Model save
Browse files- README.md +67 -0
- all_results.json +8 -0
- generation_config.json +6 -0
- train_results.json +8 -0
- trainer_state.json +1417 -0
README.md
ADDED
|
@@ -0,0 +1,67 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: transformers
|
| 3 |
+
model_name: Qwen2.5-3B-GRPO-Natural-Reasoning
|
| 4 |
+
tags:
|
| 5 |
+
- generated_from_trainer
|
| 6 |
+
- trl
|
| 7 |
+
- grpo
|
| 8 |
+
licence: license
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# Model Card for Qwen2.5-3B-GRPO-Natural-Reasoning
|
| 12 |
+
|
| 13 |
+
This model is a fine-tuned version of [None](https://huggingface.co/None).
|
| 14 |
+
It has been trained using [TRL](https://github.com/huggingface/trl).
|
| 15 |
+
|
| 16 |
+
## Quick start
|
| 17 |
+
|
| 18 |
+
```python
|
| 19 |
+
from transformers import pipeline
|
| 20 |
+
|
| 21 |
+
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?"
|
| 22 |
+
generator = pipeline("text-generation", model="qingyangzhang/Qwen2.5-3B-GRPO-Natural-Reasoning", device="cuda")
|
| 23 |
+
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
|
| 24 |
+
print(output["generated_text"])
|
| 25 |
+
```
|
| 26 |
+
|
| 27 |
+
## Training procedure
|
| 28 |
+
|
| 29 |
+
[<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/zqyoung1127-tianjin-university/huggingface/runs/o8d2wzv6)
|
| 30 |
+
|
| 31 |
+
|
| 32 |
+
This model was trained with GRPO, a method introduced in [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://huggingface.co/papers/2402.03300).
|
| 33 |
+
|
| 34 |
+
### Framework versions
|
| 35 |
+
|
| 36 |
+
- TRL: 0.14.0
|
| 37 |
+
- Transformers: 4.48.3
|
| 38 |
+
- Pytorch: 2.5.1
|
| 39 |
+
- Datasets: 3.1.0
|
| 40 |
+
- Tokenizers: 0.21.0
|
| 41 |
+
|
| 42 |
+
## Citations
|
| 43 |
+
|
| 44 |
+
Cite GRPO as:
|
| 45 |
+
|
| 46 |
+
```bibtex
|
| 47 |
+
@article{zhihong2024deepseekmath,
|
| 48 |
+
title = {{DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models}},
|
| 49 |
+
author = {Zhihong Shao and Peiyi Wang and Qihao Zhu and Runxin Xu and Junxiao Song and Mingchuan Zhang and Y. K. Li and Y. Wu and Daya Guo},
|
| 50 |
+
year = 2024,
|
| 51 |
+
eprint = {arXiv:2402.03300},
|
| 52 |
+
}
|
| 53 |
+
|
| 54 |
+
```
|
| 55 |
+
|
| 56 |
+
Cite TRL as:
|
| 57 |
+
|
| 58 |
+
```bibtex
|
| 59 |
+
@misc{vonwerra2022trl,
|
| 60 |
+
title = {{TRL: Transformer Reinforcement Learning}},
|
| 61 |
+
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},
|
| 62 |
+
year = 2020,
|
| 63 |
+
journal = {GitHub repository},
|
| 64 |
+
publisher = {GitHub},
|
| 65 |
+
howpublished = {\url{https://github.com/huggingface/trl}}
|
| 66 |
+
}
|
| 67 |
+
```
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"total_flos": 0.0,
|
| 3 |
+
"train_loss": 2.3229669920965534e-08,
|
| 4 |
+
"train_runtime": 35950.9286,
|
| 5 |
+
"train_samples": 12058,
|
| 6 |
+
"train_samples_per_second": 0.335,
|
| 7 |
+
"train_steps_per_second": 0.003
|
| 8 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token_id": 151643,
|
| 3 |
+
"eos_token_id": 151643,
|
| 4 |
+
"max_new_tokens": 2048,
|
| 5 |
+
"transformers_version": "4.48.3"
|
| 6 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"total_flos": 0.0,
|
| 3 |
+
"train_loss": 2.3229669920965534e-08,
|
| 4 |
+
"train_runtime": 35950.9286,
|
| 5 |
+
"train_samples": 12058,
|
| 6 |
+
"train_samples_per_second": 0.335,
|
| 7 |
+
"train_steps_per_second": 0.003
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1417 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9950248756218906,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 125,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"completion_length": 376.33854484558105,
|
| 13 |
+
"epoch": 0.007960199004975124,
|
| 14 |
+
"grad_norm": 0.21353177726268768,
|
| 15 |
+
"learning_rate": 1e-06,
|
| 16 |
+
"loss": 0.0,
|
| 17 |
+
"reward": 0.2578125004656613,
|
| 18 |
+
"reward_std": 0.2655366016551852,
|
| 19 |
+
"rewards/accuracy_reward": 0.2578125004656613,
|
| 20 |
+
"step": 1
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"completion_length": 359.9583339691162,
|
| 24 |
+
"epoch": 0.015920398009950248,
|
| 25 |
+
"grad_norm": 0.02424778789281845,
|
| 26 |
+
"learning_rate": 1e-06,
|
| 27 |
+
"loss": 0.0,
|
| 28 |
+
"reward": 0.2942708353511989,
|
| 29 |
+
"reward_std": 0.26902162190526724,
|
| 30 |
+
"rewards/accuracy_reward": 0.2942708353511989,
|
| 31 |
+
"step": 2
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"completion_length": 372.2769145965576,
|
| 35 |
+
"epoch": 0.023880597014925373,
|
| 36 |
+
"grad_norm": 0.08308012783527374,
|
| 37 |
+
"learning_rate": 1e-06,
|
| 38 |
+
"loss": 0.0,
|
| 39 |
+
"reward": 0.25000000349245965,
|
| 40 |
+
"reward_std": 0.2666853931732476,
|
| 41 |
+
"rewards/accuracy_reward": 0.25000000349245965,
|
| 42 |
+
"step": 3
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"completion_length": 366.7196235656738,
|
| 46 |
+
