commit from mk
Browse files- README.md +57 -0
- all_results.json +8 -0
- config.json +39 -0
- generation_config.json +6 -0
- merges.txt +0 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +6 -0
- tokenizer.json +0 -0
- tokenizer_config.json +10 -0
- train_results.json +8 -0
- trainer_state.json +2185 -0
- training_args.bin +3 -0
- vocab.json +0 -0
README.md
ADDED
|
@@ -0,0 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
tags:
|
| 4 |
+
- generated_from_trainer
|
| 5 |
+
datasets:
|
| 6 |
+
- customized
|
| 7 |
+
model-index:
|
| 8 |
+
- name: finetuned_gpt2
|
| 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 |
+
# finetuned_gpt2
|
| 16 |
+
|
| 17 |
+
This model is a fine-tuned version of [gpt2](https://huggingface.co/gpt2) on the customized dataset.
|
| 18 |
+
|
| 19 |
+
## Model description
|
| 20 |
+
|
| 21 |
+
More information needed
|
| 22 |
+
|
| 23 |
+
## Intended uses & limitations
|
| 24 |
+
|
| 25 |
+
More information needed
|
| 26 |
+
|
| 27 |
+
## Training and evaluation data
|
| 28 |
+
|
| 29 |
+
More information needed
|
| 30 |
+
|
| 31 |
+
## Training procedure
|
| 32 |
+
|
| 33 |
+
### Training hyperparameters
|
| 34 |
+
|
| 35 |
+
The following hyperparameters were used during training:
|
| 36 |
+
- learning_rate: 2e-05
|
| 37 |
+
- train_batch_size: 1
|
| 38 |
+
- eval_batch_size: 8
|
| 39 |
+
- seed: 42
|
| 40 |
+
- distributed_type: multi-GPU
|
| 41 |
+
- num_devices: 6
|
| 42 |
+
- total_train_batch_size: 6
|
| 43 |
+
- total_eval_batch_size: 48
|
| 44 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 45 |
+
- lr_scheduler_type: linear
|
| 46 |
+
- num_epochs: 5.0
|
| 47 |
+
|
| 48 |
+
### Training results
|
| 49 |
+
|
| 50 |
+
|
| 51 |
+
|
| 52 |
+
### Framework versions
|
| 53 |
+
|
| 54 |
+
- Transformers 4.28.0.dev0
|
| 55 |
+
- Pytorch 2.0.0+cu117
|
| 56 |
+
- Datasets 2.10.1
|
| 57 |
+
- Tokenizers 0.13.3
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"train_loss": 2.03250736604541,
|
| 4 |
+
"train_runtime": 4701.6865,
|
| 5 |
+
"train_samples": 8645,
|
| 6 |
+
"train_samples_per_second": 9.194,
|
| 7 |
+
"train_steps_per_second": 1.532
|
| 8 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,39 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "gpt2",
|
| 3 |
+
"activation_function": "gelu_new",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"GPT2LMHeadModel"
|
| 6 |
+
],
|
| 7 |
+
"attn_pdrop": 0.1,
|
| 8 |
+
"bos_token_id": 50256,
|
| 9 |
+
"embd_pdrop": 0.1,
|
| 10 |
+
"eos_token_id": 50256,
|
| 11 |
+
"initializer_range": 0.02,
|
| 12 |
+
"layer_norm_epsilon": 1e-05,
|
| 13 |
+
"model_type": "gpt2",
|
| 14 |
+
"n_ctx": 1024,
|
| 15 |
+
"n_embd": 768,
|
| 16 |
+
"n_head": 12,
|
| 17 |
+
"n_inner": null,
|
| 18 |
+
"n_layer": 12,
|
| 19 |
+
"n_positions": 1024,
|
| 20 |
+
"reorder_and_upcast_attn": false,
|
| 21 |
+
"resid_pdrop": 0.1,
|
| 22 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 23 |
+
"scale_attn_weights": true,
|
| 24 |
+
"summary_activation": null,
|
| 25 |
+
"summary_first_dropout": 0.1,
|
| 26 |
+
"summary_proj_to_labels": true,
|
| 27 |
+
"summary_type": "cls_index",
|
| 28 |
+
"summary_use_proj": true,
|
| 29 |
+
"task_specific_params": {
|
| 30 |
+
"text-generation": {
|
| 31 |
+
"do_sample": true,
|
| 32 |
+
"max_length": 50
|
| 33 |
+
}
|
| 34 |
+
},
|
| 35 |
+
"torch_dtype": "float16",
|
| 36 |
+
"transformers_version": "4.28.0.dev0",
|
| 37 |
+
"use_cache": true,
|
| 38 |
+
"vocab_size": 50257
|
| 39 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 50256,
|
| 4 |
+
"eos_token_id": 50256,
|
| 5 |
+
"transformers_version": "4.28.0.dev0"
|
| 6 |
+
}
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:faed4563ffd74f3117d8f3bee2b951cc0ef009a6a5c6bd126526a2495a66fcc3
|
| 3 |
+
size 261515901
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<|endoftext|>",
|
| 3 |
+
"eos_token": "<|endoftext|>",
|
| 4 |
+
"pad_token": "<|endoftext|>",
|
| 5 |
+
"unk_token": "<|endoftext|>"
|
| 6 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": false,
|
| 3 |
+
"bos_token": "<|endoftext|>",
|
| 4 |
+
"clean_up_tokenization_spaces": true,
|
| 5 |
+
"eos_token": "<|endoftext|>",
|
| 6 |
+
"model_max_length": 1024,
|
| 7 |
+
"special_tokens_map_file": null,
|
| 8 |
+
"tokenizer_class": "GPT2Tokenizer",
|
| 9 |
+
"unk_token": "<|endoftext|>"
|
| 10 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 5.0,
|
| 3 |
+
"train_loss": 2.03250736604541,
|
| 4 |
+
"train_runtime": 4701.6865,
|
| 5 |
+
"train_samples": 8645,
|
| 6 |
+
"train_samples_per_second": 9.194,
|
| 7 |
+
"train_steps_per_second": 1.532
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2185 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 5.0,
|
| 5 |
+
"global_step": 7205,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.01,
|
| 12 |
+
"learning_rate": 1.9958362248438587e-05,
|
| 13 |
+
"loss": 2.8983,
|
| 14 |
+
"step": 20
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.03,
|
| 18 |
+
"learning_rate": 1.99028452463567e-05,
|
| 19 |
+
"loss": 2.548,
|
| 20 |
+
"step": 40
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.04,
|
| 24 |
+
"learning_rate": 1.984732824427481e-05,
|
| 25 |
+
"loss": 2.4594,
|
| 26 |
+
"step": 60
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.06,
|
| 30 |
+
"learning_rate": 1.9791811242192923e-05,
|
| 31 |
+
"loss": 2.3628,
|
| 32 |
+
"step": 80
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.07,
|
| 36 |
+
"learning_rate": 1.9736294240111035e-05,
|
| 37 |
+
"loss": 2.3952,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.08,
|
| 42 |
+
"learning_rate": 1.9680777238029147e-05,
|
| 43 |
+
"loss": 2.3894,
|
| 44 |
+
"step": 120
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.1,
|
| 48 |
+
"learning_rate": 1.9625260235947262e-05,
|
| 49 |
+
"loss": 2.3013,
|
| 50 |
+
"step": 140
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.11,
|
| 54 |
+
"learning_rate": 1.9569743233865374e-05,
|
| 55 |
+
"loss": 2.3323,
|
| 56 |
+
"step": 160
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.12,
|
| 60 |
+
"learning_rate": 1.9514226231783486e-05,
|
| 61 |
+
"loss": 2.2816,
|
| 62 |
+
"step": 180
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.14,
|
| 66 |
+
"learning_rate": 1.9458709229701598e-05,
|
| 67 |
+
"loss": 2.3024,
|
| 68 |
+
"step": 200
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.15,
|
| 72 |
+
"learning_rate": 1.940319222761971e-05,
|
| 73 |
+
"loss": 2.2957,
|
| 74 |
+
"step": 220
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.17,
|
| 78 |
+
"learning_rate": 1.9347675225537822e-05,
|
| 79 |
+
"loss": 2.2605,
|
| 80 |
+
"step": 240
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.18,
|
| 84 |
+
"learning_rate": 1.9292158223455934e-05,
|
| 85 |
+
"loss": 2.2591,
|
| 86 |
+
"step": 260
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.19,
|
| 90 |
+
"learning_rate": 1.923664122137405e-05,
|
| 91 |
+
"loss": 2.2737,
|
| 92 |
+
"step": 280
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.21,
|
| 96 |
+
"learning_rate": 1.918112421929216e-05,
|
| 97 |
+
"loss": 2.2776,
|
| 98 |
+
"step": 300
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.22,
|
| 102 |
+
"learning_rate": 1.912560721721027e-05,
|
| 103 |
+
"loss": 2.2052,
|
| 104 |
+
"step": 320
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.24,
|
| 108 |
+
"learning_rate": 1.9070090215128386e-05,
|
| 109 |
+
"loss": 2.2548,
|
| 110 |
+
"step": 340
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.25,
|
| 114 |
+
"learning_rate": 1.9014573213046498e-05,
|
| 115 |
+
"loss": 2.246,
|
| 116 |
+
"step": 360
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.26,
|
| 120 |
+
"learning_rate": 1.895905621096461e-05,
|
| 121 |
+
"loss": 2.2273,
|
| 122 |
+
"step": 380
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.28,
|
| 126 |
+
"learning_rate": 1.8903539208882722e-05,
|
| 127 |
+
"loss": 2.2351,
|
| 128 |
+
"step": 400
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.29,
|
| 132 |
+
"learning_rate": 1.8848022206800834e-05,
|
| 133 |
+
"loss": 2.234,
|
| 134 |
+
"step": 420
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.31,
|
| 138 |
+
"learning_rate": 1.8792505204718946e-05,
