init commit
Browse files- README.md +56 -0
- all_results.json +14 -0
- config.json +41 -0
- eval_results.json +9 -0
- merges.txt +0 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +1 -0
- tokenizer.json +0 -0
- tokenizer_config.json +1 -0
- train_results.json +8 -0
- trainer_state.json +2053 -0
- training_args.bin +3 -0
- vocab.json +0 -0
README.md
ADDED
|
@@ -0,0 +1,56 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
tags:
|
| 3 |
+
- generated_from_trainer
|
| 4 |
+
model-index:
|
| 5 |
+
- name: argpt2-goodreads
|
| 6 |
+
results: []
|
| 7 |
+
---
|
| 8 |
+
|
| 9 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 10 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 11 |
+
|
| 12 |
+
# argpt2-goodreads
|
| 13 |
+
|
| 14 |
+
This model is a fine-tuned version of [gpt2-medium](https://huggingface.co/gpt2-medium) on an unknown dataset.
|
| 15 |
+
It achieves the following results on the evaluation set:
|
| 16 |
+
- Loss: 1.4389
|
| 17 |
+
|
| 18 |
+
## Model description
|
| 19 |
+
|
| 20 |
+
More information needed
|
| 21 |
+
|
| 22 |
+
## Intended uses & limitations
|
| 23 |
+
|
| 24 |
+
More information needed
|
| 25 |
+
|
| 26 |
+
## Training and evaluation data
|
| 27 |
+
|
| 28 |
+
More information needed
|
| 29 |
+
|
| 30 |
+
## Training procedure
|
| 31 |
+
|
| 32 |
+
### Training hyperparameters
|
| 33 |
+
|
| 34 |
+
The following hyperparameters were used during training:
|
| 35 |
+
- learning_rate: 5e-05
|
| 36 |
+
- train_batch_size: 16
|
| 37 |
+
- eval_batch_size: 16
|
| 38 |
+
- seed: 42
|
| 39 |
+
- distributed_type: tpu
|
| 40 |
+
- num_devices: 8
|
| 41 |
+
- total_train_batch_size: 128
|
| 42 |
+
- total_eval_batch_size: 128
|
| 43 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 44 |
+
- lr_scheduler_type: linear
|
| 45 |
+
- num_epochs: 20.0
|
| 46 |
+
|
| 47 |
+
### Training results
|
| 48 |
+
|
| 49 |
+
|
| 50 |
+
|
| 51 |
+
### Framework versions
|
| 52 |
+
|
| 53 |
+
- Transformers 4.13.0.dev0
|
| 54 |
+
- Pytorch 1.10.0+cu102
|
| 55 |
+
- Datasets 1.16.1
|
| 56 |
+
- Tokenizers 0.10.3
|
all_results.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 20.0,
|
| 3 |
+
"eval_loss": 1.438940167427063,
|
| 4 |
+
"eval_runtime": 277.0149,
|
| 5 |
+
"eval_samples": 27329,
|
| 6 |
+
"eval_samples_per_second": 98.655,
|
| 7 |
+
"eval_steps_per_second": 0.773,
|
| 8 |
+
"perplexity": 4.216224955689952,
|
| 9 |
+
"train_loss": 1.4740209875163242,
|
| 10 |
+
"train_runtime": 8294.5138,
|
| 11 |
+
"train_samples": 108110,
|
| 12 |
+
"train_samples_per_second": 260.678,
|
| 13 |
+
"train_steps_per_second": 2.037
|
| 14 |
+
}
|
config.json
ADDED
|
@@ -0,0 +1,41 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "gpt2-medium",
|
| 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": 1024,
|
| 16 |
+
"n_head": 16,
|
| 17 |
+
"n_inner": null,
|
| 18 |
+
"n_layer": 24,
|
| 19 |
+
"n_positions": 1024,
|
| 20 |
+
"n_special": 0,
|
| 21 |
+
"predict_special_tokens": true,
|
| 22 |
+
"reorder_and_upcast_attn": false,
|
| 23 |
+
"resid_pdrop": 0.1,
|
| 24 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 25 |
+
"scale_attn_weights": true,
|
| 26 |
+
"summary_activation": null,
|
| 27 |
+
"summary_first_dropout": 0.1,
|
| 28 |
+
"summary_proj_to_labels": true,
|
| 29 |
+
"summary_type": "cls_index",
|
| 30 |
+
"summary_use_proj": true,
|
| 31 |
+
"task_specific_params": {
|
| 32 |
+
"text-generation": {
|
| 33 |
+
"do_sample": true,
|
| 34 |
+
"max_length": 50
|
| 35 |
+
}
|
| 36 |
+
},
|
| 37 |
+
"torch_dtype": "float32",
|
| 38 |
+
"transformers_version": "4.13.0.dev0",
|
| 39 |
+
"use_cache": true,
|
| 40 |
+
"vocab_size": 50257
|
| 41 |
+
}
|
eval_results.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 20.0,
|
| 3 |
+
"eval_loss": 1.438940167427063,
|
| 4 |
+
"eval_runtime": 277.0149,
|
| 5 |
+
"eval_samples": 27329,
|
| 6 |
+
"eval_samples_per_second": 98.655,
|
| 7 |
+
"eval_steps_per_second": 0.773,
|
| 8 |
+
"perplexity": 4.216224955689952
|
| 9 |
+
}
|
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:dfc83030c315717a319bf0a748d9343d47b1722631f4accfd11b93a1c1ac4c6b
|
| 3 |
+
size 1650414999
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
| 1 |
+
{"bos_token": "<|endoftext|>", "eos_token": "<|endoftext|>", "unk_token": "<|endoftext|>"}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
| 1 |
+
{"unk_token": "<|endoftext|>", "bos_token": "<|endoftext|>", "eos_token": "<|endoftext|>", "add_prefix_space": false, "model_max_length": 1024, "special_tokens_map_file": null, "name_or_path": "gpt2-medium", "tokenizer_class": "GPT2Tokenizer"}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 20.0,
|
| 3 |
+
"train_loss": 1.4740209875163242,
|
| 4 |
+
"train_runtime": 8294.5138,
|
| 5 |
+
"train_samples": 108110,
|
| 6 |
+
"train_samples_per_second": 260.678,
|
| 7 |
+
"train_steps_per_second": 2.037
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2053 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 20.0,
|
| 5 |
+
"global_step": 16900,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.06,
|
| 12 |
+
"learning_rate": 4.985207100591716e-05,
|
| 13 |
+
"loss": 2.2222,
|
| 14 |
+
"step": 50
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.12,
|
| 18 |
+
"learning_rate": 4.970414201183432e-05,
|
| 19 |
+
"loss": 2.043,
|
| 20 |
+
"step": 100
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.18,
|
| 24 |
+
"learning_rate": 4.955621301775148e-05,
|
| 25 |
+
"loss": 1.9649,
|
| 26 |
+
"step": 150
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.24,
|
| 30 |
+
"learning_rate": 4.9408284023668644e-05,
|
| 31 |
+
"loss": 1.9144,
|
| 32 |
+
"step": 200
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.3,
|
| 36 |
+
"learning_rate": 4.92603550295858e-05,
|
| 37 |
+
"loss": 1.8759,
|
| 38 |
+
"step": 250
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.36,
|
| 42 |
+
"learning_rate": 4.9112426035502965e-05,
|
| 43 |
+
"loss": 1.8501,
|
| 44 |
+
"step": 300
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.41,
|
| 48 |
+
"learning_rate": 4.896449704142012e-05,
|
| 49 |
+
"loss": 1.8282,
|
| 50 |
+
"step": 350
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.47,
|
| 54 |
+
"learning_rate": 4.881656804733728e-05,
|
| 55 |
+
"loss": 1.8055,
|
| 56 |
+
"step": 400
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.53,
|
| 60 |
+
"learning_rate": 4.866863905325444e-05,
|
| 61 |
+
"loss": 1.7912,
|
| 62 |
+
"step": 450
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.59,
|
| 66 |
+
"learning_rate": 4.85207100591716e-05,
|
| 67 |
+
"loss": 1.7801,
|
| 68 |
+
"step": 500
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.65,
|
| 72 |
+
"learning_rate": 4.8372781065088756e-05,
|
| 73 |
+
"loss": 1.7691,
|
| 74 |
+
"step": 550
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.71,
|
| 78 |
+
"learning_rate": 4.822485207100592e-05,
|
| 79 |
+
"loss": 1.753,
|
| 80 |
+
"step": 600
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.77,
|
| 84 |
+
"learning_rate": 4.8076923076923084e-05,
|
| 85 |
+
"loss": 1.7455,
|
| 86 |
+
"step": 650
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.83,
|
| 90 |
+
"learning_rate": 4.792899408284024e-05,
|
| 91 |
+
"loss": 1.7326,
|
| 92 |
+
"step": 700
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.89,
|
| 96 |
+
"learning_rate": 4.77810650887574e-05,
|
| 97 |
+
"loss": 1.7271,
|
| 98 |
+
"step": 750
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.95,
|
| 102 |
+
"learning_rate": 4.7633136094674555e-05,
|
| 103 |
+
"loss": 1.7172,
|
| 104 |
+
"step": 800
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 1.01,
|
| 108 |
+
"learning_rate": 4.748520710059172e-05,
|
| 109 |
+
"loss": 1.7097,
|
| 110 |
+
"step": 850
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 1.07,
|
| 114 |
+
