Spaces:
Runtime error
Runtime error
Commit
•
3b0479f
1
Parent(s):
5516bfe
Upload 9 files
Browse files- robin-7b/README.md +59 -0
- robin-7b/adapter_config.json +19 -0
- robin-7b/adapter_model.bin +3 -0
- robin-7b/all_results.json +8 -0
- robin-7b/special_tokens_map.json +1 -0
- robin-7b/tokenizer.model +3 -0
- robin-7b/tokenizer_config.json +9 -0
- robin-7b/train_results.json +8 -0
- robin-7b/trainer_state.json +1717 -0
robin-7b/README.md
ADDED
|
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: other
|
| 3 |
+
tags:
|
| 4 |
+
- generated_from_trainer
|
| 5 |
+
datasets:
|
| 6 |
+
- customized
|
| 7 |
+
model-index:
|
| 8 |
+
- name: xl_031_lora
|
| 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 |
+
# xl_031_lora
|
| 16 |
+
|
| 17 |
+
This model is a fine-tuned version of [aleksickx/llama-7b-hf](https://huggingface.co/aleksickx/llama-7b-hf) 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: 0.0008
|
| 37 |
+
- train_batch_size: 4
|
| 38 |
+
- eval_batch_size: 8
|
| 39 |
+
- seed: 42
|
| 40 |
+
- distributed_type: multi-GPU
|
| 41 |
+
- num_devices: 8
|
| 42 |
+
- gradient_accumulation_steps: 4
|
| 43 |
+
- total_train_batch_size: 128
|
| 44 |
+
- total_eval_batch_size: 64
|
| 45 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 46 |
+
- lr_scheduler_type: cosine
|
| 47 |
+
- lr_scheduler_warmup_ratio: 0.03
|
| 48 |
+
- num_epochs: 10.0
|
| 49 |
+
|
| 50 |
+
### Training results
|
| 51 |
+
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
### Framework versions
|
| 55 |
+
|
| 56 |
+
- Transformers 4.28.0.dev0
|
| 57 |
+
- Pytorch 2.0.0+cu117
|
| 58 |
+
- Datasets 2.10.1
|
| 59 |
+
- Tokenizers 0.13.3
|
robin-7b/adapter_config.json
ADDED
|
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"base_model_name_or_path": "aleksickx/llama-7b-hf",
|
| 3 |
+
"bias": "none",
|
| 4 |
+
"enable_lora": null,
|
| 5 |
+
"fan_in_fan_out": false,
|
| 6 |
+
"inference_mode": true,
|
| 7 |
+
"init_lora_weights": true,
|
| 8 |
+
"lora_alpha": 32,
|
| 9 |
+
"lora_dropout": 0.1,
|
| 10 |
+
"merge_weights": false,
|
| 11 |
+
"modules_to_save": null,
|
| 12 |
+
"peft_type": "LORA",
|
| 13 |
+
"r": 32,
|
| 14 |
+
"target_modules": [
|
| 15 |
+
"q_proj",
|
| 16 |
+
"v_proj"
|
| 17 |
+
],
|
| 18 |
+
"task_type": "CAUSAL_LM"
|
| 19 |
+
}
|
robin-7b/adapter_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:320829da8b0ace9f5ada5155c6cb119b8e4c03d9c507c0103ee4765fd8c47d83
|
| 3 |
+
size 33574781
|
robin-7b/all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 9.99,
|
| 3 |
+
"train_loss": 0.8784972819876163,
|
| 4 |
+
"train_runtime": 13393.5014,
|
| 5 |
+
"train_samples": 72257,
|
| 6 |
+
"train_samples_per_second": 53.949,
|
| 7 |
+
"train_steps_per_second": 0.421
|
| 8 |
+
}
|
robin-7b/special_tokens_map.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{}
|
robin-7b/tokenizer.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
|
| 3 |
+
size 499723
|
robin-7b/tokenizer_config.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "",
|
| 3 |
+
"clean_up_tokenization_spaces": false,
|
| 4 |
+
"eos_token": "",
|
| 5 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 6 |
+
"special_tokens_map_file": "/home/xiangliu/.cache/huggingface/hub/models--aleksickx--llama-7b-hf/snapshots/d7d132438caf5e95800f35dfc46cf82c2be9b365/special_tokens_map.json",
|
| 7 |
+
"tokenizer_class": "LlamaTokenizer",
|
| 8 |
+
"unk_token": ""
|
| 9 |
+
}
|
robin-7b/train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 9.99,
|
| 3 |
+
"train_loss": 0.8784972819876163,
|
| 4 |
+
"train_runtime": 13393.5014,
|
| 5 |
+
"train_samples": 72257,
|
| 6 |
+
"train_samples_per_second": 53.949,
|
| 7 |
+
"train_steps_per_second": 0.421
|
| 8 |
+
}
|
robin-7b/trainer_state.json
ADDED
|
@@ -0,0 +1,1717 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 9.9867197875166,
|
| 5 |
+
"global_step": 5640,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.04,
|
| 12 |
+
"learning_rate": 9.411764705882353e-05,
|
| 13 |
+
"loss": 1.501,
|
| 14 |
+
"step": 20
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.07,
|
| 18 |
+
"learning_rate": 0.00018823529411764707,
|
| 19 |
+
"loss": 1.4128,
|
| 20 |
+
"step": 40
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.11,
|
| 24 |
+
"learning_rate": 0.0002823529411764706,
|
| 25 |
+
"loss": 1.325,
|
| 26 |
+
"step": 60
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.14,
|
| 30 |
+
"learning_rate": 0.00037647058823529414,
|
| 31 |
+
"loss": 1.2828,
|
| 32 |
+
"step": 80
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.18,
|
| 36 |
+
"learning_rate": 0.00047058823529411766,
|
| 37 |
+
"loss": 1.2758,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.21,
|
| 42 |
+
"learning_rate": 0.0005647058823529412,
|
| 43 |
+
"loss": 1.2667,
|
| 44 |
+
"step": 120
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.25,
|
| 48 |
+
"learning_rate": 0.0006588235294117648,
|
| 49 |
+
"loss": 1.2504,
|
| 50 |
+
"step": 140
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.28,
|
| 54 |
+
"learning_rate": 0.0007529411764705883,
|
| 55 |
+
"loss": 1.2394,
|
| 56 |
+
"step": 160
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.32,
|
| 60 |
+
"learning_rate": 0.0007999934028874321,
|
| 61 |
+
"loss": 1.228,
|
| 62 |
+
"step": 180
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.35,
|
| 66 |
+
"learning_rate": 0.0007999406272925394,
|
| 67 |
+
"loss": 1.2138,
|
| 68 |
+
"step": 200
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.39,
|
| 72 |
+
"learning_rate": 0.0007998350830660272,
|
| 73 |
+
"loss": 1.2122,
|
| 74 |
+
"step": 220
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.42,
|
| 78 |
+
"learning_rate": 0.0007996767841335234,
|
| 79 |
+
"loss": 1.219,
|
| 80 |
+
"step": 240
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.46,
|
| 84 |
+
"learning_rate": 0.0007994657513811737,
|
| 85 |
+
"loss": 1.1998,
|
| 86 |
+
"step": 260
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.5,
|
| 90 |
+
"learning_rate": 0.0007992020126528848,
|
| 91 |
+
"loss": 1.188,
|
| 92 |
+
"step": 280
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.53,
|
| 96 |
+
"learning_rate": 0.0007988856027466511,
|
| 97 |
+
"loss": 1.1931,
|
| 98 |
+
"step": 300
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.57,
|
| 102 |
+
"learning_rate": 0.000798516563409964,
|
| 103 |
+
"loss": 1.1753,
|
| 104 |
+
"step": 320