|
| 139 |
+
"loss": 2.2372,
|
| 140 |
+
"step": 440
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.32,
|
| 144 |
+
"learning_rate": 1.8736988202637058e-05,
|
| 145 |
+
"loss": 2.1749,
|
| 146 |
+
"step": 460
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.33,
|
| 150 |
+
"learning_rate": 1.8681471200555174e-05,
|
| 151 |
+
"loss": 2.1885,
|
| 152 |
+
"step": 480
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.35,
|
| 156 |
+
"learning_rate": 1.8625954198473282e-05,
|
| 157 |
+
"loss": 2.1873,
|
| 158 |
+
"step": 500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.36,
|
| 162 |
+
"learning_rate": 1.8570437196391394e-05,
|
| 163 |
+
"loss": 2.2094,
|
| 164 |
+
"step": 520
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.37,
|
| 168 |
+
"learning_rate": 1.851492019430951e-05,
|
| 169 |
+
"loss": 2.1991,
|
| 170 |
+
"step": 540
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.39,
|
| 174 |
+
"learning_rate": 1.8459403192227622e-05,
|
| 175 |
+
"loss": 2.1826,
|
| 176 |
+
"step": 560
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.4,
|
| 180 |
+
"learning_rate": 1.8403886190145734e-05,
|
| 181 |
+
"loss": 2.2183,
|
| 182 |
+
"step": 580
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.42,
|
| 186 |
+
"learning_rate": 1.8348369188063846e-05,
|
| 187 |
+
"loss": 2.1631,
|
| 188 |
+
"step": 600
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.43,
|
| 192 |
+
"learning_rate": 1.8292852185981958e-05,
|
| 193 |
+
"loss": 2.2094,
|
| 194 |
+
"step": 620
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.44,
|
| 198 |
+
"learning_rate": 1.823733518390007e-05,
|
| 199 |
+
"loss": 2.2029,
|
| 200 |
+
"step": 640
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.46,
|
| 204 |
+
"learning_rate": 1.8181818181818182e-05,
|
| 205 |
+
"loss": 2.1485,
|
| 206 |
+
"step": 660
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.47,
|
| 210 |
+
"learning_rate": 1.8126301179736298e-05,
|
| 211 |
+
"loss": 2.1773,
|
| 212 |
+
"step": 680
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.49,
|
| 216 |
+
"learning_rate": 1.807078417765441e-05,
|
| 217 |
+
"loss": 2.1284,
|
| 218 |
+
"step": 700
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.5,
|
| 222 |
+
"learning_rate": 1.8015267175572518e-05,
|
| 223 |
+
"loss": 2.1754,
|
| 224 |
+
"step": 720
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.51,
|
| 228 |
+
"learning_rate": 1.7959750173490634e-05,
|
| 229 |
+
"loss": 2.208,
|
| 230 |
+
"step": 740
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.53,
|
| 234 |
+
"learning_rate": 1.7904233171408746e-05,
|
| 235 |
+
"loss": 2.1423,
|
| 236 |
+
"step": 760
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.54,
|
| 240 |
+
"learning_rate": 1.7848716169326858e-05,
|
| 241 |
+
"loss": 2.1489,
|
| 242 |
+
"step": 780
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.56,
|
| 246 |
+
"learning_rate": 1.779319916724497e-05,
|
| 247 |
+
"loss": 2.1808,
|
| 248 |
+
"step": 800
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.57,
|
| 252 |
+
"learning_rate": 1.7737682165163085e-05,
|
| 253 |
+
"loss": 2.1837,
|
| 254 |
+
"step": 820
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.58,
|
| 258 |
+
"learning_rate": 1.7682165163081194e-05,
|
| 259 |
+
"loss": 2.135,
|
| 260 |
+
"step": 840
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.6,
|
| 264 |
+
"learning_rate": 1.7626648160999306e-05,
|
| 265 |
+
"loss": 2.1493,
|
| 266 |
+
"step": 860
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.61,
|
| 270 |
+
"learning_rate": 1.757113115891742e-05,
|
| 271 |
+
"loss": 2.1963,
|
| 272 |
+
"step": 880
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.62,
|
| 276 |
+
"learning_rate": 1.7515614156835533e-05,
|
| 277 |
+
"loss": 2.1365,
|
| 278 |
+
"step": 900
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.64,
|
| 282 |
+
"learning_rate": 1.7460097154753645e-05,
|
| 283 |
+
"loss": 2.1524,
|
| 284 |
+
"step": 920
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.65,
|
| 288 |
+
"learning_rate": 1.7404580152671757e-05,
|
| 289 |
+
"loss": 2.1455,
|
| 290 |
+
"step": 940
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.67,
|
| 294 |
+
"learning_rate": 1.734906315058987e-05,
|
| 295 |
+
"loss": 2.1359,
|
| 296 |
+
"step": 960
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.68,
|
| 300 |
+
"learning_rate": 1.729354614850798e-05,
|
| 301 |
+
"loss": 2.1394,
|
| 302 |
+
"step": 980
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.69,
|
| 306 |
+
"learning_rate": 1.7238029146426094e-05,
|
| 307 |
+
"loss": 2.1559,
|
| 308 |
+
"step": 1000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.71,
|
| 312 |
+
"learning_rate": 1.718251214434421e-05,
|
| 313 |
+
"loss": 2.1819,
|
| 314 |
+
"step": 1020
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.72,
|
| 318 |
+
"learning_rate": 1.7126995142262318e-05,
|
| 319 |
+
"loss": 2.1547,
|
| 320 |
+
"step": 1040
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.74,
|
| 324 |
+
"learning_rate": 1.707147814018043e-05,
|
| 325 |
+
"loss": 2.1226,
|
| 326 |
+
"step": 1060
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.75,
|
| 330 |
+
"learning_rate": 1.7015961138098545e-05,
|
| 331 |
+
"loss": 2.0807,
|
| 332 |
+
"step": 1080
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.76,
|
| 336 |
+
"learning_rate": 1.6960444136016657e-05,
|
| 337 |
+
"loss": 2.1243,
|
| 338 |
+
"step": 1100
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.78,
|
| 342 |
+
"learning_rate": 1.690492713393477e-05,
|
| 343 |
+
"loss": 2.1298,
|
| 344 |
+
"step": 1120
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.79,
|
| 348 |
+
"learning_rate": 1.684941013185288e-05,
|
| 349 |
+
"loss": 2.1265,
|
| 350 |
+
"step": 1140
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.8,
|
| 354 |
+
"learning_rate": 1.6793893129770993e-05,
|
| 355 |
+
"loss": 2.1234,
|
| 356 |
+
"step": 1160
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.82,
|
| 360 |
+
"learning_rate": 1.6738376127689105e-05,
|
| 361 |
+
"loss": 2.1243,
|
| 362 |
+
"step": 1180
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.83,
|
| 366 |
+
"learning_rate": 1.6682859125607217e-05,
|
| 367 |
+
"loss": 2.1369,
|
| 368 |
+
"step": 1200
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.85,
|
| 372 |
+
"learning_rate": 1.6627342123525333e-05,
|
| 373 |
+
"loss": 2.1074,
|
| 374 |
+
"step": 1220
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.86,
|
| 378 |
+
"learning_rate": 1.6571825121443445e-05,
|
| 379 |
+
"loss": 2.1411,
|
| 380 |
+
"step": 1240
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.87,
|
| 384 |
+
"learning_rate": 1.6516308119361554e-05,
|
| 385 |
+
"loss": 2.12,
|
| 386 |
+
"step": 1260
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.89,
|
| 390 |
+
"learning_rate": 1.646079111727967e-05,
|
| 391 |
+
"loss": 2.0964,
|
| 392 |
+
"step": 1280
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.9,
|
| 396 |
+
"learning_rate": 1.640527411519778e-05,
|
| 397 |
+
"loss": 2.0956,
|
| 398 |
+
"step": 1300
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.92,
|
| 402 |
+
"learning_rate": 1.6349757113115893e-05,
|
| 403 |
+
"loss": 2.1189,
|
| 404 |
+
"step": 1320
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.93,
|
| 408 |
+
"learning_rate": 1.6294240111034005e-05,
|
| 409 |
+
"loss": 2.1288,
|
| 410 |
+
"step": 1340
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.94,
|
| 414 |
+
"learning_rate": 1.6238723108952117e-05,
|
| 415 |
+
"loss": 2.0694,
|
| 416 |
+
"step": 1360
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.96,
|
| 420 |
+
"learning_rate": 1.618320610687023e-05,
|
| 421 |
+
"loss": 2.155,
|
| 422 |
+
"step": 1380
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.97,
|
| 426 |
+
"learning_rate": 1.612768910478834e-05,
|
| 427 |
+
"loss": 2.0934,
|
| 428 |
+
"step": 1400
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.99,
|
| 432 |
+
"learning_rate": 1.6072172102706457e-05,
|
| 433 |
+
"loss": 2.1193,
|
| 434 |
+
"step": 1420
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 1.0,
|
| 438 |
+
"learning_rate": 1.6019430950728662e-05,
