Model save
Browse files- README.md +58 -0
- all_results.json +8 -0
- generation_config.json +14 -0
- train_results.json +8 -0
- trainer_state.json +3038 -0
README.md
ADDED
|
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: Qwen/Qwen2.5-0.5B-Instruct
|
| 3 |
+
library_name: transformers
|
| 4 |
+
model_name: Qwen2.5-0.5B-Open-R1-Distill
|
| 5 |
+
tags:
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
- trl
|
| 8 |
+
- sft
|
| 9 |
+
licence: license
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# Model Card for Qwen2.5-0.5B-Open-R1-Distill
|
| 13 |
+
|
| 14 |
+
This model is a fine-tuned version of [Qwen/Qwen2.5-0.5B-Instruct](https://huggingface.co/Qwen/Qwen2.5-0.5B-Instruct).
|
| 15 |
+
It has been trained using [TRL](https://github.com/huggingface/trl).
|
| 16 |
+
|
| 17 |
+
## Quick start
|
| 18 |
+
|
| 19 |
+
```python
|
| 20 |
+
from transformers import pipeline
|
| 21 |
+
|
| 22 |
+
question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
|
| 23 |
+
generator = pipeline("text-generation", model="rhinocamp/Qwen2.5-0.5B-Open-R1-Distill", device="cuda")
|
| 24 |
+
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
|
| 25 |
+
print(output["generated_text"])
|
| 26 |
+
```
|
| 27 |
+
|
| 28 |
+
## Training procedure
|
| 29 |
+
|
| 30 |
+
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/ryan-guo99-meta/huggingface/runs/d1zm9cg7)
|
| 31 |
+
|
| 32 |
+
|
| 33 |
+
This model was trained with SFT.
|
| 34 |
+
|
| 35 |
+
### Framework versions
|
| 36 |
+
|
| 37 |
+
- TRL: 0.16.0.dev0
|
| 38 |
+
- Transformers: 4.49.0
|
| 39 |
+
- Pytorch: 2.7.0.dev20250310+cu124
|
| 40 |
+
- Datasets: 3.4.1
|
| 41 |
+
- Tokenizers: 0.21.1
|
| 42 |
+
|
| 43 |
+
## Citations
|
| 44 |
+
|
| 45 |
+
|
| 46 |
+
|
| 47 |
+
Cite TRL as:
|
| 48 |
+
|
| 49 |
+
```bibtex
|
| 50 |
+
@misc{vonwerra2022trl,
|
| 51 |
+
title = {{TRL: Transformer Reinforcement Learning}},
|
| 52 |
+
author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallouédec},
|
| 53 |
+
year = 2020,
|
| 54 |
+
journal = {GitHub repository},
|
| 55 |
+
publisher = {GitHub},
|
| 56 |
+
howpublished = {\url{https://github.com/huggingface/trl}}
|
| 57 |
+
}
|
| 58 |
+
```
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"total_flos": 241268013662208.0,
|
| 3 |
+
"train_loss": 0.6583015036894314,
|
| 4 |
+
"train_runtime": 13168.5767,
|
| 5 |
+
"train_samples": 93733,
|
| 6 |
+
"train_samples_per_second": 2.605,
|
| 7 |
+
"train_steps_per_second": 0.163
|
| 8 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token_id": 151643,
|
| 3 |
+
"do_sample": true,
|
| 4 |
+
"eos_token_id": [
|
| 5 |
+
151645,
|
| 6 |
+
151643
|
| 7 |
+
],
|
| 8 |
+
"pad_token_id": 151643,
|
| 9 |
+
"repetition_penalty": 1.1,
|
| 10 |
+
"temperature": 0.7,
|
| 11 |
+
"top_k": 20,
|
| 12 |
+
"top_p": 0.8,
|
| 13 |
+
"transformers_version": "4.49.0"
|
| 14 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"total_flos": 241268013662208.0,
|
| 3 |
+
"train_loss": 0.6583015036894314,
|
| 4 |
+
"train_runtime": 13168.5767,
|
| 5 |
+
"train_samples": 93733,
|
| 6 |
+
"train_samples_per_second": 2.605,
|
| 7 |
+
"train_steps_per_second": 0.163
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3038 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 2144,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0023320895522388058,
|
| 13 |
+
"grad_norm": 3.7231960217328877,
|
| 14 |
+
"learning_rate": 2.3148148148148148e-06,
|
| 15 |
+
"loss": 1.0047,
|
| 16 |
+
"step": 5
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.0046641791044776115,
|
| 20 |
+
"grad_norm": 2.8583368538866347,
|
| 21 |
+
"learning_rate": 4.6296296296296296e-06,
|
| 22 |
+
"loss": 0.9757,
|
| 23 |
+
"step": 10
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.006996268656716418,
|
| 27 |
+
"grad_norm": 2.8571738838176763,
|
| 28 |
+
"learning_rate": 6.944444444444445e-06,
|
| 29 |
+
"loss": 0.9678,
|
| 30 |
+
"step": 15
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.009328358208955223,
|
| 34 |
+
"grad_norm": 1.7369287409068157,
|
| 35 |
+
"learning_rate": 9.259259259259259e-06,
|
| 36 |
+
"loss": 0.9001,
|
| 37 |
+
"step": 20
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.01166044776119403,
|
| 41 |
+
"grad_norm": 1.1660883689238433,
|
| 42 |
+
"learning_rate": 1.1574074074074075e-05,
|
| 43 |
+
"loss": 0.8759,
|
| 44 |
+
"step": 25
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.013992537313432836,
|
| 48 |
+
"grad_norm": 1.1321862136040273,
|
| 49 |
+
"learning_rate": 1.388888888888889e-05,
|
| 50 |
+
"loss": 0.8381,
|
| 51 |
+
"step": 30
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.01632462686567164,
|
| 55 |
+
"grad_norm": 1.024631758461693,
|
| 56 |
+
"learning_rate": 1.6203703703703704e-05,
|
| 57 |
+
"loss": 0.834,
|
| 58 |
+
"step": 35
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.018656716417910446,
|
| 62 |
+
"grad_norm": 0.9053510661908479,
|
| 63 |
+
"learning_rate": 1.8518518518518518e-05,
|
| 64 |
+
"loss": 0.8172,
|
| 65 |
+
"step": 40
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.020988805970149255,
|
| 69 |
+
"grad_norm": 1.0170230987058317,
|
| 70 |
+
"learning_rate": 2.0833333333333336e-05,
|
| 71 |
+
"loss": 0.8146,
|
| 72 |
+
"step": 45
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.02332089552238806,
|
| 76 |
+
"grad_norm": 0.9764899017407078,
|
| 77 |
+
"learning_rate": 2.314814814814815e-05,
|
| 78 |
+
"loss": 0.7891,
|
| 79 |
+
"step": 50
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.025652985074626867,
|
| 83 |
+
"grad_norm": 1.0251102524307585,
|
| 84 |
+
"learning_rate": 2.5462962962962965e-05,
|
| 85 |
+
"loss": 0.7814,
|
| 86 |
+
"step": 55
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.027985074626865673,
|
| 90 |
+
"grad_norm": 1.1022491122388751,
|
| 91 |
+
"learning_rate": 2.777777777777778e-05,
|
| 92 |
+
"loss": 0.7872,
|
| 93 |
+
"step": 60
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.03031716417910448,
|
| 97 |
+
"grad_norm": 1.0992277107302175,
|
| 98 |
+
"learning_rate": 3.0092592592592593e-05,
|
| 99 |
+
"loss": 0.7726,
|
| 100 |
+
"step": 65
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.03264925373134328,
|
| 104 |
+
"grad_norm": 1.0365397617690528,
|
| 105 |
+
"learning_rate": 3.240740740740741e-05,
|
| 106 |
+
"loss": 0.7455,
|
| 107 |
+
"step": 70
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.034981343283582086,
|
| 111 |
+
"grad_norm": 1.195178900196762,
|
| 112 |
+
"learning_rate": 3.472222222222222e-05,
|
| 113 |
+
"loss": 0.756,
|
| 114 |
+
"step": 75
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.03731343283582089,
|
| 118 |
+
"grad_norm": 1.1412528086977047,
|
| 119 |
+
"learning_rate": 3.7037037037037037e-05,
|
| 120 |
+
"loss": 0.7633,
|
| 121 |
+
"step": 80
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.039645522388059705,
|
| 125 |
+
"grad_norm": 1.2472194183265157,
|
| 126 |
+
"learning_rate": 3.935185185185186e-05,
|
| 127 |
+
"loss": 0.7557,
|
| 128 |
+
"step": 85
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.04197761194029851,
|
| 132 |
+
"grad_norm": 1.0570653948605382,
|
| 133 |
+
"learning_rate": 4.166666666666667e-05,
|
| 134 |
+
"loss": 0.7601,
|
| 135 |
+
"step": 90
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.044309701492537316,
|
| 139 |
+
"grad_norm": 1.1587882371432328,
|
| 140 |
+
"learning_rate": 4.3981481481481486e-05,
|
| 141 |
+
"loss": 0.7417,
|
| 142 |
+
"step": 95
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.04664179104477612,
|
| 146 |
+
"grad_norm": 1.3162397865412059,
|
| 147 |
+
"learning_rate": 4.62962962962963e-05,
|
| 148 |
+
"loss": 0.739,
|
| 149 |
+
"step": 100
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.04897388059701493,
|
| 153 |
+
"grad_norm": 0.9985279247385236,
|
| 154 |
+
"learning_rate": 4.8611111111111115e-05,
|
| 155 |
+
"loss": 0.7687,
|
| 156 |
+
"step": 105
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.051305970149253734,
|
| 160 |
+
"grad_norm": 1.2887741163380346,
|
| 161 |
+
"learning_rate": 4.999989285883431e-05,
|
| 162 |
+
"loss": 0.7664,
|
| 163 |
+
"step": 110
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.05363805970149254,
|
| 167 |
+
"grad_norm": 1.2392518485033788,
|
| 168 |
+
"learning_rate": 4.999868753243865e-05,
|
| 169 |
+
"loss": 0.7675,
|
| 170 |
+
"step": 115
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.055970149253731345,
|
| 174 |
+
"grad_norm": 1.8916579665083808,
|
| 175 |
+
"learning_rate": 4.999614302517356e-05,
|
| 176 |
+
"loss": 0.7598,
|
| 177 |
+
"step": 120
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.05830223880597015,
|
| 181 |
+
"grad_norm": 1.5520377612517253,
|
| 182 |
+
"learning_rate": 4.9992259488494784e-05,
|
| 183 |
+
"loss": 0.7424,
|
| 184 |
+
"step": 125
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.06063432835820896,
|
| 188 |
+
"grad_norm": 1.239718216272371,
|
| 189 |
+
"learning_rate": 4.99870371535606e-05,
|
| 190 |
+
"loss": 0.7366,
|
| 191 |
+
"step": 130
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.06296641791044776,
|
| 195 |
+
"grad_norm": 0.9825341780149127,
|
| 196 |
+
"learning_rate": 4.9980476331218066e-05,
|
| 197 |
+
"loss": 0.786,
|
| 198 |
+
"step": 135
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.06529850746268656,
|
| 202 |
+
"grad_norm": 0.9155268841122092,
|
| 203 |
+
"learning_rate": 4.997257741198456e-05,
|
| 204 |
+
"loss": 0.7505,
|
| 205 |
+
"step": 140
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.06763059701492537,
|
| 209 |
+
"grad_norm": 1.082193107284934,
|
| 210 |
+
"learning_rate": 4.996334086602444e-05,
|
| 211 |
+
"loss": 0.7351,
|
| 212 |
+
"step": 145
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.06996268656716417,
|
| 216 |
+
"grad_norm": 1.0423700442715462,
|
| 217 |
+
"learning_rate": 4.9952767243121146e-05,
|
| 218 |
+
"loss": 0.7551,
|
| 219 |
+
"step": 150
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.07229477611940298,
|
| 223 |
+
"grad_norm": 0.9000144515873072,
|
| 224 |
+
"learning_rate": 4.994085717264443e-05,
|
| 225 |
+
"loss": 0.7368,
|
| 226 |
+
"step": 155
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.07462686567164178,
|
| 230 |
+
"grad_norm": 1.044303554591329,
|
| 231 |
+
"learning_rate": 4.992761136351291e-05,
|
| 232 |
+
"loss": 0.7436,
|
| 233 |
+
"step": 160
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.07695895522388059,
|
| 237 |
+
"grad_norm": 1.0470880900421262,
|
| 238 |
+
"learning_rate": 4.99130306041519e-05,
|
| 239 |
+
"loss": 0.7446,
|
| 240 |
+
"step": 165
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.07929104477611941,
|
| 244 |
+
"grad_norm": 1.1323111761895537,
|
| 245 |
+
"learning_rate": 4.989711576244639e-05,
|
| 246 |
+
"loss": 0.7226,
|
| 247 |
+
"step": 170
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.08162313432835822,
|
| 251 |
+
"grad_norm": 1.0668807832388134,
|
| 252 |
+
"learning_rate": 4.9879867785689506e-05,
|
| 253 |
+
"loss": 0.7107,
|
| 254 |
+
"step": 175
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.08395522388059702,
|
| 258 |
+
"grad_norm": 0.9414033524848218,
|
| 259 |
+
"learning_rate": 4.986128770052603e-05,
|
| 260 |
+
"loss": 0.7315,
|
| 261 |
+
"step": 180
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.08628731343283583,
|
| 265 |
+
"grad_norm": 0.9861314941949435,
|
| 266 |
+
"learning_rate": 4.9841376612891365e-05,
|
| 267 |
+
"loss": 0.7226,
|
| 268 |
+
"step": 185
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.08861940298507463,
|
| 272 |
+
"grad_norm": 0.9154465133645083,
|
| 273 |
+
"learning_rate": 4.9820135707945634e-05,
|
| 274 |
+
"loss": 0.719,
|
| 275 |
+
"step": 190
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.09095149253731344,
|
| 279 |
+
"grad_norm": 1.0049244209561472,
|
| 280 |
+
"learning_rate": 4.97975662500032e-05,
|
| 281 |
+
"loss": 0.7474,
|
| 282 |
+
"step": 195
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.09328358208955224,
|
| 286 |
+
"grad_norm": 0.9649064226232575,
|
| 287 |
+
"learning_rate": 4.9773669582457364e-05,
|
| 288 |
+
"loss": 0.7273,
|
| 289 |
+
"step": 200
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.09561567164179105,
|
| 293 |
+
"grad_norm": 1.0099484363867934,
|
| 294 |
+
"learning_rate": 4.974844712770044e-05,
|
| 295 |
+
"loss": 0.7456,
|
| 296 |
+
"step": 205
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.09794776119402986,
|
| 300 |
+
"grad_norm": 0.9092588577896343,
|
| 301 |
+
"learning_rate": 4.972190038703905e-05,
|
| 302 |
+
"loss": 0.7389,
|
| 303 |
+
"step": 210
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.10027985074626866,
