
"learning_rate": 4.597283190285877e-05,
|
| 1164 |
+
"loss": 0.8343,
|
| 1165 |
+
"step": 95500
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 1.35,
|
| 1169 |
+
"learning_rate": 4.5949141003107834e-05,
|
| 1170 |
+
"loss": 0.8316,
|
| 1171 |
+
"step": 96000
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 1.36,
|
| 1175 |
+
"learning_rate": 4.592540262660389e-05,
|
| 1176 |
+
"loss": 0.8339,
|
| 1177 |
+
"step": 96500
|
| 1178 |
+
},
|
| 1179 |
+
{
|
| 1180 |
+
"epoch": 1.37,
|
| 1181 |
+
"learning_rate": 4.590166425009994e-05,
|
| 1182 |
+
"loss": 0.8199,
|
| 1183 |
+
"step": 97000
|
| 1184 |
+
},
|
| 1185 |
+
{
|
| 1186 |
+
"epoch": 1.37,
|
| 1187 |
+
"learning_rate": 4.5877925873595994e-05,
|
| 1188 |
+
"loss": 0.8162,
|
| 1189 |
+
"step": 97500
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 1.38,
|
| 1193 |
+
"learning_rate": 4.585418749709205e-05,
|
| 1194 |
+
"loss": 0.8239,
|
| 1195 |
+
"step": 98000
|
| 1196 |
+
},
|
| 1197 |
+
{
|
| 1198 |
+
"epoch": 1.39,
|
| 1199 |
+
"learning_rate": 4.5830496597341116e-05,
|
| 1200 |
+
"loss": 0.8197,
|
| 1201 |
+
"step": 98500
|
| 1202 |
+
},
|
| 1203 |
+
{
|
| 1204 |
+
"epoch": 1.4,
|
| 1205 |
+
"learning_rate": 4.5806805697590176e-05,
|
| 1206 |
+
"loss": 0.8164,
|
| 1207 |
+
"step": 99000
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 1.4,
|
| 1211 |
+
"learning_rate": 4.578306732108623e-05,
|
| 1212 |
+
"loss": 0.818,
|
| 1213 |
+
"step": 99500
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 1.41,
|
| 1217 |
+
"learning_rate": 4.575932894458229e-05,
|
| 1218 |
+
"loss": 0.8189,
|
| 1219 |
+
"step": 100000
|
| 1220 |
+
},
|
| 1221 |
+
{
|
| 1222 |
+
"epoch": 1.42,
|
| 1223 |
+
"learning_rate": 4.573559056807834e-05,
|
| 1224 |
+
"loss": 0.8303,
|
| 1225 |
+
"step": 100500
|
| 1226 |
+
},
|
| 1227 |
+
{
|
| 1228 |
+
"epoch": 1.42,
|
| 1229 |
+
"learning_rate": 4.57118521915744e-05,
|
| 1230 |
+
"loss": 0.8174,
|
| 1231 |
+
"step": 101000
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"epoch": 1.43,
|
| 1235 |
+
"learning_rate": 4.568811381507045e-05,
|
| 1236 |
+
"loss": 0.814,
|
| 1237 |
+
"step": 101500
|
| 1238 |
+
},
|
| 1239 |
+
{
|
| 1240 |
+
"epoch": 1.44,
|
| 1241 |
+
"learning_rate": 4.566437543856651e-05,
|
| 1242 |
+
"loss": 0.8226,
|
| 1243 |
+
"step": 102000
|
| 1244 |
+
},
|
| 1245 |
+
{
|
| 1246 |
+
"epoch": 1.45,
|
| 1247 |
+
"learning_rate": 4.5640637062062564e-05,
|
| 1248 |
+
"loss": 0.818,
|
| 1249 |
+
"step": 102500
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 1.45,
|
| 1253 |
+
"learning_rate": 4.561689868555862e-05,
|
| 1254 |
+
"loss": 0.8054,
|
| 1255 |
+
"step": 103000
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 1.46,
|
| 1259 |
+
"learning_rate": 4.5593160309054675e-05,
|
| 1260 |
+
"loss": 0.8191,
|
| 1261 |
+
"step": 103500
|
| 1262 |
+
},
|
| 1263 |
+
{
|
| 1264 |
+
"epoch": 1.47,
|
| 1265 |
+
"learning_rate": 4.556946940930374e-05,
|
| 1266 |
+
"loss": 0.8134,
|
| 1267 |
+
"step": 104000
|
| 1268 |
+
},
|
| 1269 |
+
{
|
| 1270 |
+
"epoch": 1.47,
|
| 1271 |
+
"learning_rate": 4.554573103279979e-05,
|
| 1272 |
+
"loss": 0.8077,
|
| 1273 |
+
"step": 104500
|
| 1274 |
+
},
|
| 1275 |
+
{
|
| 1276 |
+
"epoch": 1.48,
|
| 1277 |
+
"learning_rate": 4.5521992656295846e-05,
|
| 1278 |
+
"loss": 0.8081,
|
| 1279 |
+
"step": 105000
|
| 1280 |
+
},
|
| 1281 |
+
{
|
| 1282 |
+
"epoch": 1.49,
|
| 1283 |
+
"learning_rate": 4.54982542797919e-05,
|
| 1284 |
+
"loss": 0.7985,
|
| 1285 |
+
"step": 105500
|
| 1286 |
+
},
|
| 1287 |
+
{
|
| 1288 |
+
"epoch": 1.49,
|
| 1289 |
+
"learning_rate": 4.547451590328796e-05,
|
| 1290 |
+
"loss": 0.8064,
|
| 1291 |
+
"step": 106000
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 1.5,
|
| 1295 |
+
"learning_rate": 4.545077752678401e-05,
|
| 1296 |
+
"loss": 0.8008,
|
| 1297 |
+
"step": 106500
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 1.51,
|
| 1301 |
+
"learning_rate": 4.542703915028007e-05,
|
| 1302 |
+
"loss": 0.7934,
|
| 1303 |
+
"step": 107000
|
| 1304 |
+
},
|
| 1305 |
+
{
|
| 1306 |
+
"epoch": 1.52,
|
| 1307 |
+
"learning_rate": 4.540330077377612e-05,
|
| 1308 |
+
"loss": 0.7915,
|
| 1309 |
+
"step": 107500
|
| 1310 |
+
},
|
| 1311 |
+
{
|
| 1312 |
+
"epoch": 1.52,
|
| 1313 |
+
"learning_rate": 4.537956239727218e-05,
|
| 1314 |
+
"loss": 0.8087,
|
| 1315 |
+
"step": 108000
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 1.53,
|
| 1319 |
+
"learning_rate": 4.5355824020768234e-05,
|
| 1320 |
+
"loss": 0.7941,
|
| 1321 |
+
"step": 108500
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"epoch": 1.54,
|
| 1325 |
+
"learning_rate": 4.533208564426429e-05,
|
| 1326 |
+
"loss": 0.8032,
|
| 1327 |
+
"step": 109000
|
| 1328 |
+
},
|
| 1329 |
+
{
|
| 1330 |
+
"epoch": 1.54,
|
| 1331 |
+
"learning_rate": 4.5308347267760345e-05,
|
| 1332 |
+
"loss": 0.8061,
|
| 1333 |
+
"step": 109500
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 1.55,
|
| 1337 |
+
"learning_rate": 4.5284608891256394e-05,
|
| 1338 |
+
"loss": 0.7974,
|
| 1339 |
+
"step": 110000
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 1.56,
|
| 1343 |
+
"learning_rate": 4.526091799150546e-05,
|
| 1344 |
+
"loss": 0.7915,
|
| 1345 |
+
"step": 110500
|
| 1346 |
+
},
|
| 1347 |
+
{
|
| 1348 |
+
"epoch": 1.57,
|
| 1349 |
+
"learning_rate": 4.523722709175453e-05,
|
| 1350 |
+
"loss": 0.7998,
|
| 1351 |
+
"step": 111000
|
| 1352 |
+
},
|
| 1353 |
+
{
|
| 1354 |
+
"epoch": 1.57,
|
| 1355 |
+
"learning_rate": 4.5213488715250576e-05,
|
| 1356 |
+
"loss": 0.8014,
|
| 1357 |
+
"step": 111500
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 1.58,
|
| 1361 |
+
"learning_rate": 4.518975033874663e-05,
|
| 1362 |
+
"loss": 0.7848,
|
| 1363 |
+
"step": 112000
|
| 1364 |
+
},
|
| 1365 |
+
{
|
| 1366 |
+
"epoch": 1.59,
|
| 1367 |
+
"learning_rate": 4.516601196224269e-05,
|
| 1368 |
+
"loss": 0.7914,
|
| 1369 |
+
"step": 112500
|
| 1370 |
+
},
|
| 1371 |
+
{
|
| 1372 |
+
"epoch": 1.59,
|
| 1373 |
+
"learning_rate": 4.5142321062491754e-05,
|
| 1374 |
+
"loss": 0.795,
|
| 1375 |
+
"step": 113000
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 1.6,
|
| 1379 |
+
"learning_rate": 4.5118630162740814e-05,
|
| 1380 |
+
"loss": 0.7913,
|
| 1381 |
+
"step": 113500
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 1.61,
|
| 1385 |
+
"learning_rate": 4.509489178623687e-05,
|
| 1386 |
+
"loss": 0.7936,
|
| 1387 |
+
"step": 114000
|
| 1388 |
+
},
|
| 1389 |
+
{
|
| 1390 |
+
"epoch": 1.61,
|
| 1391 |
+
"learning_rate": 4.5071153409732924e-05,
|
| 1392 |
+
"loss": 0.7932,
|
| 1393 |
+
"step": 114500
|
| 1394 |
+
},
|
| 1395 |
+
{
|
| 1396 |
+
"epoch": 1.62,
|
| 1397 |
+
"learning_rate": 4.504741503322898e-05,
|
| 1398 |
+
"loss": 0.7835,
|
| 1399 |
+
"step": 115000
|
| 1400 |
+
},
|
| 1401 |
+
{
|
| 1402 |
+
"epoch": 1.63,
|
| 1403 |
+
"learning_rate": 4.502367665672504e-05,
|
| 1404 |
+
"loss": 0.7942,
|
| 1405 |
+
"step": 115500
|
| 1406 |
+
},
|
| 1407 |
+
{
|
| 1408 |
+
"epoch": 1.64,
|
| 1409 |
+
"learning_rate": 4.499993828022109e-05,
|
| 1410 |
+
"loss": 0.7771,
|
| 1411 |
+
"step": 116000
|
| 1412 |
+
},
|
| 1413 |
+
{
|
| 1414 |
+
"epoch": 1.64,
|
| 1415 |
+
"learning_rate": 4.4976199903717146e-05,
|
| 1416 |
+
"loss": 0.7805,
|
| 1417 |
+
"step": 116500
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 1.65,
|
| 1421 |
+
"learning_rate": 4.4952509003966206e-05,
|
| 1422 |
+
"loss": 0.7952,
|
| 1423 |
+
"step": 117000
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 1.66,
|
| 1427 |
+
"learning_rate": 4.492877062746226e-05,
|
| 1428 |
+
"loss": 0.7799,
|
| 1429 |
+
"step": 117500
|
| 1430 |
+
},
|
| 1431 |
+
{
|
| 1432 |
+
"epoch": 1.66,
|
| 1433 |
+
"learning_rate": 4.490503225095832e-05,
|
| 1434 |
+
"loss": 0.7932,
|
| 1435 |
+
"step": 118000
|
| 1436 |
+
},
|
| 1437 |
+
{
|
| 1438 |
+
"epoch": 1.67,
|
| 1439 |
+
"learning_rate": 4.488129387445438e-05,
|
| 1440 |
+
"loss": 0.7919,
|
| 1441 |
+
"step": 118500
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 1.68,
|
| 1445 |
+
"learning_rate": 4.4857555497950435e-05,
|
| 1446 |
+
"loss": 0.7776,
|
| 1447 |
+
"step": 119000
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 1.69,
|
| 1451 |
+
"learning_rate": 4.4833864598199495e-05,
|
| 1452 |
+
"loss": 0.7731,
|
| 1453 |
+
"step": 119500
|
| 1454 |
+
},
|
| 1455 |
+
{
|
| 1456 |
+
"epoch": 1.69,
|
| 1457 |
+
"learning_rate": 4.481012622169555e-05,
|
| 1458 |
+
"loss": 0.7778,
|
| 1459 |
+
"step": 120000
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 1.7,
|
| 1463 |
+
"learning_rate": 4.47863878451916e-05,
|
| 1464 |
+
"loss": 0.7638,
|
| 1465 |
+
"step": 120500
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 1.71,
|
| 1469 |
+
"learning_rate": 4.476264946868766e-05,
|
| 1470 |
+
"loss": 0.7819,
|
| 1471 |
+
"step": 121000
|
| 1472 |
+
},
|
| 1473 |
+
{
|
| 1474 |
+
"epoch": 1.71,
|
| 1475 |
+
"learning_rate": 4.473891109218372e-05,
|
| 1476 |
+
"loss": 0.7717,
|
| 1477 |
+
"step": 121500
|
| 1478 |
+
},
|
| 1479 |
+
{
|
| 1480 |
+
"epoch": 1.72,
|
| 1481 |
+
"learning_rate": 4.471517271567977e-05,
|
| 1482 |
+
"loss": 0.7803,
|
| 1483 |
+
"step": 122000
|
| 1484 |
+
},
|
| 1485 |
+
{
|
| 1486 |
+
"epoch": 1.73,
|
| 1487 |
+
"learning_rate": 4.469143433917583e-05,
|
| 1488 |
+
"loss": 0.7807,
|
| 1489 |
+
"step": 122500
|
| 1490 |
+
},
|
| 1491 |
+
{
|
| 1492 |
+
"epoch": 1.73,
|
| 1493 |
+
"learning_rate": 4.4667695962671876e-05,
|
| 1494 |
+
"loss": 0.7651,
|
| 1495 |
+
"step": 123000
|
| 1496 |
+
},
|
| 1497 |
+
{
|
| 1498 |
+
"epoch": 1.74,
|
| 1499 |
+
"learning_rate": 4.464395758616793e-05,
|
| 1500 |
+
"loss": 0.7811,
|
| 1501 |
+
"step": 123500
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 1.75,
|
| 1505 |
+
"learning_rate": 4.4620266686417e-05,
|
| 1506 |
+
"loss": 0.7794,
|
| 1507 |
+
"step": 124000
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 1.76,
|
| 1511 |
+
"learning_rate": 4.4596528309913054e-05,
|
| 1512 |
+
"loss": 0.7763,
|
| 1513 |
+
"step": 124500
|
| 1514 |
+
},
|
| 1515 |
+
{
|
| 1516 |
+
"epoch": 1.76,
|
| 1517 |
+
"learning_rate": 4.457278993340911e-05,
|
| 1518 |
+
"loss": 0.7757,
|
| 1519 |
+
"step": 125000
|
| 1520 |
+
},
|
| 1521 |
+
{
|
| 1522 |
+
"epoch": 1.77,
|
| 1523 |
+
"learning_rate": 4.4549051556905165e-05,
|
| 1524 |
+
"loss": 0.7683,
|
| 1525 |
+
"step": 125500
|
| 1526 |
+
},
|
| 1527 |
+
{
|
| 1528 |
+
"epoch": 1.78,
|
| 1529 |
+
"learning_rate": 4.4525360657154225e-05,
|
| 1530 |
+
"loss": 0.7764,
|
| 1531 |
+
"step": 126000
|
| 1532 |
+
},
|
| 1533 |
+
{
|
| 1534 |
+
"epoch": 1.78,
|
| 1535 |
+
"learning_rate": 4.450162228065028e-05,
|
| 1536 |
+
"loss": 0.7661,
|
| 1537 |
+
"step": 126500
|
| 1538 |
+
},
|
| 1539 |
+
{
|
| 1540 |
+
"epoch": 1.79,
|
| 1541 |
+
"learning_rate": 4.4477883904146336e-05,
|
| 1542 |
+
"loss": 0.7645,
|
| 1543 |
+
"step": 127000
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 1.8,
|
| 1547 |
+
"learning_rate": 4.445414552764239e-05,
|
| 1548 |
+
"loss": 0.7703,
|
| 1549 |
+
"step": 127500
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 1.8,
|
| 1553 |
+
"learning_rate": 4.443040715113845e-05,
|
| 1554 |
+
"loss": 0.7791,
|
| 1555 |
+
"step": 128000
|
| 1556 |
+
},
|
| 1557 |
+
{
|
| 1558 |
+
"epoch": 1.81,
|
| 1559 |
+
"learning_rate": 4.44066687746345e-05,
|
| 1560 |
+
"loss": 0.7657,
|
| 1561 |
+
"step": 128500
|
| 1562 |
+
},
|
| 1563 |
+
{
|
| 1564 |
+
"epoch": 1.82,
|
| 1565 |
+
"learning_rate": 4.438293039813056e-05,
|
| 1566 |
+
"loss": 0.7651,
|
| 1567 |
+
"step": 129000
|
| 1568 |
+
},
|
| 1569 |
+
{
|
| 1570 |
+
"epoch": 1.83,
|
| 1571 |
+
"learning_rate": 4.435919202162661e-05,
|
| 1572 |
+
"loss": 0.7803,
|
| 1573 |
+
"step": 129500
|
| 1574 |
+
},
|
| 1575 |
+
{
|
| 1576 |
+
"epoch": 1.83,
|
| 1577 |
+
"learning_rate": 4.433545364512267e-05,
|
| 1578 |
+
"loss": 0.7504,
|
| 1579 |
+
"step": 130000
|
| 1580 |
+
},
|
| 1581 |
+
{
|
| 1582 |
+
"epoch": 1.84,
|
| 1583 |
+
"learning_rate": 4.4311762745371735e-05,
|
| 1584 |
+
"loss": 0.7785,
|
| 1585 |
+
"step": 130500
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 1.85,
|
| 1589 |
+
"learning_rate": 4.4288024368867784e-05,
|
| 1590 |
+
"loss": 0.7592,
|
| 1591 |
+
"step": 131000
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 1.85,
|
| 1595 |
+
"learning_rate": 4.426428599236384e-05,
|
| 1596 |
+
"loss": 0.7714,
|
| 1597 |
+
"step": 131500
|
| 1598 |
+
},
|
| 1599 |
+
{
|
| 1600 |
+
"epoch": 1.86,
|
| 1601 |
+
"learning_rate": 4.4240547615859895e-05,
|
| 1602 |
+
"loss": 0.7652,
|
| 1603 |
+
"step": 132000
|
| 1604 |
+
},
|
| 1605 |
+
{
|
| 1606 |
+
"epoch": 1.87,
|
| 1607 |
+
"learning_rate": 4.421680923935595e-05,
|
| 1608 |
+
"loss": 0.7608,
|
| 1609 |
+
"step": 132500
|
| 1610 |
+
},
|
| 1611 |
+
{
|
| 1612 |
+
"epoch": 1.88,
|
| 1613 |
+
"learning_rate": 4.4193070862852006e-05,
|
| 1614 |
+
"loss": 0.7688,
|
| 1615 |
+
"step": 133000
|
| 1616 |
+
},
|
| 1617 |
+
{
|
| 1618 |
+
"epoch": 1.88,
|
| 1619 |
+
"learning_rate": 4.416933248634806e-05,
|
| 1620 |
+
"loss": 0.79,
|
| 1621 |
+
"step": 133500
|
| 1622 |
+
},
|
| 1623 |
+
{
|
| 1624 |
+
"epoch": 1.89,
|
| 1625 |
+
"learning_rate": 4.414559410984412e-05,
|
| 1626 |
+
"loss": 0.7526,
|
| 1627 |
+
"step": 134000
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 1.9,
|
| 1631 |
+
"learning_rate": 4.412190321009318e-05,
|
| 1632 |
+
"loss": 0.765,
|
| 1633 |
+
"step": 134500
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 1.9,
|
| 1637 |
+
"learning_rate": 4.409816483358923e-05,
|
| 1638 |
+
"loss": 0.7554,
|
| 1639 |
+
"step": 135000
|
| 1640 |
+
},
|
| 1641 |
+
{
|
| 1642 |
+
"epoch": 1.91,
|
| 1643 |
+
"learning_rate": 4.407442645708529e-05,
|
| 1644 |
+
"loss": 0.7636,
|
| 1645 |
+
"step": 135500
|
| 1646 |
+
},
|
| 1647 |
+
{
|
| 1648 |
+
"epoch": 1.92,
|
| 1649 |
+
"learning_rate": 4.4050735557334354e-05,
|
| 1650 |
+
"loss": 0.7582,
|
| 1651 |
+
"step": 136000
|
| 1652 |
+
},
|
| 1653 |
+
{
|
| 1654 |
+
"epoch": 1.92,
|
| 1655 |
+
"learning_rate": 4.402699718083041e-05,
|
| 1656 |
+
"loss": 0.7504,
|
| 1657 |
+
"step": 136500
|
| 1658 |
+
},
|
| 1659 |
+
{
|
| 1660 |
+
"epoch": 1.93,
|
| 1661 |
+
"learning_rate": 4.400330628107947e-05,
|
| 1662 |
+
"loss": 0.7591,
|
| 1663 |
+
"step": 137000
|
| 1664 |
+
},
|
| 1665 |
+
{
|
| 1666 |
+
"epoch": 1.94,
|
| 1667 |
+
"learning_rate": 4.3979567904575525e-05,
|
| 1668 |
+
"loss": 0.7629,
|
| 1669 |
+
"step": 137500
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 1.95,
|
| 1673 |
+
"learning_rate": 4.395582952807158e-05,
|
| 1674 |
+
"loss": 0.7543,
|
| 1675 |
+
"step": 138000
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 1.95,
|
| 1679 |
+
"learning_rate": 4.3932091151567636e-05,
|
| 1680 |
+
"loss": 0.7634,
|
| 1681 |
+
"step": 138500
|
| 1682 |
+
},
|
| 1683 |
+
{
|
| 1684 |
+
"epoch": 1.96,
|
| 1685 |
+
"learning_rate": 4.390835277506369e-05,
|
| 1686 |
+
"loss": 0.7553,
|
| 1687 |
+
"step": 139000
|
| 1688 |
+
},
|
| 1689 |
+
{
|
| 1690 |
+
"epoch": 1.97,
|
| 1691 |
+
"learning_rate": 4.388461439855975e-05,
|
| 1692 |
+
"loss": 0.7548,
|
| 1693 |
+
"step": 139500
|
| 1694 |
+
},
|
| 1695 |
+
{
|
| 1696 |
+
"epoch": 1.97,
|
| 1697 |
+
"learning_rate": 4.38608760220558e-05,
|
| 1698 |
+
"loss": 0.7554,
|
| 1699 |
+
"step": 140000
|
| 1700 |
+
},
|
| 1701 |
+
{
|
| 1702 |
+
"epoch": 1.98,
|
| 1703 |
+
"learning_rate": 4.383713764555186e-05,
|
| 1704 |
+
"loss": 0.7539,
|
| 1705 |
+
"step": 140500
|
| 1706 |
+
},
|
| 1707 |
+
{
|
| 1708 |
+
"epoch": 1.99,
|
| 1709 |
+
"learning_rate": 4.3813399269047914e-05,
|
| 1710 |
+
"loss": 0.7562,
|
| 1711 |
+
"step": 141000
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 2.0,
|
| 1715 |
+
"learning_rate": 4.378966089254396e-05,
|
| 1716 |
+
"loss": 0.7537,
|
| 1717 |
+
"step": 141500
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 2.0,
|
| 1721 |
+
"eval_bleu": 42.7907,
|
| 1722 |
+
"eval_gen_len": 13.7941,
|
| 1723 |
+
"eval_loss": 0.8987888693809509,
|
| 1724 |
+
"eval_runtime": 10044.1408,
|
| 1725 |
+
"eval_samples_per_second": 14.121,
|
| 1726 |
+
"eval_steps_per_second": 1.765,
|
| 1727 |
+
"step": 141839
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"epoch": 2.0,
|
| 1731 |
+
"learning_rate": 4.376596999279303e-05,
|
| 1732 |
+
"loss": 0.7511,
|
| 1733 |
+
"step": 142000
|
| 1734 |
+
},
|
| 1735 |
+
{
|
| 1736 |
+
"epoch": 2.01,
|
| 1737 |
+
"learning_rate": 4.3742231616289084e-05,
|
| 1738 |
+
"loss": 0.75,
|
| 1739 |
+
"step": 142500
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 2.02,
|
| 1743 |
+
"learning_rate": 4.371849323978514e-05,
|
| 1744 |
+
"loss": 0.7489,
|
| 1745 |
+
"step": 143000
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 2.02,
|
| 1749 |
+
"learning_rate": 4.3694754863281195e-05,
|
| 1750 |
+
"loss": 0.7359,
|
| 1751 |
+
"step": 143500
|
| 1752 |
+
},
|
| 1753 |
+
{
|
| 1754 |
+
"epoch": 2.03,
|
| 1755 |
+
"learning_rate": 4.367101648677725e-05,
|
| 1756 |
+
"loss": 0.7416,
|
| 1757 |
+
"step": 144000
|
| 1758 |
+
},
|
| 1759 |
+
{
|
| 1760 |
+
"epoch": 2.04,
|
| 1761 |
+
"learning_rate": 4.364732558702632e-05,
|
| 1762 |
+
"loss": 0.7421,
|
| 1763 |
+
"step": 144500
|
| 1764 |
+
},
|
| 1765 |
+
{
|
| 1766 |
+
"epoch": 2.04,
|
| 1767 |
+
"learning_rate": 4.362358721052237e-05,
|
| 1768 |
+
"loss": 0.7448,
|
| 1769 |
+
"step": 145000
|
| 1770 |
+
},
|
| 1771 |
+
{
|
| 1772 |
+
"epoch": 2.05,
|
| 1773 |
+
"learning_rate": 4.359984883401843e-05,
|
| 1774 |
+
"loss": 0.7314,
|
| 1775 |
+
"step": 145500
|
| 1776 |
+
},
|
| 1777 |
+
{
|
| 1778 |
+
"epoch": 2.06,
|
| 1779 |
+
"learning_rate": 4.357611045751448e-05,
|
| 1780 |
+
"loss": 0.731,
|
| 1781 |
+
"step": 146000
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 2.07,
|
| 1785 |
+
"learning_rate": 4.3552419557763544e-05,
|
| 1786 |
+
"loss": 0.7307,
|
| 1787 |
+
"step": 146500
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 2.07,
|
| 1791 |
+
"learning_rate": 4.35286811812596e-05,
|
| 1792 |
+
"loss": 0.7411,
|
| 1793 |
+
"step": 147000
|
| 1794 |
+
},
|
| 1795 |
+
{
|
| 1796 |
+
"epoch": 2.08,
|
| 1797 |
+
"learning_rate": 4.3504942804755655e-05,
|
| 1798 |
+
"loss": 0.7298,
|
| 1799 |
+
"step": 147500
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"epoch": 2.09,
|
| 1803 |
+
"learning_rate": 4.348120442825171e-05,
|
| 1804 |
+
"loss": 0.7144,
|
| 1805 |
+
"step": 148000
|
| 1806 |
+
},
|
| 1807 |
+
{
|
| 1808 |
+
"epoch": 2.09,
|
| 1809 |
+
"learning_rate": 4.3457466051747766e-05,
|
| 1810 |
+
"loss": 0.7146,
|
| 1811 |
+
"step": 148500
|
| 1812 |
+
},
|
| 1813 |
+
{
|
| 1814 |
+
"epoch": 2.1,
|
| 1815 |
+
"learning_rate": 4.343372767524382e-05,
|
| 1816 |
+
"loss": 0.7157,
|
| 1817 |
+
"step": 149000
|
| 1818 |
+
},
|
| 1819 |
+
{
|
| 1820 |
+
"epoch": 2.11,
|
| 1821 |
+
"learning_rate": 4.341003677549288e-05,
|
| 1822 |
+
"loss": 0.7116,
|
| 1823 |
+
"step": 149500
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 2.12,
|
| 1827 |
+
"learning_rate": 4.338629839898894e-05,
|
| 1828 |
+
"loss": 0.7105,
|
| 1829 |
+
"step": 150000
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 2.12,
|
| 1833 |
+
"learning_rate": 4.336256002248499e-05,
|
| 1834 |
+
"loss": 0.7096,
|
| 1835 |
+
"step": 150500
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"epoch": 2.13,
|
| 1839 |
+
"learning_rate": 4.3338916599487063e-05,
|
| 1840 |
+
"loss": 0.7026,
|
| 1841 |
+
"step": 151000
|
| 1842 |
+
},
|
| 1843 |
+
{
|
| 1844 |
+
"epoch": 2.14,
|
| 1845 |
+
"learning_rate": 4.331517822298312e-05,
|
| 1846 |
+
"loss": 0.706,
|
| 1847 |
+
"step": 151500
|
| 1848 |
+
},
|
| 1849 |
+
{
|
| 1850 |
+
"epoch": 2.14,
|
| 1851 |
+
"learning_rate": 4.3291439846479174e-05,
|
| 1852 |
+
"loss": 0.6996,
|
| 1853 |
+
"step": 152000
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"epoch": 2.15,
|
| 1857 |
+
"learning_rate": 4.326770146997523e-05,
|
| 1858 |
+
"loss": 0.6965,
|
| 1859 |
+
"step": 152500
|
| 1860 |
+
},
|
| 1861 |
+
{
|
| 1862 |
+
"epoch": 2.16,
|
| 1863 |
+
"learning_rate": 4.3243963093471285e-05,
|
| 1864 |
+
"loss": 0.699,
|
| 1865 |
+
"step": 153000
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 2.16,
|
| 1869 |
+
"learning_rate": 4.322022471696734e-05,
|
| 1870 |
+
"loss": 0.6926,
|
| 1871 |
+
"step": 153500
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 2.17,
|
| 1875 |
+
"learning_rate": 4.3196486340463396e-05,
|
| 1876 |
+
"loss": 0.6925,
|
| 1877 |
+
"step": 154000
|
| 1878 |
+
},
|
| 1879 |
+
{
|
| 1880 |
+
"epoch": 2.18,
|
| 1881 |
+
"learning_rate": 4.3172747963959445e-05,
|
| 1882 |
+
"loss": 0.6825,
|
| 1883 |
+
"step": 154500
|
| 1884 |
+
},
|
| 1885 |
+
{
|
| 1886 |
+
"epoch": 2.19,
|
| 1887 |
+
"learning_rate": 4.31490095874555e-05,
|
| 1888 |
+
"loss": 0.6869,
|
| 1889 |
+
"step": 155000
|
| 1890 |
+
},
|
| 1891 |
+
{
|
| 1892 |
+
"epoch": 2.19,
|
| 1893 |
+
"learning_rate": 4.312527121095156e-05,
|
| 1894 |
+
"loss": 0.691,
|
| 1895 |
+
"step": 155500
|
| 1896 |
+
},
|
| 1897 |
+
{
|
| 1898 |
+
"epoch": 2.2,
|
| 1899 |
+
"learning_rate": 4.310153283444762e-05,
|
| 1900 |
+
"loss": 0.682,
|
| 1901 |
+
"step": 156000
|
| 1902 |
+
},
|
| 1903 |
+
{
|
| 1904 |
+
"epoch": 2.21,
|
| 1905 |
+
"learning_rate": 4.3077794457943673e-05,
|
| 1906 |
+
"loss": 0.6822,
|
| 1907 |
+
"step": 156500
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"epoch": 2.21,
|
| 1911 |
+
"learning_rate": 4.305405608143972e-05,
|
| 1912 |
+
"loss": 0.6666,
|
| 1913 |
+
"step": 157000
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 2.22,
|
| 1917 |
+
"learning_rate": 4.303031770493578e-05,
|
| 1918 |
+
"loss": 0.6749,
|
| 1919 |
+
"step": 157500
|
| 1920 |
+
},
|
| 1921 |
+
{
|
| 1922 |
+
"epoch": 2.23,
|
| 1923 |
+
"learning_rate": 4.300657932843183e-05,
|
| 1924 |
+
"loss": 0.6903,
|
| 1925 |
+
"step": 158000
|
| 1926 |
+
},
|
| 1927 |
+
{
|
| 1928 |
+
"epoch": 2.23,
|
| 1929 |
+
"learning_rate": 4.29828884286809e-05,
|
| 1930 |
+
"loss": 0.6816,
|
| 1931 |
+
"step": 158500
|
| 1932 |
+
},
|
| 1933 |
+
{
|
| 1934 |
+
"epoch": 2.24,
|
| 1935 |
+
"learning_rate": 4.2959150052176955e-05,
|
| 1936 |
+
"loss": 0.6776,
|
| 1937 |
+
"step": 159000
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"epoch": 2.25,
|
| 1941 |
+
"learning_rate": 4.2935459152426015e-05,
|
| 1942 |
+
"loss": 0.6787,
|
| 1943 |
+
"step": 159500
|
| 1944 |
+
},
|
| 1945 |
+
{
|
| 1946 |
+
"epoch": 2.26,
|
| 1947 |
+
"learning_rate": 4.2911768252675075e-05,
|
| 1948 |
+
"loss": 0.6559,
|
| 1949 |
+
"step": 160000
|
| 1950 |
+
},
|
| 1951 |
+
{
|
| 1952 |
+
"epoch": 2.26,
|
| 1953 |
+
"learning_rate": 4.288807735292414e-05,
|
| 1954 |
+
"loss": 0.668,
|
| 1955 |
+
"step": 160500
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 2.27,
|
| 1959 |
+
"learning_rate": 4.28643389764202e-05,
|
| 1960 |
+
"loss": 0.6755,
|
| 1961 |
+
"step": 161000
|
| 1962 |
+
},
|
| 1963 |
+
{
|
| 1964 |
+
"epoch": 2.28,
|
| 1965 |
+
"learning_rate": 4.284060059991625e-05,
|
| 1966 |
+
"loss": 0.6751,
|
| 1967 |
+
"step": 161500
|
| 1968 |
+
},
|
| 1969 |
+
{
|
| 1970 |
+
"epoch": 2.28,
|
| 1971 |
+
"learning_rate": 4.281686222341231e-05,
|
| 1972 |
+
"loss": 0.6608,
|
| 1973 |
+
"step": 162000
|
| 1974 |
+
},
|
| 1975 |
+
{
|
| 1976 |
+
"epoch": 2.29,
|
| 1977 |
+
"learning_rate": 4.2793123846908364e-05,
|
| 1978 |
+
"loss": 0.6562,
|
| 1979 |
+
"step": 162500
|
| 1980 |
+
},
|
| 1981 |
+
{
|
| 1982 |
+
"epoch": 2.3,
|
| 1983 |
+
"learning_rate": 4.276938547040442e-05,
|
| 1984 |
+
"loss": 0.6672,
|
| 1985 |
+
"step": 163000
|
| 1986 |
+
},
|
| 1987 |
+
{
|
| 1988 |
+
"epoch": 2.31,
|
| 1989 |
+
"learning_rate": 4.2745647093900475e-05,
|
| 1990 |
+
"loss": 0.6577,
|
| 1991 |
+
"step": 163500
|
| 1992 |
+
},
|
| 1993 |
+
{
|
| 1994 |
+
"epoch": 2.31,
|
| 1995 |
+
"learning_rate": 4.272190871739653e-05,
|
| 1996 |
+
"loss": 0.6714,
|
