Upload 12 files
Browse files- README.md +56 -0
- added_tokens.json +3 -0
- all_results.json +14 -0
- bpe.codes +0 -0
- config.json +43 -0
- eval_results.json +9 -0
- special_tokens_map.json +9 -0
- tokenizer_config.json +54 -0
- train_results.json +8 -0
- trainer_state.json +1816 -0
- training_args.bin +3 -0
- vocab.txt +0 -0
README.md
ADDED
|
@@ -0,0 +1,56 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: vinai/phobert-base
|
| 3 |
+
tags:
|
| 4 |
+
- generated_from_trainer
|
| 5 |
+
metrics:
|
| 6 |
+
- accuracy
|
| 7 |
+
model-index:
|
| 8 |
+
- name: save
|
| 9 |
+
results: []
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 13 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 14 |
+
|
| 15 |
+
# save
|
| 16 |
+
|
| 17 |
+
This model is a fine-tuned version of [vinai/phobert-base](https://huggingface.co/vinai/phobert-base) on an unknown dataset.
|
| 18 |
+
It achieves the following results on the evaluation set:
|
| 19 |
+
- Loss: 0.2260
|
| 20 |
+
- Accuracy: 0.9756
|
| 21 |
+
|
| 22 |
+
## Model description
|
| 23 |
+
|
| 24 |
+
More information needed
|
| 25 |
+
|
| 26 |
+
## Intended uses & limitations
|
| 27 |
+
|
| 28 |
+
More information needed
|
| 29 |
+
|
| 30 |
+
## Training and evaluation data
|
| 31 |
+
|
| 32 |
+
More information needed
|
| 33 |
+
|
| 34 |
+
## Training procedure
|
| 35 |
+
|
| 36 |
+
### Training hyperparameters
|
| 37 |
+
|
| 38 |
+
The following hyperparameters were used during training:
|
| 39 |
+
- learning_rate: 2e-05
|
| 40 |
+
- train_batch_size: 1
|
| 41 |
+
- eval_batch_size: 8
|
| 42 |
+
- seed: 42
|
| 43 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 44 |
+
- lr_scheduler_type: linear
|
| 45 |
+
- num_epochs: 10.0
|
| 46 |
+
|
| 47 |
+
### Training results
|
| 48 |
+
|
| 49 |
+
|
| 50 |
+
|
| 51 |
+
### Framework versions
|
| 52 |
+
|
| 53 |
+
- Transformers 4.35.0
|
| 54 |
+
- Pytorch 2.1.0+cu118
|
| 55 |
+
- Datasets 2.14.6
|
| 56 |
+
- Tokenizers 0.14.1
|
added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
all_results.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"eval_accuracy": 0.9755901287553648,
|
| 4 |
+
"eval_loss": 0.22602945566177368,
|
| 5 |
+
"eval_runtime": 10.7799,
|
| 6 |
+
"eval_samples": 3728,
|
| 7 |
+
"eval_samples_per_second": 345.828,
|
| 8 |
+
"eval_steps_per_second": 43.229,
|
| 9 |
+
"train_loss": 0.1504479006291696,
|
| 10 |
+
"train_runtime": 9647.5503,
|
| 11 |
+
"train_samples": 14911,
|
| 12 |
+
"train_samples_per_second": 15.456,
|
| 13 |
+
"train_steps_per_second": 15.456
|
| 14 |
+
}
|
bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
config.json
ADDED
|
@@ -0,0 +1,43 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "vinai/phobert-base",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"RobertaForSequenceClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"eos_token_id": 2,
|
| 10 |
+
"finetuning_task": "text-classification",
|
| 11 |
+
"gradient_checkpointing": false,
|
| 12 |
+
"hidden_act": "gelu",
|
| 13 |
+
"hidden_dropout_prob": 0.1,
|
| 14 |
+
"hidden_size": 768,
|
| 15 |
+
"id2label": {
|
| 16 |
+
"0": "AGENCY",
|
| 17 |
+
"1": "PROTENTIAL",
|
| 18 |
+
"2": "SHARING",
|
| 19 |
+
"3": "SPAM"
|
| 20 |
+
},
|
| 21 |
+
"initializer_range": 0.02,
|
| 22 |
+
"intermediate_size": 3072,
|
| 23 |
+
"label2id": {
|
| 24 |
+
"AGENCY": 0,
|
| 25 |
+
"PROTENTIAL": 1,
|
| 26 |
+
"SHARING": 2,
|
| 27 |
+
"SPAM": 3
|
| 28 |
+
},
|
| 29 |
+
"layer_norm_eps": 1e-05,
|
| 30 |
+
"max_position_embeddings": 258,
|
| 31 |
+
"model_type": "roberta",
|
| 32 |
+
"num_attention_heads": 12,
|
| 33 |
+
"num_hidden_layers": 12,
|
| 34 |
+
"pad_token_id": 1,
|
| 35 |
+
"position_embedding_type": "absolute",
|
| 36 |
+
"problem_type": "single_label_classification",
|
| 37 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 38 |
+
"torch_dtype": "float32",
|
| 39 |
+
"transformers_version": "4.35.0",
|
| 40 |
+
"type_vocab_size": 1,
|
| 41 |
+
"use_cache": true,
|
| 42 |
+
"vocab_size": 64001
|
| 43 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"eval_accuracy": 0.9755901287553648,
|
| 4 |
+
"eval_loss": 0.22602945566177368,
|
| 5 |
+
"eval_runtime": 10.7799,
|
| 6 |
+
"eval_samples": 3728,
|
| 7 |
+
"eval_samples_per_second": 345.828,
|
| 8 |
+
"eval_steps_per_second": 43.229
|
| 9 |
+
}
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,54 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": true,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"mask_token": "<mask>",
|
| 49 |
+
"model_max_length": 256,
|
| 50 |
+
"pad_token": "<pad>",
|
| 51 |
+
"sep_token": "</s>",
|
| 52 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 53 |
+
"unk_token": "<unk>"
|
| 54 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 10.0,
|
| 3 |
+
"train_loss": 0.1504479006291696,
|
| 4 |
+
"train_runtime": 9647.5503,
|
| 5 |
+
"train_samples": 14911,
|
| 6 |
+
"train_samples_per_second": 15.456,
|
| 7 |
+
"train_steps_per_second": 15.456
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1816 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 10.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 149110,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.03,
|
| 13 |
+
"learning_rate": 1.9932935416806386e-05,
|
| 14 |
+
"loss": 0.7069,
|
| 15 |
+
"step": 500
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.07,
|
| 19 |
+
"learning_rate": 1.986587083361277e-05,
|
| 20 |
+
"loss": 0.4778,
|
| 21 |
+
"step": 1000
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.1,
|
| 25 |
+
"learning_rate": 1.9798806250419156e-05,
|
| 26 |
+
"loss": 0.4614,
|
| 27 |
+
"step": 1500
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.13,
|
| 31 |
+
"learning_rate": 1.973174166722554e-05,
|
| 32 |
+
"loss": 0.395,
|
| 33 |
+
"step": 2000
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.17,
|
| 37 |
+
"learning_rate": 1.9664677084031925e-05,
|
| 38 |
+
"loss": 0.3459,
|
| 39 |
+
"step": 2500
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.2,
|
| 43 |
+
"learning_rate": 1.9597612500838306e-05,
|
| 44 |
+
"loss": 0.3942,
|
| 45 |
+
"step": 3000
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.23,
|
| 49 |
+
"learning_rate": 1.9530547917644694e-05,
|
| 50 |
+
"loss": 0.5045,
|
| 51 |
+
"step": 3500
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.27,
|
| 55 |
+
"learning_rate": 1.946348333445108e-05,
|
| 56 |
+
"loss": 0.5222,
|
| 57 |
+
"step": 4000
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.3,
|
| 61 |
+
"learning_rate": 1.9396418751257464e-05,
|
| 62 |
+
"loss": 0.6322,
|
| 63 |
+
"step": 4500
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.34,
|
| 67 |
+
"learning_rate": 1.9329354168063848e-05,
|
| 68 |
+
"loss": 0.4139,
|
| 69 |
+
"step": 5000
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.37,
|
| 73 |
+
"learning_rate": 1.926228958487023e-05,
|
| 74 |
+
"loss": 0.3958,
|
| 75 |
+
"step": 5500
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.4,
|
| 79 |
+
"learning_rate": 1.9195225001676618e-05,