|
| 307 |
+
"grad_norm": 1.0001019634384871,
|
| 308 |
+
"learning_rate": 4.96940309406048e-05,
|
| 309 |
+
"loss": 0.7334,
|
| 310 |
+
"step": 215
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.10261194029850747,
|
| 314 |
+
"grad_norm": 1.0388048983714429,
|
| 315 |
+
"learning_rate": 4.966484044726024e-05,
|
| 316 |
+
"loss": 0.7246,
|
| 317 |
+
"step": 220
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.10494402985074627,
|
| 321 |
+
"grad_norm": 1.0827378138832435,
|
| 322 |
+
"learning_rate": 4.963433064450002e-05,
|
| 323 |
+
"loss": 0.7208,
|
| 324 |
+
"step": 225
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.10727611940298508,
|
| 328 |
+
"grad_norm": 1.0427299412602653,
|
| 329 |
+
"learning_rate": 4.9602503348347625e-05,
|
| 330 |
+
"loss": 0.7133,
|
| 331 |
+
"step": 230
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.10960820895522388,
|
| 335 |
+
"grad_norm": 0.9388505104499245,
|
| 336 |
+
"learning_rate": 4.956936045324716e-05,
|
| 337 |
+
"loss": 0.6982,
|
| 338 |
+
"step": 235
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.11194029850746269,
|
| 342 |
+
"grad_norm": 0.8567679904473422,
|
| 343 |
+
"learning_rate": 4.953490393195063e-05,
|
| 344 |
+
"loss": 0.7518,
|
| 345 |
+
"step": 240
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.1142723880597015,
|
| 349 |
+
"grad_norm": 0.9211748764639448,
|
| 350 |
+
"learning_rate": 4.9499135835400526e-05,
|
| 351 |
+
"loss": 0.7287,
|
| 352 |
+
"step": 245
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.1166044776119403,
|
| 356 |
+
"grad_norm": 1.0303852066380612,
|
| 357 |
+
"learning_rate": 4.9462058292607735e-05,
|
| 358 |
+
"loss": 0.7134,
|
| 359 |
+
"step": 250
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.11893656716417911,
|
| 363 |
+
"grad_norm": 1.2197325105359402,
|
| 364 |
+
"learning_rate": 4.942367351052481e-05,
|
| 365 |
+
"loss": 0.7086,
|
| 366 |
+
"step": 255
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.12126865671641791,
|
| 370 |
+
"grad_norm": 0.8299444059232342,
|
| 371 |
+
"learning_rate": 4.938398377391461e-05,
|
| 372 |
+
"loss": 0.7187,
|
| 373 |
+
"step": 260
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.12360074626865672,
|
| 377 |
+
"grad_norm": 0.9224154095712178,
|
| 378 |
+
"learning_rate": 4.9342991445214334e-05,
|
| 379 |
+
"loss": 0.7126,
|
| 380 |
+
"step": 265
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.1259328358208955,
|
| 384 |
+
"grad_norm": 0.88090785003352,
|
| 385 |
+
"learning_rate": 4.930069896439485e-05,
|
| 386 |
+
"loss": 0.7105,
|
| 387 |
+
"step": 270
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.12826492537313433,
|
| 391 |
+
"grad_norm": 0.859476692536741,
|
| 392 |
+
"learning_rate": 4.925710884881551e-05,
|
| 393 |
+
"loss": 0.7071,
|
| 394 |
+
"step": 275
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.13059701492537312,
|
| 398 |
+
"grad_norm": 1.013753901988807,
|
| 399 |
+
"learning_rate": 4.921222369307427e-05,
|
| 400 |
+
"loss": 0.7205,
|
| 401 |
+
"step": 280
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.13292910447761194,
|
| 405 |
+
"grad_norm": 0.940182496551757,
|
| 406 |
+
"learning_rate": 4.916604616885328e-05,
|
| 407 |
+
"loss": 0.7139,
|
| 408 |
+
"step": 285
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.13526119402985073,
|
| 412 |
+
"grad_norm": 1.0259898104670397,
|
| 413 |
+
"learning_rate": 4.9118579024759854e-05,
|
| 414 |
+
"loss": 0.7194,
|
| 415 |
+
"step": 290
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.13759328358208955,
|
| 419 |
+
"grad_norm": 0.941096247874784,
|
| 420 |
+
"learning_rate": 4.9069825086162865e-05,
|
| 421 |
+
"loss": 0.694,
|
| 422 |
+
"step": 295
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.13992537313432835,
|
| 426 |
+
"grad_norm": 0.8605571897927781,
|
| 427 |
+
"learning_rate": 4.901978725502454e-05,
|
| 428 |
+
"loss": 0.7021,
|
| 429 |
+
"step": 300
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.14225746268656717,
|
| 433 |
+
"grad_norm": 1.012855599597888,
|
| 434 |
+
"learning_rate": 4.89684685097278e-05,
|
| 435 |
+
"loss": 0.6889,
|
| 436 |
+
"step": 305
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.14458955223880596,
|
| 440 |
+
"grad_norm": 0.9877119462322568,
|
| 441 |
+
"learning_rate": 4.891587190489891e-05,
|
| 442 |
+
"loss": 0.7154,
|
| 443 |
+
"step": 310
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.14692164179104478,
|
| 447 |
+
"grad_norm": 1.090755921619965,
|
| 448 |
+
"learning_rate": 4.886200057122568e-05,
|
| 449 |
+
"loss": 0.7195,
|
| 450 |
+
"step": 315
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.14925373134328357,
|
| 454 |
+
"grad_norm": 0.9811029143694675,
|
| 455 |
+
"learning_rate": 4.880685771527114e-05,
|
| 456 |
+
"loss": 0.7055,
|
| 457 |
+
"step": 320
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.1515858208955224,
|
| 461 |
+
"grad_norm": 0.8841142544985066,
|
| 462 |
+
"learning_rate": 4.8750446619282646e-05,
|
| 463 |
+
"loss": 0.7029,
|
| 464 |
+
"step": 325
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.15391791044776118,
|
| 468 |
+
"grad_norm": 0.8332150176518437,
|
| 469 |
+
"learning_rate": 4.869277064099654e-05,
|
| 470 |
+
"loss": 0.7093,
|
| 471 |
+
"step": 330
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.15625,
|
| 475 |
+
"grad_norm": 0.8744887046543461,
|
| 476 |
+
"learning_rate": 4.8633833213438254e-05,
|
| 477 |
+
"loss": 0.7065,
|
| 478 |
+
"step": 335
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.15858208955223882,
|
| 482 |
+
"grad_norm": 0.837397759622329,
|
| 483 |
+
"learning_rate": 4.8573637844718e-05,
|
| 484 |
+
"loss": 0.7325,
|
| 485 |
+
"step": 340
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.1609141791044776,
|
| 489 |
+
"grad_norm": 0.9698882033094054,
|
| 490 |
+
"learning_rate": 4.851218811782195e-05,
|
| 491 |
+
"loss": 0.7209,
|
| 492 |
+
"step": 345
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.16324626865671643,
|
| 496 |
+
"grad_norm": 0.8572480215660884,
|
| 497 |
+
"learning_rate": 4.844948769039896e-05,
|
| 498 |
+
"loss": 0.7117,
|
| 499 |
+
"step": 350
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.16557835820895522,
|
| 503 |
+
"grad_norm": 0.8647940595338452,
|
| 504 |
+
"learning_rate": 4.8385540294542855e-05,
|
| 505 |
+
"loss": 0.7059,
|
| 506 |
+
"step": 355
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.16791044776119404,
|
| 510 |
+
"grad_norm": 0.863764155230387,
|
| 511 |
+
"learning_rate": 4.83203497365703e-05,
|
| 512 |
+
"loss": 0.7114,
|
| 513 |
+
"step": 360
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.17024253731343283,
|
| 517 |
+
"grad_norm": 0.8063210763659003,
|
| 518 |
+
"learning_rate": 4.825391989679422e-05,
|
| 519 |
+
"loss": 0.7034,
|
| 520 |
+
"step": 365
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.17257462686567165,
|
| 524 |
+
"grad_norm": 0.8575257639465484,
|
| 525 |
+
"learning_rate": 4.818625472929286e-05,
|
| 526 |
+
"loss": 0.6947,
|
| 527 |
+
"step": 370
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.17490671641791045,
|
| 531 |
+
"grad_norm": 0.8200838613816108,
|
| 532 |
+
"learning_rate": 4.811735826167436e-05,
|
| 533 |
+
"loss": 0.681,
|
| 534 |
+
"step": 375
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.17723880597014927,
|
| 538 |
+
"grad_norm": 0.7823498981274288,
|
| 539 |
+
"learning_rate": 4.8047234594837143e-05,
|
| 540 |
+
"loss": 0.6958,
|
| 541 |
+
"step": 380
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.17957089552238806,
|
| 545 |
+
"grad_norm": 0.8190018216789188,
|
| 546 |
+
"learning_rate": 4.7975887902725696e-05,
|
| 547 |
+
"loss": 0.6835,
|
| 548 |
+
"step": 385
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.18190298507462688,
|
| 552 |
+
"grad_norm": 0.9214403517298152,
|
| 553 |
+
"learning_rate": 4.7903322432082185e-05,
|
| 554 |
+
"loss": 0.6984,
|
| 555 |
+
"step": 390
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.18423507462686567,
|
| 559 |
+
"grad_norm": 0.8273600510500148,
|
| 560 |
+
"learning_rate": 4.78295425021937e-05,
|
| 561 |
+
"loss": 0.6753,
|
| 562 |
+
"step": 395
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.1865671641791045,
|
| 566 |
+
"grad_norm": 0.9257746216157634,
|
| 567 |
+
"learning_rate": 4.775455250463507e-05,
|
| 568 |
+
"loss": 0.7063,
|
| 569 |
+
"step": 400
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.18889925373134328,
|
| 573 |
+
"grad_norm": 0.8553399615781841,
|
| 574 |
+
"learning_rate": 4.767835690300759e-05,
|
| 575 |
+
"loss": 0.7031,
|
| 576 |
+
"step": 405
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.1912313432835821,
|
| 580 |
+
"grad_norm": 0.8035273180959155,
|
| 581 |
+
"learning_rate": 4.760096023267322e-05,
|
| 582 |
+
"loss": 0.7098,
|
| 583 |
+
"step": 410
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.1935634328358209,
|
| 587 |
+
"grad_norm": 0.8691622488936168,
|
| 588 |
+
"learning_rate": 4.752236710048472e-05,
|
| 589 |
+
"loss": 0.6861,
|
| 590 |
+
"step": 415
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.1958955223880597,
|
| 594 |
+
"grad_norm": 0.9841920794382144,
|
| 595 |
+
"learning_rate": 4.744258218451135e-05,
|
| 596 |
+
"loss": 0.7031,
|
| 597 |
+
"step": 420
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.1982276119402985,
|
| 601 |
+
"grad_norm": 0.9498479171681536,
|
| 602 |
+
"learning_rate": 4.736161023376051e-05,
|
| 603 |
+
"loss": 0.6726,
|
| 604 |
+
"step": 425
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.20055970149253732,
|
| 608 |
+
"grad_norm": 0.8700794084478172,
|
| 609 |
+
"learning_rate": 4.7279456067895e-05,
|
| 610 |
+
"loss": 0.7153,
|
| 611 |
+
"step": 430
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.20289179104477612,
|
| 615 |
+
"grad_norm": 0.8127172667880727,
|
| 616 |
+
"learning_rate": 4.719612457694616e-05,
|
| 617 |
+
"loss": 0.6963,
|
| 618 |
+
"step": 435
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.20522388059701493,
|
| 622 |
+
"grad_norm": 0.7902973875049144,
|
| 623 |
+
"learning_rate": 4.71116207210228e-05,
|
| 624 |
+
"loss": 0.6773,
|
| 625 |
+
"step": 440
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.20755597014925373,
|
| 629 |
+
"grad_norm": 0.7788306399019677,
|
| 630 |
+
"learning_rate": 4.7025949530016e-05,
|
| 631 |
+
"loss": 0.6684,
|
| 632 |
+
"step": 445
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.20988805970149255,
|
| 636 |
+
"grad_norm": 0.7944637728792737,
|
| 637 |
+
"learning_rate": 4.6939116103299655e-05,
|
| 638 |
+
"loss": 0.6905,
|
| 639 |
+
"step": 450
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.21222014925373134,
|
| 643 |
+
"grad_norm": 0.8571602340727155,
|
| 644 |
+
"learning_rate": 4.685112560942699e-05,
|
| 645 |
+
"loss": 0.6881,
|
| 646 |
+
"step": 455
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.21455223880597016,
|
| 650 |
+
"grad_norm": 0.8780895287743348,
|
| 651 |
+
"learning_rate": 4.676198328582288e-05,
|
| 652 |
+
"loss": 0.6927,
|
| 653 |
+
"step": 460
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.21688432835820895,
|
| 657 |
+
"grad_norm": 0.9388682206384781,
|
| 658 |
+
"learning_rate": 4.6671694438472154e-05,
|
| 659 |
+
"loss": 0.7056,
|
| 660 |
+
"step": 465
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.21921641791044777,
|
| 664 |
+
"grad_norm": 0.8229322015185289,
|
| 665 |
+
"learning_rate": 4.6580264441603724e-05,
|
| 666 |
+
"loss": 0.6894,
|
| 667 |
+
"step": 470
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.22154850746268656,
|
| 671 |
+
"grad_norm": 0.9117054400216464,
|
| 672 |
+
"learning_rate": 4.648769873737071e-05,
|
| 673 |
+
"loss": 0.6972,
|
| 674 |
+
"step": 475
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.22388059701492538,
|
| 678 |
+
"grad_norm": 0.7774500802624019,
|
| 679 |
+
"learning_rate": 4.6394002835526535e-05,
|
| 680 |
+
"loss": 0.6969,
|
| 681 |
+
"step": 480
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.22621268656716417,
|
| 685 |
+
"grad_norm": 0.9974671450500393,
|
| 686 |
+
"learning_rate": 4.6299182313096916e-05,
|
| 687 |
+
"loss": 0.7168,
|
| 688 |
+
"step": 485
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.228544776119403,
|
| 692 |
+
"grad_norm": 0.7745214676810972,
|
| 693 |
+
"learning_rate": 4.6203242814047946e-05,
|
| 694 |
+
"loss": 0.6388,
|
| 695 |
+
"step": 490
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.23087686567164178,
|
| 699 |
+
"grad_norm": 0.7946997637352378,
|
| 700 |
+
"learning_rate": 4.610619004895017e-05,
|
| 701 |
+
"loss": 0.6932,
|
| 702 |
+
"step": 495