| 1997 |
+
"step": 164000
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 2.32,
|
| 2001 |
+
"learning_rate": 4.2698170340892586e-05,
|
| 2002 |
+
"loss": 0.6706,
|
| 2003 |
+
"step": 164500
|
| 2004 |
+
},
|
| 2005 |
+
{
|
| 2006 |
+
"epoch": 2.33,
|
| 2007 |
+
"learning_rate": 4.2674479441141646e-05,
|
| 2008 |
+
"loss": 0.6593,
|
| 2009 |
+
"step": 165000
|
| 2010 |
+
},
|
| 2011 |
+
{
|
| 2012 |
+
"epoch": 2.33,
|
| 2013 |
+
"learning_rate": 4.26507410646377e-05,
|
| 2014 |
+
"loss": 0.6633,
|
| 2015 |
+
"step": 165500
|
| 2016 |
+
},
|
| 2017 |
+
{
|
| 2018 |
+
"epoch": 2.34,
|
| 2019 |
+
"learning_rate": 4.262705016488677e-05,
|
| 2020 |
+
"loss": 0.6672,
|
| 2021 |
+
"step": 166000
|
| 2022 |
+
},
|
| 2023 |
+
{
|
| 2024 |
+
"epoch": 2.35,
|
| 2025 |
+
"learning_rate": 4.260331178838282e-05,
|
| 2026 |
+
"loss": 0.6584,
|
| 2027 |
+
"step": 166500
|
| 2028 |
+
},
|
| 2029 |
+
{
|
| 2030 |
+
"epoch": 2.35,
|
| 2031 |
+
"learning_rate": 4.257957341187888e-05,
|
| 2032 |
+
"loss": 0.6593,
|
| 2033 |
+
"step": 167000
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 2.36,
|
| 2037 |
+
"learning_rate": 4.255583503537493e-05,
|
| 2038 |
+
"loss": 0.6651,
|
| 2039 |
+
"step": 167500
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 2.37,
|
| 2043 |
+
"learning_rate": 4.253209665887098e-05,
|
| 2044 |
+
"loss": 0.6519,
|
| 2045 |
+
"step": 168000
|
| 2046 |
+
},
|
| 2047 |
+
{
|
| 2048 |
+
"epoch": 2.38,
|
| 2049 |
+
"learning_rate": 4.250835828236704e-05,
|
| 2050 |
+
"loss": 0.6518,
|
| 2051 |
+
"step": 168500
|
| 2052 |
+
},
|
| 2053 |
+
{
|
| 2054 |
+
"epoch": 2.38,
|
| 2055 |
+
"learning_rate": 4.24846199058631e-05,
|
| 2056 |
+
"loss": 0.6598,
|
| 2057 |
+
"step": 169000
|
| 2058 |
+
},
|
| 2059 |
+
{
|
| 2060 |
+
"epoch": 2.39,
|
| 2061 |
+
"learning_rate": 4.2460881529359156e-05,
|
| 2062 |
+
"loss": 0.6445,
|
| 2063 |
+
"step": 169500
|
| 2064 |
+
},
|
| 2065 |
+
{
|
| 2066 |
+
"epoch": 2.4,
|
| 2067 |
+
"learning_rate": 4.2437143152855205e-05,
|
| 2068 |
+
"loss": 0.6499,
|
| 2069 |
+
"step": 170000
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"epoch": 2.4,
|
| 2073 |
+
"learning_rate": 4.241340477635126e-05,
|
| 2074 |
+
"loss": 0.6516,
|
| 2075 |
+
"step": 170500
|
| 2076 |
+
},
|
| 2077 |
+
{
|
| 2078 |
+
"epoch": 2.41,
|
| 2079 |
+
"learning_rate": 4.2389666399847316e-05,
|
| 2080 |
+
"loss": 0.6553,
|
| 2081 |
+
"step": 171000
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 2.42,
|
| 2085 |
+
"learning_rate": 4.236592802334337e-05,
|
| 2086 |
+
"loss": 0.668,
|
| 2087 |
+
"step": 171500
|
| 2088 |
+
},
|
| 2089 |
+
{
|
| 2090 |
+
"epoch": 2.43,
|
| 2091 |
+
"learning_rate": 4.2342189646839427e-05,
|
| 2092 |
+
"loss": 0.6447,
|
| 2093 |
+
"step": 172000
|
| 2094 |
+
},
|
| 2095 |
+
{
|
| 2096 |
+
"epoch": 2.43,
|
| 2097 |
+
"learning_rate": 4.231845127033548e-05,
|
| 2098 |
+
"loss": 0.6474,
|
| 2099 |
+
"step": 172500
|
| 2100 |
+
},
|
| 2101 |
+
{
|
| 2102 |
+
"epoch": 2.44,
|
| 2103 |
+
"learning_rate": 4.229471289383154e-05,
|
| 2104 |
+
"loss": 0.6594,
|
| 2105 |
+
"step": 173000
|
| 2106 |
+
},
|
| 2107 |
+
{
|
| 2108 |
+
"epoch": 2.45,
|
| 2109 |
+
"learning_rate": 4.227097451732759e-05,
|
| 2110 |
+
"loss": 0.6516,
|
| 2111 |
+
"step": 173500
|
| 2112 |
+
},
|
| 2113 |
+
{
|
| 2114 |
+
"epoch": 2.45,
|
| 2115 |
+
"learning_rate": 4.224723614082365e-05,
|
| 2116 |
+
"loss": 0.6404,
|
| 2117 |
+
"step": 174000
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 2.46,
|
| 2121 |
+
"learning_rate": 4.222354524107271e-05,
|
| 2122 |
+
"loss": 0.6526,
|
| 2123 |
+
"step": 174500
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 2.47,
|
| 2127 |
+
"learning_rate": 4.2199806864568764e-05,
|
| 2128 |
+
"loss": 0.6459,
|
| 2129 |
+
"step": 175000
|
| 2130 |
+
},
|
| 2131 |
+
{
|
| 2132 |
+
"epoch": 2.47,
|
| 2133 |
+
"learning_rate": 4.2176068488064826e-05,
|
| 2134 |
+
"loss": 0.6519,
|
| 2135 |
+
"step": 175500
|
| 2136 |
+
},
|
| 2137 |
+
{
|
| 2138 |
+
"epoch": 2.48,
|
| 2139 |
+
"learning_rate": 4.2152377588313886e-05,
|
| 2140 |
+
"loss": 0.6391,
|
| 2141 |
+
"step": 176000
|
| 2142 |
+
},
|
| 2143 |
+
{
|
| 2144 |
+
"epoch": 2.49,
|
| 2145 |
+
"learning_rate": 4.212863921180994e-05,
|
| 2146 |
+
"loss": 0.6346,
|
| 2147 |
+
"step": 176500
|
| 2148 |
+
},
|
| 2149 |
+
{
|
| 2150 |
+
"epoch": 2.5,
|
| 2151 |
+
"learning_rate": 4.2104900835306e-05,
|
| 2152 |
+
"loss": 0.6395,
|
| 2153 |
+
"step": 177000
|
| 2154 |
+
},
|
| 2155 |
+
{
|
| 2156 |
+
"epoch": 2.5,
|
| 2157 |
+
"learning_rate": 4.2081162458802046e-05,
|
| 2158 |
+
"loss": 0.6409,
|
| 2159 |
+
"step": 177500
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"epoch": 2.51,
|
| 2163 |
+
"learning_rate": 4.20574240822981e-05,
|
| 2164 |
+
"loss": 0.6285,
|
| 2165 |
+
"step": 178000
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 2.52,
|
| 2169 |
+
"learning_rate": 4.2033685705794163e-05,
|
| 2170 |
+
"loss": 0.6307,
|
| 2171 |
+
"step": 178500
|
| 2172 |
+
},
|
| 2173 |
+
{
|
| 2174 |
+
"epoch": 2.52,
|
| 2175 |
+
"learning_rate": 4.200999480604322e-05,
|
| 2176 |
+
"loss": 0.6506,
|
| 2177 |
+
"step": 179000
|
| 2178 |
+
},
|
| 2179 |
+
{
|
| 2180 |
+
"epoch": 2.53,
|
| 2181 |
+
"learning_rate": 4.198625642953928e-05,
|
| 2182 |
+
"loss": 0.6307,
|
| 2183 |
+
"step": 179500
|
| 2184 |
+
},
|
| 2185 |
+
{
|
| 2186 |
+
"epoch": 2.54,
|
| 2187 |
+
"learning_rate": 4.1962518053035334e-05,
|
| 2188 |
+
"loss": 0.6446,
|
| 2189 |
+
"step": 180000
|
| 2190 |
+
},
|
| 2191 |
+
{
|
| 2192 |
+
"epoch": 2.55,
|
| 2193 |
+
"learning_rate": 4.193877967653139e-05,
|
| 2194 |
+
"loss": 0.6422,
|
| 2195 |
+
"step": 180500
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"epoch": 2.55,
|
| 2199 |
+
"learning_rate": 4.1915088776780457e-05,
|
| 2200 |
+
"loss": 0.6398,
|
| 2201 |
+
"step": 181000
|
| 2202 |
+
},
|
| 2203 |
+
{
|
| 2204 |
+
"epoch": 2.56,
|
| 2205 |
+
"learning_rate": 4.1891397877029516e-05,
|
| 2206 |
+
"loss": 0.6333,
|
| 2207 |
+
"step": 181500
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 2.57,
|
| 2211 |
+
"learning_rate": 4.186765950052557e-05,
|
| 2212 |
+
"loss": 0.6363,
|
| 2213 |
+
"step": 182000
|
| 2214 |
+
},
|
| 2215 |
+
{
|
| 2216 |
+
"epoch": 2.57,
|
| 2217 |
+
"learning_rate": 4.184392112402162e-05,
|
| 2218 |
+
"loss": 0.6319,
|
| 2219 |
+
"step": 182500
|
| 2220 |
+
},
|
| 2221 |
+
{
|
| 2222 |
+
"epoch": 2.58,
|
| 2223 |
+
"learning_rate": 4.1820182747517676e-05,
|
| 2224 |
+
"loss": 0.6294,
|
| 2225 |
+
"step": 183000
|
| 2226 |
+
},
|
| 2227 |
+
{
|
| 2228 |
+
"epoch": 2.59,
|
| 2229 |
+
"learning_rate": 4.179649184776674e-05,
|
| 2230 |
+
"loss": 0.6366,
|
| 2231 |
+
"step": 183500
|
| 2232 |
+
},
|
| 2233 |
+
{
|
| 2234 |
+
"epoch": 2.59,
|
| 2235 |
+
"learning_rate": 4.17727534712628e-05,
|
| 2236 |
+
"loss": 0.6384,
|
| 2237 |
+
"step": 184000
|
| 2238 |
+
},
|
| 2239 |
+
{
|
| 2240 |
+
"epoch": 2.6,
|
| 2241 |
+
"learning_rate": 4.1749015094758854e-05,
|
| 2242 |
+
"loss": 0.6365,
|
| 2243 |
+
"step": 184500
|
| 2244 |
+
},
|
| 2245 |
+
{
|
| 2246 |
+
"epoch": 2.61,
|
| 2247 |
+
"learning_rate": 4.172527671825491e-05,
|
| 2248 |
+
"loss": 0.6331,
|
| 2249 |
+
"step": 185000
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 2.62,
|
| 2253 |
+
"learning_rate": 4.1701538341750965e-05,
|
| 2254 |
+
"loss": 0.6327,
|
| 2255 |
+
"step": 185500
|
| 2256 |
+
},
|
| 2257 |
+
{
|
| 2258 |
+
"epoch": 2.62,
|
| 2259 |
+
"learning_rate": 4.167779996524701e-05,
|
| 2260 |
+
"loss": 0.628,
|
| 2261 |
+
"step": 186000
|
| 2262 |
+
},
|
| 2263 |
+
{
|
| 2264 |
+
"epoch": 2.63,
|
| 2265 |
+
"learning_rate": 4.1654061588743076e-05,
|
| 2266 |
+
"loss": 0.6309,
|
| 2267 |
+
"step": 186500
|
| 2268 |
+
},
|
| 2269 |
+
{
|
| 2270 |
+
"epoch": 2.64,
|
| 2271 |
+
"learning_rate": 4.163032321223913e-05,
|
| 2272 |
+
"loss": 0.6159,
|
| 2273 |
+
"step": 187000
|
| 2274 |
+
},
|
| 2275 |
+
{
|
| 2276 |
+
"epoch": 2.64,
|
| 2277 |
+
"learning_rate": 4.1606584835735187e-05,
|
| 2278 |
+
"loss": 0.6307,
|
| 2279 |
+
"step": 187500
|
| 2280 |
+
},
|
| 2281 |
+
{
|
| 2282 |
+
"epoch": 2.65,
|
| 2283 |
+
"learning_rate": 4.158284645923124e-05,
|
| 2284 |
+
"loss": 0.6356,
|
| 2285 |
+
"step": 188000
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"epoch": 2.66,
|
| 2289 |
+
"learning_rate": 4.155910808272729e-05,
|
| 2290 |
+
"loss": 0.6269,
|
| 2291 |
+
"step": 188500
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 2.66,
|
| 2295 |
+
"learning_rate": 4.1535369706223346e-05,
|
| 2296 |
+
"loss": 0.6368,
|
| 2297 |
+
"step": 189000
|
| 2298 |
+
},
|
| 2299 |
+
{
|
| 2300 |
+
"epoch": 2.67,
|
| 2301 |
+
"learning_rate": 4.15116313297194e-05,
|
| 2302 |
+
"loss": 0.6282,
|
| 2303 |
+
"step": 189500
|
| 2304 |
+
},
|
| 2305 |
+
{
|
| 2306 |
+
"epoch": 2.68,
|
| 2307 |
+
"learning_rate": 4.1487892953215464e-05,
|
| 2308 |
+
"loss": 0.6226,
|
| 2309 |
+
"step": 190000
|
| 2310 |
+
},
|
| 2311 |
+
{
|
| 2312 |
+
"epoch": 2.69,
|
| 2313 |
+
"learning_rate": 4.146415457671152e-05,
|
| 2314 |
+
"loss": 0.6181,
|
| 2315 |
+
"step": 190500
|
| 2316 |
+
},
|
| 2317 |
+
{
|
| 2318 |
+
"epoch": 2.69,
|
| 2319 |
+
"learning_rate": 4.144041620020757e-05,
|
| 2320 |
+
"loss": 0.6138,
|
| 2321 |
+
"step": 191000
|
| 2322 |
+
},
|
| 2323 |
+
{
|
| 2324 |
+
"epoch": 2.7,
|
| 2325 |
+
"learning_rate": 4.1416677823703623e-05,
|
| 2326 |
+
"loss": 0.6184,
|
| 2327 |
+
"step": 191500
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 2.71,
|
| 2331 |
+
"learning_rate": 4.139298692395269e-05,
|
| 2332 |
+
"loss": 0.624,
|
| 2333 |
+
"step": 192000
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 2.71,
|
| 2337 |
+
"learning_rate": 4.136924854744874e-05,
|
| 2338 |
+
"loss": 0.6125,
|
| 2339 |
+
"step": 192500
|
| 2340 |
+
},
|
| 2341 |
+
{
|
| 2342 |
+
"epoch": 2.72,
|
| 2343 |
+
"learning_rate": 4.13455101709448e-05,
|
| 2344 |
+
"loss": 0.6261,
|
| 2345 |
+
"step": 193000
|
| 2346 |
+
},
|
| 2347 |
+
{
|
| 2348 |
+
"epoch": 2.73,
|
| 2349 |
+
"learning_rate": 4.1321771794440857e-05,
|
| 2350 |
+
"loss": 0.6189,
|
| 2351 |
+
"step": 193500
|
| 2352 |
+
},
|
| 2353 |
+
{
|
| 2354 |
+
"epoch": 2.74,
|
| 2355 |
+
"learning_rate": 4.129803341793691e-05,
|
| 2356 |
+
"loss": 0.6149,
|
| 2357 |
+
"step": 194000
|
| 2358 |
+
},
|
| 2359 |
+
{
|
| 2360 |
+
"epoch": 2.74,
|
| 2361 |
+
"learning_rate": 4.127429504143297e-05,
|
| 2362 |
+
"loss": 0.6296,
|
| 2363 |
+
"step": 194500
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 2.75,
|
| 2367 |
+
"learning_rate": 4.1250556664929016e-05,
|
| 2368 |
+
"loss": 0.6263,
|
| 2369 |
+
"step": 195000
|
| 2370 |
+
},
|
| 2371 |
+
{
|
| 2372 |
+
"epoch": 2.76,
|
| 2373 |
+
"learning_rate": 4.122681828842507e-05,
|
| 2374 |
+
"loss": 0.6298,
|
| 2375 |
+
"step": 195500
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 2.76,
|
| 2379 |
+
"learning_rate": 4.120312738867414e-05,
|
| 2380 |
+
"loss": 0.6102,
|
| 2381 |
+
"step": 196000
|
| 2382 |
+
},
|
| 2383 |
+
{
|
| 2384 |
+
"epoch": 2.77,
|
| 2385 |
+
"learning_rate": 4.1179389012170194e-05,
|
| 2386 |
+
"loss": 0.6248,
|
| 2387 |
+
"step": 196500
|
| 2388 |
+
},
|
| 2389 |
+
{
|
| 2390 |
+
"epoch": 2.78,
|
| 2391 |
+
"learning_rate": 4.115565063566625e-05,
|
| 2392 |
+
"loss": 0.6121,
|
| 2393 |
+
"step": 197000
|
| 2394 |
+
},
|
| 2395 |
+
{
|
| 2396 |
+
"epoch": 2.78,
|
| 2397 |
+
"learning_rate": 4.1131912259162305e-05,
|
| 2398 |
+
"loss": 0.6156,
|
| 2399 |
+
"step": 197500
|
| 2400 |
+
},
|
| 2401 |
+
{
|
| 2402 |
+
"epoch": 2.79,
|
| 2403 |
+
"learning_rate": 4.1108221359411365e-05,
|
| 2404 |
+
"loss": 0.6076,
|
| 2405 |
+
"step": 198000
|
| 2406 |
+
},
|
| 2407 |
+
{
|
| 2408 |
+
"epoch": 2.8,
|
| 2409 |
+
"learning_rate": 4.1084577936413436e-05,
|
| 2410 |
+
"loss": 0.6231,
|
| 2411 |
+
"step": 198500
|
| 2412 |
+
},
|
| 2413 |
+
{
|
| 2414 |
+
"epoch": 2.81,
|
| 2415 |
+
"learning_rate": 4.106083955990949e-05,
|
| 2416 |
+
"loss": 0.6281,
|
| 2417 |
+
"step": 199000
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 2.81,
|
| 2421 |
+
"learning_rate": 4.103710118340555e-05,
|
| 2422 |
+
"loss": 0.6157,
|
| 2423 |
+
"step": 199500
|
| 2424 |
+
},
|
| 2425 |
+
{
|
| 2426 |
+
"epoch": 2.82,
|
| 2427 |
+
"learning_rate": 4.10133628069016e-05,
|
| 2428 |
+
"loss": 0.6086,
|
| 2429 |
+
"step": 200000
|
| 2430 |
+
},
|
| 2431 |
+
{
|
| 2432 |
+
"epoch": 2.83,
|
| 2433 |
+
"learning_rate": 4.098962443039766e-05,
|
| 2434 |
+
"loss": 0.6225,
|
| 2435 |
+
"step": 200500
|
| 2436 |
+
},
|
| 2437 |
+
{
|
| 2438 |
+
"epoch": 2.83,
|
| 2439 |
+
"learning_rate": 4.096588605389371e-05,
|
| 2440 |
+
"loss": 0.6078,
|
| 2441 |
+
"step": 201000
|
| 2442 |
+
},
|
| 2443 |
+
{
|
| 2444 |
+
"epoch": 2.84,
|
| 2445 |
+
"learning_rate": 4.094214767738977e-05,
|
| 2446 |
+
"loss": 0.6217,
|
| 2447 |
+
"step": 201500
|
| 2448 |
+
},
|
| 2449 |
+
{
|
| 2450 |
+
"epoch": 2.85,
|
| 2451 |
+
"learning_rate": 4.0918409300885824e-05,
|
| 2452 |
+
"loss": 0.6096,
|
| 2453 |
+
"step": 202000
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"epoch": 2.86,
|
| 2457 |
+
"learning_rate": 4.089467092438188e-05,
|
| 2458 |
+
"loss": 0.6253,
|
| 2459 |
+
"step": 202500
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 2.86,
|
| 2463 |
+
"learning_rate": 4.087102750138395e-05,
|
| 2464 |
+
"loss": 0.6068,
|
| 2465 |
+
"step": 203000
|
| 2466 |
+
},
|
| 2467 |
+
{
|
| 2468 |
+
"epoch": 2.87,
|
| 2469 |
+
"learning_rate": 4.0847289124880006e-05,
|
| 2470 |
+
"loss": 0.613,
|
| 2471 |
+
"step": 203500
|
| 2472 |
+
},
|
| 2473 |
+
{
|
| 2474 |
+
"epoch": 2.88,
|
| 2475 |
+
"learning_rate": 4.082355074837606e-05,
|
| 2476 |
+
"loss": 0.6179,
|
| 2477 |
+
"step": 204000
|
| 2478 |
+
},
|
| 2479 |
+
{
|
| 2480 |
+
"epoch": 2.88,
|
| 2481 |
+
"learning_rate": 4.079981237187212e-05,
|
| 2482 |
+
"loss": 0.6312,
|
| 2483 |
+
"step": 204500
|
| 2484 |
+
},
|
| 2485 |
+
{
|
| 2486 |
+
"epoch": 2.89,
|
| 2487 |
+
"learning_rate": 4.077612147212118e-05,
|
| 2488 |
+
"loss": 0.601,
|
| 2489 |
+
"step": 205000
|
| 2490 |
+
},
|
| 2491 |
+
{
|
| 2492 |
+
"epoch": 2.9,
|
| 2493 |
+
"learning_rate": 4.075238309561723e-05,
|
| 2494 |
+
"loss": 0.615,
|
| 2495 |
+
"step": 205500
|
| 2496 |
+
},
|
| 2497 |
+
{
|
| 2498 |
+
"epoch": 2.9,
|
| 2499 |
+
"learning_rate": 4.072864471911329e-05,
|
| 2500 |
+
"loss": 0.6071,
|
| 2501 |
+
"step": 206000
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 2.91,
|
| 2505 |
+
"learning_rate": 4.0704906342609344e-05,
|
| 2506 |
+
"loss": 0.6135,
|
| 2507 |
+
"step": 206500
|
| 2508 |
+
},
|
| 2509 |
+
{
|
| 2510 |
+
"epoch": 2.92,
|
| 2511 |
+
"learning_rate": 4.06811679661054e-05,
|
| 2512 |
+
"loss": 0.6073,
|
| 2513 |
+
"step": 207000
|
| 2514 |
+
},
|
| 2515 |
+
{
|
| 2516 |
+
"epoch": 2.93,
|
| 2517 |
+
"learning_rate": 4.065747706635446e-05,
|
| 2518 |
+
"loss": 0.6073,
|
| 2519 |
+
"step": 207500
|
| 2520 |
+
},
|
| 2521 |
+
{
|
| 2522 |
+
"epoch": 2.93,
|
| 2523 |
+
"learning_rate": 4.0633738689850515e-05,
|
| 2524 |
+
"loss": 0.6053,
|
| 2525 |
+
"step": 208000
|
| 2526 |
+
},
|
| 2527 |
+
{
|
| 2528 |
+
"epoch": 2.94,
|
| 2529 |
+
"learning_rate": 4.061000031334658e-05,
|
| 2530 |
+
"loss": 0.6116,
|
| 2531 |
+
"step": 208500
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 2.95,
|
| 2535 |
+
"learning_rate": 4.058626193684263e-05,
|
| 2536 |
+
"loss": 0.606,
|
| 2537 |
+
"step": 209000
|
| 2538 |
+
},
|
| 2539 |
+
{
|
| 2540 |
+
"epoch": 2.95,
|
| 2541 |
+
"learning_rate": 4.056252356033868e-05,
|
| 2542 |
+
"loss": 0.6183,
|
| 2543 |
+
"step": 209500
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 2.96,
|
| 2547 |
+
"learning_rate": 4.0538785183834736e-05,
|
| 2548 |
+
"loss": 0.601,
|
| 2549 |
+
"step": 210000
|
| 2550 |
+
},
|
| 2551 |
+
{
|
| 2552 |
+
"epoch": 2.97,
|
| 2553 |
+
"learning_rate": 4.051504680733079e-05,
|
| 2554 |
+
"loss": 0.6041,
|
| 2555 |
+
"step": 210500
|
| 2556 |
+
},
|
| 2557 |
+
{
|
| 2558 |
+
"epoch": 2.98,
|
| 2559 |
+
"learning_rate": 4.049130843082685e-05,
|
| 2560 |
+
"loss": 0.6073,
|
| 2561 |
+
"step": 211000
|
| 2562 |
+
},
|
| 2563 |
+
{
|
| 2564 |
+
"epoch": 2.98,
|
| 2565 |
+
"learning_rate": 4.04675700543229e-05,
|
| 2566 |
+
"loss": 0.6106,
|
| 2567 |
+
"step": 211500
|
| 2568 |
+
},
|
| 2569 |
+
{
|
| 2570 |
+
"epoch": 2.99,
|
| 2571 |
+
"learning_rate": 4.044383167781896e-05,
|
| 2572 |
+
"loss": 0.6105,
|
| 2573 |
+
"step": 212000
|
| 2574 |
+
},
|
| 2575 |
+
{
|
| 2576 |
+
"epoch": 3.0,
|
| 2577 |
+
"learning_rate": 4.0420093301315014e-05,
|
| 2578 |
+
"loss": 0.5973,
|
| 2579 |
+
"step": 212500
|
| 2580 |
+
},
|
| 2581 |
+
{
|
| 2582 |
+
"epoch": 3.0,
|
| 2583 |
+
"eval_bleu": 43.4697,
|
| 2584 |
+
"eval_gen_len": 13.6556,
|
| 2585 |
+
"eval_loss": 0.9142627120018005,
|
| 2586 |
+
"eval_runtime": 9819.0968,
|
| 2587 |
+
"eval_samples_per_second": 14.445,
|
| 2588 |
+
"eval_steps_per_second": 1.806,
|
| 2589 |
+
"step": 212758
|
| 2590 |
+
},
|
| 2591 |
+
{
|
| 2592 |
+
"epoch": 3.0,
|
| 2593 |
+
"learning_rate": 4.039635492481107e-05,
|
| 2594 |
+
"loss": 0.6014,
|
| 2595 |
+
"step": 213000
|
| 2596 |
+
},
|
| 2597 |
+
{
|
| 2598 |
+
"epoch": 3.01,
|
| 2599 |
+
"learning_rate": 4.037266402506013e-05,
|
| 2600 |
+
"loss": 0.6099,
|
| 2601 |
+
"step": 213500
|
| 2602 |
+
},
|
| 2603 |
+
{
|
| 2604 |
+
"epoch": 3.02,
|
| 2605 |
+
"learning_rate": 4.0348925648556185e-05,
|
| 2606 |
+
"loss": 0.5951,
|
| 2607 |
+
"step": 214000
|
| 2608 |
+
},
|
| 2609 |
+
{
|
| 2610 |
+
"epoch": 3.02,
|
| 2611 |
+
"learning_rate": 4.032518727205224e-05,
|
| 2612 |
+
"loss": 0.597,
|
| 2613 |
+
"step": 214500
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 3.03,
|
| 2617 |
+
"learning_rate": 4.03014488955483e-05,
|
| 2618 |
+
"loss": 0.596,
|
| 2619 |
+
"step": 215000
|
| 2620 |
+
},
|
| 2621 |
+
{
|
| 2622 |
+
"epoch": 3.04,
|
| 2623 |
+
"learning_rate": 4.027771051904435e-05,
|
| 2624 |
+
"loss": 0.5914,
|
| 2625 |
+
"step": 215500
|
| 2626 |
+
},
|
| 2627 |
+
{
|
| 2628 |
+
"epoch": 3.05,
|
| 2629 |
+
"learning_rate": 4.0253972142540406e-05,
|
| 2630 |
+
"loss": 0.5979,
|
| 2631 |
+
"step": 216000
|
| 2632 |
+
},
|
| 2633 |
+
{
|
| 2634 |
+
"epoch": 3.05,
|
| 2635 |
+
"learning_rate": 4.023023376603646e-05,
|
| 2636 |
+
"loss": 0.5855,
|
| 2637 |
+
"step": 216500
|
| 2638 |
+
},
|
| 2639 |
+
{
|
| 2640 |
+
"epoch": 3.06,
|
| 2641 |
+
"learning_rate": 4.020649538953252e-05,
|
| 2642 |
+
"loss": 0.5887,
|
| 2643 |
+
"step": 217000
|
| 2644 |
+
},
|
| 2645 |
+
{
|
| 2646 |
+
"epoch": 3.07,
|
| 2647 |
+
"learning_rate": 4.018275701302857e-05,
|
| 2648 |
+
"loss": 0.5853,
|
| 2649 |
+
"step": 217500
|
| 2650 |
+
},
|
| 2651 |
+
{
|
| 2652 |
+
"epoch": 3.07,
|
| 2653 |
+
"learning_rate": 4.015901863652463e-05,
|
| 2654 |
+
"loss": 0.5921,
|
| 2655 |
+
"step": 218000
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 3.08,
|
| 2659 |
+
"learning_rate": 4.013528026002068e-05,
|
| 2660 |
+
"loss": 0.5873,
|
| 2661 |
+
"step": 218500
|
| 2662 |
+
},
|
| 2663 |
+
{
|
| 2664 |
+
"epoch": 3.09,
|
| 2665 |
+
"learning_rate": 4.0111589360269744e-05,
|
| 2666 |
+
"loss": 0.5708,
|
| 2667 |
+
"step": 219000
|
| 2668 |
+
},
|
| 2669 |
+
{
|
| 2670 |
+
"epoch": 3.1,
|
| 2671 |
+
"learning_rate": 4.00878509837658e-05,
|
| 2672 |
+
"loss": 0.5777,
|
| 2673 |
+
"step": 219500
|
| 2674 |
+
},
|
| 2675 |
+
{
|
| 2676 |
+
"epoch": 3.1,
|
| 2677 |
+
"learning_rate": 4.0064112607261855e-05,
|
| 2678 |
+
"loss": 0.5772,
|
| 2679 |
+
"step": 220000
|
| 2680 |
+
},
|
| 2681 |
+
{
|
| 2682 |
+
"epoch": 3.11,
|
| 2683 |
+
"learning_rate": 4.004037423075791e-05,
|
| 2684 |
+
"loss": 0.573,
|
| 2685 |
+
"step": 220500
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 3.12,
|
| 2689 |
+
"learning_rate": 4.0016635854253966e-05,
|
| 2690 |
+
"loss": 0.5734,
|
| 2691 |
+
"step": 221000
|
| 2692 |
+
},
|
| 2693 |
+
{
|
| 2694 |
+
"epoch": 3.12,
|
| 2695 |
+
"learning_rate": 3.999289747775002e-05,
|
| 2696 |
+
"loss": 0.5682,
|
| 2697 |
+
"step": 221500
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 3.13,
|
| 2701 |
+
"learning_rate": 3.996920657799909e-05,
|
| 2702 |
+
"loss": 0.5628,
|
| 2703 |
+
"step": 222000
|
| 2704 |
+
},
|
| 2705 |
+
{
|
| 2706 |
+
"epoch": 3.14,
|
| 2707 |
+
"learning_rate": 3.9945468201495136e-05,
|
| 2708 |
+
"loss": 0.5687,
|
| 2709 |
+
"step": 222500
|
| 2710 |
+
},
|
| 2711 |
+
{
|
| 2712 |
+
"epoch": 3.14,
|
| 2713 |
+
"learning_rate": 3.992172982499119e-05,
|
| 2714 |
+
"loss": 0.5569,
|
| 2715 |
+
"step": 223000
|
| 2716 |
+
},
|
| 2717 |
+
{
|
| 2718 |
+
"epoch": 3.15,
|
| 2719 |
+
"learning_rate": 3.989799144848725e-05,
|
| 2720 |
+
"loss": 0.5571,
|
| 2721 |
+
"step": 223500
|
| 2722 |
+
},
|
| 2723 |
+
{
|
| 2724 |
+
"epoch": 3.16,
|
| 2725 |
+
"learning_rate": 3.98742530719833e-05,
|
| 2726 |
+
"loss": 0.5619,
|
| 2727 |
+
"step": 224000
|
| 2728 |
+
},
|
| 2729 |
+
{
|
| 2730 |
+
"epoch": 3.17,
|
| 2731 |
+
"learning_rate": 3.9850514695479365e-05,
|
| 2732 |
+
"loss": 0.5635,
|
| 2733 |
+
"step": 224500
|
| 2734 |
+
},
|
| 2735 |
+
{
|
| 2736 |
+
"epoch": 3.17,
|
| 2737 |
+
"learning_rate": 3.982677631897542e-05,
|
| 2738 |
+
"loss": 0.5538,
|
| 2739 |
+
"step": 225000
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 3.18,
|
| 2743 |
+
"learning_rate": 3.980303794247147e-05,
|
| 2744 |
+
"loss": 0.5494,
|
| 2745 |
+
"step": 225500
|
| 2746 |
+
},
|
| 2747 |
+
{
|
| 2748 |
+
"epoch": 3.19,
|
| 2749 |
+
"learning_rate": 3.9779299565967525e-05,
|
| 2750 |
+
"loss": 0.5532,
|
| 2751 |
+
"step": 226000
|
| 2752 |
+
},
|
| 2753 |
+
{
|
| 2754 |
+
"epoch": 3.19,
|
| 2755 |
+
"learning_rate": 3.975556118946358e-05,
|
| 2756 |
+
"loss": 0.5576,
|
| 2757 |
+
"step": 226500
|
| 2758 |
+
},
|
| 2759 |
+
{
|
| 2760 |
+
"epoch": 3.2,
|
| 2761 |
+
"learning_rate": 3.973187028971264e-05,
|
| 2762 |
+
"loss": 0.5499,
|
| 2763 |
+
"step": 227000
|
| 2764 |
+
},
|
| 2765 |
+
{
|
| 2766 |
+
"epoch": 3.21,
|
| 2767 |
+
"learning_rate": 3.970822686671472e-05,
|
| 2768 |
+
"loss": 0.5434,
|
| 2769 |
+
"step": 227500
|
| 2770 |
+
},
|
| 2771 |
+
{
|
| 2772 |
+
"epoch": 3.21,
|
| 2773 |
+
"learning_rate": 3.968448849021077e-05,
|
| 2774 |
+
"loss": 0.5421,
|
| 2775 |
+
"step": 228000
|
| 2776 |
+
},
|
| 2777 |
+
{
|
| 2778 |
+
"epoch": 3.22,
|
| 2779 |
+
"learning_rate": 3.966075011370682e-05,
|
| 2780 |
+
"loss": 0.5377,
|
| 2781 |
+
"step": 228500
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 3.23,
|
| 2785 |
+
"learning_rate": 3.963701173720288e-05,
|
| 2786 |
+
"loss": 0.5614,
|
| 2787 |
+
"step": 229000
|
| 2788 |
+
},
|
| 2789 |
+
{
|
| 2790 |
+
"epoch": 3.24,
|
| 2791 |
+
"learning_rate": 3.961327336069894e-05,
|
| 2792 |
+
"loss": 0.5458,
|
| 2793 |
+
"step": 229500
|
| 2794 |
+
},
|
| 2795 |
+
{
|
| 2796 |
+
"epoch": 3.24,
|
| 2797 |
+
"learning_rate": 3.9589534984194995e-05,
|
| 2798 |
+
"loss": 0.5483,
|
| 2799 |
+
"step": 230000
|
| 2800 |
+
},
|
| 2801 |
+
{
|
| 2802 |
+
"epoch": 3.25,
|
| 2803 |
+
"learning_rate": 3.9565796607691044e-05,
|
| 2804 |
+
"loss": 0.5423,
|
| 2805 |
+
"step": 230500
|
| 2806 |
+
},
|
| 2807 |
+
{
|
| 2808 |
+
"epoch": 3.26,
|
| 2809 |
+
"learning_rate": 3.95420582311871e-05,
|
| 2810 |
+
"loss": 0.5346,
|
| 2811 |
+
"step": 231000
|
| 2812 |
+
},
|
| 2813 |
+
{
|
| 2814 |
+
"epoch": 3.26,
|
| 2815 |
+
"learning_rate": 3.9518319854683155e-05,
|
| 2816 |
+
"loss": 0.5334,
|
| 2817 |
+
"step": 231500
|
| 2818 |
+
},
|
| 2819 |
+
{
|
| 2820 |
+
"epoch": 3.27,
|
| 2821 |
+
"learning_rate": 3.9494628954932215e-05,
|
| 2822 |
+
"loss": 0.5451,
|
| 2823 |
+
"step": 232000
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 3.28,
|
| 2827 |
+
"learning_rate": 3.947089057842828e-05,
|
| 2828 |
+
"loss": 0.5378,
|
| 2829 |
+
"step": 232500
|
| 2830 |
+
},
|
| 2831 |
+
{
|
| 2832 |
+
"epoch": 3.29,
|
| 2833 |
+