"epoch": 0.031840796019900496,
|
| 47 |
+
"grad_norm": 0.035361409187316895,
|
| 48 |
+
"learning_rate": 1e-06,
|
| 49 |
+
"loss": 0.0,
|
| 50 |
+
"reward": 0.30295139644294977,
|
| 51 |
+
"reward_std": 0.2662010188214481,
|
| 52 |
+
"rewards/accuracy_reward": 0.30295139644294977,
|
| 53 |
+
"step": 4
|
| 54 |
+
},
|
| 55 |
+
{
|
| 56 |
+
"completion_length": 358.4774341583252,
|
| 57 |
+
"epoch": 0.03980099502487562,
|
| 58 |
+
"grad_norm": 0.05183703452348709,
|
| 59 |
+
"learning_rate": 1e-06,
|
| 60 |
+
"loss": 0.0,
|
| 61 |
+
"reward": 0.3211805585306138,
|
| 62 |
+
"reward_std": 0.3101580459624529,
|
| 63 |
+
"rewards/accuracy_reward": 0.3211805585306138,
|
| 64 |
+
"step": 5
|
| 65 |
+
},
|
| 66 |
+
{
|
| 67 |
+
"completion_length": 359.42882347106934,
|
| 68 |
+
"epoch": 0.04776119402985075,
|
| 69 |
+
"grad_norm": 0.024956537410616875,
|
| 70 |
+
"learning_rate": 1e-06,
|
| 71 |
+
"loss": 0.0,
|
| 72 |
+
"reward": 0.37065972574055195,
|
| 73 |
+
"reward_std": 0.3160043712705374,
|
| 74 |
+
"rewards/accuracy_reward": 0.37065972574055195,
|
| 75 |
+
"step": 6
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"completion_length": 386.2057285308838,
|
| 79 |
+
"epoch": 0.05572139303482587,
|
| 80 |
+
"grad_norm": 0.02574196085333824,
|
| 81 |
+
"learning_rate": 1e-06,
|
| 82 |
+
"loss": 0.0,
|
| 83 |
+
"reward": 0.33159722574055195,
|
| 84 |
+
"reward_std": 0.30505089182406664,
|
| 85 |
+
"rewards/accuracy_reward": 0.33159722574055195,
|
| 86 |
+
"step": 7
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"completion_length": 374.69618797302246,
|
| 90 |
+
"epoch": 0.06368159203980099,
|
| 91 |
+
"grad_norm": 0.03463900834321976,
|
| 92 |
+
"learning_rate": 1e-06,
|
| 93 |
+
"loss": 0.0,
|
| 94 |
+
"reward": 0.3211805559694767,
|
| 95 |
+
"reward_std": 0.2687861230224371,
|
| 96 |
+
"rewards/accuracy_reward": 0.3211805559694767,
|
| 97 |
+
"step": 8
|
| 98 |
+
},
|
| 99 |
+
{
|
| 100 |
+
"completion_length": 366.6770896911621,
|
| 101 |
+
"epoch": 0.07164179104477612,
|
| 102 |
+
"grad_norm": 0.09361224621534348,
|
| 103 |
+
"learning_rate": 1e-06,
|
| 104 |
+
"loss": 0.0,
|
| 105 |
+
"reward": 0.41319445334374905,
|
| 106 |
+
"reward_std": 0.3116005442570895,
|
| 107 |
+
"rewards/accuracy_reward": 0.41319445334374905,
|
| 108 |
+
"step": 9
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"completion_length": 391.65625381469727,
|
| 112 |
+
"epoch": 0.07960199004975124,
|
| 113 |
+
"grad_norm": 0.01774750091135502,
|
| 114 |
+
"learning_rate": 1e-06,
|
| 115 |
+
"loss": 0.0,
|
| 116 |
+
"reward": 0.37326389690861106,
|
| 117 |
+
"reward_std": 0.3113201856613159,
|
| 118 |
+
"rewards/accuracy_reward": 0.37326389690861106,
|
| 119 |
+
"step": 10
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"completion_length": 406.9496593475342,
|
| 123 |
+
"epoch": 0.08756218905472637,
|
| 124 |
+
"grad_norm": 0.1700211465358734,
|
| 125 |
+
"learning_rate": 1e-06,
|
| 126 |
+
"loss": 0.0,
|
| 127 |
+
"reward": 0.36979166616220027,
|
| 128 |
+
"reward_std": 0.3051337222568691,
|
| 129 |
+
"rewards/accuracy_reward": 0.36979166616220027,
|
| 130 |
+
"step": 11
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"completion_length": 407.82986640930176,
|
| 134 |
+
"epoch": 0.0955223880597015,
|
| 135 |
+
"grad_norm": 0.01841021701693535,
|
| 136 |
+
"learning_rate": 1e-06,
|
| 137 |
+
"loss": 0.0,
|
| 138 |
+
"reward": 0.3515625,
|
| 139 |
+
"reward_std": 0.27109134290367365,
|
| 140 |
+
"rewards/accuracy_reward": 0.3515625,
|
| 141 |
+
"step": 12
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"completion_length": 438.81944847106934,
|
| 145 |
+
"epoch": 0.10348258706467661,
|
| 146 |
+
"grad_norm": 0.014034698717296124,
|
| 147 |
+
"learning_rate": 1e-06,
|
| 148 |
+
"loss": 0.0,
|
| 149 |
+
"reward": 0.4001736156642437,
|
| 150 |
+
"reward_std": 0.29678336903452873,
|
| 151 |
+
"rewards/accuracy_reward": 0.4001736156642437,
|
| 152 |
+
"step": 13
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"completion_length": 419.9982662200928,
|
| 156 |
+
"epoch": 0.11144278606965174,
|
| 157 |
+
"grad_norm": 0.02144305780529976,
|
| 158 |
+
"learning_rate": 1e-06,
|
| 159 |
+
"loss": 0.0,
|
| 160 |
+
"reward": 0.33767361380159855,
|
| 161 |
+
"reward_std": 0.2604841380380094,
|
| 162 |
+
"rewards/accuracy_reward": 0.33767361380159855,
|
| 163 |
+
"step": 14
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"completion_length": 425.06771659851074,
|
| 167 |
+
"epoch": 0.11940298507462686,
|
| 168 |
+
"grad_norm": 0.010681218467652798,
|
| 169 |
+
"learning_rate": 1e-06,
|
| 170 |
+
"loss": 0.0,
|
| 171 |
+
"reward": 0.40451388992369175,
|
| 172 |
+
"reward_std": 0.25912062590941787,
|
| 173 |
+
"rewards/accuracy_reward": 0.40451388992369175,
|
| 174 |
+
"step": 15
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"completion_length": 431.0529499053955,
|
| 178 |
+
"epoch": 0.12736318407960198,