"learning_rate": 4.7337278106508875e-05,
|
| 115 |
+
"loss": 1.698,
|
| 116 |
+
"step": 900
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 1.12,
|
| 120 |
+
"learning_rate": 4.718934911242604e-05,
|
| 121 |
+
"loss": 1.6881,
|
| 122 |
+
"step": 950
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 1.18,
|
| 126 |
+
"learning_rate": 4.7041420118343196e-05,
|
| 127 |
+
"loss": 1.6812,
|
| 128 |
+
"step": 1000
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 1.24,
|
| 132 |
+
"learning_rate": 4.689349112426036e-05,
|
| 133 |
+
"loss": 1.6711,
|
| 134 |
+
"step": 1050
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 1.3,
|
| 138 |
+
"learning_rate": 4.674556213017752e-05,
|
| 139 |
+
"loss": 1.6749,
|
| 140 |
+
"step": 1100
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 1.36,
|
| 144 |
+
"learning_rate": 4.6597633136094674e-05,
|
| 145 |
+
"loss": 1.6684,
|
| 146 |
+
"step": 1150
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 1.42,
|
| 150 |
+
"learning_rate": 4.644970414201184e-05,
|
| 151 |
+
"loss": 1.6656,
|
| 152 |
+
"step": 1200
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 1.48,
|
| 156 |
+
"learning_rate": 4.6301775147928994e-05,
|
| 157 |
+
"loss": 1.6604,
|
| 158 |
+
"step": 1250
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 1.54,
|
| 162 |
+
"learning_rate": 4.615384615384616e-05,
|
| 163 |
+
"loss": 1.6578,
|
| 164 |
+
"step": 1300
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 1.6,
|
| 168 |
+
"learning_rate": 4.6005917159763315e-05,
|
| 169 |
+
"loss": 1.6527,
|
| 170 |
+
"step": 1350
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 1.66,
|
| 174 |
+
"learning_rate": 4.585798816568048e-05,
|
| 175 |
+
"loss": 1.6502,
|
| 176 |
+
"step": 1400
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 1.72,
|
| 180 |
+
"learning_rate": 4.5710059171597636e-05,
|
| 181 |
+
"loss": 1.6485,
|
| 182 |
+
"step": 1450
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 1.78,
|
| 186 |
+
"learning_rate": 4.556213017751479e-05,
|
| 187 |
+
"loss": 1.6418,
|
| 188 |
+
"step": 1500
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 1.83,
|
| 192 |
+
"learning_rate": 4.5414201183431957e-05,
|
| 193 |
+
"loss": 1.6397,
|
| 194 |
+
"step": 1550
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 1.89,
|
| 198 |
+
"learning_rate": 4.5266272189349114e-05,
|
| 199 |
+
"loss": 1.6301,
|
| 200 |
+
"step": 1600
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 1.95,
|
| 204 |
+
"learning_rate": 4.511834319526627e-05,
|
| 205 |
+
"loss": 1.6273,
|
| 206 |
+
"step": 1650
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 2.01,
|
| 210 |
+
"learning_rate": 4.4970414201183434e-05,
|
| 211 |
+
"loss": 1.6227,
|
| 212 |
+
"step": 1700
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 2.07,
|
| 216 |
+
"learning_rate": 4.48224852071006e-05,
|
| 217 |
+
"loss": 1.6148,
|
| 218 |
+
"step": 1750
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 2.13,
|
| 222 |
+
"learning_rate": 4.4674556213017755e-05,
|
| 223 |
+
"loss": 1.6132,
|
| 224 |
+
"step": 1800
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 2.19,
|
| 228 |
+
"learning_rate": 4.452662721893491e-05,
|
| 229 |
+
"loss": 1.6094,
|
| 230 |
+
"step": 1850
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 2.25,
|
| 234 |
+
"learning_rate": 4.437869822485207e-05,
|
| 235 |
+
"loss": 1.605,
|
| 236 |
+
"step": 1900
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 2.31,
|
| 240 |
+
"learning_rate": 4.423076923076923e-05,
|
| 241 |
+
"loss": 1.6058,
|
| 242 |
+
"step": 1950
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 2.37,
|
| 246 |
+
"learning_rate": 4.408284023668639e-05,
|
| 247 |
+
"loss": 1.6,
|
| 248 |
+
"step": 2000
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 2.43,
|
| 252 |
+
"learning_rate": 4.393491124260355e-05,
|
| 253 |
+
"loss": 1.597,
|
| 254 |
+
"step": 2050
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 2.49,
|
| 258 |
+
"learning_rate": 4.378698224852072e-05,
|
| 259 |
+
"loss": 1.5985,
|
| 260 |
+
"step": 2100
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 2.54,
|
| 264 |
+
"learning_rate": 4.3639053254437874e-05,
|
| 265 |
+
"loss": 1.5941,
|
| 266 |
+
"step": 2150
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 2.6,
|
| 270 |
+
"learning_rate": 4.349112426035503e-05,
|
| 271 |
+
"loss": 1.5953,
|
| 272 |
+
"step": 2200
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 2.66,
|
| 276 |
+
"learning_rate": 4.334319526627219e-05,
|
| 277 |
+
"loss": 1.5922,
|
| 278 |
+
"step": 2250
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 2.72,
|
| 282 |
+
"learning_rate": 4.319526627218935e-05,
|
| 283 |
+
"loss": 1.5908,
|
| 284 |
+
"step": 2300
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 2.78,
|
| 288 |
+
"learning_rate": 4.304733727810651e-05,
|
| 289 |
+
"loss": 1.5827,
|
| 290 |
+
"step": 2350
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 2.84,
|
| 294 |
+
"learning_rate": 4.289940828402367e-05,
|
| 295 |
+
"loss": 1.585,
|
| 296 |
+
"step": 2400
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 2.9,
|
| 300 |
+
"learning_rate": 4.275147928994083e-05,
|
| 301 |
+
"loss": 1.5853,
|
| 302 |
+
"step": 2450
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 2.96,
|
| 306 |
+
"learning_rate": 4.260355029585799e-05,
|
| 307 |
+
"loss": 1.5837,
|
| 308 |
+
"step": 2500
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 3.02,
|
| 312 |
+
"learning_rate": 4.245562130177515e-05,
|
| 313 |
+
"loss": 1.5704,
|
| 314 |
+
"step": 2550
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 3.08,
|
| 318 |
+
"learning_rate": 4.230769230769231e-05,
|
| 319 |
+
"loss": 1.5664,
|
| 320 |
+
"step": 2600
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 3.14,
|
| 324 |
+
"learning_rate": 4.215976331360947e-05,
|
| 325 |
+
"loss": 1.5649,
|
| 326 |
+
"step": 2650
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 3.2,
|
| 330 |
+
"learning_rate": 4.201183431952663e-05,
|
| 331 |
+
"loss": 1.5639,
|
| 332 |
+
"step": 2700
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 3.25,
|
| 336 |
+
"learning_rate": 4.1863905325443785e-05,
|
| 337 |
+
"loss": 1.5623,
|
| 338 |
+
"step": 2750
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 3.31,
|
| 342 |
+
"learning_rate": 4.171597633136095e-05,
|
| 343 |
+
"loss": 1.5616,
|
| 344 |
+
"step": 2800
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 3.37,
|
| 348 |
+
"learning_rate": 4.156804733727811e-05,
|
| 349 |
+
"loss": 1.5603,
|
| 350 |
+
"step": 2850
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 3.43,
|
| 354 |
+
"learning_rate": 4.142011834319527e-05,
|
| 355 |
+
"loss": 1.56,
|
| 356 |
+
"step": 2900
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 3.49,
|
| 360 |
+
"learning_rate": 4.1272189349112426e-05,
|
| 361 |
+
"loss": 1.5552,
|
| 362 |
+
"step": 2950
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 3.55,
|
| 366 |
+
"learning_rate": 4.112426035502959e-05,
|
| 367 |
+
"loss": 1.5563,
|
| 368 |
+
"step": 3000
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 3.61,
|
| 372 |
+
"learning_rate": 4.097633136094675e-05,
|
| 373 |
+
"loss": 1.5531,
|
| 374 |
+
"step": 3050
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 3.67,
|
| 378 |
+
"learning_rate": 4.0828402366863904e-05,
|
| 379 |
+
"loss": 1.5525,
|
| 380 |
+
"step": 3100
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 3.73,
|
| 384 |
+
"learning_rate": 4.068047337278107e-05,
|
| 385 |
+
"loss": 1.5487,
|
| 386 |
+
"step": 3150
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 3.79,
|
| 390 |
+
"learning_rate": 4.053254437869823e-05,
|
| 391 |
+
"loss": 1.557,
|
| 392 |
+
"step": 3200
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 3.85,
|
| 396 |
+
"learning_rate": 4.038461538461539e-05,
|
| 397 |