|
| 439 |
+
"loss": 2.157,
|
| 440 |
+
"step": 1440
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 1.01,
|
| 444 |
+
"learning_rate": 1.5963913948646774e-05,
|
| 445 |
+
"loss": 2.0529,
|
| 446 |
+
"step": 1460
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 1.03,
|
| 450 |
+
"learning_rate": 1.5908396946564886e-05,
|
| 451 |
+
"loss": 2.0998,
|
| 452 |
+
"step": 1480
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 1.04,
|
| 456 |
+
"learning_rate": 1.5852879944482998e-05,
|
| 457 |
+
"loss": 2.0727,
|
| 458 |
+
"step": 1500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 1.05,
|
| 462 |
+
"learning_rate": 1.579736294240111e-05,
|
| 463 |
+
"loss": 2.1137,
|
| 464 |
+
"step": 1520
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 1.07,
|
| 468 |
+
"learning_rate": 1.5741845940319226e-05,
|
| 469 |
+
"loss": 2.0415,
|
| 470 |
+
"step": 1540
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 1.08,
|
| 474 |
+
"learning_rate": 1.5686328938237338e-05,
|
| 475 |
+
"loss": 2.0903,
|
| 476 |
+
"step": 1560
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 1.1,
|
| 480 |
+
"learning_rate": 1.5630811936155446e-05,
|
| 481 |
+
"loss": 2.0646,
|
| 482 |
+
"step": 1580
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 1.11,
|
| 486 |
+
"learning_rate": 1.557529493407356e-05,
|
| 487 |
+
"loss": 2.093,
|
| 488 |
+
"step": 1600
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 1.12,
|
| 492 |
+
"learning_rate": 1.5519777931991674e-05,
|
| 493 |
+
"loss": 2.0981,
|
| 494 |
+
"step": 1620
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 1.14,
|
| 498 |
+
"learning_rate": 1.5464260929909786e-05,
|
| 499 |
+
"loss": 2.0702,
|
| 500 |
+
"step": 1640
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 1.15,
|
| 504 |
+
"learning_rate": 1.5408743927827898e-05,
|
| 505 |
+
"loss": 2.0566,
|
| 506 |
+
"step": 1660
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 1.17,
|
| 510 |
+
"learning_rate": 1.5353226925746013e-05,
|
| 511 |
+
"loss": 2.1073,
|
| 512 |
+
"step": 1680
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 1.18,
|
| 516 |
+
"learning_rate": 1.5297709923664122e-05,
|
| 517 |
+
"loss": 2.0682,
|
| 518 |
+
"step": 1700
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 1.19,
|
| 522 |
+
"learning_rate": 1.5242192921582236e-05,
|
| 523 |
+
"loss": 2.0639,
|
| 524 |
+
"step": 1720
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 1.21,
|
| 528 |
+
"learning_rate": 1.518667591950035e-05,
|
| 529 |
+
"loss": 2.0975,
|
| 530 |
+
"step": 1740
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 1.22,
|
| 534 |
+
"learning_rate": 1.513115891741846e-05,
|
| 535 |
+
"loss": 2.068,
|
| 536 |
+
"step": 1760
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 1.24,
|
| 540 |
+
"learning_rate": 1.5075641915336573e-05,
|
| 541 |
+
"loss": 2.0548,
|
| 542 |
+
"step": 1780
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 1.25,
|
| 546 |
+
"learning_rate": 1.5020124913254685e-05,
|
| 547 |
+
"loss": 2.0739,
|
| 548 |
+
"step": 1800
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 1.26,
|
| 552 |
+
"learning_rate": 1.4964607911172798e-05,
|
| 553 |
+
"loss": 2.009,
|
| 554 |
+
"step": 1820
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 1.28,
|
| 558 |
+
"learning_rate": 1.4909090909090911e-05,
|
| 559 |
+
"loss": 2.0673,
|
| 560 |
+
"step": 1840
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 1.29,
|
| 564 |
+
"learning_rate": 1.4853573907009022e-05,
|
| 565 |
+
"loss": 2.0708,
|
| 566 |
+
"step": 1860
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 1.3,
|
| 570 |
+
"learning_rate": 1.4798056904927135e-05,
|
| 571 |
+
"loss": 2.0457,
|
| 572 |
+
"step": 1880
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 1.32,
|
| 576 |
+
"learning_rate": 1.4742539902845247e-05,
|
| 577 |
+
"loss": 2.0952,
|
| 578 |
+
"step": 1900
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 1.33,
|
| 582 |
+
"learning_rate": 1.468702290076336e-05,
|
| 583 |
+
"loss": 2.0415,
|
| 584 |
+
"step": 1920
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 1.35,
|
| 588 |
+
"learning_rate": 1.4631505898681473e-05,
|
| 589 |
+
"loss": 2.0762,
|
| 590 |
+
"step": 1940
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 1.36,
|
| 594 |
+
"learning_rate": 1.4575988896599585e-05,
|
| 595 |
+
"loss": 2.0755,
|
| 596 |
+
"step": 1960
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 1.37,
|
| 600 |
+
"learning_rate": 1.4520471894517699e-05,
|
| 601 |
+
"loss": 2.0555,
|
| 602 |
+
"step": 1980
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 1.39,
|
| 606 |
+
"learning_rate": 1.446495489243581e-05,
|
| 607 |
+
"loss": 2.0714,
|
| 608 |
+
"step": 2000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 1.4,
|
| 612 |
+
"learning_rate": 1.4409437890353921e-05,
|
| 613 |
+
"loss": 2.055,
|
| 614 |
+
"step": 2020
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 1.42,
|
| 618 |
+
"learning_rate": 1.4353920888272035e-05,
|
| 619 |
+
"loss": 2.0359,
|
| 620 |
+
"step": 2040
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 1.43,
|
| 624 |
+
"learning_rate": 1.4298403886190147e-05,
|
| 625 |
+
"loss": 2.0178,
|
| 626 |
+
"step": 2060
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 1.44,
|
| 630 |
+
"learning_rate": 1.4242886884108261e-05,
|
| 631 |
+
"loss": 1.9921,
|
| 632 |
+
"step": 2080
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 1.46,
|
| 636 |
+
"learning_rate": 1.4187369882026371e-05,
|
| 637 |
+
"loss": 2.0643,
|
| 638 |
+
"step": 2100
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 1.47,
|
| 642 |
+
"learning_rate": 1.4131852879944483e-05,
|
| 643 |
+
"loss": 2.055,
|
| 644 |
+
"step": 2120
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 1.49,
|
| 648 |
+
"learning_rate": 1.4076335877862597e-05,
|
| 649 |
+
"loss": 2.0363,
|
| 650 |
+
"step": 2140
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 1.5,
|
| 654 |
+
"learning_rate": 1.4020818875780709e-05,
|
| 655 |
+
"loss": 2.0468,
|
| 656 |
+
"step": 2160
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 1.51,
|
| 660 |
+
"learning_rate": 1.3965301873698823e-05,
|
| 661 |
+
"loss": 2.0616,
|
| 662 |
+
"step": 2180
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 1.53,
|
| 666 |
+
"learning_rate": 1.3909784871616933e-05,
|
| 667 |
+
"loss": 2.0377,
|
| 668 |
+
"step": 2200
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 1.54,
|
| 672 |
+
"learning_rate": 1.3854267869535045e-05,
|
| 673 |
+
"loss": 2.0459,
|
| 674 |
+
"step": 2220
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 1.55,
|
| 678 |
+
"learning_rate": 1.3798750867453159e-05,
|
| 679 |
+
"loss": 2.0649,
|
| 680 |
+
"step": 2240
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 1.57,
|
| 684 |
+
"learning_rate": 1.3743233865371271e-05,
|
| 685 |
+
"loss": 2.0639,
|
| 686 |
+
"step": 2260
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 1.58,
|
| 690 |
+
"learning_rate": 1.3687716863289385e-05,
|
| 691 |
+
"loss": 2.045,
|
| 692 |
+
"step": 2280
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 1.6,
|
| 696 |
+
"learning_rate": 1.3632199861207495e-05,
|
| 697 |
+
"loss": 2.0513,
|
| 698 |
+
"step": 2300
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 1.61,
|
| 702 |
+
"learning_rate": 1.3576682859125607e-05,
|
| 703 |
+
"loss": 2.0429,
|
| 704 |
+
"step": 2320
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 1.62,
|
| 708 |
+
"learning_rate": 1.3523941707147814e-05,
|
| 709 |
+
"loss": 2.0545,
|
| 710 |
+
"step": 2340
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 1.64,
|
| 714 |
+
"learning_rate": 1.3468424705065928e-05,
|
| 715 |
+
"loss": 2.0548,
|
| 716 |
+
"step": 2360
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 1.65,
|
| 720 |
+
"learning_rate": 1.341290770298404e-05,
|
| 721 |
+
"loss": 2.071,
|
| 722 |
+
"step": 2380
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 1.67,
|
| 726 |
+
"learning_rate": 1.3357390700902154e-05,
|
| 727 |
+
"loss": 2.0398,
|
| 728 |
+
"step": 2400
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 1.68,
|
| 732 |
+
"learning_rate": 1.3301873698820264e-05,
|
| 733 |
+
"loss": 2.0098,
|
| 734 |
+