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.6,
|
| 108 |
+
"learning_rate": 0.0007980949433343026,
|
| 109 |
+
"loss": 1.173,
|
| 110 |
+
"step": 340
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.64,
|
| 114 |
+
"learning_rate": 0.0007976207981487104,
|
| 115 |
+
"loss": 1.1637,
|
| 116 |
+
"step": 360
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.67,
|
| 120 |
+
"learning_rate": 0.0007970941904124546,
|
| 121 |
+
"loss": 1.1651,
|
| 122 |
+
"step": 380
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.71,
|
| 126 |
+
"learning_rate": 0.0007965151896067728,
|
| 127 |
+
"loss": 1.1559,
|
| 128 |
+
"step": 400
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.74,
|
| 132 |
+
"learning_rate": 0.0007958838721257046,
|
| 133 |
+
"loss": 1.1658,
|
| 134 |
+
"step": 420
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.78,
|
| 138 |
+
"learning_rate": 0.0007952003212660127,
|
| 139 |
+
"loss": 1.1496,
|
| 140 |
+
"step": 440
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.81,
|
| 144 |
+
"learning_rate": 0.0007944646272161933,
|
| 145 |
+
"loss": 1.1471,
|
| 146 |
+
"step": 460
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.85,
|
| 150 |
+
"learning_rate": 0.0007936768870445747,
|
| 151 |
+
"loss": 1.1311,
|
| 152 |
+
"step": 480
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.89,
|
| 156 |
+
"learning_rate": 0.0007928372046865116,
|
| 157 |
+
"loss": 1.1301,
|
| 158 |
+
"step": 500
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.92,
|
| 162 |
+
"learning_rate": 0.0007919456909306711,
|
| 163 |
+
"loss": 1.134,
|
| 164 |
+
"step": 520
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.96,
|
| 168 |
+
"learning_rate": 0.0007910024634044154,
|
| 169 |
+
"loss": 1.1235,
|
| 170 |
+
"step": 540
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.99,
|
| 174 |
+
"learning_rate": 0.0007900076465582816,
|
| 175 |
+
"loss": 1.1239,
|
| 176 |
+
"step": 560
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 1.03,
|
| 180 |
+
"learning_rate": 0.0007889613716495616,
|
| 181 |
+
"loss": 1.0878,
|
| 182 |
+
"step": 580
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 1.06,
|
| 186 |
+
"learning_rate": 0.0007878637767249839,
|
| 187 |
+
"loss": 1.0879,
|
| 188 |
+
"step": 600
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 1.1,
|
| 192 |
+
"learning_rate": 0.0007867150066024996,
|
| 193 |
+
"loss": 1.0671,
|
| 194 |
+
"step": 620
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 1.13,
|
| 198 |
+
"learning_rate": 0.0007855152128521754,
|
| 199 |
+
"loss": 1.0689,
|
| 200 |
+
"step": 640
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 1.17,
|
| 204 |
+
"learning_rate": 0.0007842645537761941,
|
| 205 |
+
"loss": 1.0794,
|
| 206 |
+
"step": 660
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 1.2,
|
| 210 |
+
"learning_rate": 0.0007829631943879694,
|
| 211 |
+
"loss": 1.0653,
|
| 212 |
+
"step": 680
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 1.24,
|
| 216 |
+
"learning_rate": 0.0007816113063903726,
|
| 217 |
+
"loss": 1.066,
|
| 218 |
+
"step": 700
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 1.27,
|
| 222 |
+
"learning_rate": 0.0007802090681530788,
|
| 223 |
+
"loss": 1.0675,
|
| 224 |
+
"step": 720
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.31,
|
| 228 |
+
"learning_rate": 0.0007787566646890325,
|
| 229 |
+
"loss": 1.0598,
|
| 230 |
+
"step": 740
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 1.35,
|
| 234 |
+
"learning_rate": 0.0007772542876300359,
|
| 235 |
+
"loss": 1.0669,
|
| 236 |
+
"step": 760
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 1.38,
|
| 240 |
+
"learning_rate": 0.0007757021352014663,
|
| 241 |
+
"loss": 1.0558,
|
| 242 |
+
"step": 780
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 1.42,
|
| 246 |
+
"learning_rate": 0.0007741004121961207,
|
| 247 |
+
"loss": 1.0578,
|
| 248 |
+
"step": 800
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 1.45,
|
| 252 |
+
"learning_rate": 0.0007724493299471956,
|
| 253 |
+
"loss": 1.056,
|
| 254 |
+
"step": 820
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 1.49,
|
| 258 |
+
"learning_rate": 0.0007707491063004035,
|
| 259 |
+
"loss": 1.0491,
|
| 260 |
+
"step": 840
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 1.52,
|
| 264 |
+
"learning_rate": 0.0007689999655852306,
|
| 265 |
+
"loss": 1.0497,
|
| 266 |
+
"step": 860
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 1.56,
|
| 270 |
+
"learning_rate": 0.0007672021385853376,
|
| 271 |
+
"loss": 1.0393,
|
| 272 |
+
"step": 880
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 1.59,
|
| 276 |
+
"learning_rate": 0.0007653558625081099,
|
| 277 |
+
"loss": 1.0379,
|
| 278 |
+
"step": 900
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 1.63,
|
| 282 |
+
"learning_rate": 0.0007634613809533613,
|
| 283 |
+
"loss": 1.049,
|
| 284 |
+
"step": 920
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 1.66,
|
| 288 |
+
"learning_rate": 0.0007615189438811918,
|
| 289 |
+
"loss": 1.0594,
|
| 290 |
+
"step": 940
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 1.7,
|
| 294 |
+
"learning_rate": 0.0007595288075790085,
|
| 295 |
+
"loss": 1.0375,
|
| 296 |
+
"step": 960
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.74,
|
| 300 |
+
"learning_rate": 0.0007574912346277103,
|
| 301 |
+
"loss": 1.0453,
|
| 302 |
+
"step": 980
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.77,
|
| 306 |
+
"learning_rate": 0.0007554064938670426,
|
| 307 |
+
"loss": 1.0348,
|
| 308 |
+
"step": 1000
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 1.81,
|
| 312 |
+
"learning_rate": 0.0007532748603601265,
|
| 313 |
+
"loss": 1.0441,
|
| 314 |
+
"step": 1020
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 1.84,
|
| 318 |
+
"learning_rate": 0.0007510966153571667,
|
| 319 |
+
"loss": 1.0237,
|
| 320 |
+
"step": 1040
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 1.88,
|
| 324 |
+
"learning_rate": 0.000748872046258343,
|
| 325 |
+
"loss": 1.0478,
|
| 326 |
+
"step": 1060
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 1.91,
|
| 330 |
+
"learning_rate": 0.0007466014465758899,
|
| 331 |
+
"loss": 1.0289,
|
| 332 |
+
"step": 1080
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 1.95,
|
| 336 |
+
"learning_rate": 0.0007442851158953712,
|
| 337 |
+
"loss": 1.026,
|
| 338 |
+
"step": 1100
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 1.98,