+
"loss": 1.5508,
|
| 398 |
+
"step": 3250
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 3.91,
|
| 402 |
+
"learning_rate": 4.0236686390532545e-05,
|
| 403 |
+
"loss": 1.5467,
|
| 404 |
+
"step": 3300
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 3.96,
|
| 408 |
+
"learning_rate": 4.00887573964497e-05,
|
| 409 |
+
"loss": 1.546,
|
| 410 |
+
"step": 3350
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 4.02,
|
| 414 |
+
"learning_rate": 3.9940828402366866e-05,
|
| 415 |
+
"loss": 1.5368,
|
| 416 |
+
"step": 3400
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 4.08,
|
| 420 |
+
"learning_rate": 3.979289940828402e-05,
|
| 421 |
+
"loss": 1.5326,
|
| 422 |
+
"step": 3450
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 4.14,
|
| 426 |
+
"learning_rate": 3.964497041420119e-05,
|
| 427 |
+
"loss": 1.5352,
|
| 428 |
+
"step": 3500
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 4.2,
|
| 432 |
+
"learning_rate": 3.9497041420118344e-05,
|
| 433 |
+
"loss": 1.5311,
|
| 434 |
+
"step": 3550
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 4.26,
|
| 438 |
+
"learning_rate": 3.934911242603551e-05,
|
| 439 |
+
"loss": 1.5299,
|
| 440 |
+
"step": 3600
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 4.32,
|
| 444 |
+
"learning_rate": 3.9201183431952664e-05,
|
| 445 |
+
"loss": 1.5301,
|
| 446 |
+
"step": 3650
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 4.38,
|
| 450 |
+
"learning_rate": 3.905325443786982e-05,
|
| 451 |
+
"loss": 1.5334,
|
| 452 |
+
"step": 3700
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 4.44,
|
| 456 |
+
"learning_rate": 3.8905325443786985e-05,
|
| 457 |
+
"loss": 1.5256,
|
| 458 |
+
"step": 3750
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 4.5,
|
| 462 |
+
"learning_rate": 3.875739644970414e-05,
|
| 463 |
+
"loss": 1.5246,
|
| 464 |
+
"step": 3800
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 4.56,
|
| 468 |
+
"learning_rate": 3.86094674556213e-05,
|
| 469 |
+
"loss": 1.5305,
|
| 470 |
+
"step": 3850
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 4.62,
|
| 474 |
+
"learning_rate": 3.846153846153846e-05,
|
| 475 |
+
"loss": 1.5234,
|
| 476 |
+
"step": 3900
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 4.67,
|
| 480 |
+
"learning_rate": 3.8313609467455627e-05,
|
| 481 |
+
"loss": 1.5262,
|
| 482 |
+
"step": 3950
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 4.73,
|
| 486 |
+
"learning_rate": 3.8165680473372784e-05,
|
| 487 |
+
"loss": 1.5267,
|
| 488 |
+
"step": 4000
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 4.79,
|
| 492 |
+
"learning_rate": 3.801775147928994e-05,
|
| 493 |
+
"loss": 1.5209,
|
| 494 |
+
"step": 4050
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 4.85,
|
| 498 |
+
"learning_rate": 3.7869822485207104e-05,
|
| 499 |
+
"loss": 1.5196,
|
| 500 |
+
"step": 4100
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 4.91,
|
| 504 |
+
"learning_rate": 3.772189349112426e-05,
|
| 505 |
+
"loss": 1.5222,
|
| 506 |
+
"step": 4150
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 4.97,
|
| 510 |
+
"learning_rate": 3.757396449704142e-05,
|
| 511 |
+
"loss": 1.5196,
|
| 512 |
+
"step": 4200
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 5.03,
|
| 516 |
+
"learning_rate": 3.742603550295858e-05,
|
| 517 |
+
"loss": 1.5127,
|
| 518 |
+
"step": 4250
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 5.09,
|
| 522 |
+
"learning_rate": 3.7278106508875746e-05,
|
| 523 |
+
"loss": 1.5052,
|
| 524 |
+
"step": 4300
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 5.15,
|
| 528 |
+
"learning_rate": 3.71301775147929e-05,
|
| 529 |
+
"loss": 1.5066,
|
| 530 |
+
"step": 4350
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 5.21,
|
| 534 |
+
"learning_rate": 3.698224852071006e-05,
|
| 535 |
+
"loss": 1.5059,
|
| 536 |
+
"step": 4400
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 5.27,
|
| 540 |
+
"learning_rate": 3.6834319526627223e-05,
|
| 541 |
+
"loss": 1.5123,
|
| 542 |
+
"step": 4450
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 5.33,
|
| 546 |
+
"learning_rate": 3.668639053254438e-05,
|
| 547 |
+
"loss": 1.5058,
|
| 548 |
+
"step": 4500
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 5.38,
|
| 552 |
+
"learning_rate": 3.653846153846154e-05,
|
| 553 |
+
"loss": 1.5025,
|
| 554 |
+
"step": 4550
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 5.44,
|
| 558 |
+
"learning_rate": 3.63905325443787e-05,
|
| 559 |
+
"loss": 1.5006,
|
| 560 |
+
"step": 4600
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 5.5,
|
| 564 |
+
"learning_rate": 3.6242603550295865e-05,
|
| 565 |
+
"loss": 1.5022,
|
| 566 |
+
"step": 4650
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 5.56,
|
| 570 |
+
"learning_rate": 3.609467455621302e-05,
|
| 571 |
+
"loss": 1.5061,
|
| 572 |
+
"step": 4700
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 5.62,
|
| 576 |
+
"learning_rate": 3.594674556213018e-05,
|
| 577 |
+
"loss": 1.5057,
|
| 578 |
+
"step": 4750
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 5.68,
|
| 582 |
+
"learning_rate": 3.5798816568047336e-05,
|
| 583 |
+
"loss": 1.5022,
|
| 584 |
+
"step": 4800
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 5.74,
|
| 588 |
+
"learning_rate": 3.56508875739645e-05,
|
| 589 |
+
"loss": 1.5022,
|
| 590 |
+
"step": 4850
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 5.8,
|
| 594 |
+
"learning_rate": 3.5502958579881656e-05,
|
| 595 |
+
"loss": 1.5007,
|
| 596 |
+
"step": 4900
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 5.86,
|
| 600 |
+
"learning_rate": 3.5355029585798813e-05,
|
| 601 |
+
"loss": 1.4965,
|
| 602 |
+
"step": 4950
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 5.92,
|
| 606 |
+
"learning_rate": 3.520710059171598e-05,
|
| 607 |
+
"loss": 1.5002,
|
| 608 |
+
"step": 5000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 5.98,
|
| 612 |
+
"learning_rate": 3.505917159763314e-05,
|
| 613 |
+
"loss": 1.4991,
|
| 614 |
+
"step": 5050
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 6.04,
|
| 618 |
+
"learning_rate": 3.49112426035503e-05,
|
| 619 |
+
"loss": 1.4943,
|
| 620 |
+
"step": 5100
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 6.09,
|
| 624 |
+
"learning_rate": 3.4763313609467455e-05,
|
| 625 |
+
"loss": 1.486,
|
| 626 |
+
"step": 5150
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 6.15,
|
| 630 |
+
"learning_rate": 3.461538461538462e-05,
|
| 631 |
+
"loss": 1.486,
|
| 632 |
+
"step": 5200
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 6.21,
|
| 636 |
+
"learning_rate": 3.4467455621301776e-05,
|
| 637 |
+
"loss": 1.4871,
|
| 638 |
+
"step": 5250
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 6.27,
|
| 642 |
+
"learning_rate": 3.431952662721893e-05,
|
| 643 |
+
"loss": 1.4844,
|
| 644 |
+
"step": 5300
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 6.33,
|
| 648 |
+
"learning_rate": 3.4171597633136096e-05,
|
| 649 |
+
"loss": 1.4845,
|
| 650 |
+
"step": 5350
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 6.39,
|
| 654 |
+
"learning_rate": 3.402366863905326e-05,
|
| 655 |
+
"loss": 1.4859,
|
| 656 |
+
"step": 5400
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 6.45,
|
| 660 |
+
"learning_rate": 3.387573964497042e-05,
|
| 661 |
+
"loss": 1.4872,
|
| 662 |
+
"step": 5450
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 6.51,
|
| 666 |
+
"learning_rate": 3.3727810650887574e-05,
|
| 667 |
+
"loss": 1.4816,
|
| 668 |
+
"step": 5500
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 6.57,
|
| 672 |
+
"learning_rate": 3.357988165680474e-05,
|
| 673 |
+
"loss": 1.4868,
|
| 674 |
+
"step": 5550
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 6.63,
|
| 678 |
+
"learning_rate": 3.3431952662721895e-05,
|
| 679 |
+
"loss": 1.4816,
|