"learning_rate": 3.944715220192433e-05,
|
| 2834 |
+
"loss": 0.5323,
|
| 2835 |
+
"step": 233000
|
| 2836 |
+
},
|
| 2837 |
+
{
|
| 2838 |
+
"epoch": 3.29,
|
| 2839 |
+
"learning_rate": 3.942346130217339e-05,
|
| 2840 |
+
"loss": 0.5346,
|
| 2841 |
+
"step": 233500
|
| 2842 |
+
},
|
| 2843 |
+
{
|
| 2844 |
+
"epoch": 3.3,
|
| 2845 |
+
"learning_rate": 3.939972292566945e-05,
|
| 2846 |
+
"loss": 0.5334,
|
| 2847 |
+
"step": 234000
|
| 2848 |
+
},
|
| 2849 |
+
{
|
| 2850 |
+
"epoch": 3.31,
|
| 2851 |
+
"learning_rate": 3.9375984549165504e-05,
|
| 2852 |
+
"loss": 0.5292,
|
| 2853 |
+
"step": 234500
|
| 2854 |
+
},
|
| 2855 |
+
{
|
| 2856 |
+
"epoch": 3.31,
|
| 2857 |
+
"learning_rate": 3.935224617266155e-05,
|
| 2858 |
+
"loss": 0.5442,
|
| 2859 |
+
"step": 235000
|
| 2860 |
+
},
|
| 2861 |
+
{
|
| 2862 |
+
"epoch": 3.32,
|
| 2863 |
+
"learning_rate": 3.9328507796157615e-05,
|
| 2864 |
+
"loss": 0.5423,
|
| 2865 |
+
"step": 235500
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 3.33,
|
| 2869 |
+
"learning_rate": 3.930476941965367e-05,
|
| 2870 |
+
"loss": 0.53,
|
| 2871 |
+
"step": 236000
|
| 2872 |
+
},
|
| 2873 |
+
{
|
| 2874 |
+
"epoch": 3.33,
|
| 2875 |
+
"learning_rate": 3.9281031043149725e-05,
|
| 2876 |
+
"loss": 0.5359,
|
| 2877 |
+
"step": 236500
|
| 2878 |
+
},
|
| 2879 |
+
{
|
| 2880 |
+
"epoch": 3.34,
|
| 2881 |
+
"learning_rate": 3.9257340143398785e-05,
|
| 2882 |
+
"loss": 0.5432,
|
| 2883 |
+
"step": 237000
|
| 2884 |
+
},
|
| 2885 |
+
{
|
| 2886 |
+
"epoch": 3.35,
|
| 2887 |
+
"learning_rate": 3.923360176689484e-05,
|
| 2888 |
+
"loss": 0.5325,
|
| 2889 |
+
"step": 237500
|
| 2890 |
+
},
|
| 2891 |
+
{
|
| 2892 |
+
"epoch": 3.36,
|
| 2893 |
+
"learning_rate": 3.9209863390390896e-05,
|
| 2894 |
+
"loss": 0.53,
|
| 2895 |
+
"step": 238000
|
| 2896 |
+
},
|
| 2897 |
+
{
|
| 2898 |
+
"epoch": 3.36,
|
| 2899 |
+
"learning_rate": 3.918612501388695e-05,
|
| 2900 |
+
"loss": 0.5347,
|
| 2901 |
+
"step": 238500
|
| 2902 |
+
},
|
| 2903 |
+
{
|
| 2904 |
+
"epoch": 3.37,
|
| 2905 |
+
"learning_rate": 3.916238663738301e-05,
|
| 2906 |
+
"loss": 0.5268,
|
| 2907 |
+
"step": 239000
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 3.38,
|
| 2911 |
+
"learning_rate": 3.913864826087906e-05,
|
| 2912 |
+
"loss": 0.5233,
|
| 2913 |
+
"step": 239500
|
| 2914 |
+
},
|
| 2915 |
+
{
|
| 2916 |
+
"epoch": 3.38,
|
| 2917 |
+
"learning_rate": 3.911490988437512e-05,
|
| 2918 |
+
"loss": 0.5281,
|
| 2919 |
+
"step": 240000
|
| 2920 |
+
},
|
| 2921 |
+
{
|
| 2922 |
+
"epoch": 3.39,
|
| 2923 |
+
"learning_rate": 3.9091171507871174e-05,
|
| 2924 |
+
"loss": 0.5191,
|
| 2925 |
+
"step": 240500
|
| 2926 |
+
},
|
| 2927 |
+
{
|
| 2928 |
+
"epoch": 3.4,
|
| 2929 |
+
"learning_rate": 3.906743313136723e-05,
|
| 2930 |
+
"loss": 0.5226,
|
| 2931 |
+
"step": 241000
|
| 2932 |
+
},
|
| 2933 |
+
{
|
| 2934 |
+
"epoch": 3.41,
|
| 2935 |
+
"learning_rate": 3.904369475486328e-05,
|
| 2936 |
+
"loss": 0.525,
|
| 2937 |
+
"step": 241500
|
| 2938 |
+
},
|
| 2939 |
+
{
|
| 2940 |
+
"epoch": 3.41,
|
| 2941 |
+
"learning_rate": 3.901995637835934e-05,
|
| 2942 |
+
"loss": 0.5252,
|
| 2943 |
+
"step": 242000
|
| 2944 |
+
},
|
| 2945 |
+
{
|
| 2946 |
+
"epoch": 3.42,
|
| 2947 |
+
"learning_rate": 3.8996218001855396e-05,
|
| 2948 |
+
"loss": 0.5406,
|
| 2949 |
+
"step": 242500
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 3.43,
|
| 2953 |
+
"learning_rate": 3.897257457885746e-05,
|
| 2954 |
+
"loss": 0.5168,
|
| 2955 |
+
"step": 243000
|
| 2956 |
+
},
|
| 2957 |
+
{
|
| 2958 |
+
"epoch": 3.43,
|
| 2959 |
+
"learning_rate": 3.8948836202353515e-05,
|
| 2960 |
+
"loss": 0.5218,
|
| 2961 |
+
"step": 243500
|
| 2962 |
+
},
|
| 2963 |
+
{
|
| 2964 |
+
"epoch": 3.44,
|
| 2965 |
+
"learning_rate": 3.892509782584958e-05,
|
| 2966 |
+
"loss": 0.5304,
|
| 2967 |
+
"step": 244000
|
| 2968 |
+
},
|
| 2969 |
+
{
|
| 2970 |
+
"epoch": 3.45,
|
| 2971 |
+
"learning_rate": 3.890135944934563e-05,
|
| 2972 |
+
"loss": 0.5217,
|
| 2973 |
+
"step": 244500
|
| 2974 |
+
},
|
| 2975 |
+
{
|
| 2976 |
+
"epoch": 3.45,
|
| 2977 |
+
"learning_rate": 3.887762107284169e-05,
|
| 2978 |
+
"loss": 0.5143,
|
| 2979 |
+
"step": 245000
|
| 2980 |
+
},
|
| 2981 |
+
{
|
| 2982 |
+
"epoch": 3.46,
|
| 2983 |
+
"learning_rate": 3.885388269633774e-05,
|
| 2984 |
+
"loss": 0.5326,
|
| 2985 |
+
"step": 245500
|
| 2986 |
+
},
|
| 2987 |
+
{
|
| 2988 |
+
"epoch": 3.47,
|
| 2989 |
+
"learning_rate": 3.883014431983379e-05,
|
| 2990 |
+
"loss": 0.5152,
|
| 2991 |
+
"step": 246000
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 3.48,
|
| 2995 |
+
"learning_rate": 3.880640594332985e-05,
|
| 2996 |
+
"loss": 0.5288,
|
| 2997 |
+
"step": 246500
|
| 2998 |
+
},
|
| 2999 |
+
{
|
| 3000 |
+
"epoch": 3.48,
|
| 3001 |
+
"learning_rate": 3.8782667566825904e-05,
|
| 3002 |
+
"loss": 0.517,
|
| 3003 |
+
"step": 247000
|
| 3004 |
+
},
|
| 3005 |
+
{
|
| 3006 |
+
"epoch": 3.49,
|
| 3007 |
+
"learning_rate": 3.8758929190321966e-05,
|
| 3008 |
+
"loss": 0.5092,
|
| 3009 |
+
"step": 247500
|
| 3010 |
+
},
|
| 3011 |
+
{
|
| 3012 |
+
"epoch": 3.5,
|
| 3013 |
+
"learning_rate": 3.8735238290571026e-05,
|
| 3014 |
+
"loss": 0.5147,
|
| 3015 |
+
"step": 248000
|
| 3016 |
+
},
|
| 3017 |
+
{
|
| 3018 |
+
"epoch": 3.5,
|
| 3019 |
+
"learning_rate": 3.871149991406708e-05,
|
| 3020 |
+
"loss": 0.5175,
|
| 3021 |
+
"step": 248500
|
| 3022 |
+
},
|
| 3023 |
+
{
|
| 3024 |
+
"epoch": 3.51,
|
| 3025 |
+
"learning_rate": 3.868780901431614e-05,
|
| 3026 |
+
"loss": 0.4983,
|
| 3027 |
+
"step": 249000
|
| 3028 |
+
},
|
| 3029 |
+
{
|
| 3030 |
+
"epoch": 3.52,
|
| 3031 |
+
"learning_rate": 3.866411811456521e-05,
|
| 3032 |
+
"loss": 0.5145,
|
| 3033 |
+
"step": 249500
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 3.53,
|
| 3037 |
+
"learning_rate": 3.8640379738061264e-05,
|
| 3038 |
+
"loss": 0.5226,
|
| 3039 |
+
"step": 250000
|
| 3040 |
+
},
|
| 3041 |
+
{
|
| 3042 |
+
"epoch": 3.53,
|
| 3043 |
+
"learning_rate": 3.861664136155731e-05,
|
| 3044 |
+
"loss": 0.5148,
|
| 3045 |
+
"step": 250500
|
| 3046 |
+
},
|
| 3047 |
+
{
|
| 3048 |
+
"epoch": 3.54,
|
| 3049 |
+
"learning_rate": 3.859290298505337e-05,
|
| 3050 |
+
"loss": 0.5172,
|
| 3051 |
+
"step": 251000
|
| 3052 |
+
},
|
| 3053 |
+
{
|
| 3054 |
+
"epoch": 3.55,
|
| 3055 |
+
"learning_rate": 3.856916460854942e-05,
|
| 3056 |
+
"loss": 0.5245,
|
| 3057 |
+
"step": 251500
|
| 3058 |
+
},
|
| 3059 |
+
{
|
| 3060 |
+
"epoch": 3.55,
|
| 3061 |
+
"learning_rate": 3.854547370879849e-05,
|
| 3062 |
+
"loss": 0.5164,
|
| 3063 |
+
"step": 252000
|
| 3064 |
+
},
|
| 3065 |
+
{
|
| 3066 |
+
"epoch": 3.56,
|
| 3067 |
+
"learning_rate": 3.8521735332294545e-05,
|
| 3068 |
+
"loss": 0.5087,
|
| 3069 |
+
"step": 252500
|
| 3070 |
+
},
|
| 3071 |
+
{
|
| 3072 |
+
"epoch": 3.57,
|
| 3073 |
+
"learning_rate": 3.84979969557906e-05,
|
| 3074 |
+
"loss": 0.517,
|
| 3075 |
+
"step": 253000
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 3.57,
|
| 3079 |
+
"learning_rate": 3.8474258579286656e-05,
|
| 3080 |
+
"loss": 0.5089,
|
| 3081 |
+
"step": 253500
|
| 3082 |
+
},
|
| 3083 |
+
{
|
| 3084 |
+
"epoch": 3.58,
|
| 3085 |
+
"learning_rate": 3.8450520202782705e-05,
|
| 3086 |
+
"loss": 0.5045,
|
| 3087 |
+
"step": 254000
|
| 3088 |
+
},
|
| 3089 |
+
{
|
| 3090 |
+
"epoch": 3.59,
|
| 3091 |
+
"learning_rate": 3.842682930303177e-05,
|
| 3092 |
+
"loss": 0.5169,
|
| 3093 |
+
"step": 254500
|
| 3094 |
+
},
|
| 3095 |
+
{
|
| 3096 |
+
"epoch": 3.6,
|
| 3097 |
+
"learning_rate": 3.840309092652783e-05,
|
| 3098 |
+
"loss": 0.5204,
|
| 3099 |
+
"step": 255000
|
| 3100 |
+
},
|
| 3101 |
+
{
|
| 3102 |
+
"epoch": 3.6,
|
| 3103 |
+
"learning_rate": 3.837935255002388e-05,
|
| 3104 |
+
"loss": 0.5064,
|
| 3105 |
+
"step": 255500
|
| 3106 |
+
},
|
| 3107 |
+
{
|
| 3108 |
+
"epoch": 3.61,
|
| 3109 |
+
"learning_rate": 3.835561417351994e-05,
|
| 3110 |
+
"loss": 0.5074,
|
| 3111 |
+
"step": 256000
|
| 3112 |
+
},
|
| 3113 |
+
{
|
| 3114 |
+
"epoch": 3.62,
|
| 3115 |
+
"learning_rate": 3.8331875797015994e-05,
|
| 3116 |
+
"loss": 0.5147,
|
| 3117 |
+
"step": 256500
|
| 3118 |
+
},
|
| 3119 |
+
{
|
| 3120 |
+
"epoch": 3.62,
|
| 3121 |
+
"learning_rate": 3.830813742051205e-05,
|
| 3122 |
+
"loss": 0.5031,
|
| 3123 |
+
"step": 257000
|
| 3124 |
+
},
|
| 3125 |
+
{
|
| 3126 |
+
"epoch": 3.63,
|
| 3127 |
+
"learning_rate": 3.8284399044008104e-05,
|
| 3128 |
+
"loss": 0.5046,
|
| 3129 |
+
"step": 257500
|
| 3130 |
+
},
|
| 3131 |
+
{
|
| 3132 |
+
"epoch": 3.64,
|
| 3133 |
+
"learning_rate": 3.826066066750415e-05,
|
| 3134 |
+
"loss": 0.5023,
|
| 3135 |
+
"step": 258000
|
| 3136 |
+
},
|
| 3137 |
+
{
|
| 3138 |
+
"epoch": 3.64,
|
| 3139 |
+
"learning_rate": 3.8236922291000215e-05,
|
| 3140 |
+
"loss": 0.5053,
|
| 3141 |
+
"step": 258500
|
| 3142 |
+
},
|
| 3143 |
+
{
|
| 3144 |
+
"epoch": 3.65,
|
| 3145 |
+
"learning_rate": 3.821318391449627e-05,
|
| 3146 |
+
"loss": 0.5145,
|
| 3147 |
+
"step": 259000
|
| 3148 |
+
},
|
| 3149 |
+
{
|
| 3150 |
+
"epoch": 3.66,
|
| 3151 |
+
"learning_rate": 3.8189445537992326e-05,
|
| 3152 |
+
"loss": 0.5037,
|
| 3153 |
+
"step": 259500
|
| 3154 |
+
},
|
| 3155 |
+
{
|
| 3156 |
+
"epoch": 3.67,
|
| 3157 |
+
"learning_rate": 3.816570716148838e-05,
|
| 3158 |
+
"loss": 0.5164,
|
| 3159 |
+
"step": 260000
|
| 3160 |
+
},
|
| 3161 |
+
{
|
| 3162 |
+
"epoch": 3.67,
|
| 3163 |
+
"learning_rate": 3.814196878498443e-05,
|
| 3164 |
+
"loss": 0.5089,
|
| 3165 |
+
"step": 260500
|
| 3166 |
+
},
|
| 3167 |
+
{
|
| 3168 |
+
"epoch": 3.68,
|
| 3169 |
+
"learning_rate": 3.8118230408480486e-05,
|
| 3170 |
+
"loss": 0.499,
|
| 3171 |
+
"step": 261000
|
| 3172 |
+
},
|
| 3173 |
+
{
|
| 3174 |
+
"epoch": 3.69,
|
| 3175 |
+
"learning_rate": 3.809453950872955e-05,
|
| 3176 |
+
"loss": 0.5004,
|
| 3177 |
+
"step": 261500
|
| 3178 |
+
},
|
| 3179 |
+
{
|
| 3180 |
+
"epoch": 3.69,
|
| 3181 |
+
"learning_rate": 3.807084860897861e-05,
|
| 3182 |
+
"loss": 0.4955,
|
| 3183 |
+
"step": 262000
|
| 3184 |
+
},
|
| 3185 |
+
{
|
| 3186 |
+
"epoch": 3.7,
|
| 3187 |
+
"learning_rate": 3.804715770922768e-05,
|
| 3188 |
+
"loss": 0.5021,
|
| 3189 |
+
"step": 262500
|
| 3190 |
+
},
|
| 3191 |
+
{
|
| 3192 |
+
"epoch": 3.71,
|
| 3193 |
+
"learning_rate": 3.802341933272373e-05,
|
| 3194 |
+
"loss": 0.5064,
|
| 3195 |
+
"step": 263000
|
| 3196 |
+
},
|
| 3197 |
+
{
|
| 3198 |
+
"epoch": 3.72,
|
| 3199 |
+
"learning_rate": 3.799968095621979e-05,
|
| 3200 |
+
"loss": 0.4947,
|
| 3201 |
+
"step": 263500
|
| 3202 |
+
},
|
| 3203 |
+
{
|
| 3204 |
+
"epoch": 3.72,
|
| 3205 |
+
"learning_rate": 3.7975942579715846e-05,
|
| 3206 |
+
"loss": 0.5033,
|
| 3207 |
+
"step": 264000
|
| 3208 |
+
},
|
| 3209 |
+
{
|
| 3210 |
+
"epoch": 3.73,
|
| 3211 |
+
"learning_rate": 3.79522042032119e-05,
|
| 3212 |
+
"loss": 0.5022,
|
| 3213 |
+
"step": 264500
|
| 3214 |
+
},
|
| 3215 |
+
{
|
| 3216 |
+
"epoch": 3.74,
|
| 3217 |
+
"learning_rate": 3.792846582670796e-05,
|
| 3218 |
+
"loss": 0.4956,
|
| 3219 |
+
"step": 265000
|
| 3220 |
+
},
|
| 3221 |
+
{
|
| 3222 |
+
"epoch": 3.74,
|
| 3223 |
+
"learning_rate": 3.790477492695702e-05,
|
| 3224 |
+
"loss": 0.5102,
|
| 3225 |
+
"step": 265500
|
| 3226 |
+
},
|
| 3227 |
+
{
|
| 3228 |
+
"epoch": 3.75,
|
| 3229 |
+
"learning_rate": 3.788103655045307e-05,
|
| 3230 |
+
"loss": 0.509,
|
| 3231 |
+
"step": 266000
|
| 3232 |
+
},
|
| 3233 |
+
{
|
| 3234 |
+
"epoch": 3.76,
|
| 3235 |
+
"learning_rate": 3.785729817394913e-05,
|
| 3236 |
+
"loss": 0.5048,
|
| 3237 |
+
"step": 266500
|
| 3238 |
+
},
|
| 3239 |
+
{
|
| 3240 |
+
"epoch": 3.76,
|
| 3241 |
+
"learning_rate": 3.783355979744518e-05,
|
| 3242 |
+
"loss": 0.4954,
|
| 3243 |
+
"step": 267000
|
| 3244 |
+
},
|
| 3245 |
+
{
|
| 3246 |
+
"epoch": 3.77,
|
| 3247 |
+
"learning_rate": 3.780982142094124e-05,
|
| 3248 |
+
"loss": 0.5009,
|
| 3249 |
+
"step": 267500
|
| 3250 |
+
},
|
| 3251 |
+
{
|
| 3252 |
+
"epoch": 3.78,
|
| 3253 |
+
"learning_rate": 3.7786083044437294e-05,
|
| 3254 |
+
"loss": 0.4909,
|
| 3255 |
+
"step": 268000
|
| 3256 |
+
},
|
| 3257 |
+
{
|
| 3258 |
+
"epoch": 3.79,
|
| 3259 |
+
"learning_rate": 3.776234466793335e-05,
|
| 3260 |
+
"loss": 0.5018,
|
| 3261 |
+
"step": 268500
|
| 3262 |
+
},
|
| 3263 |
+
{
|
| 3264 |
+
"epoch": 3.79,
|
| 3265 |
+
"learning_rate": 3.77386062914294e-05,
|
| 3266 |
+
"loss": 0.4903,
|
| 3267 |
+
"step": 269000
|
| 3268 |
+
},
|
| 3269 |
+
{
|
| 3270 |
+
"epoch": 3.8,
|
| 3271 |
+
"learning_rate": 3.7714867914925454e-05,
|
| 3272 |
+
"loss": 0.504,
|
| 3273 |
+
"step": 269500
|
| 3274 |
+
},
|
| 3275 |
+
{
|
| 3276 |
+
"epoch": 3.81,
|
| 3277 |
+
"learning_rate": 3.7691129538421516e-05,
|
| 3278 |
+
"loss": 0.5088,
|
| 3279 |
+
"step": 270000
|
| 3280 |
+
},
|
| 3281 |
+
{
|
| 3282 |
+
"epoch": 3.81,
|
| 3283 |
+
"learning_rate": 3.7667438638670576e-05,
|
| 3284 |
+
"loss": 0.5009,
|
| 3285 |
+
"step": 270500
|
| 3286 |
+
},
|
| 3287 |
+
{
|
| 3288 |
+
"epoch": 3.82,
|
| 3289 |
+
"learning_rate": 3.7643747738919636e-05,
|
| 3290 |
+
"loss": 0.4924,
|
| 3291 |
+
"step": 271000
|
| 3292 |
+
},
|
| 3293 |
+
{
|
| 3294 |
+
"epoch": 3.83,
|
| 3295 |
+
"learning_rate": 3.76200568391687e-05,
|
| 3296 |
+
"loss": 0.4999,
|
| 3297 |
+
"step": 271500
|
| 3298 |
+
},
|
| 3299 |
+
{
|
| 3300 |
+
"epoch": 3.84,
|
| 3301 |
+
"learning_rate": 3.759631846266476e-05,
|
| 3302 |
+
"loss": 0.4997,
|
| 3303 |
+
"step": 272000
|
| 3304 |
+
},
|
| 3305 |
+
{
|
| 3306 |
+
"epoch": 3.84,
|
| 3307 |
+
"learning_rate": 3.7572580086160813e-05,
|
| 3308 |
+
"loss": 0.4968,
|
| 3309 |
+
"step": 272500
|
| 3310 |
+
},
|
| 3311 |
+
{
|
| 3312 |
+
"epoch": 3.85,
|
| 3313 |
+
"learning_rate": 3.754884170965687e-05,
|
| 3314 |
+
"loss": 0.4905,
|
| 3315 |
+
"step": 273000
|
| 3316 |
+
},
|
| 3317 |
+
{
|
| 3318 |
+
"epoch": 3.86,
|
| 3319 |
+
"learning_rate": 3.7525103333152924e-05,
|
| 3320 |
+
"loss": 0.5055,
|
| 3321 |
+
"step": 273500
|
| 3322 |
+
},
|
| 3323 |
+
{
|
| 3324 |
+
"epoch": 3.86,
|
| 3325 |
+
"learning_rate": 3.750136495664898e-05,
|
| 3326 |
+
"loss": 0.4918,
|
| 3327 |
+
"step": 274000
|
| 3328 |
+
},
|
| 3329 |
+
{
|
| 3330 |
+
"epoch": 3.87,
|
| 3331 |
+
"learning_rate": 3.747762658014503e-05,
|
| 3332 |
+
"loss": 0.4947,
|
| 3333 |
+
"step": 274500
|
| 3334 |
+
},
|
| 3335 |
+
{
|
| 3336 |
+
"epoch": 3.88,
|
| 3337 |
+
"learning_rate": 3.745388820364109e-05,
|
| 3338 |
+
"loss": 0.5016,
|
| 3339 |
+
"step": 275000
|
| 3340 |
+
},
|
| 3341 |
+
{
|
| 3342 |
+
"epoch": 3.88,
|
| 3343 |
+
"learning_rate": 3.7430149827137146e-05,
|
| 3344 |
+
"loss": 0.5084,
|
| 3345 |
+
"step": 275500
|
| 3346 |
+
},
|
| 3347 |
+
{
|
| 3348 |
+
"epoch": 3.89,
|
| 3349 |
+
"learning_rate": 3.74064114506332e-05,
|
| 3350 |
+
"loss": 0.4882,
|
| 3351 |
+
"step": 276000
|
| 3352 |
+
},
|
| 3353 |
+
{
|
| 3354 |
+
"epoch": 3.9,
|
| 3355 |
+
"learning_rate": 3.738272055088226e-05,
|
| 3356 |
+
"loss": 0.4948,
|
| 3357 |
+
"step": 276500
|
| 3358 |
+
},
|
| 3359 |
+
{
|
| 3360 |
+
"epoch": 3.91,
|
| 3361 |
+
"learning_rate": 3.735898217437832e-05,
|
| 3362 |
+
"loss": 0.4971,
|
| 3363 |
+
"step": 277000
|
| 3364 |
+
},
|
| 3365 |
+
{
|
| 3366 |
+
"epoch": 3.91,
|
| 3367 |
+
"learning_rate": 3.733524379787437e-05,
|
| 3368 |
+
"loss": 0.4867,
|
| 3369 |
+
"step": 277500
|
| 3370 |
+
},
|
| 3371 |
+
{
|
| 3372 |
+
"epoch": 3.92,
|
| 3373 |
+
"learning_rate": 3.731150542137043e-05,
|
| 3374 |
+
"loss": 0.4917,
|
| 3375 |
+
"step": 278000
|
| 3376 |
+
},
|
| 3377 |
+
{
|
| 3378 |
+
"epoch": 3.93,
|
| 3379 |
+
"learning_rate": 3.7287767044866484e-05,
|
| 3380 |
+
"loss": 0.4853,
|
| 3381 |
+
"step": 278500
|
| 3382 |
+
},
|
| 3383 |
+
{
|
| 3384 |
+
"epoch": 3.93,
|
| 3385 |
+
"learning_rate": 3.726402866836254e-05,
|
| 3386 |
+
"loss": 0.4948,
|
| 3387 |
+
"step": 279000
|
| 3388 |
+
},
|
| 3389 |
+
{
|
| 3390 |
+
"epoch": 3.94,
|
| 3391 |
+
"learning_rate": 3.7240290291858594e-05,
|
| 3392 |
+
"loss": 0.4935,
|
| 3393 |
+
"step": 279500
|
| 3394 |
+
},
|
| 3395 |
+
{
|
| 3396 |
+
"epoch": 3.95,
|
| 3397 |
+
"learning_rate": 3.721655191535465e-05,
|
| 3398 |
+
"loss": 0.4906,
|
| 3399 |
+
"step": 280000
|
| 3400 |
+
},
|
| 3401 |
+
{
|
| 3402 |
+
"epoch": 3.96,
|
| 3403 |
+
"learning_rate": 3.71928135388507e-05,
|
| 3404 |
+
"loss": 0.4992,
|
| 3405 |
+
"step": 280500
|
| 3406 |
+
},
|
| 3407 |
+
{
|
| 3408 |
+
"epoch": 3.96,
|
| 3409 |
+
"learning_rate": 3.716907516234676e-05,
|
| 3410 |
+
"loss": 0.4814,
|
| 3411 |
+
"step": 281000
|
| 3412 |
+
},
|
| 3413 |
+
{
|
| 3414 |
+
"epoch": 3.97,
|
| 3415 |
+
"learning_rate": 3.7145336785842816e-05,
|
| 3416 |
+
"loss": 0.4903,
|
| 3417 |
+
"step": 281500
|
| 3418 |
+
},
|
| 3419 |
+
{
|
| 3420 |
+
"epoch": 3.98,
|
| 3421 |
+
"learning_rate": 3.712159840933887e-05,
|
| 3422 |
+
"loss": 0.4869,
|
| 3423 |
+
"step": 282000
|
| 3424 |
+
},
|
| 3425 |
+
{
|
| 3426 |
+
"epoch": 3.98,
|
| 3427 |
+
"learning_rate": 3.709786003283493e-05,
|
| 3428 |
+
"loss": 0.493,
|
| 3429 |
+
"step": 282500
|
| 3430 |
+
},
|
| 3431 |
+
{
|
| 3432 |
+
"epoch": 3.99,
|
| 3433 |
+
"learning_rate": 3.7074121656330976e-05,
|
| 3434 |
+
"loss": 0.4909,
|
| 3435 |
+
"step": 283000
|
| 3436 |
+
},
|
| 3437 |
+
{
|
| 3438 |
+
"epoch": 4.0,
|
| 3439 |
+
"learning_rate": 3.705038327982703e-05,
|
| 3440 |
+
"loss": 0.4873,
|
| 3441 |
+
"step": 283500
|
| 3442 |
+
},
|
| 3443 |
+
{
|
| 3444 |
+
"epoch": 4.0,
|
| 3445 |
+
"eval_bleu": 43.6153,
|
| 3446 |
+
"eval_gen_len": 13.6544,
|
| 3447 |
+
"eval_loss": 0.9757916927337646,
|
| 3448 |
+
"eval_runtime": 9810.7798,
|
| 3449 |
+
"eval_samples_per_second": 14.457,
|
| 3450 |
+
"eval_steps_per_second": 1.807,
|
| 3451 |
+
"step": 283678
|
| 3452 |
+
}
|
| 3453 |
+
],
|
| 3454 |
+
"logging_steps": 500,
|
| 3455 |
+
"max_steps": 1063785,
|
| 3456 |
+
"num_train_epochs": 15,
|
| 3457 |
+
"save_steps": 500,
|
| 3458 |
+
"total_flos": 9.836272613473649e+18,
|
| 3459 |
+
"trial_name": null,
|
| 3460 |
+
"trial_params": null
|
| 3461 |
+
}
|
checkpoint-283678/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0adf1a980c6128833811b7e6eb546e117ffd3efb8c21dc7de95b5e76a5b21b8d
|
| 3 |
+
size 4728
|
checkpoint-354597/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": "4.35.2",
|
| 57 |
+
"use_cache": true,
|
| 58 |
+
"vocab_size": 250054
|
| 59 |
+
}
|
checkpoint-354597/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": "4.35.2"
|
| 12 |
+
}
|
checkpoint-354597/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9ce9e9eb263fd3cc9b36e2fdc0918ce420101f814a240d91ff77a49f82e41db0
|
| 3 |
+
size 2444578688
|
checkpoint-354597/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:25bb0d2ae19fe22e534b6fa5026eaf9d323d2eca340fcc7e01dbf5601c39754e
|
| 3 |
+
size 4887473903
|
checkpoint-354597/rng_state_0.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dd09c3aa3224ca285750e8bdd73139b56096392004cf3cd2dcac1a92be8986bb
|
| 3 |
+
size 15024
|
checkpoint-354597/rng_state_1.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:476f94429eeae0e99cf9e686bcf4841b726c55b9a4f5b4a19553928f1f8fa1da
|
| 3 |
+
size 15024
|
checkpoint-354597/rng_state_2.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:95d07add4fc4c4215c42f55301a6b8884a9fab1e2fcf651ecb89d629cecfc0c2
|
| 3 |
+
size 15024
|
checkpoint-354597/rng_state_3.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bc76d4ca55054826774d54f7bfa258338bf25a99ce8d5b5dd025a77f653f14ae
|
| 3 |
+
size 15024
|
checkpoint-354597/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3bed9e4256d2ef527f867c56827da062cd2b96d2d096480d53aa75d2f1d2e19e
|
| 3 |
+
size 1064
|
checkpoint-354597/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-354597/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-354597/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1d91c41f87c8dbce15b820b41232d0dcd26ba285c22362400d3dd771a711417d
|
| 3 |
+
size 17110107
|
checkpoint-354597/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": "zh_CN",
|
| 525 |
+
"tgt_lang": "ko_KR",
|
| 526 |
+
"tokenizer_class": "MBart50Tokenizer",
|
| 527 |
+
"unk_token": "<unk>"
|
| 528 |
+
}
|
checkpoint-354597/trainer_state.json
ADDED
|
@@ -0,0 +1,4323 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.8987888693809509,
|
| 3 |
+
"best_model_checkpoint": "./zhko_mbartLarge_100p_run1/checkpoint-141839",
|
| 4 |
+
"epoch": 4.9999929497528885,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 354597,
|
| 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": 2.335965407031397e-06,
|
| 14 |
+
"loss": 2.879,
|
| 15 |
+
"step": 500
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.01,
|
| 19 |
+
"learning_rate": 4.6860312088738486e-06,
|
| 20 |
+
"loss": 2.0241,
|
| 21 |
+
"step": 1000
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.02,
|
| 25 |
+
"learning_rate": 7.0360970107162996e-06,
|
| 26 |
+
"loss": 1.8819,
|
| 27 |
+
"step": 1500
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.03,
|
| 31 |
+
"learning_rate": 9.386162812558751e-06,
|
| 32 |
+
"loss": 1.8009,
|
| 33 |
+
"step": 2000
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.04,
|
| 37 |
+
"learning_rate": 1.1736228614401204e-05,
|
| 38 |
+
"loss": 1.7072,
|
| 39 |
+
"step": 2500
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.04,
|
| 43 |
+
"learning_rate": 1.4086294416243657e-05,
|
| 44 |
+
"loss": 1.682,
|
| 45 |
+
"step": 3000
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.05,
|
| 49 |
+
"learning_rate": 1.643636021808611e-05,
|
| 50 |
+
"loss": 1.6188,
|
| 51 |
+
"step": 3500
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.06,
|
| 55 |
+
"learning_rate": 1.878642601992856e-05,
|
| 56 |
+
"loss": 1.5802,
|
| 57 |
+
"step": 4000
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.06,
|
| 61 |
+
"learning_rate": 2.1136491821771008e-05,
|
| 62 |
+
"loss": 1.5547,
|
| 63 |
+
"step": 4500
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.07,
|
| 67 |
+
"learning_rate": 2.348655762361346e-05,
|
| 68 |
+
"loss": 1.5422,
|
| 69 |
+
"step": 5000
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.08,
|
| 73 |
+
"learning_rate": 2.5831923293852228e-05,
|
| 74 |
+
"loss": 1.5194,
|
| 75 |
+
"step": 5500
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.08,
|
| 79 |
+
"learning_rate": 2.8181989095694684e-05,
|
| 80 |
+
"loss": 1.4815,
|
| 81 |
+
"step": 6000
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.09,
|
| 85 |
+
"learning_rate": 3.053205489753713e-05,
|
| 86 |
+
"loss": 1.4509,
|
| 87 |
+
"step": 6500
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.1,
|
| 91 |
+
"learning_rate": 3.288212069937958e-05,
|
| 92 |
+
"loss": 1.458,
|
| 93 |
+
"step": 7000
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.11,
|
| 97 |
+
"learning_rate": 3.5227486369618354e-05,
|
| 98 |
+
"loss": 1.4415,
|
| 99 |
+
"step": 7500
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.11,
|
| 103 |
+
"learning_rate": 3.757285203985712e-05,
|
| 104 |
+
"loss": 1.4284,
|
| 105 |
+
"step": 8000
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.12,
|
| 109 |
+
"learning_rate": 3.991821771009588e-05,
|
| 110 |
+
"loss": 1.4297,
|
| 111 |
+
"step": 8500
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.13,
|
| 115 |
+
"learning_rate": 4.2268283511938334e-05,
|
| 116 |
+
"loss": 1.4307,
|
| 117 |
+
"step": 9000
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.13,
|
| 121 |
+
"learning_rate": 4.461834931378079e-05,
|
| 122 |
+
"loss": 1.3971,
|
| 123 |
+
"step": 9500
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.14,
|
| 127 |
+
"learning_rate": 4.696371498401955e-05,
|
| 128 |
+
"loss": 1.4106,
|
| 129 |
+
"step": 10000
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.15,
|
| 133 |
+
"learning_rate": 4.9313780785862004e-05,
|
| 134 |
+
"loss": 1.3999,
|
| 135 |
+
"step": 10500
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.16,
|
| 139 |
+
"learning_rate": 4.998319322943521e-05,