|
| 179 |
+
"grad_norm": 0.009904686361551285,
|
| 180 |
+
"learning_rate": 1e-06,
|
| 181 |
+
"loss": 0.0,
|
| 182 |
+
"reward": 0.4123263955116272,
|
| 183 |
+
"reward_std": 0.254815224558115,
|
| 184 |
+
"rewards/accuracy_reward": 0.4123263955116272,
|
| 185 |
+
"step": 16
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"completion_length": 431.28211784362793,
|
| 189 |
+
"epoch": 0.13532338308457711,
|
| 190 |
+
"grad_norm": 0.010264623910188675,
|
| 191 |
+
"learning_rate": 1e-06,
|
| 192 |
+
"loss": 0.0,
|
| 193 |
+
"reward": 0.4027777807787061,
|
| 194 |
+
"reward_std": 0.22909809951670468,
|
| 195 |
+
"rewards/accuracy_reward": 0.4027777807787061,
|
| 196 |
+
"step": 17
|
| 197 |
+
},
|
| 198 |
+
{
|
| 199 |
+
"completion_length": 410.7699718475342,
|
| 200 |
+
"epoch": 0.14328358208955225,
|
| 201 |
+
"grad_norm": 0.012061933986842632,
|
| 202 |
+
"learning_rate": 1e-06,
|
| 203 |
+
"loss": 0.0,
|
| 204 |
+
"reward": 0.44531250186264515,
|
| 205 |
+
"reward_std": 0.258999613346532,
|
| 206 |
+
"rewards/accuracy_reward": 0.44531250186264515,
|
| 207 |
+
"step": 18
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"completion_length": 461.3133716583252,
|
| 211 |
+
"epoch": 0.15124378109452735,
|
| 212 |
+
"grad_norm": 0.008844373747706413,
|
| 213 |
+
"learning_rate": 1e-06,
|
| 214 |
+
"loss": 0.0,
|
| 215 |
+
"reward": 0.3836805624887347,
|
| 216 |
+
"reward_std": 0.232815052382648,
|
| 217 |
+
"rewards/accuracy_reward": 0.3836805624887347,
|
| 218 |
+
"step": 19
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"completion_length": 434.8680610656738,
|
| 222 |
+
"epoch": 0.15920398009950248,
|
| 223 |
+
"grad_norm": 0.008117014542222023,
|
| 224 |
+
"learning_rate": 1e-06,
|
| 225 |
+
"loss": 0.0,
|
| 226 |
+
"reward": 0.4652777798473835,
|
| 227 |
+
"reward_std": 0.24379803240299225,
|
| 228 |
+
"rewards/accuracy_reward": 0.4652777798473835,
|
| 229 |
+
"step": 20
|
| 230 |
+
},
|
| 231 |
+
{
|
| 232 |
+
"completion_length": 432.0590362548828,
|
| 233 |
+
"epoch": 0.16716417910447762,
|
| 234 |
+
"grad_norm": 0.020449288189411163,
|
| 235 |
+
"learning_rate": 1e-06,
|
| 236 |
+
"loss": 0.0,
|
| 237 |
+
"reward": 0.4314236156642437,
|
| 238 |
+
"reward_std": 0.26850314904004335,
|
| 239 |
+
"rewards/accuracy_reward": 0.4314236156642437,
|
| 240 |
+
"step": 21
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"completion_length": 444.558162689209,
|
| 244 |
+
"epoch": 0.17512437810945275,
|
| 245 |
+
"grad_norm": 0.008622939698398113,
|
| 246 |
+
"learning_rate": 1e-06,
|
| 247 |
+
"loss": 0.0,
|
| 248 |
+
"reward": 0.4166666679084301,
|
| 249 |
+
"reward_std": 0.22840349189937115,
|
| 250 |
+
"rewards/accuracy_reward": 0.4166666679084301,
|
| 251 |
+
"step": 22
|
| 252 |
+
},
|
| 253 |
+
{
|
| 254 |
+
"completion_length": 468.4470462799072,
|
| 255 |
+
"epoch": 0.18308457711442785,
|
| 256 |
+
"grad_norm": 0.008969324640929699,
|
| 257 |
+
"learning_rate": 1e-06,
|
| 258 |
+
"loss": 0.0,
|
| 259 |
+
"reward": 0.4010416716337204,
|
| 260 |
+
"reward_std": 0.2720437094103545,
|
| 261 |
+
"rewards/accuracy_reward": 0.4010416716337204,
|
| 262 |
+
"step": 23
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"completion_length": 432.6501770019531,
|
| 266 |
+
"epoch": 0.191044776119403,
|
| 267 |
+
"grad_norm": 0.009219987317919731,
|
| 268 |
+
"learning_rate": 1e-06,
|
| 269 |
+
"loss": 0.0,
|
| 270 |
+
"reward": 0.493055559694767,
|
| 271 |
+
"reward_std": 0.277205478399992,
|
| 272 |
+
"rewards/accuracy_reward": 0.493055559694767,
|
| 273 |
+
"step": 24
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"completion_length": 462.31423568725586,
|
| 277 |
+
"epoch": 0.19900497512437812,
|
| 278 |
+
"grad_norm": 0.00965342577546835,
|
| 279 |
+
"learning_rate": 1e-06,
|
| 280 |
+
"loss": 0.0,
|
| 281 |
+
"reward": 0.45052083767950535,
|
| 282 |
+
"reward_std": 0.27437643241137266,
|
| 283 |
+
"rewards/accuracy_reward": 0.45052083767950535,
|
| 284 |
+
"step": 25
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"completion_length": 483.3194465637207,
|
| 288 |
+
"epoch": 0.20696517412935322,
|
| 289 |
+
"grad_norm": 0.006745666265487671,
|
| 290 |
+
"learning_rate": 1e-06,
|
| 291 |
+
"loss": 0.0,
|
| 292 |
+
"reward": 0.4088541737291962,
|
| 293 |
+
"reward_std": 0.1909226190764457,
|
| 294 |
+
"rewards/accuracy_reward": 0.4088541737291962,
|
| 295 |
+
"step": 26
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"completion_length": 457.39236640930176,
|
| 299 |
+
"epoch": 0.21492537313432836,
|
| 300 |
+
"grad_norm": 0.008856388740241528,
|
| 301 |
+
"learning_rate": 1e-06,
|
| 302 |
+
"loss": 0.0,
|
| 303 |
+
"reward": 0.5269097276031971,
|
| 304 |
+
"reward_std": 0.23024124139919877,
|
| 305 |
+
"rewards/accuracy_reward": 0.5269097276031971,
|
| 306 |
+
"step": 27
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"completion_length": 474.4479236602783,
|