"step": 2420
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 1.69,
|
| 738 |
+
"learning_rate": 1.3246356696738378e-05,
|
| 739 |
+
"loss": 2.0455,
|
| 740 |
+
"step": 2440
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 1.71,
|
| 744 |
+
"learning_rate": 1.319083969465649e-05,
|
| 745 |
+
"loss": 2.0545,
|
| 746 |
+
"step": 2460
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 1.72,
|
| 750 |
+
"learning_rate": 1.3135322692574602e-05,
|
| 751 |
+
"loss": 2.0368,
|
| 752 |
+
"step": 2480
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 1.73,
|
| 756 |
+
"learning_rate": 1.3079805690492715e-05,
|
| 757 |
+
"loss": 2.0542,
|
| 758 |
+
"step": 2500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 1.75,
|
| 762 |
+
"learning_rate": 1.3024288688410826e-05,
|
| 763 |
+
"loss": 2.0422,
|
| 764 |
+
"step": 2520
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 1.76,
|
| 768 |
+
"learning_rate": 1.296877168632894e-05,
|
| 769 |
+
"loss": 2.0206,
|
| 770 |
+
"step": 2540
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 1.78,
|
| 774 |
+
"learning_rate": 1.2913254684247052e-05,
|
| 775 |
+
"loss": 2.043,
|
| 776 |
+
"step": 2560
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 1.79,
|
| 780 |
+
"learning_rate": 1.2857737682165164e-05,
|
| 781 |
+
"loss": 2.0716,
|
| 782 |
+
"step": 2580
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 1.8,
|
| 786 |
+
"learning_rate": 1.2802220680083277e-05,
|
| 787 |
+
"loss": 2.0472,
|
| 788 |
+
"step": 2600
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 1.82,
|
| 792 |
+
"learning_rate": 1.274670367800139e-05,
|
| 793 |
+
"loss": 2.0248,
|
| 794 |
+
"step": 2620
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 1.83,
|
| 798 |
+
"learning_rate": 1.2691186675919503e-05,
|
| 799 |
+
"loss": 2.0106,
|
| 800 |
+
"step": 2640
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 1.85,
|
| 804 |
+
"learning_rate": 1.2635669673837614e-05,
|
| 805 |
+
"loss": 2.0243,
|
| 806 |
+
"step": 2660
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 1.86,
|
| 810 |
+
"learning_rate": 1.2580152671755726e-05,
|
| 811 |
+
"loss": 2.0447,
|
| 812 |
+
"step": 2680
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 1.87,
|
| 816 |
+
"learning_rate": 1.252463566967384e-05,
|
| 817 |
+
"loss": 2.037,
|
| 818 |
+
"step": 2700
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 1.89,
|
| 822 |
+
"learning_rate": 1.2469118667591951e-05,
|
| 823 |
+
"loss": 2.0408,
|
| 824 |
+
"step": 2720
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 1.9,
|
| 828 |
+
"learning_rate": 1.2413601665510065e-05,
|
| 829 |
+
"loss": 2.0394,
|
| 830 |
+
"step": 2740
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 1.92,
|
| 834 |
+
"learning_rate": 1.2358084663428175e-05,
|
| 835 |
+
"loss": 2.0685,
|
| 836 |
+
"step": 2760
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 1.93,
|
| 840 |
+
"learning_rate": 1.2302567661346287e-05,
|
| 841 |
+
"loss": 2.0259,
|
| 842 |
+
"step": 2780
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 1.94,
|
| 846 |
+
"learning_rate": 1.2247050659264401e-05,
|
| 847 |
+
"loss": 2.0026,
|
| 848 |
+
"step": 2800
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 1.96,
|
| 852 |
+
"learning_rate": 1.2191533657182513e-05,
|
| 853 |
+
"loss": 2.0251,
|
| 854 |
+
"step": 2820
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 1.97,
|
| 858 |
+
"learning_rate": 1.2136016655100627e-05,
|
| 859 |
+
"loss": 2.0177,
|
| 860 |
+
"step": 2840
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 1.98,
|
| 864 |
+
"learning_rate": 1.2080499653018737e-05,
|
| 865 |
+
"loss": 2.0575,
|
| 866 |
+
"step": 2860
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 2.0,
|
| 870 |
+
"learning_rate": 1.202498265093685e-05,
|
| 871 |
+
"loss": 2.0655,
|
| 872 |
+
"step": 2880
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 2.01,
|
| 876 |
+
"learning_rate": 1.1969465648854963e-05,
|
| 877 |
+
"loss": 2.0154,
|
| 878 |
+
"step": 2900
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 2.03,
|
| 882 |
+
"learning_rate": 1.1913948646773075e-05,
|
| 883 |
+
"loss": 2.0024,
|
| 884 |
+
"step": 2920
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 2.04,
|
| 888 |
+
"learning_rate": 1.1858431644691189e-05,
|
| 889 |
+
"loss": 2.0017,
|
| 890 |
+
"step": 2940
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 2.05,
|
| 894 |
+
"learning_rate": 1.18029146426093e-05,
|
| 895 |
+
"loss": 1.9705,
|
| 896 |
+
"step": 2960
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 2.07,
|
| 900 |
+
"learning_rate": 1.1747397640527411e-05,
|
| 901 |
+
"loss": 2.0569,
|
| 902 |
+
"step": 2980
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 2.08,
|
| 906 |
+
"learning_rate": 1.1691880638445525e-05,
|
| 907 |
+
"loss": 2.0121,
|
| 908 |
+
"step": 3000
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 2.1,
|
| 912 |
+
"learning_rate": 1.1636363636363637e-05,
|
| 913 |
+
"loss": 1.9899,
|
| 914 |
+
"step": 3020
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 2.11,
|
| 918 |
+
"learning_rate": 1.158084663428175e-05,
|
| 919 |
+
"loss": 1.9694,
|
| 920 |
+
"step": 3040
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 2.12,
|
| 924 |
+
"learning_rate": 1.1525329632199863e-05,
|
| 925 |
+
"loss": 2.0268,
|
| 926 |
+
"step": 3060
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 2.14,
|
| 930 |
+
"learning_rate": 1.1469812630117973e-05,
|
| 931 |
+
"loss": 1.9967,
|
| 932 |
+
"step": 3080
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 2.15,
|
| 936 |
+
"learning_rate": 1.1414295628036087e-05,
|
| 937 |
+
"loss": 2.0055,
|
| 938 |
+
"step": 3100
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 2.17,
|
| 942 |
+
"learning_rate": 1.1358778625954199e-05,
|
| 943 |
+
"loss": 1.9879,
|
| 944 |
+
"step": 3120
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 2.18,
|
| 948 |
+
"learning_rate": 1.1303261623872313e-05,
|
| 949 |
+
"loss": 1.9885,
|
| 950 |
+
"step": 3140
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 2.19,
|
| 954 |
+
"learning_rate": 1.1247744621790425e-05,
|
| 955 |
+
"loss": 2.0225,
|
| 956 |
+
"step": 3160
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 2.21,
|
| 960 |
+
"learning_rate": 1.1192227619708535e-05,
|
| 961 |
+
"loss": 2.0058,
|
| 962 |
+
"step": 3180
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 2.22,
|
| 966 |
+
"learning_rate": 1.1136710617626649e-05,
|
| 967 |
+
"loss": 2.0204,
|
| 968 |
+
"step": 3200
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 2.23,
|
| 972 |
+
"learning_rate": 1.1081193615544761e-05,
|
| 973 |
+
"loss": 1.983,
|
| 974 |
+
"step": 3220
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 2.25,
|
| 978 |
+
"learning_rate": 1.1025676613462875e-05,
|
| 979 |
+
"loss": 1.9868,
|
| 980 |
+
"step": 3240
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 2.26,
|
| 984 |
+
"learning_rate": 1.0970159611380987e-05,
|
| 985 |
+
"loss": 2.0118,
|
| 986 |
+
"step": 3260
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 2.28,
|
| 990 |
+
"learning_rate": 1.09146426092991e-05,
|
| 991 |
+
"loss": 2.0158,
|
| 992 |
+
"step": 3280
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 2.29,
|
| 996 |
+
"learning_rate": 1.085912560721721e-05,
|
| 997 |
+
"loss": 2.029,
|
| 998 |
+
"step": 3300
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 2.3,
|
| 1002 |
+
"learning_rate": 1.0803608605135323e-05,
|
| 1003 |
+
"loss": 1.9814,
|
| 1004 |
+
"step": 3320
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 2.32,
|
| 1008 |
+
"learning_rate": 1.0748091603053437e-05,
|
| 1009 |
+
"loss": 2.0236,
|
| 1010 |
+
"step": 3340
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 2.33,
|
| 1014 |
+
"learning_rate": 1.0692574600971549e-05,
|
| 1015 |
+
"loss": 1.9839,
|
| 1016 |
+
"step": 3360
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 2.35,
|
| 1020 |
+
"learning_rate": 1.0637057598889662e-05,
|
| 1021 |
+
"loss": 2.0041,
|
| 1022 |
+
"step": 3380
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 2.36,
|
| 1026 |
+
"learning_rate": 1.0581540596807773e-05,
|
| 1027 |
+
"loss": 2.0303,
|
| 1028 |
+
"step": 3400