|
| 140 |
+
"loss": 1.3937,
|
| 141 |
+
"step": 11000
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.16,
|
| 145 |
+
"learning_rate": 4.995945485293127e-05,
|
| 146 |
+
"loss": 1.3773,
|
| 147 |
+
"step": 11500
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.17,
|
| 151 |
+
"learning_rate": 4.9935716476427316e-05,
|
| 152 |
+
"loss": 1.3621,
|
| 153 |
+
"step": 12000
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.18,
|
| 157 |
+
"learning_rate": 4.991197809992337e-05,
|
| 158 |
+
"loss": 1.3552,
|
| 159 |
+
"step": 12500
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.18,
|
| 163 |
+
"learning_rate": 4.988823972341943e-05,
|
| 164 |
+
"loss": 1.3312,
|
| 165 |
+
"step": 13000
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.19,
|
| 169 |
+
"learning_rate": 4.986450134691549e-05,
|
| 170 |
+
"loss": 1.3331,
|
| 171 |
+
"step": 13500
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.2,
|
| 175 |
+
"learning_rate": 4.984081044716455e-05,
|
| 176 |
+
"loss": 1.3259,
|
| 177 |
+
"step": 14000
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.2,
|
| 181 |
+
"learning_rate": 4.9817072070660604e-05,
|
| 182 |
+
"loss": 1.2983,
|
| 183 |
+
"step": 14500
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.21,
|
| 187 |
+
"learning_rate": 4.979333369415666e-05,
|
| 188 |
+
"loss": 1.2828,
|
| 189 |
+
"step": 15000
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.22,
|
| 193 |
+
"learning_rate": 4.976959531765271e-05,
|
| 194 |
+
"loss": 1.2793,
|
| 195 |
+
"step": 15500
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.23,
|
| 199 |
+
"learning_rate": 4.9745856941148764e-05,
|
| 200 |
+
"loss": 1.2968,
|
| 201 |
+
"step": 16000
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.23,
|
| 205 |
+
"learning_rate": 4.9722118564644826e-05,
|
| 206 |
+
"loss": 1.2774,
|
| 207 |
+
"step": 16500
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.24,
|
| 211 |
+
"learning_rate": 4.969838018814088e-05,
|
| 212 |
+
"loss": 1.2507,
|
| 213 |
+
"step": 17000
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.25,
|
| 217 |
+
"learning_rate": 4.967464181163694e-05,
|
| 218 |
+
"loss": 1.2521,
|
| 219 |
+
"step": 17500
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.25,
|
| 223 |
+
"learning_rate": 4.965090343513299e-05,
|
| 224 |
+
"loss": 1.2337,
|
| 225 |
+
"step": 18000
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.26,
|
| 229 |
+
"learning_rate": 4.962716505862904e-05,
|
| 230 |
+
"loss": 1.2238,
|
| 231 |
+
"step": 18500
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 0.27,
|
| 235 |
+
"learning_rate": 4.96034266821251e-05,
|
| 236 |
+
"loss": 1.2111,
|
| 237 |
+
"step": 19000
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.27,
|
| 241 |
+
"learning_rate": 4.957968830562115e-05,
|
| 242 |
+
"loss": 1.2336,
|
| 243 |
+
"step": 19500
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.28,
|
| 247 |
+
"learning_rate": 4.955599740587022e-05,
|
| 248 |
+
"loss": 1.1975,
|
| 249 |
+
"step": 20000
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.29,
|
| 253 |
+
"learning_rate": 4.9532259029366274e-05,
|
| 254 |
+
"loss": 1.2092,
|
| 255 |
+
"step": 20500
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.3,
|
| 259 |
+
"learning_rate": 4.950852065286233e-05,
|
| 260 |
+
"loss": 1.2079,
|
| 261 |
+
"step": 21000
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.3,
|
| 265 |
+
"learning_rate": 4.948482975311139e-05,
|
| 266 |
+
"loss": 1.1831,
|
| 267 |
+
"step": 21500
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.31,
|
| 271 |
+
"learning_rate": 4.946109137660745e-05,
|
| 272 |
+
"loss": 1.1965,
|
| 273 |
+
"step": 22000
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 0.32,
|
| 277 |
+
"learning_rate": 4.94373530001035e-05,
|
| 278 |
+
"loss": 1.1928,
|
| 279 |
+
"step": 22500
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.32,
|
| 283 |
+
"learning_rate": 4.9413614623599556e-05,
|
| 284 |
+
"loss": 1.1779,
|
| 285 |
+
"step": 23000
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.33,
|
| 289 |
+
"learning_rate": 4.938987624709561e-05,
|
| 290 |
+
"loss": 1.1759,
|
| 291 |
+
"step": 23500
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.34,
|
| 295 |
+
"learning_rate": 4.936613787059167e-05,
|
| 296 |
+
"loss": 1.1871,
|
| 297 |
+
"step": 24000
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.35,
|
| 301 |
+
"learning_rate": 4.934239949408772e-05,
|
| 302 |
+
"loss": 1.1782,
|
| 303 |
+
"step": 24500
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.35,
|
| 307 |
+
"learning_rate": 4.931866111758378e-05,
|
| 308 |
+
"loss": 1.1619,
|
| 309 |
+
"step": 25000
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.36,
|
| 313 |
+
"learning_rate": 4.9294922741079834e-05,
|
| 314 |
+
"loss": 1.1615,
|
| 315 |
+
"step": 25500
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 0.37,
|
| 319 |
+
"learning_rate": 4.927118436457589e-05,
|
| 320 |
+
"loss": 1.1476,
|
| 321 |
+
"step": 26000
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 0.37,
|
| 325 |
+
"learning_rate": 4.924749346482495e-05,
|
| 326 |
+
"loss": 1.1507,
|
| 327 |
+
"step": 26500
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 0.38,
|
| 331 |
+
"learning_rate": 4.922380256507401e-05,
|
| 332 |
+
"loss": 1.1489,
|
| 333 |
+
"step": 27000
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 0.39,
|
| 337 |
+
"learning_rate": 4.9200064188570064e-05,
|
| 338 |
+
"loss": 1.144,
|
| 339 |
+
"step": 27500
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.39,
|
| 343 |
+
"learning_rate": 4.917632581206613e-05,
|
| 344 |
+
"loss": 1.1223,
|
| 345 |
+
"step": 28000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.4,
|
| 349 |
+
"learning_rate": 4.915258743556218e-05,
|
| 350 |
+
"loss": 1.1319,
|
| 351 |
+
"step": 28500
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 0.41,
|
| 355 |
+
"learning_rate": 4.912884905905824e-05,
|
| 356 |
+
"loss": 1.1235,
|
| 357 |
+
"step": 29000
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 0.42,
|
| 361 |
+
"learning_rate": 4.9105110682554286e-05,
|
| 362 |
+
"loss": 1.1408,
|
| 363 |
+
"step": 29500
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 0.42,
|
| 367 |
+
"learning_rate": 4.908141978280335e-05,
|
| 368 |
+
"loss": 1.1299,
|
| 369 |
+
"step": 30000
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.43,
|
| 373 |
+
"learning_rate": 4.905772888305242e-05,
|
| 374 |
+
"loss": 1.1172,
|
| 375 |
+
"step": 30500
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 0.44,
|
| 379 |
+
"learning_rate": 4.903399050654847e-05,
|
| 380 |
+
"loss": 1.1316,
|
| 381 |
+
"step": 31000
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.44,
|
| 385 |
+
"learning_rate": 4.9010299606797535e-05,
|
| 386 |
+
"loss": 1.1154,
|
| 387 |
+
"step": 31500
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.45,
|
| 391 |
+
"learning_rate": 4.8986561230293584e-05,
|
| 392 |
+
"loss": 1.1162,
|
| 393 |
+
"step": 32000
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 0.46,
|
| 397 |
+
"learning_rate": 4.896282285378964e-05,
|
| 398 |
+
"loss": 1.101,
|
| 399 |
+
"step": 32500
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 0.47,
|
| 403 |
+
"learning_rate": 4.89390844772857e-05,
|
| 404 |
+
"loss": 1.1153,
|
| 405 |
+
"step": 33000
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 0.47,
|
| 409 |
+
"learning_rate": 4.891534610078176e-05,
|
| 410 |
+
"loss": 1.1023,
|
| 411 |
+
"step": 33500
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 0.48,
|
| 415 |
+
"learning_rate": 4.889160772427781e-05,
|
| 416 |
+
"loss": 1.0935,
|
| 417 |
+
"step": 34000
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 0.49,
|
| 421 |
+
"learning_rate": 4.886786934777387e-05,
|
| 422 |
+
"loss": 1.0898,
|
| 423 |
+
"step": 34500
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.49,
|
| 427 |
+
"learning_rate": 4.884413097126992e-05,
|
| 428 |
+
"loss": 1.0846,
|
| 429 |
+
"step": 35000
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.5,
|
| 433 |
+
"learning_rate": 4.882039259476597e-05,
|
| 434 |
+
"loss": 1.0963,
|
| 435 |
+
"step": 35500
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 0.51,
|
| 439 |
+
"learning_rate": 4.879665421826203e-05,
|
| 440 |
+
"loss": 1.0764,
|
| 441 |
+
"step": 36000
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 0.51,
|
| 445 |
+
"learning_rate": 4.877291584175809e-05,
|
| 446 |
+
"loss": 1.0649,
|
| 447 |
+
"step": 36500
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 0.52,
|
| 451 |
+
"learning_rate": 4.8749177465254145e-05,
|
| 452 |
+
"loss": 1.0935,
|
| 453 |
+
"step": 37000
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 0.53,
|
| 457 |
+
"learning_rate": 4.8725439088750194e-05,
|
| 458 |
+
"loss": 1.0739,
|
| 459 |
+
"step": 37500
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.54,
|
| 463 |
+
"learning_rate": 4.870170071224625e-05,
|
| 464 |
+
"loss": 1.0779,
|
| 465 |
+
"step": 38000
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.54,
|
| 469 |
+
"learning_rate": 4.8677962335742305e-05,
|
| 470 |
+
"loss": 1.083,
|
| 471 |
+
"step": 38500
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.55,
|
| 475 |
+
"learning_rate": 4.865422395923836e-05,
|
| 476 |
+
"loss": 1.0729,
|
| 477 |
+
"step": 39000
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 0.56,
|
| 481 |
+
"learning_rate": 4.8630485582734416e-05,
|
| 482 |
+
"loss": 1.0587,
|
| 483 |
+
"step": 39500
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 0.56,
|
| 487 |
+
"learning_rate": 4.860679468298348e-05,
|
| 488 |
+
"loss": 1.078,
|
| 489 |
+
"step": 40000
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 0.57,
|
| 493 |
+
"learning_rate": 4.858305630647954e-05,
|
| 494 |
+
"loss": 1.0802,
|
| 495 |
+
"step": 40500
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 0.58,
|
| 499 |
+
"learning_rate": 4.855931792997559e-05,
|
| 500 |
+
"loss": 1.0538,
|
| 501 |
+
"step": 41000
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 0.59,
|
| 505 |
+
"learning_rate": 4.853557955347164e-05,
|
| 506 |
+
"loss": 1.0562,
|
| 507 |
+
"step": 41500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.59,
|
| 511 |
+
"learning_rate": 4.85118411769677e-05,
|
| 512 |
+
"loss": 1.0516,
|
| 513 |
+
"step": 42000
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.6,
|
| 517 |
+
"learning_rate": 4.848810280046375e-05,
|
| 518 |
+
"loss": 1.0624,
|
| 519 |
+
"step": 42500
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 0.61,
|
| 523 |
+
"learning_rate": 4.8464364423959815e-05,
|
| 524 |
+
"loss": 1.0587,
|
| 525 |
+
"step": 43000
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 0.61,
|
| 529 |
+
"learning_rate": 4.8440626047455864e-05,
|
| 530 |
+
"loss": 1.0544,
|
| 531 |
+
"step": 43500
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 0.62,
|
| 535 |
+
"learning_rate": 4.841688767095192e-05,
|
| 536 |
+
"loss": 1.0467,
|
| 537 |
+
"step": 44000
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 0.63,
|
| 541 |
+
"learning_rate": 4.8393291724707e-05,
|
| 542 |
+
"loss": 1.0613,
|
| 543 |
+
"step": 44500
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 0.63,
|
| 547 |
+
"learning_rate": 4.836955334820306e-05,
|
| 548 |
+
"loss": 1.0365,
|
| 549 |
+
"step": 45000
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.64,
|
| 553 |
+
"learning_rate": 4.834581497169911e-05,
|
| 554 |
+
"loss": 1.0424,
|
| 555 |
+
"step": 45500
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.65,
|
| 559 |
+
"learning_rate": 4.832207659519516e-05,
|
| 560 |
+
"loss": 1.0442,
|
| 561 |
+
"step": 46000
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 0.66,
|
| 565 |
+
"learning_rate": 4.829833821869122e-05,
|
| 566 |
+
"loss": 1.0354,
|
| 567 |
+
"step": 46500
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 0.66,
|
| 571 |
+
"learning_rate": 4.8274789749199304e-05,
|
| 572 |
+
"loss": 1.0543,
|
| 573 |
+
"step": 47000
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 0.67,
|
| 577 |
+
"learning_rate": 4.825105137269536e-05,
|
| 578 |
+
"loss": 1.0476,
|
| 579 |
+
"step": 47500
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 0.68,
|
| 583 |
+
"learning_rate": 4.8227312996191415e-05,
|
| 584 |
+
"loss": 1.0427,
|
| 585 |
+
"step": 48000
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.68,
|
| 589 |
+
"learning_rate": 4.820357461968748e-05,
|
| 590 |
+
"loss": 1.0174,
|
| 591 |
+
"step": 48500
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 0.69,
|
| 595 |
+
"learning_rate": 4.8179836243183526e-05,
|
| 596 |
+
"loss": 1.0325,
|
| 597 |
+
"step": 49000
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.7,
|
| 601 |
+
"learning_rate": 4.815609786667958e-05,
|
| 602 |
+
"loss": 1.0126,
|
| 603 |
+
"step": 49500
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 0.71,
|
| 607 |
+
"learning_rate": 4.813240696692865e-05,
|
| 608 |
+
"loss": 1.0335,
|
| 609 |
+
"step": 50000
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 0.71,
|
| 613 |
+
"learning_rate": 4.81086685904247e-05,
|
| 614 |
+
"loss": 1.0219,
|
| 615 |
+
"step": 50500
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 0.72,
|
| 619 |
+
"learning_rate": 4.808493021392075e-05,
|
| 620 |
+
"loss": 1.0291,
|
| 621 |
+
"step": 51000
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 0.73,
|
| 625 |
+
"learning_rate": 4.8061191837416815e-05,
|
| 626 |
+
"loss": 1.0251,
|
| 627 |
+
"step": 51500
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 0.73,
|
| 631 |
+
"learning_rate": 4.803745346091287e-05,
|
| 632 |
+
"loss": 1.01,
|
| 633 |
+
"step": 52000
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 0.74,
|
| 637 |
+
"learning_rate": 4.8013715084408925e-05,
|
| 638 |
+
"loss": 1.0224,
|
| 639 |
+
"step": 52500
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.75,
|
| 643 |
+
"learning_rate": 4.7989976707904974e-05,
|
| 644 |
+
"loss": 1.0163,
|
| 645 |
+
"step": 53000
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 0.75,
|
| 649 |
+
"learning_rate": 4.796623833140103e-05,
|
| 650 |
+
"loss": 1.0214,
|
| 651 |
+
"step": 53500
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 0.76,
|
| 655 |
+
"learning_rate": 4.7942499954897085e-05,
|
| 656 |
+
"loss": 1.0187,
|
| 657 |
+
"step": 54000
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 0.77,
|
| 661 |
+
"learning_rate": 4.791876157839314e-05,
|
| 662 |
+
"loss": 1.0091,
|
| 663 |
+
"step": 54500
|
| 664 |
+
},
|
| 665 |
+
{
|
| 666 |
+
"epoch": 0.78,
|
| 667 |
+
"learning_rate": 4.78950232018892e-05,
|
| 668 |
+
"loss": 1.0256,
|
| 669 |
+
"step": 55000
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 0.78,
|
| 673 |
+
"learning_rate": 4.787128482538525e-05,
|
| 674 |
+
"loss": 1.0025,
|
| 675 |
+
"step": 55500
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 0.79,
|
| 679 |
+
"learning_rate": 4.784754644888131e-05,
|
| 680 |
+
"loss": 1.0013,
|
| 681 |
+
"step": 56000
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.8,
|
| 685 |
+
"learning_rate": 4.782380807237736e-05,
|
| 686 |
+
"loss": 1.0098,
|
| 687 |
+
"step": 56500
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 0.8,
|
| 691 |
+
"learning_rate": 4.780006969587342e-05,
|
| 692 |
+
"loss": 1.017,
|
| 693 |
+
"step": 57000
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 0.81,
|
| 697 |
+
"learning_rate": 4.777637879612248e-05,
|
| 698 |
+
"loss": 1.0114,
|
| 699 |
+
"step": 57500
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 0.82,
|
| 703 |
+
"learning_rate": 4.775264041961854e-05,
|
| 704 |
+
"loss": 0.9977,
|
| 705 |
+
"step": 58000
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 0.82,
|
| 709 |
+
"learning_rate": 4.7728902043114596e-05,
|
| 710 |
+
"loss": 1.0107,
|
| 711 |
+
"step": 58500
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 0.83,
|
| 715 |
+
"learning_rate": 4.7705163666610644e-05,
|
| 716 |
+
"loss": 0.9869,
|
| 717 |
+
"step": 59000
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 0.84,
|
| 721 |
+
"learning_rate": 4.76814252901067e-05,
|
| 722 |
+
"loss": 1.008,
|
| 723 |
+
"step": 59500
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.85,
|
| 727 |
+
"learning_rate": 4.7657686913602755e-05,
|
| 728 |
+
"loss": 1.0025,
|
| 729 |
+
"step": 60000
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 0.85,
|
| 733 |
+
"learning_rate": 4.763394853709881e-05,
|
| 734 |
+
"loss": 0.9964,
|
| 735 |
+
"step": 60500
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 0.86,
|
| 739 |
+
"learning_rate": 4.7610210160594866e-05,
|
| 740 |
+
"loss": 0.9921,
|
| 741 |
+
"step": 61000
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 0.87,
|
| 745 |
+
"learning_rate": 4.758647178409092e-05,
|
| 746 |
+
"loss": 0.9938,
|
| 747 |
+
"step": 61500
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 0.87,
|
| 751 |
+
"learning_rate": 4.756273340758698e-05,
|
| 752 |
+
"loss": 0.9934,
|
| 753 |
+
"step": 62000
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 0.88,
|
| 757 |
+
"learning_rate": 4.753899503108303e-05,
|
| 758 |
+
"loss": 1.0249,
|
| 759 |
+
"step": 62500
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 0.89,
|
| 763 |
+
"learning_rate": 4.751530413133209e-05,
|
| 764 |
+
"loss": 0.9832,
|
| 765 |
+
"step": 63000
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.9,
|
| 769 |
+
"learning_rate": 4.749156575482815e-05,
|
| 770 |
+
"loss": 1.001,
|
| 771 |
+
"step": 63500
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 0.9,
|
| 775 |
+
"learning_rate": 4.74678273783242e-05,
|
| 776 |
+
"loss": 0.9755,
|
| 777 |
+
"step": 64000
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 0.91,
|
| 781 |
+
"learning_rate": 4.7444089001820266e-05,
|
| 782 |
+
"loss": 0.995,
|
| 783 |
+
"step": 64500
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 0.92,
|
| 787 |
+
"learning_rate": 4.742035062531632e-05,
|
| 788 |
+
"loss": 0.9824,
|
| 789 |
+
"step": 65000
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 0.92,
|
| 793 |
+
"learning_rate": 4.739661224881237e-05,
|
| 794 |
+
"loss": 0.9759,
|
| 795 |
+
"step": 65500
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 0.93,
|
| 799 |
+
"learning_rate": 4.7372873872308425e-05,
|
| 800 |
+
"loss": 0.98,
|
| 801 |
+
"step": 66000
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 0.94,
|
| 805 |
+
"learning_rate": 4.734913549580448e-05,
|
| 806 |
+
"loss": 0.9905,
|
| 807 |
+
"step": 66500
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.94,
|
| 811 |
+
"learning_rate": 4.7325397119300536e-05,
|
| 812 |
+
"loss": 0.9811,
|
| 813 |
+
"step": 67000
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 0.95,
|
| 817 |
+
"learning_rate": 4.730165874279659e-05,
|
| 818 |
+
"loss": 0.9873,
|
| 819 |
+
"step": 67500
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 0.96,
|
| 823 |
+
"learning_rate": 4.727792036629265e-05,
|
| 824 |
+
"loss": 0.9804,
|
| 825 |
+
"step": 68000
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 0.97,
|
| 829 |
+
"learning_rate": 4.7254229466541714e-05,
|
| 830 |
+
"loss": 0.9695,
|
| 831 |
+
"step": 68500
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 0.97,
|
| 835 |
+
"learning_rate": 4.723049109003776e-05,
|
| 836 |
+
"loss": 0.9751,
|
| 837 |
+
"step": 69000
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 0.98,
|
| 841 |
+
"learning_rate": 4.720675271353382e-05,
|
| 842 |
+
"loss": 0.9735,
|
| 843 |
+
"step": 69500
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 0.99,
|
| 847 |
+
"learning_rate": 4.7183061813782885e-05,
|
| 848 |
+
"loss": 0.9832,
|
| 849 |
+
"step": 70000
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.99,
|
| 853 |
+
"learning_rate": 4.7159370914031945e-05,
|
| 854 |
+
"loss": 0.9719,
|
| 855 |
+
"step": 70500
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 1.0,
|
| 859 |
+
"eval_bleu": 40.8492,
|
| 860 |
+
"eval_gen_len": 13.8028,
|
| 861 |
+
"eval_loss": 0.9435123801231384,
|
| 862 |
+
"eval_runtime": 10127.6159,
|
| 863 |
+
"eval_samples_per_second": 14.005,
|
| 864 |
+
"eval_steps_per_second": 1.751,
|
| 865 |
+
"step": 70919
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 1.0,
|
| 869 |
+
"learning_rate": 4.7135632537528e-05,
|
| 870 |
+
"loss": 0.9687,
|
| 871 |
+
"step": 71000
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 1.01,
|
| 875 |
+
"learning_rate": 4.7111894161024056e-05,
|
| 876 |
+
"loss": 0.9639,
|
| 877 |
+
"step": 71500
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 1.02,
|
| 881 |
+
"learning_rate": 4.708815578452011e-05,
|
| 882 |
+
"loss": 0.9669,
|
| 883 |
+
"step": 72000
|
| 884 |
+
},
|
| 885 |
+
{
|
| 886 |
+
"epoch": 1.02,
|
| 887 |
+
"learning_rate": 4.7064417408016166e-05,
|
| 888 |
+
"loss": 0.956,
|
| 889 |
+
"step": 72500
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"epoch": 1.03,
|
| 893 |
+
"learning_rate": 4.704067903151222e-05,
|
| 894 |
+
"loss": 0.9563,
|
| 895 |
+
"step": 73000
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"epoch": 1.04,
|
| 899 |
+
"learning_rate": 4.701694065500828e-05,
|
| 900 |
+
"loss": 0.9564,
|
| 901 |
+
"step": 73500
|
| 902 |
+
},
|
| 903 |
+
{
|
| 904 |
+
"epoch": 1.04,
|
| 905 |
+
"learning_rate": 4.699320227850433e-05,
|
| 906 |
+
"loss": 0.9555,
|
| 907 |
+
"step": 74000
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 1.05,
|
| 911 |
+
"learning_rate": 4.696946390200039e-05,
|
| 912 |
+
"loss": 0.9408,
|
| 913 |
+
"step": 74500
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 1.06,
|
| 917 |
+
"learning_rate": 4.6945725525496444e-05,
|
| 918 |
+
"loss": 0.9347,
|
| 919 |
+
"step": 75000
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 1.06,
|
| 923 |
+
"learning_rate": 4.69219871489925e-05,
|
| 924 |
+
"loss": 0.9368,
|
| 925 |
+
"step": 75500
|
| 926 |
+
},
|
| 927 |
+
{
|
| 928 |
+
"epoch": 1.07,
|
| 929 |
+
"learning_rate": 4.6898296249241566e-05,
|
| 930 |
+
"loss": 0.9401,
|
| 931 |
+
"step": 76000
|
| 932 |
+
},
|
| 933 |
+
{
|
| 934 |
+
"epoch": 1.08,
|
| 935 |
+
"learning_rate": 4.6874557872737615e-05,
|
| 936 |
+
"loss": 0.9383,
|
| 937 |
+
"step": 76500
|
| 938 |
+
},
|
| 939 |
+
{
|
| 940 |
+
"epoch": 1.09,
|
| 941 |
+
"learning_rate": 4.685081949623367e-05,
|
| 942 |
+
"loss": 0.9232,
|
| 943 |
+
"step": 77000
|
| 944 |
+
},
|
| 945 |
+
{
|
| 946 |
+
"epoch": 1.09,
|
| 947 |
+
"learning_rate": 4.6827081119729726e-05,
|
| 948 |
+
"loss": 0.9046,
|
| 949 |
+
"step": 77500
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 1.1,
|
| 953 |
+
"learning_rate": 4.680334274322578e-05,
|
| 954 |
+
"loss": 0.9236,
|
| 955 |
+
"step": 78000
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 1.11,
|
| 959 |
+
"learning_rate": 4.6779604366721837e-05,
|
| 960 |
+
"loss": 0.9091,
|
| 961 |
+
"step": 78500
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 1.11,
|
| 965 |
+
"learning_rate": 4.675586599021789e-05,
|
| 966 |
+
"loss": 0.9036,
|
| 967 |
+
"step": 79000
|
| 968 |
+
},
|
| 969 |
+
{
|
| 970 |
+
"epoch": 1.12,
|
| 971 |
+
"learning_rate": 4.673212761371395e-05,
|
| 972 |
+
"loss": 0.909,
|
| 973 |
+
"step": 79500
|
| 974 |
+
},
|
| 975 |
+
{
|
| 976 |
+
"epoch": 1.13,
|
| 977 |
+
"learning_rate": 4.670838923721e-05,
|
| 978 |
+
"loss": 0.9003,
|
| 979 |
+
"step": 80000
|
| 980 |
+
},
|
| 981 |
+
{
|
| 982 |
+
"epoch": 1.14,
|
| 983 |
+
"learning_rate": 4.668469833745906e-05,
|
| 984 |
+
"loss": 0.8957,
|
| 985 |
+
"step": 80500
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 1.14,
|
| 989 |
+
"learning_rate": 4.666095996095512e-05,
|
| 990 |
+
"loss": 0.8928,
|
| 991 |
+
"step": 81000
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 1.15,
|
| 995 |
+
"learning_rate": 4.66373640147102e-05,
|
| 996 |
+
"loss": 0.8842,
|
| 997 |
+
"step": 81500