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.2332089552238806,
|
| 706 |
+
"grad_norm": 0.8376007460886711,
|
| 707 |
+
"learning_rate": 4.6008029794638596e-05,
|
| 708 |
+
"loss": 0.6626,
|
| 709 |
+
"step": 500
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.2355410447761194,
|
| 713 |
+
"grad_norm": 0.7750371283803145,
|
| 714 |
+
"learning_rate": 4.590876789386893e-05,
|
| 715 |
+
"loss": 0.6681,
|
| 716 |
+
"step": 505
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.23787313432835822,
|
| 720 |
+
"grad_norm": 0.8455168484121804,
|
| 721 |
+
"learning_rate": 4.580841025496974e-05,
|
| 722 |
+
"loss": 0.6786,
|
| 723 |
+
"step": 510
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.240205223880597,
|
| 727 |
+
"grad_norm": 0.7240530893309742,
|
| 728 |
+
"learning_rate": 4.570696285149084e-05,
|
| 729 |
+
"loss": 0.6825,
|
| 730 |
+
"step": 515
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.24253731343283583,
|
| 734 |
+
"grad_norm": 0.7536963649856001,
|
| 735 |
+
"learning_rate": 4.560443172184763e-05,
|
| 736 |
+
"loss": 0.6856,
|
| 737 |
+
"step": 520
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.24486940298507462,
|
| 741 |
+
"grad_norm": 0.7793205383613011,
|
| 742 |
+
"learning_rate": 4.55008229689618e-05,
|
| 743 |
+
"loss": 0.6667,
|
| 744 |
+
"step": 525
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.24720149253731344,
|
| 748 |
+
"grad_norm": 0.7157463633792304,
|
| 749 |
+
"learning_rate": 4.539614275989793e-05,
|
| 750 |
+
"loss": 0.6754,
|
| 751 |
+
"step": 530
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.24953358208955223,
|
| 755 |
+
"grad_norm": 0.7890577436144398,
|
| 756 |
+
"learning_rate": 4.529039732549653e-05,
|
| 757 |
+
"loss": 0.675,
|
| 758 |
+
"step": 535
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.251865671641791,
|
| 762 |
+
"grad_norm": 0.7846004685785093,
|
| 763 |
+
"learning_rate": 4.5183592960003104e-05,
|
| 764 |
+
"loss": 0.6635,
|
| 765 |
+
"step": 540
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.25419776119402987,
|
| 769 |
+
"grad_norm": 0.8071994271154869,
|
| 770 |
+
"learning_rate": 4.507573602069351e-05,
|
| 771 |
+
"loss": 0.6773,
|
| 772 |
+
"step": 545
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.25652985074626866,
|
| 776 |
+
"grad_norm": 0.8233537619340792,
|
| 777 |
+
"learning_rate": 4.496683292749555e-05,
|
| 778 |
+
"loss": 0.6719,
|
| 779 |
+
"step": 550
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.25886194029850745,
|
| 783 |
+
"grad_norm": 0.7715749039189433,
|
| 784 |
+
"learning_rate": 4.485689016260686e-05,
|
| 785 |
+
"loss": 0.6643,
|
| 786 |
+
"step": 555
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.26119402985074625,
|
| 790 |
+
"grad_norm": 0.7396872015907794,
|
| 791 |
+
"learning_rate": 4.4745914270109055e-05,
|
| 792 |
+
"loss": 0.6943,
|
| 793 |
+
"step": 560
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.2635261194029851,
|
| 797 |
+
"grad_norm": 0.7904102147449842,
|
| 798 |
+
"learning_rate": 4.463391185557822e-05,
|
| 799 |
+
"loss": 0.6819,
|
| 800 |
+
"step": 565
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.2658582089552239,
|
| 804 |
+
"grad_norm": 0.7165693906401905,
|
| 805 |
+
"learning_rate": 4.4520889585691705e-05,
|
| 806 |
+
"loss": 0.6865,
|
| 807 |
+
"step": 570
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.2681902985074627,
|
| 811 |
+
"grad_norm": 0.8804513560159245,
|
| 812 |
+
"learning_rate": 4.440685418783135e-05,
|
| 813 |
+
"loss": 0.6631,
|
| 814 |
+
"step": 575
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.27052238805970147,
|
| 818 |
+
"grad_norm": 0.7607710965405562,
|
| 819 |
+
"learning_rate": 4.429181244968301e-05,
|
| 820 |
+
"loss": 0.6699,
|
| 821 |
+
"step": 580
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.2728544776119403,
|
| 825 |
+
"grad_norm": 0.7620322976601013,
|
| 826 |
+
"learning_rate": 4.417577121883256e-05,
|
| 827 |
+
"loss": 0.669,
|
| 828 |
+
"step": 585
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.2751865671641791,
|
| 832 |
+
"grad_norm": 0.7328194761484526,
|
| 833 |
+
"learning_rate": 4.4058737402358295e-05,
|
| 834 |
+
"loss": 0.6488,
|
| 835 |
+
"step": 590
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.2775186567164179,
|
| 839 |
+
"grad_norm": 0.7634679687008787,
|
| 840 |
+
"learning_rate": 4.394071796641983e-05,
|
| 841 |
+
"loss": 0.6729,
|
| 842 |
+
"step": 595
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.2798507462686567,
|
| 846 |
+
"grad_norm": 0.8531001469934757,
|
| 847 |
+
"learning_rate": 4.38217199358434e-05,
|
| 848 |
+
"loss": 0.685,
|
| 849 |
+
"step": 600
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.28218283582089554,
|
| 853 |
+
"grad_norm": 0.7264481750331164,
|
| 854 |
+
"learning_rate": 4.3701750393703786e-05,
|
| 855 |
+
"loss": 0.6461,
|
| 856 |
+
"step": 605
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.28451492537313433,
|
| 860 |
+
"grad_norm": 0.7533827654944563,
|
| 861 |
+
"learning_rate": 4.3580816480902656e-05,
|
| 862 |
+
"loss": 0.6745,
|
| 863 |
+
"step": 610
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.2868470149253731,
|
| 867 |
+
"grad_norm": 0.7458225351946263,
|
| 868 |
+
"learning_rate": 4.345892539574359e-05,
|
| 869 |
+
"loss": 0.6831,
|
| 870 |
+
"step": 615
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.2891791044776119,
|
| 874 |
+
"grad_norm": 0.724365300513045,
|
| 875 |
+
"learning_rate": 4.3336084393503545e-05,
|
| 876 |
+
"loss": 0.6546,
|
| 877 |
+
"step": 620
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.29151119402985076,
|
| 881 |
+
"grad_norm": 0.7838535311430127,
|
| 882 |
+
"learning_rate": 4.3212300786001045e-05,
|
| 883 |
+
"loss": 0.6577,
|
| 884 |
+
"step": 625
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.29384328358208955,
|
| 888 |
+
"grad_norm": 0.7715960756766049,
|
| 889 |
+
"learning_rate": 4.308758194116094e-05,
|
| 890 |
+
"loss": 0.684,
|
| 891 |
+
"step": 630
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.29617537313432835,
|
| 895 |
+
"grad_norm": 0.7160624711169831,
|
| 896 |
+
"learning_rate": 4.296193528257586e-05,
|
| 897 |
+
"loss": 0.6893,
|
| 898 |
+
"step": 635
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.29850746268656714,
|
| 902 |
+
"grad_norm": 0.7745707050758446,
|
| 903 |
+
"learning_rate": 4.283536828906436e-05,
|
| 904 |
+
"loss": 0.6789,
|
| 905 |
+
"step": 640
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.300839552238806,
|
| 909 |
+
"grad_norm": 0.8251172241406618,
|
| 910 |
+
"learning_rate": 4.270788849422572e-05,
|
| 911 |
+
"loss": 0.6623,
|
| 912 |
+
"step": 645
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.3031716417910448,
|
| 916 |
+
"grad_norm": 0.8355188301078915,
|
| 917 |
+
"learning_rate": 4.2579503485991567e-05,
|
| 918 |
+
"loss": 0.6724,
|
| 919 |
+
"step": 650
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.30550373134328357,
|
| 923 |
+
"grad_norm": 0.8380841544851293,
|
| 924 |
+
"learning_rate": 4.245022090617418e-05,
|
| 925 |
+
"loss": 0.6765,
|
| 926 |
+
"step": 655
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.30783582089552236,
|
| 930 |
+
"grad_norm": 0.7456377128567805,
|
| 931 |
+
"learning_rate": 4.2320048450011684e-05,
|
| 932 |
+
"loss": 0.6741,
|
| 933 |
+
"step": 660
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.3101679104477612,
|
| 937 |
+
"grad_norm": 0.6884827740191439,
|
| 938 |
+
"learning_rate": 4.218899386570994e-05,
|
| 939 |
+
"loss": 0.6556,
|
| 940 |
+
"step": 665
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.3125,
|
| 944 |
+
"grad_norm": 0.6695176963355354,
|
| 945 |
+
"learning_rate": 4.205706495398143e-05,
|
| 946 |
+
"loss": 0.6824,
|
| 947 |
+
"step": 670
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.3148320895522388,
|
| 951 |
+
"grad_norm": 0.6518330480049027,
|
| 952 |
+
"learning_rate": 4.192426956758085e-05,
|
| 953 |
+
"loss": 0.6452,
|
| 954 |
+
"step": 675
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.31716417910447764,
|
| 958 |
+
"grad_norm": 0.7023106383456971,
|
| 959 |
+
"learning_rate": 4.179061561083777e-05,
|
| 960 |
+
"loss": 0.6573,
|
| 961 |
+
"step": 680
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.31949626865671643,
|
| 965 |
+
"grad_norm": 0.6796008979702891,
|
| 966 |
+
"learning_rate": 4.165611103918612e-05,
|
| 967 |
+
"loss": 0.6813,
|
| 968 |
+
"step": 685
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.3218283582089552,
|
| 972 |
+
"grad_norm": 0.7234212222083802,
|
| 973 |
+
"learning_rate": 4.1520763858690644e-05,
|
| 974 |
+
"loss": 0.6563,
|
| 975 |
+
"step": 690
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.324160447761194,
|
| 979 |
+
"grad_norm": 0.8232766133796315,
|
| 980 |
+
"learning_rate": 4.138458212557038e-05,
|
| 981 |
+
"loss": 0.6789,
|
| 982 |
+
"step": 695
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.32649253731343286,
|
| 986 |
+
"grad_norm": 0.892744295966519,
|
| 987 |
+
"learning_rate": 4.124757394571914e-05,
|
| 988 |
+
"loss": 0.6511,
|
| 989 |
+
"step": 700
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.32882462686567165,
|
| 993 |
+
"grad_norm": 0.7465253057638759,
|
| 994 |
+
"learning_rate": 4.110974747422299e-05,
|
| 995 |
+
"loss": 0.6609,
|
| 996 |
+
"step": 705
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.33115671641791045,
|
| 1000 |
+
"grad_norm": 0.9710864864368132,
|
| 1001 |
+
"learning_rate": 4.097111091487486e-05,
|
| 1002 |
+
"loss": 0.6567,
|
| 1003 |
+
"step": 710
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.33348880597014924,
|
| 1007 |
+
"grad_norm": 0.7586336649039006,
|
| 1008 |
+
"learning_rate": 4.083167251968625e-05,
|
| 1009 |
+
"loss": 0.6868,
|
| 1010 |
+
"step": 715
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.3358208955223881,
|
| 1014 |
+
"grad_norm": 0.7411454559784593,
|
| 1015 |
+
"learning_rate": 4.069144058839605e-05,
|
| 1016 |
+
"loss": 0.6715,
|
| 1017 |
+
"step": 720
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.3381529850746269,
|
| 1021 |
+
"grad_norm": 0.7406787177182912,
|
| 1022 |
+
"learning_rate": 4.055042346797643e-05,
|
| 1023 |
+
"loss": 0.6655,
|
| 1024 |
+
"step": 725
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.34048507462686567,
|
| 1028 |
+
"grad_norm": 0.806810167383283,
|
| 1029 |
+
"learning_rate": 4.040862955213615e-05,
|
| 1030 |
+
"loss": 0.6792,
|
| 1031 |
+
"step": 730
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.34281716417910446,
|
| 1035 |
+
"grad_norm": 0.684339497670028,
|
| 1036 |
+
"learning_rate": 4.026606728082082e-05,
|
| 1037 |
+
"loss": 0.6643,
|
| 1038 |
+
"step": 735
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.3451492537313433,
|
| 1042 |
+
"grad_norm": 0.7645437701144697,
|
| 1043 |
+
"learning_rate": 4.012274513971061e-05,
|
| 1044 |
+
"loss": 0.6839,
|
| 1045 |
+
"step": 740
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.3474813432835821,
|
| 1049 |
+
"grad_norm": 0.6793956256373396,
|
| 1050 |
+
"learning_rate": 3.997867165971512e-05,
|
| 1051 |
+
"loss": 0.6507,
|
| 1052 |
+
"step": 745
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.3498134328358209,
|
| 1056 |
+
"grad_norm": 0.720694420633502,
|
| 1057 |
+
"learning_rate": 3.9833855416465624e-05,
|
| 1058 |
+
"loss": 0.6643,
|
| 1059 |
+
"step": 750
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.3521455223880597,
|
| 1063 |
+
"grad_norm": 0.7523164344936248,
|
| 1064 |
+
"learning_rate": 3.968830502980459e-05,
|
| 1065 |
+
"loss": 0.6591,
|
| 1066 |
+
"step": 755
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.35447761194029853,
|
| 1070 |
+
"grad_norm": 0.7805853823171353,
|
| 1071 |
+
"learning_rate": 3.954202916327264e-05,
|
| 1072 |
+
"loss": 0.6549,
|
| 1073 |
+
"step": 760
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.3568097014925373,
|
| 1077 |
+
"grad_norm": 0.7234820701835654,
|
| 1078 |
+
"learning_rate": 3.939503652359287e-05,
|
| 1079 |
+
"loss": 0.6572,
|
| 1080 |
+
"step": 765
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.3591417910447761,
|
| 1084 |
+
"grad_norm": 0.7688015369932021,
|
| 1085 |
+
"learning_rate": 3.924733586015257e-05,
|
| 1086 |
+
"loss": 0.6533,
|
| 1087 |
+
"step": 770
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.3614738805970149,
|
| 1091 |
+
"grad_norm": 0.708671520124952,
|
| 1092 |
+
"learning_rate": 3.9098935964482476e-05,
|
| 1093 |
+
"loss": 0.6551,
|
| 1094 |
+
"step": 775
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.36380597014925375,
|
| 1098 |
+
"grad_norm": 0.7183227720244021,
|
| 1099 |
+
"learning_rate": 3.894984566973346e-05,