|
| 342 |
+
"learning_rate": 0.0007419233598361512,
|
| 343 |
+
"loss": 1.0244,
|
| 344 |
+
"step": 1120
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 2.02,
|
| 348 |
+
"learning_rate": 0.0007395164900110721,
|
| 349 |
+
"loss": 0.9968,
|
| 350 |
+
"step": 1140
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 2.05,
|
| 354 |
+
"learning_rate": 0.0007370648239853385,
|
| 355 |
+
"loss": 0.9798,
|
| 356 |
+
"step": 1160
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 2.09,
|
| 360 |
+
"learning_rate": 0.0007345686852346176,
|
| 361 |
+
"loss": 0.9529,
|
| 362 |
+
"step": 1180
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 2.12,
|
| 366 |
+
"learning_rate": 0.0007320284031023603,
|
| 367 |
+
"loss": 0.9666,
|
| 368 |
+
"step": 1200
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 2.16,
|
| 372 |
+
"learning_rate": 0.000729444312756346,
|
| 373 |
+
"loss": 0.9797,
|
| 374 |
+
"step": 1220
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 2.2,
|
| 378 |
+
"learning_rate": 0.0007268167551444611,
|
| 379 |
+
"loss": 0.9674,
|
| 380 |
+
"step": 1240
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 2.23,
|
| 384 |
+
"learning_rate": 0.0007241460769497138,
|
| 385 |
+
"loss": 0.9666,
|
| 386 |
+
"step": 1260
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 2.27,
|
| 390 |
+
"learning_rate": 0.0007214326305444917,
|
| 391 |
+
"loss": 0.9785,
|
| 392 |
+
"step": 1280
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 2.3,
|
| 396 |
+
"learning_rate": 0.0007186767739440701,
|
| 397 |
+
"loss": 0.9629,
|
| 398 |
+
"step": 1300
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 2.34,
|
| 402 |
+
"learning_rate": 0.0007158788707593748,
|
| 403 |
+
"loss": 0.973,
|
| 404 |
+
"step": 1320
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 2.37,
|
| 408 |
+
"learning_rate": 0.0007130392901490069,
|
| 409 |
+
"loss": 0.9649,
|
| 410 |
+
"step": 1340
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 2.41,
|
| 414 |
+
"learning_rate": 0.0007101584067705355,
|
| 415 |
+
"loss": 0.9766,
|
| 416 |
+
"step": 1360
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 2.44,
|
| 420 |
+
"learning_rate": 0.0007072366007310646,
|
| 421 |
+
"loss": 0.954,
|
| 422 |
+
"step": 1380
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 2.48,
|
| 426 |
+
"learning_rate": 0.0007042742575370822,
|
| 427 |
+
"loss": 0.9576,
|
| 428 |
+
"step": 1400
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 2.51,
|
| 432 |
+
"learning_rate": 0.0007012717680435956,
|
| 433 |
+
"loss": 0.9783,
|
| 434 |
+
"step": 1420
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 2.55,
|
| 438 |
+
"learning_rate": 0.0006982295284025612,
|
| 439 |
+
"loss": 0.9553,
|
| 440 |
+
"step": 1440
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 2.59,
|
| 444 |
+
"learning_rate": 0.0006951479400106161,
|
| 445 |
+
"loss": 0.951,
|
| 446 |
+
"step": 1460
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 2.62,
|
| 450 |
+
"learning_rate": 0.000692027409456118,
|
| 451 |
+
"loss": 0.9647,
|
| 452 |
+
"step": 1480
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 2.66,
|
| 456 |
+
"learning_rate": 0.0006888683484654981,
|
| 457 |
+
"loss": 0.9656,
|
| 458 |
+
"step": 1500
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 2.69,
|
| 462 |
+
"learning_rate": 0.0006856711738489386,
|
| 463 |
+
"loss": 0.9507,
|
| 464 |
+
"step": 1520
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 2.73,
|
| 468 |
+
"learning_rate": 0.0006824363074453778,
|
| 469 |
+
"loss": 0.9496,
|
| 470 |
+
"step": 1540
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 2.76,
|
| 474 |
+
"learning_rate": 0.0006791641760668519,
|
| 475 |
+
"loss": 0.9571,
|
| 476 |
+
"step": 1560
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 2.8,
|
| 480 |
+
"learning_rate": 0.0006758552114421815,
|
| 481 |
+
"loss": 0.9494,
|
| 482 |
+
"step": 1580
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 2.83,
|
| 486 |
+
"learning_rate": 0.0006725098501600088,
|
| 487 |
+
"loss": 0.9556,
|
| 488 |
+
"step": 1600
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 2.87,
|
| 492 |
+
"learning_rate": 0.0006691285336111928,
|
| 493 |
+
"loss": 0.9411,
|
| 494 |
+
"step": 1620
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 2.9,
|
| 498 |
+
"learning_rate": 0.0006657117079305725,
|
| 499 |
+
"loss": 0.9501,
|
| 500 |
+
"step": 1640
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 2.94,
|
| 504 |
+
"learning_rate": 0.0006622598239381033,
|
| 505 |
+
"loss": 0.9598,
|
| 506 |
+
"step": 1660
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 2.97,
|
| 510 |
+
"learning_rate": 0.0006587733370793743,
|
| 511 |
+
"loss": 0.9599,
|
| 512 |
+
"step": 1680
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 3.01,
|
| 516 |
+
"learning_rate": 0.0006552527073655178,
|
| 517 |
+
"loss": 0.9306,
|
| 518 |
+
"step": 1700
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 3.05,
|
| 522 |
+
"learning_rate": 0.0006516983993125138,
|
| 523 |
+
"loss": 0.9013,
|
| 524 |
+
"step": 1720
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 3.08,
|
| 528 |
+
"learning_rate": 0.0006481108818799015,
|
| 529 |
+
"loss": 0.8798,
|
| 530 |
+
"step": 1740
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 3.12,
|
| 534 |
+
"learning_rate": 0.0006444906284089044,
|
| 535 |
+
"loss": 0.903,
|
| 536 |
+
"step": 1760
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 3.15,
|
| 540 |
+
"learning_rate": 0.000640838116559977,
|
| 541 |
+
"loss": 0.9017,
|
| 542 |
+
"step": 1780
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 3.19,
|
| 546 |
+
"learning_rate": 0.0006371538282497815,
|
| 547 |
+
"loss": 0.9051,
|
| 548 |
+
"step": 1800
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 3.22,
|
| 552 |
+
"learning_rate": 0.0006334382495876036,
|
| 553 |
+
"loss": 0.8966,
|
| 554 |
+
"step": 1820
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 3.26,
|
| 558 |
+
"learning_rate": 0.0006296918708112143,
|
| 559 |
+
"loss": 0.8863,
|
| 560 |
+
"step": 1840
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 3.29,
|
| 564 |
+
"learning_rate": 0.0006259151862221875,
|
| 565 |
+
"loss": 0.8926,
|
| 566 |
+
"step": 1860
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 3.33,
|
| 570 |
+
"learning_rate": 0.0006221086941206817,
|
| 571 |
+
"loss": 0.908,
|
| 572 |
+
"step": 1880