| 310 |
+
"epoch": 0.2228855721393035,
|
| 311 |
+
"grad_norm": 0.007453792728483677,
|
| 312 |
+
"learning_rate": 1e-06,
|
| 313 |
+
"loss": 0.0,
|
| 314 |
+
"reward": 0.43750000558793545,
|
| 315 |
+
"reward_std": 0.231502128764987,
|
| 316 |
+
"rewards/accuracy_reward": 0.43750000558793545,
|
| 317 |
+
"step": 28
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"completion_length": 485.6154537200928,
|
| 321 |
+
"epoch": 0.2308457711442786,
|
| 322 |
+
"grad_norm": 0.009123899042606354,
|
| 323 |
+
"learning_rate": 1e-06,
|
| 324 |
+
"loss": 0.0,
|
| 325 |
+
"reward": 0.41059028171002865,
|
| 326 |
+
"reward_std": 0.2188729103654623,
|
| 327 |
+
"rewards/accuracy_reward": 0.41059028171002865,
|
| 328 |
+
"step": 29
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"completion_length": 510.6953182220459,
|
| 332 |
+
"epoch": 0.23880597014925373,
|
| 333 |
+
"grad_norm": 0.008295816369354725,
|
| 334 |
+
"learning_rate": 1e-06,
|
| 335 |
+
"loss": 0.0,
|
| 336 |
+
"reward": 0.39930556435137987,
|
| 337 |
+
"reward_std": 0.2771527715958655,
|
| 338 |
+
"rewards/accuracy_reward": 0.39930556435137987,
|
| 339 |
+
"step": 30
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"completion_length": 480.7777843475342,
|
| 343 |
+
"epoch": 0.24676616915422886,
|
| 344 |
+
"grad_norm": 0.007505806162953377,
|
| 345 |
+
"learning_rate": 1e-06,
|
| 346 |
+
"loss": 0.0,
|
| 347 |
+
"reward": 0.4817708386108279,
|
| 348 |
+
"reward_std": 0.23485751589760184,
|
| 349 |
+
"rewards/accuracy_reward": 0.4817708386108279,
|
| 350 |
+
"step": 31
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"completion_length": 515.5642375946045,
|
| 354 |
+
"epoch": 0.25472636815920396,
|
| 355 |
+
"grad_norm": 0.008023527450859547,
|
| 356 |
+
"learning_rate": 1e-06,
|
| 357 |
+
"loss": 0.0,
|
| 358 |
+
"reward": 0.4557291707023978,
|
| 359 |
+
"reward_std": 0.24294046964496374,
|
| 360 |
+
"rewards/accuracy_reward": 0.4557291707023978,
|
| 361 |
+
"step": 32
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"completion_length": 483.67101097106934,
|
| 365 |
+
"epoch": 0.2626865671641791,
|
| 366 |
+
"grad_norm": 0.008072705008089542,
|
| 367 |
+
"learning_rate": 1e-06,
|
| 368 |
+
"loss": 0.0,
|
| 369 |
+
"reward": 0.48437500558793545,
|
| 370 |
+
"reward_std": 0.2514351741410792,
|
| 371 |
+
"rewards/accuracy_reward": 0.48437500558793545,
|
| 372 |
+
"step": 33
|
| 373 |
+
},
|
| 374 |
+
{
|
| 375 |
+
"completion_length": 479.74914169311523,
|
| 376 |
+
"epoch": 0.27064676616915423,
|
| 377 |
+
"grad_norm": 0.007777619641274214,
|
| 378 |
+
"learning_rate": 1e-06,
|
| 379 |
+
"loss": 0.0,
|
| 380 |
+
"reward": 0.49305556155741215,
|
| 381 |
+
"reward_std": 0.20784300100058317,
|
| 382 |
+
"rewards/accuracy_reward": 0.49305556155741215,
|
| 383 |
+
"step": 34
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"completion_length": 487.4548645019531,
|
| 387 |
+
"epoch": 0.27860696517412936,
|
| 388 |
+
"grad_norm": 0.007740811910480261,
|
| 389 |
+
"learning_rate": 1e-06,
|
| 390 |
+
"loss": 0.0,
|
| 391 |
+
"reward": 0.44878472946584225,
|
| 392 |
+
"reward_std": 0.21900415536947548,
|
| 393 |
+
"rewards/accuracy_reward": 0.44878472946584225,
|
| 394 |
+
"step": 35
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"completion_length": 494.6788215637207,
|
| 398 |
+
"epoch": 0.2865671641791045,
|
| 399 |
+
"grad_norm": 0.007440795190632343,
|
| 400 |
+
"learning_rate": 1e-06,
|
| 401 |
+
"loss": 0.0,
|
| 402 |
+
"reward": 0.4652777789160609,
|
| 403 |
+
"reward_std": 0.20190619095228612,
|
| 404 |
+
"rewards/accuracy_reward": 0.4652777789160609,
|
| 405 |
+
"step": 36
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"completion_length": 503.41319274902344,
|
| 409 |
+
"epoch": 0.2945273631840796,
|
| 410 |
+
"grad_norm": 0.007378404960036278,
|
| 411 |
+
"learning_rate": 1e-06,
|
| 412 |
+
"loss": 0.0,
|
| 413 |
+
"reward": 0.4782986221835017,
|
| 414 |
+
"reward_std": 0.21211031870916486,
|
| 415 |
+
"rewards/accuracy_reward": 0.4782986221835017,
|
| 416 |
+
"step": 37
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"completion_length": 520.353307723999,
|
| 420 |
+
"epoch": 0.3024875621890547,
|
| 421 |
+
"grad_norm": 0.01668449677526951,
|
| 422 |
+
"learning_rate": 1e-06,
|
| 423 |
+
"loss": 0.0,
|
| 424 |
+
"reward": 0.4401041737291962,
|
| 425 |
+
"reward_std": 0.22917978325858712,
|
| 426 |
+
"rewards/accuracy_reward": 0.4401041737291962,
|
| 427 |
+
"step": 38
|
| 428 |
+
},
|
| 429 |
+
{
|
| 430 |
+
"completion_length": 483.45833587646484,
|
| 431 |
+
"epoch": 0.31044776119402984,
|
| 432 |
+
"grad_norm": 0.006350772920995951,
|
| 433 |
+
"learning_rate": 1e-06,
|
| 434 |
+
"loss": 0.0,
|
| 435 |
+
"reward": 0.4539930587634444,
|
| 436 |
+
"reward_std": 0.18883798900060356,
|
| 437 |
+
"rewards/accuracy_reward": 0.4539930587634444,
|
| 438 |
+
"step": 39