|
| 1100 |
+
"loss": 0.6646,
|
| 1101 |
+
"step": 780
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.36613805970149255,
|
| 1105 |
+
"grad_norm": 0.6737145813609308,
|
| 1106 |
+
"learning_rate": 3.880007385015075e-05,
|
| 1107 |
+
"loss": 0.6511,
|
| 1108 |
+
"step": 785
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.36847014925373134,
|
| 1112 |
+
"grad_norm": 0.6920229984205842,
|
| 1113 |
+
"learning_rate": 3.864962942054572e-05,
|
| 1114 |
+
"loss": 0.6818,
|
| 1115 |
+
"step": 790
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.37080223880597013,
|
| 1119 |
+
"grad_norm": 0.775113876143003,
|
| 1120 |
+
"learning_rate": 3.849852133576527e-05,
|
| 1121 |
+
"loss": 0.6435,
|
| 1122 |
+
"step": 795
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 0.373134328358209,
|
| 1126 |
+
"grad_norm": 0.7394720394307068,
|
| 1127 |
+
"learning_rate": 3.834675859015876e-05,
|
| 1128 |
+
"loss": 0.671,
|
| 1129 |
+
"step": 800
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.37546641791044777,
|
| 1133 |
+
"grad_norm": 0.7324224028801083,
|
| 1134 |
+
"learning_rate": 3.819435021704274e-05,
|
| 1135 |
+
"loss": 0.66,
|
| 1136 |
+
"step": 805
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.37779850746268656,
|
| 1140 |
+
"grad_norm": 0.7482922307683427,
|
| 1141 |
+
"learning_rate": 3.804130528816312e-05,
|
| 1142 |
+
"loss": 0.6605,
|
| 1143 |
+
"step": 810
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.38013059701492535,
|
| 1147 |
+
"grad_norm": 0.693797019256493,
|
| 1148 |
+
"learning_rate": 3.7887632913155355e-05,
|
| 1149 |
+
"loss": 0.6459,
|
| 1150 |
+
"step": 815
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.3824626865671642,
|
| 1154 |
+
"grad_norm": 0.7332686963960172,
|
| 1155 |
+
"learning_rate": 3.77333422390021e-05,
|
| 1156 |
+
"loss": 0.6386,
|
| 1157 |
+
"step": 820
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.384794776119403,
|
| 1161 |
+
"grad_norm": 0.6722013452671626,
|
| 1162 |
+
"learning_rate": 3.75784424494888e-05,
|
| 1163 |
+
"loss": 0.6509,
|
| 1164 |
+
"step": 825
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.3871268656716418,
|
| 1168 |
+
"grad_norm": 0.6844841017486335,
|
| 1169 |
+
"learning_rate": 3.7422942764657054e-05,
|
| 1170 |
+
"loss": 0.6624,
|
| 1171 |
+
"step": 830
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.3894589552238806,
|
| 1175 |
+
"grad_norm": 0.709365449407489,
|
| 1176 |
+
"learning_rate": 3.726685244025578e-05,
|
| 1177 |
+
"loss": 0.6662,
|
| 1178 |
+
"step": 835
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.3917910447761194,
|
| 1182 |
+
"grad_norm": 0.7295699828504463,
|
| 1183 |
+
"learning_rate": 3.711018076719034e-05,
|
| 1184 |
+
"loss": 0.6578,
|
| 1185 |
+
"step": 840
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.3941231343283582,
|
| 1189 |
+
"grad_norm": 0.7263503777736142,
|
| 1190 |
+
"learning_rate": 3.695293707096947e-05,
|
| 1191 |
+
"loss": 0.6708,
|
| 1192 |
+
"step": 845
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.396455223880597,
|
| 1196 |
+
"grad_norm": 0.7812017284771229,
|
| 1197 |
+
"learning_rate": 3.679513071115025e-05,
|
| 1198 |
+
"loss": 0.6617,
|
| 1199 |
+
"step": 850
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.3987873134328358,
|
| 1203 |
+
"grad_norm": 0.6765328713271164,
|
| 1204 |
+
"learning_rate": 3.663677108078094e-05,
|
| 1205 |
+
"loss": 0.6674,
|
| 1206 |
+
"step": 855
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 0.40111940298507465,
|
| 1210 |
+
"grad_norm": 0.6874658416764338,
|
| 1211 |
+
"learning_rate": 3.647786760584194e-05,
|
| 1212 |
+
"loss": 0.657,
|
| 1213 |
+
"step": 860
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.40345149253731344,
|
| 1217 |
+
"grad_norm": 0.6396850329707013,
|
| 1218 |
+
"learning_rate": 3.6318429744684676e-05,
|
| 1219 |
+
"loss": 0.6728,
|
| 1220 |
+
"step": 865
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.40578358208955223,
|
| 1224 |
+
"grad_norm": 0.7376511879685435,
|
| 1225 |
+
"learning_rate": 3.615846698746869e-05,
|
| 1226 |
+
"loss": 0.6626,
|
| 1227 |
+
"step": 870
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.408115671641791,
|
| 1231 |
+
"grad_norm": 0.7037616699556043,
|
| 1232 |
+
"learning_rate": 3.599798885559667e-05,
|
| 1233 |
+
"loss": 0.677,
|
| 1234 |
+
"step": 875
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 0.41044776119402987,
|
| 1238 |
+
"grad_norm": 0.6519988831184151,
|
| 1239 |
+
"learning_rate": 3.583700490114776e-05,
|
| 1240 |
+
"loss": 0.6385,
|
| 1241 |
+
"step": 880
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.41277985074626866,
|
| 1245 |
+
"grad_norm": 0.6836720339790674,
|
| 1246 |
+
"learning_rate": 3.5675524706309014e-05,
|
| 1247 |
+
"loss": 0.6397,
|
| 1248 |
+
"step": 885
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 0.41511194029850745,
|
| 1252 |
+
"grad_norm": 0.6577966791961214,
|
| 1253 |
+
"learning_rate": 3.5513557882805e-05,
|
| 1254 |
+
"loss": 0.6718,
|
| 1255 |
+
"step": 890
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.41744402985074625,
|
| 1259 |
+
"grad_norm": 0.6676911965920004,
|
| 1260 |
+
"learning_rate": 3.5351114071325696e-05,
|
| 1261 |
+
"loss": 0.6552,
|
| 1262 |
+
"step": 895
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.4197761194029851,
|
| 1266 |
+
"grad_norm": 0.6775036363841046,
|
| 1267 |
+
"learning_rate": 3.518820294095267e-05,
|
| 1268 |
+
"loss": 0.66,
|
| 1269 |
+
"step": 900
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.4221082089552239,
|
| 1273 |
+
"grad_norm": 0.7274179785465997,
|
| 1274 |
+
"learning_rate": 3.50248341885835e-05,
|
| 1275 |
+
"loss": 0.6457,
|
| 1276 |
+
"step": 905
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 0.4244402985074627,
|
| 1280 |
+
"grad_norm": 0.6365285914594158,
|
| 1281 |
+
"learning_rate": 3.486101753835468e-05,
|
| 1282 |
+
"loss": 0.6549,
|
| 1283 |
+
"step": 910
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.42677238805970147,
|
| 1287 |
+
"grad_norm": 0.7015328692797856,
|
| 1288 |
+
"learning_rate": 3.469676274106271e-05,
|
| 1289 |
+
"loss": 0.6446,
|
| 1290 |
+
"step": 915
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 0.4291044776119403,
|
| 1294 |
+
"grad_norm": 0.6897225314313983,
|
| 1295 |
+
"learning_rate": 3.453207957358377e-05,
|
| 1296 |
+
"loss": 0.6481,
|
| 1297 |
+
"step": 920
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.4314365671641791,
|
| 1301 |
+
"grad_norm": 0.7018612270350465,
|
| 1302 |
+
"learning_rate": 3.436697783829178e-05,
|
| 1303 |
+
"loss": 0.6421,
|
| 1304 |
+
"step": 925
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.4337686567164179,
|
| 1308 |
+
"grad_norm": 0.683607419776046,
|
| 1309 |
+
"learning_rate": 3.420146736247487e-05,
|
| 1310 |
+
"loss": 0.6262,
|
| 1311 |
+
"step": 930
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.4361007462686567,
|
| 1315 |
+
"grad_norm": 0.6817625776138896,
|
| 1316 |
+
"learning_rate": 3.4035557997750506e-05,
|
| 1317 |
+
"loss": 0.6655,
|
| 1318 |
+
"step": 935
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 0.43843283582089554,
|
| 1322 |
+
"grad_norm": 0.7113749823554164,
|
| 1323 |
+
"learning_rate": 3.386925961947906e-05,
|
| 1324 |
+
"loss": 0.6589,
|
| 1325 |
+
"step": 940
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.44076492537313433,
|
| 1329 |
+
"grad_norm": 0.701429407427016,
|
| 1330 |
+
"learning_rate": 3.370258212617602e-05,
|
| 1331 |
+
"loss": 0.6756,
|
| 1332 |
+
"step": 945
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 0.4430970149253731,
|
| 1336 |
+
"grad_norm": 0.6831638841691235,
|
| 1337 |
+
"learning_rate": 3.353553543892277e-05,
|
| 1338 |
+
"loss": 0.645,
|
| 1339 |
+
"step": 950
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.4454291044776119,
|
| 1343 |
+
"grad_norm": 0.6951969526491697,
|
| 1344 |
+
"learning_rate": 3.336812950077611e-05,
|
| 1345 |
+
"loss": 0.653,
|
| 1346 |
+
"step": 955
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.44776119402985076,
|
| 1350 |
+
"grad_norm": 0.712881849067669,
|
| 1351 |
+
"learning_rate": 3.320037427617639e-05,
|
| 1352 |
+
"loss": 0.6503,
|
| 1353 |
+
"step": 960
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.45009328358208955,
|
| 1357 |
+
"grad_norm": 0.677080067582334,
|
| 1358 |
+
"learning_rate": 3.30322797503544e-05,
|
| 1359 |
+
"loss": 0.6634,
|
| 1360 |
+
"step": 965
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 0.45242537313432835,
|
| 1364 |
+
"grad_norm": 0.673026887857187,
|
| 1365 |
+
"learning_rate": 3.2863855928737026e-05,
|
| 1366 |
+
"loss": 0.6299,
|
| 1367 |
+
"step": 970
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.45475746268656714,
|
| 1371 |
+
"grad_norm": 0.6976163186906711,
|
| 1372 |
+
"learning_rate": 3.2695112836351703e-05,
|
| 1373 |
+
"loss": 0.6514,
|
| 1374 |
+
"step": 975
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.457089552238806,
|
| 1378 |
+
"grad_norm": 0.6835684109054025,
|
| 1379 |
+
"learning_rate": 3.252606051722972e-05,
|
| 1380 |
+
"loss": 0.6715,
|
| 1381 |
+
"step": 980
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.4594216417910448,
|
| 1385 |
+
"grad_norm": 0.634542987869755,
|
| 1386 |
+
"learning_rate": 3.235670903380832e-05,
|
| 1387 |
+
"loss": 0.6452,
|
| 1388 |
+
"step": 985
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 0.46175373134328357,
|
| 1392 |
+
"grad_norm": 0.6700081127629496,
|
| 1393 |
+
"learning_rate": 3.218706846633183e-05,
|
| 1394 |
+
"loss": 0.6424,
|
| 1395 |
+
"step": 990
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.46408582089552236,
|
| 1399 |
+
"grad_norm": 0.674844701562939,
|
| 1400 |
+
"learning_rate": 3.201714891225156e-05,
|
| 1401 |
+
"loss": 0.64,
|
| 1402 |
+
"step": 995
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 0.4664179104477612,
|
| 1406 |
+
"grad_norm": 0.6901945362990005,
|
| 1407 |
+
"learning_rate": 3.1846960485624886e-05,
|
| 1408 |
+
"loss": 0.6359,
|
| 1409 |
+
"step": 1000
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.46875,
|
| 1413 |
+
"grad_norm": 0.711302010737976,
|
| 1414 |
+
"learning_rate": 3.1676513316513156e-05,
|
| 1415 |
+
"loss": 0.6613,
|
| 1416 |
+
"step": 1005
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 0.4710820895522388,
|
| 1420 |
+
"grad_norm": 0.6415681002352712,
|
| 1421 |
+
"learning_rate": 3.150581755037877e-05,
|
| 1422 |
+
"loss": 0.6448,
|
| 1423 |
+
"step": 1010
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.47341417910447764,
|
| 1427 |
+
"grad_norm": 0.6730615759692896,
|
| 1428 |
+
"learning_rate": 3.133488334748125e-05,
|
| 1429 |
+
"loss": 0.6569,
|
| 1430 |
+
"step": 1015
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 0.47574626865671643,
|
| 1434 |
+
"grad_norm": 0.6995346562904475,
|
| 1435 |
+
"learning_rate": 3.1163720882272516e-05,
|
| 1436 |
+
"loss": 0.637,
|
| 1437 |
+
"step": 1020
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.4780783582089552,
|
| 1441 |
+
"grad_norm": 0.7048313864139962,
|
| 1442 |
+
"learning_rate": 3.0992340342791246e-05,
|
| 1443 |
+
"loss": 0.6496,
|
| 1444 |
+
"step": 1025
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 0.480410447761194,
|
| 1448 |
+
"grad_norm": 0.6599686029534936,
|
| 1449 |
+
"learning_rate": 3.08207519300565e-05,
|
| 1450 |
+
"loss": 0.642,
|
| 1451 |
+
"step": 1030
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.48274253731343286,
|
| 1455 |
+
"grad_norm": 0.676508421822385,
|
| 1456 |
+
"learning_rate": 3.064896585746045e-05,
|
| 1457 |
+
"loss": 0.6307,
|
| 1458 |
+
"step": 1035
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 0.48507462686567165,
|
| 1462 |
+
"grad_norm": 0.6701675191331543,
|
| 1463 |
+
"learning_rate": 3.047699235016056e-05,
|
| 1464 |
+
"loss": 0.6384,
|
| 1465 |
+
"step": 1040
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.48740671641791045,
|
| 1469 |
+
"grad_norm": 0.7258698708508784,
|
| 1470 |
+
"learning_rate": 3.030484164447085e-05,
|
| 1471 |
+
"loss": 0.6283,
|
| 1472 |
+
"step": 1045
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 0.48973880597014924,
|
| 1476 |
+
"grad_norm": 0.7407558467480371,
|
| 1477 |
+
"learning_rate": 3.0132523987252658e-05,
|
| 1478 |
+
"loss": 0.6495,
|
| 1479 |
+
"step": 1050
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.4920708955223881,
|
| 1483 |
+
"grad_norm": 0.689449626576069,
|
| 1484 |
+
"learning_rate": 2.9960049635304755e-05,
|
| 1485 |
+
"loss": 0.6219,
|
| 1486 |
+
"step": 1055
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 0.4944029850746269,
|
| 1490 |
+
"grad_norm": 0.6605235489630851,