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 3.36,
|
| 576 |
+
"learning_rate": 0.0006182728967396925,
|
| 577 |
+
"loss": 0.9007,
|
| 578 |
+
"step": 1900
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 3.4,
|
| 582 |
+
"learning_rate": 0.0006144083001787886,
|
| 583 |
+
"loss": 0.8872,
|
| 584 |
+
"step": 1920
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 3.44,
|
| 588 |
+
"learning_rate": 0.0006105154143373362,
|
| 589 |
+
"loss": 0.8984,
|
| 590 |
+
"step": 1940
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 3.47,
|
| 594 |
+
"learning_rate": 0.0006065947528472215,
|
| 595 |
+
"loss": 0.9123,
|
| 596 |
+
"step": 1960
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 3.51,
|
| 600 |
+
"learning_rate": 0.0006026468330050827,
|
| 601 |
+
"loss": 0.8929,
|
| 602 |
+
"step": 1980
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 3.54,
|
| 606 |
+
"learning_rate": 0.0005986721757040564,
|
| 607 |
+
"loss": 0.9145,
|
| 608 |
+
"step": 2000
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 3.58,
|
| 612 |
+
"learning_rate": 0.0005946713053650507,
|
| 613 |
+
"loss": 0.8867,
|
| 614 |
+
"step": 2020
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 3.61,
|
| 618 |
+
"learning_rate": 0.0005906447498675521,
|
| 619 |
+
"loss": 0.8914,
|
| 620 |
+
"step": 2040
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 3.65,
|
| 624 |
+
"learning_rate": 0.0005865930404799774,
|
| 625 |
+
"loss": 0.8946,
|
| 626 |
+
"step": 2060
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 3.68,
|
| 630 |
+
"learning_rate": 0.0005825167117895765,
|
| 631 |
+
"loss": 0.892,
|
| 632 |
+
"step": 2080
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 3.72,
|
| 636 |
+
"learning_rate": 0.0005784163016318987,
|
| 637 |
+
"loss": 0.8875,
|
| 638 |
+
"step": 2100
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 3.75,
|
| 642 |
+
"learning_rate": 0.0005742923510198303,
|
| 643 |
+
"loss": 0.888,
|
| 644 |
+
"step": 2120
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 3.79,
|
| 648 |
+
"learning_rate": 0.0005701454040722124,
|
| 649 |
+
"loss": 0.9078,
|
| 650 |
+
"step": 2140
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 3.82,
|
| 654 |
+
"learning_rate": 0.0005659760079420498,
|
| 655 |
+
"loss": 0.9027,
|
| 656 |
+
"step": 2160
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 3.86,
|
| 660 |
+
"learning_rate": 0.000561784712744318,
|
| 661 |
+
"loss": 0.8997,
|
| 662 |
+
"step": 2180
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 3.9,
|
| 666 |
+
"learning_rate": 0.0005575720714833808,
|
| 667 |
+
"loss": 0.9053,
|
| 668 |
+
"step": 2200
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 3.93,
|
| 672 |
+
"learning_rate": 0.0005533386399800275,
|
| 673 |
+
"loss": 0.9054,
|
| 674 |
+
"step": 2220
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 3.97,
|
| 678 |
+
"learning_rate": 0.0005490849767981348,
|
| 679 |
+
"loss": 0.8988,
|
| 680 |
+
"step": 2240
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 4.0,
|
| 684 |
+
"learning_rate": 0.0005448116431709716,
|
| 685 |
+
"loss": 0.8903,
|
| 686 |
+
"step": 2260
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 4.04,
|
| 690 |
+
"learning_rate": 0.0005405192029271477,
|
| 691 |
+
"loss": 0.8373,
|
| 692 |
+
"step": 2280
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 4.07,
|
| 696 |
+
"learning_rate": 0.0005362082224162223,
|
| 697 |
+
"loss": 0.8336,
|
| 698 |
+
"step": 2300
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 4.11,
|
| 702 |
+
"learning_rate": 0.0005318792704339792,
|
| 703 |
+
"loss": 0.8483,
|
| 704 |
+
"step": 2320
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 4.14,
|
| 708 |
+
"learning_rate": 0.0005275329181473787,
|
| 709 |
+
"loss": 0.8453,
|
| 710 |
+
"step": 2340
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 4.18,
|
| 714 |
+
"learning_rate": 0.0005231697390191976,
|
| 715 |
+
"loss": 0.8351,
|
| 716 |
+
"step": 2360
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 4.21,
|
| 720 |
+
"learning_rate": 0.000518790308732366,
|
| 721 |
+
"loss": 0.8329,
|
| 722 |
+
"step": 2380
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 4.25,
|
| 726 |
+
"learning_rate": 0.0005143952051140103,
|
| 727 |
+
"loss": 0.8394,
|
| 728 |
+
"step": 2400
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 4.29,
|
| 732 |
+
"learning_rate": 0.000509985008059215,
|
| 733 |
+
"loss": 0.8526,
|
| 734 |
+
"step": 2420
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 4.32,
|
| 738 |
+
"learning_rate": 0.0005055602994545098,
|
| 739 |
+
"loss": 0.826,
|
| 740 |
+
"step": 2440
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 4.36,
|
| 744 |
+
"learning_rate": 0.0005011216631010953,
|
| 745 |
+
"loss": 0.849,
|
| 746 |
+
"step": 2460
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 4.39,
|
| 750 |
+
"learning_rate": 0.0004966696846378156,
|
| 751 |
+
"loss": 0.8507,
|
| 752 |
+
"step": 2480
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 4.43,
|
| 756 |
+
"learning_rate": 0.000492204951463888,
|
| 757 |
+
"loss": 0.8461,
|
| 758 |
+
"step": 2500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 4.46,
|
| 762 |
+
"learning_rate": 0.00048772805266140154,
|
| 763 |
+
"loss": 0.8533,
|
| 764 |
+
"step": 2520
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 4.5,
|
| 768 |
+
"learning_rate": 0.00048323957891759203,
|
| 769 |
+
"loss": 0.8384,
|
| 770 |
+
"step": 2540
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 4.53,
|
| 774 |
+
"learning_rate": 0.00047874012244690696,
|
| 775 |
+
"loss": 0.842,
|
| 776 |
+
"step": 2560
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 4.57,
|
| 780 |
+
"learning_rate": 0.000474230276912867,
|
| 781 |
+
"loss": 0.8608,
|
| 782 |
+
"step": 2580
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 4.6,
|
| 786 |
+
"learning_rate": 0.00046971063734973833,
|
| 787 |
+
"loss": 0.8562,
|
| 788 |
+
"step": 2600
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 4.64,
|
| 792 |
+
"learning_rate": 0.0004651818000840229,
|
| 793 |
+
"loss": 0.8594,
|
| 794 |
+
"step": 2620
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 4.67,
|
| 798 |
+
"learning_rate": 0.0004606443626557778,
|
| 799 |
+
"loss": 0.8608,
|
| 800 |
+
"step": 2640
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 4.71,
|
| 804 |
+
"learning_rate": 0.0004560989237397758,
|
| 805 |
+
"loss": 0.8486,