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 2.37,
|
| 1032 |
+
"learning_rate": 1.0526023594725885e-05,
|
| 1033 |
+
"loss": 1.9959,
|
| 1034 |
+
"step": 3420
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 2.39,
|
| 1038 |
+
"learning_rate": 1.0470506592643998e-05,
|
| 1039 |
+
"loss": 1.9582,
|
| 1040 |
+
"step": 3440
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 2.4,
|
| 1044 |
+
"learning_rate": 1.041498959056211e-05,
|
| 1045 |
+
"loss": 2.0094,
|
| 1046 |
+
"step": 3460
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 2.41,
|
| 1050 |
+
"learning_rate": 1.0359472588480224e-05,
|
| 1051 |
+
"loss": 2.0021,
|
| 1052 |
+
"step": 3480
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 2.43,
|
| 1056 |
+
"learning_rate": 1.0303955586398335e-05,
|
| 1057 |
+
"loss": 1.9736,
|
| 1058 |
+
"step": 3500
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 2.44,
|
| 1062 |
+
"learning_rate": 1.0248438584316447e-05,
|
| 1063 |
+
"loss": 2.0042,
|
| 1064 |
+
"step": 3520
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 2.46,
|
| 1068 |
+
"learning_rate": 1.019292158223456e-05,
|
| 1069 |
+
"loss": 2.0099,
|
| 1070 |
+
"step": 3540
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 2.47,
|
| 1074 |
+
"learning_rate": 1.0137404580152672e-05,
|
| 1075 |
+
"loss": 2.0219,
|
| 1076 |
+
"step": 3560
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 2.48,
|
| 1080 |
+
"learning_rate": 1.0081887578070786e-05,
|
| 1081 |
+
"loss": 1.984,
|
| 1082 |
+
"step": 3580
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 2.5,
|
| 1086 |
+
"learning_rate": 1.0026370575988898e-05,
|
| 1087 |
+
"loss": 1.9987,
|
| 1088 |
+
"step": 3600
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 2.51,
|
| 1092 |
+
"learning_rate": 9.97085357390701e-06,
|
| 1093 |
+
"loss": 1.9931,
|
| 1094 |
+
"step": 3620
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 2.53,
|
| 1098 |
+
"learning_rate": 9.915336571825122e-06,
|
| 1099 |
+
"loss": 1.9698,
|
| 1100 |
+
"step": 3640
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 2.54,
|
| 1104 |
+
"learning_rate": 9.859819569743234e-06,
|
| 1105 |
+
"loss": 1.9966,
|
| 1106 |
+
"step": 3660
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 2.55,
|
| 1110 |
+
"learning_rate": 9.804302567661346e-06,
|
| 1111 |
+
"loss": 1.9893,
|
| 1112 |
+
"step": 3680
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 2.57,
|
| 1116 |
+
"learning_rate": 9.74878556557946e-06,
|
| 1117 |
+
"loss": 1.9571,
|
| 1118 |
+
"step": 3700
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 2.58,
|
| 1122 |
+
"learning_rate": 9.693268563497572e-06,
|
| 1123 |
+
"loss": 1.9974,
|
| 1124 |
+
"step": 3720
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 2.6,
|
| 1128 |
+
"learning_rate": 9.637751561415684e-06,
|
| 1129 |
+
"loss": 1.9828,
|
| 1130 |
+
"step": 3740
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 2.61,
|
| 1134 |
+
"learning_rate": 9.582234559333796e-06,
|
| 1135 |
+
"loss": 1.971,
|
| 1136 |
+
"step": 3760
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 2.62,
|
| 1140 |
+
"learning_rate": 9.526717557251908e-06,
|
| 1141 |
+
"loss": 1.9799,
|
| 1142 |
+
"step": 3780
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 2.64,
|
| 1146 |
+
"learning_rate": 9.471200555170022e-06,
|
| 1147 |
+
"loss": 1.9926,
|
| 1148 |
+
"step": 3800
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 2.65,
|
| 1152 |
+
"learning_rate": 9.415683553088134e-06,
|
| 1153 |
+
"loss": 1.9738,
|
| 1154 |
+
"step": 3820
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 2.66,
|
| 1158 |
+
"learning_rate": 9.360166551006246e-06,
|
| 1159 |
+
"loss": 1.9917,
|
| 1160 |
+
"step": 3840
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 2.68,
|
| 1164 |
+
"learning_rate": 9.30464954892436e-06,
|
| 1165 |
+
"loss": 2.0237,
|
| 1166 |
+
"step": 3860
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 2.69,
|
| 1170 |
+
"learning_rate": 9.24913254684247e-06,
|
| 1171 |
+
"loss": 2.0106,
|
| 1172 |
+
"step": 3880
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 2.71,
|
| 1176 |
+
"learning_rate": 9.193615544760584e-06,
|
| 1177 |
+
"loss": 2.0059,
|
| 1178 |
+
"step": 3900
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 2.72,
|
| 1182 |
+
"learning_rate": 9.138098542678696e-06,
|
| 1183 |
+
"loss": 2.0034,
|
| 1184 |
+
"step": 3920
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 2.73,
|
| 1188 |
+
"learning_rate": 9.082581540596808e-06,
|
| 1189 |
+
"loss": 2.0007,
|
| 1190 |
+
"step": 3940
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 2.75,
|
| 1194 |
+
"learning_rate": 9.027064538514922e-06,
|
| 1195 |
+
"loss": 1.9945,
|
| 1196 |
+
"step": 3960
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 2.76,
|
| 1200 |
+
"learning_rate": 8.971547536433032e-06,
|
| 1201 |
+
"loss": 1.9635,
|
| 1202 |
+
"step": 3980
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 2.78,
|
| 1206 |
+
"learning_rate": 8.916030534351146e-06,
|
| 1207 |
+
"loss": 2.0221,
|
| 1208 |
+
"step": 4000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 2.79,
|
| 1212 |
+
"learning_rate": 8.860513532269258e-06,
|
| 1213 |
+
"loss": 1.9817,
|
| 1214 |
+
"step": 4020
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 2.8,
|
| 1218 |
+
"learning_rate": 8.804996530187372e-06,
|
| 1219 |
+
"loss": 1.9781,
|
| 1220 |
+
"step": 4040
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 2.82,
|
| 1224 |
+
"learning_rate": 8.749479528105484e-06,
|
| 1225 |
+
"loss": 2.0199,
|
| 1226 |
+
"step": 4060
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 2.83,
|
| 1230 |
+
"learning_rate": 8.693962526023596e-06,
|
| 1231 |
+
"loss": 1.9616,
|
| 1232 |
+
"step": 4080
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 2.85,
|
| 1236 |
+
"learning_rate": 8.638445523941708e-06,
|
| 1237 |
+
"loss": 1.9641,
|
| 1238 |
+
"step": 4100
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 2.86,
|
| 1242 |
+
"learning_rate": 8.58292852185982e-06,
|
| 1243 |
+
"loss": 2.0155,
|
| 1244 |
+
"step": 4120
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 2.87,
|
| 1248 |
+
"learning_rate": 8.527411519777934e-06,
|
| 1249 |
+
"loss": 1.978,
|
| 1250 |
+
"step": 4140
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 2.89,
|
| 1254 |
+
"learning_rate": 8.471894517696046e-06,
|
| 1255 |
+
"loss": 1.9652,
|
| 1256 |
+
"step": 4160
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 2.9,
|
| 1260 |
+
"learning_rate": 8.416377515614158e-06,
|
| 1261 |
+
"loss": 1.9862,
|
| 1262 |
+
"step": 4180
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 2.91,
|
| 1266 |
+
"learning_rate": 8.36086051353227e-06,
|
| 1267 |
+
"loss": 2.0246,
|
| 1268 |
+
"step": 4200
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 2.93,
|
| 1272 |
+
"learning_rate": 8.305343511450382e-06,
|
| 1273 |
+
"loss": 1.9726,
|
| 1274 |
+
"step": 4220
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 2.94,
|
| 1278 |
+
"learning_rate": 8.249826509368495e-06,
|
| 1279 |
+
"loss": 2.0039,
|
| 1280 |
+
"step": 4240
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 2.96,
|
| 1284 |
+
"learning_rate": 8.194309507286607e-06,
|
| 1285 |
+
"loss": 1.9942,
|
| 1286 |
+
"step": 4260
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 2.97,
|
| 1290 |
+
"learning_rate": 8.13879250520472e-06,
|
| 1291 |
+
"loss": 2.0113,
|
| 1292 |
+
"step": 4280
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 2.98,
|
| 1296 |
+
"learning_rate": 8.083275503122832e-06,
|
| 1297 |
+
"loss": 1.9691,
|
| 1298 |
+
"step": 4300
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 3.0,
|
| 1302 |
+
"learning_rate": 8.027758501040944e-06,
|
| 1303 |
+
"loss": 1.994,
|
| 1304 |
+
"step": 4320
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 3.01,
|
| 1308 |
+
"learning_rate": 7.972241498959057e-06,
|
| 1309 |
+
"loss": 2.0017,
|
| 1310 |
+
"step": 4340
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 3.03,
|
| 1314 |
+
"learning_rate": 7.91672449687717e-06,
|
| 1315 |
+
"loss": 1.9494,
|
| 1316 |
+
"step": 4360
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 3.04,
|
| 1320 |
+
"learning_rate": 7.863983344899376e-06,
|