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 1.16,
|
| 1001 |
+
"learning_rate": 4.6613625638206256e-05,
|
| 1002 |
+
"loss": 0.885,
|
| 1003 |
+
"step": 82000
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 1.16,
|
| 1007 |
+
"learning_rate": 4.658988726170231e-05,
|
| 1008 |
+
"loss": 0.8821,
|
| 1009 |
+
"step": 82500
|
| 1010 |
+
},
|
| 1011 |
+
{
|
| 1012 |
+
"epoch": 1.17,
|
| 1013 |
+
"learning_rate": 4.656614888519836e-05,
|
| 1014 |
+
"loss": 0.8769,
|
| 1015 |
+
"step": 83000
|
| 1016 |
+
},
|
| 1017 |
+
{
|
| 1018 |
+
"epoch": 1.18,
|
| 1019 |
+
"learning_rate": 4.6542410508694416e-05,
|
| 1020 |
+
"loss": 0.8681,
|
| 1021 |
+
"step": 83500
|
| 1022 |
+
},
|
| 1023 |
+
{
|
| 1024 |
+
"epoch": 1.18,
|
| 1025 |
+
"learning_rate": 4.651867213219048e-05,
|
| 1026 |
+
"loss": 0.8767,
|
| 1027 |
+
"step": 84000
|
| 1028 |
+
},
|
| 1029 |
+
{
|
| 1030 |
+
"epoch": 1.19,
|
| 1031 |
+
"learning_rate": 4.6494933755686534e-05,
|
| 1032 |
+
"loss": 0.8702,
|
| 1033 |
+
"step": 84500
|
| 1034 |
+
},
|
| 1035 |
+
{
|
| 1036 |
+
"epoch": 1.2,
|
| 1037 |
+
"learning_rate": 4.647119537918259e-05,
|
| 1038 |
+
"loss": 0.8753,
|
| 1039 |
+
"step": 85000
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 1.21,
|
| 1043 |
+
"learning_rate": 4.644745700267864e-05,
|
| 1044 |
+
"loss": 0.8587,
|
| 1045 |
+
"step": 85500
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 1.21,
|
| 1049 |
+
"learning_rate": 4.642371862617469e-05,
|
| 1050 |
+
"loss": 0.8459,
|
| 1051 |
+
"step": 86000
|
| 1052 |
+
},
|
| 1053 |
+
{
|
| 1054 |
+
"epoch": 1.22,
|
| 1055 |
+
"learning_rate": 4.639998024967075e-05,
|
| 1056 |
+
"loss": 0.853,
|
| 1057 |
+
"step": 86500
|
| 1058 |
+
},
|
| 1059 |
+
{
|
| 1060 |
+
"epoch": 1.23,
|
| 1061 |
+
"learning_rate": 4.6376241873166804e-05,
|
| 1062 |
+
"loss": 0.8811,
|
| 1063 |
+
"step": 87000
|
| 1064 |
+
},
|
| 1065 |
+
{
|
| 1066 |
+
"epoch": 1.23,
|
| 1067 |
+
"learning_rate": 4.635255097341587e-05,
|
| 1068 |
+
"loss": 0.8618,
|
| 1069 |
+
"step": 87500
|
| 1070 |
+
},
|
| 1071 |
+
{
|
| 1072 |
+
"epoch": 1.24,
|
| 1073 |
+
"learning_rate": 4.6328812596911926e-05,
|
| 1074 |
+
"loss": 0.8493,
|
| 1075 |
+
"step": 88000
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"epoch": 1.25,
|
| 1079 |
+
"learning_rate": 4.630507422040798e-05,
|
| 1080 |
+
"loss": 0.8603,
|
| 1081 |
+
"step": 88500
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 1.25,
|
| 1085 |
+
"learning_rate": 4.628133584390403e-05,
|
| 1086 |
+
"loss": 0.8333,
|
| 1087 |
+
"step": 89000
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 1.26,
|
| 1091 |
+
"learning_rate": 4.6257597467400086e-05,
|
| 1092 |
+
"loss": 0.8386,
|
| 1093 |
+
"step": 89500
|
| 1094 |
+
},
|
| 1095 |
+
{
|
| 1096 |
+
"epoch": 1.27,
|
| 1097 |
+
"learning_rate": 4.623385909089614e-05,
|
| 1098 |
+
"loss": 0.8424,
|
| 1099 |
+
"step": 90000
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 1.28,
|
| 1103 |
+
"learning_rate": 4.6210120714392204e-05,
|
| 1104 |
+
"loss": 0.8519,
|
| 1105 |
+
"step": 90500
|
| 1106 |
+
},
|
| 1107 |
+
{
|
| 1108 |
+
"epoch": 1.28,
|
| 1109 |
+
"learning_rate": 4.618638233788826e-05,
|
| 1110 |
+
"loss": 0.8297,
|
| 1111 |
+
"step": 91000
|
| 1112 |
+
},
|
| 1113 |
+
{
|
| 1114 |
+
"epoch": 1.29,
|
| 1115 |
+
"learning_rate": 4.616264396138431e-05,
|
| 1116 |
+
"loss": 0.8406,
|
| 1117 |
+
"step": 91500
|
| 1118 |
+
},
|
| 1119 |
+
{
|
| 1120 |
+
"epoch": 1.3,
|
| 1121 |
+
"learning_rate": 4.613890558488036e-05,
|
| 1122 |
+
"loss": 0.8392,
|
| 1123 |
+
"step": 92000
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 1.3,
|
| 1127 |
+
"learning_rate": 4.611516720837642e-05,
|
| 1128 |
+
"loss": 0.8269,
|
| 1129 |
+
"step": 92500
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 1.31,
|
| 1133 |
+
"learning_rate": 4.6091476308625486e-05,
|
| 1134 |
+
"loss": 0.8392,
|
| 1135 |
+
"step": 93000
|
| 1136 |
+
},
|
| 1137 |
+
{
|
| 1138 |
+
"epoch": 1.32,
|
| 1139 |
+
"learning_rate": 4.6067785408874545e-05,
|
| 1140 |
+
"loss": 0.8474,
|
| 1141 |
+
"step": 93500
|
| 1142 |
+
},
|
| 1143 |
+
{
|
| 1144 |
+
"epoch": 1.33,
|
| 1145 |
+
"learning_rate": 4.60440470323706e-05,
|
| 1146 |
+
"loss": 0.8329,
|
| 1147 |
+
"step": 94000
|
| 1148 |
+
},
|
| 1149 |
+
{
|
| 1150 |
+
"epoch": 1.33,
|
| 1151 |
+
"learning_rate": 4.6020308655866656e-05,
|
| 1152 |
+
"loss": 0.8327,
|
| 1153 |
+
"step": 94500
|
| 1154 |
+
},
|
| 1155 |
+
{
|
| 1156 |
+
"epoch": 1.34,
|
| 1157 |
+
"learning_rate": 4.599657027936271e-05,
|
| 1158 |
+
"loss": 0.8435,
|
| 1159 |
+
"step": 95000
|
| 1160 |
+
},
|
| 1161 |
+
{
|
| 1162 |
+
"epoch": 1.35,
|
| 1163 |
+
"learning_rate": 4.597283190285877e-05,
|
| 1164 |
+
"loss": 0.8343,
|
| 1165 |
+
"step": 95500
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 1.35,
|
| 1169 |
+
"learning_rate": 4.5949141003107834e-05,
|
| 1170 |
+
"loss": 0.8316,
|
| 1171 |
+
"step": 96000
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 1.36,
|
| 1175 |
+
"learning_rate": 4.592540262660389e-05,
|
| 1176 |
+
"loss": 0.8339,
|
| 1177 |
+
"step": 96500
|
| 1178 |
+
},
|
| 1179 |
+
{
|
| 1180 |
+
"epoch": 1.37,
|
| 1181 |
+
"learning_rate": 4.590166425009994e-05,
|
| 1182 |
+
"loss": 0.8199,
|
| 1183 |
+
"step": 97000
|
| 1184 |
+
},
|
| 1185 |
+
{
|
| 1186 |
+
"epoch": 1.37,
|
| 1187 |
+
"learning_rate": 4.5877925873595994e-05,
|
| 1188 |
+
"loss": 0.8162,
|
| 1189 |
+
"step": 97500
|
| 1190 |
+
},
|
| 1191 |
+
{
|
| 1192 |
+
"epoch": 1.38,
|
| 1193 |
+
"learning_rate": 4.585418749709205e-05,
|
| 1194 |
+
"loss": 0.8239,
|
| 1195 |
+
"step": 98000
|
| 1196 |
+
},
|
| 1197 |
+
{
|
| 1198 |
+
"epoch": 1.39,
|
| 1199 |
+
"learning_rate": 4.5830496597341116e-05,
|
| 1200 |
+
"loss": 0.8197,
|
| 1201 |
+
"step": 98500
|
| 1202 |
+
},
|
| 1203 |
+
{
|
| 1204 |
+
"epoch": 1.4,
|
| 1205 |
+
"learning_rate": 4.5806805697590176e-05,
|
| 1206 |
+
"loss": 0.8164,
|
| 1207 |
+
"step": 99000
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 1.4,
|
| 1211 |
+
"learning_rate": 4.578306732108623e-05,
|
| 1212 |
+
"loss": 0.818,
|
| 1213 |
+
"step": 99500
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 1.41,
|
| 1217 |
+
"learning_rate": 4.575932894458229e-05,
|
| 1218 |
+
"loss": 0.8189,
|
| 1219 |
+
"step": 100000
|
| 1220 |
+
},
|
| 1221 |
+
{
|
| 1222 |
+
"epoch": 1.42,
|
| 1223 |
+
"learning_rate": 4.573559056807834e-05,
|
| 1224 |
+
"loss": 0.8303,
|
| 1225 |
+
"step": 100500
|
| 1226 |
+
},
|
| 1227 |
+
{
|
| 1228 |
+
"epoch": 1.42,
|
| 1229 |
+
"learning_rate": 4.57118521915744e-05,
|
| 1230 |
+
"loss": 0.8174,
|
| 1231 |
+
"step": 101000
|
| 1232 |
+
},
|
| 1233 |
+
{
|
| 1234 |
+
"epoch": 1.43,
|
| 1235 |
+
"learning_rate": 4.568811381507045e-05,
|
| 1236 |
+
"loss": 0.814,
|
| 1237 |
+
"step": 101500
|
| 1238 |
+
},
|
| 1239 |
+
{
|
| 1240 |
+
"epoch": 1.44,
|
| 1241 |
+
"learning_rate": 4.566437543856651e-05,
|
| 1242 |
+
"loss": 0.8226,
|
| 1243 |
+
"step": 102000
|
| 1244 |
+
},
|
| 1245 |
+
{
|
| 1246 |
+
"epoch": 1.45,
|
| 1247 |
+
"learning_rate": 4.5640637062062564e-05,
|
| 1248 |
+
"loss": 0.818,
|
| 1249 |
+
"step": 102500
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 1.45,
|
| 1253 |
+
"learning_rate": 4.561689868555862e-05,
|
| 1254 |
+
"loss": 0.8054,
|
| 1255 |
+
"step": 103000
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 1.46,
|
| 1259 |
+
"learning_rate": 4.5593160309054675e-05,
|
| 1260 |
+
"loss": 0.8191,
|
| 1261 |
+
"step": 103500
|
| 1262 |
+
},
|
| 1263 |
+
{
|
| 1264 |
+
"epoch": 1.47,
|
| 1265 |
+
"learning_rate": 4.556946940930374e-05,
|
| 1266 |
+
"loss": 0.8134,
|
| 1267 |
+
"step": 104000
|
| 1268 |
+
},
|
| 1269 |
+
{
|
| 1270 |
+
"epoch": 1.47,
|
| 1271 |
+
"learning_rate": 4.554573103279979e-05,
|
| 1272 |
+
"loss": 0.8077,
|
| 1273 |
+
"step": 104500
|
| 1274 |
+
},
|
| 1275 |
+
{
|
| 1276 |
+
"epoch": 1.48,
|
| 1277 |
+
"learning_rate": 4.5521992656295846e-05,
|
| 1278 |
+
"loss": 0.8081,
|
| 1279 |
+
"step": 105000
|
| 1280 |
+
},
|
| 1281 |
+
{
|
| 1282 |
+
"epoch": 1.49,
|
| 1283 |
+
"learning_rate": 4.54982542797919e-05,
|
| 1284 |
+
"loss": 0.7985,
|
| 1285 |
+
"step": 105500
|
| 1286 |
+
},
|
| 1287 |
+
{
|
| 1288 |
+
"epoch": 1.49,
|
| 1289 |
+
"learning_rate": 4.547451590328796e-05,
|
| 1290 |
+
"loss": 0.8064,
|
| 1291 |
+
"step": 106000
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 1.5,
|
| 1295 |
+
"learning_rate": 4.545077752678401e-05,
|
| 1296 |
+
"loss": 0.8008,
|
| 1297 |
+
"step": 106500
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 1.51,
|
| 1301 |
+
"learning_rate": 4.542703915028007e-05,
|
| 1302 |
+
"loss": 0.7934,
|
| 1303 |
+
"step": 107000
|
| 1304 |
+
},
|
| 1305 |
+
{
|
| 1306 |
+
"epoch": 1.52,
|
| 1307 |
+
"learning_rate": 4.540330077377612e-05,
|
| 1308 |
+
"loss": 0.7915,
|
| 1309 |
+
"step": 107500
|
| 1310 |
+
},
|
| 1311 |
+
{
|
| 1312 |
+
"epoch": 1.52,
|
| 1313 |
+
"learning_rate": 4.537956239727218e-05,
|
| 1314 |
+
"loss": 0.8087,
|
| 1315 |
+
"step": 108000
|
| 1316 |
+
},
|
| 1317 |
+
{
|
| 1318 |
+
"epoch": 1.53,
|
| 1319 |
+
"learning_rate": 4.5355824020768234e-05,
|
| 1320 |
+
"loss": 0.7941,
|
| 1321 |
+
"step": 108500
|
| 1322 |
+
},
|
| 1323 |
+
{
|
| 1324 |
+
"epoch": 1.54,
|
| 1325 |
+
"learning_rate": 4.533208564426429e-05,
|
| 1326 |
+
"loss": 0.8032,
|
| 1327 |
+
"step": 109000
|
| 1328 |
+
},
|
| 1329 |
+
{
|
| 1330 |
+
"epoch": 1.54,
|
| 1331 |
+
"learning_rate": 4.5308347267760345e-05,
|
| 1332 |
+
"loss": 0.8061,
|
| 1333 |
+
"step": 109500
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 1.55,
|
| 1337 |
+
"learning_rate": 4.5284608891256394e-05,
|
| 1338 |
+
"loss": 0.7974,
|
| 1339 |
+
"step": 110000
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 1.56,
|
| 1343 |
+
"learning_rate": 4.526091799150546e-05,
|
| 1344 |
+
"loss": 0.7915,
|
| 1345 |
+
"step": 110500
|
| 1346 |
+
},
|
| 1347 |
+
{
|
| 1348 |
+
"epoch": 1.57,
|
| 1349 |
+
"learning_rate": 4.523722709175453e-05,
|
| 1350 |
+
"loss": 0.7998,
|
| 1351 |
+
"step": 111000
|
| 1352 |
+
},
|
| 1353 |
+
{
|
| 1354 |
+
"epoch": 1.57,
|
| 1355 |
+
"learning_rate": 4.5213488715250576e-05,
|
| 1356 |
+
"loss": 0.8014,
|
| 1357 |
+
"step": 111500
|
| 1358 |
+
},
|
| 1359 |
+
{
|
| 1360 |
+
"epoch": 1.58,
|
| 1361 |
+
"learning_rate": 4.518975033874663e-05,
|
| 1362 |
+
"loss": 0.7848,
|
| 1363 |
+
"step": 112000
|
| 1364 |
+
},
|
| 1365 |
+
{
|
| 1366 |
+
"epoch": 1.59,
|
| 1367 |
+
"learning_rate": 4.516601196224269e-05,
|
| 1368 |
+
"loss": 0.7914,
|
| 1369 |
+
"step": 112500
|
| 1370 |
+
},
|
| 1371 |
+
{
|
| 1372 |
+
"epoch": 1.59,
|
| 1373 |
+
"learning_rate": 4.5142321062491754e-05,
|
| 1374 |
+
"loss": 0.795,
|
| 1375 |
+
"step": 113000
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 1.6,
|
| 1379 |
+
"learning_rate": 4.5118630162740814e-05,
|
| 1380 |
+
"loss": 0.7913,
|
| 1381 |
+
"step": 113500
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 1.61,
|
| 1385 |
+
"learning_rate": 4.509489178623687e-05,
|
| 1386 |
+
"loss": 0.7936,
|
| 1387 |
+
"step": 114000
|
| 1388 |
+
},
|
| 1389 |
+
{
|
| 1390 |
+
"epoch": 1.61,
|
| 1391 |
+
"learning_rate": 4.5071153409732924e-05,
|
| 1392 |
+
"loss": 0.7932,
|
| 1393 |
+
"step": 114500
|
| 1394 |
+
},
|
| 1395 |
+
{
|
| 1396 |
+
"epoch": 1.62,
|
| 1397 |
+
"learning_rate": 4.504741503322898e-05,
|
| 1398 |
+
"loss": 0.7835,
|
| 1399 |
+
"step": 115000
|
| 1400 |
+
},
|
| 1401 |
+
{
|
| 1402 |
+
"epoch": 1.63,
|
| 1403 |
+
"learning_rate": 4.502367665672504e-05,
|
| 1404 |
+
"loss": 0.7942,
|
| 1405 |
+
"step": 115500
|
| 1406 |
+
},
|
| 1407 |
+
{
|
| 1408 |
+
"epoch": 1.64,
|
| 1409 |
+
"learning_rate": 4.499993828022109e-05,
|
| 1410 |
+
"loss": 0.7771,
|
| 1411 |
+
"step": 116000
|
| 1412 |
+
},
|
| 1413 |
+
{
|
| 1414 |
+
"epoch": 1.64,
|
| 1415 |
+
"learning_rate": 4.4976199903717146e-05,
|
| 1416 |
+
"loss": 0.7805,
|
| 1417 |
+
"step": 116500
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 1.65,
|
| 1421 |
+
"learning_rate": 4.4952509003966206e-05,
|
| 1422 |
+
"loss": 0.7952,
|
| 1423 |
+
"step": 117000
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 1.66,
|
| 1427 |
+
"learning_rate": 4.492877062746226e-05,
|
| 1428 |
+
"loss": 0.7799,
|
| 1429 |
+
"step": 117500
|
| 1430 |
+
},
|
| 1431 |
+
{
|
| 1432 |
+
"epoch": 1.66,
|
| 1433 |
+
"learning_rate": 4.490503225095832e-05,
|
| 1434 |
+
"loss": 0.7932,
|
| 1435 |
+
"step": 118000
|
| 1436 |
+
},
|
| 1437 |
+
{
|
| 1438 |
+
"epoch": 1.67,
|
| 1439 |
+
"learning_rate": 4.488129387445438e-05,
|
| 1440 |
+
"loss": 0.7919,
|
| 1441 |
+
"step": 118500
|
| 1442 |
+
},
|
| 1443 |
+
{
|
| 1444 |
+
"epoch": 1.68,
|
| 1445 |
+
"learning_rate": 4.4857555497950435e-05,
|
| 1446 |
+
"loss": 0.7776,
|
| 1447 |
+
"step": 119000
|
| 1448 |
+
},
|
| 1449 |
+
{
|
| 1450 |
+
"epoch": 1.69,
|
| 1451 |
+
"learning_rate": 4.4833864598199495e-05,
|
| 1452 |
+
"loss": 0.7731,
|
| 1453 |
+
"step": 119500
|
| 1454 |
+
},
|
| 1455 |
+
{
|
| 1456 |
+
"epoch": 1.69,
|
| 1457 |
+
"learning_rate": 4.481012622169555e-05,
|
| 1458 |
+
"loss": 0.7778,
|
| 1459 |
+
"step": 120000
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 1.7,
|
| 1463 |
+
"learning_rate": 4.47863878451916e-05,
|
| 1464 |
+
"loss": 0.7638,
|
| 1465 |
+
"step": 120500
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 1.71,
|
| 1469 |
+
"learning_rate": 4.476264946868766e-05,
|
| 1470 |
+
"loss": 0.7819,
|
| 1471 |
+
"step": 121000
|
| 1472 |
+
},
|
| 1473 |
+
{
|
| 1474 |
+
"epoch": 1.71,
|
| 1475 |
+
"learning_rate": 4.473891109218372e-05,
|
| 1476 |
+
"loss": 0.7717,
|
| 1477 |
+
"step": 121500
|
| 1478 |
+
},
|
| 1479 |
+
{
|
| 1480 |
+
"epoch": 1.72,
|
| 1481 |
+
"learning_rate": 4.471517271567977e-05,
|
| 1482 |
+
"loss": 0.7803,
|
| 1483 |
+
"step": 122000
|
| 1484 |
+
},
|
| 1485 |
+
{
|
| 1486 |
+
"epoch": 1.73,
|
| 1487 |
+
"learning_rate": 4.469143433917583e-05,
|
| 1488 |
+
"loss": 0.7807,
|
| 1489 |
+
"step": 122500
|
| 1490 |
+
},
|
| 1491 |
+
{
|
| 1492 |
+
"epoch": 1.73,
|
| 1493 |
+
"learning_rate": 4.4667695962671876e-05,
|
| 1494 |
+
"loss": 0.7651,
|
| 1495 |
+
"step": 123000
|
| 1496 |
+
},
|
| 1497 |
+
{
|
| 1498 |
+
"epoch": 1.74,
|
| 1499 |
+
"learning_rate": 4.464395758616793e-05,
|
| 1500 |
+
"loss": 0.7811,
|
| 1501 |
+
"step": 123500
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 1.75,
|
| 1505 |
+
"learning_rate": 4.4620266686417e-05,
|
| 1506 |
+
"loss": 0.7794,
|
| 1507 |
+
"step": 124000
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 1.76,
|
| 1511 |
+
"learning_rate": 4.4596528309913054e-05,
|
| 1512 |
+
"loss": 0.7763,
|
| 1513 |
+
"step": 124500
|
| 1514 |
+
},
|
| 1515 |
+
{
|
| 1516 |
+
"epoch": 1.76,
|
| 1517 |
+
"learning_rate": 4.457278993340911e-05,
|
| 1518 |
+
"loss": 0.7757,
|
| 1519 |
+
"step": 125000
|
| 1520 |
+
},
|
| 1521 |
+
{
|
| 1522 |
+
"epoch": 1.77,
|
| 1523 |
+
"learning_rate": 4.4549051556905165e-05,
|
| 1524 |
+
"loss": 0.7683,
|
| 1525 |
+
"step": 125500
|
| 1526 |
+
},
|
| 1527 |
+
{
|
| 1528 |
+
"epoch": 1.78,
|
| 1529 |
+
"learning_rate": 4.4525360657154225e-05,
|
| 1530 |
+
"loss": 0.7764,
|
| 1531 |
+
"step": 126000
|
| 1532 |
+
},
|
| 1533 |
+
{
|
| 1534 |
+
"epoch": 1.78,
|
| 1535 |
+
"learning_rate": 4.450162228065028e-05,
|
| 1536 |
+
"loss": 0.7661,
|
| 1537 |
+
"step": 126500
|
| 1538 |
+
},
|
| 1539 |
+
{
|
| 1540 |
+
"epoch": 1.79,
|
| 1541 |
+
"learning_rate": 4.4477883904146336e-05,
|
| 1542 |
+
"loss": 0.7645,
|
| 1543 |
+
"step": 127000
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 1.8,
|
| 1547 |
+
"learning_rate": 4.445414552764239e-05,
|
| 1548 |
+
"loss": 0.7703,
|
| 1549 |
+
"step": 127500
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 1.8,
|
| 1553 |
+
"learning_rate": 4.443040715113845e-05,
|
| 1554 |
+
"loss": 0.7791,
|
| 1555 |
+
"step": 128000
|
| 1556 |
+
},
|
| 1557 |
+
{
|
| 1558 |
+
"epoch": 1.81,
|
| 1559 |
+
"learning_rate": 4.44066687746345e-05,
|
| 1560 |
+
"loss": 0.7657,
|
| 1561 |
+
"step": 128500
|
| 1562 |
+
},
|
| 1563 |
+
{
|
| 1564 |
+
"epoch": 1.82,
|
| 1565 |
+
"learning_rate": 4.438293039813056e-05,
|
| 1566 |
+
"loss": 0.7651,
|
| 1567 |
+
"step": 129000
|
| 1568 |
+
},
|
| 1569 |
+
{
|
| 1570 |
+
"epoch": 1.83,
|
| 1571 |
+
"learning_rate": 4.435919202162661e-05,
|
| 1572 |
+
"loss": 0.7803,
|
| 1573 |
+
"step": 129500
|
| 1574 |
+
},
|
| 1575 |
+
{
|
| 1576 |
+
"epoch": 1.83,
|
| 1577 |
+
"learning_rate": 4.433545364512267e-05,
|
| 1578 |
+
"loss": 0.7504,
|
| 1579 |
+
"step": 130000
|
| 1580 |
+
},
|
| 1581 |
+
{
|
| 1582 |
+
"epoch": 1.84,
|
| 1583 |
+
"learning_rate": 4.4311762745371735e-05,
|
| 1584 |
+
"loss": 0.7785,
|
| 1585 |
+
"step": 130500
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 1.85,
|
| 1589 |
+
"learning_rate": 4.4288024368867784e-05,
|
| 1590 |
+
"loss": 0.7592,
|
| 1591 |
+
"step": 131000
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 1.85,
|
| 1595 |
+
"learning_rate": 4.426428599236384e-05,
|
| 1596 |
+
"loss": 0.7714,
|
| 1597 |
+
"step": 131500
|
| 1598 |
+
},
|
| 1599 |
+
{
|
| 1600 |
+
"epoch": 1.86,
|
| 1601 |
+
"learning_rate": 4.4240547615859895e-05,
|
| 1602 |
+
"loss": 0.7652,
|
| 1603 |
+
"step": 132000
|
| 1604 |
+
},
|
| 1605 |
+
{
|
| 1606 |
+
"epoch": 1.87,
|
| 1607 |
+
"learning_rate": 4.421680923935595e-05,
|
| 1608 |
+
"loss": 0.7608,
|
| 1609 |
+
"step": 132500
|
| 1610 |
+
},
|
| 1611 |
+
{
|
| 1612 |
+
"epoch": 1.88,
|
| 1613 |
+
"learning_rate": 4.4193070862852006e-05,
|
| 1614 |
+
"loss": 0.7688,
|
| 1615 |
+
"step": 133000
|
| 1616 |
+
},
|
| 1617 |
+
{
|
| 1618 |
+
"epoch": 1.88,
|
| 1619 |
+
"learning_rate": 4.416933248634806e-05,
|
| 1620 |
+
"loss": 0.79,
|
| 1621 |
+
"step": 133500
|
| 1622 |
+
},
|
| 1623 |
+
{
|
| 1624 |
+
"epoch": 1.89,
|
| 1625 |
+
"learning_rate": 4.414559410984412e-05,
|
| 1626 |
+
"loss": 0.7526,
|
| 1627 |
+
"step": 134000
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 1.9,
|
| 1631 |
+
"learning_rate": 4.412190321009318e-05,
|
| 1632 |
+
"loss": 0.765,
|
| 1633 |
+
"step": 134500
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 1.9,
|
| 1637 |
+
"learning_rate": 4.409816483358923e-05,
|
| 1638 |
+
"loss": 0.7554,
|
| 1639 |
+
"step": 135000
|
| 1640 |
+
},
|
| 1641 |
+
{
|
| 1642 |
+
"epoch": 1.91,
|
| 1643 |
+
"learning_rate": 4.407442645708529e-05,
|
| 1644 |
+
"loss": 0.7636,
|
| 1645 |
+
"step": 135500
|
| 1646 |
+
},
|
| 1647 |
+
{
|
| 1648 |
+
"epoch": 1.92,
|
| 1649 |
+
"learning_rate": 4.4050735557334354e-05,
|
| 1650 |
+
"loss": 0.7582,
|
| 1651 |
+
"step": 136000
|
| 1652 |
+
},
|
| 1653 |
+
{
|
| 1654 |
+
"epoch": 1.92,
|
| 1655 |
+
"learning_rate": 4.402699718083041e-05,
|
| 1656 |
+
"loss": 0.7504,
|
| 1657 |
+
"step": 136500
|
| 1658 |
+
},
|
| 1659 |
+
{
|
| 1660 |
+
"epoch": 1.93,
|
| 1661 |
+
"learning_rate": 4.400330628107947e-05,
|
| 1662 |
+
"loss": 0.7591,
|
| 1663 |
+
"step": 137000
|
| 1664 |
+
},
|
| 1665 |
+
{
|
| 1666 |
+
"epoch": 1.94,
|
| 1667 |
+
"learning_rate": 4.3979567904575525e-05,
|
| 1668 |
+
"loss": 0.7629,
|
| 1669 |
+
"step": 137500
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 1.95,
|
| 1673 |
+
"learning_rate": 4.395582952807158e-05,
|
| 1674 |
+
"loss": 0.7543,
|
| 1675 |
+
"step": 138000
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 1.95,
|
| 1679 |
+
"learning_rate": 4.3932091151567636e-05,
|
| 1680 |
+
"loss": 0.7634,
|
| 1681 |
+
"step": 138500
|
| 1682 |
+
},
|
| 1683 |
+
{
|
| 1684 |
+
"epoch": 1.96,
|
| 1685 |
+
"learning_rate": 4.390835277506369e-05,
|
| 1686 |
+
"loss": 0.7553,
|
| 1687 |
+
"step": 139000
|
| 1688 |
+
},
|
| 1689 |
+
{
|
| 1690 |
+
"epoch": 1.97,
|
| 1691 |
+
"learning_rate": 4.388461439855975e-05,
|
| 1692 |
+
"loss": 0.7548,
|
| 1693 |
+
"step": 139500
|
| 1694 |
+
},
|
| 1695 |
+
{
|
| 1696 |
+
"epoch": 1.97,
|
| 1697 |
+
"learning_rate": 4.38608760220558e-05,
|
| 1698 |
+
"loss": 0.7554,
|
| 1699 |
+
"step": 140000
|
| 1700 |
+
},
|
| 1701 |
+
{
|
| 1702 |
+
"epoch": 1.98,
|
| 1703 |
+
"learning_rate": 4.383713764555186e-05,
|
| 1704 |
+
"loss": 0.7539,
|
| 1705 |
+
"step": 140500
|
| 1706 |
+
},
|
| 1707 |
+
{
|
| 1708 |
+
"epoch": 1.99,
|
| 1709 |
+
"learning_rate": 4.3813399269047914e-05,
|
| 1710 |
+
"loss": 0.7562,
|
| 1711 |
+
"step": 141000
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 2.0,
|
| 1715 |
+
"learning_rate": 4.378966089254396e-05,
|
| 1716 |
+
"loss": 0.7537,
|
| 1717 |
+
"step": 141500
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 2.0,
|
| 1721 |
+
"eval_bleu": 42.7907,
|
| 1722 |
+
"eval_gen_len": 13.7941,
|
| 1723 |
+
"eval_loss": 0.8987888693809509,
|
| 1724 |
+
"eval_runtime": 10044.1408,
|
| 1725 |
+
"eval_samples_per_second": 14.121,
|
| 1726 |
+
"eval_steps_per_second": 1.765,
|
| 1727 |
+
"step": 141839
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"epoch": 2.0,
|
| 1731 |
+
"learning_rate": 4.376596999279303e-05,
|
| 1732 |
+
"loss": 0.7511,
|
| 1733 |
+
"step": 142000
|
| 1734 |
+
},
|
| 1735 |
+
{
|
| 1736 |
+
"epoch": 2.01,
|
| 1737 |
+
"learning_rate": 4.3742231616289084e-05,
|
| 1738 |
+
"loss": 0.75,
|
| 1739 |
+
"step": 142500
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 2.02,
|
| 1743 |
+
"learning_rate": 4.371849323978514e-05,
|
| 1744 |
+
"loss": 0.7489,
|
| 1745 |
+
"step": 143000
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 2.02,
|
| 1749 |
+
"learning_rate": 4.3694754863281195e-05,
|
| 1750 |
+
"loss": 0.7359,
|
| 1751 |
+
"step": 143500
|
| 1752 |
+
},
|
| 1753 |
+
{
|
| 1754 |
+
"epoch": 2.03,
|
| 1755 |
+
"learning_rate": 4.367101648677725e-05,
|
| 1756 |
+
"loss": 0.7416,
|
| 1757 |
+
"step": 144000
|
| 1758 |
+
},
|
| 1759 |
+
{
|
| 1760 |
+
"epoch": 2.04,
|
| 1761 |
+
"learning_rate": 4.364732558702632e-05,
|
| 1762 |
+
"loss": 0.7421,
|
| 1763 |
+
"step": 144500
|
| 1764 |
+
},
|
| 1765 |
+
{
|
| 1766 |
+
"epoch": 2.04,
|
| 1767 |
+
"learning_rate": 4.362358721052237e-05,
|
| 1768 |
+
"loss": 0.7448,
|
| 1769 |
+
"step": 145000
|
| 1770 |
+
},
|
| 1771 |
+
{
|
| 1772 |
+
"epoch": 2.05,
|
| 1773 |
+
"learning_rate": 4.359984883401843e-05,
|
| 1774 |
+
"loss": 0.7314,
|
| 1775 |
+
"step": 145500
|
| 1776 |
+
},
|
| 1777 |
+
{
|
| 1778 |
+
"epoch": 2.06,
|
| 1779 |
+
"learning_rate": 4.357611045751448e-05,
|
| 1780 |
+
"loss": 0.731,
|
| 1781 |
+
"step": 146000
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 2.07,
|
| 1785 |
+
"learning_rate": 4.3552419557763544e-05,
|
| 1786 |
+
"loss": 0.7307,
|
| 1787 |
+
"step": 146500
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 2.07,
|
| 1791 |
+
"learning_rate": 4.35286811812596e-05,
|
| 1792 |
+
"loss": 0.7411,
|
| 1793 |
+
"step": 147000
|
| 1794 |
+
},
|
| 1795 |
+
{
|
| 1796 |
+
"epoch": 2.08,
|
| 1797 |
+
"learning_rate": 4.3504942804755655e-05,
|
| 1798 |
+
"loss": 0.7298,
|
| 1799 |
+
"step": 147500
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"epoch": 2.09,
|
| 1803 |
+
"learning_rate": 4.348120442825171e-05,
|
| 1804 |
+
"loss": 0.7144,
|
| 1805 |
+
"step": 148000
|
| 1806 |
+
},
|
| 1807 |
+
{
|
| 1808 |
+
"epoch": 2.09,
|
| 1809 |
+
"learning_rate": 4.3457466051747766e-05,
|
| 1810 |
+
"loss": 0.7146,
|
| 1811 |
+
"step": 148500
|
| 1812 |
+
},
|
| 1813 |
+
{
|
| 1814 |
+
"epoch": 2.1,
|
| 1815 |
+
"learning_rate": 4.343372767524382e-05,
|
| 1816 |
+
"loss": 0.7157,
|
| 1817 |
+
"step": 149000
|
| 1818 |
+
},
|
| 1819 |
+
{
|
| 1820 |
+
"epoch": 2.11,
|
| 1821 |
+
"learning_rate": 4.341003677549288e-05,
|
| 1822 |
+
"loss": 0.7116,
|
| 1823 |
+
"step": 149500
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 2.12,
|
| 1827 |
+
"learning_rate": 4.338629839898894e-05,
|
| 1828 |
+
"loss": 0.7105,
|
| 1829 |
+
"step": 150000
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 2.12,
|
| 1833 |
+
"learning_rate": 4.336256002248499e-05,
|
| 1834 |
+
"loss": 0.7096,
|
| 1835 |
+