|
| 806 |
+
"step": 2660
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 4.75,
|
| 810 |
+
"learning_rate": 0.00045154608306651514,
|
| 811 |
+
"loss": 0.869,
|
| 812 |
+
"step": 2680
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 4.78,
|
| 816 |
+
"learning_rate": 0.0004469864413430907,
|
| 817 |
+
"loss": 0.8482,
|
| 818 |
+
"step": 2700
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 4.82,
|
| 822 |
+
"learning_rate": 0.00044242060017393573,
|
| 823 |
+
"loss": 0.8583,
|
| 824 |
+
"step": 2720
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 4.85,
|
| 828 |
+
"learning_rate": 0.00043784916198144543,
|
| 829 |
+
"loss": 0.8582,
|
| 830 |
+
"step": 2740
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 4.89,
|
| 834 |
+
"learning_rate": 0.00043327272992649317,
|
| 835 |
+
"loss": 0.8504,
|
| 836 |
+
"step": 2760
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 4.92,
|
| 840 |
+
"learning_rate": 0.00042869190782884794,
|
| 841 |
+
"loss": 0.8592,
|
| 842 |
+
"step": 2780
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 4.96,
|
| 846 |
+
"learning_rate": 0.00042410730008750623,
|
| 847 |
+
"loss": 0.8545,
|
| 848 |
+
"step": 2800
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 4.99,
|
| 852 |
+
"learning_rate": 0.00041951951160094664,
|
| 853 |
+
"loss": 0.855,
|
| 854 |
+
"step": 2820
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 5.03,
|
| 858 |
+
"learning_rate": 0.00041492914768731927,
|
| 859 |
+
"loss": 0.7869,
|
| 860 |
+
"step": 2840
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 5.06,
|
| 864 |
+
"learning_rate": 0.0004103368140045789,
|
| 865 |
+
"loss": 0.8083,
|
| 866 |
+
"step": 2860
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 5.1,
|
| 870 |
+
"learning_rate": 0.00040574311647057366,
|
| 871 |
+
"loss": 0.8108,
|
| 872 |
+
"step": 2880
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 5.14,
|
| 876 |
+
"learning_rate": 0.00040114866118310045,
|
| 877 |
+
"loss": 0.7968,
|
| 878 |
+
"step": 2900
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 5.17,
|
| 882 |
+
"learning_rate": 0.0003965540543399344,
|
| 883 |
+
"loss": 0.8016,
|
| 884 |
+
"step": 2920
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 5.21,
|
| 888 |
+
"learning_rate": 0.00039195990215884756,
|
| 889 |
+
"loss": 0.7967,
|
| 890 |
+
"step": 2940
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 5.24,
|
| 894 |
+
"learning_rate": 0.00038736681079762293,
|
| 895 |
+
"loss": 0.8096,
|
| 896 |
+
"step": 2960
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 5.28,
|
| 900 |
+
"learning_rate": 0.0003827753862740779,
|
| 901 |
+
"loss": 0.8073,
|
| 902 |
+
"step": 2980
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 5.31,
|
| 906 |
+
"learning_rate": 0.0003781862343861055,
|
| 907 |
+
"loss": 0.804,
|
| 908 |
+
"step": 3000
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 5.35,
|
| 912 |
+
"learning_rate": 0.00037359996063174425,
|
| 913 |
+
"loss": 0.8119,
|
| 914 |
+
"step": 3020
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 5.38,
|
| 918 |
+
"learning_rate": 0.0003690171701292887,
|
| 919 |
+
"loss": 0.7997,
|
| 920 |
+
"step": 3040
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 5.42,
|
| 924 |
+
"learning_rate": 0.0003644384675374489,
|
| 925 |
+
"loss": 0.8202,
|
| 926 |
+
"step": 3060
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 5.45,
|
| 930 |
+
"learning_rate": 0.0003598644569755713,
|
| 931 |
+
"loss": 0.815,
|
| 932 |
+
"step": 3080
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 5.49,
|
| 936 |
+
"learning_rate": 0.00035529574194393033,
|
| 937 |
+
"loss": 0.825,
|
| 938 |
+
"step": 3100
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 5.52,
|
| 942 |
+
"learning_rate": 0.00035073292524410207,
|
| 943 |
+
"loss": 0.8171,
|
| 944 |
+
"step": 3120
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 5.56,
|
| 948 |
+
"learning_rate": 0.00034617660889943,
|
| 949 |
+
"loss": 0.7921,
|
| 950 |
+
"step": 3140
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 5.6,
|
| 954 |
+
"learning_rate": 0.00034162739407559285,
|
| 955 |
+
"loss": 0.8299,
|
| 956 |
+
"step": 3160
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 5.63,
|
| 960 |
+
"learning_rate": 0.0003370858810012869,
|
| 961 |
+
"loss": 0.811,
|
| 962 |
+
"step": 3180
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 5.67,
|
| 966 |
+
"learning_rate": 0.00033255266888903006,
|
| 967 |
+
"loss": 0.8093,
|
| 968 |
+
"step": 3200
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 5.7,
|
| 972 |
+
"learning_rate": 0.00032802835585610225,
|
| 973 |
+
"loss": 0.8106,
|
| 974 |
+
"step": 3220
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 5.74,
|
| 978 |
+
"learning_rate": 0.00032351353884562783,
|
| 979 |
+
"loss": 0.8053,
|
| 980 |
+
"step": 3240
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 5.77,
|
| 984 |
+
"learning_rate": 0.00031900881354781556,
|
| 985 |
+
"loss": 0.8161,
|
| 986 |
+
"step": 3260
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 5.81,
|
| 990 |
+
"learning_rate": 0.00031451477432136154,
|
| 991 |
+
"loss": 0.8186,
|
| 992 |
+
"step": 3280
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 5.84,
|
| 996 |
+
"learning_rate": 0.0003100320141150293,
|
| 997 |
+
"loss": 0.8046,
|
| 998 |
+
"step": 3300
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 5.88,
|
| 1002 |
+
"learning_rate": 0.00030556112438941526,
|
| 1003 |
+
"loss": 0.8236,
|
| 1004 |
+
"step": 3320
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 5.91,
|
| 1008 |
+
"learning_rate": 0.00030110269503891084,
|
| 1009 |
+
"loss": 0.8057,
|
| 1010 |
+
"step": 3340
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 5.95,
|
| 1014 |
+
"learning_rate": 0.0002966573143138713,
|
| 1015 |
+
"loss": 0.8109,
|
| 1016 |
+
"step": 3360
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 5.98,
|
| 1020 |
+
"learning_rate": 0.00029222556874300036,
|
| 1021 |
+
"loss": 0.8163,
|
| 1022 |
+
"step": 3380
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 6.02,
|
| 1026 |
+
"learning_rate": 0.0002878080430559646,
|
| 1027 |
+
"loss": 0.7901,
|
| 1028 |
+
"step": 3400
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 6.06,
|
| 1032 |
+
"learning_rate": 0.0002834053201062417,
|
| 1033 |
+
"loss": 0.7749,
|
| 1034 |
+
"step": 3420
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 6.09,