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"completion_length": 514.628475189209,
|
| 442 |
+
"epoch": 0.31840796019900497,
|
| 443 |
+
"grad_norm": 0.007026695180684328,
|
| 444 |
+
"learning_rate": 1e-06,
|
| 445 |
+
"loss": 0.0,
|
| 446 |
+
"reward": 0.5251736100763083,
|
| 447 |
+
"reward_std": 0.23443278204649687,
|
| 448 |
+
"rewards/accuracy_reward": 0.5251736100763083,
|
| 449 |
+
"step": 40
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"completion_length": 518.1093788146973,
|
| 453 |
+
"epoch": 0.3263681592039801,
|
| 454 |
+
"grad_norm": 0.008940315805375576,
|
| 455 |
+
"learning_rate": 1e-06,
|
| 456 |
+
"loss": 0.0,
|
| 457 |
+
"reward": 0.44010417396202683,
|
| 458 |
+
"reward_std": 0.17670530593022704,
|
| 459 |
+
"rewards/accuracy_reward": 0.44010417396202683,
|
| 460 |
+
"step": 41
|
| 461 |
+
},
|
| 462 |
+
{
|
| 463 |
+
"completion_length": 477.1041679382324,
|
| 464 |
+
"epoch": 0.33432835820895523,
|
| 465 |
+
"grad_norm": 0.006643439643085003,
|
| 466 |
+
"learning_rate": 1e-06,
|
| 467 |
+
"loss": 0.0,
|
| 468 |
+
"reward": 0.5312499990686774,
|
| 469 |
+
"reward_std": 0.20215208712033927,
|
| 470 |
+
"rewards/accuracy_reward": 0.5312499990686774,
|
| 471 |
+
"step": 42
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"completion_length": 507.63542556762695,
|
| 475 |
+
"epoch": 0.34228855721393037,
|
| 476 |
+
"grad_norm": 0.007287212181836367,
|
| 477 |
+
"learning_rate": 1e-06,
|
| 478 |
+
"loss": 0.0,
|
| 479 |
+
"reward": 0.5182291716337204,
|
| 480 |
+
"reward_std": 0.2360649723559618,
|
| 481 |
+
"rewards/accuracy_reward": 0.5182291716337204,
|
| 482 |
+
"step": 43
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"completion_length": 496.23959159851074,
|
| 486 |
+
"epoch": 0.3502487562189055,
|
| 487 |
+
"grad_norm": 0.007122528273612261,
|
| 488 |
+
"learning_rate": 1e-06,
|
| 489 |
+
"loss": 0.0,
|
| 490 |
+
"reward": 0.45312500931322575,
|
| 491 |
+
"reward_std": 0.20658620377071202,
|
| 492 |
+
"rewards/accuracy_reward": 0.45312500931322575,
|
| 493 |
+
"step": 44
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"completion_length": 485.8533020019531,
|
| 497 |
+
"epoch": 0.3582089552238806,
|
| 498 |
+
"grad_norm": 0.007156004197895527,
|
| 499 |
+
"learning_rate": 1e-06,
|
| 500 |
+
"loss": 0.0,
|
| 501 |
+
"reward": 0.530381953343749,
|
| 502 |
+
"reward_std": 0.24199844780378044,
|
| 503 |
+
"rewards/accuracy_reward": 0.530381953343749,
|
| 504 |
+
"step": 45
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"completion_length": 502.89844512939453,
|
| 508 |
+
"epoch": 0.3661691542288557,
|
| 509 |
+
"grad_norm": 0.0069762468338012695,
|
| 510 |
+
"learning_rate": 1e-06,
|
| 511 |
+
"loss": 0.0,
|
| 512 |
+
"reward": 0.4661458395421505,
|
| 513 |
+
"reward_std": 0.22123132436536252,
|
| 514 |
+
"rewards/accuracy_reward": 0.4661458395421505,
|
| 515 |
+
"step": 46
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"completion_length": 500.3958339691162,
|
| 519 |
+
"epoch": 0.37412935323383084,
|
| 520 |
+
"grad_norm": 0.006156955845654011,
|
| 521 |
+
"learning_rate": 1e-06,
|
| 522 |
+
"loss": 0.0,
|
| 523 |
+
"reward": 0.5052083404734731,
|
| 524 |
+
"reward_std": 0.20680285105481744,
|
| 525 |
+
"rewards/accuracy_reward": 0.5052083404734731,
|
| 526 |
+
"step": 47
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"completion_length": 494.1701469421387,
|
| 530 |
+
"epoch": 0.382089552238806,
|
| 531 |
+
"grad_norm": 0.005795426666736603,
|
| 532 |
+
"learning_rate": 1e-06,
|
| 533 |
+
"loss": 0.0,
|
| 534 |
+
"reward": 0.46527777798473835,
|
| 535 |
+
"reward_std": 0.14974681939929724,
|
| 536 |
+
"rewards/accuracy_reward": 0.46527777798473835,
|
| 537 |
+
"step": 48
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"completion_length": 489.7239627838135,
|
| 541 |
+
"epoch": 0.3900497512437811,
|
| 542 |
+
"grad_norm": 0.006671064533293247,
|
| 543 |
+
"learning_rate": 1e-06,
|
| 544 |
+
"loss": 0.0,
|
| 545 |
+
"reward": 0.44010416977107525,
|
| 546 |
+
"reward_std": 0.2161168558523059,
|
| 547 |
+
"rewards/accuracy_reward": 0.44010416977107525,
|
| 548 |
+
"step": 49
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"completion_length": 519.2543487548828,
|
| 552 |
+
"epoch": 0.39800995024875624,
|
| 553 |
+
"grad_norm": 0.025584502145648003,
|
| 554 |
+
"learning_rate": 1e-06,
|
| 555 |
+
"loss": 0.0,
|
| 556 |
+
"reward": 0.42795139038935304,
|
| 557 |
+
"reward_std": 0.20397222600877285,
|
| 558 |
+
"rewards/accuracy_reward": 0.42795139038935304,
|
| 559 |
+
"step": 50
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"completion_length": 500.26996994018555,
|
| 563 |
+
"epoch": 0.4059701492537313,
|
| 564 |
+
"grad_norm": 0.008419223129749298,
|
| 565 |
+
"learning_rate": 1e-06,
|
| 566 |
+
"loss": 0.0,
|
| 567 |
+
"reward": 0.45399306155741215,
|
| 568 |
+
"reward_std": 0.23704699613153934,
|
| 569 |
+
"rewards/accuracy_reward": 0.45399306155741215,