"step": 150500
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"epoch": 2.13,
|
| 1839 |
+
"learning_rate": 4.3338916599487063e-05,
|
| 1840 |
+
"loss": 0.7026,
|
| 1841 |
+
"step": 151000
|
| 1842 |
+
},
|
| 1843 |
+
{
|
| 1844 |
+
"epoch": 2.14,
|
| 1845 |
+
"learning_rate": 4.331517822298312e-05,
|
| 1846 |
+
"loss": 0.706,
|
| 1847 |
+
"step": 151500
|
| 1848 |
+
},
|
| 1849 |
+
{
|
| 1850 |
+
"epoch": 2.14,
|
| 1851 |
+
"learning_rate": 4.3291439846479174e-05,
|
| 1852 |
+
"loss": 0.6996,
|
| 1853 |
+
"step": 152000
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"epoch": 2.15,
|
| 1857 |
+
"learning_rate": 4.326770146997523e-05,
|
| 1858 |
+
"loss": 0.6965,
|
| 1859 |
+
"step": 152500
|
| 1860 |
+
},
|
| 1861 |
+
{
|
| 1862 |
+
"epoch": 2.16,
|
| 1863 |
+
"learning_rate": 4.3243963093471285e-05,
|
| 1864 |
+
"loss": 0.699,
|
| 1865 |
+
"step": 153000
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 2.16,
|
| 1869 |
+
"learning_rate": 4.322022471696734e-05,
|
| 1870 |
+
"loss": 0.6926,
|
| 1871 |
+
"step": 153500
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 2.17,
|
| 1875 |
+
"learning_rate": 4.3196486340463396e-05,
|
| 1876 |
+
"loss": 0.6925,
|
| 1877 |
+
"step": 154000
|
| 1878 |
+
},
|
| 1879 |
+
{
|
| 1880 |
+
"epoch": 2.18,
|
| 1881 |
+
"learning_rate": 4.3172747963959445e-05,
|
| 1882 |
+
"loss": 0.6825,
|
| 1883 |
+
"step": 154500
|
| 1884 |
+
},
|
| 1885 |
+
{
|
| 1886 |
+
"epoch": 2.19,
|
| 1887 |
+
"learning_rate": 4.31490095874555e-05,
|
| 1888 |
+
"loss": 0.6869,
|
| 1889 |
+
"step": 155000
|
| 1890 |
+
},
|
| 1891 |
+
{
|
| 1892 |
+
"epoch": 2.19,
|
| 1893 |
+
"learning_rate": 4.312527121095156e-05,
|
| 1894 |
+
"loss": 0.691,
|
| 1895 |
+
"step": 155500
|
| 1896 |
+
},
|
| 1897 |
+
{
|
| 1898 |
+
"epoch": 2.2,
|
| 1899 |
+
"learning_rate": 4.310153283444762e-05,
|
| 1900 |
+
"loss": 0.682,
|
| 1901 |
+
"step": 156000
|
| 1902 |
+
},
|
| 1903 |
+
{
|
| 1904 |
+
"epoch": 2.21,
|
| 1905 |
+
"learning_rate": 4.3077794457943673e-05,
|
| 1906 |
+
"loss": 0.6822,
|
| 1907 |
+
"step": 156500
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"epoch": 2.21,
|
| 1911 |
+
"learning_rate": 4.305405608143972e-05,
|
| 1912 |
+
"loss": 0.6666,
|
| 1913 |
+
"step": 157000
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 2.22,
|
| 1917 |
+
"learning_rate": 4.303031770493578e-05,
|
| 1918 |
+
"loss": 0.6749,
|
| 1919 |
+
"step": 157500
|
| 1920 |
+
},
|
| 1921 |
+
{
|
| 1922 |
+
"epoch": 2.23,
|
| 1923 |
+
"learning_rate": 4.300657932843183e-05,
|
| 1924 |
+
"loss": 0.6903,
|
| 1925 |
+
"step": 158000
|
| 1926 |
+
},
|
| 1927 |
+
{
|
| 1928 |
+
"epoch": 2.23,
|
| 1929 |
+
"learning_rate": 4.29828884286809e-05,
|
| 1930 |
+
"loss": 0.6816,
|
| 1931 |
+
"step": 158500
|
| 1932 |
+
},
|
| 1933 |
+
{
|
| 1934 |
+
"epoch": 2.24,
|
| 1935 |
+
"learning_rate": 4.2959150052176955e-05,
|
| 1936 |
+
"loss": 0.6776,
|
| 1937 |
+
"step": 159000
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"epoch": 2.25,
|
| 1941 |
+
"learning_rate": 4.2935459152426015e-05,
|
| 1942 |
+
"loss": 0.6787,
|
| 1943 |
+
"step": 159500
|
| 1944 |
+
},
|
| 1945 |
+
{
|
| 1946 |
+
"epoch": 2.26,
|
| 1947 |
+
"learning_rate": 4.2911768252675075e-05,
|
| 1948 |
+
"loss": 0.6559,
|
| 1949 |
+
"step": 160000
|
| 1950 |
+
},
|
| 1951 |
+
{
|
| 1952 |
+
"epoch": 2.26,
|
| 1953 |
+
"learning_rate": 4.288807735292414e-05,
|
| 1954 |
+
"loss": 0.668,
|
| 1955 |
+
"step": 160500
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 2.27,
|
| 1959 |
+
"learning_rate": 4.28643389764202e-05,
|
| 1960 |
+
"loss": 0.6755,
|
| 1961 |
+
"step": 161000
|
| 1962 |
+
},
|
| 1963 |
+
{
|
| 1964 |
+
"epoch": 2.28,
|
| 1965 |
+
"learning_rate": 4.284060059991625e-05,
|
| 1966 |
+
"loss": 0.6751,
|
| 1967 |
+
"step": 161500
|
| 1968 |
+
},
|
| 1969 |
+
{
|
| 1970 |
+
"epoch": 2.28,
|
| 1971 |
+
"learning_rate": 4.281686222341231e-05,
|
| 1972 |
+
"loss": 0.6608,
|
| 1973 |
+
"step": 162000
|
| 1974 |
+
},
|
| 1975 |
+
{
|
| 1976 |
+
"epoch": 2.29,
|
| 1977 |
+
"learning_rate": 4.2793123846908364e-05,
|
| 1978 |
+
"loss": 0.6562,
|
| 1979 |
+
"step": 162500
|
| 1980 |
+
},
|
| 1981 |
+
{
|
| 1982 |
+
"epoch": 2.3,
|
| 1983 |
+
"learning_rate": 4.276938547040442e-05,
|
| 1984 |
+
"loss": 0.6672,
|
| 1985 |
+
"step": 163000
|
| 1986 |
+
},
|
| 1987 |
+
{
|
| 1988 |
+
"epoch": 2.31,
|
| 1989 |
+
"learning_rate": 4.2745647093900475e-05,
|
| 1990 |
+
"loss": 0.6577,
|
| 1991 |
+
"step": 163500
|
| 1992 |
+
},
|
| 1993 |
+
{
|
| 1994 |
+
"epoch": 2.31,
|
| 1995 |
+
"learning_rate": 4.272190871739653e-05,
|
| 1996 |
+
"loss": 0.6714,
|
| 1997 |
+
"step": 164000
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 2.32,
|
| 2001 |
+
"learning_rate": 4.2698170340892586e-05,
|
| 2002 |
+
"loss": 0.6706,
|
| 2003 |
+
"step": 164500
|
| 2004 |
+
},
|
| 2005 |
+
{
|
| 2006 |
+
"epoch": 2.33,
|
| 2007 |
+
"learning_rate": 4.2674479441141646e-05,
|
| 2008 |
+
"loss": 0.6593,
|
| 2009 |
+
"step": 165000
|
| 2010 |
+
},
|
| 2011 |
+
{
|
| 2012 |
+
"epoch": 2.33,
|
| 2013 |
+
"learning_rate": 4.26507410646377e-05,
|
| 2014 |
+
"loss": 0.6633,
|
| 2015 |
+
"step": 165500
|
| 2016 |
+
},
|
| 2017 |
+
{
|
| 2018 |
+
"epoch": 2.34,
|
| 2019 |
+
"learning_rate": 4.262705016488677e-05,
|
| 2020 |
+
"loss": 0.6672,
|
| 2021 |
+
"step": 166000
|
| 2022 |
+
},
|
| 2023 |
+
{
|
| 2024 |
+
"epoch": 2.35,
|
| 2025 |
+
"learning_rate": 4.260331178838282e-05,
|
| 2026 |
+
"loss": 0.6584,
|
| 2027 |
+
"step": 166500
|
| 2028 |
+
},
|
| 2029 |
+
{
|
| 2030 |
+
"epoch": 2.35,
|
| 2031 |
+
"learning_rate": 4.257957341187888e-05,
|
| 2032 |
+
"loss": 0.6593,
|
| 2033 |
+
"step": 167000
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 2.36,
|
| 2037 |
+
"learning_rate": 4.255583503537493e-05,
|
| 2038 |
+
"loss": 0.6651,
|
| 2039 |
+
"step": 167500
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 2.37,
|
| 2043 |
+
"learning_rate": 4.253209665887098e-05,
|
| 2044 |
+
"loss": 0.6519,
|
| 2045 |
+
"step": 168000
|
| 2046 |
+
},
|
| 2047 |
+
{
|
| 2048 |
+
"epoch": 2.38,
|
| 2049 |
+
"learning_rate": 4.250835828236704e-05,
|
| 2050 |
+
"loss": 0.6518,
|
| 2051 |
+
"step": 168500
|
| 2052 |
+
},
|
| 2053 |
+
{
|
| 2054 |
+
"epoch": 2.38,
|
| 2055 |
+
"learning_rate": 4.24846199058631e-05,
|
| 2056 |
+
"loss": 0.6598,
|
| 2057 |
+
"step": 169000
|
| 2058 |
+
},
|
| 2059 |
+
{
|
| 2060 |
+
"epoch": 2.39,
|
| 2061 |
+
"learning_rate": 4.2460881529359156e-05,
|
| 2062 |
+
"loss": 0.6445,
|
| 2063 |
+
"step": 169500
|
| 2064 |
+
},
|
| 2065 |
+
{
|
| 2066 |
+
"epoch": 2.4,
|
| 2067 |
+
"learning_rate": 4.2437143152855205e-05,
|
| 2068 |
+
"loss": 0.6499,
|
| 2069 |
+
"step": 170000
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"epoch": 2.4,
|
| 2073 |
+
"learning_rate": 4.241340477635126e-05,
|
| 2074 |
+
"loss": 0.6516,
|
| 2075 |
+
"step": 170500
|
| 2076 |
+
},
|
| 2077 |
+
{
|
| 2078 |
+
"epoch": 2.41,
|
| 2079 |
+
"learning_rate": 4.2389666399847316e-05,
|
| 2080 |
+
"loss": 0.6553,
|
| 2081 |
+
"step": 171000
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 2.42,
|
| 2085 |
+
"learning_rate": 4.236592802334337e-05,
|
| 2086 |
+
"loss": 0.668,
|
| 2087 |
+
"step": 171500
|
| 2088 |
+
},
|
| 2089 |
+
{
|
| 2090 |
+
"epoch": 2.43,
|
| 2091 |
+
"learning_rate": 4.2342189646839427e-05,
|
| 2092 |
+
"loss": 0.6447,
|
| 2093 |
+
"step": 172000
|
| 2094 |
+
},
|
| 2095 |
+
{
|
| 2096 |
+
"epoch": 2.43,
|
| 2097 |
+
"learning_rate": 4.231845127033548e-05,
|
| 2098 |
+
"loss": 0.6474,
|
| 2099 |
+
"step": 172500
|
| 2100 |
+
},
|
| 2101 |
+
{
|
| 2102 |
+
"epoch": 2.44,
|
| 2103 |
+
"learning_rate": 4.229471289383154e-05,
|
| 2104 |
+
"loss": 0.6594,
|
| 2105 |
+
"step": 173000
|
| 2106 |
+
},
|
| 2107 |
+
{
|
| 2108 |
+
"epoch": 2.45,
|
| 2109 |
+
"learning_rate": 4.227097451732759e-05,
|
| 2110 |
+
"loss": 0.6516,
|
| 2111 |
+
"step": 173500
|
| 2112 |
+
},
|
| 2113 |
+
{
|
| 2114 |
+
"epoch": 2.45,
|
| 2115 |
+
"learning_rate": 4.224723614082365e-05,
|
| 2116 |
+
"loss": 0.6404,
|
| 2117 |
+
"step": 174000
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 2.46,
|
| 2121 |
+
"learning_rate": 4.222354524107271e-05,
|
| 2122 |
+
"loss": 0.6526,
|
| 2123 |
+
"step": 174500
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 2.47,
|
| 2127 |
+
"learning_rate": 4.2199806864568764e-05,
|
| 2128 |
+
"loss": 0.6459,
|
| 2129 |
+
"step": 175000
|
| 2130 |
+
},
|
| 2131 |
+
{
|
| 2132 |
+
"epoch": 2.47,
|
| 2133 |
+
"learning_rate": 4.2176068488064826e-05,
|
| 2134 |
+
"loss": 0.6519,
|
| 2135 |
+
"step": 175500
|
| 2136 |
+
},
|
| 2137 |
+
{
|
| 2138 |
+
"epoch": 2.48,
|
| 2139 |
+
"learning_rate": 4.2152377588313886e-05,
|
| 2140 |
+
"loss": 0.6391,
|
| 2141 |
+
"step": 176000
|
| 2142 |
+
},
|
| 2143 |
+
{
|
| 2144 |
+
"epoch": 2.49,
|
| 2145 |
+
"learning_rate": 4.212863921180994e-05,
|
| 2146 |
+
"loss": 0.6346,
|
| 2147 |
+
"step": 176500
|
| 2148 |
+
},
|
| 2149 |
+
{
|
| 2150 |
+
"epoch": 2.5,
|
| 2151 |
+
"learning_rate": 4.2104900835306e-05,
|
| 2152 |
+
"loss": 0.6395,
|
| 2153 |
+
"step": 177000
|
| 2154 |
+
},
|
| 2155 |
+
{
|
| 2156 |
+
"epoch": 2.5,
|
| 2157 |
+
"learning_rate": 4.2081162458802046e-05,
|
| 2158 |
+
"loss": 0.6409,
|
| 2159 |
+
"step": 177500
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"epoch": 2.51,
|
| 2163 |
+
"learning_rate": 4.20574240822981e-05,
|
| 2164 |
+
"loss": 0.6285,
|
| 2165 |
+
"step": 178000
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 2.52,
|
| 2169 |
+
"learning_rate": 4.2033685705794163e-05,
|
| 2170 |
+
"loss": 0.6307,
|
| 2171 |
+
"step": 178500
|
| 2172 |
+
},
|
| 2173 |
+
{
|
| 2174 |
+
"epoch": 2.52,
|
| 2175 |
+
"learning_rate": 4.200999480604322e-05,
|
| 2176 |
+
"loss": 0.6506,
|
| 2177 |
+
"step": 179000
|
| 2178 |
+
},
|
| 2179 |
+
{
|
| 2180 |
+
"epoch": 2.53,
|
| 2181 |
+
"learning_rate": 4.198625642953928e-05,
|
| 2182 |
+
"loss": 0.6307,
|
| 2183 |
+
"step": 179500
|
| 2184 |
+
},
|
| 2185 |
+
{
|
| 2186 |
+
"epoch": 2.54,
|
| 2187 |
+
"learning_rate": 4.1962518053035334e-05,
|
| 2188 |
+
"loss": 0.6446,
|
| 2189 |
+
"step": 180000
|
| 2190 |
+
},
|
| 2191 |
+
{
|
| 2192 |
+
"epoch": 2.55,
|
| 2193 |
+
"learning_rate": 4.193877967653139e-05,
|
| 2194 |
+
"loss": 0.6422,
|
| 2195 |
+
"step": 180500
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"epoch": 2.55,
|
| 2199 |
+
"learning_rate": 4.1915088776780457e-05,
|
| 2200 |
+
"loss": 0.6398,
|
| 2201 |
+
"step": 181000
|
| 2202 |
+
},
|
| 2203 |
+
{
|
| 2204 |
+
"epoch": 2.56,
|
| 2205 |
+
"learning_rate": 4.1891397877029516e-05,
|
| 2206 |
+
"loss": 0.6333,
|
| 2207 |
+
"step": 181500
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 2.57,
|
| 2211 |
+
"learning_rate": 4.186765950052557e-05,
|
| 2212 |
+
"loss": 0.6363,
|
| 2213 |
+
"step": 182000
|
| 2214 |
+
},
|
| 2215 |
+
{
|
| 2216 |
+
"epoch": 2.57,
|
| 2217 |
+
"learning_rate": 4.184392112402162e-05,
|
| 2218 |
+
"loss": 0.6319,
|
| 2219 |
+
"step": 182500
|
| 2220 |
+
},
|
| 2221 |
+
{
|
| 2222 |
+
"epoch": 2.58,
|
| 2223 |
+
"learning_rate": 4.1820182747517676e-05,
|
| 2224 |
+
"loss": 0.6294,
|
| 2225 |
+
"step": 183000
|
| 2226 |
+
},
|
| 2227 |
+
{
|
| 2228 |
+
"epoch": 2.59,
|
| 2229 |
+
"learning_rate": 4.179649184776674e-05,
|
| 2230 |
+
"loss": 0.6366,
|
| 2231 |
+
"step": 183500
|
| 2232 |
+
},
|
| 2233 |
+
{
|
| 2234 |
+
"epoch": 2.59,
|
| 2235 |
+
"learning_rate": 4.17727534712628e-05,
|
| 2236 |
+
"loss": 0.6384,
|
| 2237 |
+
"step": 184000
|
| 2238 |
+
},
|
| 2239 |
+
{
|
| 2240 |
+
"epoch": 2.6,
|
| 2241 |
+
"learning_rate": 4.1749015094758854e-05,
|
| 2242 |
+
"loss": 0.6365,
|
| 2243 |
+
"step": 184500
|
| 2244 |
+
},
|
| 2245 |
+
{
|
| 2246 |
+
"epoch": 2.61,
|
| 2247 |
+
"learning_rate": 4.172527671825491e-05,
|
| 2248 |
+
"loss": 0.6331,
|
| 2249 |
+
"step": 185000
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 2.62,
|
| 2253 |
+
"learning_rate": 4.1701538341750965e-05,
|
| 2254 |
+
"loss": 0.6327,
|
| 2255 |
+
"step": 185500
|
| 2256 |
+
},
|
| 2257 |
+
{
|
| 2258 |
+
"epoch": 2.62,
|
| 2259 |
+
"learning_rate": 4.167779996524701e-05,
|
| 2260 |
+
"loss": 0.628,
|
| 2261 |
+
"step": 186000
|
| 2262 |
+
},
|
| 2263 |
+
{
|
| 2264 |
+
"epoch": 2.63,
|
| 2265 |
+
"learning_rate": 4.1654061588743076e-05,
|
| 2266 |
+
"loss": 0.6309,
|
| 2267 |
+
"step": 186500
|
| 2268 |
+
},
|
| 2269 |
+
{
|
| 2270 |
+
"epoch": 2.64,
|
| 2271 |
+
"learning_rate": 4.163032321223913e-05,
|
| 2272 |
+
"loss": 0.6159,
|
| 2273 |
+
"step": 187000
|
| 2274 |
+
},
|
| 2275 |
+
{
|
| 2276 |
+
"epoch": 2.64,
|
| 2277 |
+
"learning_rate": 4.1606584835735187e-05,
|
| 2278 |
+
"loss": 0.6307,
|
| 2279 |
+
"step": 187500
|
| 2280 |
+
},
|
| 2281 |
+
{
|
| 2282 |
+
"epoch": 2.65,
|
| 2283 |
+
"learning_rate": 4.158284645923124e-05,
|
| 2284 |
+
"loss": 0.6356,
|
| 2285 |
+
"step": 188000
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"epoch": 2.66,
|
| 2289 |
+
"learning_rate": 4.155910808272729e-05,
|
| 2290 |
+
"loss": 0.6269,
|
| 2291 |
+
"step": 188500
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 2.66,
|
| 2295 |
+
"learning_rate": 4.1535369706223346e-05,
|
| 2296 |
+
"loss": 0.6368,
|
| 2297 |
+
"step": 189000
|
| 2298 |
+
},
|
| 2299 |
+
{
|
| 2300 |
+
"epoch": 2.67,
|
| 2301 |
+
"learning_rate": 4.15116313297194e-05,
|
| 2302 |
+
"loss": 0.6282,
|
| 2303 |
+
"step": 189500
|
| 2304 |
+
},
|
| 2305 |
+
{
|
| 2306 |
+
"epoch": 2.68,
|
| 2307 |
+
"learning_rate": 4.1487892953215464e-05,
|
| 2308 |
+
"loss": 0.6226,
|
| 2309 |
+
"step": 190000
|
| 2310 |
+
},
|
| 2311 |
+
{
|
| 2312 |
+
"epoch": 2.69,
|
| 2313 |
+
"learning_rate": 4.146415457671152e-05,
|
| 2314 |
+
"loss": 0.6181,
|
| 2315 |
+
"step": 190500
|
| 2316 |
+
},
|
| 2317 |
+
{
|
| 2318 |
+
"epoch": 2.69,
|
| 2319 |
+
"learning_rate": 4.144041620020757e-05,
|
| 2320 |
+
"loss": 0.6138,
|
| 2321 |
+
"step": 191000
|
| 2322 |
+
},
|
| 2323 |
+
{
|
| 2324 |
+
"epoch": 2.7,
|
| 2325 |
+
"learning_rate": 4.1416677823703623e-05,
|
| 2326 |
+
"loss": 0.6184,
|
| 2327 |
+
"step": 191500
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 2.71,
|
| 2331 |
+
"learning_rate": 4.139298692395269e-05,
|
| 2332 |
+
"loss": 0.624,
|
| 2333 |
+
"step": 192000
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 2.71,
|
| 2337 |
+
"learning_rate": 4.136924854744874e-05,
|
| 2338 |
+
"loss": 0.6125,
|
| 2339 |
+
"step": 192500
|
| 2340 |
+
},
|
| 2341 |
+
{
|
| 2342 |
+
"epoch": 2.72,
|
| 2343 |
+
"learning_rate": 4.13455101709448e-05,
|
| 2344 |
+
"loss": 0.6261,
|
| 2345 |
+
"step": 193000
|
| 2346 |
+
},
|
| 2347 |
+
{
|
| 2348 |
+
"epoch": 2.73,
|
| 2349 |
+
"learning_rate": 4.1321771794440857e-05,
|
| 2350 |
+
"loss": 0.6189,
|
| 2351 |
+
"step": 193500
|
| 2352 |
+
},
|
| 2353 |
+
{
|
| 2354 |
+
"epoch": 2.74,
|
| 2355 |
+
"learning_rate": 4.129803341793691e-05,
|
| 2356 |
+
"loss": 0.6149,
|
| 2357 |
+
"step": 194000
|
| 2358 |
+
},
|
| 2359 |
+
{
|
| 2360 |
+
"epoch": 2.74,
|
| 2361 |
+
"learning_rate": 4.127429504143297e-05,
|
| 2362 |
+
"loss": 0.6296,
|
| 2363 |
+
"step": 194500
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 2.75,
|
| 2367 |
+
"learning_rate": 4.1250556664929016e-05,
|
| 2368 |
+
"loss": 0.6263,
|
| 2369 |
+
"step": 195000
|
| 2370 |
+
},
|
| 2371 |
+
{
|
| 2372 |
+
"epoch": 2.76,
|
| 2373 |
+
"learning_rate": 4.122681828842507e-05,
|
| 2374 |
+
"loss": 0.6298,
|
| 2375 |
+
"step": 195500
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 2.76,
|
| 2379 |
+
"learning_rate": 4.120312738867414e-05,
|
| 2380 |
+
"loss": 0.6102,
|
| 2381 |
+
"step": 196000
|
| 2382 |
+
},
|
| 2383 |
+
{
|
| 2384 |
+
"epoch": 2.77,
|
| 2385 |
+
"learning_rate": 4.1179389012170194e-05,
|
| 2386 |
+
"loss": 0.6248,
|
| 2387 |
+
"step": 196500
|
| 2388 |
+
},
|
| 2389 |
+
{
|
| 2390 |
+
"epoch": 2.78,
|
| 2391 |
+
"learning_rate": 4.115565063566625e-05,
|
| 2392 |
+
"loss": 0.6121,
|
| 2393 |
+
"step": 197000
|
| 2394 |
+
},
|
| 2395 |
+
{
|
| 2396 |
+
"epoch": 2.78,
|
| 2397 |
+
"learning_rate": 4.1131912259162305e-05,
|
| 2398 |
+
"loss": 0.6156,
|
| 2399 |
+
"step": 197500
|
| 2400 |
+
},
|
| 2401 |
+
{
|
| 2402 |
+
"epoch": 2.79,
|
| 2403 |
+
"learning_rate": 4.1108221359411365e-05,
|
| 2404 |
+
"loss": 0.6076,
|
| 2405 |
+
"step": 198000
|
| 2406 |
+
},
|
| 2407 |
+
{
|
| 2408 |
+
"epoch": 2.8,
|
| 2409 |
+
"learning_rate": 4.1084577936413436e-05,
|
| 2410 |
+
"loss": 0.6231,
|
| 2411 |
+
"step": 198500
|
| 2412 |
+
},
|
| 2413 |
+
{
|
| 2414 |
+
"epoch": 2.81,
|
| 2415 |
+
"learning_rate": 4.106083955990949e-05,
|
| 2416 |
+
"loss": 0.6281,
|
| 2417 |
+
"step": 199000
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 2.81,
|
| 2421 |
+
"learning_rate": 4.103710118340555e-05,
|
| 2422 |
+
"loss": 0.6157,
|
| 2423 |
+
"step": 199500
|
| 2424 |
+
},
|
| 2425 |
+
{
|
| 2426 |
+
"epoch": 2.82,
|
| 2427 |
+
"learning_rate": 4.10133628069016e-05,
|
| 2428 |
+
"loss": 0.6086,
|
| 2429 |
+
"step": 200000
|
| 2430 |
+
},
|
| 2431 |
+
{
|
| 2432 |
+
"epoch": 2.83,
|
| 2433 |
+
"learning_rate": 4.098962443039766e-05,
|
| 2434 |
+
"loss": 0.6225,
|
| 2435 |
+
"step": 200500
|
| 2436 |
+
},
|
| 2437 |
+
{
|
| 2438 |
+
"epoch": 2.83,
|
| 2439 |
+
"learning_rate": 4.096588605389371e-05,
|
| 2440 |
+
"loss": 0.6078,
|
| 2441 |
+
"step": 201000
|
| 2442 |
+
},
|
| 2443 |
+
{
|
| 2444 |
+
"epoch": 2.84,
|
| 2445 |
+
"learning_rate": 4.094214767738977e-05,
|
| 2446 |
+
"loss": 0.6217,
|
| 2447 |
+
"step": 201500
|
| 2448 |
+
},
|
| 2449 |
+
{
|
| 2450 |
+
"epoch": 2.85,
|
| 2451 |
+
"learning_rate": 4.0918409300885824e-05,
|
| 2452 |
+
"loss": 0.6096,
|
| 2453 |
+
"step": 202000
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"epoch": 2.86,
|
| 2457 |
+
"learning_rate": 4.089467092438188e-05,
|
| 2458 |
+
"loss": 0.6253,
|
| 2459 |
+
"step": 202500
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 2.86,
|
| 2463 |
+
"learning_rate": 4.087102750138395e-05,
|
| 2464 |
+
"loss": 0.6068,
|
| 2465 |
+
"step": 203000
|
| 2466 |
+
},
|
| 2467 |
+
{
|
| 2468 |
+
"epoch": 2.87,
|
| 2469 |
+
"learning_rate": 4.0847289124880006e-05,
|
| 2470 |
+
"loss": 0.613,
|
| 2471 |
+
"step": 203500
|
| 2472 |
+
},
|
| 2473 |
+
{
|
| 2474 |
+
"epoch": 2.88,
|
| 2475 |
+
"learning_rate": 4.082355074837606e-05,
|
| 2476 |
+
"loss": 0.6179,
|
| 2477 |
+
"step": 204000
|
| 2478 |
+
},
|
| 2479 |
+
{
|
| 2480 |
+
"epoch": 2.88,
|
| 2481 |
+
"learning_rate": 4.079981237187212e-05,
|
| 2482 |
+
"loss": 0.6312,
|
| 2483 |
+
"step": 204500
|
| 2484 |
+
},
|
| 2485 |
+
{
|
| 2486 |
+
"epoch": 2.89,
|
| 2487 |
+
"learning_rate": 4.077612147212118e-05,
|
| 2488 |
+
"loss": 0.601,
|
| 2489 |
+
"step": 205000
|
| 2490 |
+
},
|
| 2491 |
+
{
|
| 2492 |
+
"epoch": 2.9,
|
| 2493 |
+
"learning_rate": 4.075238309561723e-05,
|
| 2494 |
+
"loss": 0.615,
|
| 2495 |
+
"step": 205500
|
| 2496 |
+
},
|
| 2497 |
+
{
|
| 2498 |
+
"epoch": 2.9,
|
| 2499 |
+
"learning_rate": 4.072864471911329e-05,
|
| 2500 |
+
"loss": 0.6071,
|
| 2501 |
+
"step": 206000
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 2.91,
|
| 2505 |
+
"learning_rate": 4.0704906342609344e-05,
|
| 2506 |
+
"loss": 0.6135,
|
| 2507 |
+
"step": 206500
|
| 2508 |
+
},
|
| 2509 |
+
{
|
| 2510 |
+
"epoch": 2.92,
|
| 2511 |
+
"learning_rate": 4.06811679661054e-05,
|
| 2512 |
+
"loss": 0.6073,
|
| 2513 |
+
"step": 207000
|
| 2514 |
+
},
|
| 2515 |
+
{
|
| 2516 |
+
"epoch": 2.93,
|
| 2517 |
+
"learning_rate": 4.065747706635446e-05,
|
| 2518 |
+
"loss": 0.6073,
|
| 2519 |
+
"step": 207500
|
| 2520 |
+
},
|
| 2521 |
+
{
|
| 2522 |
+
"epoch": 2.93,
|
| 2523 |
+
"learning_rate": 4.0633738689850515e-05,
|
| 2524 |
+
"loss": 0.6053,
|
| 2525 |
+
"step": 208000
|
| 2526 |
+
},
|
| 2527 |
+
{
|
| 2528 |
+
"epoch": 2.94,
|
| 2529 |
+
"learning_rate": 4.061000031334658e-05,
|
| 2530 |
+
"loss": 0.6116,
|
| 2531 |
+
"step": 208500
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 2.95,
|
| 2535 |
+
"learning_rate": 4.058626193684263e-05,
|
| 2536 |
+
"loss": 0.606,
|
| 2537 |
+
"step": 209000
|
| 2538 |
+
},
|
| 2539 |
+
{
|
| 2540 |
+
"epoch": 2.95,
|
| 2541 |
+
"learning_rate": 4.056252356033868e-05,
|
| 2542 |
+
"loss": 0.6183,
|
| 2543 |
+
"step": 209500
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 2.96,
|
| 2547 |
+
"learning_rate": 4.0538785183834736e-05,
|
| 2548 |
+
"loss": 0.601,
|
| 2549 |
+
"step": 210000
|
| 2550 |
+
},
|
| 2551 |
+
{
|
| 2552 |
+
"epoch": 2.97,
|
| 2553 |
+
"learning_rate": 4.051504680733079e-05,
|
| 2554 |
+
"loss": 0.6041,
|
| 2555 |
+
"step": 210500
|
| 2556 |
+
},
|
| 2557 |
+
{
|
| 2558 |
+
"epoch": 2.98,
|
| 2559 |
+
"learning_rate": 4.049130843082685e-05,
|
| 2560 |
+
"loss": 0.6073,
|
| 2561 |
+
"step": 211000
|
| 2562 |
+
},
|
| 2563 |
+
{
|
| 2564 |
+
"epoch": 2.98,
|
| 2565 |
+
"learning_rate": 4.04675700543229e-05,
|
| 2566 |
+
"loss": 0.6106,
|
| 2567 |
+
"step": 211500
|
| 2568 |
+
},
|
| 2569 |
+
{
|
| 2570 |
+
"epoch": 2.99,
|
| 2571 |
+
"learning_rate": 4.044383167781896e-05,
|
| 2572 |
+
"loss": 0.6105,
|
| 2573 |
+
"step": 212000
|
| 2574 |
+
},
|
| 2575 |
+
{
|
| 2576 |
+
"epoch": 3.0,
|
| 2577 |
+
"learning_rate": 4.0420093301315014e-05,
|
| 2578 |
+
"loss": 0.5973,
|
| 2579 |
+
"step": 212500
|
| 2580 |
+
},
|
| 2581 |
+
{
|
| 2582 |
+
"epoch": 3.0,
|
| 2583 |
+
"eval_bleu": 43.4697,
|
| 2584 |
+
"eval_gen_len": 13.6556,
|
| 2585 |
+
"eval_loss": 0.9142627120018005,
|
| 2586 |
+
"eval_runtime": 9819.0968,
|
| 2587 |
+
"eval_samples_per_second": 14.445,
|
| 2588 |
+
"eval_steps_per_second": 1.806,
|
| 2589 |
+
"step": 212758
|
| 2590 |
+
},
|
| 2591 |
+
{
|
| 2592 |
+
"epoch": 3.0,
|
| 2593 |
+
"learning_rate": 4.039635492481107e-05,
|
| 2594 |
+
"loss": 0.6014,
|
| 2595 |
+
"step": 213000
|
| 2596 |
+
},
|
| 2597 |
+
{
|
| 2598 |
+
"epoch": 3.01,
|
| 2599 |
+
"learning_rate": 4.037266402506013e-05,
|
| 2600 |
+
"loss": 0.6099,
|
| 2601 |
+
"step": 213500
|
| 2602 |
+
},
|
| 2603 |
+
{
|
| 2604 |
+
"epoch": 3.02,
|
| 2605 |
+
"learning_rate": 4.0348925648556185e-05,
|
| 2606 |
+
"loss": 0.5951,
|
| 2607 |
+
"step": 214000
|
| 2608 |
+
},
|
| 2609 |
+
{
|
| 2610 |
+
"epoch": 3.02,
|
| 2611 |
+
"learning_rate": 4.032518727205224e-05,
|
| 2612 |
+
"loss": 0.597,
|
| 2613 |
+
"step": 214500
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 3.03,
|
| 2617 |
+
"learning_rate": 4.03014488955483e-05,
|
| 2618 |
+
"loss": 0.596,
|
| 2619 |
+
"step": 215000
|
| 2620 |
+
},
|
| 2621 |
+
{
|
| 2622 |
+
"epoch": 3.04,
|
| 2623 |
+
"learning_rate": 4.027771051904435e-05,
|
| 2624 |
+
"loss": 0.5914,
|
| 2625 |
+
"step": 215500
|
| 2626 |
+
},
|
| 2627 |
+
{
|
| 2628 |
+
"epoch": 3.05,
|
| 2629 |
+
"learning_rate": 4.0253972142540406e-05,
|
| 2630 |
+
"loss": 0.5979,
|
| 2631 |
+
"step": 216000
|
| 2632 |
+
},
|
| 2633 |
+
{
|
| 2634 |
+
"epoch": 3.05,
|
| 2635 |
+
"learning_rate": 4.023023376603646e-05,
|
| 2636 |
+
"loss": 0.5855,
|
| 2637 |
+
"step": 216500
|
| 2638 |
+
},
|
| 2639 |
+
{
|
| 2640 |
+
"epoch": 3.06,
|
| 2641 |
+
"learning_rate": 4.020649538953252e-05,
|
| 2642 |
+
"loss": 0.5887,
|
| 2643 |
+
"step": 217000
|
| 2644 |
+
},
|
| 2645 |
+
{
|
| 2646 |
+
"epoch": 3.07,
|
| 2647 |
+
"learning_rate": 4.018275701302857e-05,
|
| 2648 |
+
"loss": 0.5853,
|
| 2649 |
+
"step": 217500
|
| 2650 |
+
},
|
| 2651 |
+
{
|
| 2652 |
+
"epoch": 3.07,
|
| 2653 |
+
"learning_rate": 4.015901863652463e-05,
|
| 2654 |
+
"loss": 0.5921,
|
| 2655 |
+
"step": 218000
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 3.08,
|
| 2659 |
+
"learning_rate": 4.013528026002068e-05,
|
| 2660 |
+
"loss": 0.5873,
|
| 2661 |
+
"step": 218500
|
| 2662 |
+
},
|
| 2663 |
+
{
|
| 2664 |
+
"epoch": 3.09,
|
| 2665 |
+
"learning_rate": 4.0111589360269744e-05,
|
| 2666 |
+
"loss": 0.5708,
|
| 2667 |
+
"step": 219000
|
| 2668 |
+
},
|
| 2669 |
+
{
|
| 2670 |
+
"epoch": 3.1,
|
| 2671 |