|
| 80 |
+
"loss": 0.3241,
|
| 81 |
+
"step": 6000
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.44,
|
| 85 |
+
"learning_rate": 1.9128160418483002e-05,
|
| 86 |
+
"loss": 0.3645,
|
| 87 |
+
"step": 6500
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.47,
|
| 91 |
+
"learning_rate": 1.9061095835289383e-05,
|
| 92 |
+
"loss": 0.4296,
|
| 93 |
+
"step": 7000
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.5,
|
| 97 |
+
"learning_rate": 1.899403125209577e-05,
|
| 98 |
+
"loss": 0.4467,
|
| 99 |
+
"step": 7500
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.54,
|
| 103 |
+
"learning_rate": 1.8926966668902153e-05,
|
| 104 |
+
"loss": 0.499,
|
| 105 |
+
"step": 8000
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.57,
|
| 109 |
+
"learning_rate": 1.8859902085708537e-05,
|
| 110 |
+
"loss": 0.3124,
|
| 111 |
+
"step": 8500
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.6,
|
| 115 |
+
"learning_rate": 1.8792837502514922e-05,
|
| 116 |
+
"loss": 0.4169,
|
| 117 |
+
"step": 9000
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.64,
|
| 121 |
+
"learning_rate": 1.8725772919321307e-05,
|
| 122 |
+
"loss": 0.4071,
|
| 123 |
+
"step": 9500
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.67,
|
| 127 |
+
"learning_rate": 1.8658708336127695e-05,
|
| 128 |
+
"loss": 0.2717,
|
| 129 |
+
"step": 10000
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.7,
|
| 133 |
+
"learning_rate": 1.8591643752934076e-05,
|
| 134 |
+
"loss": 0.3581,
|
| 135 |
+
"step": 10500
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.74,
|
| 139 |
+
"learning_rate": 1.852457916974046e-05,
|
| 140 |
+
"loss": 0.3181,
|
| 141 |
+
"step": 11000
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.77,
|
| 145 |
+
"learning_rate": 1.8457514586546845e-05,
|
| 146 |
+
"loss": 0.3488,
|
| 147 |
+
"step": 11500
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.8,
|
| 151 |
+
"learning_rate": 1.839045000335323e-05,
|
| 152 |
+
"loss": 0.3342,
|
| 153 |
+
"step": 12000
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.84,
|
| 157 |
+
"learning_rate": 1.8323385420159615e-05,
|
| 158 |
+
"loss": 0.4063,
|
| 159 |
+
"step": 12500
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.87,
|
| 163 |
+
"learning_rate": 1.8256320836966e-05,
|
| 164 |
+
"loss": 0.3071,
|
| 165 |
+
"step": 13000
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.91,
|
| 169 |
+
"learning_rate": 1.8189256253772384e-05,
|
| 170 |
+
"loss": 0.2998,
|
| 171 |
+
"step": 13500
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.94,
|
| 175 |
+
"learning_rate": 1.812219167057877e-05,
|
| 176 |
+
"loss": 0.2989,
|
| 177 |
+
"step": 14000
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.97,
|
| 181 |
+
"learning_rate": 1.8055127087385153e-05,
|
| 182 |
+
"loss": 0.2705,
|
| 183 |
+
"step": 14500
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 1.01,
|
| 187 |
+
"learning_rate": 1.7988062504191538e-05,
|
| 188 |
+
"loss": 0.3,
|
| 189 |
+
"step": 15000
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 1.04,
|
| 193 |
+
"learning_rate": 1.7920997920997923e-05,
|
| 194 |
+
"loss": 0.3253,
|
| 195 |
+
"step": 15500
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 1.07,
|
| 199 |
+
"learning_rate": 1.7853933337804307e-05,
|
| 200 |
+
"loss": 0.2612,
|
| 201 |
+
"step": 16000
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 1.11,
|
| 205 |
+
"learning_rate": 1.7786868754610692e-05,
|
| 206 |
+
"loss": 0.3027,
|
| 207 |
+
"step": 16500
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 1.14,
|
| 211 |
+
"learning_rate": 1.7719804171417077e-05,
|
| 212 |
+
"loss": 0.3124,
|
| 213 |
+
"step": 17000
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 1.17,
|
| 217 |
+
"learning_rate": 1.765273958822346e-05,
|
| 218 |
+
"loss": 0.2997,
|
| 219 |
+
"step": 17500
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 1.21,
|
| 223 |
+
"learning_rate": 1.7585675005029846e-05,
|
| 224 |
+
"loss": 0.3177,
|
| 225 |
+
"step": 18000
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 1.24,
|
| 229 |
+
"learning_rate": 1.751861042183623e-05,
|
| 230 |
+
"loss": 0.3636,
|
| 231 |
+
"step": 18500
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 1.27,
|
| 235 |
+
"learning_rate": 1.7451545838642615e-05,
|
| 236 |
+
"loss": 0.3025,
|
| 237 |
+
"step": 19000
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 1.31,
|
| 241 |
+
"learning_rate": 1.7384481255449e-05,
|
| 242 |
+
"loss": 0.4098,
|
| 243 |
+
"step": 19500
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 1.34,
|
| 247 |
+
"learning_rate": 1.731741667225538e-05,
|
| 248 |
+
"loss": 0.3093,
|
| 249 |
+
"step": 20000
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 1.37,
|
| 253 |
+
"learning_rate": 1.725035208906177e-05,
|
| 254 |
+
"loss": 0.3261,
|
| 255 |
+
"step": 20500
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 1.41,
|
| 259 |
+
"learning_rate": 1.718328750586815e-05,
|
| 260 |
+
"loss": 0.2338,
|
| 261 |
+
"step": 21000
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 1.44,
|
| 265 |
+
"learning_rate": 1.711622292267454e-05,
|
| 266 |
+
"loss": 0.2204,
|
| 267 |
+
"step": 21500
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 1.48,
|
| 271 |
+
"learning_rate": 1.7049158339480923e-05,
|
| 272 |
+
"loss": 0.2963,
|
| 273 |
+
"step": 22000
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 1.51,
|
| 277 |
+
"learning_rate": 1.6982093756287304e-05,
|
| 278 |
+
"loss": 0.1368,
|
| 279 |
+
"step": 22500
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 1.54,
|
| 283 |
+
"learning_rate": 1.6915029173093693e-05,
|
| 284 |
+
"loss": 0.307,
|
| 285 |
+
"step": 23000
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 1.58,
|
| 289 |
+
"learning_rate": 1.6847964589900074e-05,
|
| 290 |
+
"loss": 0.2184,
|
| 291 |
+
"step": 23500
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 1.61,
|
| 295 |
+
"learning_rate": 1.678090000670646e-05,
|
| 296 |
+
"loss": 0.2967,
|
| 297 |
+
"step": 24000
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 1.64,
|
| 301 |
+
"learning_rate": 1.6713835423512847e-05,
|
| 302 |
+
"loss": 0.2879,
|
| 303 |
+
"step": 24500
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 1.68,
|
| 307 |
+
"learning_rate": 1.6646770840319228e-05,
|
| 308 |
+
"loss": 0.2899,
|
| 309 |
+
"step": 25000
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 1.71,
|
| 313 |
+
"learning_rate": 1.6579706257125612e-05,
|
| 314 |
+
"loss": 0.2182,
|
| 315 |
+
"step": 25500
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 1.74,
|
| 319 |
+
"learning_rate": 1.6512641673931997e-05,
|
| 320 |
+
"loss": 0.1737,
|
| 321 |
+
"step": 26000
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 1.78,
|
| 325 |
+
"learning_rate": 1.6445577090738382e-05,
|
| 326 |
+
"loss": 0.2756,
|
| 327 |
+
"step": 26500
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 1.81,
|
| 331 |
+
"learning_rate": 1.6378512507544766e-05,
|
| 332 |
+
"loss": 0.284,
|