|
| 570 |
+
"step": 51
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"completion_length": 537.690107345581,
|
| 574 |
+
"epoch": 0.41393034825870645,
|
| 575 |
+
"grad_norm": 0.0072258333675563335,
|
| 576 |
+
"learning_rate": 1e-06,
|
| 577 |
+
"loss": 0.0,
|
| 578 |
+
"reward": 0.461805553175509,
|
| 579 |
+
"reward_std": 0.2027341139037162,
|
| 580 |
+
"rewards/accuracy_reward": 0.461805553175509,
|
| 581 |
+
"step": 52
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"completion_length": 513.0434074401855,
|
| 585 |
+
"epoch": 0.4218905472636816,
|
| 586 |
+
"grad_norm": 0.006870542652904987,
|
| 587 |
+
"learning_rate": 1e-06,
|
| 588 |
+
"loss": 0.0,
|
| 589 |
+
"reward": 0.471354172565043,
|
| 590 |
+
"reward_std": 0.20462280698120594,
|
| 591 |
+
"rewards/accuracy_reward": 0.471354172565043,
|
| 592 |
+
"step": 53
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"completion_length": 499.86979484558105,
|
| 596 |
+
"epoch": 0.4298507462686567,
|
| 597 |
+
"grad_norm": 0.007057087495923042,
|
| 598 |
+
"learning_rate": 1e-06,
|
| 599 |
+
"loss": 0.0,
|
| 600 |
+
"reward": 0.4947916716337204,
|
| 601 |
+
"reward_std": 0.2298571434803307,
|
| 602 |
+
"rewards/accuracy_reward": 0.4947916716337204,
|
| 603 |
+
"step": 54
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"completion_length": 499.2500057220459,
|
| 607 |
+
"epoch": 0.43781094527363185,
|
| 608 |
+
"grad_norm": 0.008050983771681786,
|
| 609 |
+
"learning_rate": 1e-06,
|
| 610 |
+
"loss": 0.0,
|
| 611 |
+
"reward": 0.49826389364898205,
|
| 612 |
+
"reward_std": 0.21837170561775565,
|
| 613 |
+
"rewards/accuracy_reward": 0.49826389364898205,
|
| 614 |
+
"step": 55
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"completion_length": 485.90799140930176,
|
| 618 |
+
"epoch": 0.445771144278607,
|
| 619 |
+
"grad_norm": 0.006851210258901119,
|
| 620 |
+
"learning_rate": 1e-06,
|
| 621 |
+
"loss": 0.0,
|
| 622 |
+
"reward": 0.47569445241242647,
|
| 623 |
+
"reward_std": 0.18343511526472867,
|
| 624 |
+
"rewards/accuracy_reward": 0.47569445241242647,
|
| 625 |
+
"step": 56
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"completion_length": 473.76736068725586,
|
| 629 |
+
"epoch": 0.4537313432835821,
|
| 630 |
+
"grad_norm": 0.006152069661766291,
|
| 631 |
+
"learning_rate": 1e-06,
|
| 632 |
+
"loss": 0.0,
|
| 633 |
+
"reward": 0.5147569477558136,
|
| 634 |
+
"reward_std": 0.18534683482721448,
|
| 635 |
+
"rewards/accuracy_reward": 0.5147569477558136,
|
| 636 |
+
"step": 57
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"completion_length": 509.58506965637207,
|
| 640 |
+
"epoch": 0.4616915422885572,
|
| 641 |
+
"grad_norm": 0.010075108148157597,
|
| 642 |
+
"learning_rate": 1e-06,
|
| 643 |
+
"loss": 0.0,
|
| 644 |
+
"reward": 0.5260416697710752,
|
| 645 |
+
"reward_std": 0.18553904327563941,
|
| 646 |
+
"rewards/accuracy_reward": 0.5260416697710752,
|
| 647 |
+
"step": 58
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"completion_length": 522.5798645019531,
|
| 651 |
+
"epoch": 0.4696517412935323,
|
| 652 |
+
"grad_norm": 0.006305539049208164,
|
| 653 |
+
"learning_rate": 1e-06,
|
| 654 |
+
"loss": 0.0,
|
| 655 |
+
"reward": 0.4357638917863369,
|
| 656 |
+
"reward_std": 0.19109400524757802,
|
| 657 |
+
"rewards/accuracy_reward": 0.4357638917863369,
|
| 658 |
+
"step": 59
|
| 659 |
+
},
|
| 660 |
+
{
|
| 661 |
+
"completion_length": 488.6762180328369,
|
| 662 |
+
"epoch": 0.47761194029850745,
|
| 663 |
+
"grad_norm": 0.007000816985964775,
|
| 664 |
+
"learning_rate": 1e-06,
|
| 665 |
+
"loss": 0.0,
|
| 666 |
+
"reward": 0.4236111119389534,
|
| 667 |
+
"reward_std": 0.18203387153334916,
|
| 668 |
+
"rewards/accuracy_reward": 0.4236111119389534,
|
| 669 |
+
"step": 60
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"completion_length": 513.9618091583252,
|
| 673 |
+
"epoch": 0.4855721393034826,
|
| 674 |
+
"grad_norm": 0.006217929068952799,
|
| 675 |
+
"learning_rate": 1e-06,
|
| 676 |
+
"loss": 0.0,
|
| 677 |
+
"reward": 0.4435763927176595,
|
| 678 |
+
"reward_std": 0.18900510389357805,
|
| 679 |
+
"rewards/accuracy_reward": 0.4435763927176595,
|
| 680 |
+
"step": 61
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"completion_length": 519.9574699401855,
|
| 684 |
+
"epoch": 0.4935323383084577,
|
| 685 |
+
"grad_norm": 0.006258189212530851,
|
| 686 |
+
"learning_rate": 1e-06,
|
| 687 |
+
"loss": 0.0,
|
| 688 |
+
"reward": 0.4739583358168602,
|
| 689 |
+
"reward_std": 0.20483243186026812,
|
| 690 |
+
"rewards/accuracy_reward": 0.4739583358168602,
|
| 691 |
+
"step": 62
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"completion_length": 491.98351287841797,
|
| 695 |
+
"epoch": 0.5014925373134328,
|
| 696 |
+
"grad_norm": 0.006038163788616657,
|
| 697 |
+
"learning_rate": 1e-06,
|
| 698 |
+
"loss": 0.0,
|
| 699 |
+
"reward": 0.4887152789160609,
|
| 700 |
+
"reward_std": 0.16513075795955956,