|
| 1491 |
+
"learning_rate": 2.9787428854752736e-05,
|
| 1492 |
+
"loss": 0.6328,
|
| 1493 |
+
"step": 1060
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.49673507462686567,
|
| 1497 |
+
"grad_norm": 0.6521941117929255,
|
| 1498 |
+
"learning_rate": 2.961467192043807e-05,
|
| 1499 |
+
"loss": 0.6445,
|
| 1500 |
+
"step": 1065
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 0.49906716417910446,
|
| 1504 |
+
"grad_norm": 0.719415886577355,
|
| 1505 |
+
"learning_rate": 2.9441789115306402e-05,
|
| 1506 |
+
"loss": 0.6601,
|
| 1507 |
+
"step": 1070
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.5013992537313433,
|
| 1511 |
+
"grad_norm": 0.7143547584326713,
|
| 1512 |
+
"learning_rate": 2.926879072979558e-05,
|
| 1513 |
+
"loss": 0.6477,
|
| 1514 |
+
"step": 1075
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 0.503731343283582,
|
| 1518 |
+
"grad_norm": 0.6805215530249611,
|
| 1519 |
+
"learning_rate": 2.9095687061223058e-05,
|
| 1520 |
+
"loss": 0.6456,
|
| 1521 |
+
"step": 1080
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.5060634328358209,
|
| 1525 |
+
"grad_norm": 0.6951914121970127,
|
| 1526 |
+
"learning_rate": 2.8922488413173053e-05,
|
| 1527 |
+
"loss": 0.6441,
|
| 1528 |
+
"step": 1085
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 0.5083955223880597,
|
| 1532 |
+
"grad_norm": 0.706237459032283,
|
| 1533 |
+
"learning_rate": 2.874920509488319e-05,
|
| 1534 |
+
"loss": 0.6574,
|
| 1535 |
+
"step": 1090
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 0.5107276119402985,
|
| 1539 |
+
"grad_norm": 0.7346760125128584,
|
| 1540 |
+
"learning_rate": 2.8575847420630887e-05,
|
| 1541 |
+
"loss": 0.6314,
|
| 1542 |
+
"step": 1095
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 0.5130597014925373,
|
| 1546 |
+
"grad_norm": 0.6542912159857234,
|
| 1547 |
+
"learning_rate": 2.8402425709119435e-05,
|
| 1548 |
+
"loss": 0.659,
|
| 1549 |
+
"step": 1100
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 0.5153917910447762,
|
| 1553 |
+
"grad_norm": 0.6819702914670763,
|
| 1554 |
+
"learning_rate": 2.8228950282863776e-05,
|
| 1555 |
+
"loss": 0.6492,
|
| 1556 |
+
"step": 1105
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 0.5177238805970149,
|
| 1560 |
+
"grad_norm": 0.6770174696103765,
|
| 1561 |
+
"learning_rate": 2.8055431467576106e-05,
|
| 1562 |
+
"loss": 0.638,
|
| 1563 |
+
"step": 1110
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.5200559701492538,
|
| 1567 |
+
"grad_norm": 0.662818262088358,
|
| 1568 |
+
"learning_rate": 2.788187959155124e-05,
|
| 1569 |
+
"loss": 0.63,
|
| 1570 |
+
"step": 1115
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 0.5223880597014925,
|
| 1574 |
+
"grad_norm": 0.7072513898861241,
|
| 1575 |
+
"learning_rate": 2.7708304985051868e-05,
|
| 1576 |
+
"loss": 0.6623,
|
| 1577 |
+
"step": 1120
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 0.5247201492537313,
|
| 1581 |
+
"grad_norm": 0.711447160492985,
|
| 1582 |
+
"learning_rate": 2.7534717979693647e-05,
|
| 1583 |
+
"loss": 0.6344,
|
| 1584 |
+
"step": 1125
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 0.5270522388059702,
|
| 1588 |
+
"grad_norm": 0.6918500244833096,
|
| 1589 |
+
"learning_rate": 2.7361128907830253e-05,
|
| 1590 |
+
"loss": 0.6385,
|
| 1591 |
+
"step": 1130
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 0.5293843283582089,
|
| 1595 |
+
"grad_norm": 0.6722193101035042,
|
| 1596 |
+
"learning_rate": 2.7187548101938353e-05,
|
| 1597 |
+
"loss": 0.6256,
|
| 1598 |
+
"step": 1135
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 0.5317164179104478,
|
| 1602 |
+
"grad_norm": 0.6569908241938373,
|
| 1603 |
+
"learning_rate": 2.7013985894002623e-05,
|
| 1604 |
+
"loss": 0.6236,
|
| 1605 |
+
"step": 1140
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.5340485074626866,
|
| 1609 |
+
"grad_norm": 0.695528007373876,
|
| 1610 |
+
"learning_rate": 2.6840452614900726e-05,
|
| 1611 |
+
"loss": 0.6533,
|
| 1612 |
+
"step": 1145
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 0.5363805970149254,
|
| 1616 |
+
"grad_norm": 0.6182593924769839,
|
| 1617 |
+
"learning_rate": 2.6666958593788405e-05,
|
| 1618 |
+
"loss": 0.631,
|
| 1619 |
+
"step": 1150
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 0.5387126865671642,
|
| 1623 |
+
"grad_norm": 0.6254875427778447,
|
| 1624 |
+
"learning_rate": 2.649351415748466e-05,
|
| 1625 |
+
"loss": 0.6528,
|
| 1626 |
+
"step": 1155
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 0.5410447761194029,
|
| 1630 |
+
"grad_norm": 0.6195472352586545,
|
| 1631 |
+
"learning_rate": 2.6320129629857093e-05,
|
| 1632 |
+
"loss": 0.6421,
|
| 1633 |
+
"step": 1160
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 0.5433768656716418,
|
| 1637 |
+
"grad_norm": 0.6406233353457038,
|
| 1638 |
+
"learning_rate": 2.6146815331207358e-05,
|
| 1639 |
+
"loss": 0.6183,
|
| 1640 |
+
"step": 1165
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 0.5457089552238806,
|
| 1644 |
+
"grad_norm": 0.6456118637481933,
|
| 1645 |
+
"learning_rate": 2.597358157765692e-05,
|
| 1646 |
+
"loss": 0.6474,
|
| 1647 |
+
"step": 1170
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.5480410447761194,
|
| 1651 |
+
"grad_norm": 0.6505557028284191,
|
| 1652 |
+
"learning_rate": 2.5800438680532974e-05,
|
| 1653 |
+
"loss": 0.6307,
|
| 1654 |
+
"step": 1175
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 0.5503731343283582,
|
| 1658 |
+
"grad_norm": 0.6673856033395356,
|
| 1659 |
+
"learning_rate": 2.56273969457547e-05,
|
| 1660 |
+
"loss": 0.6305,
|
| 1661 |
+
"step": 1180
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 0.5527052238805971,
|
| 1665 |
+
"grad_norm": 0.6976453369750254,
|
| 1666 |
+
"learning_rate": 2.545446667321984e-05,
|
| 1667 |
+
"loss": 0.6511,
|
| 1668 |
+
"step": 1185
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 0.5550373134328358,
|
| 1672 |
+
"grad_norm": 0.6523315253145765,
|
| 1673 |
+
"learning_rate": 2.528165815619162e-05,
|
| 1674 |
+
"loss": 0.6326,
|
| 1675 |
+
"step": 1190
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 0.5573694029850746,
|
| 1679 |
+
"grad_norm": 0.6513887450128372,
|
| 1680 |
+
"learning_rate": 2.5108981680686035e-05,
|
| 1681 |
+
"loss": 0.6357,
|
| 1682 |
+
"step": 1195
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 0.5597014925373134,
|
| 1686 |
+
"grad_norm": 0.6686846233513335,
|
| 1687 |
+
"learning_rate": 2.4936447524859625e-05,
|
| 1688 |
+
"loss": 0.6243,
|
| 1689 |
+
"step": 1200
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.5620335820895522,
|
| 1693 |
+
"grad_norm": 0.6254057596195436,
|
| 1694 |
+
"learning_rate": 2.4764065958397715e-05,
|
| 1695 |
+
"loss": 0.6273,
|
| 1696 |
+
"step": 1205
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 0.5643656716417911,
|
| 1700 |
+
"grad_norm": 0.6496921416294746,
|
| 1701 |
+
"learning_rate": 2.459184724190308e-05,
|
| 1702 |
+
"loss": 0.6179,
|
| 1703 |
+
"step": 1210
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.5666977611940298,
|
| 1707 |
+
"grad_norm": 0.6568624775107711,
|
| 1708 |
+
"learning_rate": 2.441980162628527e-05,
|
| 1709 |
+
"loss": 0.6344,
|
| 1710 |
+
"step": 1215
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 0.5690298507462687,
|
| 1714 |
+
"grad_norm": 0.6567469938905354,
|
| 1715 |
+
"learning_rate": 2.4247939352150386e-05,
|
| 1716 |
+
"loss": 0.6449,
|
| 1717 |
+
"step": 1220
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.5713619402985075,
|
| 1721 |
+
"grad_norm": 0.61556362280737,
|
| 1722 |
+
"learning_rate": 2.4076270649191573e-05,
|
| 1723 |
+
"loss": 0.6555,
|
| 1724 |
+
"step": 1225
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 0.5736940298507462,
|
| 1728 |
+
"grad_norm": 0.6394081674254839,
|
| 1729 |
+
"learning_rate": 2.390480573558012e-05,
|
| 1730 |
+
"loss": 0.6363,
|
| 1731 |
+
"step": 1230
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.5760261194029851,
|
| 1735 |
+
"grad_norm": 0.6051477641244504,
|
| 1736 |
+
"learning_rate": 2.3733554817357246e-05,
|
| 1737 |
+
"loss": 0.6348,
|
| 1738 |
+
"step": 1235
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 0.5783582089552238,
|
| 1742 |
+
"grad_norm": 0.5930962238087756,
|
| 1743 |
+
"learning_rate": 2.3562528087826573e-05,
|
| 1744 |
+
"loss": 0.6206,
|
| 1745 |
+
"step": 1240
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.5806902985074627,
|
| 1749 |
+
"grad_norm": 0.698655785577389,
|
| 1750 |
+
"learning_rate": 2.339173572694746e-05,
|
| 1751 |
+
"loss": 0.6273,
|
| 1752 |
+
"step": 1245
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 0.5830223880597015,
|
| 1756 |
+
"grad_norm": 0.6327913417184158,
|
| 1757 |
+
"learning_rate": 2.3221187900729003e-05,
|
| 1758 |
+
"loss": 0.637,
|
| 1759 |
+
"step": 1250
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.5853544776119403,
|
| 1763 |
+
"grad_norm": 0.6339231296671493,
|
| 1764 |
+
"learning_rate": 2.3050894760624982e-05,
|
| 1765 |
+
"loss": 0.641,
|
| 1766 |
+
"step": 1255
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 0.5876865671641791,
|
| 1770 |
+
"grad_norm": 0.6051505920325168,
|
| 1771 |
+
"learning_rate": 2.2880866442929544e-05,
|
| 1772 |
+
"loss": 0.6265,
|
| 1773 |
+
"step": 1260
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.590018656716418,
|
| 1777 |
+
"grad_norm": 0.6637235396372423,
|
| 1778 |
+
"learning_rate": 2.271111306817396e-05,
|
| 1779 |
+
"loss": 0.6482,
|
| 1780 |
+
"step": 1265
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 0.5923507462686567,
|
| 1784 |
+
"grad_norm": 0.6469789610641435,
|
| 1785 |
+
"learning_rate": 2.254164474052416e-05,
|
| 1786 |
+
"loss": 0.6266,
|
| 1787 |
+
"step": 1270
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 0.5946828358208955,
|
| 1791 |
+
"grad_norm": 0.5897603262670164,
|
| 1792 |
+
"learning_rate": 2.237247154717932e-05,
|
| 1793 |
+
"loss": 0.6188,
|
| 1794 |
+
"step": 1275
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 0.5970149253731343,
|
| 1798 |
+
"grad_norm": 0.6115820772990138,
|
| 1799 |
+
"learning_rate": 2.2203603557771447e-05,
|
| 1800 |
+
"loss": 0.6198,
|
| 1801 |
+
"step": 1280
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 0.5993470149253731,
|
| 1805 |
+
"grad_norm": 0.6159243257321477,
|
| 1806 |
+
"learning_rate": 2.2035050823766008e-05,
|
| 1807 |
+
"loss": 0.6313,
|
| 1808 |
+
"step": 1285
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 0.601679104477612,
|
| 1812 |
+
"grad_norm": 0.6683096146103435,
|
| 1813 |
+
"learning_rate": 2.186682337786365e-05,
|
| 1814 |
+
"loss": 0.6369,
|
| 1815 |
+
"step": 1290
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.6040111940298507,
|
| 1819 |
+
"grad_norm": 0.603613728907185,
|
| 1820 |
+
"learning_rate": 2.1698931233403013e-05,
|
| 1821 |
+
"loss": 0.6151,
|
| 1822 |
+
"step": 1295
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 0.6063432835820896,
|
| 1826 |
+
"grad_norm": 0.6004781166951644,
|
| 1827 |
+
"learning_rate": 2.153138438376473e-05,
|
| 1828 |
+
"loss": 0.622,
|
| 1829 |
+
"step": 1300
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 0.6086753731343284,
|
| 1833 |
+
"grad_norm": 18.22261692025831,
|
| 1834 |
+
"learning_rate": 2.136419280177655e-05,
|
| 1835 |
+
"loss": 0.6161,
|
| 1836 |
+
"step": 1305
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 0.6110074626865671,
|
| 1840 |
+
"grad_norm": 0.5978990585115841,
|
| 1841 |
+
"learning_rate": 2.119736643911979e-05,
|
| 1842 |
+
"loss": 0.6416,
|
| 1843 |
+
"step": 1310
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 0.613339552238806,
|
| 1847 |
+
"grad_norm": 0.597454038032284,
|
| 1848 |
+
"learning_rate": 2.1030915225736947e-05,
|
| 1849 |
+
"loss": 0.6288,
|
| 1850 |
+
"step": 1315
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 0.6156716417910447,
|
| 1854 |
+
"grad_norm": 0.6430793847859734,
|
| 1855 |
+
"learning_rate": 2.0864849069240645e-05,
|
| 1856 |
+
"loss": 0.6383,
|
| 1857 |
+
"step": 1320
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 0.6180037313432836,
|
| 1861 |
+
"grad_norm": 0.5831365581687563,
|
| 1862 |
+
"learning_rate": 2.0699177854323902e-05,
|
| 1863 |
+
"loss": 0.6295,
|
| 1864 |
+
"step": 1325
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 0.6203358208955224,
|
| 1868 |
+
"grad_norm": 0.5998744003452269,
|
| 1869 |
+
"learning_rate": 2.0533911442171805e-05,
|
| 1870 |
+
"loss": 0.6287,
|
| 1871 |
+
"step": 1330
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 0.6226679104477612,
|
| 1875 |
+
"grad_norm": 0.6643837943700802,
|
| 1876 |
+
"learning_rate": 2.036905966987449e-05,
|
| 1877 |