+
"learning_rate": 4.00878509837658e-05,
|
| 2672 |
+
"loss": 0.5777,
|
| 2673 |
+
"step": 219500
|
| 2674 |
+
},
|
| 2675 |
+
{
|
| 2676 |
+
"epoch": 3.1,
|
| 2677 |
+
"learning_rate": 4.0064112607261855e-05,
|
| 2678 |
+
"loss": 0.5772,
|
| 2679 |
+
"step": 220000
|
| 2680 |
+
},
|
| 2681 |
+
{
|
| 2682 |
+
"epoch": 3.11,
|
| 2683 |
+
"learning_rate": 4.004037423075791e-05,
|
| 2684 |
+
"loss": 0.573,
|
| 2685 |
+
"step": 220500
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 3.12,
|
| 2689 |
+
"learning_rate": 4.0016635854253966e-05,
|
| 2690 |
+
"loss": 0.5734,
|
| 2691 |
+
"step": 221000
|
| 2692 |
+
},
|
| 2693 |
+
{
|
| 2694 |
+
"epoch": 3.12,
|
| 2695 |
+
"learning_rate": 3.999289747775002e-05,
|
| 2696 |
+
"loss": 0.5682,
|
| 2697 |
+
"step": 221500
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 3.13,
|
| 2701 |
+
"learning_rate": 3.996920657799909e-05,
|
| 2702 |
+
"loss": 0.5628,
|
| 2703 |
+
"step": 222000
|
| 2704 |
+
},
|
| 2705 |
+
{
|
| 2706 |
+
"epoch": 3.14,
|
| 2707 |
+
"learning_rate": 3.9945468201495136e-05,
|
| 2708 |
+
"loss": 0.5687,
|
| 2709 |
+
"step": 222500
|
| 2710 |
+
},
|
| 2711 |
+
{
|
| 2712 |
+
"epoch": 3.14,
|
| 2713 |
+
"learning_rate": 3.992172982499119e-05,
|
| 2714 |
+
"loss": 0.5569,
|
| 2715 |
+
"step": 223000
|
| 2716 |
+
},
|
| 2717 |
+
{
|
| 2718 |
+
"epoch": 3.15,
|
| 2719 |
+
"learning_rate": 3.989799144848725e-05,
|
| 2720 |
+
"loss": 0.5571,
|
| 2721 |
+
"step": 223500
|
| 2722 |
+
},
|
| 2723 |
+
{
|
| 2724 |
+
"epoch": 3.16,
|
| 2725 |
+
"learning_rate": 3.98742530719833e-05,
|
| 2726 |
+
"loss": 0.5619,
|
| 2727 |
+
"step": 224000
|
| 2728 |
+
},
|
| 2729 |
+
{
|
| 2730 |
+
"epoch": 3.17,
|
| 2731 |
+
"learning_rate": 3.9850514695479365e-05,
|
| 2732 |
+
"loss": 0.5635,
|
| 2733 |
+
"step": 224500
|
| 2734 |
+
},
|
| 2735 |
+
{
|
| 2736 |
+
"epoch": 3.17,
|
| 2737 |
+
"learning_rate": 3.982677631897542e-05,
|
| 2738 |
+
"loss": 0.5538,
|
| 2739 |
+
"step": 225000
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 3.18,
|
| 2743 |
+
"learning_rate": 3.980303794247147e-05,
|
| 2744 |
+
"loss": 0.5494,
|
| 2745 |
+
"step": 225500
|
| 2746 |
+
},
|
| 2747 |
+
{
|
| 2748 |
+
"epoch": 3.19,
|
| 2749 |
+
"learning_rate": 3.9779299565967525e-05,
|
| 2750 |
+
"loss": 0.5532,
|
| 2751 |
+
"step": 226000
|
| 2752 |
+
},
|
| 2753 |
+
{
|
| 2754 |
+
"epoch": 3.19,
|
| 2755 |
+
"learning_rate": 3.975556118946358e-05,
|
| 2756 |
+
"loss": 0.5576,
|
| 2757 |
+
"step": 226500
|
| 2758 |
+
},
|
| 2759 |
+
{
|
| 2760 |
+
"epoch": 3.2,
|
| 2761 |
+
"learning_rate": 3.973187028971264e-05,
|
| 2762 |
+
"loss": 0.5499,
|
| 2763 |
+
"step": 227000
|
| 2764 |
+
},
|
| 2765 |
+
{
|
| 2766 |
+
"epoch": 3.21,
|
| 2767 |
+
"learning_rate": 3.970822686671472e-05,
|
| 2768 |
+
"loss": 0.5434,
|
| 2769 |
+
"step": 227500
|
| 2770 |
+
},
|
| 2771 |
+
{
|
| 2772 |
+
"epoch": 3.21,
|
| 2773 |
+
"learning_rate": 3.968448849021077e-05,
|
| 2774 |
+
"loss": 0.5421,
|
| 2775 |
+
"step": 228000
|
| 2776 |
+
},
|
| 2777 |
+
{
|
| 2778 |
+
"epoch": 3.22,
|
| 2779 |
+
"learning_rate": 3.966075011370682e-05,
|
| 2780 |
+
"loss": 0.5377,
|
| 2781 |
+
"step": 228500
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 3.23,
|
| 2785 |
+
"learning_rate": 3.963701173720288e-05,
|
| 2786 |
+
"loss": 0.5614,
|
| 2787 |
+
"step": 229000
|
| 2788 |
+
},
|
| 2789 |
+
{
|
| 2790 |
+
"epoch": 3.24,
|
| 2791 |
+
"learning_rate": 3.961327336069894e-05,
|
| 2792 |
+
"loss": 0.5458,
|
| 2793 |
+
"step": 229500
|
| 2794 |
+
},
|
| 2795 |
+
{
|
| 2796 |
+
"epoch": 3.24,
|
| 2797 |
+
"learning_rate": 3.9589534984194995e-05,
|
| 2798 |
+
"loss": 0.5483,
|
| 2799 |
+
"step": 230000
|
| 2800 |
+
},
|
| 2801 |
+
{
|
| 2802 |
+
"epoch": 3.25,
|
| 2803 |
+
"learning_rate": 3.9565796607691044e-05,
|
| 2804 |
+
"loss": 0.5423,
|
| 2805 |
+
"step": 230500
|
| 2806 |
+
},
|
| 2807 |
+
{
|
| 2808 |
+
"epoch": 3.26,
|
| 2809 |
+
"learning_rate": 3.95420582311871e-05,
|
| 2810 |
+
"loss": 0.5346,
|
| 2811 |
+
"step": 231000
|
| 2812 |
+
},
|
| 2813 |
+
{
|
| 2814 |
+
"epoch": 3.26,
|
| 2815 |
+
"learning_rate": 3.9518319854683155e-05,
|
| 2816 |
+
"loss": 0.5334,
|
| 2817 |
+
"step": 231500
|
| 2818 |
+
},
|
| 2819 |
+
{
|
| 2820 |
+
"epoch": 3.27,
|
| 2821 |
+
"learning_rate": 3.9494628954932215e-05,
|
| 2822 |
+
"loss": 0.5451,
|
| 2823 |
+
"step": 232000
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 3.28,
|
| 2827 |
+
"learning_rate": 3.947089057842828e-05,
|
| 2828 |
+
"loss": 0.5378,
|
| 2829 |
+
"step": 232500
|
| 2830 |
+
},
|
| 2831 |
+
{
|
| 2832 |
+
"epoch": 3.29,
|
| 2833 |
+
"learning_rate": 3.944715220192433e-05,
|
| 2834 |
+
"loss": 0.5323,
|
| 2835 |
+
"step": 233000
|
| 2836 |
+
},
|
| 2837 |
+
{
|
| 2838 |
+
"epoch": 3.29,
|
| 2839 |
+
"learning_rate": 3.942346130217339e-05,
|
| 2840 |
+
"loss": 0.5346,
|
| 2841 |
+
"step": 233500
|
| 2842 |
+
},
|
| 2843 |
+
{
|
| 2844 |
+
"epoch": 3.3,
|
| 2845 |
+
"learning_rate": 3.939972292566945e-05,
|
| 2846 |
+
"loss": 0.5334,
|
| 2847 |
+
"step": 234000
|
| 2848 |
+
},
|
| 2849 |
+
{
|
| 2850 |
+
"epoch": 3.31,
|
| 2851 |
+
"learning_rate": 3.9375984549165504e-05,
|
| 2852 |
+
"loss": 0.5292,
|
| 2853 |
+
"step": 234500
|
| 2854 |
+
},
|
| 2855 |
+
{
|
| 2856 |
+
"epoch": 3.31,
|
| 2857 |
+
"learning_rate": 3.935224617266155e-05,
|
| 2858 |
+
"loss": 0.5442,
|
| 2859 |
+
"step": 235000
|
| 2860 |
+
},
|
| 2861 |
+
{
|
| 2862 |
+
"epoch": 3.32,
|
| 2863 |
+
"learning_rate": 3.9328507796157615e-05,
|
| 2864 |
+
"loss": 0.5423,
|
| 2865 |
+
"step": 235500
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 3.33,
|
| 2869 |
+
"learning_rate": 3.930476941965367e-05,
|
| 2870 |
+
"loss": 0.53,
|
| 2871 |
+
"step": 236000
|
| 2872 |
+
},
|
| 2873 |
+
{
|
| 2874 |
+
"epoch": 3.33,
|
| 2875 |
+
"learning_rate": 3.9281031043149725e-05,
|
| 2876 |
+
"loss": 0.5359,
|
| 2877 |
+
"step": 236500
|
| 2878 |
+
},
|
| 2879 |
+
{
|
| 2880 |
+
"epoch": 3.34,
|
| 2881 |
+
"learning_rate": 3.9257340143398785e-05,
|
| 2882 |
+
"loss": 0.5432,
|
| 2883 |
+
"step": 237000
|
| 2884 |
+
},
|
| 2885 |
+
{
|
| 2886 |
+
"epoch": 3.35,
|
| 2887 |
+
"learning_rate": 3.923360176689484e-05,
|
| 2888 |
+
"loss": 0.5325,
|
| 2889 |
+
"step": 237500
|
| 2890 |
+
},
|
| 2891 |
+
{
|
| 2892 |
+
"epoch": 3.36,
|
| 2893 |
+
"learning_rate": 3.9209863390390896e-05,
|
| 2894 |
+
"loss": 0.53,
|
| 2895 |
+
"step": 238000
|
| 2896 |
+
},
|
| 2897 |
+
{
|
| 2898 |
+
"epoch": 3.36,
|
| 2899 |
+
"learning_rate": 3.918612501388695e-05,
|
| 2900 |
+
"loss": 0.5347,
|
| 2901 |
+
"step": 238500
|
| 2902 |
+
},
|
| 2903 |
+
{
|
| 2904 |
+
"epoch": 3.37,
|
| 2905 |
+
"learning_rate": 3.916238663738301e-05,
|
| 2906 |
+
"loss": 0.5268,
|
| 2907 |
+
"step": 239000
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 3.38,
|
| 2911 |
+
"learning_rate": 3.913864826087906e-05,
|
| 2912 |
+
"loss": 0.5233,
|
| 2913 |
+
"step": 239500
|
| 2914 |
+
},
|
| 2915 |
+
{
|
| 2916 |
+
"epoch": 3.38,
|
| 2917 |
+
"learning_rate": 3.911490988437512e-05,
|
| 2918 |
+
"loss": 0.5281,
|
| 2919 |
+
"step": 240000
|
| 2920 |
+
},
|
| 2921 |
+
{
|
| 2922 |
+
"epoch": 3.39,
|
| 2923 |
+
"learning_rate": 3.9091171507871174e-05,
|
| 2924 |
+
"loss": 0.5191,
|
| 2925 |
+
"step": 240500
|
| 2926 |
+
},
|
| 2927 |
+
{
|
| 2928 |
+
"epoch": 3.4,
|
| 2929 |
+
"learning_rate": 3.906743313136723e-05,
|
| 2930 |
+
"loss": 0.5226,
|
| 2931 |
+
"step": 241000
|
| 2932 |
+
},
|
| 2933 |
+
{
|
| 2934 |
+
"epoch": 3.41,
|
| 2935 |
+
"learning_rate": 3.904369475486328e-05,
|
| 2936 |
+
"loss": 0.525,
|
| 2937 |
+
"step": 241500
|
| 2938 |
+
},
|
| 2939 |
+
{
|
| 2940 |
+
"epoch": 3.41,
|
| 2941 |
+
"learning_rate": 3.901995637835934e-05,
|
| 2942 |
+
"loss": 0.5252,
|
| 2943 |
+
"step": 242000
|
| 2944 |
+
},
|
| 2945 |
+
{
|
| 2946 |
+
"epoch": 3.42,
|
| 2947 |
+
"learning_rate": 3.8996218001855396e-05,
|
| 2948 |
+
"loss": 0.5406,
|
| 2949 |
+
"step": 242500
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 3.43,
|
| 2953 |
+
"learning_rate": 3.897257457885746e-05,
|
| 2954 |
+
"loss": 0.5168,
|
| 2955 |
+
"step": 243000
|
| 2956 |
+
},
|
| 2957 |
+
{
|
| 2958 |
+
"epoch": 3.43,
|
| 2959 |
+
"learning_rate": 3.8948836202353515e-05,
|
| 2960 |
+
"loss": 0.5218,
|
| 2961 |
+
"step": 243500
|
| 2962 |
+
},
|
| 2963 |
+
{
|
| 2964 |
+
"epoch": 3.44,
|
| 2965 |
+
"learning_rate": 3.892509782584958e-05,
|
| 2966 |
+
"loss": 0.5304,
|
| 2967 |
+
"step": 244000
|
| 2968 |
+
},
|
| 2969 |
+
{
|
| 2970 |
+
"epoch": 3.45,
|
| 2971 |
+
"learning_rate": 3.890135944934563e-05,
|
| 2972 |
+
"loss": 0.5217,
|
| 2973 |
+
"step": 244500
|
| 2974 |
+
},
|
| 2975 |
+
{
|
| 2976 |
+
"epoch": 3.45,
|
| 2977 |
+
"learning_rate": 3.887762107284169e-05,
|
| 2978 |
+
"loss": 0.5143,
|
| 2979 |
+
"step": 245000
|
| 2980 |
+
},
|
| 2981 |
+
{
|
| 2982 |
+
"epoch": 3.46,
|
| 2983 |
+
"learning_rate": 3.885388269633774e-05,
|
| 2984 |
+
"loss": 0.5326,
|
| 2985 |
+
"step": 245500
|
| 2986 |
+
},
|
| 2987 |
+
{
|
| 2988 |
+
"epoch": 3.47,
|
| 2989 |
+
"learning_rate": 3.883014431983379e-05,
|
| 2990 |
+
"loss": 0.5152,
|
| 2991 |
+
"step": 246000
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 3.48,
|
| 2995 |
+
"learning_rate": 3.880640594332985e-05,
|
| 2996 |
+
"loss": 0.5288,
|
| 2997 |
+
"step": 246500
|
| 2998 |
+
},
|
| 2999 |
+
{
|
| 3000 |
+
"epoch": 3.48,
|
| 3001 |
+
"learning_rate": 3.8782667566825904e-05,
|
| 3002 |
+
"loss": 0.517,
|
| 3003 |
+
"step": 247000
|
| 3004 |
+
},
|
| 3005 |
+
{
|
| 3006 |
+
"epoch": 3.49,
|
| 3007 |
+
"learning_rate": 3.8758929190321966e-05,
|
| 3008 |
+
"loss": 0.5092,
|
| 3009 |
+
"step": 247500
|
| 3010 |
+
},
|
| 3011 |
+
{
|
| 3012 |
+
"epoch": 3.5,
|
| 3013 |
+
"learning_rate": 3.8735238290571026e-05,
|
| 3014 |
+
"loss": 0.5147,
|
| 3015 |
+
"step": 248000
|
| 3016 |
+
},
|
| 3017 |
+
{
|
| 3018 |
+
"epoch": 3.5,
|
| 3019 |
+
"learning_rate": 3.871149991406708e-05,
|
| 3020 |
+
"loss": 0.5175,
|
| 3021 |
+
"step": 248500
|
| 3022 |
+
},
|
| 3023 |
+
{
|
| 3024 |
+
"epoch": 3.51,
|
| 3025 |
+
"learning_rate": 3.868780901431614e-05,
|
| 3026 |
+
"loss": 0.4983,
|
| 3027 |
+
"step": 249000
|
| 3028 |
+
},
|
| 3029 |
+
{
|
| 3030 |
+
"epoch": 3.52,
|
| 3031 |
+
"learning_rate": 3.866411811456521e-05,
|
| 3032 |
+
"loss": 0.5145,
|
| 3033 |
+
"step": 249500
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 3.53,
|
| 3037 |
+
"learning_rate": 3.8640379738061264e-05,
|
| 3038 |
+
"loss": 0.5226,
|
| 3039 |
+
"step": 250000
|
| 3040 |
+
},
|
| 3041 |
+
{
|
| 3042 |
+
"epoch": 3.53,
|
| 3043 |
+
"learning_rate": 3.861664136155731e-05,
|
| 3044 |
+
"loss": 0.5148,
|
| 3045 |
+
"step": 250500
|
| 3046 |
+
},
|
| 3047 |
+
{
|
| 3048 |
+
"epoch": 3.54,
|
| 3049 |
+
"learning_rate": 3.859290298505337e-05,
|
| 3050 |
+
"loss": 0.5172,
|
| 3051 |
+
"step": 251000
|
| 3052 |
+
},
|
| 3053 |
+
{
|
| 3054 |
+
"epoch": 3.55,
|
| 3055 |
+
"learning_rate": 3.856916460854942e-05,
|
| 3056 |
+
"loss": 0.5245,
|
| 3057 |
+
"step": 251500
|
| 3058 |
+
},
|
| 3059 |
+
{
|
| 3060 |
+
"epoch": 3.55,
|
| 3061 |
+
"learning_rate": 3.854547370879849e-05,
|
| 3062 |
+
"loss": 0.5164,
|
| 3063 |
+
"step": 252000
|
| 3064 |
+
},
|
| 3065 |
+
{
|
| 3066 |
+
"epoch": 3.56,
|
| 3067 |
+
"learning_rate": 3.8521735332294545e-05,
|
| 3068 |
+
"loss": 0.5087,
|
| 3069 |
+
"step": 252500
|
| 3070 |
+
},
|
| 3071 |
+
{
|
| 3072 |
+
"epoch": 3.57,
|
| 3073 |
+
"learning_rate": 3.84979969557906e-05,
|
| 3074 |
+
"loss": 0.517,
|
| 3075 |
+
"step": 253000
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 3.57,
|
| 3079 |
+
"learning_rate": 3.8474258579286656e-05,
|
| 3080 |
+
"loss": 0.5089,
|
| 3081 |
+
"step": 253500
|
| 3082 |
+
},
|
| 3083 |
+
{
|
| 3084 |
+
"epoch": 3.58,
|
| 3085 |
+
"learning_rate": 3.8450520202782705e-05,
|
| 3086 |
+
"loss": 0.5045,
|
| 3087 |
+
"step": 254000
|
| 3088 |
+
},
|
| 3089 |
+
{
|
| 3090 |
+
"epoch": 3.59,
|
| 3091 |
+
"learning_rate": 3.842682930303177e-05,
|
| 3092 |
+
"loss": 0.5169,
|
| 3093 |
+
"step": 254500
|
| 3094 |
+
},
|
| 3095 |
+
{
|
| 3096 |
+
"epoch": 3.6,
|
| 3097 |
+
"learning_rate": 3.840309092652783e-05,
|
| 3098 |
+
"loss": 0.5204,
|
| 3099 |
+
"step": 255000
|
| 3100 |
+
},
|
| 3101 |
+
{
|
| 3102 |
+
"epoch": 3.6,
|
| 3103 |
+
"learning_rate": 3.837935255002388e-05,
|
| 3104 |
+
"loss": 0.5064,
|
| 3105 |
+
"step": 255500
|
| 3106 |
+
},
|
| 3107 |
+
{
|
| 3108 |
+
"epoch": 3.61,
|
| 3109 |
+
"learning_rate": 3.835561417351994e-05,
|
| 3110 |
+
"loss": 0.5074,
|
| 3111 |
+
"step": 256000
|
| 3112 |
+
},
|
| 3113 |
+
{
|
| 3114 |
+
"epoch": 3.62,
|
| 3115 |
+
"learning_rate": 3.8331875797015994e-05,
|
| 3116 |
+
"loss": 0.5147,
|
| 3117 |
+
"step": 256500
|
| 3118 |
+
},
|
| 3119 |
+
{
|
| 3120 |
+
"epoch": 3.62,
|
| 3121 |
+
"learning_rate": 3.830813742051205e-05,
|
| 3122 |
+
"loss": 0.5031,
|
| 3123 |
+
"step": 257000
|
| 3124 |
+
},
|
| 3125 |
+
{
|
| 3126 |
+
"epoch": 3.63,
|
| 3127 |
+
"learning_rate": 3.8284399044008104e-05,
|
| 3128 |
+
"loss": 0.5046,
|
| 3129 |
+
"step": 257500
|
| 3130 |
+
},
|
| 3131 |
+
{
|
| 3132 |
+
"epoch": 3.64,
|
| 3133 |
+
"learning_rate": 3.826066066750415e-05,
|
| 3134 |
+
"loss": 0.5023,
|
| 3135 |
+
"step": 258000
|
| 3136 |
+
},
|
| 3137 |
+
{
|
| 3138 |
+
"epoch": 3.64,
|
| 3139 |
+
"learning_rate": 3.8236922291000215e-05,
|
| 3140 |
+
"loss": 0.5053,
|
| 3141 |
+
"step": 258500
|
| 3142 |
+
},
|
| 3143 |
+
{
|
| 3144 |
+
"epoch": 3.65,
|
| 3145 |
+
"learning_rate": 3.821318391449627e-05,
|
| 3146 |
+
"loss": 0.5145,
|
| 3147 |
+
"step": 259000
|
| 3148 |
+
},
|
| 3149 |
+
{
|
| 3150 |
+
"epoch": 3.66,
|
| 3151 |
+
"learning_rate": 3.8189445537992326e-05,
|
| 3152 |
+
"loss": 0.5037,
|
| 3153 |
+
"step": 259500
|
| 3154 |
+
},
|
| 3155 |
+
{
|
| 3156 |
+
"epoch": 3.67,
|
| 3157 |
+
"learning_rate": 3.816570716148838e-05,
|
| 3158 |
+
"loss": 0.5164,
|
| 3159 |
+
"step": 260000
|
| 3160 |
+
},
|
| 3161 |
+
{
|
| 3162 |
+
"epoch": 3.67,
|
| 3163 |
+
"learning_rate": 3.814196878498443e-05,
|
| 3164 |
+
"loss": 0.5089,
|
| 3165 |
+
"step": 260500
|
| 3166 |
+
},
|
| 3167 |
+
{
|
| 3168 |
+
"epoch": 3.68,
|
| 3169 |
+
"learning_rate": 3.8118230408480486e-05,
|
| 3170 |
+
"loss": 0.499,
|
| 3171 |
+
"step": 261000
|
| 3172 |
+
},
|
| 3173 |
+
{
|
| 3174 |
+
"epoch": 3.69,
|
| 3175 |
+
"learning_rate": 3.809453950872955e-05,
|
| 3176 |
+
"loss": 0.5004,
|
| 3177 |
+
"step": 261500
|
| 3178 |
+
},
|
| 3179 |
+
{
|
| 3180 |
+
"epoch": 3.69,
|
| 3181 |
+
"learning_rate": 3.807084860897861e-05,
|
| 3182 |
+
"loss": 0.4955,
|
| 3183 |
+
"step": 262000
|
| 3184 |
+
},
|
| 3185 |
+
{
|
| 3186 |
+
"epoch": 3.7,
|
| 3187 |
+
"learning_rate": 3.804715770922768e-05,
|
| 3188 |
+
"loss": 0.5021,
|
| 3189 |
+
"step": 262500
|
| 3190 |
+
},
|
| 3191 |
+
{
|
| 3192 |
+
"epoch": 3.71,
|
| 3193 |
+
"learning_rate": 3.802341933272373e-05,
|
| 3194 |
+
"loss": 0.5064,
|
| 3195 |
+
"step": 263000
|
| 3196 |
+
},
|
| 3197 |
+
{
|
| 3198 |
+
"epoch": 3.72,
|
| 3199 |
+
"learning_rate": 3.799968095621979e-05,
|
| 3200 |
+
"loss": 0.4947,
|
| 3201 |
+
"step": 263500
|
| 3202 |
+
},
|
| 3203 |
+
{
|
| 3204 |
+
"epoch": 3.72,
|
| 3205 |
+
"learning_rate": 3.7975942579715846e-05,
|
| 3206 |
+
"loss": 0.5033,
|
| 3207 |
+
"step": 264000
|
| 3208 |
+
},
|
| 3209 |
+
{
|
| 3210 |
+
"epoch": 3.73,
|
| 3211 |
+
"learning_rate": 3.79522042032119e-05,
|
| 3212 |
+
"loss": 0.5022,
|
| 3213 |
+
"step": 264500
|
| 3214 |
+
},
|
| 3215 |
+
{
|
| 3216 |
+
"epoch": 3.74,
|
| 3217 |
+
"learning_rate": 3.792846582670796e-05,
|
| 3218 |
+
"loss": 0.4956,
|
| 3219 |
+
"step": 265000
|
| 3220 |
+
},
|
| 3221 |
+
{
|
| 3222 |
+
"epoch": 3.74,
|
| 3223 |
+
"learning_rate": 3.790477492695702e-05,
|
| 3224 |
+
"loss": 0.5102,
|
| 3225 |
+
"step": 265500
|
| 3226 |
+
},
|
| 3227 |
+
{
|
| 3228 |
+
"epoch": 3.75,
|
| 3229 |
+
"learning_rate": 3.788103655045307e-05,
|
| 3230 |
+
"loss": 0.509,
|
| 3231 |
+
"step": 266000
|
| 3232 |
+
},
|
| 3233 |
+
{
|
| 3234 |
+
"epoch": 3.76,
|
| 3235 |
+
"learning_rate": 3.785729817394913e-05,
|
| 3236 |
+
"loss": 0.5048,
|
| 3237 |
+
"step": 266500
|
| 3238 |
+
},
|
| 3239 |
+
{
|
| 3240 |
+
"epoch": 3.76,
|
| 3241 |
+
"learning_rate": 3.783355979744518e-05,
|
| 3242 |
+
"loss": 0.4954,
|
| 3243 |
+
"step": 267000
|
| 3244 |
+
},
|
| 3245 |
+
{
|
| 3246 |
+
"epoch": 3.77,
|
| 3247 |
+
"learning_rate": 3.780982142094124e-05,
|
| 3248 |
+
"loss": 0.5009,
|
| 3249 |
+
"step": 267500
|
| 3250 |
+
},
|
| 3251 |
+
{
|
| 3252 |
+
"epoch": 3.78,
|
| 3253 |
+
"learning_rate": 3.7786083044437294e-05,
|
| 3254 |
+
"loss": 0.4909,
|
| 3255 |
+
"step": 268000
|
| 3256 |
+
},
|
| 3257 |
+
{
|
| 3258 |
+
"epoch": 3.79,
|
| 3259 |
+
"learning_rate": 3.776234466793335e-05,
|
| 3260 |
+
"loss": 0.5018,
|
| 3261 |
+
"step": 268500
|
| 3262 |
+
},
|
| 3263 |
+
{
|
| 3264 |
+
"epoch": 3.79,
|
| 3265 |
+
"learning_rate": 3.77386062914294e-05,
|
| 3266 |
+
"loss": 0.4903,
|
| 3267 |
+
"step": 269000
|
| 3268 |
+
},
|
| 3269 |
+
{
|
| 3270 |
+
"epoch": 3.8,
|
| 3271 |
+
"learning_rate": 3.7714867914925454e-05,
|
| 3272 |
+
"loss": 0.504,
|
| 3273 |
+
"step": 269500
|
| 3274 |
+
},
|
| 3275 |
+
{
|
| 3276 |
+
"epoch": 3.81,
|
| 3277 |
+
"learning_rate": 3.7691129538421516e-05,
|
| 3278 |
+
"loss": 0.5088,
|
| 3279 |
+
"step": 270000
|
| 3280 |
+
},
|
| 3281 |
+
{
|
| 3282 |
+
"epoch": 3.81,
|
| 3283 |
+
"learning_rate": 3.7667438638670576e-05,
|
| 3284 |
+
"loss": 0.5009,
|
| 3285 |
+
"step": 270500
|
| 3286 |
+
},
|
| 3287 |
+
{
|
| 3288 |
+
"epoch": 3.82,
|
| 3289 |
+
"learning_rate": 3.7643747738919636e-05,
|
| 3290 |
+
"loss": 0.4924,
|
| 3291 |
+
"step": 271000
|
| 3292 |
+
},
|
| 3293 |
+
{
|
| 3294 |
+
"epoch": 3.83,
|
| 3295 |
+
"learning_rate": 3.76200568391687e-05,
|
| 3296 |
+
"loss": 0.4999,
|
| 3297 |
+
"step": 271500
|
| 3298 |
+
},
|
| 3299 |
+
{
|
| 3300 |
+
"epoch": 3.84,
|
| 3301 |
+
"learning_rate": 3.759631846266476e-05,
|
| 3302 |
+
"loss": 0.4997,
|
| 3303 |
+
"step": 272000
|
| 3304 |
+
},
|
| 3305 |
+
{
|
| 3306 |
+
"epoch": 3.84,
|
| 3307 |
+
"learning_rate": 3.7572580086160813e-05,
|
| 3308 |
+
"loss": 0.4968,
|
| 3309 |
+
"step": 272500
|
| 3310 |
+
},
|
| 3311 |
+
{
|
| 3312 |
+
"epoch": 3.85,
|
| 3313 |
+
"learning_rate": 3.754884170965687e-05,
|
| 3314 |
+
"loss": 0.4905,
|
| 3315 |
+
"step": 273000
|
| 3316 |
+
},
|
| 3317 |
+
{
|
| 3318 |
+
"epoch": 3.86,
|
| 3319 |
+
"learning_rate": 3.7525103333152924e-05,
|
| 3320 |
+
"loss": 0.5055,
|
| 3321 |
+
"step": 273500
|
| 3322 |
+
},
|
| 3323 |
+
{
|
| 3324 |
+
"epoch": 3.86,
|
| 3325 |
+
"learning_rate": 3.750136495664898e-05,
|
| 3326 |
+
"loss": 0.4918,
|
| 3327 |
+
"step": 274000
|
| 3328 |
+
},
|
| 3329 |
+
{
|
| 3330 |
+
"epoch": 3.87,
|
| 3331 |
+
"learning_rate": 3.747762658014503e-05,
|
| 3332 |
+
"loss": 0.4947,
|
| 3333 |
+
"step": 274500
|
| 3334 |
+
},
|
| 3335 |
+
{
|
| 3336 |
+
"epoch": 3.88,
|
| 3337 |
+
"learning_rate": 3.745388820364109e-05,
|
| 3338 |
+
"loss": 0.5016,
|
| 3339 |
+
"step": 275000
|
| 3340 |
+
},
|
| 3341 |
+
{
|
| 3342 |
+
"epoch": 3.88,
|
| 3343 |
+
"learning_rate": 3.7430149827137146e-05,
|
| 3344 |
+
"loss": 0.5084,
|
| 3345 |
+
"step": 275500
|
| 3346 |
+
},
|
| 3347 |
+
{
|
| 3348 |
+
"epoch": 3.89,
|
| 3349 |
+
"learning_rate": 3.74064114506332e-05,
|
| 3350 |
+
"loss": 0.4882,
|
| 3351 |
+
"step": 276000
|
| 3352 |
+
},
|
| 3353 |
+
{
|
| 3354 |
+
"epoch": 3.9,
|
| 3355 |
+
"learning_rate": 3.738272055088226e-05,
|
| 3356 |
+
"loss": 0.4948,
|
| 3357 |
+
"step": 276500
|
| 3358 |
+
},
|
| 3359 |
+
{
|
| 3360 |
+
"epoch": 3.91,
|
| 3361 |
+
"learning_rate": 3.735898217437832e-05,
|
| 3362 |
+
"loss": 0.4971,
|
| 3363 |
+
"step": 277000
|
| 3364 |
+
},
|
| 3365 |
+
{
|
| 3366 |
+
"epoch": 3.91,
|
| 3367 |
+
"learning_rate": 3.733524379787437e-05,
|
| 3368 |
+
"loss": 0.4867,
|
| 3369 |
+
"step": 277500
|
| 3370 |
+
},
|
| 3371 |
+
{
|
| 3372 |
+
"epoch": 3.92,
|
| 3373 |
+
"learning_rate": 3.731150542137043e-05,
|
| 3374 |
+
"loss": 0.4917,
|
| 3375 |
+
"step": 278000
|
| 3376 |
+
},
|
| 3377 |
+
{
|
| 3378 |
+
"epoch": 3.93,
|
| 3379 |
+
"learning_rate": 3.7287767044866484e-05,
|
| 3380 |
+
"loss": 0.4853,
|
| 3381 |
+
"step": 278500
|
| 3382 |
+
},
|
| 3383 |
+
{
|
| 3384 |
+
"epoch": 3.93,
|
| 3385 |
+
"learning_rate": 3.726402866836254e-05,
|
| 3386 |
+
"loss": 0.4948,
|
| 3387 |
+
"step": 279000
|
| 3388 |
+
},
|
| 3389 |
+
{
|
| 3390 |
+
"epoch": 3.94,
|
| 3391 |
+
"learning_rate": 3.7240290291858594e-05,
|
| 3392 |
+
"loss": 0.4935,
|
| 3393 |
+
"step": 279500
|
| 3394 |
+
},
|
| 3395 |
+
{
|
| 3396 |
+
"epoch": 3.95,
|
| 3397 |
+
"learning_rate": 3.721655191535465e-05,
|
| 3398 |
+
"loss": 0.4906,
|
| 3399 |
+
"step": 280000
|
| 3400 |
+
},
|
| 3401 |
+
{
|
| 3402 |
+
"epoch": 3.96,
|
| 3403 |
+
"learning_rate": 3.71928135388507e-05,
|
| 3404 |
+
"loss": 0.4992,
|
| 3405 |
+
"step": 280500
|
| 3406 |
+
},
|
| 3407 |
+
{
|
| 3408 |
+
"epoch": 3.96,
|
| 3409 |
+
"learning_rate": 3.716907516234676e-05,
|
| 3410 |
+
"loss": 0.4814,
|
| 3411 |
+
"step": 281000
|
| 3412 |
+
},
|
| 3413 |
+
{
|
| 3414 |
+
"epoch": 3.97,
|
| 3415 |
+
"learning_rate": 3.7145336785842816e-05,
|
| 3416 |
+
"loss": 0.4903,
|
| 3417 |
+
"step": 281500
|
| 3418 |
+
},
|
| 3419 |
+
{
|
| 3420 |
+
"epoch": 3.98,
|
| 3421 |
+
"learning_rate": 3.712159840933887e-05,
|
| 3422 |
+
"loss": 0.4869,
|
| 3423 |
+
"step": 282000
|
| 3424 |
+
},
|
| 3425 |
+
{
|
| 3426 |
+
"epoch": 3.98,
|
| 3427 |
+
"learning_rate": 3.709786003283493e-05,
|
| 3428 |
+
"loss": 0.493,
|
| 3429 |
+
"step": 282500
|
| 3430 |
+
},
|
| 3431 |
+
{
|
| 3432 |
+
"epoch": 3.99,
|
| 3433 |
+
"learning_rate": 3.7074121656330976e-05,
|
| 3434 |
+
"loss": 0.4909,
|
| 3435 |
+
"step": 283000
|
| 3436 |
+
},
|
| 3437 |
+
{
|
| 3438 |
+
"epoch": 4.0,
|
| 3439 |
+
"learning_rate": 3.705038327982703e-05,
|
| 3440 |
+
"loss": 0.4873,
|
| 3441 |
+
"step": 283500
|
| 3442 |
+
},
|
| 3443 |
+
{
|
| 3444 |
+
"epoch": 4.0,
|
| 3445 |
+
"eval_bleu": 43.6153,
|
| 3446 |
+
"eval_gen_len": 13.6544,
|
| 3447 |
+
"eval_loss": 0.9757916927337646,
|
| 3448 |
+
"eval_runtime": 9810.7798,
|
| 3449 |
+
"eval_samples_per_second": 14.457,
|
| 3450 |
+
"eval_steps_per_second": 1.807,
|
| 3451 |
+
"step": 283678
|
| 3452 |
+
},
|
| 3453 |
+
{
|
| 3454 |
+
"epoch": 4.0,
|
| 3455 |
+
"learning_rate": 3.70266923800761e-05,
|
| 3456 |
+
"loss": 0.4799,
|
| 3457 |
+
"step": 284000
|
| 3458 |
+
},
|
| 3459 |
+
{
|
| 3460 |
+
"epoch": 4.01,
|
| 3461 |
+
"learning_rate": 3.700300148032516e-05,
|
| 3462 |
+
"loss": 0.4853,
|
| 3463 |
+
"step": 284500
|
| 3464 |
+
},
|
| 3465 |
+
{
|
| 3466 |
+
"epoch": 4.02,
|
| 3467 |
+
"learning_rate": 3.6979263103821214e-05,
|
| 3468 |
+
"loss": 0.4872,
|
| 3469 |
+
"step": 285000
|
| 3470 |
+
},
|
| 3471 |
+
{
|
| 3472 |
+
"epoch": 4.03,
|
| 3473 |
+
"learning_rate": 3.695552472731727e-05,
|
| 3474 |
+
"loss": 0.4809,
|
| 3475 |
+
"step": 285500
|
| 3476 |
+
},
|
| 3477 |
+
{
|
| 3478 |
+
"epoch": 4.03,
|
| 3479 |
+
"learning_rate": 3.6931786350813324e-05,
|
| 3480 |
+
"loss": 0.4787,
|
| 3481 |
+
"step": 286000
|
| 3482 |
+
},
|
| 3483 |
+
{
|
| 3484 |
+
"epoch": 4.04,
|
| 3485 |
+
"learning_rate": 3.690804797430938e-05,
|
| 3486 |
+
"loss": 0.4797,
|
| 3487 |
+
"step": 286500
|
| 3488 |
+
},
|
| 3489 |
+
{
|
| 3490 |
+
"epoch": 4.05,
|
| 3491 |
+
"learning_rate": 3.6884309597805435e-05,
|
| 3492 |
+
"loss": 0.4811,
|
| 3493 |
+
"step": 287000
|
| 3494 |
+
},
|
| 3495 |
+
{
|
| 3496 |
+
"epoch": 4.05,
|
| 3497 |
+
"learning_rate": 3.686057122130149e-05,
|
| 3498 |
+
"loss": 0.4728,
|
| 3499 |
+
"step": 287500
|
| 3500 |
+
},
|
| 3501 |
+
{
|
| 3502 |
+
"epoch": 4.06,
|
| 3503 |
+
"learning_rate": 3.6836832844797546e-05,
|
| 3504 |
+
"loss": 0.4756,
|
| 3505 |
+
"step": 288000
|
| 3506 |
+
},
|
| 3507 |
+
{
|
| 3508 |
+
"epoch": 4.07,
|