|
| 1038 |
+
"learning_rate": 0.00027901798079421977,
|
| 1039 |
+
"loss": 0.7775,
|
| 1040 |
+
"step": 3440
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 6.13,
|
| 1044 |
+
"learning_rate": 0.0002746466039905513,
|
| 1045 |
+
"loss": 0.7589,
|
| 1046 |
+
"step": 3460
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 6.16,
|
| 1050 |
+
"learning_rate": 0.000270291766459777,
|
| 1051 |
+
"loss": 0.7742,
|
| 1052 |
+
"step": 3480
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 6.2,
|
| 1056 |
+
"learning_rate": 0.00026595404278422684,
|
| 1057 |
+
"loss": 0.7914,
|
| 1058 |
+
"step": 3500
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 6.23,
|
| 1062 |
+
"learning_rate": 0.00026163400528820836,
|
| 1063 |
+
"loss": 0.7909,
|
| 1064 |
+
"step": 3520
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 6.27,
|
| 1068 |
+
"learning_rate": 0.0002573322239624947,
|
| 1069 |
+
"loss": 0.7653,
|
| 1070 |
+
"step": 3540
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 6.3,
|
| 1074 |
+
"learning_rate": 0.000253049266389118,
|
| 1075 |
+
"loss": 0.7779,
|
| 1076 |
+
"step": 3560
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 6.34,
|
| 1080 |
+
"learning_rate": 0.0002487856976664831,
|
| 1081 |
+
"loss": 0.7771,
|
| 1082 |
+
"step": 3580
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 6.37,
|
| 1086 |
+
"learning_rate": 0.00024454208033480683,
|
| 1087 |
+
"loss": 0.771,
|
| 1088 |
+
"step": 3600
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 6.41,
|
| 1092 |
+
"learning_rate": 0.00024031897430189695,
|
| 1093 |
+
"loss": 0.7788,
|
| 1094 |
+
"step": 3620
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 6.45,
|
| 1098 |
+
"learning_rate": 0.00023611693676927606,
|
| 1099 |
+
"loss": 0.7798,
|
| 1100 |
+
"step": 3640
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 6.48,
|
| 1104 |
+
"learning_rate": 0.00023193652215866429,
|
| 1105 |
+
"loss": 0.7779,
|
| 1106 |
+
"step": 3660
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 6.52,
|
| 1110 |
+
"learning_rate": 0.00022777828203882875,
|
| 1111 |
+
"loss": 0.7784,
|
| 1112 |
+
"step": 3680
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 6.55,
|
| 1116 |
+
"learning_rate": 0.00022364276505280794,
|
| 1117 |
+
"loss": 0.7946,
|
| 1118 |
+
"step": 3700
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 6.59,
|
| 1122 |
+
"learning_rate": 0.0002195305168455239,
|
| 1123 |
+
"loss": 0.774,
|
| 1124 |
+
"step": 3720
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 6.62,
|
| 1128 |
+
"learning_rate": 0.00021544207999178917,
|
| 1129 |
+
"loss": 0.7823,
|
| 1130 |
+
"step": 3740
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 6.66,
|
| 1134 |
+
"learning_rate": 0.00021137799392471814,
|
| 1135 |
+
"loss": 0.7779,
|
| 1136 |
+
"step": 3760
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 6.69,
|
| 1140 |
+
"learning_rate": 0.00020733879486455433,
|
| 1141 |
+
"loss": 0.7867,
|
| 1142 |
+
"step": 3780
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 6.73,
|
| 1146 |
+
"learning_rate": 0.0002033250157479206,
|
| 1147 |
+
"loss": 0.7738,
|
| 1148 |
+
"step": 3800
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 6.76,
|
| 1152 |
+
"learning_rate": 0.0001993371861575028,
|
| 1153 |
+
"loss": 0.7814,
|
| 1154 |
+
"step": 3820
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 6.8,
|
| 1158 |
+
"learning_rate": 0.00019537583225217605,
|
| 1159 |
+
"loss": 0.7695,
|
| 1160 |
+
"step": 3840
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 6.83,
|
| 1164 |
+
"learning_rate": 0.00019144147669758322,
|
| 1165 |
+
"loss": 0.7846,
|
| 1166 |
+
"step": 3860
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 6.87,
|
| 1170 |
+
"learning_rate": 0.00018753463859717283,
|
| 1171 |
+
"loss": 0.7779,
|
| 1172 |
+
"step": 3880
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 6.91,
|
| 1176 |
+
"learning_rate": 0.0001836558334237088,
|
| 1177 |
+
"loss": 0.7769,
|
| 1178 |
+
"step": 3900
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 6.94,
|
| 1182 |
+
"learning_rate": 0.0001798055729512579,
|
| 1183 |
+
"loss": 0.782,
|
| 1184 |
+
"step": 3920
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 6.98,
|
| 1188 |
+
"learning_rate": 0.00017598436518766596,
|
| 1189 |
+
"loss": 0.7683,
|
| 1190 |
+
"step": 3940
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 7.01,
|
| 1194 |
+
"learning_rate": 0.0001721927143075305,
|
| 1195 |
+
"loss": 0.7674,
|
| 1196 |
+
"step": 3960
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 7.05,
|
| 1200 |
+
"learning_rate": 0.00016843112058567935,
|
| 1201 |
+
"loss": 0.7427,
|
| 1202 |
+
"step": 3980
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 7.08,
|
| 1206 |
+
"learning_rate": 0.00016470008033116443,
|
| 1207 |
+
"loss": 0.7627,
|
| 1208 |
+
"step": 4000
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 7.12,
|
| 1212 |
+
"learning_rate": 0.00016100008582177705,
|
| 1213 |
+
"loss": 0.7541,
|
| 1214 |
+
"step": 4020
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 7.15,
|
| 1218 |
+
"learning_rate": 0.00015733162523909707,
|
| 1219 |
+
"loss": 0.7654,
|
| 1220 |
+
"step": 4040
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 7.19,
|
| 1224 |
+
"learning_rate": 0.0001536951826040813,
|
| 1225 |
+
"loss": 0.7382,
|
| 1226 |
+
"step": 4060
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 7.22,
|
| 1230 |
+
"learning_rate": 0.0001500912377132013,
|
| 1231 |
+
"loss": 0.7555,
|
| 1232 |
+
"step": 4080
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 7.26,
|
| 1236 |
+
"learning_rate": 0.00014652026607513848,
|
| 1237 |
+
"loss": 0.748,
|
| 1238 |
+
"step": 4100
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 7.3,
|
| 1242 |
+
"learning_rate": 0.00014298273884804478,
|
| 1243 |
+
"loss": 0.753,
|
| 1244 |
+
"step": 4120
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 7.33,
|
| 1248 |
+
"learning_rate": 0.00013947912277737808,
|
| 1249 |
+
"loss": 0.7594,
|
| 1250 |
+
"step": 4140
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 7.37,
|
| 1254 |
+
"learning_rate": 0.00013600988013431832,
|
| 1255 |
+
"loss": 0.7387,
|
| 1256 |
+
"step": 4160
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 7.4,
|
| 1260 |
+
"learning_rate": 0.00013257546865477572,
|
| 1261 |
+
"loss": 0.7551,
|
| 1262 |
+
"step": 4180
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 7.44,