| 680 |
+
"step": 5600
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 6.69,
|
| 684 |
+
"learning_rate": 3.328402366863905e-05,
|
| 685 |
+
"loss": 1.4838,
|
| 686 |
+
"step": 5650
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 6.75,
|
| 690 |
+
"learning_rate": 3.3136094674556215e-05,
|
| 691 |
+
"loss": 1.4808,
|
| 692 |
+
"step": 5700
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 6.8,
|
| 696 |
+
"learning_rate": 3.298816568047338e-05,
|
| 697 |
+
"loss": 1.4821,
|
| 698 |
+
"step": 5750
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 6.86,
|
| 702 |
+
"learning_rate": 3.2840236686390536e-05,
|
| 703 |
+
"loss": 1.4828,
|
| 704 |
+
"step": 5800
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 6.92,
|
| 708 |
+
"learning_rate": 3.269230769230769e-05,
|
| 709 |
+
"loss": 1.4768,
|
| 710 |
+
"step": 5850
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 6.98,
|
| 714 |
+
"learning_rate": 3.254437869822485e-05,
|
| 715 |
+
"loss": 1.4822,
|
| 716 |
+
"step": 5900
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 7.04,
|
| 720 |
+
"learning_rate": 3.2396449704142014e-05,
|
| 721 |
+
"loss": 1.4681,
|
| 722 |
+
"step": 5950
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 7.1,
|
| 726 |
+
"learning_rate": 3.224852071005917e-05,
|
| 727 |
+
"loss": 1.4655,
|
| 728 |
+
"step": 6000
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 7.16,
|
| 732 |
+
"learning_rate": 3.210059171597633e-05,
|
| 733 |
+
"loss": 1.4679,
|
| 734 |
+
"step": 6050
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 7.22,
|
| 738 |
+
"learning_rate": 3.195266272189349e-05,
|
| 739 |
+
"loss": 1.4687,
|
| 740 |
+
"step": 6100
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 7.28,
|
| 744 |
+
"learning_rate": 3.1804733727810655e-05,
|
| 745 |
+
"loss": 1.4695,
|
| 746 |
+
"step": 6150
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 7.34,
|
| 750 |
+
"learning_rate": 3.165680473372781e-05,
|
| 751 |
+
"loss": 1.4684,
|
| 752 |
+
"step": 6200
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 7.4,
|
| 756 |
+
"learning_rate": 3.150887573964497e-05,
|
| 757 |
+
"loss": 1.4689,
|
| 758 |
+
"step": 6250
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 7.46,
|
| 762 |
+
"learning_rate": 3.136094674556213e-05,
|
| 763 |
+
"loss": 1.4684,
|
| 764 |
+
"step": 6300
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 7.51,
|
| 768 |
+
"learning_rate": 3.121301775147929e-05,
|
| 769 |
+
"loss": 1.4662,
|
| 770 |
+
"step": 6350
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 7.57,
|
| 774 |
+
"learning_rate": 3.106508875739645e-05,
|
| 775 |
+
"loss": 1.4654,
|
| 776 |
+
"step": 6400
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 7.63,
|
| 780 |
+
"learning_rate": 3.091715976331361e-05,
|
| 781 |
+
"loss": 1.4662,
|
| 782 |
+
"step": 6450
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 7.69,
|
| 786 |
+
"learning_rate": 3.0769230769230774e-05,
|
| 787 |
+
"loss": 1.4669,
|
| 788 |
+
"step": 6500
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 7.75,
|
| 792 |
+
"learning_rate": 3.062130177514793e-05,
|
| 793 |
+
"loss": 1.4671,
|
| 794 |
+
"step": 6550
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 7.81,
|
| 798 |
+
"learning_rate": 3.047337278106509e-05,
|
| 799 |
+
"loss": 1.4647,
|
| 800 |
+
"step": 6600
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 7.87,
|
| 804 |
+
"learning_rate": 3.032544378698225e-05,
|
| 805 |
+
"loss": 1.4669,
|
| 806 |
+
"step": 6650
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 7.93,
|
| 810 |
+
"learning_rate": 3.017751479289941e-05,
|
| 811 |
+
"loss": 1.4637,
|
| 812 |
+
"step": 6700
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 7.99,
|
| 816 |
+
"learning_rate": 3.0029585798816566e-05,
|
| 817 |
+
"loss": 1.4706,
|
| 818 |
+
"step": 6750
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 8.05,
|
| 822 |
+
"learning_rate": 2.9881656804733733e-05,
|
| 823 |
+
"loss": 1.4526,
|
| 824 |
+
"step": 6800
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 8.11,
|
| 828 |
+
"learning_rate": 2.973372781065089e-05,
|
| 829 |
+
"loss": 1.4544,
|
| 830 |
+
"step": 6850
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 8.17,
|
| 834 |
+
"learning_rate": 2.958579881656805e-05,
|
| 835 |
+
"loss": 1.4539,
|
| 836 |
+
"step": 6900
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 8.22,
|
| 840 |
+
"learning_rate": 2.9437869822485207e-05,
|
| 841 |
+
"loss": 1.4521,
|
| 842 |
+
"step": 6950
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 8.28,
|
| 846 |
+
"learning_rate": 2.9289940828402368e-05,
|
| 847 |
+
"loss": 1.4562,
|
| 848 |
+
"step": 7000
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 8.34,
|
| 852 |
+
"learning_rate": 2.9142011834319528e-05,
|
| 853 |
+
"loss": 1.4511,
|
| 854 |
+
"step": 7050
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 8.4,
|
| 858 |
+
"learning_rate": 2.8994082840236685e-05,
|
| 859 |
+
"loss": 1.4522,
|
| 860 |
+
"step": 7100
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 8.46,
|
| 864 |
+
"learning_rate": 2.8846153846153845e-05,
|
| 865 |
+
"loss": 1.4518,
|
| 866 |
+
"step": 7150
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 8.52,
|
| 870 |
+
"learning_rate": 2.869822485207101e-05,
|
| 871 |
+
"loss": 1.4501,
|
| 872 |
+
"step": 7200
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 8.58,
|
| 876 |
+
"learning_rate": 2.855029585798817e-05,
|
| 877 |
+
"loss": 1.4577,
|
| 878 |
+
"step": 7250
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 8.64,
|
| 882 |
+
"learning_rate": 2.8402366863905327e-05,
|
| 883 |
+
"loss": 1.4475,
|
| 884 |
+
"step": 7300
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 8.7,
|
| 888 |
+
"learning_rate": 2.8254437869822487e-05,
|
| 889 |
+
"loss": 1.4528,
|
| 890 |
+
"step": 7350
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 8.76,
|
| 894 |
+
"learning_rate": 2.8106508875739644e-05,
|
| 895 |
+
"loss": 1.4495,
|
| 896 |
+
"step": 7400
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 8.82,
|
| 900 |
+
"learning_rate": 2.7958579881656804e-05,
|
| 901 |
+
"loss": 1.4552,
|
| 902 |
+
"step": 7450
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 8.88,
|
| 906 |
+
"learning_rate": 2.7810650887573965e-05,
|
| 907 |
+
"loss": 1.4544,
|
| 908 |
+
"step": 7500
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 8.93,
|
| 912 |
+
"learning_rate": 2.766272189349113e-05,
|
| 913 |
+
"loss": 1.4512,
|
| 914 |
+
"step": 7550
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 8.99,
|
| 918 |
+
"learning_rate": 2.751479289940829e-05,
|
| 919 |
+
"loss": 1.4531,
|
| 920 |
+
"step": 7600
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 9.05,
|
| 924 |
+
"learning_rate": 2.7366863905325446e-05,
|
| 925 |
+
"loss": 1.4462,
|
| 926 |
+
"step": 7650
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 9.11,
|
| 930 |
+
"learning_rate": 2.7218934911242606e-05,
|
| 931 |
+
"loss": 1.4399,
|
| 932 |
+
"step": 7700
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 9.17,
|
| 936 |
+
"learning_rate": 2.7071005917159763e-05,
|
| 937 |
+
"loss": 1.4394,
|
| 938 |
+
"step": 7750
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 9.23,
|
| 942 |
+
"learning_rate": 2.6923076923076923e-05,
|
| 943 |
+
"loss": 1.4366,
|
| 944 |
+
"step": 7800
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 9.29,
|
| 948 |
+
"learning_rate": 2.6775147928994084e-05,
|
| 949 |
+
"loss": 1.4416,
|
| 950 |
+
"step": 7850
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 9.35,
|
| 954 |
+
"learning_rate": 2.6627218934911247e-05,
|
| 955 |
+
"loss": 1.4405,
|
| 956 |
+
"step": 7900
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 9.41,
|
| 960 |
+
"learning_rate": 2.6479289940828404e-05,
|
| 961 |
+
"loss": 1.4387,
|
| 962 |
+