| 1321 |
+
"loss": 1.9718,
|
| 1322 |
+
"step": 4380
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 3.05,
|
| 1326 |
+
"learning_rate": 7.808466342817488e-06,
|
| 1327 |
+
"loss": 1.9305,
|
| 1328 |
+
"step": 4400
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 3.07,
|
| 1332 |
+
"learning_rate": 7.7529493407356e-06,
|
| 1333 |
+
"loss": 1.9697,
|
| 1334 |
+
"step": 4420
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 3.08,
|
| 1338 |
+
"learning_rate": 7.697432338653712e-06,
|
| 1339 |
+
"loss": 1.9658,
|
| 1340 |
+
"step": 4440
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 3.1,
|
| 1344 |
+
"learning_rate": 7.641915336571826e-06,
|
| 1345 |
+
"loss": 1.9424,
|
| 1346 |
+
"step": 4460
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 3.11,
|
| 1350 |
+
"learning_rate": 7.586398334489938e-06,
|
| 1351 |
+
"loss": 1.9312,
|
| 1352 |
+
"step": 4480
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 3.12,
|
| 1356 |
+
"learning_rate": 7.530881332408051e-06,
|
| 1357 |
+
"loss": 2.0042,
|
| 1358 |
+
"step": 4500
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 3.14,
|
| 1362 |
+
"learning_rate": 7.475364330326162e-06,
|
| 1363 |
+
"loss": 1.9868,
|
| 1364 |
+
"step": 4520
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 3.15,
|
| 1368 |
+
"learning_rate": 7.419847328244275e-06,
|
| 1369 |
+
"loss": 1.9679,
|
| 1370 |
+
"step": 4540
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 3.16,
|
| 1374 |
+
"learning_rate": 7.364330326162388e-06,
|
| 1375 |
+
"loss": 1.9448,
|
| 1376 |
+
"step": 4560
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 3.18,
|
| 1380 |
+
"learning_rate": 7.3088133240805e-06,
|
| 1381 |
+
"loss": 1.988,
|
| 1382 |
+
"step": 4580
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 3.19,
|
| 1386 |
+
"learning_rate": 7.253296321998613e-06,
|
| 1387 |
+
"loss": 1.9676,
|
| 1388 |
+
"step": 4600
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 3.21,
|
| 1392 |
+
"learning_rate": 7.197779319916726e-06,
|
| 1393 |
+
"loss": 1.9859,
|
| 1394 |
+
"step": 4620
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 3.22,
|
| 1398 |
+
"learning_rate": 7.142262317834837e-06,
|
| 1399 |
+
"loss": 1.9846,
|
| 1400 |
+
"step": 4640
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 3.23,
|
| 1404 |
+
"learning_rate": 7.08674531575295e-06,
|
| 1405 |
+
"loss": 1.9827,
|
| 1406 |
+
"step": 4660
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 3.25,
|
| 1410 |
+
"learning_rate": 7.031228313671062e-06,
|
| 1411 |
+
"loss": 1.9648,
|
| 1412 |
+
"step": 4680
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 3.26,
|
| 1416 |
+
"learning_rate": 6.975711311589175e-06,
|
| 1417 |
+
"loss": 1.9662,
|
| 1418 |
+
"step": 4700
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 3.28,
|
| 1422 |
+
"learning_rate": 6.920194309507288e-06,
|
| 1423 |
+
"loss": 1.9911,
|
| 1424 |
+
"step": 4720
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 3.29,
|
| 1428 |
+
"learning_rate": 6.864677307425399e-06,
|
| 1429 |
+
"loss": 2.0097,
|
| 1430 |
+
"step": 4740
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 3.3,
|
| 1434 |
+
"learning_rate": 6.809160305343512e-06,
|
| 1435 |
+
"loss": 1.9571,
|
| 1436 |
+
"step": 4760
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 3.32,
|
| 1440 |
+
"learning_rate": 6.753643303261625e-06,
|
| 1441 |
+
"loss": 1.935,
|
| 1442 |
+
"step": 4780
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 3.33,
|
| 1446 |
+
"learning_rate": 6.698126301179737e-06,
|
| 1447 |
+
"loss": 1.9674,
|
| 1448 |
+
"step": 4800
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 3.34,
|
| 1452 |
+
"learning_rate": 6.64260929909785e-06,
|
| 1453 |
+
"loss": 1.9723,
|
| 1454 |
+
"step": 4820
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 3.36,
|
| 1458 |
+
"learning_rate": 6.587092297015961e-06,
|
| 1459 |
+
"loss": 1.9424,
|
| 1460 |
+
"step": 4840
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 3.37,
|
| 1464 |
+
"learning_rate": 6.531575294934074e-06,
|
| 1465 |
+
"loss": 1.9555,
|
| 1466 |
+
"step": 4860
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 3.39,
|
| 1470 |
+
"learning_rate": 6.476058292852187e-06,
|
| 1471 |
+
"loss": 1.9312,
|
| 1472 |
+
"step": 4880
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 3.4,
|
| 1476 |
+
"learning_rate": 6.420541290770299e-06,
|
| 1477 |
+
"loss": 1.9737,
|
| 1478 |
+
"step": 4900
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 3.41,
|
| 1482 |
+
"learning_rate": 6.365024288688412e-06,
|
| 1483 |
+
"loss": 1.9727,
|
| 1484 |
+
"step": 4920
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 3.43,
|
| 1488 |
+
"learning_rate": 6.312283136710619e-06,
|
| 1489 |
+
"loss": 2.0054,
|
| 1490 |
+
"step": 4940
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 3.44,
|
| 1494 |
+
"learning_rate": 6.25676613462873e-06,
|
| 1495 |
+
"loss": 1.962,
|
| 1496 |
+
"step": 4960
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 3.46,
|
| 1500 |
+
"learning_rate": 6.201249132546843e-06,
|
| 1501 |
+
"loss": 1.9729,
|
| 1502 |
+
"step": 4980
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 3.47,
|
| 1506 |
+
"learning_rate": 6.145732130464956e-06,
|
| 1507 |
+
"loss": 1.9911,
|
| 1508 |
+
"step": 5000
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 3.48,
|
| 1512 |
+
"learning_rate": 6.090215128383068e-06,
|
| 1513 |
+
"loss": 1.9298,
|
| 1514 |
+
"step": 5020
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 3.5,
|
| 1518 |
+
"learning_rate": 6.0346981263011805e-06,
|
| 1519 |
+
"loss": 1.9564,
|
| 1520 |
+
"step": 5040
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 3.51,
|
| 1524 |
+
"learning_rate": 5.9791811242192934e-06,
|
| 1525 |
+
"loss": 1.944,
|
| 1526 |
+
"step": 5060
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 3.53,
|
| 1530 |
+
"learning_rate": 5.923664122137405e-06,
|
| 1531 |
+
"loss": 1.946,
|
| 1532 |
+
"step": 5080
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 3.54,
|
| 1536 |
+
"learning_rate": 5.8681471200555175e-06,
|
| 1537 |
+
"loss": 2.0002,
|
| 1538 |
+
"step": 5100
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 3.55,
|
| 1542 |
+
"learning_rate": 5.8126301179736296e-06,
|
| 1543 |
+
"loss": 1.9727,
|
| 1544 |
+
"step": 5120
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 3.57,
|
| 1548 |
+
"learning_rate": 5.7571131158917425e-06,
|
| 1549 |
+
"loss": 1.9354,
|
| 1550 |
+
"step": 5140
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 3.58,
|
| 1554 |
+
"learning_rate": 5.701596113809855e-06,
|
| 1555 |
+
"loss": 1.9635,
|
| 1556 |
+
"step": 5160
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 3.59,
|
| 1560 |
+
"learning_rate": 5.6460791117279665e-06,
|
| 1561 |
+
"loss": 1.9595,
|
| 1562 |
+
"step": 5180
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 3.61,
|
| 1566 |
+
"learning_rate": 5.5905621096460794e-06,
|
| 1567 |
+
"loss": 1.9583,
|
| 1568 |
+
"step": 5200
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 3.62,
|
| 1572 |
+
"learning_rate": 5.535045107564192e-06,
|
| 1573 |
+
"loss": 1.959,
|
| 1574 |
+
"step": 5220
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 3.64,
|
| 1578 |
+
"learning_rate": 5.479528105482304e-06,
|
| 1579 |
+
"loss": 1.9449,
|
| 1580 |
+
"step": 5240
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 3.65,
|
| 1584 |
+
"learning_rate": 5.424011103400417e-06,
|
| 1585 |
+
"loss": 1.9484,
|
| 1586 |
+
"step": 5260
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 3.66,
|
| 1590 |
+
"learning_rate": 5.3684941013185285e-06,
|
| 1591 |
+
"loss": 1.9413,
|
| 1592 |
+
"step": 5280
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 3.68,
|
| 1596 |
+
"learning_rate": 5.312977099236641e-06,
|
| 1597 |
+
"loss": 1.9902,
|
| 1598 |
+
"step": 5300
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 3.69,
|
| 1602 |
+
"learning_rate": 5.257460097154754e-06,
|
| 1603 |
+
"loss": 1.9673,
|
| 1604 |
+
"step": 5320
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 3.71,
|
| 1608 |
+
"learning_rate": 5.201943095072866e-06,
|
| 1609 |
+
"loss": 1.9432,
|
| 1610 |
+
"step": 5340
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 3.72,