| 333 |
+
"step": 27000
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 1.84,
|
| 337 |
+
"learning_rate": 1.631144792435115e-05,
|
| 338 |
+
"loss": 0.2672,
|
| 339 |
+
"step": 27500
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 1.88,
|
| 343 |
+
"learning_rate": 1.6244383341157536e-05,
|
| 344 |
+
"loss": 0.2016,
|
| 345 |
+
"step": 28000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 1.91,
|
| 349 |
+
"learning_rate": 1.617731875796392e-05,
|
| 350 |
+
"loss": 0.2257,
|
| 351 |
+
"step": 28500
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 1.94,
|
| 355 |
+
"learning_rate": 1.6110254174770305e-05,
|
| 356 |
+
"loss": 0.3434,
|
| 357 |
+
"step": 29000
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 1.98,
|
| 361 |
+
"learning_rate": 1.604318959157669e-05,
|
| 362 |
+
"loss": 0.1699,
|
| 363 |
+
"step": 29500
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 2.01,
|
| 367 |
+
"learning_rate": 1.5976125008383074e-05,
|
| 368 |
+
"loss": 0.2511,
|
| 369 |
+
"step": 30000
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 2.05,
|
| 373 |
+
"learning_rate": 1.590906042518946e-05,
|
| 374 |
+
"loss": 0.2026,
|
| 375 |
+
"step": 30500
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 2.08,
|
| 379 |
+
"learning_rate": 1.5841995841995844e-05,
|
| 380 |
+
"loss": 0.2522,
|
| 381 |
+
"step": 31000
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 2.11,
|
| 385 |
+
"learning_rate": 1.577493125880223e-05,
|
| 386 |
+
"loss": 0.2572,
|
| 387 |
+
"step": 31500
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 2.15,
|
| 391 |
+
"learning_rate": 1.5707866675608613e-05,
|
| 392 |
+
"loss": 0.2237,
|
| 393 |
+
"step": 32000
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 2.18,
|
| 397 |
+
"learning_rate": 1.5640802092414998e-05,
|
| 398 |
+
"loss": 0.195,
|
| 399 |
+
"step": 32500
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 2.21,
|
| 403 |
+
"learning_rate": 1.557373750922138e-05,
|
| 404 |
+
"loss": 0.218,
|
| 405 |
+
"step": 33000
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 2.25,
|
| 409 |
+
"learning_rate": 1.5506672926027767e-05,
|
| 410 |
+
"loss": 0.1463,
|
| 411 |
+
"step": 33500
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 2.28,
|
| 415 |
+
"learning_rate": 1.543960834283415e-05,
|
| 416 |
+
"loss": 0.1723,
|
| 417 |
+
"step": 34000
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 2.31,
|
| 421 |
+
"learning_rate": 1.5372543759640536e-05,
|
| 422 |
+
"loss": 0.1874,
|
| 423 |
+
"step": 34500
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 2.35,
|
| 427 |
+
"learning_rate": 1.530547917644692e-05,
|
| 428 |
+
"loss": 0.2871,
|
| 429 |
+
"step": 35000
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 2.38,
|
| 433 |
+
"learning_rate": 1.5238414593253304e-05,
|
| 434 |
+
"loss": 0.3132,
|
| 435 |
+
"step": 35500
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 2.41,
|
| 439 |
+
"learning_rate": 1.5171350010059689e-05,
|
| 440 |
+
"loss": 0.2591,
|
| 441 |
+
"step": 36000
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 2.45,
|
| 445 |
+
"learning_rate": 1.5104285426866075e-05,
|
| 446 |
+
"loss": 0.1916,
|
| 447 |
+
"step": 36500
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 2.48,
|
| 451 |
+
"learning_rate": 1.5037220843672458e-05,
|
| 452 |
+
"loss": 0.2998,
|
| 453 |
+
"step": 37000
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 2.51,
|
| 457 |
+
"learning_rate": 1.4970156260478843e-05,
|
| 458 |
+
"loss": 0.1397,
|
| 459 |
+
"step": 37500
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 2.55,
|
| 463 |
+
"learning_rate": 1.4903091677285225e-05,
|
| 464 |
+
"loss": 0.2588,
|
| 465 |
+
"step": 38000
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 2.58,
|
| 469 |
+
"learning_rate": 1.4836027094091612e-05,
|
| 470 |
+
"loss": 0.2526,
|
| 471 |
+
"step": 38500
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 2.62,
|
| 475 |
+
"learning_rate": 1.4768962510897995e-05,
|
| 476 |
+
"loss": 0.1548,
|
| 477 |
+
"step": 39000
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 2.65,
|
| 481 |
+
"learning_rate": 1.4701897927704381e-05,
|
| 482 |
+
"loss": 0.2322,
|
| 483 |
+
"step": 39500
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 2.68,
|
| 487 |
+
"learning_rate": 1.4634833344510766e-05,
|
| 488 |
+
"loss": 0.1839,
|
| 489 |
+
"step": 40000
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 2.72,
|
| 493 |
+
"learning_rate": 1.4567768761317149e-05,
|
| 494 |
+
"loss": 0.1062,
|
| 495 |
+
"step": 40500
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 2.75,
|
| 499 |
+
"learning_rate": 1.4500704178123535e-05,
|
| 500 |
+
"loss": 0.1758,
|
| 501 |
+
"step": 41000
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 2.78,
|
| 505 |
+
"learning_rate": 1.4433639594929918e-05,
|
| 506 |
+
"loss": 0.2261,
|
| 507 |
+
"step": 41500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 2.82,
|
| 511 |
+
"learning_rate": 1.4366575011736303e-05,
|
| 512 |
+
"loss": 0.1591,
|
| 513 |
+
"step": 42000
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 2.85,
|
| 517 |
+
"learning_rate": 1.4299510428542687e-05,
|
| 518 |
+
"loss": 0.2023,
|
| 519 |
+
"step": 42500
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 2.88,
|
| 523 |
+
"learning_rate": 1.4232445845349072e-05,
|
| 524 |
+
"loss": 0.2378,
|
| 525 |
+
"step": 43000
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 2.92,
|
| 529 |
+
"learning_rate": 1.4165381262155458e-05,
|
| 530 |
+
"loss": 0.1869,
|
| 531 |
+
"step": 43500
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 2.95,
|
| 535 |
+
"learning_rate": 1.4098316678961841e-05,
|
| 536 |
+
"loss": 0.2229,
|
| 537 |
+
"step": 44000
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 2.98,
|
| 541 |
+
"learning_rate": 1.4031252095768226e-05,
|
| 542 |
+
"loss": 0.1752,
|
| 543 |
+
"step": 44500
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 3.02,
|
| 547 |
+
"learning_rate": 1.3964187512574609e-05,
|
| 548 |
+
"loss": 0.2121,
|
| 549 |
+
"step": 45000
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 3.05,
|
| 553 |
+
"learning_rate": 1.3897122929380995e-05,
|
| 554 |
+
"loss": 0.1928,
|
| 555 |
+
"step": 45500
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 3.08,
|
| 559 |
+
"learning_rate": 1.383005834618738e-05,
|
| 560 |
+
"loss": 0.0827,
|
| 561 |
+
"step": 46000
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 3.12,
|
| 565 |
+
"learning_rate": 1.3762993762993763e-05,
|
| 566 |
+
"loss": 0.2259,
|
| 567 |
+
"step": 46500
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 3.15,
|
| 571 |
+
"learning_rate": 1.369592917980015e-05,
|
| 572 |
+
"loss": 0.1447,
|
| 573 |
+
"step": 47000
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 3.19,
|
| 577 |
+
"learning_rate": 1.3628864596606532e-05,
|
| 578 |
+
"loss": 0.1481,
|
| 579 |
+
"step": 47500
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 3.22,
|
| 583 |
+
"learning_rate": 1.3561800013412919e-05,
|
| 584 |