"step": 7950
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 9.47,
|
| 966 |
+
"learning_rate": 2.6331360946745565e-05,
|
| 967 |
+
"loss": 1.4418,
|
| 968 |
+
"step": 8000
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 9.53,
|
| 972 |
+
"learning_rate": 2.6183431952662725e-05,
|
| 973 |
+
"loss": 1.4351,
|
| 974 |
+
"step": 8050
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 9.59,
|
| 978 |
+
"learning_rate": 2.6035502958579882e-05,
|
| 979 |
+
"loss": 1.4399,
|
| 980 |
+
"step": 8100
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 9.64,
|
| 984 |
+
"learning_rate": 2.5887573964497042e-05,
|
| 985 |
+
"loss": 1.4389,
|
| 986 |
+
"step": 8150
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 9.7,
|
| 990 |
+
"learning_rate": 2.57396449704142e-05,
|
| 991 |
+
"loss": 1.4382,
|
| 992 |
+
"step": 8200
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 9.76,
|
| 996 |
+
"learning_rate": 2.559171597633136e-05,
|
| 997 |
+
"loss": 1.4391,
|
| 998 |
+
"step": 8250
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 9.82,
|
| 1002 |
+
"learning_rate": 2.5443786982248524e-05,
|
| 1003 |
+
"loss": 1.4429,
|
| 1004 |
+
"step": 8300
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 9.88,
|
| 1008 |
+
"learning_rate": 2.5295857988165684e-05,
|
| 1009 |
+
"loss": 1.4393,
|
| 1010 |
+
"step": 8350
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 9.94,
|
| 1014 |
+
"learning_rate": 2.514792899408284e-05,
|
| 1015 |
+
"loss": 1.4407,
|
| 1016 |
+
"step": 8400
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 10.0,
|
| 1020 |
+
"learning_rate": 2.5e-05,
|
| 1021 |
+
"loss": 1.4418,
|
| 1022 |
+
"step": 8450
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 10.06,
|
| 1026 |
+
"learning_rate": 2.485207100591716e-05,
|
| 1027 |
+
"loss": 1.4246,
|
| 1028 |
+
"step": 8500
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 10.12,
|
| 1032 |
+
"learning_rate": 2.4704142011834322e-05,
|
| 1033 |
+
"loss": 1.4319,
|
| 1034 |
+
"step": 8550
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 10.18,
|
| 1038 |
+
"learning_rate": 2.4556213017751482e-05,
|
| 1039 |
+
"loss": 1.4263,
|
| 1040 |
+
"step": 8600
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 10.24,
|
| 1044 |
+
"learning_rate": 2.440828402366864e-05,
|
| 1045 |
+
"loss": 1.4334,
|
| 1046 |
+
"step": 8650
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 10.3,
|
| 1050 |
+
"learning_rate": 2.42603550295858e-05,
|
| 1051 |
+
"loss": 1.4287,
|
| 1052 |
+
"step": 8700
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 10.36,
|
| 1056 |
+
"learning_rate": 2.411242603550296e-05,
|
| 1057 |
+
"loss": 1.4289,
|
| 1058 |
+
"step": 8750
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 10.41,
|
| 1062 |
+
"learning_rate": 2.396449704142012e-05,
|
| 1063 |
+
"loss": 1.4285,
|
| 1064 |
+
"step": 8800
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 10.47,
|
| 1068 |
+
"learning_rate": 2.3816568047337277e-05,
|
| 1069 |
+
"loss": 1.4286,
|
| 1070 |
+
"step": 8850
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 10.53,
|
| 1074 |
+
"learning_rate": 2.3668639053254438e-05,
|
| 1075 |
+
"loss": 1.4273,
|
| 1076 |
+
"step": 8900
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 10.59,
|
| 1080 |
+
"learning_rate": 2.3520710059171598e-05,
|
| 1081 |
+
"loss": 1.4291,
|
| 1082 |
+
"step": 8950
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 10.65,
|
| 1086 |
+
"learning_rate": 2.337278106508876e-05,
|
| 1087 |
+
"loss": 1.4268,
|
| 1088 |
+
"step": 9000
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 10.71,
|
| 1092 |
+
"learning_rate": 2.322485207100592e-05,
|
| 1093 |
+
"loss": 1.4302,
|
| 1094 |
+
"step": 9050
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 10.77,
|
| 1098 |
+
"learning_rate": 2.307692307692308e-05,
|
| 1099 |
+
"loss": 1.4313,
|
| 1100 |
+
"step": 9100
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 10.83,
|
| 1104 |
+
"learning_rate": 2.292899408284024e-05,
|
| 1105 |
+
"loss": 1.4274,
|
| 1106 |
+
"step": 9150
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 10.89,
|
| 1110 |
+
"learning_rate": 2.2781065088757396e-05,
|
| 1111 |
+
"loss": 1.4286,
|
| 1112 |
+
"step": 9200
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 10.95,
|
| 1116 |
+
"learning_rate": 2.2633136094674557e-05,
|
| 1117 |
+
"loss": 1.4278,
|
| 1118 |
+
"step": 9250
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 11.01,
|
| 1122 |
+
"learning_rate": 2.2485207100591717e-05,
|
| 1123 |
+
"loss": 1.4265,
|
| 1124 |
+
"step": 9300
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 11.07,
|
| 1128 |
+
"learning_rate": 2.2337278106508877e-05,
|
| 1129 |
+
"loss": 1.4156,
|
| 1130 |
+
"step": 9350
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 11.12,
|
| 1134 |
+
"learning_rate": 2.2189349112426034e-05,
|
| 1135 |
+
"loss": 1.4174,
|
| 1136 |
+
"step": 9400
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 11.18,
|
| 1140 |
+
"learning_rate": 2.2041420118343195e-05,
|
| 1141 |
+
"loss": 1.4161,
|
| 1142 |
+
"step": 9450
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 11.24,
|
| 1146 |
+
"learning_rate": 2.189349112426036e-05,
|
| 1147 |
+
"loss": 1.4216,
|
| 1148 |
+
"step": 9500
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 11.3,
|
| 1152 |
+
"learning_rate": 2.1745562130177516e-05,
|
| 1153 |
+
"loss": 1.4167,
|
| 1154 |
+
"step": 9550
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 11.36,
|
| 1158 |
+
"learning_rate": 2.1597633136094676e-05,
|
| 1159 |
+
"loss": 1.4188,
|
| 1160 |
+
"step": 9600
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 11.42,
|
| 1164 |
+
"learning_rate": 2.1449704142011836e-05,
|
| 1165 |
+
"loss": 1.4242,
|
| 1166 |
+
"step": 9650
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 11.48,
|
| 1170 |
+
"learning_rate": 2.1301775147928997e-05,
|
| 1171 |
+
"loss": 1.4177,
|
| 1172 |
+
"step": 9700
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 11.54,
|
| 1176 |
+
"learning_rate": 2.1153846153846154e-05,
|
| 1177 |
+
"loss": 1.4188,
|
| 1178 |
+
"step": 9750
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 11.6,
|
| 1182 |
+
"learning_rate": 2.1005917159763314e-05,
|
| 1183 |
+
"loss": 1.4202,
|
| 1184 |
+
"step": 9800
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 11.66,
|
| 1188 |
+
"learning_rate": 2.0857988165680474e-05,
|
| 1189 |
+
"loss": 1.4209,
|
| 1190 |
+
"step": 9850
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 11.72,
|
| 1194 |
+
"learning_rate": 2.0710059171597635e-05,
|
| 1195 |
+
"loss": 1.4155,
|
| 1196 |
+
"step": 9900
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 11.78,
|
| 1200 |
+
"learning_rate": 2.0562130177514795e-05,
|
| 1201 |
+
"loss": 1.4214,
|
| 1202 |
+
"step": 9950
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 11.83,
|
| 1206 |
+
"learning_rate": 2.0414201183431952e-05,
|
| 1207 |
+
"loss": 1.4201,
|
| 1208 |
+
"step": 10000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 11.89,
|
| 1212 |
+
"learning_rate": 2.0266272189349116e-05,
|
| 1213 |
+
"loss": 1.4175,
|
| 1214 |
+
"step": 10050
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 11.95,
|
| 1218 |
+
"learning_rate": 2.0118343195266273e-05,
|
| 1219 |
+
"loss": 1.4171,
|
| 1220 |
+
"step": 10100
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 12.01,
|
| 1224 |
+
"learning_rate": 1.9970414201183433e-05,
|
| 1225 |
+
"loss": 1.4188,
|
| 1226 |
+
"step": 10150
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 12.07,
|
| 1230 |
+
"learning_rate": 1.9822485207100593e-05,
|
| 1231 |
+
"loss": 1.4123,
|
| 1232 |
+
"step": 10200
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 12.13,
|
| 1236 |
+
"learning_rate": 1.9674556213017754e-05,
|
| 1237 |
+
"loss": 1.4079,
|
| 1238 |
+
"step": 10250