|
| 701 |
+
"rewards/accuracy_reward": 0.4887152789160609,
|
| 702 |
+
"step": 63
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"completion_length": 497.38976097106934,
|
| 706 |
+
"epoch": 0.5094527363184079,
|
| 707 |
+
"grad_norm": 0.007668279577046633,
|
| 708 |
+
"learning_rate": 1e-06,
|
| 709 |
+
"loss": 0.0,
|
| 710 |
+
"reward": 0.509548619389534,
|
| 711 |
+
"reward_std": 0.22911294596269727,
|
| 712 |
+
"rewards/accuracy_reward": 0.509548619389534,
|
| 713 |
+
"step": 64
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"completion_length": 493.03125190734863,
|
| 717 |
+
"epoch": 0.5174129353233831,
|
| 718 |
+
"grad_norm": 0.03523726761341095,
|
| 719 |
+
"learning_rate": 1e-06,
|
| 720 |
+
"loss": 0.0,
|
| 721 |
+
"reward": 0.5390625074505806,
|
| 722 |
+
"reward_std": 0.15510419360361993,
|
| 723 |
+
"rewards/accuracy_reward": 0.5390625074505806,
|
| 724 |
+
"step": 65
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"completion_length": 501.24392890930176,
|
| 728 |
+
"epoch": 0.5253731343283582,
|
| 729 |
+
"grad_norm": 0.005824146326631308,
|
| 730 |
+
"learning_rate": 1e-06,
|
| 731 |
+
"loss": 0.0,
|
| 732 |
+
"reward": 0.45659722574055195,
|
| 733 |
+
"reward_std": 0.18801256455481052,
|
| 734 |
+
"rewards/accuracy_reward": 0.45659722574055195,
|
| 735 |
+
"step": 66
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"completion_length": 507.4540042877197,
|
| 739 |
+
"epoch": 0.5333333333333333,
|
| 740 |
+
"grad_norm": 0.018976733088493347,
|
| 741 |
+
"learning_rate": 1e-06,
|
| 742 |
+
"loss": 0.0,
|
| 743 |
+
"reward": 0.46180555783212185,
|
| 744 |
+
"reward_std": 0.16340018948540092,
|
| 745 |
+
"rewards/accuracy_reward": 0.46180555783212185,
|
| 746 |
+
"step": 67
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"completion_length": 521.8541736602783,
|
| 750 |
+
"epoch": 0.5412935323383085,
|
| 751 |
+
"grad_norm": 0.006117444485425949,
|
| 752 |
+
"learning_rate": 1e-06,
|
| 753 |
+
"loss": 0.0,
|
| 754 |
+
"reward": 0.5338541683740914,
|
| 755 |
+
"reward_std": 0.21227112039923668,
|
| 756 |
+
"rewards/accuracy_reward": 0.5338541683740914,
|
| 757 |
+
"step": 68
|
| 758 |
+
},
|
| 759 |
+
{
|
| 760 |
+
"completion_length": 518.9166679382324,
|
| 761 |
+
"epoch": 0.5492537313432836,
|
| 762 |
+
"grad_norm": 0.006700050085783005,
|
| 763 |
+
"learning_rate": 1e-06,
|
| 764 |
+
"loss": 0.0,
|
| 765 |
+
"reward": 0.4574652770534158,
|
| 766 |
+
"reward_std": 0.22972481418401003,
|
| 767 |
+
"rewards/accuracy_reward": 0.4574652770534158,
|
| 768 |
+
"step": 69
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"completion_length": 498.01736068725586,
|
| 772 |
+
"epoch": 0.5572139303482587,
|
| 773 |
+
"grad_norm": 0.005739975720643997,
|
| 774 |
+
"learning_rate": 1e-06,
|
| 775 |
+
"loss": 0.0,
|
| 776 |
+
"reward": 0.4522569514811039,
|
| 777 |
+
"reward_std": 0.19061035430058837,
|
| 778 |
+
"rewards/accuracy_reward": 0.4522569514811039,
|
| 779 |
+
"step": 70
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"completion_length": 537.0442790985107,
|
| 783 |
+
"epoch": 0.5651741293532339,
|
| 784 |
+
"grad_norm": 0.00948350690305233,
|
| 785 |
+
"learning_rate": 1e-06,
|
| 786 |
+
"loss": 0.0,
|
| 787 |
+
"reward": 0.42708334140479565,
|
| 788 |
+
"reward_std": 0.15355130773968995,
|
| 789 |
+
"rewards/accuracy_reward": 0.42708334140479565,
|
| 790 |
+
"step": 71
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"completion_length": 474.55295753479004,
|
| 794 |
+
"epoch": 0.573134328358209,
|
| 795 |
+
"grad_norm": 0.007795177400112152,
|
| 796 |
+
"learning_rate": 1e-06,
|
| 797 |
+
"loss": 0.0,
|
| 798 |
+
"reward": 0.543402774259448,
|
| 799 |
+
"reward_std": 0.19562758482061327,
|
| 800 |
+
"rewards/accuracy_reward": 0.543402774259448,
|
| 801 |
+
"step": 72
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"completion_length": 525.5338649749756,
|
| 805 |
+
"epoch": 0.5810945273631841,
|
| 806 |
+
"grad_norm": 0.005621105432510376,
|
| 807 |
+
"learning_rate": 1e-06,
|
| 808 |
+
"loss": 0.0,
|
| 809 |
+
"reward": 0.4956597238779068,
|
| 810 |
+
"reward_std": 0.207592043094337,
|
| 811 |
+
"rewards/accuracy_reward": 0.4956597238779068,
|
| 812 |
+
"step": 73
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"completion_length": 473.47743797302246,
|
| 816 |
+
"epoch": 0.5890547263681593,
|
| 817 |
+
"grad_norm": 0.0060116685926914215,
|
| 818 |
+
"learning_rate": 1e-06,
|
| 819 |
+
"loss": 0.0,
|
| 820 |
+
"reward": 0.599826393648982,
|
| 821 |
+
"reward_std": 0.18716048658825457,
|
| 822 |
+
"rewards/accuracy_reward": 0.599826393648982,
|
| 823 |
+
"step": 74
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"completion_length": 493.51215744018555,
|
| 827 |
+
"epoch": 0.5970149253731343,
|
| 828 |
+
"grad_norm": 0.005957755260169506,
|
| 829 |
+
"learning_rate": 1e-06,
|
| 830 |
+
"loss": 0.0,
|
| 831 |
+