+
"loss": 0.2343,
|
| 585 |
+
"step": 48000
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 3.25,
|
| 589 |
+
"learning_rate": 1.3494735430219302e-05,
|
| 590 |
+
"loss": 0.1454,
|
| 591 |
+
"step": 48500
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 3.29,
|
| 595 |
+
"learning_rate": 1.3427670847025686e-05,
|
| 596 |
+
"loss": 0.1152,
|
| 597 |
+
"step": 49000
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 3.32,
|
| 601 |
+
"learning_rate": 1.3360606263832073e-05,
|
| 602 |
+
"loss": 0.2361,
|
| 603 |
+
"step": 49500
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 3.35,
|
| 607 |
+
"learning_rate": 1.3293541680638456e-05,
|
| 608 |
+
"loss": 0.091,
|
| 609 |
+
"step": 50000
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 3.39,
|
| 613 |
+
"learning_rate": 1.322647709744484e-05,
|
| 614 |
+
"loss": 0.1912,
|
| 615 |
+
"step": 50500
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 3.42,
|
| 619 |
+
"learning_rate": 1.3159412514251225e-05,
|
| 620 |
+
"loss": 0.1125,
|
| 621 |
+
"step": 51000
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 3.45,
|
| 625 |
+
"learning_rate": 1.309234793105761e-05,
|
| 626 |
+
"loss": 0.1295,
|
| 627 |
+
"step": 51500
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 3.49,
|
| 631 |
+
"learning_rate": 1.3025283347863994e-05,
|
| 632 |
+
"loss": 0.106,
|
| 633 |
+
"step": 52000
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 3.52,
|
| 637 |
+
"learning_rate": 1.2958218764670379e-05,
|
| 638 |
+
"loss": 0.0838,
|
| 639 |
+
"step": 52500
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 3.55,
|
| 643 |
+
"learning_rate": 1.2891154181476764e-05,
|
| 644 |
+
"loss": 0.1616,
|
| 645 |
+
"step": 53000
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 3.59,
|
| 649 |
+
"learning_rate": 1.2824089598283147e-05,
|
| 650 |
+
"loss": 0.1084,
|
| 651 |
+
"step": 53500
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 3.62,
|
| 655 |
+
"learning_rate": 1.2757025015089533e-05,
|
| 656 |
+
"loss": 0.2535,
|
| 657 |
+
"step": 54000
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 3.66,
|
| 661 |
+
"learning_rate": 1.2689960431895916e-05,
|
| 662 |
+
"loss": 0.1363,
|
| 663 |
+
"step": 54500
|
| 664 |
+
},
|
| 665 |
+
{
|
| 666 |
+
"epoch": 3.69,
|
| 667 |
+
"learning_rate": 1.26228958487023e-05,
|
| 668 |
+
"loss": 0.2317,
|
| 669 |
+
"step": 55000
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 3.72,
|
| 673 |
+
"learning_rate": 1.2555831265508687e-05,
|
| 674 |
+
"loss": 0.2266,
|
| 675 |
+
"step": 55500
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 3.76,
|
| 679 |
+
"learning_rate": 1.248876668231507e-05,
|
| 680 |
+
"loss": 0.1864,
|
| 681 |
+
"step": 56000
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 3.79,
|
| 685 |
+
"learning_rate": 1.2421702099121456e-05,
|
| 686 |
+
"loss": 0.0946,
|
| 687 |
+
"step": 56500
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 3.82,
|
| 691 |
+
"learning_rate": 1.2354637515927839e-05,
|
| 692 |
+
"loss": 0.1919,
|
| 693 |
+
"step": 57000
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 3.86,
|
| 697 |
+
"learning_rate": 1.2287572932734224e-05,
|
| 698 |
+
"loss": 0.1106,
|
| 699 |
+
"step": 57500
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 3.89,
|
| 703 |
+
"learning_rate": 1.222050834954061e-05,
|
| 704 |
+
"loss": 0.1769,
|
| 705 |
+
"step": 58000
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 3.92,
|
| 709 |
+
"learning_rate": 1.2153443766346993e-05,
|
| 710 |
+
"loss": 0.089,
|
| 711 |
+
"step": 58500
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 3.96,
|
| 715 |
+
"learning_rate": 1.2086379183153378e-05,
|
| 716 |
+
"loss": 0.1087,
|
| 717 |
+
"step": 59000
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 3.99,
|
| 721 |
+
"learning_rate": 1.2019314599959762e-05,
|
| 722 |
+
"loss": 0.1879,
|
| 723 |
+
"step": 59500
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 4.02,
|
| 727 |
+
"learning_rate": 1.1952250016766147e-05,
|
| 728 |
+
"loss": 0.14,
|
| 729 |
+
"step": 60000
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 4.06,
|
| 733 |
+
"learning_rate": 1.188518543357253e-05,
|
| 734 |
+
"loss": 0.1331,
|
| 735 |
+
"step": 60500
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 4.09,
|
| 739 |
+
"learning_rate": 1.1818120850378916e-05,
|
| 740 |
+
"loss": 0.0841,
|
| 741 |
+
"step": 61000
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 4.12,
|
| 745 |
+
"learning_rate": 1.1751056267185301e-05,
|
| 746 |
+
"loss": 0.134,
|
| 747 |
+
"step": 61500
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 4.16,
|
| 751 |
+
"learning_rate": 1.1683991683991684e-05,
|
| 752 |
+
"loss": 0.1823,
|
| 753 |
+
"step": 62000
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 4.19,
|
| 757 |
+
"learning_rate": 1.161692710079807e-05,
|
| 758 |
+
"loss": 0.1005,
|
| 759 |
+
"step": 62500
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 4.23,
|
| 763 |
+
"learning_rate": 1.1549862517604453e-05,
|
| 764 |
+
"loss": 0.1003,
|
| 765 |
+
"step": 63000
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 4.26,
|
| 769 |
+
"learning_rate": 1.1482797934410838e-05,
|
| 770 |
+
"loss": 0.1729,
|
| 771 |
+
"step": 63500
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 4.29,
|
| 775 |
+
"learning_rate": 1.1415733351217224e-05,
|
| 776 |
+
"loss": 0.155,
|
| 777 |
+
"step": 64000
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 4.33,
|
| 781 |
+
"learning_rate": 1.1348668768023607e-05,
|
| 782 |
+
"loss": 0.1039,
|
| 783 |
+
"step": 64500
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 4.36,
|
| 787 |
+
"learning_rate": 1.1281604184829994e-05,
|
| 788 |
+
"loss": 0.1536,
|
| 789 |
+
"step": 65000
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 4.39,
|
| 793 |
+
"learning_rate": 1.1214539601636377e-05,
|
| 794 |
+
"loss": 0.1576,
|
| 795 |
+
"step": 65500
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 4.43,
|
| 799 |
+
"learning_rate": 1.1147475018442761e-05,
|
| 800 |
+
"loss": 0.1298,
|
| 801 |
+
"step": 66000
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 4.46,
|
| 805 |
+
"learning_rate": 1.1080410435249144e-05,
|
| 806 |
+
"loss": 0.1465,
|
| 807 |
+
"step": 66500
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 4.49,
|
| 811 |
+
"learning_rate": 1.101334585205553e-05,
|
| 812 |
+
"loss": 0.1322,
|
| 813 |
+
"step": 67000
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 4.53,
|
| 817 |
+
"learning_rate": 1.0946281268861915e-05,
|
| 818 |
+
"loss": 0.1322,
|
| 819 |
+
"step": 67500
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 4.56,
|
| 823 |
+
"learning_rate": 1.0879216685668298e-05,
|
| 824 |
+
"loss": 0.1597,
|
| 825 |
+
"step": 68000
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 4.59,
|
| 829 |
+
"learning_rate": 1.0812152102474685e-05,
|
| 830 |
+
"loss": 0.1188,
|
| 831 |
+
"step": 68500
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 4.63,
|
| 835 |
+
"learning_rate": 1.0745087519281068e-05,