|
| 1614 |
+
"learning_rate": 5.146426092990979e-06,
|
| 1615 |
+
"loss": 1.9196,
|
| 1616 |
+
"step": 5360
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 3.73,
|
| 1620 |
+
"learning_rate": 5.090909090909091e-06,
|
| 1621 |
+
"loss": 2.0023,
|
| 1622 |
+
"step": 5380
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 3.75,
|
| 1626 |
+
"learning_rate": 5.035392088827203e-06,
|
| 1627 |
+
"loss": 1.9332,
|
| 1628 |
+
"step": 5400
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 3.76,
|
| 1632 |
+
"learning_rate": 4.979875086745316e-06,
|
| 1633 |
+
"loss": 1.9341,
|
| 1634 |
+
"step": 5420
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 3.78,
|
| 1638 |
+
"learning_rate": 4.924358084663429e-06,
|
| 1639 |
+
"loss": 1.9848,
|
| 1640 |
+
"step": 5440
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 3.79,
|
| 1644 |
+
"learning_rate": 4.868841082581541e-06,
|
| 1645 |
+
"loss": 1.9639,
|
| 1646 |
+
"step": 5460
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 3.8,
|
| 1650 |
+
"learning_rate": 4.813324080499653e-06,
|
| 1651 |
+
"loss": 1.9417,
|
| 1652 |
+
"step": 5480
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 3.82,
|
| 1656 |
+
"learning_rate": 4.757807078417766e-06,
|
| 1657 |
+
"loss": 1.9722,
|
| 1658 |
+
"step": 5500
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 3.83,
|
| 1662 |
+
"learning_rate": 4.702290076335878e-06,
|
| 1663 |
+
"loss": 1.9817,
|
| 1664 |
+
"step": 5520
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 3.84,
|
| 1668 |
+
"learning_rate": 4.646773074253991e-06,
|
| 1669 |
+
"loss": 1.9237,
|
| 1670 |
+
"step": 5540
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 3.86,
|
| 1674 |
+
"learning_rate": 4.591256072172103e-06,
|
| 1675 |
+
"loss": 1.9498,
|
| 1676 |
+
"step": 5560
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 3.87,
|
| 1680 |
+
"learning_rate": 4.535739070090215e-06,
|
| 1681 |
+
"loss": 1.9873,
|
| 1682 |
+
"step": 5580
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 3.89,
|
| 1686 |
+
"learning_rate": 4.480222068008328e-06,
|
| 1687 |
+
"loss": 1.9558,
|
| 1688 |
+
"step": 5600
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 3.9,
|
| 1692 |
+
"learning_rate": 4.42470506592644e-06,
|
| 1693 |
+
"loss": 1.9454,
|
| 1694 |
+
"step": 5620
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 3.91,
|
| 1698 |
+
"learning_rate": 4.369188063844553e-06,
|
| 1699 |
+
"loss": 1.965,
|
| 1700 |
+
"step": 5640
|
| 1701 |
+
},
|
| 1702 |
+
{
|
| 1703 |
+
"epoch": 3.93,
|
| 1704 |
+
"learning_rate": 4.313671061762666e-06,
|
| 1705 |
+
"loss": 1.9538,
|
| 1706 |
+
"step": 5660
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 3.94,
|
| 1710 |
+
"learning_rate": 4.258154059680778e-06,
|
| 1711 |
+
"loss": 1.9655,
|
| 1712 |
+
"step": 5680
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 3.96,
|
| 1716 |
+
"learning_rate": 4.20263705759889e-06,
|
| 1717 |
+
"loss": 1.9635,
|
| 1718 |
+
"step": 5700
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 3.97,
|
| 1722 |
+
"learning_rate": 4.147120055517003e-06,
|
| 1723 |
+
"loss": 1.9721,
|
| 1724 |
+
"step": 5720
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 3.98,
|
| 1728 |
+
"learning_rate": 4.091603053435115e-06,
|
| 1729 |
+
"loss": 2.001,
|
| 1730 |
+
"step": 5740
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 4.0,
|
| 1734 |
+
"learning_rate": 4.036086051353228e-06,
|
| 1735 |
+
"loss": 1.9389,
|
| 1736 |
+
"step": 5760
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 4.01,
|
| 1740 |
+
"learning_rate": 3.98056904927134e-06,
|
| 1741 |
+
"loss": 1.9536,
|
| 1742 |
+
"step": 5780
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 4.02,
|
| 1746 |
+
"learning_rate": 3.925052047189452e-06,
|
| 1747 |
+
"loss": 1.9084,
|
| 1748 |
+
"step": 5800
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 4.04,
|
| 1752 |
+
"learning_rate": 3.869535045107565e-06,
|
| 1753 |
+
"loss": 1.9742,
|
| 1754 |
+
"step": 5820
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 4.05,
|
| 1758 |
+
"learning_rate": 3.8140180430256767e-06,
|
| 1759 |
+
"loss": 1.9404,
|
| 1760 |
+
"step": 5840
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 4.07,
|
| 1764 |
+
"learning_rate": 3.7585010409437896e-06,
|
| 1765 |
+
"loss": 1.9314,
|
| 1766 |
+
"step": 5860
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 4.08,
|
| 1770 |
+
"learning_rate": 3.7029840388619016e-06,
|
| 1771 |
+
"loss": 1.9285,
|
| 1772 |
+
"step": 5880
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 4.09,
|
| 1776 |
+
"learning_rate": 3.647467036780014e-06,
|
| 1777 |
+
"loss": 1.9508,
|
| 1778 |
+
"step": 5900
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 4.11,
|
| 1782 |
+
"learning_rate": 3.5919500346981266e-06,
|
| 1783 |
+
"loss": 1.9263,
|
| 1784 |
+
"step": 5920
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 4.12,
|
| 1788 |
+
"learning_rate": 3.536433032616239e-06,
|
| 1789 |
+
"loss": 1.9733,
|
| 1790 |
+
"step": 5940
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 4.14,
|
| 1794 |
+
"learning_rate": 3.4809160305343515e-06,
|
| 1795 |
+
"loss": 1.9549,
|
| 1796 |
+
"step": 5960
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 4.15,
|
| 1800 |
+
"learning_rate": 3.425399028452464e-06,
|
| 1801 |
+
"loss": 1.9557,
|
| 1802 |
+
"step": 5980
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 4.16,
|
| 1806 |
+
"learning_rate": 3.369882026370576e-06,
|
| 1807 |
+
"loss": 1.9279,
|
| 1808 |
+
"step": 6000
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 4.18,
|
| 1812 |
+
"learning_rate": 3.314365024288689e-06,
|
| 1813 |
+
"loss": 1.907,
|
| 1814 |
+
"step": 6020
|
| 1815 |
+
},
|
| 1816 |
+
{
|
| 1817 |
+
"epoch": 4.19,
|
| 1818 |
+
"learning_rate": 3.258848022206801e-06,
|
| 1819 |
+
"loss": 1.933,
|
| 1820 |
+
"step": 6040
|
| 1821 |
+
},
|
| 1822 |
+
{
|
| 1823 |
+
"epoch": 4.21,
|
| 1824 |
+
"learning_rate": 3.2033310201249134e-06,
|
| 1825 |
+
"loss": 1.9399,
|
| 1826 |
+
"step": 6060
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"epoch": 4.22,
|
| 1830 |
+
"learning_rate": 3.1478140180430263e-06,
|
| 1831 |
+
"loss": 1.9493,
|
| 1832 |
+
"step": 6080
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 4.23,
|
| 1836 |
+
"learning_rate": 3.0922970159611383e-06,
|
| 1837 |
+
"loss": 1.9823,
|
| 1838 |
+
"step": 6100
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"epoch": 4.25,
|
| 1842 |
+
"learning_rate": 3.036780013879251e-06,
|
| 1843 |
+
"loss": 1.9645,
|
| 1844 |
+
"step": 6120
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 4.26,
|
| 1848 |
+
"learning_rate": 2.9812630117973633e-06,
|
| 1849 |
+
"loss": 1.9479,
|
| 1850 |
+
"step": 6140
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 4.27,
|
| 1854 |
+
"learning_rate": 2.9257460097154753e-06,
|
| 1855 |
+
"loss": 1.9512,
|
| 1856 |
+
"step": 6160
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"epoch": 4.29,
|
| 1860 |
+
"learning_rate": 2.870229007633588e-06,
|
| 1861 |
+
"loss": 1.921,
|
| 1862 |
+
"step": 6180
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 4.3,
|
| 1866 |
+
"learning_rate": 2.8147120055517002e-06,
|
| 1867 |
+
"loss": 1.926,
|
| 1868 |
+
"step": 6200
|
| 1869 |
+
},
|
| 1870 |
+
{
|
| 1871 |
+
"epoch": 4.32,
|
| 1872 |
+
"learning_rate": 2.7591950034698127e-06,
|
| 1873 |
+
"loss": 1.9395,
|
| 1874 |
+
"step": 6220
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 4.33,
|
| 1878 |
+
"learning_rate": 2.7036780013879256e-06,
|
| 1879 |
+
"loss": 1.9876,
|
| 1880 |
+
"step": 6240
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 4.34,
|
| 1884 |
+
"learning_rate": 2.6481609993060377e-06,
|
| 1885 |
+
"loss": 1.9327,
|
| 1886 |
+
"step": 6260
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 4.36,
|
| 1890 |
+
"learning_rate": 2.595419847328244e-06,
|
| 1891 |
+
"loss": 1.9638,
|
| 1892 |
+
"step": 6280
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 4.37,
|
| 1896 |
+
"learning_rate": 2.539902845246357e-06,
|
| 1897 |
+
"loss": 1.959,
|
| 1898 |
+
"step": 6300
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 4.39,
|
| 1902 |
+
"learning_rate": 2.4843858431644695e-06,
|
| 1903 |