|
| 1266 |
+
"learning_rate": 0.00012917634147899607,
|
| 1267 |
+
"loss": 0.7609,
|
| 1268 |
+
"step": 4200
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 7.47,
|
| 1272 |
+
"learning_rate": 0.00012581294709177327,
|
| 1273 |
+
"loss": 0.762,
|
| 1274 |
+
"step": 4220
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 7.51,
|
| 1278 |
+
"learning_rate": 0.00012248572926327537,
|
| 1279 |
+
"loss": 0.755,
|
| 1280 |
+
"step": 4240
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 7.54,
|
| 1284 |
+
"learning_rate": 0.00011919512699049314,
|
| 1285 |
+
"loss": 0.753,
|
| 1286 |
+
"step": 4260
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 7.58,
|
| 1290 |
+
"learning_rate": 0.00011594157443931872,
|
| 1291 |
+
"loss": 0.7603,
|
| 1292 |
+
"step": 4280
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 7.61,
|
| 1296 |
+
"learning_rate": 0.0001127255008872604,
|
| 1297 |
+
"loss": 0.7565,
|
| 1298 |
+
"step": 4300
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 7.65,
|
| 1302 |
+
"learning_rate": 0.00010954733066680401,
|
| 1303 |
+
"loss": 0.7542,
|
| 1304 |
+
"step": 4320
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 7.68,
|
| 1308 |
+
"learning_rate": 0.00010640748310942559,
|
| 1309 |
+
"loss": 0.7674,
|
| 1310 |
+
"step": 4340
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 7.72,
|
| 1314 |
+
"learning_rate": 0.00010330637249026445,
|
| 1315 |
+
"loss": 0.7447,
|
| 1316 |
+
"step": 4360
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 7.76,
|
| 1320 |
+
"learning_rate": 0.00010024440797346324,
|
| 1321 |
+
"loss": 0.7582,
|
| 1322 |
+
"step": 4380
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 7.79,
|
| 1326 |
+
"learning_rate": 9.722199355818227e-05,
|
| 1327 |
+
"loss": 0.7598,
|
| 1328 |
+
"step": 4400
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 7.83,
|
| 1332 |
+
"learning_rate": 9.423952802529564e-05,
|
| 1333 |
+
"loss": 0.7389,
|
| 1334 |
+
"step": 4420
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 7.86,
|
| 1338 |
+
"learning_rate": 9.129740488477518e-05,
|
| 1339 |
+
"loss": 0.759,
|
| 1340 |
+
"step": 4440
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 7.9,
|
| 1344 |
+
"learning_rate": 8.83960123237706e-05,
|
| 1345 |
+
"loss": 0.7442,
|
| 1346 |
+
"step": 4460
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 7.93,
|
| 1350 |
+
"learning_rate": 8.553573315539188e-05,
|
| 1351 |
+
"loss": 0.7627,
|
| 1352 |
+
"step": 4480
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 7.97,
|
| 1356 |
+
"learning_rate": 8.271694476819956e-05,
|
| 1357 |
+
"loss": 0.7531,
|
| 1358 |
+
"step": 4500
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 8.0,
|
| 1362 |
+
"learning_rate": 7.994001907641262e-05,
|
| 1363 |
+
"loss": 0.7567,
|
| 1364 |
+
"step": 4520
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 8.04,
|
| 1368 |
+
"learning_rate": 7.720532247083743e-05,
|
| 1369 |
+
"loss": 0.7274,
|
| 1370 |
+
"step": 4540
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 8.07,
|
| 1374 |
+
"learning_rate": 7.451321577052533e-05,
|
| 1375 |
+
"loss": 0.7347,
|
| 1376 |
+
"step": 4560
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 8.11,
|
| 1380 |
+
"learning_rate": 7.18640541751661e-05,
|
| 1381 |
+
"loss": 0.752,
|
| 1382 |
+
"step": 4580
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 8.15,
|
| 1386 |
+
"learning_rate": 6.925818721822239e-05,
|
| 1387 |
+
"loss": 0.7293,
|
| 1388 |
+
"step": 4600
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 8.18,
|
| 1392 |
+
"learning_rate": 6.669595872081211e-05,
|
| 1393 |
+
"loss": 0.7213,
|
| 1394 |
+
"step": 4620
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 8.22,
|
| 1398 |
+
"learning_rate": 6.417770674634365e-05,
|
| 1399 |
+
"loss": 0.7246,
|
| 1400 |
+
"step": 4640
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 8.25,
|
| 1404 |
+
"learning_rate": 6.170376355591204e-05,
|
| 1405 |
+
"loss": 0.7353,
|
| 1406 |
+
"step": 4660
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 8.29,
|
| 1410 |
+
"learning_rate": 5.9274455564459896e-05,
|
| 1411 |
+
"loss": 0.7396,
|
| 1412 |
+
"step": 4680
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 8.32,
|
| 1416 |
+
"learning_rate": 5.689010329770965e-05,
|
| 1417 |
+
"loss": 0.7401,
|
| 1418 |
+
"step": 4700
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 8.36,
|
| 1422 |
+
"learning_rate": 5.455102134987304e-05,
|
| 1423 |
+
"loss": 0.7399,
|
| 1424 |
+
"step": 4720
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 8.39,
|
| 1428 |
+
"learning_rate": 5.225751834214339e-05,
|
| 1429 |
+
"loss": 0.7427,
|
| 1430 |
+
"step": 4740
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 8.43,
|
| 1434 |
+
"learning_rate": 5.000989688197555e-05,
|
| 1435 |
+
"loss": 0.7289,
|
| 1436 |
+
"step": 4760
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 8.46,
|
| 1440 |
+
"learning_rate": 4.780845352315968e-05,
|
| 1441 |
+
"loss": 0.7398,
|
| 1442 |
+
"step": 4780
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 8.5,
|
| 1446 |
+
"learning_rate": 4.565347872669339e-05,
|
| 1447 |
+
"loss": 0.7332,
|
| 1448 |
+
"step": 4800
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 8.53,
|
| 1452 |
+
"learning_rate": 4.3545256822458445e-05,
|
| 1453 |
+
"loss": 0.7365,
|
| 1454 |
+
"step": 4820
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 8.57,
|
| 1458 |
+
"learning_rate": 4.148406597170529e-05,
|
| 1459 |
+
"loss": 0.7347,
|
| 1460 |
+
"step": 4840
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 8.61,
|
| 1464 |
+
"learning_rate": 3.947017813035254e-05,
|
| 1465 |
+
"loss": 0.7515,
|
| 1466 |
+
"step": 4860
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 8.64,
|
| 1470 |
+
"learning_rate": 3.7503859013104806e-05,
|
| 1471 |
+
"loss": 0.7397,
|
| 1472 |
+
"step": 4880
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 8.68,
|
| 1476 |
+
"learning_rate": 3.5585368058393834e-05,
|
| 1477 |
+
"loss": 0.7597,
|
| 1478 |
+
"step": 4900
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 8.71,
|
| 1482 |
+
"learning_rate": 3.3714958394147975e-05,
|
| 1483 |
+
"loss": 0.7391,
|
| 1484 |
+
"step": 4920
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 8.75,
|
| 1488 |
+
"learning_rate": 3.1892876804394144e-05,
|
| 1489 |
+
"loss": 0.7359,
|
| 1490 |
+
"step": 4940
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 8.78,