|
| 836 |
+
"loss": 0.1614,
|
| 837 |
+
"step": 69000
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 4.66,
|
| 841 |
+
"learning_rate": 1.0678022936087454e-05,
|
| 842 |
+
"loss": 0.1174,
|
| 843 |
+
"step": 69500
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 4.69,
|
| 847 |
+
"learning_rate": 1.0610958352893839e-05,
|
| 848 |
+
"loss": 0.1336,
|
| 849 |
+
"step": 70000
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 4.73,
|
| 853 |
+
"learning_rate": 1.0543893769700222e-05,
|
| 854 |
+
"loss": 0.0941,
|
| 855 |
+
"step": 70500
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 4.76,
|
| 859 |
+
"learning_rate": 1.0476829186506608e-05,
|
| 860 |
+
"loss": 0.0724,
|
| 861 |
+
"step": 71000
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 4.8,
|
| 865 |
+
"learning_rate": 1.040976460331299e-05,
|
| 866 |
+
"loss": 0.1419,
|
| 867 |
+
"step": 71500
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 4.83,
|
| 871 |
+
"learning_rate": 1.0342700020119375e-05,
|
| 872 |
+
"loss": 0.1308,
|
| 873 |
+
"step": 72000
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 4.86,
|
| 877 |
+
"learning_rate": 1.027563543692576e-05,
|
| 878 |
+
"loss": 0.1128,
|
| 879 |
+
"step": 72500
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 4.9,
|
| 883 |
+
"learning_rate": 1.0208570853732145e-05,
|
| 884 |
+
"loss": 0.0805,
|
| 885 |
+
"step": 73000
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 4.93,
|
| 889 |
+
"learning_rate": 1.0141506270538531e-05,
|
| 890 |
+
"loss": 0.0361,
|
| 891 |
+
"step": 73500
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 4.96,
|
| 895 |
+
"learning_rate": 1.0074441687344914e-05,
|
| 896 |
+
"loss": 0.1846,
|
| 897 |
+
"step": 74000
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 5.0,
|
| 901 |
+
"learning_rate": 1.0007377104151299e-05,
|
| 902 |
+
"loss": 0.0949,
|
| 903 |
+
"step": 74500
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 5.03,
|
| 907 |
+
"learning_rate": 9.940312520957683e-06,
|
| 908 |
+
"loss": 0.1941,
|
| 909 |
+
"step": 75000
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 5.06,
|
| 913 |
+
"learning_rate": 9.873247937764068e-06,
|
| 914 |
+
"loss": 0.154,
|
| 915 |
+
"step": 75500
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 5.1,
|
| 919 |
+
"learning_rate": 9.806183354570453e-06,
|
| 920 |
+
"loss": 0.1002,
|
| 921 |
+
"step": 76000
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 5.13,
|
| 925 |
+
"learning_rate": 9.739118771376836e-06,
|
| 926 |
+
"loss": 0.0869,
|
| 927 |
+
"step": 76500
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 5.16,
|
| 931 |
+
"learning_rate": 9.67205418818322e-06,
|
| 932 |
+
"loss": 0.0587,
|
| 933 |
+
"step": 77000
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 5.2,
|
| 937 |
+
"learning_rate": 9.604989604989607e-06,
|
| 938 |
+
"loss": 0.1265,
|
| 939 |
+
"step": 77500
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 5.23,
|
| 943 |
+
"learning_rate": 9.537925021795991e-06,
|
| 944 |
+
"loss": 0.1247,
|
| 945 |
+
"step": 78000
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 5.26,
|
| 949 |
+
"learning_rate": 9.470860438602374e-06,
|
| 950 |
+
"loss": 0.0767,
|
| 951 |
+
"step": 78500
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 5.3,
|
| 955 |
+
"learning_rate": 9.403795855408759e-06,
|
| 956 |
+
"loss": 0.0519,
|
| 957 |
+
"step": 79000
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 5.33,
|
| 961 |
+
"learning_rate": 9.336731272215144e-06,
|
| 962 |
+
"loss": 0.0949,
|
| 963 |
+
"step": 79500
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 5.37,
|
| 967 |
+
"learning_rate": 9.269666689021528e-06,
|
| 968 |
+
"loss": 0.085,
|
| 969 |
+
"step": 80000
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 5.4,
|
| 973 |
+
"learning_rate": 9.202602105827913e-06,
|
| 974 |
+
"loss": 0.1271,
|
| 975 |
+
"step": 80500
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 5.43,
|
| 979 |
+
"learning_rate": 9.135537522634298e-06,
|
| 980 |
+
"loss": 0.2085,
|
| 981 |
+
"step": 81000
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 5.47,
|
| 985 |
+
"learning_rate": 9.068472939440682e-06,
|
| 986 |
+
"loss": 0.0912,
|
| 987 |
+
"step": 81500
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 5.5,
|
| 991 |
+
"learning_rate": 9.001408356247067e-06,
|
| 992 |
+
"loss": 0.1413,
|
| 993 |
+
"step": 82000
|
| 994 |
+
},
|
| 995 |
+
{
|
| 996 |
+
"epoch": 5.53,
|
| 997 |
+
"learning_rate": 8.934343773053452e-06,
|
| 998 |
+
"loss": 0.1015,
|
| 999 |
+
"step": 82500
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 5.57,
|
| 1003 |
+
"learning_rate": 8.867279189859835e-06,
|
| 1004 |
+
"loss": 0.1277,
|
| 1005 |
+
"step": 83000
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 5.6,
|
| 1009 |
+
"learning_rate": 8.80021460666622e-06,
|
| 1010 |
+
"loss": 0.0239,
|
| 1011 |
+
"step": 83500
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 5.63,
|
| 1015 |
+
"learning_rate": 8.733150023472606e-06,
|
| 1016 |
+
"loss": 0.1947,
|
| 1017 |
+
"step": 84000
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 5.67,
|
| 1021 |
+
"learning_rate": 8.66608544027899e-06,
|
| 1022 |
+
"loss": 0.0589,
|
| 1023 |
+
"step": 84500
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 5.7,
|
| 1027 |
+
"learning_rate": 8.599020857085373e-06,
|
| 1028 |
+
"loss": 0.112,
|
| 1029 |
+
"step": 85000
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 5.73,
|
| 1033 |
+
"learning_rate": 8.531956273891758e-06,
|
| 1034 |
+
"loss": 0.0843,
|
| 1035 |
+
"step": 85500
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 5.77,
|
| 1039 |
+
"learning_rate": 8.464891690698143e-06,
|
| 1040 |
+
"loss": 0.151,
|
| 1041 |
+
"step": 86000
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 5.8,
|
| 1045 |
+
"learning_rate": 8.397827107504527e-06,
|
| 1046 |
+
"loss": 0.1011,
|
| 1047 |
+
"step": 86500
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 5.83,
|
| 1051 |
+
"learning_rate": 8.330762524310912e-06,
|
| 1052 |
+
"loss": 0.045,
|
| 1053 |
+
"step": 87000
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 5.87,
|
| 1057 |
+
"learning_rate": 8.263697941117297e-06,
|
| 1058 |
+
"loss": 0.0533,
|
| 1059 |
+
"step": 87500
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 5.9,
|
| 1063 |
+
"learning_rate": 8.196633357923681e-06,
|
| 1064 |
+
"loss": 0.0512,
|
| 1065 |
+
"step": 88000
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 5.94,
|
| 1069 |
+
"learning_rate": 8.129568774730066e-06,
|
| 1070 |
+
"loss": 0.2159,
|
| 1071 |
+
"step": 88500
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 5.97,
|
| 1075 |
+
"learning_rate": 8.06250419153645e-06,
|
| 1076 |
+
"loss": 0.1077,
|
| 1077 |
+
"step": 89000
|
| 1078 |
+
},
|
| 1079 |
+
{
|
| 1080 |
+
"epoch": 6.0,
|
| 1081 |
+
"learning_rate": 7.995439608342833e-06,
|
| 1082 |
+
"loss": 0.11,
|
| 1083 |
+
"step": 89500
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 6.04,
|
| 1087 |
+