+
"loss": 1.9385,
|
| 1904 |
+
"step": 6320
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 4.4,
|
| 1908 |
+
"learning_rate": 2.431644691186676e-06,
|
| 1909 |
+
"loss": 1.9325,
|
| 1910 |
+
"step": 6340
|
| 1911 |
+
},
|
| 1912 |
+
{
|
| 1913 |
+
"epoch": 4.41,
|
| 1914 |
+
"learning_rate": 2.3761276891047885e-06,
|
| 1915 |
+
"loss": 1.9624,
|
| 1916 |
+
"step": 6360
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 4.43,
|
| 1920 |
+
"learning_rate": 2.320610687022901e-06,
|
| 1921 |
+
"loss": 1.9417,
|
| 1922 |
+
"step": 6380
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 4.44,
|
| 1926 |
+
"learning_rate": 2.2650936849410135e-06,
|
| 1927 |
+
"loss": 1.9376,
|
| 1928 |
+
"step": 6400
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 4.46,
|
| 1932 |
+
"learning_rate": 2.209576682859126e-06,
|
| 1933 |
+
"loss": 1.9207,
|
| 1934 |
+
"step": 6420
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 4.47,
|
| 1938 |
+
"learning_rate": 2.1540596807772384e-06,
|
| 1939 |
+
"loss": 1.9202,
|
| 1940 |
+
"step": 6440
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 4.48,
|
| 1944 |
+
"learning_rate": 2.0985426786953504e-06,
|
| 1945 |
+
"loss": 1.9181,
|
| 1946 |
+
"step": 6460
|
| 1947 |
+
},
|
| 1948 |
+
{
|
| 1949 |
+
"epoch": 4.5,
|
| 1950 |
+
"learning_rate": 2.043025676613463e-06,
|
| 1951 |
+
"loss": 1.9712,
|
| 1952 |
+
"step": 6480
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 4.51,
|
| 1956 |
+
"learning_rate": 1.9875086745315754e-06,
|
| 1957 |
+
"loss": 1.9433,
|
| 1958 |
+
"step": 6500
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 4.52,
|
| 1962 |
+
"learning_rate": 1.931991672449688e-06,
|
| 1963 |
+
"loss": 1.9602,
|
| 1964 |
+
"step": 6520
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 4.54,
|
| 1968 |
+
"learning_rate": 1.8764746703678003e-06,
|
| 1969 |
+
"loss": 1.9049,
|
| 1970 |
+
"step": 6540
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 4.55,
|
| 1974 |
+
"learning_rate": 1.8209576682859126e-06,
|
| 1975 |
+
"loss": 1.8855,
|
| 1976 |
+
"step": 6560
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 4.57,
|
| 1980 |
+
"learning_rate": 1.7654406662040252e-06,
|
| 1981 |
+
"loss": 1.9369,
|
| 1982 |
+
"step": 6580
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 4.58,
|
| 1986 |
+
"learning_rate": 1.7099236641221375e-06,
|
| 1987 |
+
"loss": 1.9644,
|
| 1988 |
+
"step": 6600
|
| 1989 |
+
},
|
| 1990 |
+
{
|
| 1991 |
+
"epoch": 4.59,
|
| 1992 |
+
"learning_rate": 1.65440666204025e-06,
|
| 1993 |
+
"loss": 1.9247,
|
| 1994 |
+
"step": 6620
|
| 1995 |
+
},
|
| 1996 |
+
{
|
| 1997 |
+
"epoch": 4.61,
|
| 1998 |
+
"learning_rate": 1.5988896599583622e-06,
|
| 1999 |
+
"loss": 1.9326,
|
| 2000 |
+
"step": 6640
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 4.62,
|
| 2004 |
+
"learning_rate": 1.543372657876475e-06,
|
| 2005 |
+
"loss": 1.9486,
|
| 2006 |
+
"step": 6660
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 4.64,
|
| 2010 |
+
"learning_rate": 1.4878556557945872e-06,
|
| 2011 |
+
"loss": 1.9217,
|
| 2012 |
+
"step": 6680
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 4.65,
|
| 2016 |
+
"learning_rate": 1.4323386537126996e-06,
|
| 2017 |
+
"loss": 1.9574,
|
| 2018 |
+
"step": 6700
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 4.66,
|
| 2022 |
+
"learning_rate": 1.3768216516308119e-06,
|
| 2023 |
+
"loss": 1.9716,
|
| 2024 |
+
"step": 6720
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 4.68,
|
| 2028 |
+
"learning_rate": 1.3213046495489246e-06,
|
| 2029 |
+
"loss": 1.9439,
|
| 2030 |
+
"step": 6740
|
| 2031 |
+
},
|
| 2032 |
+
{
|
| 2033 |
+
"epoch": 4.69,
|
| 2034 |
+
"learning_rate": 1.265787647467037e-06,
|
| 2035 |
+
"loss": 1.9521,
|
| 2036 |
+
"step": 6760
|
| 2037 |
+
},
|
| 2038 |
+
{
|
| 2039 |
+
"epoch": 4.71,
|
| 2040 |
+
"learning_rate": 1.2102706453851493e-06,
|
| 2041 |
+
"loss": 1.9322,
|
| 2042 |
+
"step": 6780
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 4.72,
|
| 2046 |
+
"learning_rate": 1.1547536433032617e-06,
|
| 2047 |
+
"loss": 1.9674,
|
| 2048 |
+
"step": 6800
|
| 2049 |
+
},
|
| 2050 |
+
{
|
| 2051 |
+
"epoch": 4.73,
|
| 2052 |
+
"learning_rate": 1.0992366412213742e-06,
|
| 2053 |
+
"loss": 1.9346,
|
| 2054 |
+
"step": 6820
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 4.75,
|
| 2058 |
+
"learning_rate": 1.0437196391394867e-06,
|
| 2059 |
+
"loss": 1.965,
|
| 2060 |
+
"step": 6840
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 4.76,
|
| 2064 |
+
"learning_rate": 9.88202637057599e-07,
|
| 2065 |
+
"loss": 1.9198,
|
| 2066 |
+
"step": 6860
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 4.77,
|
| 2070 |
+
"learning_rate": 9.326856349757114e-07,
|
| 2071 |
+
"loss": 1.9703,
|
| 2072 |
+
"step": 6880
|
| 2073 |
+
},
|
| 2074 |
+
{
|
| 2075 |
+
"epoch": 4.79,
|
| 2076 |
+
"learning_rate": 8.771686328938238e-07,
|
| 2077 |
+
"loss": 1.9539,
|
| 2078 |
+
"step": 6900
|
| 2079 |
+
},
|
| 2080 |
+
{
|
| 2081 |
+
"epoch": 4.8,
|
| 2082 |
+
"learning_rate": 8.216516308119362e-07,
|
| 2083 |
+
"loss": 1.9268,
|
| 2084 |
+
"step": 6920
|
| 2085 |
+
},
|
| 2086 |
+
{
|
| 2087 |
+
"epoch": 4.82,
|
| 2088 |
+
"learning_rate": 7.661346287300486e-07,
|
| 2089 |
+
"loss": 1.9492,
|
| 2090 |
+
"step": 6940
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 4.83,
|
| 2094 |
+
"learning_rate": 7.106176266481612e-07,
|
| 2095 |
+
"loss": 1.9533,
|
| 2096 |
+
"step": 6960
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 4.84,
|
| 2100 |
+
"learning_rate": 6.551006245662734e-07,
|
| 2101 |
+
"loss": 1.9536,
|
| 2102 |
+
"step": 6980
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 4.86,
|
| 2106 |
+
"learning_rate": 5.995836224843859e-07,
|
| 2107 |
+
"loss": 1.9381,
|
| 2108 |
+
"step": 7000
|
| 2109 |
+
},
|
| 2110 |
+
{
|
| 2111 |
+
"epoch": 4.87,
|
| 2112 |
+
"learning_rate": 5.440666204024984e-07,
|
| 2113 |
+
"loss": 1.9477,
|
| 2114 |
+
"step": 7020
|
| 2115 |
+
},
|
| 2116 |
+
{
|
| 2117 |
+
"epoch": 4.89,
|
| 2118 |
+
"learning_rate": 4.885496183206107e-07,
|
| 2119 |
+
"loss": 1.9502,
|
| 2120 |
+
"step": 7040
|
| 2121 |
+
},
|
| 2122 |
+
{
|
| 2123 |
+
"epoch": 4.9,
|
| 2124 |
+
"learning_rate": 4.3303261623872313e-07,
|
| 2125 |
+
"loss": 1.9621,
|
| 2126 |
+
"step": 7060
|
| 2127 |
+
},
|
| 2128 |
+
{
|
| 2129 |
+
"epoch": 4.91,
|
| 2130 |
+
"learning_rate": 3.7751561415683554e-07,
|
| 2131 |
+
"loss": 1.9642,
|
| 2132 |
+
"step": 7080
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 4.93,
|
| 2136 |
+
"learning_rate": 3.2199861207494796e-07,
|
| 2137 |
+
"loss": 1.9513,
|
| 2138 |
+
"step": 7100
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 4.94,
|
| 2142 |
+
"learning_rate": 2.664816099930604e-07,
|
| 2143 |
+
"loss": 1.9607,
|
| 2144 |
+
"step": 7120
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 4.95,
|
| 2148 |
+
"learning_rate": 2.109646079111728e-07,
|
| 2149 |
+
"loss": 1.943,
|
| 2150 |
+
"step": 7140
|
| 2151 |
+
},
|
| 2152 |
+
{
|
| 2153 |
+
"epoch": 4.97,
|
| 2154 |
+
"learning_rate": 1.5544760582928525e-07,
|
| 2155 |
+
"loss": 1.9397,
|
| 2156 |
+
"step": 7160
|
| 2157 |
+
},
|
| 2158 |
+
{
|
| 2159 |
+
"epoch": 4.98,
|
| 2160 |
+
"learning_rate": 9.993060374739764e-08,
|
| 2161 |
+
"loss": 1.9777,
|
| 2162 |
+
"step": 7180
|
| 2163 |
+
},
|
| 2164 |
+
{
|
| 2165 |
+
"epoch": 5.0,
|
| 2166 |
+
"learning_rate": 4.441360166551007e-08,
|
| 2167 |
+
"loss": 1.9451,
|
| 2168 |
+
"step": 7200
|
| 2169 |
+
},
|
| 2170 |
+
{
|
| 2171 |
+
"epoch": 5.0,
|
| 2172 |
+
"step": 7205,
|
| 2173 |
+
"total_flos": 16112557228032.0,
|
| 2174 |
+
"train_loss": 2.03250736604541,
|
| 2175 |
+
"train_runtime": 4701.6865,
|
| 2176 |
+
"train_samples_per_second": 9.194,
|
| 2177 |
+
"train_steps_per_second": 1.532
|
| 2178 |
+
}
|
| 2179 |
+
],
|
| 2180 |
+
"max_steps": 7205,
|
| 2181 |
+
"num_train_epochs": 5,
|
| 2182 |
+
"total_flos": 16112557228032.0,
|
| 2183 |
+
"trial_name": null,
|
| 2184 |
+
"trial_params": null
|
| 2185 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:70881724598f9f47142eb17dfcfe20197a7e2ac911ecbc805300519460143aec
|
| 3 |
+
size 4859
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|