+
"loss": 0.6343,
|
| 1878 |
+
"step": 1335
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 0.625,
|
| 1882 |
+
"grad_norm": 0.6438429216570205,
|
| 1883 |
+
"learning_rate": 2.0204632349841667e-05,
|
| 1884 |
+
"loss": 0.616,
|
| 1885 |
+
"step": 1340
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 0.6273320895522388,
|
| 1889 |
+
"grad_norm": 0.6110144914242343,
|
| 1890 |
+
"learning_rate": 2.0040639269218532e-05,
|
| 1891 |
+
"loss": 0.6212,
|
| 1892 |
+
"step": 1345
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 0.6296641791044776,
|
| 1896 |
+
"grad_norm": 0.6615405256014992,
|
| 1897 |
+
"learning_rate": 1.9877090189303182e-05,
|
| 1898 |
+
"loss": 0.6463,
|
| 1899 |
+
"step": 1350
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 0.6319962686567164,
|
| 1903 |
+
"grad_norm": 0.6183940678157902,
|
| 1904 |
+
"learning_rate": 1.9713994844965657e-05,
|
| 1905 |
+
"loss": 0.6183,
|
| 1906 |
+
"step": 1355
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 0.6343283582089553,
|
| 1910 |
+
"grad_norm": 0.6746036398298787,
|
| 1911 |
+
"learning_rate": 1.9551362944068462e-05,
|
| 1912 |
+
"loss": 0.6301,
|
| 1913 |
+
"step": 1360
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 0.636660447761194,
|
| 1917 |
+
"grad_norm": 0.648307373321323,
|
| 1918 |
+
"learning_rate": 1.938920416688874e-05,
|
| 1919 |
+
"loss": 0.6308,
|
| 1920 |
+
"step": 1365
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 0.6389925373134329,
|
| 1924 |
+
"grad_norm": 0.6069645644995993,
|
| 1925 |
+
"learning_rate": 1.922752816554204e-05,
|
| 1926 |
+
"loss": 0.6205,
|
| 1927 |
+
"step": 1370
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 0.6413246268656716,
|
| 1931 |
+
"grad_norm": 0.5822850576614598,
|
| 1932 |
+
"learning_rate": 1.9066344563407856e-05,
|
| 1933 |
+
"loss": 0.6295,
|
| 1934 |
+
"step": 1375
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 0.6436567164179104,
|
| 1938 |
+
"grad_norm": 0.6496746526806308,
|
| 1939 |
+
"learning_rate": 1.890566295455678e-05,
|
| 1940 |
+
"loss": 0.605,
|
| 1941 |
+
"step": 1380
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 0.6459888059701493,
|
| 1945 |
+
"grad_norm": 0.5866035567882427,
|
| 1946 |
+
"learning_rate": 1.874549290317946e-05,
|
| 1947 |
+
"loss": 0.608,
|
| 1948 |
+
"step": 1385
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 0.648320895522388,
|
| 1952 |
+
"grad_norm": 0.6041775047865613,
|
| 1953 |
+
"learning_rate": 1.858584394301728e-05,
|
| 1954 |
+
"loss": 0.6356,
|
| 1955 |
+
"step": 1390
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 0.6506529850746269,
|
| 1959 |
+
"grad_norm": 0.5932846773816811,
|
| 1960 |
+
"learning_rate": 1.8426725576794918e-05,
|
| 1961 |
+
"loss": 0.6576,
|
| 1962 |
+
"step": 1395
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 0.6529850746268657,
|
| 1966 |
+
"grad_norm": 0.6021211057686333,
|
| 1967 |
+
"learning_rate": 1.8268147275654707e-05,
|
| 1968 |
+
"loss": 0.6276,
|
| 1969 |
+
"step": 1400
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 0.6553171641791045,
|
| 1973 |
+
"grad_norm": 0.5706056220756198,
|
| 1974 |
+
"learning_rate": 1.8110118478592915e-05,
|
| 1975 |
+
"loss": 0.6129,
|
| 1976 |
+
"step": 1405
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 0.6576492537313433,
|
| 1980 |
+
"grad_norm": 0.6303230581411896,
|
| 1981 |
+
"learning_rate": 1.7952648591897858e-05,
|
| 1982 |
+
"loss": 0.6296,
|
| 1983 |
+
"step": 1410
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 0.659981343283582,
|
| 1987 |
+
"grad_norm": 0.6069553160577528,
|
| 1988 |
+
"learning_rate": 1.7795746988590027e-05,
|
| 1989 |
+
"loss": 0.6162,
|
| 1990 |
+
"step": 1415
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 0.6623134328358209,
|
| 1994 |
+
"grad_norm": 0.6233052721164215,
|
| 1995 |
+
"learning_rate": 1.7639423007864252e-05,
|
| 1996 |
+
"loss": 0.6071,
|
| 1997 |
+
"step": 1420
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 0.6646455223880597,
|
| 2001 |
+
"grad_norm": 0.5792817379292262,
|
| 2002 |
+
"learning_rate": 1.7483685954533692e-05,
|
| 2003 |
+
"loss": 0.6186,
|
| 2004 |
+
"step": 1425
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 0.6669776119402985,
|
| 2008 |
+
"grad_norm": 0.598455839355501,
|
| 2009 |
+
"learning_rate": 1.7328545098476106e-05,
|
| 2010 |
+
"loss": 0.6126,
|
| 2011 |
+
"step": 1430
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 0.6693097014925373,
|
| 2015 |
+
"grad_norm": 0.5887965955039507,
|
| 2016 |
+
"learning_rate": 1.717400967408196e-05,
|
| 2017 |
+
"loss": 0.6219,
|
| 2018 |
+
"step": 1435
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 0.6716417910447762,
|
| 2022 |
+
"grad_norm": 0.6037531009113533,
|
| 2023 |
+
"learning_rate": 1.702008887970491e-05,
|
| 2024 |
+
"loss": 0.6164,
|
| 2025 |
+
"step": 1440
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 0.6739738805970149,
|
| 2029 |
+
"grad_norm": 0.5608388614868615,
|
| 2030 |
+
"learning_rate": 1.6866791877114165e-05,
|
| 2031 |
+
"loss": 0.6246,
|
| 2032 |
+
"step": 1445
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 0.6763059701492538,
|
| 2036 |
+
"grad_norm": 0.5732233279593167,
|
| 2037 |
+
"learning_rate": 1.671412779094926e-05,
|
| 2038 |
+
"loss": 0.622,
|
| 2039 |
+
"step": 1450
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 0.6786380597014925,
|
| 2043 |
+
"grad_norm": 0.604276726888298,
|
| 2044 |
+
"learning_rate": 1.656210570817685e-05,
|
| 2045 |
+
"loss": 0.6227,
|
| 2046 |
+
"step": 1455
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 0.6809701492537313,
|
| 2050 |
+
"grad_norm": 0.6084928698482973,
|
| 2051 |
+
"learning_rate": 1.6410734677549872e-05,
|
| 2052 |
+
"loss": 0.6156,
|
| 2053 |
+
"step": 1460
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 0.6833022388059702,
|
| 2057 |
+
"grad_norm": 0.6167514863041356,
|
| 2058 |
+
"learning_rate": 1.6260023709068932e-05,
|
| 2059 |
+
"loss": 0.622,
|
| 2060 |
+
"step": 1465
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 0.6856343283582089,
|
| 2064 |
+
"grad_norm": 0.6360423400540859,
|
| 2065 |
+
"learning_rate": 1.6109981773446036e-05,
|
| 2066 |
+
"loss": 0.6183,
|
| 2067 |
+
"step": 1470
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 0.6879664179104478,
|
| 2071 |
+
"grad_norm": 0.6092909418703362,
|
| 2072 |
+
"learning_rate": 1.5960617801570555e-05,
|
| 2073 |
+
"loss": 0.6162,
|
| 2074 |
+
"step": 1475
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 0.6902985074626866,
|
| 2078 |
+
"grad_norm": 0.6033346353614264,
|
| 2079 |
+
"learning_rate": 1.58119406839777e-05,
|
| 2080 |
+
"loss": 0.6221,
|
| 2081 |
+
"step": 1480
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 0.6926305970149254,
|
| 2085 |
+
"grad_norm": 0.5497901412692066,
|
| 2086 |
+
"learning_rate": 1.566395927031932e-05,
|
| 2087 |
+
"loss": 0.6094,
|
| 2088 |
+
"step": 1485
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 0.6949626865671642,
|
| 2092 |
+
"grad_norm": 0.5726565331606571,
|
| 2093 |
+
"learning_rate": 1.5516682368837133e-05,
|
| 2094 |
+
"loss": 0.6165,
|
| 2095 |
+
"step": 1490
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 0.6972947761194029,
|
| 2099 |
+
"grad_norm": 0.6379827482434526,
|
| 2100 |
+
"learning_rate": 1.5370118745838453e-05,
|
| 2101 |
+
"loss": 0.6404,
|
| 2102 |
+
"step": 1495
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 0.6996268656716418,
|
| 2106 |
+
"grad_norm": 0.6035346967740655,
|
| 2107 |
+
"learning_rate": 1.5224277125174388e-05,
|
| 2108 |
+
"loss": 0.6238,
|
| 2109 |
+
"step": 1500
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 0.7019589552238806,
|
| 2113 |
+
"grad_norm": 0.6046299493263414,
|
| 2114 |
+
"learning_rate": 1.5079166187720561e-05,
|
| 2115 |
+
"loss": 0.6233,
|
| 2116 |
+
"step": 1505
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 0.7042910447761194,
|
| 2120 |
+
"grad_norm": 0.5813425376508258,
|
| 2121 |
+
"learning_rate": 1.4934794570860416e-05,
|
| 2122 |
+
"loss": 0.6406,
|
| 2123 |
+
"step": 1510
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 0.7066231343283582,
|
| 2127 |
+
"grad_norm": 0.6005101213362166,
|
| 2128 |
+
"learning_rate": 1.4791170867971132e-05,
|
| 2129 |
+
"loss": 0.6167,
|
| 2130 |
+
"step": 1515
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 0.7089552238805971,
|
| 2134 |
+
"grad_norm": 0.5968043199699391,
|
| 2135 |
+
"learning_rate": 1.464830362791204e-05,
|
| 2136 |
+
"loss": 0.6154,
|
| 2137 |
+
"step": 1520
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 0.7112873134328358,
|
| 2141 |
+
"grad_norm": 0.585701695377086,
|
| 2142 |
+
"learning_rate": 1.450620135451585e-05,
|
| 2143 |
+
"loss": 0.6225,
|
| 2144 |
+
"step": 1525
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 0.7136194029850746,
|
| 2148 |
+
"grad_norm": 0.5957247172214722,
|
| 2149 |
+
"learning_rate": 1.4364872506082425e-05,
|
| 2150 |
+
"loss": 0.6282,
|
| 2151 |
+
"step": 1530
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 0.7159514925373134,
|
| 2155 |
+
"grad_norm": 0.5967837516043032,
|
| 2156 |
+
"learning_rate": 1.4224325494875385e-05,
|
| 2157 |
+
"loss": 0.6399,
|
| 2158 |
+
"step": 1535
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 0.7182835820895522,
|
| 2162 |
+
"grad_norm": 0.587594664717925,
|
| 2163 |
+
"learning_rate": 1.4084568686621314e-05,
|
| 2164 |
+
"loss": 0.6201,
|
| 2165 |
+
"step": 1540
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 0.7206156716417911,
|
| 2169 |
+
"grad_norm": 0.5617516377764582,
|
| 2170 |
+
"learning_rate": 1.3945610400011851e-05,
|
| 2171 |
+
"loss": 0.633,
|
| 2172 |
+
"step": 1545
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 0.7229477611940298,
|
| 2176 |
+
"grad_norm": 0.6036131494081576,
|
| 2177 |
+
"learning_rate": 1.3807458906208546e-05,
|
| 2178 |
+
"loss": 0.6421,
|
| 2179 |
+
"step": 1550
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 0.7252798507462687,
|
| 2183 |
+
"grad_norm": 0.5619681271472718,
|
| 2184 |
+
"learning_rate": 1.3670122428350521e-05,
|
| 2185 |
+
"loss": 0.618,
|
| 2186 |
+
"step": 1555
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 0.7276119402985075,
|
| 2190 |
+
"grad_norm": 0.5870654465012264,
|
| 2191 |
+
"learning_rate": 1.3533609141065008e-05,
|
| 2192 |
+
"loss": 0.6101,
|
| 2193 |
+
"step": 1560
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 0.7299440298507462,
|
| 2197 |
+
"grad_norm": 0.6499759452581642,
|
| 2198 |
+
"learning_rate": 1.3397927169980773e-05,
|
| 2199 |
+
"loss": 0.6161,
|
| 2200 |
+
"step": 1565
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 0.7322761194029851,
|
| 2204 |
+
"grad_norm": 0.5802369273091457,
|
| 2205 |
+
"learning_rate": 1.326308459124447e-05,
|
| 2206 |
+
"loss": 0.6156,
|
| 2207 |
+
"step": 1570
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 0.7346082089552238,
|
| 2211 |
+
"grad_norm": 0.5862527640011228,
|
| 2212 |
+
"learning_rate": 1.3129089431039931e-05,
|
| 2213 |
+
"loss": 0.6218,
|
| 2214 |
+
"step": 1575
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 0.7369402985074627,
|
| 2218 |
+
"grad_norm": 0.6041233460220358,
|
| 2219 |
+
"learning_rate": 1.299594966511038e-05,
|
| 2220 |
+
"loss": 0.6344,
|
| 2221 |
+
"step": 1580
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 0.7392723880597015,
|
| 2225 |
+
"grad_norm": 0.5761625096468084,
|
| 2226 |
+
"learning_rate": 1.2863673218283783e-05,
|
| 2227 |
+
"loss": 0.617,
|
| 2228 |
+
"step": 1585
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 0.7416044776119403,
|
| 2232 |
+
"grad_norm": 0.6001779302881813,
|
| 2233 |
+
"learning_rate": 1.2732267964001033e-05,
|
| 2234 |
+
"loss": 0.6101,
|
| 2235 |
+
"step": 1590
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 0.7439365671641791,
|
| 2239 |
+
"grad_norm": 0.5964143297371681,
|
| 2240 |
+
"learning_rate": 1.26017417238474e-05,
|
| 2241 |
+
"loss": 0.6212,
|
| 2242 |
+
"step": 1595
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 0.746268656716418,
|
| 2246 |
+
"grad_norm": 0.5964222072799055,
|
| 2247 |
+
"learning_rate": 1.2472102267086904e-05,
|
| 2248 |
+
"loss": 0.6211,
|
| 2249 |
+
"step": 1600
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 0.7486007462686567,
|
| 2253 |
+
"grad_norm": 0.6476516452528284,
|
| 2254 |
+
"learning_rate": 1.2343357310199925e-05,
|
| 2255 |
+
"loss": 0.6389,
|
| 2256 |
+
"step": 1605
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 0.7509328358208955,
|
| 2260 |
+
"grad_norm": 0.6211226051748449,
|
| 2261 |
+
"learning_rate": 1.2215514516423813e-05,
|
| 2262 |
+
"loss": 0.6345,
|
| 2263 |
+
"step": 1610