"learning_rate": 7.92837502514922e-06,
|
| 1088 |
+
"loss": 0.1365,
|
| 1089 |
+
"step": 90000
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 6.07,
|
| 1093 |
+
"learning_rate": 7.861310441955604e-06,
|
| 1094 |
+
"loss": 0.0742,
|
| 1095 |
+
"step": 90500
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 6.1,
|
| 1099 |
+
"learning_rate": 7.794245858761989e-06,
|
| 1100 |
+
"loss": 0.0716,
|
| 1101 |
+
"step": 91000
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 6.14,
|
| 1105 |
+
"learning_rate": 7.727181275568372e-06,
|
| 1106 |
+
"loss": 0.1481,
|
| 1107 |
+
"step": 91500
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 6.17,
|
| 1111 |
+
"learning_rate": 7.660116692374757e-06,
|
| 1112 |
+
"loss": 0.008,
|
| 1113 |
+
"step": 92000
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 6.2,
|
| 1117 |
+
"learning_rate": 7.593052109181141e-06,
|
| 1118 |
+
"loss": 0.0615,
|
| 1119 |
+
"step": 92500
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 6.24,
|
| 1123 |
+
"learning_rate": 7.525987525987527e-06,
|
| 1124 |
+
"loss": 0.106,
|
| 1125 |
+
"step": 93000
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 6.27,
|
| 1129 |
+
"learning_rate": 7.4589229427939116e-06,
|
| 1130 |
+
"loss": 0.0534,
|
| 1131 |
+
"step": 93500
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 6.3,
|
| 1135 |
+
"learning_rate": 7.391858359600295e-06,
|
| 1136 |
+
"loss": 0.1034,
|
| 1137 |
+
"step": 94000
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 6.34,
|
| 1141 |
+
"learning_rate": 7.32479377640668e-06,
|
| 1142 |
+
"loss": 0.107,
|
| 1143 |
+
"step": 94500
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 6.37,
|
| 1147 |
+
"learning_rate": 7.257729193213065e-06,
|
| 1148 |
+
"loss": 0.0235,
|
| 1149 |
+
"step": 95000
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 6.4,
|
| 1153 |
+
"learning_rate": 7.1906646100194485e-06,
|
| 1154 |
+
"loss": 0.0844,
|
| 1155 |
+
"step": 95500
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 6.44,
|
| 1159 |
+
"learning_rate": 7.123600026825834e-06,
|
| 1160 |
+
"loss": 0.0495,
|
| 1161 |
+
"step": 96000
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 6.47,
|
| 1165 |
+
"learning_rate": 7.056535443632219e-06,
|
| 1166 |
+
"loss": 0.0572,
|
| 1167 |
+
"step": 96500
|
| 1168 |
+
},
|
| 1169 |
+
{
|
| 1170 |
+
"epoch": 6.51,
|
| 1171 |
+
"learning_rate": 6.989470860438603e-06,
|
| 1172 |
+
"loss": 0.024,
|
| 1173 |
+
"step": 97000
|
| 1174 |
+
},
|
| 1175 |
+
{
|
| 1176 |
+
"epoch": 6.54,
|
| 1177 |
+
"learning_rate": 6.922406277244987e-06,
|
| 1178 |
+
"loss": 0.1605,
|
| 1179 |
+
"step": 97500
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 6.57,
|
| 1183 |
+
"learning_rate": 6.855341694051372e-06,
|
| 1184 |
+
"loss": 0.1724,
|
| 1185 |
+
"step": 98000
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 6.61,
|
| 1189 |
+
"learning_rate": 6.7882771108577564e-06,
|
| 1190 |
+
"loss": 0.0505,
|
| 1191 |
+
"step": 98500
|
| 1192 |
+
},
|
| 1193 |
+
{
|
| 1194 |
+
"epoch": 6.64,
|
| 1195 |
+
"learning_rate": 6.721212527664142e-06,
|
| 1196 |
+
"loss": 0.0549,
|
| 1197 |
+
"step": 99000
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 6.67,
|
| 1201 |
+
"learning_rate": 6.654147944470526e-06,
|
| 1202 |
+
"loss": 0.1133,
|
| 1203 |
+
"step": 99500
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 6.71,
|
| 1207 |
+
"learning_rate": 6.58708336127691e-06,
|
| 1208 |
+
"loss": 0.0441,
|
| 1209 |
+
"step": 100000
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 6.74,
|
| 1213 |
+
"learning_rate": 6.520018778083294e-06,
|
| 1214 |
+
"loss": 0.0775,
|
| 1215 |
+
"step": 100500
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 6.77,
|
| 1219 |
+
"learning_rate": 6.452954194889679e-06,
|
| 1220 |
+
"loss": 0.0319,
|
| 1221 |
+
"step": 101000
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 6.81,
|
| 1225 |
+
"learning_rate": 6.3858896116960635e-06,
|
| 1226 |
+
"loss": 0.1535,
|
| 1227 |
+
"step": 101500
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 6.84,
|
| 1231 |
+
"learning_rate": 6.318825028502449e-06,
|
| 1232 |
+
"loss": 0.0782,
|
| 1233 |
+
"step": 102000
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 6.87,
|
| 1237 |
+
"learning_rate": 6.251760445308833e-06,
|
| 1238 |
+
"loss": 0.1007,
|
| 1239 |
+
"step": 102500
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 6.91,
|
| 1243 |
+
"learning_rate": 6.1846958621152175e-06,
|
| 1244 |
+
"loss": 0.0677,
|
| 1245 |
+
"step": 103000
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 6.94,
|
| 1249 |
+
"learning_rate": 6.117631278921602e-06,
|
| 1250 |
+
"loss": 0.0279,
|
| 1251 |
+
"step": 103500
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"epoch": 6.97,
|
| 1255 |
+
"learning_rate": 6.050566695727986e-06,
|
| 1256 |
+
"loss": 0.0685,
|
| 1257 |
+
"step": 104000
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"epoch": 7.01,
|
| 1261 |
+
"learning_rate": 5.983502112534371e-06,
|
| 1262 |
+
"loss": 0.0705,
|
| 1263 |
+
"step": 104500
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 7.04,
|
| 1267 |
+
"learning_rate": 5.916437529340756e-06,
|
| 1268 |
+
"loss": 0.0993,
|
| 1269 |
+
"step": 105000
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 7.08,
|
| 1273 |
+
"learning_rate": 5.849372946147141e-06,
|
| 1274 |
+
"loss": 0.0558,
|
| 1275 |
+
"step": 105500
|
| 1276 |
+
},
|
| 1277 |
+
{
|
| 1278 |
+
"epoch": 7.11,
|
| 1279 |
+
"learning_rate": 5.782308362953525e-06,
|
| 1280 |
+
"loss": 0.0672,
|
| 1281 |
+
"step": 106000
|
| 1282 |
+
},
|
| 1283 |
+
{
|
| 1284 |
+
"epoch": 7.14,
|
| 1285 |
+
"learning_rate": 5.715243779759909e-06,
|
| 1286 |
+
"loss": 0.0796,
|
| 1287 |
+
"step": 106500
|
| 1288 |
+
},
|
| 1289 |
+
{
|
| 1290 |
+
"epoch": 7.18,
|
| 1291 |
+
"learning_rate": 5.648179196566294e-06,
|
| 1292 |
+
"loss": 0.038,
|
| 1293 |
+
"step": 107000
|
| 1294 |
+
},
|
| 1295 |
+
{
|
| 1296 |
+
"epoch": 7.21,
|
| 1297 |
+
"learning_rate": 5.581114613372678e-06,
|
| 1298 |
+
"loss": 0.0612,
|
| 1299 |
+
"step": 107500
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 7.24,
|
| 1303 |
+
"learning_rate": 5.514050030179063e-06,
|
| 1304 |
+
"loss": 0.0537,
|
| 1305 |
+
"step": 108000
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 7.28,
|
| 1309 |
+
"learning_rate": 5.446985446985448e-06,
|
| 1310 |
+
"loss": 0.0291,
|
| 1311 |
+
"step": 108500
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 7.31,
|
| 1315 |
+
"learning_rate": 5.379920863791832e-06,
|
| 1316 |
+
"loss": 0.0772,
|
| 1317 |
+
"step": 109000
|
| 1318 |
+
},
|
| 1319 |
+
{
|
| 1320 |
+
"epoch": 7.34,
|
| 1321 |
+
"learning_rate": 5.312856280598216e-06,
|
| 1322 |
+
"loss": 0.0997,
|
| 1323 |
+
"step": 109500
|
| 1324 |
+
},
|
| 1325 |
+
{
|
| 1326 |
+
"epoch": 7.38,
|
| 1327 |
+
"learning_rate": 5.245791697404601e-06,
|
| 1328 |
+
"loss": 0.0505,
|
| 1329 |
+
"step": 110000
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"epoch": 7.41,
|
| 1333 |
+
"learning_rate": 5.178727114210985e-06,