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 12.19,
|
| 1242 |
+
"learning_rate": 1.952662721893491e-05,
|
| 1243 |
+
"loss": 1.4057,
|
| 1244 |
+
"step": 10300
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 12.25,
|
| 1248 |
+
"learning_rate": 1.937869822485207e-05,
|
| 1249 |
+
"loss": 1.4053,
|
| 1250 |
+
"step": 10350
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 12.31,
|
| 1254 |
+
"learning_rate": 1.923076923076923e-05,
|
| 1255 |
+
"loss": 1.41,
|
| 1256 |
+
"step": 10400
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 12.37,
|
| 1260 |
+
"learning_rate": 1.9082840236686392e-05,
|
| 1261 |
+
"loss": 1.4113,
|
| 1262 |
+
"step": 10450
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 12.43,
|
| 1266 |
+
"learning_rate": 1.8934911242603552e-05,
|
| 1267 |
+
"loss": 1.4073,
|
| 1268 |
+
"step": 10500
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 12.49,
|
| 1272 |
+
"learning_rate": 1.878698224852071e-05,
|
| 1273 |
+
"loss": 1.416,
|
| 1274 |
+
"step": 10550
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 12.54,
|
| 1278 |
+
"learning_rate": 1.8639053254437873e-05,
|
| 1279 |
+
"loss": 1.4113,
|
| 1280 |
+
"step": 10600
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 12.6,
|
| 1284 |
+
"learning_rate": 1.849112426035503e-05,
|
| 1285 |
+
"loss": 1.4095,
|
| 1286 |
+
"step": 10650
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 12.66,
|
| 1290 |
+
"learning_rate": 1.834319526627219e-05,
|
| 1291 |
+
"loss": 1.4056,
|
| 1292 |
+
"step": 10700
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 12.72,
|
| 1296 |
+
"learning_rate": 1.819526627218935e-05,
|
| 1297 |
+
"loss": 1.4081,
|
| 1298 |
+
"step": 10750
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 12.78,
|
| 1302 |
+
"learning_rate": 1.804733727810651e-05,
|
| 1303 |
+
"loss": 1.4123,
|
| 1304 |
+
"step": 10800
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 12.84,
|
| 1308 |
+
"learning_rate": 1.7899408284023668e-05,
|
| 1309 |
+
"loss": 1.4106,
|
| 1310 |
+
"step": 10850
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 12.9,
|
| 1314 |
+
"learning_rate": 1.7751479289940828e-05,
|
| 1315 |
+
"loss": 1.4123,
|
| 1316 |
+
"step": 10900
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 12.96,
|
| 1320 |
+
"learning_rate": 1.760355029585799e-05,
|
| 1321 |
+
"loss": 1.4109,
|
| 1322 |
+
"step": 10950
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 13.02,
|
| 1326 |
+
"learning_rate": 1.745562130177515e-05,
|
| 1327 |
+
"loss": 1.4044,
|
| 1328 |
+
"step": 11000
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 13.08,
|
| 1332 |
+
"learning_rate": 1.730769230769231e-05,
|
| 1333 |
+
"loss": 1.4006,
|
| 1334 |
+
"step": 11050
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 13.14,
|
| 1338 |
+
"learning_rate": 1.7159763313609466e-05,
|
| 1339 |
+
"loss": 1.4004,
|
| 1340 |
+
"step": 11100
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 13.2,
|
| 1344 |
+
"learning_rate": 1.701183431952663e-05,
|
| 1345 |
+
"loss": 1.3985,
|
| 1346 |
+
"step": 11150
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 13.25,
|
| 1350 |
+
"learning_rate": 1.6863905325443787e-05,
|
| 1351 |
+
"loss": 1.4041,
|
| 1352 |
+
"step": 11200
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 13.31,
|
| 1356 |
+
"learning_rate": 1.6715976331360947e-05,
|
| 1357 |
+
"loss": 1.4045,
|
| 1358 |
+
"step": 11250
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 13.37,
|
| 1362 |
+
"learning_rate": 1.6568047337278108e-05,
|
| 1363 |
+
"loss": 1.4003,
|
| 1364 |
+
"step": 11300
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 13.43,
|
| 1368 |
+
"learning_rate": 1.6420118343195268e-05,
|
| 1369 |
+
"loss": 1.3994,
|
| 1370 |
+
"step": 11350
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 13.49,
|
| 1374 |
+
"learning_rate": 1.6272189349112425e-05,
|
| 1375 |
+
"loss": 1.403,
|
| 1376 |
+
"step": 11400
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 13.55,
|
| 1380 |
+
"learning_rate": 1.6124260355029585e-05,
|
| 1381 |
+
"loss": 1.4027,
|
| 1382 |
+
"step": 11450
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 13.61,
|
| 1386 |
+
"learning_rate": 1.5976331360946746e-05,
|
| 1387 |
+
"loss": 1.4023,
|
| 1388 |
+
"step": 11500
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 13.67,
|
| 1392 |
+
"learning_rate": 1.5828402366863906e-05,
|
| 1393 |
+
"loss": 1.4011,
|
| 1394 |
+
"step": 11550
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 13.73,
|
| 1398 |
+
"learning_rate": 1.5680473372781066e-05,
|
| 1399 |
+
"loss": 1.3999,
|
| 1400 |
+
"step": 11600
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 13.79,
|
| 1404 |
+
"learning_rate": 1.5532544378698223e-05,
|
| 1405 |
+
"loss": 1.4028,
|
| 1406 |
+
"step": 11650
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 13.85,
|
| 1410 |
+
"learning_rate": 1.5384615384615387e-05,
|
| 1411 |
+
"loss": 1.404,
|
| 1412 |
+
"step": 11700
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 13.91,
|
| 1416 |
+
"learning_rate": 1.5236686390532546e-05,
|
| 1417 |
+
"loss": 1.4037,
|
| 1418 |
+
"step": 11750
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 13.96,
|
| 1422 |
+
"learning_rate": 1.5088757396449705e-05,
|
| 1423 |
+
"loss": 1.4025,
|
| 1424 |
+
"step": 11800
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 14.02,
|
| 1428 |
+
"learning_rate": 1.4940828402366867e-05,
|
| 1429 |
+
"loss": 1.4015,
|
| 1430 |
+
"step": 11850
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 14.08,
|
| 1434 |
+
"learning_rate": 1.4792899408284025e-05,
|
| 1435 |
+
"loss": 1.3906,
|
| 1436 |
+
"step": 11900
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 14.14,
|
| 1440 |
+
"learning_rate": 1.4644970414201184e-05,
|
| 1441 |
+
"loss": 1.3972,
|
| 1442 |
+
"step": 11950
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 14.2,
|
| 1446 |
+
"learning_rate": 1.4497041420118343e-05,
|
| 1447 |
+
"loss": 1.3938,
|
| 1448 |
+
"step": 12000
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 14.26,
|
| 1452 |
+
"learning_rate": 1.4349112426035505e-05,
|
| 1453 |
+
"loss": 1.3925,
|
| 1454 |
+
"step": 12050
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 14.32,
|
| 1458 |
+
"learning_rate": 1.4201183431952663e-05,
|
| 1459 |
+
"loss": 1.3954,
|
| 1460 |
+
"step": 12100
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 14.38,
|
| 1464 |
+
"learning_rate": 1.4053254437869822e-05,
|
| 1465 |
+
"loss": 1.3917,
|
| 1466 |
+
"step": 12150
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 14.44,
|
| 1470 |
+
"learning_rate": 1.3905325443786982e-05,
|
| 1471 |
+
"loss": 1.3955,
|
| 1472 |
+
"step": 12200
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 14.5,
|
| 1476 |
+
"learning_rate": 1.3757396449704144e-05,
|
| 1477 |
+
"loss": 1.3981,
|
| 1478 |
+
"step": 12250
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 14.56,
|
| 1482 |
+
"learning_rate": 1.3609467455621303e-05,
|
| 1483 |
+
"loss": 1.3969,
|
| 1484 |
+
"step": 12300
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 14.62,
|
| 1488 |
+
"learning_rate": 1.3461538461538462e-05,
|
| 1489 |
+
"loss": 1.3936,
|
| 1490 |
+
"step": 12350
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 14.67,
|
| 1494 |
+
"learning_rate": 1.3313609467455624e-05,
|
| 1495 |
+
"loss": 1.3959,
|
| 1496 |
+
"step": 12400
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 14.73,
|
| 1500 |
+
"learning_rate": 1.3165680473372782e-05,
|
| 1501 |
+
"loss": 1.3921,
|
| 1502 |
+
"step": 12450
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 14.79,
|
| 1506 |
+
"learning_rate": 1.3017751479289941e-05,
|
| 1507 |
+
"loss": 1.396,
|
| 1508 |
+
"step": 12500
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 14.85,
|
| 1512 |
+
"learning_rate": 1.28698224852071e-05,
|
| 1513 |