|
| 1494 |
+
"learning_rate": 3.0119363696697078e-05,
|
| 1495 |
+
"loss": 0.7412,
|
| 1496 |
+
"step": 4960
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 8.82,
|
| 1500 |
+
"learning_rate": 2.839465307043927e-05,
|
| 1501 |
+
"loss": 0.7485,
|
| 1502 |
+
"step": 4980
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 8.85,
|
| 1506 |
+
"learning_rate": 2.6718972485947037e-05,
|
| 1507 |
+
"loss": 0.7534,
|
| 1508 |
+
"step": 5000
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 8.89,
|
| 1512 |
+
"learning_rate": 2.5092543034466264e-05,
|
| 1513 |
+
"loss": 0.7405,
|
| 1514 |
+
"step": 5020
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 8.92,
|
| 1518 |
+
"learning_rate": 2.3515579308990597e-05,
|
| 1519 |
+
"loss": 0.7415,
|
| 1520 |
+
"step": 5040
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 8.96,
|
| 1524 |
+
"learning_rate": 2.1988289375948524e-05,
|
| 1525 |
+
"loss": 0.7309,
|
| 1526 |
+
"step": 5060
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 9.0,
|
| 1530 |
+
"learning_rate": 2.0510874747750575e-05,
|
| 1531 |
+
"loss": 0.7418,
|
| 1532 |
+
"step": 5080
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 9.03,
|
| 1536 |
+
"learning_rate": 1.9083530356201407e-05,
|
| 1537 |
+
"loss": 0.7469,
|
| 1538 |
+
"step": 5100
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 9.07,
|
| 1542 |
+
"learning_rate": 1.7706444526780585e-05,
|
| 1543 |
+
"loss": 0.7385,
|
| 1544 |
+
"step": 5120
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 9.1,
|
| 1548 |
+
"learning_rate": 1.637979895379429e-05,
|
| 1549 |
+
"loss": 0.7333,
|
| 1550 |
+
"step": 5140
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 9.14,
|
| 1554 |
+
"learning_rate": 1.5103768676402885e-05,
|
| 1555 |
+
"loss": 0.7314,
|
| 1556 |
+
"step": 5160
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 9.17,
|
| 1560 |
+
"learning_rate": 1.38785220555254e-05,
|
| 1561 |
+
"loss": 0.74,
|
| 1562 |
+
"step": 5180
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 9.21,
|
| 1566 |
+
"learning_rate": 1.270422075162645e-05,
|
| 1567 |
+
"loss": 0.7481,
|
| 1568 |
+
"step": 5200
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 9.24,
|
| 1572 |
+
"learning_rate": 1.1581019703386143e-05,
|
| 1573 |
+
"loss": 0.7222,
|
| 1574 |
+
"step": 5220
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 9.28,
|
| 1578 |
+
"learning_rate": 1.0509067107257365e-05,
|
| 1579 |
+
"loss": 0.7259,
|
| 1580 |
+
"step": 5240
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 9.31,
|
| 1584 |
+
"learning_rate": 9.488504397912712e-06,
|
| 1585 |
+
"loss": 0.7309,
|
| 1586 |
+
"step": 5260
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 9.35,
|
| 1590 |
+
"learning_rate": 8.51946622958324e-06,
|
| 1591 |
+
"loss": 0.7212,
|
| 1592 |
+
"step": 5280
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 9.38,
|
| 1596 |
+
"learning_rate": 7.602080458292227e-06,
|
| 1597 |
+
"loss": 0.7254,
|
| 1598 |
+
"step": 5300
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 9.42,
|
| 1602 |
+
"learning_rate": 6.7364681249854735e-06,
|
| 1603 |
+
"loss": 0.728,
|
| 1604 |
+
"step": 5320
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 9.46,
|
| 1608 |
+
"learning_rate": 5.922743439561229e-06,
|
| 1609 |
+
"loss": 0.7341,
|
| 1610 |
+
"step": 5340
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 9.49,
|
| 1614 |
+
"learning_rate": 5.161013765801137e-06,
|
| 1615 |
+
"loss": 0.721,
|
| 1616 |
+
"step": 5360
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 9.53,
|
| 1620 |
+
"learning_rate": 4.451379607204453e-06,
|
| 1621 |
+
"loss": 0.725,
|
| 1622 |
+
"step": 5380
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 9.56,
|
| 1626 |
+
"learning_rate": 3.7939345937275884e-06,
|
| 1627 |
+
"loss": 0.7399,
|
| 1628 |
+
"step": 5400
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 9.6,
|
| 1632 |
+
"learning_rate": 3.1887654694303883e-06,
|
| 1633 |
+
"loss": 0.7301,
|
| 1634 |
+
"step": 5420
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 9.63,
|
| 1638 |
+
"learning_rate": 2.635952081031201e-06,
|
| 1639 |
+
"loss": 0.7222,
|
| 1640 |
+
"step": 5440
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 9.67,
|
| 1644 |
+
"learning_rate": 2.1355673673715716e-06,
|
| 1645 |
+
"loss": 0.7308,
|
| 1646 |
+
"step": 5460
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 9.7,
|
| 1650 |
+
"learning_rate": 1.6876773497926046e-06,
|
| 1651 |
+
"loss": 0.7417,
|
| 1652 |
+
"step": 5480
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 9.74,
|
| 1656 |
+
"learning_rate": 1.292341123424423e-06,
|
| 1657 |
+
"loss": 0.7348,
|
| 1658 |
+
"step": 5500
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 9.77,
|
| 1662 |
+
"learning_rate": 9.496108493884936e-07,
|
| 1663 |
+
"loss": 0.7278,
|
| 1664 |
+
"step": 5520
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 9.81,
|
| 1668 |
+
"learning_rate": 6.595317479159313e-07,
|
| 1669 |
+
"loss": 0.7257,
|
| 1670 |
+
"step": 5540
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 9.85,
|
| 1674 |
+
"learning_rate": 4.2214209238085054e-07,
|
| 1675 |
+
"loss": 0.7296,
|
| 1676 |
+
"step": 5560
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 9.88,
|
| 1680 |
+
"learning_rate": 2.3747320425053786e-07,
|
| 1681 |
+
"loss": 0.7301,
|
| 1682 |
+
"step": 5580
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 9.92,
|
| 1686 |
+
"learning_rate": 1.0554944895293517e-07,
|
| 1687 |
+
"loss": 0.739,
|
| 1688 |
+
"step": 5600
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 9.95,
|
| 1692 |
+
"learning_rate": 2.638823266174484e-08,
|
| 1693 |
+
"loss": 0.7284,
|
| 1694 |
+
"step": 5620
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 9.99,
|
| 1698 |
+
"learning_rate": 0.0,
|
| 1699 |
+
"loss": 0.7294,
|
| 1700 |
+
"step": 5640
|
| 1701 |
+
},
|
| 1702 |
+
{
|
| 1703 |
+
"epoch": 9.99,
|
| 1704 |
+
"step": 5640,
|
| 1705 |
+
"total_flos": 1.4686171215861645e+19,
|
| 1706 |
+
"train_loss": 0.8784972819876163,
|
| 1707 |
+
"train_runtime": 13393.5014,
|
| 1708 |
+
"train_samples_per_second": 53.949,
|
| 1709 |
+
"train_steps_per_second": 0.421
|
| 1710 |
+
}
|
| 1711 |
+
],
|
| 1712 |
+
"max_steps": 5640,
|
| 1713 |
+
"num_train_epochs": 10,
|
| 1714 |
+
"total_flos": 1.4686171215861645e+19,
|
| 1715 |
+
"trial_name": null,
|
| 1716 |
+
"trial_params": null
|
| 1717 |
+
}
|