|
| 1334 |
+
"loss": 0.0636,
|
| 1335 |
+
"step": 110500
|
| 1336 |
+
},
|
| 1337 |
+
{
|
| 1338 |
+
"epoch": 7.44,
|
| 1339 |
+
"learning_rate": 5.11166253101737e-06,
|
| 1340 |
+
"loss": 0.0314,
|
| 1341 |
+
"step": 111000
|
| 1342 |
+
},
|
| 1343 |
+
{
|
| 1344 |
+
"epoch": 7.48,
|
| 1345 |
+
"learning_rate": 5.044597947823755e-06,
|
| 1346 |
+
"loss": 0.1147,
|
| 1347 |
+
"step": 111500
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 7.51,
|
| 1351 |
+
"learning_rate": 4.97753336463014e-06,
|
| 1352 |
+
"loss": 0.0815,
|
| 1353 |
+
"step": 112000
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 7.54,
|
| 1357 |
+
"learning_rate": 4.9104687814365235e-06,
|
| 1358 |
+
"loss": 0.016,
|
| 1359 |
+
"step": 112500
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"epoch": 7.58,
|
| 1363 |
+
"learning_rate": 4.843404198242908e-06,
|
| 1364 |
+
"loss": 0.0509,
|
| 1365 |
+
"step": 113000
|
| 1366 |
+
},
|
| 1367 |
+
{
|
| 1368 |
+
"epoch": 7.61,
|
| 1369 |
+
"learning_rate": 4.776339615049293e-06,
|
| 1370 |
+
"loss": 0.0365,
|
| 1371 |
+
"step": 113500
|
| 1372 |
+
},
|
| 1373 |
+
{
|
| 1374 |
+
"epoch": 7.65,
|
| 1375 |
+
"learning_rate": 4.7092750318556775e-06,
|
| 1376 |
+
"loss": 0.0374,
|
| 1377 |
+
"step": 114000
|
| 1378 |
+
},
|
| 1379 |
+
{
|
| 1380 |
+
"epoch": 7.68,
|
| 1381 |
+
"learning_rate": 4.642210448662062e-06,
|
| 1382 |
+
"loss": 0.1001,
|
| 1383 |
+
"step": 114500
|
| 1384 |
+
},
|
| 1385 |
+
{
|
| 1386 |
+
"epoch": 7.71,
|
| 1387 |
+
"learning_rate": 4.575145865468447e-06,
|
| 1388 |
+
"loss": 0.021,
|
| 1389 |
+
"step": 115000
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 7.75,
|
| 1393 |
+
"learning_rate": 4.508081282274831e-06,
|
| 1394 |
+
"loss": 0.0171,
|
| 1395 |
+
"step": 115500
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 7.78,
|
| 1399 |
+
"learning_rate": 4.441016699081215e-06,
|
| 1400 |
+
"loss": 0.0759,
|
| 1401 |
+
"step": 116000
|
| 1402 |
+
},
|
| 1403 |
+
{
|
| 1404 |
+
"epoch": 7.81,
|
| 1405 |
+
"learning_rate": 4.3739521158876e-06,
|
| 1406 |
+
"loss": 0.0804,
|
| 1407 |
+
"step": 116500
|
| 1408 |
+
},
|
| 1409 |
+
{
|
| 1410 |
+
"epoch": 7.85,
|
| 1411 |
+
"learning_rate": 4.3068875326939846e-06,
|
| 1412 |
+
"loss": 0.0418,
|
| 1413 |
+
"step": 117000
|
| 1414 |
+
},
|
| 1415 |
+
{
|
| 1416 |
+
"epoch": 7.88,
|
| 1417 |
+
"learning_rate": 4.239822949500369e-06,
|
| 1418 |
+
"loss": 0.1122,
|
| 1419 |
+
"step": 117500
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"epoch": 7.91,
|
| 1423 |
+
"learning_rate": 4.172758366306754e-06,
|
| 1424 |
+
"loss": 0.0737,
|
| 1425 |
+
"step": 118000
|
| 1426 |
+
},
|
| 1427 |
+
{
|
| 1428 |
+
"epoch": 7.95,
|
| 1429 |
+
"learning_rate": 4.1056937831131385e-06,
|
| 1430 |
+
"loss": 0.0526,
|
| 1431 |
+
"step": 118500
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 7.98,
|
| 1435 |
+
"learning_rate": 4.038629199919522e-06,
|
| 1436 |
+
"loss": 0.0483,
|
| 1437 |
+
"step": 119000
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 8.01,
|
| 1441 |
+
"learning_rate": 3.971564616725908e-06,
|
| 1442 |
+
"loss": 0.037,
|
| 1443 |
+
"step": 119500
|
| 1444 |
+
},
|
| 1445 |
+
{
|
| 1446 |
+
"epoch": 8.05,
|
| 1447 |
+
"learning_rate": 3.904500033532292e-06,
|
| 1448 |
+
"loss": 0.0687,
|
| 1449 |
+
"step": 120000
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 8.08,
|
| 1453 |
+
"learning_rate": 3.837435450338676e-06,
|
| 1454 |
+
"loss": 0.023,
|
| 1455 |
+
"step": 120500
|
| 1456 |
+
},
|
| 1457 |
+
{
|
| 1458 |
+
"epoch": 8.11,
|
| 1459 |
+
"learning_rate": 3.770370867145061e-06,
|
| 1460 |
+
"loss": 0.1,
|
| 1461 |
+
"step": 121000
|
| 1462 |
+
},
|
| 1463 |
+
{
|
| 1464 |
+
"epoch": 8.15,
|
| 1465 |
+
"learning_rate": 3.7033062839514456e-06,
|
| 1466 |
+
"loss": 0.0659,
|
| 1467 |
+
"step": 121500
|
| 1468 |
+
},
|
| 1469 |
+
{
|
| 1470 |
+
"epoch": 8.18,
|
| 1471 |
+
"learning_rate": 3.63624170075783e-06,
|
| 1472 |
+
"loss": 0.0989,
|
| 1473 |
+
"step": 122000
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 8.22,
|
| 1477 |
+
"learning_rate": 3.569177117564215e-06,
|
| 1478 |
+
"loss": 0.0603,
|
| 1479 |
+
"step": 122500
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 8.25,
|
| 1483 |
+
"learning_rate": 3.502112534370599e-06,
|
| 1484 |
+
"loss": 0.0321,
|
| 1485 |
+
"step": 123000
|
| 1486 |
+
},
|
| 1487 |
+
{
|
| 1488 |
+
"epoch": 8.28,
|
| 1489 |
+
"learning_rate": 3.4350479511769834e-06,
|
| 1490 |
+
"loss": 0.0709,
|
| 1491 |
+
"step": 123500
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 8.32,
|
| 1495 |
+
"learning_rate": 3.3679833679833685e-06,
|
| 1496 |
+
"loss": 0.0389,
|
| 1497 |
+
"step": 124000
|
| 1498 |
+
},
|
| 1499 |
+
{
|
| 1500 |
+
"epoch": 8.35,
|
| 1501 |
+
"learning_rate": 3.3009187847897527e-06,
|
| 1502 |
+
"loss": 0.036,
|
| 1503 |
+
"step": 124500
|
| 1504 |
+
},
|
| 1505 |
+
{
|
| 1506 |
+
"epoch": 8.38,
|
| 1507 |
+
"learning_rate": 3.233854201596137e-06,
|
| 1508 |
+
"loss": 0.04,
|
| 1509 |
+
"step": 125000
|
| 1510 |
+
},
|
| 1511 |
+
{
|
| 1512 |
+
"epoch": 8.42,
|
| 1513 |
+
"learning_rate": 3.166789618402522e-06,
|
| 1514 |
+
"loss": 0.0265,
|
| 1515 |
+
"step": 125500
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 8.45,
|
| 1519 |
+
"learning_rate": 3.0997250352089063e-06,
|
| 1520 |
+
"loss": 0.0434,
|
| 1521 |
+
"step": 126000
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 8.48,
|
| 1525 |
+
"learning_rate": 3.032660452015291e-06,
|
| 1526 |
+
"loss": 0.0001,
|
| 1527 |
+
"step": 126500
|
| 1528 |
+
},
|
| 1529 |
+
{
|
| 1530 |
+
"epoch": 8.52,
|
| 1531 |
+
"learning_rate": 2.9655958688216756e-06,
|
| 1532 |
+
"loss": 0.0224,
|
| 1533 |
+
"step": 127000
|
| 1534 |
+
},
|
| 1535 |
+
{
|
| 1536 |
+
"epoch": 8.55,
|
| 1537 |
+
"learning_rate": 2.8985312856280603e-06,
|
| 1538 |
+
"loss": 0.0199,
|
| 1539 |
+
"step": 127500
|
| 1540 |
+
},
|
| 1541 |
+
{
|
| 1542 |
+
"epoch": 8.58,
|
| 1543 |
+
"learning_rate": 2.8314667024344445e-06,
|
| 1544 |
+
"loss": 0.0482,
|
| 1545 |
+
"step": 128000
|
| 1546 |
+
},
|
| 1547 |
+
{
|
| 1548 |
+
"epoch": 8.62,
|
| 1549 |
+
"learning_rate": 2.764402119240829e-06,
|
| 1550 |
+
"loss": 0.0329,
|
| 1551 |
+
"step": 128500
|
| 1552 |
+
},
|
| 1553 |
+
{
|
| 1554 |
+
"epoch": 8.65,
|
| 1555 |
+
"learning_rate": 2.697337536047214e-06,
|
| 1556 |
+
"loss": 0.0515,
|
| 1557 |
+
"step": 129000
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 8.68,
|
| 1561 |
+
"learning_rate": 2.630272952853598e-06,
|
| 1562 |
+
"loss": 0.0473,
|
| 1563 |
+
"step": 129500
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 8.72,
|
| 1567 |
+
"learning_rate": 2.563208369659983e-06,
|
| 1568 |
+
"loss": 0.1201,
|
| 1569 |
+
"step": 130000
|
| 1570 |
+
},
|
| 1571 |
+
{
|
| 1572 |
+
"epoch": 8.75,
|
| 1573 |
+
"learning_rate": 2.4961437864663674e-06,
|
| 1574 |
+
"loss": 0.0295,
|
| 1575 |
+
"step": 130500
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 8.79,
|
| 1579 |
+
"learning_rate": 2.429079203272752e-06,