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 0.7532649253731343,
|
| 2267 |
+
"grad_norm": 0.6009531989706945,
|
| 2268 |
+
"learning_rate": 1.2088581495296852e-05,
|
| 2269 |
+
"loss": 0.6106,
|
| 2270 |
+
"step": 1615
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 0.7555970149253731,
|
| 2274 |
+
"grad_norm": 0.5457398269496171,
|
| 2275 |
+
"learning_rate": 1.1962565802205255e-05,
|
| 2276 |
+
"loss": 0.6107,
|
| 2277 |
+
"step": 1620
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 0.757929104477612,
|
| 2281 |
+
"grad_norm": 0.5485455623765262,
|
| 2282 |
+
"learning_rate": 1.1837474937933464e-05,
|
| 2283 |
+
"loss": 0.6194,
|
| 2284 |
+
"step": 1625
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 0.7602611940298507,
|
| 2288 |
+
"grad_norm": 0.5644743687543846,
|
| 2289 |
+
"learning_rate": 1.1713316348217673e-05,
|
| 2290 |
+
"loss": 0.6136,
|
| 2291 |
+
"step": 1630
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 0.7625932835820896,
|
| 2295 |
+
"grad_norm": 0.5944262234833658,
|
| 2296 |
+
"learning_rate": 1.1590097423302684e-05,
|
| 2297 |
+
"loss": 0.5942,
|
| 2298 |
+
"step": 1635
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 0.7649253731343284,
|
| 2302 |
+
"grad_norm": 0.5616590097486719,
|
| 2303 |
+
"learning_rate": 1.1467825497501954e-05,
|
| 2304 |
+
"loss": 0.6026,
|
| 2305 |
+
"step": 1640
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 0.7672574626865671,
|
| 2309 |
+
"grad_norm": 0.6146723556271906,
|
| 2310 |
+
"learning_rate": 1.1346507848761077e-05,
|
| 2311 |
+
"loss": 0.6156,
|
| 2312 |
+
"step": 1645
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 0.769589552238806,
|
| 2316 |
+
"grad_norm": 0.5965680064276904,
|
| 2317 |
+
"learning_rate": 1.1226151698224597e-05,
|
| 2318 |
+
"loss": 0.633,
|
| 2319 |
+
"step": 1650
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 0.7719216417910447,
|
| 2323 |
+
"grad_norm": 0.5614974770289655,
|
| 2324 |
+
"learning_rate": 1.1106764209806127e-05,
|
| 2325 |
+
"loss": 0.6194,
|
| 2326 |
+
"step": 1655
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 0.7742537313432836,
|
| 2330 |
+
"grad_norm": 0.6034383421575543,
|
| 2331 |
+
"learning_rate": 1.0988352489762006e-05,
|
| 2332 |
+
"loss": 0.6359,
|
| 2333 |
+
"step": 1660
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 0.7765858208955224,
|
| 2337 |
+
"grad_norm": 0.5568415876195435,
|
| 2338 |
+
"learning_rate": 1.0870923586268245e-05,
|
| 2339 |
+
"loss": 0.6143,
|
| 2340 |
+
"step": 1665
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 0.7789179104477612,
|
| 2344 |
+
"grad_norm": 0.5686287199324069,
|
| 2345 |
+
"learning_rate": 1.0754484489001085e-05,
|
| 2346 |
+
"loss": 0.6186,
|
| 2347 |
+
"step": 1670
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 0.78125,
|
| 2351 |
+
"grad_norm": 0.5739930134377701,
|
| 2352 |
+
"learning_rate": 1.0639042128720847e-05,
|
| 2353 |
+
"loss": 0.6235,
|
| 2354 |
+
"step": 1675
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 0.7835820895522388,
|
| 2358 |
+
"grad_norm": 0.5795603868856798,
|
| 2359 |
+
"learning_rate": 1.052460337685951e-05,
|
| 2360 |
+
"loss": 0.6049,
|
| 2361 |
+
"step": 1680
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 0.7859141791044776,
|
| 2365 |
+
"grad_norm": 0.56395308762799,
|
| 2366 |
+
"learning_rate": 1.0411175045111602e-05,
|
| 2367 |
+
"loss": 0.6032,
|
| 2368 |
+
"step": 1685
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 0.7882462686567164,
|
| 2372 |
+
"grad_norm": 0.58081174474857,
|
| 2373 |
+
"learning_rate": 1.0298763885028839e-05,
|
| 2374 |
+
"loss": 0.6178,
|
| 2375 |
+
"step": 1690
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 0.7905783582089553,
|
| 2379 |
+
"grad_norm": 0.5588983254320253,
|
| 2380 |
+
"learning_rate": 1.018737658761817e-05,
|
| 2381 |
+
"loss": 0.6105,
|
| 2382 |
+
"step": 1695
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 0.792910447761194,
|
| 2386 |
+
"grad_norm": 0.5364326303324629,
|
| 2387 |
+
"learning_rate": 1.0077019782943584e-05,
|
| 2388 |
+
"loss": 0.6237,
|
| 2389 |
+
"step": 1700
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 0.7952425373134329,
|
| 2393 |
+
"grad_norm": 0.5535399918900618,
|
| 2394 |
+
"learning_rate": 9.967700039731427e-06,
|
| 2395 |
+
"loss": 0.6129,
|
| 2396 |
+
"step": 1705
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 0.7975746268656716,
|
| 2400 |
+
"grad_norm": 0.5357340402470735,
|
| 2401 |
+
"learning_rate": 9.859423864979441e-06,
|
| 2402 |
+
"loss": 0.5947,
|
| 2403 |
+
"step": 1710
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 0.7999067164179104,
|
| 2407 |
+
"grad_norm": 0.6043622414636253,
|
| 2408 |
+
"learning_rate": 9.752197703569422e-06,
|
| 2409 |
+
"loss": 0.6341,
|
| 2410 |
+
"step": 1715
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 0.8022388059701493,
|
| 2414 |
+
"grad_norm": 0.5721045372713984,
|
| 2415 |
+
"learning_rate": 9.646027937883622e-06,
|
| 2416 |
+
"loss": 0.6161,
|
| 2417 |
+
"step": 1720
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 0.804570895522388,
|
| 2421 |
+
"grad_norm": 0.5312555490120068,
|
| 2422 |
+
"learning_rate": 9.54092088742485e-06,
|
| 2423 |
+
"loss": 0.64,
|
| 2424 |
+
"step": 1725
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 0.8069029850746269,
|
| 2428 |
+
"grad_norm": 0.5439729568974323,
|
| 2429 |
+
"learning_rate": 9.436882808440334e-06,
|
| 2430 |
+
"loss": 0.6165,
|
| 2431 |
+
"step": 1730
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 0.8092350746268657,
|
| 2435 |
+
"grad_norm": 0.5911843430242844,
|
| 2436 |
+
"learning_rate": 9.333919893549294e-06,
|
| 2437 |
+
"loss": 0.6099,
|
| 2438 |
+
"step": 1735
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 0.8115671641791045,
|
| 2442 |
+
"grad_norm": 0.5499239897213614,
|
| 2443 |
+
"learning_rate": 9.232038271374377e-06,
|
| 2444 |
+
"loss": 0.601,
|
| 2445 |
+
"step": 1740
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 0.8138992537313433,
|
| 2449 |
+
"grad_norm": 0.5917277497776632,
|
| 2450 |
+
"learning_rate": 9.131244006176846e-06,
|
| 2451 |
+
"loss": 0.6181,
|
| 2452 |
+
"step": 1745
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 0.816231343283582,
|
| 2456 |
+
"grad_norm": 0.5790530037092708,
|
| 2457 |
+
"learning_rate": 9.031543097495638e-06,
|
| 2458 |
+
"loss": 0.6033,
|
| 2459 |
+
"step": 1750
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 0.8185634328358209,
|
| 2463 |
+
"grad_norm": 0.556634899711528,
|
| 2464 |
+
"learning_rate": 8.93294147979023e-06,
|
| 2465 |
+
"loss": 0.5975,
|
| 2466 |
+
"step": 1755
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 0.8208955223880597,
|
| 2470 |
+
"grad_norm": 0.525574565943637,
|
| 2471 |
+
"learning_rate": 8.835445022087426e-06,
|
| 2472 |
+
"loss": 0.6194,
|
| 2473 |
+
"step": 1760
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 0.8232276119402985,
|
| 2477 |
+
"grad_norm": 0.542460572064194,
|
| 2478 |
+
"learning_rate": 8.739059527631999e-06,
|
| 2479 |
+
"loss": 0.6194,
|
| 2480 |
+
"step": 1765
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 0.8255597014925373,
|
| 2484 |
+
"grad_norm": 0.5961740451690235,
|
| 2485 |
+
"learning_rate": 8.6437907335413e-06,
|
| 2486 |
+
"loss": 0.6232,
|
| 2487 |
+
"step": 1770
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 0.8278917910447762,
|
| 2491 |
+
"grad_norm": 0.5156288871553688,
|
| 2492 |
+
"learning_rate": 8.549644310463717e-06,
|
| 2493 |
+
"loss": 0.625,
|
| 2494 |
+
"step": 1775
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 0.8302238805970149,
|
| 2498 |
+
"grad_norm": 0.5827304049610953,
|
| 2499 |
+
"learning_rate": 8.456625862241193e-06,
|
| 2500 |
+
"loss": 0.623,
|
| 2501 |
+
"step": 1780
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 0.8325559701492538,
|
| 2505 |
+
"grad_norm": 0.5473226394306203,
|
| 2506 |
+
"learning_rate": 8.364740925575643e-06,
|
| 2507 |
+
"loss": 0.6211,
|
| 2508 |
+
"step": 1785
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 0.8348880597014925,
|
| 2512 |
+
"grad_norm": 0.5663858177158172,
|
| 2513 |
+
"learning_rate": 8.273994969699394e-06,
|
| 2514 |
+
"loss": 0.6115,
|
| 2515 |
+
"step": 1790
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 0.8372201492537313,
|
| 2519 |
+
"grad_norm": 0.5667136142959207,
|
| 2520 |
+
"learning_rate": 8.184393396049675e-06,
|
| 2521 |
+
"loss": 0.6127,
|
| 2522 |
+
"step": 1795
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 0.8395522388059702,
|
| 2526 |
+
"grad_norm": 0.5603856211987771,
|
| 2527 |
+
"learning_rate": 8.095941537947057e-06,
|
| 2528 |
+
"loss": 0.6274,
|
| 2529 |
+
"step": 1800
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 0.8418843283582089,
|
| 2533 |
+
"grad_norm": 0.5594249365684848,
|
| 2534 |
+
"learning_rate": 8.008644660278051e-06,
|
| 2535 |
+
"loss": 0.5925,
|
| 2536 |
+
"step": 1805
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 0.8442164179104478,
|
| 2540 |
+
"grad_norm": 0.5635626874281433,
|
| 2541 |
+
"learning_rate": 7.922507959181673e-06,
|
| 2542 |
+
"loss": 0.6189,
|
| 2543 |
+
"step": 1810
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 0.8465485074626866,
|
| 2547 |
+
"grad_norm": 0.560313283634603,
|
| 2548 |
+
"learning_rate": 7.837536561740225e-06,
|
| 2549 |
+
"loss": 0.6097,
|
| 2550 |
+
"step": 1815
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 0.8488805970149254,
|
| 2554 |
+
"grad_norm": 0.6202519240674869,
|
| 2555 |
+
"learning_rate": 7.753735525674059e-06,
|
| 2556 |
+
"loss": 0.6174,
|
| 2557 |
+
"step": 1820
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 0.8512126865671642,
|
| 2561 |
+
"grad_norm": 0.5713772506301511,
|
| 2562 |
+
"learning_rate": 7.671109839040547e-06,
|
| 2563 |
+
"loss": 0.621,
|
| 2564 |
+
"step": 1825
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 0.8535447761194029,
|
| 2568 |
+
"grad_norm": 0.5237097428909497,
|
| 2569 |
+
"learning_rate": 7.58966441993719e-06,
|
| 2570 |
+
"loss": 0.5981,
|
| 2571 |
+
"step": 1830
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 0.8558768656716418,
|
| 2575 |
+
"grad_norm": 0.5483521318490681,
|
| 2576 |
+
"learning_rate": 7.509404116208868e-06,
|
| 2577 |
+
"loss": 0.6138,
|
| 2578 |
+
"step": 1835
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 0.8582089552238806,
|
| 2582 |
+
"grad_norm": 0.5778220550523655,
|
| 2583 |
+
"learning_rate": 7.430333705159286e-06,
|
| 2584 |
+
"loss": 0.6522,
|
| 2585 |
+
"step": 1840
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 0.8605410447761194,
|
| 2589 |
+
"grad_norm": 0.5486890952901237,
|
| 2590 |
+
"learning_rate": 7.352457893266627e-06,
|
| 2591 |
+
"loss": 0.6429,
|
| 2592 |
+
"step": 1845
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 0.8628731343283582,
|
| 2596 |
+
"grad_norm": 0.5693041316007504,
|
| 2597 |
+
"learning_rate": 7.275781315903374e-06,
|
| 2598 |
+
"loss": 0.6225,
|
| 2599 |
+
"step": 1850
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 0.8652052238805971,
|
| 2603 |
+
"grad_norm": 0.5594174594138699,
|
| 2604 |
+
"learning_rate": 7.20030853706046e-06,
|
| 2605 |
+
"loss": 0.6174,
|
| 2606 |
+
"step": 1855
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 0.8675373134328358,
|
| 2610 |
+
"grad_norm": 0.5408573140361542,
|
| 2611 |
+
"learning_rate": 7.126044049075548e-06,
|
| 2612 |
+
"loss": 0.6239,
|
| 2613 |
+
"step": 1860
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 0.8698694029850746,
|
| 2617 |
+
"grad_norm": 0.5625658917979015,
|
| 2618 |
+
"learning_rate": 7.052992272365681e-06,
|
| 2619 |
+
"loss": 0.6149,
|
| 2620 |
+
"step": 1865
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 0.8722014925373134,
|
| 2624 |
+
"grad_norm": 0.5795869183892941,
|
| 2625 |
+
"learning_rate": 6.9811575551641224e-06,
|
| 2626 |
+
"loss": 0.6014,
|
| 2627 |
+
"step": 1870
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 0.8745335820895522,
|
| 2631 |
+
"grad_norm": 0.5908193742703406,
|
| 2632 |
+
"learning_rate": 6.910544173261588e-06,
|
| 2633 |
+
"loss": 0.6047,
|
| 2634 |
+
"step": 1875
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 0.8768656716417911,
|
| 2638 |
+
"grad_norm": 0.5468229515625878,
|
| 2639 |
+
"learning_rate": 6.8411563297516995e-06,
|
| 2640 |
+
"loss": 0.5956,
|
| 2641 |
+
"step": 1880
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 0.8791977611940298,
|
| 2645 |
+
"grad_norm": 0.5746605159059298,
|
| 2646 |
+
"learning_rate": 6.772998154780832e-06,
|
| 2647 |
+
"loss": 0.612,
|
| 2648 |
+
"step": 1885
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 0.8815298507462687,