+
"loss": 1.4021,
|
| 1514 |
+
"step": 12550
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 14.91,
|
| 1518 |
+
"learning_rate": 1.2721893491124262e-05,
|
| 1519 |
+
"loss": 1.3903,
|
| 1520 |
+
"step": 12600
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 14.97,
|
| 1524 |
+
"learning_rate": 1.257396449704142e-05,
|
| 1525 |
+
"loss": 1.3984,
|
| 1526 |
+
"step": 12650
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 15.03,
|
| 1530 |
+
"learning_rate": 1.242603550295858e-05,
|
| 1531 |
+
"loss": 1.3901,
|
| 1532 |
+
"step": 12700
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 15.09,
|
| 1536 |
+
"learning_rate": 1.2278106508875741e-05,
|
| 1537 |
+
"loss": 1.3892,
|
| 1538 |
+
"step": 12750
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 15.15,
|
| 1542 |
+
"learning_rate": 1.21301775147929e-05,
|
| 1543 |
+
"loss": 1.3915,
|
| 1544 |
+
"step": 12800
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 15.21,
|
| 1548 |
+
"learning_rate": 1.198224852071006e-05,
|
| 1549 |
+
"loss": 1.3855,
|
| 1550 |
+
"step": 12850
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 15.27,
|
| 1554 |
+
"learning_rate": 1.1834319526627219e-05,
|
| 1555 |
+
"loss": 1.3946,
|
| 1556 |
+
"step": 12900
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 15.33,
|
| 1560 |
+
"learning_rate": 1.168639053254438e-05,
|
| 1561 |
+
"loss": 1.3925,
|
| 1562 |
+
"step": 12950
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 15.38,
|
| 1566 |
+
"learning_rate": 1.153846153846154e-05,
|
| 1567 |
+
"loss": 1.3904,
|
| 1568 |
+
"step": 13000
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 15.44,
|
| 1572 |
+
"learning_rate": 1.1390532544378698e-05,
|
| 1573 |
+
"loss": 1.3887,
|
| 1574 |
+
"step": 13050
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 15.5,
|
| 1578 |
+
"learning_rate": 1.1242603550295859e-05,
|
| 1579 |
+
"loss": 1.3902,
|
| 1580 |
+
"step": 13100
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 15.56,
|
| 1584 |
+
"learning_rate": 1.1094674556213017e-05,
|
| 1585 |
+
"loss": 1.3915,
|
| 1586 |
+
"step": 13150
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 15.62,
|
| 1590 |
+
"learning_rate": 1.094674556213018e-05,
|
| 1591 |
+
"loss": 1.3895,
|
| 1592 |
+
"step": 13200
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 15.68,
|
| 1596 |
+
"learning_rate": 1.0798816568047338e-05,
|
| 1597 |
+
"loss": 1.3852,
|
| 1598 |
+
"step": 13250
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 15.74,
|
| 1602 |
+
"learning_rate": 1.0650887573964498e-05,
|
| 1603 |
+
"loss": 1.3863,
|
| 1604 |
+
"step": 13300
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 15.8,
|
| 1608 |
+
"learning_rate": 1.0502958579881657e-05,
|
| 1609 |
+
"loss": 1.3872,
|
| 1610 |
+
"step": 13350
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 15.86,
|
| 1614 |
+
"learning_rate": 1.0355029585798817e-05,
|
| 1615 |
+
"loss": 1.3875,
|
| 1616 |
+
"step": 13400
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 15.92,
|
| 1620 |
+
"learning_rate": 1.0207100591715976e-05,
|
| 1621 |
+
"loss": 1.3919,
|
| 1622 |
+
"step": 13450
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 15.98,
|
| 1626 |
+
"learning_rate": 1.0059171597633136e-05,
|
| 1627 |
+
"loss": 1.3883,
|
| 1628 |
+
"step": 13500
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 16.04,
|
| 1632 |
+
"learning_rate": 9.911242603550297e-06,
|
| 1633 |
+
"loss": 1.3864,
|
| 1634 |
+
"step": 13550
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 16.09,
|
| 1638 |
+
"learning_rate": 9.763313609467455e-06,
|
| 1639 |
+
"loss": 1.3801,
|
| 1640 |
+
"step": 13600
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 16.15,
|
| 1644 |
+
"learning_rate": 9.615384615384616e-06,
|
| 1645 |
+
"loss": 1.3839,
|
| 1646 |
+
"step": 13650
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 16.21,
|
| 1650 |
+
"learning_rate": 9.467455621301776e-06,
|
| 1651 |
+
"loss": 1.3796,
|
| 1652 |
+
"step": 13700
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 16.27,
|
| 1656 |
+
"learning_rate": 9.319526627218936e-06,
|
| 1657 |
+
"loss": 1.3852,
|
| 1658 |
+
"step": 13750
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 16.33,
|
| 1662 |
+
"learning_rate": 9.171597633136095e-06,
|
| 1663 |
+
"loss": 1.384,
|
| 1664 |
+
"step": 13800
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 16.39,
|
| 1668 |
+
"learning_rate": 9.023668639053255e-06,
|
| 1669 |
+
"loss": 1.3887,
|
| 1670 |
+
"step": 13850
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 16.45,
|
| 1674 |
+
"learning_rate": 8.875739644970414e-06,
|
| 1675 |
+
"loss": 1.3861,
|
| 1676 |
+
"step": 13900
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 16.51,
|
| 1680 |
+
"learning_rate": 8.727810650887574e-06,
|
| 1681 |
+
"loss": 1.3819,
|
| 1682 |
+
"step": 13950
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 16.57,
|
| 1686 |
+
"learning_rate": 8.579881656804733e-06,
|
| 1687 |
+
"loss": 1.3845,
|
| 1688 |
+
"step": 14000
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 16.63,
|
| 1692 |
+
"learning_rate": 8.431952662721893e-06,
|
| 1693 |
+
"loss": 1.3875,
|
| 1694 |
+
"step": 14050
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 16.69,
|
| 1698 |
+
"learning_rate": 8.284023668639054e-06,
|
| 1699 |
+
"loss": 1.3842,
|
| 1700 |
+
"step": 14100
|
| 1701 |
+
},
|
| 1702 |
+
{
|
| 1703 |
+
"epoch": 16.75,
|
| 1704 |
+
"learning_rate": 8.136094674556213e-06,
|
| 1705 |
+
"loss": 1.3803,
|
| 1706 |
+
"step": 14150
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 16.8,
|
| 1710 |
+
"learning_rate": 7.988165680473373e-06,
|
| 1711 |
+
"loss": 1.3822,
|
| 1712 |
+
"step": 14200
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 16.86,
|
| 1716 |
+
"learning_rate": 7.840236686390533e-06,
|
| 1717 |
+
"loss": 1.3879,
|
| 1718 |
+
"step": 14250
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 16.92,
|
| 1722 |
+
"learning_rate": 7.692307692307694e-06,
|
| 1723 |
+
"loss": 1.3829,
|
| 1724 |
+
"step": 14300
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 16.98,
|
| 1728 |
+
"learning_rate": 7.544378698224852e-06,
|
| 1729 |
+
"loss": 1.3861,
|
| 1730 |
+
"step": 14350
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 17.04,
|
| 1734 |
+
"learning_rate": 7.396449704142013e-06,
|
| 1735 |
+
"loss": 1.3802,
|
| 1736 |
+
"step": 14400
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 17.1,
|
| 1740 |
+
"learning_rate": 7.248520710059171e-06,
|
| 1741 |
+
"loss": 1.3734,
|
| 1742 |
+
"step": 14450
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 17.16,
|
| 1746 |
+
"learning_rate": 7.100591715976332e-06,
|
| 1747 |
+
"loss": 1.3826,
|
| 1748 |
+
"step": 14500
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 17.22,
|
| 1752 |
+
"learning_rate": 6.952662721893491e-06,
|
| 1753 |
+
"loss": 1.3858,
|
| 1754 |
+
"step": 14550
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 17.28,
|
| 1758 |
+
"learning_rate": 6.8047337278106515e-06,
|
| 1759 |
+
"loss": 1.3781,
|
| 1760 |
+
"step": 14600
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 17.34,
|
| 1764 |
+
"learning_rate": 6.656804733727812e-06,
|
| 1765 |
+
"loss": 1.3831,
|
| 1766 |
+
"step": 14650
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 17.4,
|
| 1770 |
+
"learning_rate": 6.5088757396449705e-06,
|
| 1771 |
+
"loss": 1.3755,
|
| 1772 |
+
"step": 14700
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 17.46,
|
| 1776 |
+
"learning_rate": 6.360946745562131e-06,
|
| 1777 |
+
"loss": 1.3841,
|
| 1778 |
+
"step": 14750
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 17.51,
|
| 1782 |
+
"learning_rate": 6.21301775147929e-06,
|
| 1783 |
+
"loss": 1.3796,
|
| 1784 |
+
"step": 14800
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 17.57,
|
| 1788 |
+