| 3509 |
+
"learning_rate": 3.68130944682936e-05,
|
| 3510 |
+
"loss": 0.4728,
|
| 3511 |
+
"step": 288500
|
| 3512 |
+
},
|
| 3513 |
+
{
|
| 3514 |
+
"epoch": 4.08,
|
| 3515 |
+
"learning_rate": 3.678935609178966e-05,
|
| 3516 |
+
"loss": 0.4844,
|
| 3517 |
+
"step": 289000
|
| 3518 |
+
},
|
| 3519 |
+
{
|
| 3520 |
+
"epoch": 4.08,
|
| 3521 |
+
"learning_rate": 3.676561771528571e-05,
|
| 3522 |
+
"loss": 0.4677,
|
| 3523 |
+
"step": 289500
|
| 3524 |
+
},
|
| 3525 |
+
{
|
| 3526 |
+
"epoch": 4.09,
|
| 3527 |
+
"learning_rate": 3.674192681553478e-05,
|
| 3528 |
+
"loss": 0.4631,
|
| 3529 |
+
"step": 290000
|
| 3530 |
+
},
|
| 3531 |
+
{
|
| 3532 |
+
"epoch": 4.1,
|
| 3533 |
+
"learning_rate": 3.6718188439030835e-05,
|
| 3534 |
+
"loss": 0.4626,
|
| 3535 |
+
"step": 290500
|
| 3536 |
+
},
|
| 3537 |
+
{
|
| 3538 |
+
"epoch": 4.1,
|
| 3539 |
+
"learning_rate": 3.6694450062526884e-05,
|
| 3540 |
+
"loss": 0.4705,
|
| 3541 |
+
"step": 291000
|
| 3542 |
+
},
|
| 3543 |
+
{
|
| 3544 |
+
"epoch": 4.11,
|
| 3545 |
+
"learning_rate": 3.667071168602294e-05,
|
| 3546 |
+
"loss": 0.4582,
|
| 3547 |
+
"step": 291500
|
| 3548 |
+
},
|
| 3549 |
+
{
|
| 3550 |
+
"epoch": 4.12,
|
| 3551 |
+
"learning_rate": 3.6646973309518994e-05,
|
| 3552 |
+
"loss": 0.4668,
|
| 3553 |
+
"step": 292000
|
| 3554 |
+
},
|
| 3555 |
+
{
|
| 3556 |
+
"epoch": 4.12,
|
| 3557 |
+
"learning_rate": 3.662323493301505e-05,
|
| 3558 |
+
"loss": 0.4597,
|
| 3559 |
+
"step": 292500
|
| 3560 |
+
},
|
| 3561 |
+
{
|
| 3562 |
+
"epoch": 4.13,
|
| 3563 |
+
"learning_rate": 3.6599496556511105e-05,
|
| 3564 |
+
"loss": 0.4518,
|
| 3565 |
+
"step": 293000
|
| 3566 |
+
},
|
| 3567 |
+
{
|
| 3568 |
+
"epoch": 4.14,
|
| 3569 |
+
"learning_rate": 3.657575818000716e-05,
|
| 3570 |
+
"loss": 0.4586,
|
| 3571 |
+
"step": 293500
|
| 3572 |
+
},
|
| 3573 |
+
{
|
| 3574 |
+
"epoch": 4.15,
|
| 3575 |
+
"learning_rate": 3.6552019803503216e-05,
|
| 3576 |
+
"loss": 0.4413,
|
| 3577 |
+
"step": 294000
|
| 3578 |
+
},
|
| 3579 |
+
{
|
| 3580 |
+
"epoch": 4.15,
|
| 3581 |
+
"learning_rate": 3.6528328903752276e-05,
|
| 3582 |
+
"loss": 0.4528,
|
| 3583 |
+
"step": 294500
|
| 3584 |
+
},
|
| 3585 |
+
{
|
| 3586 |
+
"epoch": 4.16,
|
| 3587 |
+
"learning_rate": 3.650459052724833e-05,
|
| 3588 |
+
"loss": 0.4512,
|
| 3589 |
+
"step": 295000
|
| 3590 |
+
},
|
| 3591 |
+
{
|
| 3592 |
+
"epoch": 4.17,
|
| 3593 |
+
"learning_rate": 3.64808996274974e-05,
|
| 3594 |
+
"loss": 0.4555,
|
| 3595 |
+
"step": 295500
|
| 3596 |
+
},
|
| 3597 |
+
{
|
| 3598 |
+
"epoch": 4.17,
|
| 3599 |
+
"learning_rate": 3.6457161250993454e-05,
|
| 3600 |
+
"loss": 0.4474,
|
| 3601 |
+
"step": 296000
|
| 3602 |
+
},
|
| 3603 |
+
{
|
| 3604 |
+
"epoch": 4.18,
|
| 3605 |
+
"learning_rate": 3.643342287448951e-05,
|
| 3606 |
+
"loss": 0.4426,
|
| 3607 |
+
"step": 296500
|
| 3608 |
+
},
|
| 3609 |
+
{
|
| 3610 |
+
"epoch": 4.19,
|
| 3611 |
+
"learning_rate": 3.6409684497985565e-05,
|
| 3612 |
+
"loss": 0.4459,
|
| 3613 |
+
"step": 297000
|
| 3614 |
+
},
|
| 3615 |
+
{
|
| 3616 |
+
"epoch": 4.19,
|
| 3617 |
+
"learning_rate": 3.6385993598234625e-05,
|
| 3618 |
+
"loss": 0.4473,
|
| 3619 |
+
"step": 297500
|
| 3620 |
+
},
|
| 3621 |
+
{
|
| 3622 |
+
"epoch": 4.2,
|
| 3623 |
+
"learning_rate": 3.636225522173068e-05,
|
| 3624 |
+
"loss": 0.4426,
|
| 3625 |
+
"step": 298000
|
| 3626 |
+
},
|
| 3627 |
+
{
|
| 3628 |
+
"epoch": 4.21,
|
| 3629 |
+
"learning_rate": 3.6338516845226736e-05,
|
| 3630 |
+
"loss": 0.4413,
|
| 3631 |
+
"step": 298500
|
| 3632 |
+
},
|
| 3633 |
+
{
|
| 3634 |
+
"epoch": 4.22,
|
| 3635 |
+
"learning_rate": 3.631477846872279e-05,
|
| 3636 |
+
"loss": 0.4335,
|
| 3637 |
+
"step": 299000
|
| 3638 |
+
},
|
| 3639 |
+
{
|
| 3640 |
+
"epoch": 4.22,
|
| 3641 |
+
"learning_rate": 3.629104009221885e-05,
|
| 3642 |
+
"loss": 0.4422,
|
| 3643 |
+
"step": 299500
|
| 3644 |
+
},
|
| 3645 |
+
{
|
| 3646 |
+
"epoch": 4.23,
|
| 3647 |
+
"learning_rate": 3.626734919246791e-05,
|
| 3648 |
+
"loss": 0.4566,
|
| 3649 |
+
"step": 300000
|
| 3650 |
+
},
|
| 3651 |
+
{
|
| 3652 |
+
"epoch": 4.24,
|
| 3653 |
+
"learning_rate": 3.624361081596396e-05,
|
| 3654 |
+
"loss": 0.4375,
|
| 3655 |
+
"step": 300500
|
| 3656 |
+
},
|
| 3657 |
+
{
|
| 3658 |
+
"epoch": 4.24,
|
| 3659 |
+
"learning_rate": 3.621987243946002e-05,
|
| 3660 |
+
"loss": 0.4433,
|
| 3661 |
+
"step": 301000
|
| 3662 |
+
},
|
| 3663 |
+
{
|
| 3664 |
+
"epoch": 4.25,
|
| 3665 |
+
"learning_rate": 3.6196181539709084e-05,
|
| 3666 |
+
"loss": 0.4311,
|
| 3667 |
+
"step": 301500
|
| 3668 |
+
},
|
| 3669 |
+
{
|
| 3670 |
+
"epoch": 4.26,
|
| 3671 |
+
"learning_rate": 3.617244316320514e-05,
|
| 3672 |
+
"loss": 0.4282,
|
| 3673 |
+
"step": 302000
|
| 3674 |
+
},
|
| 3675 |
+
{
|
| 3676 |
+
"epoch": 4.27,
|
| 3677 |
+
"learning_rate": 3.6148704786701195e-05,
|
| 3678 |
+
"loss": 0.4302,
|
| 3679 |
+
"step": 302500
|
| 3680 |
+
},
|
| 3681 |
+
{
|
| 3682 |
+
"epoch": 4.27,
|
| 3683 |
+
"learning_rate": 3.612496641019725e-05,
|
| 3684 |
+
"loss": 0.446,
|
| 3685 |
+
"step": 303000
|
| 3686 |
+
},
|
| 3687 |
+
{
|
| 3688 |
+
"epoch": 4.28,
|
| 3689 |
+
"learning_rate": 3.61012280336933e-05,
|
| 3690 |
+
"loss": 0.4282,
|
| 3691 |
+
"step": 303500
|
| 3692 |
+
},
|
| 3693 |
+
{
|
| 3694 |
+
"epoch": 4.29,
|
| 3695 |
+
"learning_rate": 3.607748965718936e-05,
|
| 3696 |
+
"loss": 0.4335,
|
| 3697 |
+
"step": 304000
|
| 3698 |
+
},
|
| 3699 |
+
{
|
| 3700 |
+
"epoch": 4.29,
|
| 3701 |
+
"learning_rate": 3.605375128068542e-05,
|
| 3702 |
+
"loss": 0.433,
|
| 3703 |
+
"step": 304500
|
| 3704 |
+
},
|
| 3705 |
+
{
|
| 3706 |
+
"epoch": 4.3,
|
| 3707 |
+
"learning_rate": 3.603001290418147e-05,
|
| 3708 |
+
"loss": 0.4168,
|
| 3709 |
+
"step": 305000
|
| 3710 |
+
},
|
| 3711 |
+
{
|
| 3712 |
+
"epoch": 4.31,
|
| 3713 |
+
"learning_rate": 3.600627452767753e-05,
|
| 3714 |
+
"loss": 0.4391,
|
| 3715 |
+
"step": 305500
|
| 3716 |
+
},
|
| 3717 |
+
{
|
| 3718 |
+
"epoch": 4.31,
|
| 3719 |
+
"learning_rate": 3.598253615117358e-05,
|
| 3720 |
+
"loss": 0.433,
|
| 3721 |
+
"step": 306000
|
| 3722 |
+
},
|
| 3723 |
+
{
|
| 3724 |
+
"epoch": 4.32,
|
| 3725 |
+
"learning_rate": 3.595879777466963e-05,
|
| 3726 |
+
"loss": 0.4391,
|
| 3727 |
+
"step": 306500
|
| 3728 |
+
},
|
| 3729 |
+
{
|
| 3730 |
+
"epoch": 4.33,
|
| 3731 |
+
"learning_rate": 3.593505939816569e-05,
|
| 3732 |
+
"loss": 0.4276,
|
| 3733 |
+
"step": 307000
|
| 3734 |
+
},
|
| 3735 |
+
{
|
| 3736 |
+
"epoch": 4.34,
|
| 3737 |
+
"learning_rate": 3.591132102166174e-05,
|
| 3738 |
+
"loss": 0.4343,
|
| 3739 |
+
"step": 307500
|
| 3740 |
+
},
|
| 3741 |
+
{
|
| 3742 |
+
"epoch": 4.34,
|
| 3743 |
+
"learning_rate": 3.5887582645157805e-05,
|
| 3744 |
+
"loss": 0.4407,
|
| 3745 |
+
"step": 308000
|
| 3746 |
+
},
|
| 3747 |
+
{
|
| 3748 |
+
"epoch": 4.35,
|
| 3749 |
+
"learning_rate": 3.5863891745406865e-05,
|
| 3750 |
+
"loss": 0.4258,
|
| 3751 |
+
"step": 308500
|
| 3752 |
+
},
|
| 3753 |
+
{
|
| 3754 |
+
"epoch": 4.36,
|
| 3755 |
+
"learning_rate": 3.584015336890292e-05,
|
| 3756 |
+
"loss": 0.4285,
|
| 3757 |
+
"step": 309000
|
| 3758 |
+
},
|
| 3759 |
+
{
|
| 3760 |
+
"epoch": 4.36,
|
| 3761 |
+
"learning_rate": 3.581641499239897e-05,
|
| 3762 |
+
"loss": 0.4316,
|
| 3763 |
+
"step": 309500
|
| 3764 |
+
},
|
| 3765 |
+
{
|
| 3766 |
+
"epoch": 4.37,
|
| 3767 |
+
"learning_rate": 3.5792676615895025e-05,
|
| 3768 |
+
"loss": 0.4249,
|
| 3769 |
+
"step": 310000
|
| 3770 |
+
},
|
| 3771 |
+
{
|
| 3772 |
+
"epoch": 4.38,
|
| 3773 |
+
"learning_rate": 3.576898571614409e-05,
|
| 3774 |
+
"loss": 0.4276,
|
| 3775 |
+
"step": 310500
|
| 3776 |
+
},
|
| 3777 |
+
{
|
| 3778 |
+
"epoch": 4.39,
|
| 3779 |
+
"learning_rate": 3.574524733964015e-05,
|
| 3780 |
+
"loss": 0.4286,
|
| 3781 |
+
"step": 311000
|
| 3782 |
+
},
|
| 3783 |
+
{
|
| 3784 |
+
"epoch": 4.39,
|
| 3785 |
+
"learning_rate": 3.57215089631362e-05,
|
| 3786 |
+
"loss": 0.4195,
|
| 3787 |
+
"step": 311500
|
| 3788 |
+
},
|
| 3789 |
+
{
|
| 3790 |
+
"epoch": 4.4,
|
| 3791 |
+
"learning_rate": 3.569777058663226e-05,
|
| 3792 |
+
"loss": 0.419,
|
| 3793 |
+
"step": 312000
|
| 3794 |
+
},
|
| 3795 |
+
{
|
| 3796 |
+
"epoch": 4.41,
|
| 3797 |
+
"learning_rate": 3.567407968688132e-05,
|
| 3798 |
+
"loss": 0.422,
|
| 3799 |
+
"step": 312500
|
| 3800 |
+
},
|
| 3801 |
+
{
|
| 3802 |
+
"epoch": 4.41,
|
| 3803 |
+
"learning_rate": 3.5650388787130385e-05,
|
| 3804 |
+
"loss": 0.4299,
|
| 3805 |
+
"step": 313000
|
| 3806 |
+
},
|
| 3807 |
+
{
|
| 3808 |
+
"epoch": 4.42,
|
| 3809 |
+
"learning_rate": 3.562665041062644e-05,
|
| 3810 |
+
"loss": 0.4341,
|
| 3811 |
+
"step": 313500
|
| 3812 |
+
},
|
| 3813 |
+
{
|
| 3814 |
+
"epoch": 4.43,
|
| 3815 |
+
"learning_rate": 3.5602912034122496e-05,
|
| 3816 |
+
"loss": 0.4122,
|
| 3817 |
+
"step": 314000
|
| 3818 |
+
},
|
| 3819 |
+
{
|
| 3820 |
+
"epoch": 4.43,
|
| 3821 |
+
"learning_rate": 3.5579173657618544e-05,
|
| 3822 |
+
"loss": 0.4251,
|
| 3823 |
+
"step": 314500
|
| 3824 |
+
},
|
| 3825 |
+
{
|
| 3826 |
+
"epoch": 4.44,
|
| 3827 |
+
"learning_rate": 3.55554352811146e-05,
|
| 3828 |
+
"loss": 0.4298,
|
| 3829 |
+
"step": 315000
|
| 3830 |
+
},
|
| 3831 |
+
{
|
| 3832 |
+
"epoch": 4.45,
|
| 3833 |
+
"learning_rate": 3.553169690461066e-05,
|
| 3834 |
+
"loss": 0.4183,
|
| 3835 |
+
"step": 315500
|
| 3836 |
+
},
|
| 3837 |
+
{
|
| 3838 |
+
"epoch": 4.46,
|
| 3839 |
+
"learning_rate": 3.550795852810672e-05,
|
| 3840 |
+
"loss": 0.4199,
|
| 3841 |
+
"step": 316000
|
| 3842 |
+
},
|
| 3843 |
+
{
|
| 3844 |
+
"epoch": 4.46,
|
| 3845 |
+
"learning_rate": 3.548426762835578e-05,
|
| 3846 |
+
"loss": 0.4281,
|
| 3847 |
+
"step": 316500
|
| 3848 |
+
},
|
| 3849 |
+
{
|
| 3850 |
+
"epoch": 4.47,
|
| 3851 |
+
"learning_rate": 3.546052925185183e-05,
|
| 3852 |
+
"loss": 0.4158,
|
| 3853 |
+
"step": 317000
|
| 3854 |
+
},
|
| 3855 |
+
{
|
| 3856 |
+
"epoch": 4.48,
|
| 3857 |
+
"learning_rate": 3.543679087534789e-05,
|
| 3858 |
+
"loss": 0.4267,
|
| 3859 |
+
"step": 317500
|
| 3860 |
+
},
|
| 3861 |
+
{
|
| 3862 |
+
"epoch": 4.48,
|
| 3863 |
+
"learning_rate": 3.541305249884394e-05,
|
| 3864 |
+
"loss": 0.4193,
|
| 3865 |
+
"step": 318000
|
| 3866 |
+
},
|
| 3867 |
+
{
|
| 3868 |
+
"epoch": 4.49,
|
| 3869 |
+
"learning_rate": 3.538931412234e-05,
|
| 3870 |
+
"loss": 0.4106,
|
| 3871 |
+
"step": 318500
|
| 3872 |
+
},
|
| 3873 |
+
{
|
| 3874 |
+
"epoch": 4.5,
|
| 3875 |
+
"learning_rate": 3.5365575745836055e-05,
|
| 3876 |
+
"loss": 0.4186,
|
| 3877 |
+
"step": 319000
|
| 3878 |
+
},
|
| 3879 |
+
{
|
| 3880 |
+
"epoch": 4.51,
|
| 3881 |
+
"learning_rate": 3.534183736933211e-05,
|
| 3882 |
+
"loss": 0.4141,
|
| 3883 |
+
"step": 319500
|
| 3884 |
+
},
|
| 3885 |
+
{
|
| 3886 |
+
"epoch": 4.51,
|
| 3887 |
+
"learning_rate": 3.5318098992828166e-05,
|
| 3888 |
+
"loss": 0.398,
|
| 3889 |
+
"step": 320000
|
| 3890 |
+
},
|
| 3891 |
+
{
|
| 3892 |
+
"epoch": 4.52,
|
| 3893 |
+
"learning_rate": 3.529436061632422e-05,
|
| 3894 |
+
"loss": 0.4241,
|
| 3895 |
+
"step": 320500
|
| 3896 |
+
},
|
| 3897 |
+
{
|
| 3898 |
+
"epoch": 4.53,
|
| 3899 |
+
"learning_rate": 3.527062223982027e-05,
|
| 3900 |
+
"loss": 0.422,
|
| 3901 |
+
"step": 321000
|
| 3902 |
+
},
|
| 3903 |
+
{
|
| 3904 |
+
"epoch": 4.53,
|
| 3905 |
+
"learning_rate": 3.5246883863316325e-05,
|
| 3906 |
+
"loss": 0.4167,
|
| 3907 |
+
"step": 321500
|
| 3908 |
+
},
|
| 3909 |
+
{
|
| 3910 |
+
"epoch": 4.54,
|
| 3911 |
+
"learning_rate": 3.522319296356539e-05,
|
| 3912 |
+
"loss": 0.4214,
|
| 3913 |
+
"step": 322000
|
| 3914 |
+
},
|
| 3915 |
+
{
|
| 3916 |
+
"epoch": 4.55,
|
| 3917 |
+
"learning_rate": 3.519950206381445e-05,
|
| 3918 |
+
"loss": 0.4171,
|
| 3919 |
+
"step": 322500
|
| 3920 |
+
},
|
| 3921 |
+
{
|
| 3922 |
+
"epoch": 4.55,
|
| 3923 |
+
"learning_rate": 3.517576368731051e-05,
|
| 3924 |
+
"loss": 0.4211,
|
| 3925 |
+
"step": 323000
|
| 3926 |
+
},
|
| 3927 |
+
{
|
| 3928 |
+
"epoch": 4.56,
|
| 3929 |
+
"learning_rate": 3.5152072787559574e-05,
|
| 3930 |
+
"loss": 0.41,
|
| 3931 |
+
"step": 323500
|
| 3932 |
+
},
|
| 3933 |
+
{
|
| 3934 |
+
"epoch": 4.57,
|
| 3935 |
+
"learning_rate": 3.512833441105563e-05,
|
| 3936 |
+
"loss": 0.4243,
|
| 3937 |
+
"step": 324000
|
| 3938 |
+
},
|
| 3939 |
+
{
|
| 3940 |
+
"epoch": 4.58,
|
| 3941 |
+
"learning_rate": 3.5104596034551685e-05,
|
| 3942 |
+
"loss": 0.4066,
|
| 3943 |
+
"step": 324500
|
| 3944 |
+
},
|
| 3945 |
+
{
|
| 3946 |
+
"epoch": 4.58,
|
| 3947 |
+
"learning_rate": 3.508085765804774e-05,
|
| 3948 |
+
"loss": 0.4055,
|
| 3949 |
+
"step": 325000
|
| 3950 |
+
},
|
| 3951 |
+
{
|
| 3952 |
+
"epoch": 4.59,
|
| 3953 |
+
"learning_rate": 3.5057119281543796e-05,
|
| 3954 |
+
"loss": 0.419,
|
| 3955 |
+
"step": 325500
|
| 3956 |
+
},
|
| 3957 |
+
{
|
| 3958 |
+
"epoch": 4.6,
|
| 3959 |
+
"learning_rate": 3.5033380905039845e-05,
|
| 3960 |
+
"loss": 0.4222,
|
| 3961 |
+
"step": 326000
|
| 3962 |
+
},
|
| 3963 |
+
{
|
| 3964 |
+
"epoch": 4.6,
|
| 3965 |
+
"learning_rate": 3.50096425285359e-05,
|
| 3966 |
+
"loss": 0.4092,
|
| 3967 |
+
"step": 326500
|
| 3968 |
+
},
|
| 3969 |
+
{
|
| 3970 |
+
"epoch": 4.61,
|
| 3971 |
+
"learning_rate": 3.498595162878497e-05,
|
| 3972 |
+
"loss": 0.4094,
|
| 3973 |
+
"step": 327000
|
| 3974 |
+
},
|
| 3975 |
+
{
|
| 3976 |
+
"epoch": 4.62,
|
| 3977 |
+
"learning_rate": 3.496221325228102e-05,
|
| 3978 |
+
"loss": 0.4102,
|
| 3979 |
+
"step": 327500
|
| 3980 |
+
},
|
| 3981 |
+
{
|
| 3982 |
+
"epoch": 4.62,
|
| 3983 |
+
"learning_rate": 3.493847487577708e-05,
|
| 3984 |
+
"loss": 0.4096,
|
| 3985 |
+
"step": 328000
|
| 3986 |
+
},
|
| 3987 |
+
{
|
| 3988 |
+
"epoch": 4.63,
|
| 3989 |
+
"learning_rate": 3.4914736499273133e-05,
|
| 3990 |
+
"loss": 0.4052,
|
| 3991 |
+
"step": 328500
|
| 3992 |
+
},
|
| 3993 |
+
{
|
| 3994 |
+
"epoch": 4.64,
|
| 3995 |
+
"learning_rate": 3.489099812276919e-05,
|
| 3996 |
+
"loss": 0.4076,
|
| 3997 |
+
"step": 329000
|
| 3998 |
+
},
|
| 3999 |
+
{
|
| 4000 |
+
"epoch": 4.65,
|
| 4001 |
+
"learning_rate": 3.486725974626524e-05,
|
| 4002 |
+
"loss": 0.414,
|
| 4003 |
+
"step": 329500
|
| 4004 |
+
},
|
| 4005 |
+
{
|
| 4006 |
+
"epoch": 4.65,
|
| 4007 |
+
"learning_rate": 3.48435213697613e-05,
|
| 4008 |
+
"loss": 0.4084,
|
| 4009 |
+
"step": 330000
|
| 4010 |
+
},
|
| 4011 |
+
{
|
| 4012 |
+
"epoch": 4.66,
|
| 4013 |
+
"learning_rate": 3.4819782993257355e-05,
|
| 4014 |
+
"loss": 0.405,
|
| 4015 |
+
"step": 330500
|
| 4016 |
+
},
|
| 4017 |
+
{
|
| 4018 |
+
"epoch": 4.67,
|
| 4019 |
+
"learning_rate": 3.479604461675341e-05,
|
| 4020 |
+
"loss": 0.4167,
|
| 4021 |
+
"step": 331000
|
| 4022 |
+
},
|
| 4023 |
+
{
|
| 4024 |
+
"epoch": 4.67,
|
| 4025 |
+
"learning_rate": 3.4772306240249466e-05,
|
| 4026 |
+
"loss": 0.4113,
|
| 4027 |
+
"step": 331500
|
| 4028 |
+
},
|
| 4029 |
+
{
|
| 4030 |
+
"epoch": 4.68,
|
| 4031 |
+
"learning_rate": 3.4748615340498526e-05,
|
| 4032 |
+
"loss": 0.403,
|
| 4033 |
+
"step": 332000
|
| 4034 |
+
},
|
| 4035 |
+
{
|
| 4036 |
+
"epoch": 4.69,
|
| 4037 |
+
"learning_rate": 3.472492444074759e-05,
|
| 4038 |
+
"loss": 0.4066,
|
| 4039 |
+
"step": 332500
|
| 4040 |
+
},
|
| 4041 |
+
{
|
| 4042 |
+
"epoch": 4.7,
|
| 4043 |
+
"learning_rate": 3.470118606424365e-05,
|
| 4044 |
+
"loss": 0.3965,
|
| 4045 |
+
"step": 333000
|
| 4046 |
+
},
|
| 4047 |
+
{
|
| 4048 |
+
"epoch": 4.7,
|
| 4049 |
+
"learning_rate": 3.467749516449271e-05,
|
| 4050 |
+
"loss": 0.4052,
|
| 4051 |
+
"step": 333500
|
| 4052 |
+
},
|
| 4053 |
+
{
|
| 4054 |
+
"epoch": 4.71,
|
| 4055 |
+
"learning_rate": 3.4653756787988764e-05,
|
| 4056 |
+
"loss": 0.4093,
|
| 4057 |
+
"step": 334000
|
| 4058 |
+
},
|
| 4059 |
+
{
|
| 4060 |
+
"epoch": 4.72,
|
| 4061 |
+
"learning_rate": 3.463001841148481e-05,
|
| 4062 |
+
"loss": 0.4015,
|
| 4063 |
+
"step": 334500
|
| 4064 |
+
},
|
| 4065 |
+
{
|
| 4066 |
+
"epoch": 4.72,
|
| 4067 |
+
"learning_rate": 3.4606280034980875e-05,
|
| 4068 |
+
"loss": 0.4096,
|
| 4069 |
+
"step": 335000
|
| 4070 |
+
},
|
| 4071 |
+
{
|
| 4072 |
+
"epoch": 4.73,
|
| 4073 |
+
"learning_rate": 3.458254165847693e-05,
|
| 4074 |
+
"loss": 0.4041,
|
| 4075 |
+
"step": 335500
|
| 4076 |
+
},
|
| 4077 |
+
{
|
| 4078 |
+
"epoch": 4.74,
|
| 4079 |
+
"learning_rate": 3.4558803281972986e-05,
|
| 4080 |
+
"loss": 0.4046,
|
| 4081 |
+
"step": 336000
|
| 4082 |
+
},
|
| 4083 |
+
{
|
| 4084 |
+
"epoch": 4.74,
|
| 4085 |
+
"learning_rate": 3.453506490546904e-05,
|
| 4086 |
+
"loss": 0.4112,
|
| 4087 |
+
"step": 336500
|
| 4088 |
+
},
|
| 4089 |
+
{
|
| 4090 |
+
"epoch": 4.75,
|
| 4091 |
+
"learning_rate": 3.4511326528965097e-05,
|
| 4092 |
+
"loss": 0.4107,
|
| 4093 |
+
"step": 337000
|
| 4094 |
+
},
|
| 4095 |
+
{
|
| 4096 |
+
"epoch": 4.76,
|
| 4097 |
+
"learning_rate": 3.4487588152461145e-05,
|
| 4098 |
+
"loss": 0.406,
|
| 4099 |
+
"step": 337500
|
| 4100 |
+
},
|
| 4101 |
+
{
|
| 4102 |
+
"epoch": 4.77,
|
| 4103 |
+
"learning_rate": 3.44638497759572e-05,
|
| 4104 |
+
"loss": 0.4008,
|
| 4105 |
+
"step": 338000
|
| 4106 |
+
},
|
| 4107 |
+
{
|
| 4108 |
+
"epoch": 4.77,
|
| 4109 |
+
"learning_rate": 3.444011139945326e-05,
|
| 4110 |
+
"loss": 0.4088,
|
| 4111 |
+
"step": 338500
|
| 4112 |
+
},
|
| 4113 |
+
{
|
| 4114 |
+
"epoch": 4.78,
|
| 4115 |
+
"learning_rate": 3.441637302294932e-05,
|
| 4116 |
+
"loss": 0.4006,
|
| 4117 |
+
"step": 339000
|
| 4118 |
+
},
|
| 4119 |
+
{
|
| 4120 |
+
"epoch": 4.79,
|
| 4121 |
+
"learning_rate": 3.4392634646445374e-05,
|
| 4122 |
+
"loss": 0.3962,
|
| 4123 |
+
"step": 339500
|
| 4124 |
+
},
|
| 4125 |
+
{
|
| 4126 |
+
"epoch": 4.79,
|
| 4127 |
+
"learning_rate": 3.436889626994142e-05,
|
| 4128 |
+
"loss": 0.4048,
|
| 4129 |
+
"step": 340000
|
| 4130 |
+
},
|
| 4131 |
+
{
|
| 4132 |
+
"epoch": 4.8,
|
| 4133 |
+
"learning_rate": 3.434515789343748e-05,
|
| 4134 |
+
"loss": 0.4076,
|
| 4135 |
+
"step": 340500
|
| 4136 |
+
},
|
| 4137 |
+
{
|
| 4138 |
+
"epoch": 4.81,
|
| 4139 |
+
"learning_rate": 3.4321419516933533e-05,
|
| 4140 |
+
"loss": 0.4108,
|
| 4141 |
+
"step": 341000
|
| 4142 |
+
},
|
| 4143 |
+
{
|
| 4144 |
+
"epoch": 4.82,
|
| 4145 |
+
"learning_rate": 3.429768114042959e-05,
|
| 4146 |
+
"loss": 0.4037,
|
| 4147 |
+
"step": 341500
|
| 4148 |
+
},
|
| 4149 |
+
{
|
| 4150 |
+
"epoch": 4.82,
|
| 4151 |
+
"learning_rate": 3.4273942763925644e-05,
|
| 4152 |
+
"loss": 0.4019,
|
| 4153 |
+
"step": 342000
|
| 4154 |
+
},
|
| 4155 |
+
{
|
| 4156 |
+
"epoch": 4.83,
|
| 4157 |
+
"learning_rate": 3.42502043874217e-05,
|
| 4158 |
+
"loss": 0.3984,
|
| 4159 |
+
"step": 342500
|
| 4160 |
+
},
|
| 4161 |
+
{
|
| 4162 |
+
"epoch": 4.84,
|
| 4163 |
+
"learning_rate": 3.4226466010917755e-05,
|
| 4164 |
+
"loss": 0.4065,
|
| 4165 |
+
"step": 343000
|
| 4166 |
+
},
|
| 4167 |
+
{
|
| 4168 |
+
"epoch": 4.84,
|
| 4169 |
+
"learning_rate": 3.4202775111166815e-05,
|
| 4170 |
+
"loss": 0.3995,
|
| 4171 |
+
"step": 343500
|
| 4172 |
+
},
|
| 4173 |
+
{
|
| 4174 |
+
"epoch": 4.85,
|
| 4175 |
+
"learning_rate": 3.417903673466287e-05,
|
| 4176 |
+
"loss": 0.3979,
|
| 4177 |
+
"step": 344000
|
| 4178 |
+
},
|
| 4179 |
+
{
|
| 4180 |
+
"epoch": 4.86,
|
| 4181 |
+
"learning_rate": 3.4155298358158926e-05,
|
| 4182 |
+
"loss": 0.4088,
|
| 4183 |
+
"step": 344500
|
| 4184 |
+
},
|
| 4185 |
+
{
|
| 4186 |
+
"epoch": 4.86,
|
| 4187 |
+
"learning_rate": 3.413155998165499e-05,
|
| 4188 |
+
"loss": 0.3951,
|
| 4189 |
+
"step": 345000
|
| 4190 |
+
},
|
| 4191 |
+
{
|
| 4192 |
+
"epoch": 4.87,
|
| 4193 |
+
"learning_rate": 3.4107821605151044e-05,
|
| 4194 |
+
"loss": 0.4011,
|
| 4195 |
+
"step": 345500
|
| 4196 |
+
},
|
| 4197 |
+
{
|
| 4198 |
+
"epoch": 4.88,
|
| 4199 |
+
"learning_rate": 3.408408322864709e-05,
|
| 4200 |
+
"loss": 0.4073,
|
| 4201 |
+
"step": 346000
|
| 4202 |
+
},
|
| 4203 |
+
{
|
| 4204 |
+
"epoch": 4.89,
|
| 4205 |
+
"learning_rate": 3.406034485214315e-05,
|
| 4206 |
+
"loss": 0.4042,
|
| 4207 |
+
"step": 346500
|
| 4208 |
+
},
|
| 4209 |
+
{
|
| 4210 |
+
"epoch": 4.89,
|
| 4211 |
+
"learning_rate": 3.4036606475639203e-05,
|
| 4212 |
+
"loss": 0.3924,
|
| 4213 |
+
"step": 347000
|
| 4214 |
+
},
|
| 4215 |
+
{
|
| 4216 |
+
"epoch": 4.9,
|
| 4217 |
+
"learning_rate": 3.401286809913526e-05,
|
| 4218 |
+
"loss": 0.3928,
|
| 4219 |
+
"step": 347500
|
| 4220 |
+
},
|
| 4221 |
+
{
|
| 4222 |
+
"epoch": 4.91,
|
| 4223 |
+
"learning_rate": 3.3989177199384326e-05,
|
| 4224 |
+
"loss": 0.4044,
|
| 4225 |
+
"step": 348000
|
| 4226 |
+
},
|
| 4227 |
+
{
|
| 4228 |
+
"epoch": 4.91,
|
| 4229 |
+
"learning_rate": 3.396543882288038e-05,
|
| 4230 |
+
"loss": 0.3934,
|
| 4231 |
+
"step": 348500
|
| 4232 |
+
},
|
| 4233 |
+
{
|
| 4234 |
+
"epoch": 4.92,
|
| 4235 |
+
"learning_rate": 3.394170044637644e-05,
|
| 4236 |
+
"loss": 0.3946,
|
| 4237 |
+
"step": 349000
|
| 4238 |
+
},
|
| 4239 |
+
{
|
| 4240 |
+
"epoch": 4.93,
|
| 4241 |
+
"learning_rate": 3.391796206987249e-05,
|
| 4242 |
+
"loss": 0.3907,
|
| 4243 |
+
"step": 349500
|
| 4244 |
+
},
|
| 4245 |
+
{
|
| 4246 |
+
"epoch": 4.94,
|
| 4247 |
+
"learning_rate": 3.389427117012155e-05,
|
| 4248 |
+
"loss": 0.3947,
|
| 4249 |
+
"step": 350000
|
| 4250 |
+
},
|
| 4251 |
+
{
|
| 4252 |
+
"epoch": 4.94,
|
| 4253 |
+
"learning_rate": 3.387062774712362e-05,
|
| 4254 |
+
"loss": 0.3962,
|
| 4255 |
+
"step": 350500
|
| 4256 |
+
},
|
| 4257 |
+
{
|
| 4258 |
+
"epoch": 4.95,
|
| 4259 |
+
"learning_rate": 3.384688937061968e-05,
|
| 4260 |
+
"loss": 0.3961,
|
| 4261 |
+
"step": 351000
|
| 4262 |
+
},
|
| 4263 |
+
{
|
| 4264 |
+
"epoch": 4.96,
|
| 4265 |
+
"learning_rate": 3.3823150994115734e-05,
|
| 4266 |
+
"loss": 0.4004,
|
| 4267 |
+
"step": 351500
|
| 4268 |
+
},
|
| 4269 |
+
{
|
| 4270 |
+
"epoch": 4.96,
|
| 4271 |
+
"learning_rate": 3.379941261761179e-05,
|
| 4272 |
+
"loss": 0.3917,
|
| 4273 |
+
"step": 352000
|
| 4274 |
+
},
|
| 4275 |
+
{
|
| 4276 |
+
"epoch": 4.97,
|
| 4277 |
+
"learning_rate": 3.3775721717860856e-05,
|
| 4278 |
+
"loss": 0.3939,
|
| 4279 |
+
"step": 352500
|
| 4280 |
+
},
|
| 4281 |
+
{
|
| 4282 |
+
"epoch": 4.98,
|
| 4283 |
+
"learning_rate": 3.3752030818109916e-05,
|
| 4284 |
+
"loss": 0.3942,
|
| 4285 |
+
"step": 353000
|
| 4286 |
+
},
|
| 4287 |
+
{
|
| 4288 |
+
"epoch": 4.98,
|
| 4289 |
+
"learning_rate": 3.372829244160597e-05,
|
| 4290 |
+
"loss": 0.3975,
|
| 4291 |
+
"step": 353500
|
| 4292 |
+
},
|
| 4293 |
+
{
|
| 4294 |
+
"epoch": 4.99,
|
| 4295 |
+
"learning_rate": 3.370455406510202e-05,
|
| 4296 |
+
"loss": 0.4001,
|
| 4297 |
+
"step": 354000
|
| 4298 |
+
},
|
| 4299 |
+
{
|
| 4300 |
+
"epoch": 5.0,
|
| 4301 |
+
"learning_rate": 3.3680815688598076e-05,
|
| 4302 |
+
"loss": 0.388,
|
| 4303 |
+
"step": 354500
|
| 4304 |
+
},
|
| 4305 |
+
{
|
| 4306 |
+
"epoch": 5.0,
|
| 4307 |
+
"eval_bleu": 43.3763,
|
| 4308 |
+
"eval_gen_len": 13.6067,
|
| 4309 |
+
"eval_loss": 1.0399531126022339,
|
| 4310 |
+
"eval_runtime": 9774.1481,
|
| 4311 |
+
"eval_samples_per_second": 14.512,
|
| 4312 |
+
"eval_steps_per_second": 1.814,
|
| 4313 |
+
"step": 354597
|
| 4314 |
+
}
|
| 4315 |
+
],
|
| 4316 |
+
"logging_steps": 500,
|
| 4317 |
+
"max_steps": 1063785,
|
| 4318 |
+
"num_train_epochs": 15,
|
| 4319 |
+
"save_steps": 500,
|
| 4320 |
+
"total_flos": 1.2295340767110496e+19,
|
| 4321 |
+
"trial_name": null,
|
| 4322 |
+
"trial_params": null
|
| 4323 |
+
}
|
checkpoint-354597/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0adf1a980c6128833811b7e6eb546e117ffd3efb8c21dc7de95b5e76a5b21b8d
|
| 3 |
+
size 4728
|
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": "4.35.2",
|
| 57 |
+
"use_cache": true,
|
| 58 |
+
"vocab_size": 250054
|
| 59 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"eval_bleu": 42.7907,
|
| 4 |
+
"eval_gen_len": 13.7941,
|
| 5 |
+
"eval_loss": 0.8987888693809509,
|
| 6 |
+
"eval_runtime": 10014.5277,
|
| 7 |
+
"eval_samples": 141838,
|
| 8 |
+
"eval_samples_per_second": 14.163,
|
| 9 |
+
"eval_steps_per_second": 1.77
|
| 10 |
+
}
|