|
| 1580 |
+
"loss": 0.0001,
|
| 1581 |
+
"step": 131000
|
| 1582 |
+
},
|
| 1583 |
+
{
|
| 1584 |
+
"epoch": 8.82,
|
| 1585 |
+
"learning_rate": 2.3620146200791363e-06,
|
| 1586 |
+
"loss": 0.0537,
|
| 1587 |
+
"step": 131500
|
| 1588 |
+
},
|
| 1589 |
+
{
|
| 1590 |
+
"epoch": 8.85,
|
| 1591 |
+
"learning_rate": 2.294950036885521e-06,
|
| 1592 |
+
"loss": 0.0821,
|
| 1593 |
+
"step": 132000
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 8.89,
|
| 1597 |
+
"learning_rate": 2.2278854536919056e-06,
|
| 1598 |
+
"loss": 0.0665,
|
| 1599 |
+
"step": 132500
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 8.92,
|
| 1603 |
+
"learning_rate": 2.16082087049829e-06,
|
| 1604 |
+
"loss": 0.0391,
|
| 1605 |
+
"step": 133000
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 8.95,
|
| 1609 |
+
"learning_rate": 2.0937562873046745e-06,
|
| 1610 |
+
"loss": 0.0315,
|
| 1611 |
+
"step": 133500
|
| 1612 |
+
},
|
| 1613 |
+
{
|
| 1614 |
+
"epoch": 8.99,
|
| 1615 |
+
"learning_rate": 2.026691704111059e-06,
|
| 1616 |
+
"loss": 0.0431,
|
| 1617 |
+
"step": 134000
|
| 1618 |
+
},
|
| 1619 |
+
{
|
| 1620 |
+
"epoch": 9.02,
|
| 1621 |
+
"learning_rate": 1.959627120917444e-06,
|
| 1622 |
+
"loss": 0.0328,
|
| 1623 |
+
"step": 134500
|
| 1624 |
+
},
|
| 1625 |
+
{
|
| 1626 |
+
"epoch": 9.05,
|
| 1627 |
+
"learning_rate": 1.8925625377238282e-06,
|
| 1628 |
+
"loss": 0.0126,
|
| 1629 |
+
"step": 135000
|
| 1630 |
+
},
|
| 1631 |
+
{
|
| 1632 |
+
"epoch": 9.09,
|
| 1633 |
+
"learning_rate": 1.8254979545302127e-06,
|
| 1634 |
+
"loss": 0.0086,
|
| 1635 |
+
"step": 135500
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 9.12,
|
| 1639 |
+
"learning_rate": 1.7584333713365973e-06,
|
| 1640 |
+
"loss": 0.0242,
|
| 1641 |
+
"step": 136000
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 9.15,
|
| 1645 |
+
"learning_rate": 1.6913687881429818e-06,
|
| 1646 |
+
"loss": 0.0983,
|
| 1647 |
+
"step": 136500
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 9.19,
|
| 1651 |
+
"learning_rate": 1.6243042049493662e-06,
|
| 1652 |
+
"loss": 0.0471,
|
| 1653 |
+
"step": 137000
|
| 1654 |
+
},
|
| 1655 |
+
{
|
| 1656 |
+
"epoch": 9.22,
|
| 1657 |
+
"learning_rate": 1.557239621755751e-06,
|
| 1658 |
+
"loss": 0.0347,
|
| 1659 |
+
"step": 137500
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 9.25,
|
| 1663 |
+
"learning_rate": 1.4901750385621356e-06,
|
| 1664 |
+
"loss": 0.0387,
|
| 1665 |
+
"step": 138000
|
| 1666 |
+
},
|
| 1667 |
+
{
|
| 1668 |
+
"epoch": 9.29,
|
| 1669 |
+
"learning_rate": 1.4231104553685198e-06,
|
| 1670 |
+
"loss": 0.0398,
|
| 1671 |
+
"step": 138500
|
| 1672 |
+
},
|
| 1673 |
+
{
|
| 1674 |
+
"epoch": 9.32,
|
| 1675 |
+
"learning_rate": 1.3560458721749044e-06,
|
| 1676 |
+
"loss": 0.062,
|
| 1677 |
+
"step": 139000
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 9.36,
|
| 1681 |
+
"learning_rate": 1.288981288981289e-06,
|
| 1682 |
+
"loss": 0.0003,
|
| 1683 |
+
"step": 139500
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 9.39,
|
| 1687 |
+
"learning_rate": 1.2219167057876738e-06,
|
| 1688 |
+
"loss": 0.0466,
|
| 1689 |
+
"step": 140000
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 9.42,
|
| 1693 |
+
"learning_rate": 1.1548521225940582e-06,
|
| 1694 |
+
"loss": 0.0385,
|
| 1695 |
+
"step": 140500
|
| 1696 |
+
},
|
| 1697 |
+
{
|
| 1698 |
+
"epoch": 9.46,
|
| 1699 |
+
"learning_rate": 1.0877875394004427e-06,
|
| 1700 |
+
"loss": 0.0581,
|
| 1701 |
+
"step": 141000
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 9.49,
|
| 1705 |
+
"learning_rate": 1.0207229562068273e-06,
|
| 1706 |
+
"loss": 0.016,
|
| 1707 |
+
"step": 141500
|
| 1708 |
+
},
|
| 1709 |
+
{
|
| 1710 |
+
"epoch": 9.52,
|
| 1711 |
+
"learning_rate": 9.536583730132118e-07,
|
| 1712 |
+
"loss": 0.0202,
|
| 1713 |
+
"step": 142000
|
| 1714 |
+
},
|
| 1715 |
+
{
|
| 1716 |
+
"epoch": 9.56,
|
| 1717 |
+
"learning_rate": 8.865937898195963e-07,
|
| 1718 |
+
"loss": 0.0409,
|
| 1719 |
+
"step": 142500
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 9.59,
|
| 1723 |
+
"learning_rate": 8.19529206625981e-07,
|
| 1724 |
+
"loss": 0.0327,
|
| 1725 |
+
"step": 143000
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 9.62,
|
| 1729 |
+
"learning_rate": 7.524646234323654e-07,
|
| 1730 |
+
"loss": 0.0101,
|
| 1731 |
+
"step": 143500
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 9.66,
|
| 1735 |
+
"learning_rate": 6.8540004023875e-07,
|
| 1736 |
+
"loss": 0.0539,
|
| 1737 |
+
"step": 144000
|
| 1738 |
+
},
|
| 1739 |
+
{
|
| 1740 |
+
"epoch": 9.69,
|
| 1741 |
+
"learning_rate": 6.183354570451345e-07,
|
| 1742 |
+
"loss": 0.0001,
|
| 1743 |
+
"step": 144500
|
| 1744 |
+
},
|
| 1745 |
+
{
|
| 1746 |
+
"epoch": 9.72,
|
| 1747 |
+
"learning_rate": 5.512708738515191e-07,
|
| 1748 |
+
"loss": 0.0521,
|
| 1749 |
+
"step": 145000
|
| 1750 |
+
},
|
| 1751 |
+
{
|
| 1752 |
+
"epoch": 9.76,
|
| 1753 |
+
"learning_rate": 4.842062906579036e-07,
|
| 1754 |
+
"loss": 0.0702,
|
| 1755 |
+
"step": 145500
|
| 1756 |
+
},
|
| 1757 |
+
{
|
| 1758 |
+
"epoch": 9.79,
|
| 1759 |
+
"learning_rate": 4.1714170746428813e-07,
|
| 1760 |
+
"loss": 0.0559,
|
| 1761 |
+
"step": 146000
|
| 1762 |
+
},
|
| 1763 |
+
{
|
| 1764 |
+
"epoch": 9.82,
|
| 1765 |
+
"learning_rate": 3.500771242706727e-07,
|
| 1766 |
+
"loss": 0.0394,
|
| 1767 |
+
"step": 146500
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 9.86,
|
| 1771 |
+
"learning_rate": 2.8301254107705723e-07,
|
| 1772 |
+
"loss": 0.0164,
|
| 1773 |
+
"step": 147000
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 9.89,
|
| 1777 |
+
"learning_rate": 2.1594795788344176e-07,
|
| 1778 |
+
"loss": 0.0194,
|
| 1779 |
+
"step": 147500
|
| 1780 |
+
},
|
| 1781 |
+
{
|
| 1782 |
+
"epoch": 9.93,
|
| 1783 |
+
"learning_rate": 1.488833746898263e-07,
|
| 1784 |
+
"loss": 0.0652,
|
| 1785 |
+
"step": 148000
|
| 1786 |
+
},
|
| 1787 |
+
{
|
| 1788 |
+
"epoch": 9.96,
|
| 1789 |
+
"learning_rate": 8.181879149621086e-08,
|
| 1790 |
+
"loss": 0.077,
|
| 1791 |
+
"step": 148500
|
| 1792 |
+
},
|
| 1793 |
+
{
|
| 1794 |
+
"epoch": 9.99,
|
| 1795 |
+
"learning_rate": 1.4754208302595401e-08,
|
| 1796 |
+
"loss": 0.0659,
|
| 1797 |
+
"step": 149000
|
| 1798 |
+
},
|
| 1799 |
+
{
|
| 1800 |
+
"epoch": 10.0,
|
| 1801 |
+
"step": 149110,
|
| 1802 |
+
"total_flos": 9808298492528640.0,
|
| 1803 |
+
"train_loss": 0.1504479006291696,
|
| 1804 |
+
"train_runtime": 9647.5503,
|
| 1805 |
+
"train_samples_per_second": 15.456,
|
| 1806 |
+
"train_steps_per_second": 15.456
|
| 1807 |
+
}
|
| 1808 |
+
],
|
| 1809 |
+
"logging_steps": 500,
|
| 1810 |
+
"max_steps": 149110,
|
| 1811 |
+
"num_train_epochs": 10,
|
| 1812 |
+
"save_steps": 500,
|
| 1813 |
+
"total_flos": 9808298492528640.0,
|
| 1814 |
+
"trial_name": null,
|
| 1815 |
+
"trial_params": null
|
| 1816 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:751a4ead01d883aec4db1006679da3d8c2ff0e2d854a233a1157acdb75cf4a60
|
| 3 |
+
size 4536
|
vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|