"learning_rate": 6.06508875739645e-06,
|
| 1789 |
+
"loss": 1.3799,
|
| 1790 |
+
"step": 14850
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 17.63,
|
| 1794 |
+
"learning_rate": 5.917159763313609e-06,
|
| 1795 |
+
"loss": 1.3814,
|
| 1796 |
+
"step": 14900
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 17.69,
|
| 1800 |
+
"learning_rate": 5.76923076923077e-06,
|
| 1801 |
+
"loss": 1.3783,
|
| 1802 |
+
"step": 14950
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 17.75,
|
| 1806 |
+
"learning_rate": 5.621301775147929e-06,
|
| 1807 |
+
"loss": 1.378,
|
| 1808 |
+
"step": 15000
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 17.81,
|
| 1812 |
+
"learning_rate": 5.47337278106509e-06,
|
| 1813 |
+
"loss": 1.379,
|
| 1814 |
+
"step": 15050
|
| 1815 |
+
},
|
| 1816 |
+
{
|
| 1817 |
+
"epoch": 17.87,
|
| 1818 |
+
"learning_rate": 5.325443786982249e-06,
|
| 1819 |
+
"loss": 1.3806,
|
| 1820 |
+
"step": 15100
|
| 1821 |
+
},
|
| 1822 |
+
{
|
| 1823 |
+
"epoch": 17.93,
|
| 1824 |
+
"learning_rate": 5.177514792899409e-06,
|
| 1825 |
+
"loss": 1.3806,
|
| 1826 |
+
"step": 15150
|
| 1827 |
+
},
|
| 1828 |
+
{
|
| 1829 |
+
"epoch": 17.99,
|
| 1830 |
+
"learning_rate": 5.029585798816568e-06,
|
| 1831 |
+
"loss": 1.3809,
|
| 1832 |
+
"step": 15200
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"epoch": 18.05,
|
| 1836 |
+
"learning_rate": 4.881656804733728e-06,
|
| 1837 |
+
"loss": 1.3748,
|
| 1838 |
+
"step": 15250
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"epoch": 18.11,
|
| 1842 |
+
"learning_rate": 4.733727810650888e-06,
|
| 1843 |
+
"loss": 1.38,
|
| 1844 |
+
"step": 15300
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 18.17,
|
| 1848 |
+
"learning_rate": 4.5857988165680475e-06,
|
| 1849 |
+
"loss": 1.3729,
|
| 1850 |
+
"step": 15350
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 18.22,
|
| 1854 |
+
"learning_rate": 4.437869822485207e-06,
|
| 1855 |
+
"loss": 1.3798,
|
| 1856 |
+
"step": 15400
|
| 1857 |
+
},
|
| 1858 |
+
{
|
| 1859 |
+
"epoch": 18.28,
|
| 1860 |
+
"learning_rate": 4.2899408284023666e-06,
|
| 1861 |
+
"loss": 1.3779,
|
| 1862 |
+
"step": 15450
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 18.34,
|
| 1866 |
+
"learning_rate": 4.142011834319527e-06,
|
| 1867 |
+
"loss": 1.377,
|
| 1868 |
+
"step": 15500
|
| 1869 |
+
},
|
| 1870 |
+
{
|
| 1871 |
+
"epoch": 18.4,
|
| 1872 |
+
"learning_rate": 3.9940828402366864e-06,
|
| 1873 |
+
"loss": 1.3722,
|
| 1874 |
+
"step": 15550
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 18.46,
|
| 1878 |
+
"learning_rate": 3.846153846153847e-06,
|
| 1879 |
+
"loss": 1.377,
|
| 1880 |
+
"step": 15600
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 18.52,
|
| 1884 |
+
"learning_rate": 3.6982248520710063e-06,
|
| 1885 |
+
"loss": 1.3774,
|
| 1886 |
+
"step": 15650
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 18.58,
|
| 1890 |
+
"learning_rate": 3.550295857988166e-06,
|
| 1891 |
+
"loss": 1.3803,
|
| 1892 |
+
"step": 15700
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 18.64,
|
| 1896 |
+
"learning_rate": 3.4023668639053257e-06,
|
| 1897 |
+
"loss": 1.3754,
|
| 1898 |
+
"step": 15750
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 18.7,
|
| 1902 |
+
"learning_rate": 3.2544378698224853e-06,
|
| 1903 |
+
"loss": 1.3825,
|
| 1904 |
+
"step": 15800
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 18.76,
|
| 1908 |
+
"learning_rate": 3.106508875739645e-06,
|
| 1909 |
+
"loss": 1.3737,
|
| 1910 |
+
"step": 15850
|
| 1911 |
+
},
|
| 1912 |
+
{
|
| 1913 |
+
"epoch": 18.82,
|
| 1914 |
+
"learning_rate": 2.9585798816568047e-06,
|
| 1915 |
+
"loss": 1.3765,
|
| 1916 |
+
"step": 15900
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 18.88,
|
| 1920 |
+
"learning_rate": 2.8106508875739646e-06,
|
| 1921 |
+
"loss": 1.3763,
|
| 1922 |
+
"step": 15950
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 18.93,
|
| 1926 |
+
"learning_rate": 2.6627218934911246e-06,
|
| 1927 |
+
"loss": 1.3733,
|
| 1928 |
+
"step": 16000
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 18.99,
|
| 1932 |
+
"learning_rate": 2.514792899408284e-06,
|
| 1933 |
+
"loss": 1.3763,
|
| 1934 |
+
"step": 16050
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 19.05,
|
| 1938 |
+
"learning_rate": 2.366863905325444e-06,
|
| 1939 |
+
"loss": 1.3739,
|
| 1940 |
+
"step": 16100
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 19.11,
|
| 1944 |
+
"learning_rate": 2.2189349112426035e-06,
|
| 1945 |
+
"loss": 1.3749,
|
| 1946 |
+
"step": 16150
|
| 1947 |
+
},
|
| 1948 |
+
{
|
| 1949 |
+
"epoch": 19.17,
|
| 1950 |
+
"learning_rate": 2.0710059171597635e-06,
|
| 1951 |
+
"loss": 1.3695,
|
| 1952 |
+
"step": 16200
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 19.23,
|
| 1956 |
+
"learning_rate": 1.9230769230769234e-06,
|
| 1957 |
+
"loss": 1.3725,
|
| 1958 |
+
"step": 16250
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 19.29,
|
| 1962 |
+
"learning_rate": 1.775147928994083e-06,
|
| 1963 |
+
"loss": 1.373,
|
| 1964 |
+
"step": 16300
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 19.35,
|
| 1968 |
+
"learning_rate": 1.6272189349112426e-06,
|
| 1969 |
+
"loss": 1.3751,
|
| 1970 |
+
"step": 16350
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 19.41,
|
| 1974 |
+
"learning_rate": 1.4792899408284024e-06,
|
| 1975 |
+
"loss": 1.375,
|
| 1976 |
+
"step": 16400
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 19.47,
|
| 1980 |
+
"learning_rate": 1.3313609467455623e-06,
|
| 1981 |
+
"loss": 1.3767,
|
| 1982 |
+
"step": 16450
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 19.53,
|
| 1986 |
+
"learning_rate": 1.183431952662722e-06,
|
| 1987 |
+
"loss": 1.369,
|
| 1988 |
+
"step": 16500
|
| 1989 |
+
},
|
| 1990 |
+
{
|
| 1991 |
+
"epoch": 19.59,
|
| 1992 |
+
"learning_rate": 1.0355029585798817e-06,
|
| 1993 |
+
"loss": 1.3773,
|
| 1994 |
+
"step": 16550
|
| 1995 |
+
},
|
| 1996 |
+
{
|
| 1997 |
+
"epoch": 19.64,
|
| 1998 |
+
"learning_rate": 8.875739644970415e-07,
|
| 1999 |
+
"loss": 1.3749,
|
| 2000 |
+
"step": 16600
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 19.7,
|
| 2004 |
+
"learning_rate": 7.396449704142012e-07,
|
| 2005 |
+
"loss": 1.3725,
|
| 2006 |
+
"step": 16650
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 19.76,
|
| 2010 |
+
"learning_rate": 5.91715976331361e-07,
|
| 2011 |
+
"loss": 1.374,
|
| 2012 |
+
"step": 16700
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 19.82,
|
| 2016 |
+
"learning_rate": 4.4378698224852073e-07,
|
| 2017 |
+
"loss": 1.3735,
|
| 2018 |
+
"step": 16750
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 19.88,
|
| 2022 |
+
"learning_rate": 2.958579881656805e-07,
|
| 2023 |
+
"loss": 1.3759,
|
| 2024 |
+
"step": 16800
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 19.94,
|
| 2028 |
+
"learning_rate": 1.4792899408284025e-07,
|
| 2029 |
+
"loss": 1.3747,
|
| 2030 |
+
"step": 16850
|
| 2031 |
+
},
|
| 2032 |
+
{
|
| 2033 |
+
"epoch": 20.0,
|
| 2034 |
+
"learning_rate": 0.0,
|
| 2035 |
+
"loss": 1.3793,
|
| 2036 |
+
"step": 16900
|
| 2037 |
+
},
|
| 2038 |
+
{
|
| 2039 |
+
"epoch": 20.0,
|
| 2040 |
+
"step": 16900,
|
| 2041 |
+
"total_flos": 6.27801669500928e+16,
|
| 2042 |
+
"train_loss": 1.4740209875163242,
|
| 2043 |
+
"train_runtime": 8294.5138,
|
| 2044 |
+
"train_samples_per_second": 260.678,
|
| 2045 |
+
"train_steps_per_second": 2.037
|
| 2046 |
+
}
|
| 2047 |
+
],
|
| 2048 |
+
"max_steps": 16900,
|
| 2049 |
+
"num_train_epochs": 20,
|
| 2050 |
+
"total_flos": 6.27801669500928e+16,
|
| 2051 |
+
"trial_name": null,
|
| 2052 |
+
"trial_params": null
|
| 2053 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:55447a38a2a38d9c7025e300832207b3bef29b9168a5a8cd7310f3745165da2f
|
| 3 |
+
size 2799
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|