|
| 2652 |
+
"grad_norm": 0.5756230793946113,
|
| 2653 |
+
"learning_rate": 6.706073705302254e-06,
|
| 2654 |
+
"loss": 0.6145,
|
| 2655 |
+
"step": 1890
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 0.8838619402985075,
|
| 2659 |
+
"grad_norm": 0.5662725596320226,
|
| 2660 |
+
"learning_rate": 6.6403869648346634e-06,
|
| 2661 |
+
"loss": 0.6116,
|
| 2662 |
+
"step": 1895
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 0.8861940298507462,
|
| 2666 |
+
"grad_norm": 0.6638472478598192,
|
| 2667 |
+
"learning_rate": 6.575941843225068e-06,
|
| 2668 |
+
"loss": 0.5928,
|
| 2669 |
+
"step": 1900
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 0.8885261194029851,
|
| 2673 |
+
"grad_norm": 0.5392491951943847,
|
| 2674 |
+
"learning_rate": 6.5127421764160685e-06,
|
| 2675 |
+
"loss": 0.5994,
|
| 2676 |
+
"step": 1905
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 0.8908582089552238,
|
| 2680 |
+
"grad_norm": 0.5501297394583888,
|
| 2681 |
+
"learning_rate": 6.450791726217538e-06,
|
| 2682 |
+
"loss": 0.623,
|
| 2683 |
+
"step": 1910
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 0.8931902985074627,
|
| 2687 |
+
"grad_norm": 0.5561658935766048,
|
| 2688 |
+
"learning_rate": 6.390094180082694e-06,
|
| 2689 |
+
"loss": 0.6137,
|
| 2690 |
+
"step": 1915
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 0.8955223880597015,
|
| 2694 |
+
"grad_norm": 0.5486158970271028,
|
| 2695 |
+
"learning_rate": 6.330653150888617e-06,
|
| 2696 |
+
"loss": 0.6153,
|
| 2697 |
+
"step": 1920
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 0.8978544776119403,
|
| 2701 |
+
"grad_norm": 0.5624023243356174,
|
| 2702 |
+
"learning_rate": 6.272472176721207e-06,
|
| 2703 |
+
"loss": 0.6207,
|
| 2704 |
+
"step": 1925
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 0.9001865671641791,
|
| 2708 |
+
"grad_norm": 0.5469455949584299,
|
| 2709 |
+
"learning_rate": 6.215554720664598e-06,
|
| 2710 |
+
"loss": 0.6276,
|
| 2711 |
+
"step": 1930
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 0.902518656716418,
|
| 2715 |
+
"grad_norm": 0.5335851595481946,
|
| 2716 |
+
"learning_rate": 6.159904170594982e-06,
|
| 2717 |
+
"loss": 0.6151,
|
| 2718 |
+
"step": 1935
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 0.9048507462686567,
|
| 2722 |
+
"grad_norm": 0.5451933732807914,
|
| 2723 |
+
"learning_rate": 6.105523838979022e-06,
|
| 2724 |
+
"loss": 0.6302,
|
| 2725 |
+
"step": 1940
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 0.9071828358208955,
|
| 2729 |
+
"grad_norm": 0.5362545841210454,
|
| 2730 |
+
"learning_rate": 6.052416962676621e-06,
|
| 2731 |
+
"loss": 0.6058,
|
| 2732 |
+
"step": 1945
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 0.9095149253731343,
|
| 2736 |
+
"grad_norm": 0.5230617545385562,
|
| 2737 |
+
"learning_rate": 6.000586702748301e-06,
|
| 2738 |
+
"loss": 0.603,
|
| 2739 |
+
"step": 1950
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 0.9118470149253731,
|
| 2743 |
+
"grad_norm": 0.582021589520335,
|
| 2744 |
+
"learning_rate": 5.950036144267021e-06,
|
| 2745 |
+
"loss": 0.6111,
|
| 2746 |
+
"step": 1955
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 0.914179104477612,
|
| 2750 |
+
"grad_norm": 0.519728015454707,
|
| 2751 |
+
"learning_rate": 5.900768296134551e-06,
|
| 2752 |
+
"loss": 0.6049,
|
| 2753 |
+
"step": 1960
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 0.9165111940298507,
|
| 2757 |
+
"grad_norm": 0.592109342645659,
|
| 2758 |
+
"learning_rate": 5.852786090902383e-06,
|
| 2759 |
+
"loss": 0.6199,
|
| 2760 |
+
"step": 1965
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 0.9188432835820896,
|
| 2764 |
+
"grad_norm": 0.5517752644641193,
|
| 2765 |
+
"learning_rate": 5.8060923845971825e-06,
|
| 2766 |
+
"loss": 0.6113,
|
| 2767 |
+
"step": 1970
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 0.9211753731343284,
|
| 2771 |
+
"grad_norm": 0.5374604103547962,
|
| 2772 |
+
"learning_rate": 5.760689956550763e-06,
|
| 2773 |
+
"loss": 0.6051,
|
| 2774 |
+
"step": 1975
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 0.9235074626865671,
|
| 2778 |
+
"grad_norm": 0.5317691219584879,
|
| 2779 |
+
"learning_rate": 5.7165815092346825e-06,
|
| 2780 |
+
"loss": 0.5952,
|
| 2781 |
+
"step": 1980
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 0.925839552238806,
|
| 2785 |
+
"grad_norm": 0.5553747148697364,
|
| 2786 |
+
"learning_rate": 5.673769668099364e-06,
|
| 2787 |
+
"loss": 0.6088,
|
| 2788 |
+
"step": 1985
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 0.9281716417910447,
|
| 2792 |
+
"grad_norm": 0.5606224712873208,
|
| 2793 |
+
"learning_rate": 5.632256981417845e-06,
|
| 2794 |
+
"loss": 0.6134,
|
| 2795 |
+
"step": 1990
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 0.9305037313432836,
|
| 2799 |
+
"grad_norm": 0.5375797889998934,
|
| 2800 |
+
"learning_rate": 5.59204592013407e-06,
|
| 2801 |
+
"loss": 0.5973,
|
| 2802 |
+
"step": 1995
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 0.9328358208955224,
|
| 2806 |
+
"grad_norm": 0.5508292133907728,
|
| 2807 |
+
"learning_rate": 5.553138877715833e-06,
|
| 2808 |
+
"loss": 0.6298,
|
| 2809 |
+
"step": 2000
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 0.9351679104477612,
|
| 2813 |
+
"grad_norm": 0.5407565378019897,
|
| 2814 |
+
"learning_rate": 5.515538170012309e-06,
|
| 2815 |
+
"loss": 0.6085,
|
| 2816 |
+
"step": 2005
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 0.9375,
|
| 2820 |
+
"grad_norm": 0.5516140680428332,
|
| 2821 |
+
"learning_rate": 5.479246035116201e-06,
|
| 2822 |
+
"loss": 0.5998,
|
| 2823 |
+
"step": 2010
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 0.9398320895522388,
|
| 2827 |
+
"grad_norm": 0.5871082563627422,
|
| 2828 |
+
"learning_rate": 5.444264633230531e-06,
|
| 2829 |
+
"loss": 0.6146,
|
| 2830 |
+
"step": 2015
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 0.9421641791044776,
|
| 2834 |
+
"grad_norm": 0.5579174490831493,
|
| 2835 |
+
"learning_rate": 5.410596046540051e-06,
|
| 2836 |
+
"loss": 0.6008,
|
| 2837 |
+
"step": 2020
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 0.9444962686567164,
|
| 2841 |
+
"grad_norm": 0.5535279013398093,
|
| 2842 |
+
"learning_rate": 5.378242279087314e-06,
|
| 2843 |
+
"loss": 0.6071,
|
| 2844 |
+
"step": 2025
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 0.9468283582089553,
|
| 2848 |
+
"grad_norm": 0.5542818715981681,
|
| 2849 |
+
"learning_rate": 5.347205256653387e-06,
|
| 2850 |
+
"loss": 0.6131,
|
| 2851 |
+
"step": 2030
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 0.949160447761194,
|
| 2855 |
+
"grad_norm": 0.5507958590165508,
|
| 2856 |
+
"learning_rate": 5.317486826643219e-06,
|
| 2857 |
+
"loss": 0.5896,
|
| 2858 |
+
"step": 2035
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 0.9514925373134329,
|
| 2862 |
+
"grad_norm": 0.5386294413734031,
|
| 2863 |
+
"learning_rate": 5.28908875797568e-06,
|
| 2864 |
+
"loss": 0.6164,
|
| 2865 |
+
"step": 2040
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 0.9538246268656716,
|
| 2869 |
+
"grad_norm": 0.5657594223236935,
|
| 2870 |
+
"learning_rate": 5.262012740978269e-06,
|
| 2871 |
+
"loss": 0.6171,
|
| 2872 |
+
"step": 2045
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 0.9561567164179104,
|
| 2876 |
+
"grad_norm": 0.545809161265838,
|
| 2877 |
+
"learning_rate": 5.236260387286509e-06,
|
| 2878 |
+
"loss": 0.6014,
|
| 2879 |
+
"step": 2050
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 0.9584888059701493,
|
| 2883 |
+
"grad_norm": 0.5400541424481552,
|
| 2884 |
+
"learning_rate": 5.2118332297480105e-06,
|
| 2885 |
+
"loss": 0.6195,
|
| 2886 |
+
"step": 2055
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 0.960820895522388,
|
| 2890 |
+
"grad_norm": 0.5301287222550025,
|
| 2891 |
+
"learning_rate": 5.1887327223312296e-06,
|
| 2892 |
+
"loss": 0.5894,
|
| 2893 |
+
"step": 2060
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 0.9631529850746269,
|
| 2897 |
+
"grad_norm": 0.5826747467153147,
|
| 2898 |
+
"learning_rate": 5.166960240038937e-06,
|
| 2899 |
+
"loss": 0.6103,
|
| 2900 |
+
"step": 2065
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 0.9654850746268657,
|
| 2904 |
+
"grad_norm": 0.5624482325072565,
|
| 2905 |
+
"learning_rate": 5.1465170788263595e-06,
|
| 2906 |
+
"loss": 0.6205,
|
| 2907 |
+
"step": 2070
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 0.9678171641791045,
|
| 2911 |
+
"grad_norm": 0.5710832690138076,
|
| 2912 |
+
"learning_rate": 5.1274044555240525e-06,
|
| 2913 |
+
"loss": 0.6083,
|
| 2914 |
+
"step": 2075
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 0.9701492537313433,
|
| 2918 |
+
"grad_norm": 0.5427558552652122,
|
| 2919 |
+
"learning_rate": 5.109623507765466e-06,
|
| 2920 |
+
"loss": 0.6113,
|
| 2921 |
+
"step": 2080
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 0.972481343283582,
|
| 2925 |
+
"grad_norm": 0.5356447254957378,
|
| 2926 |
+
"learning_rate": 5.093175293919228e-06,
|
| 2927 |
+
"loss": 0.6138,
|
| 2928 |
+
"step": 2085
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 0.9748134328358209,
|
| 2932 |
+
"grad_norm": 0.554507085372338,
|
| 2933 |
+
"learning_rate": 5.07806079302615e-06,
|
| 2934 |
+
"loss": 0.6209,
|
| 2935 |
+
"step": 2090
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 0.9771455223880597,
|
| 2939 |
+
"grad_norm": 0.5347894717484287,
|
| 2940 |
+
"learning_rate": 5.064280904740953e-06,
|
| 2941 |
+
"loss": 0.6187,
|
| 2942 |
+
"step": 2095
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 0.9794776119402985,
|
| 2946 |
+
"grad_norm": 0.548171827658312,
|
| 2947 |
+
"learning_rate": 5.051836449278715e-06,
|
| 2948 |
+
"loss": 0.6069,
|
| 2949 |
+
"step": 2100
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 0.9818097014925373,
|
| 2953 |
+
"grad_norm": 0.562019041400098,
|
| 2954 |
+
"learning_rate": 5.040728167366057e-06,
|
| 2955 |
+
"loss": 0.6028,
|
| 2956 |
+
"step": 2105
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 0.9841417910447762,
|
| 2960 |
+
"grad_norm": 0.5554495773144189,
|
| 2961 |
+
"learning_rate": 5.030956720197035e-06,
|
| 2962 |
+
"loss": 0.6039,
|
| 2963 |
+
"step": 2110
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 0.9864738805970149,
|
| 2967 |
+
"grad_norm": 0.5847204114674615,
|
| 2968 |
+
"learning_rate": 5.022522689393809e-06,
|
| 2969 |
+
"loss": 0.6377,
|
| 2970 |
+
"step": 2115
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 0.9888059701492538,
|
| 2974 |
+
"grad_norm": 0.5293713892212091,
|
| 2975 |
+
"learning_rate": 5.015426576972003e-06,
|
| 2976 |
+
"loss": 0.6099,
|
| 2977 |
+
"step": 2120
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 0.9911380597014925,
|
| 2981 |
+
"grad_norm": 0.5377836319550625,
|
| 2982 |
+
"learning_rate": 5.009668805310832e-06,
|
| 2983 |
+
"loss": 0.6063,
|
| 2984 |
+
"step": 2125
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 0.9934701492537313,
|
| 2988 |
+
"grad_norm": 0.5661212299229181,
|
| 2989 |
+
"learning_rate": 5.005249717127964e-06,
|
| 2990 |
+
"loss": 0.6199,
|
| 2991 |
+
"step": 2130
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 0.9958022388059702,
|
| 2995 |
+
"grad_norm": 0.5383491265897701,
|
| 2996 |
+
"learning_rate": 5.002169575459111e-06,
|
| 2997 |
+
"loss": 0.6107,
|
| 2998 |
+
"step": 2135
|
| 2999 |
+
},
|
| 3000 |
+
{
|
| 3001 |
+
"epoch": 0.9981343283582089,
|
| 3002 |
+
"grad_norm": 0.5598221695919163,
|
| 3003 |
+
"learning_rate": 5.000428563642382e-06,
|
| 3004 |
+
"loss": 0.6166,
|
| 3005 |
+
"step": 2140
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 1.0,
|
| 3009 |
+
"step": 2144,
|
| 3010 |
+
"total_flos": 241268013662208.0,
|
| 3011 |
+
"train_loss": 0.6583015036894314,
|
| 3012 |
+
"train_runtime": 13168.5767,
|
| 3013 |
+
"train_samples_per_second": 2.605,
|
| 3014 |
+
"train_steps_per_second": 0.163
|
| 3015 |
+
}
|
| 3016 |
+
],
|
| 3017 |
+
"logging_steps": 5,
|
| 3018 |
+
"max_steps": 2144,
|
| 3019 |
+
"num_input_tokens_seen": 0,
|
| 3020 |
+
"num_train_epochs": 1,
|
| 3021 |
+
"save_steps": 100,
|
| 3022 |
+
"stateful_callbacks": {
|
| 3023 |
+
"TrainerControl": {
|
| 3024 |
+
"args": {
|
| 3025 |
+
"should_epoch_stop": false,
|
| 3026 |
+
"should_evaluate": false,
|
| 3027 |
+
"should_log": false,
|
| 3028 |
+
"should_save": true,
|
| 3029 |
+
"should_training_stop": true
|
| 3030 |
+
},
|
| 3031 |
+
"attributes": {}
|
| 3032 |
+
}
|
| 3033 |
+
},
|
| 3034 |
+
"total_flos": 241268013662208.0,
|
| 3035 |
+
"train_batch_size": 16,
|
| 3036 |
+
"trial_name": null,
|
| 3037 |
+
"trial_params": null
|
| 3038 |
+
}
|