"reward": 0.5390625055879354,
|
| 832 |
+
"reward_std": 0.18902742909267545,
|
| 833 |
+
"rewards/accuracy_reward": 0.5390625055879354,
|
| 834 |
+
"step": 75
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"completion_length": 501.5486183166504,
|
| 838 |
+
"epoch": 0.6049751243781094,
|
| 839 |
+
"grad_norm": 0.007860912010073662,
|
| 840 |
+
"learning_rate": 1e-06,
|
| 841 |
+
"loss": 0.0,
|
| 842 |
+
"reward": 0.4644097303971648,
|
| 843 |
+
"reward_std": 0.19782971846871078,
|
| 844 |
+
"rewards/accuracy_reward": 0.4644097303971648,
|
| 845 |
+
"step": 76
|
| 846 |
+
},
|
| 847 |
+
{
|
| 848 |
+
"completion_length": 488.24045181274414,
|
| 849 |
+
"epoch": 0.6129353233830845,
|
| 850 |
+
"grad_norm": 0.005519668105989695,
|
| 851 |
+
"learning_rate": 1e-06,
|
| 852 |
+
"loss": 0.0,
|
| 853 |
+
"reward": 0.5008680624887347,
|
| 854 |
+
"reward_std": 0.15759161300957203,
|
| 855 |
+
"rewards/accuracy_reward": 0.5008680624887347,
|
| 856 |
+
"step": 77
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"completion_length": 499.68663787841797,
|
| 860 |
+
"epoch": 0.6208955223880597,
|
| 861 |
+
"grad_norm": 0.006401211954653263,
|
| 862 |
+
"learning_rate": 1e-06,
|
| 863 |
+
"loss": 0.0,
|
| 864 |
+
"reward": 0.5434027835726738,
|
| 865 |
+
"reward_std": 0.2069430819246918,
|
| 866 |
+
"rewards/accuracy_reward": 0.5434027835726738,
|
| 867 |
+
"step": 78
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"completion_length": 494.0677146911621,
|
| 871 |
+
"epoch": 0.6288557213930348,
|
| 872 |
+
"grad_norm": 0.006467628292739391,
|
| 873 |
+
"learning_rate": 1e-06,
|
| 874 |
+
"loss": 0.0,
|
| 875 |
+
"reward": 0.4635416744276881,
|
| 876 |
+
"reward_std": 0.20270957378670573,
|
| 877 |
+
"rewards/accuracy_reward": 0.4635416744276881,
|
| 878 |
+
"step": 79
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"completion_length": 509.86806297302246,
|
| 882 |
+
"epoch": 0.6368159203980099,
|
| 883 |
+
"grad_norm": 0.005684923380613327,
|
| 884 |
+
"learning_rate": 1e-06,
|
| 885 |
+
"loss": 0.0,
|
| 886 |
+
"reward": 0.469618059694767,
|
| 887 |
+
"reward_std": 0.19833743665367365,
|
| 888 |
+
"rewards/accuracy_reward": 0.469618059694767,
|
| 889 |
+
"step": 80
|
| 890 |
+
},
|
| 891 |
+
{
|
| 892 |
+
"completion_length": 502.8680648803711,
|
| 893 |
+
"epoch": 0.6447761194029851,
|
| 894 |
+
"grad_norm": 0.005998033564537764,
|
| 895 |
+
"learning_rate": 1e-06,
|
| 896 |
+
"loss": 0.0,
|
| 897 |
+
"reward": 0.5711805634200573,
|
| 898 |
+
"reward_std": 0.18520855018869042,
|
| 899 |
+
"rewards/accuracy_reward": 0.5711805634200573,
|
| 900 |
+
"step": 81
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"completion_length": 508.05382347106934,
|
| 904 |
+
"epoch": 0.6527363184079602,
|
| 905 |
+
"grad_norm": 0.005817387718707323,
|
| 906 |
+
"learning_rate": 1e-06,
|
| 907 |
+
"loss": 0.0,
|
| 908 |
+
"reward": 0.5017361203208566,
|
| 909 |
+
"reward_std": 0.19694075919687748,
|
| 910 |
+
"rewards/accuracy_reward": 0.5017361203208566,
|
| 911 |
+
"step": 82
|
| 912 |
+
},
|
| 913 |
+
{
|
| 914 |
+
"completion_length": 529.6744899749756,
|
| 915 |
+
"epoch": 0.6606965174129353,
|
| 916 |
+
"grad_norm": 0.005400301422923803,
|
| 917 |
+
"learning_rate": 1e-06,
|
| 918 |
+
"loss": 0.0,
|
| 919 |
+
"reward": 0.5060763908550143,
|
| 920 |
+
"reward_std": 0.17399667436257005,
|
| 921 |
+
"rewards/accuracy_reward": 0.5060763908550143,
|
| 922 |
+
"step": 83
|
| 923 |
+
},
|
| 924 |
+
{
|
| 925 |
+
"completion_length": 502.5468807220459,
|
| 926 |
+
"epoch": 0.6686567164179105,
|
| 927 |
+
"grad_norm": 0.0058842068538069725,
|
| 928 |
+
"learning_rate": 1e-06,
|
| 929 |
+
"loss": 0.0,
|
| 930 |
+
"reward": 0.4739583395421505,
|
| 931 |
+
"reward_std": 0.16876469319686294,
|
| 932 |
+
"rewards/accuracy_reward": 0.4739583395421505,
|
| 933 |
+
"step": 84
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"completion_length": 518.564245223999,
|
| 937 |
+
"epoch": 0.6766169154228856,
|
| 938 |
+
"grad_norm": 0.005277169402688742,
|
| 939 |
+
"learning_rate": 1e-06,
|
| 940 |
+
"loss": 0.0,
|
| 941 |
+
"reward": 0.47916667512618005,
|
| 942 |
+
"reward_std": 0.19327577715739608,
|
| 943 |
+
"rewards/accuracy_reward": 0.47916667512618005,
|
| 944 |
+
"step": 85
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"completion_length": 475.6883716583252,
|
| 948 |
+
"epoch": 0.6845771144278607,
|
| 949 |
+
"grad_norm": 0.0056811547838151455,
|
| 950 |
+
"learning_rate": 1e-06,
|
| 951 |
+
"loss": 0.0,
|
| 952 |
+
"reward": 0.45746528450399637,
|
| 953 |
+
"reward_std": 0.15142568410374224,
|
| 954 |
+
"rewards/accuracy_reward": 0.45746528450399637,
|
| 955 |
+
"step": 86
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"completion_length": 540.3324699401855,
|
| 959 |
+
"epoch": 0.6925373134328359,
|
| 960 |
+
"grad_norm": 0.009126154705882072,
|
| 961 |
+
"learning_rate": 1e-06,
|
| 962 |
+
"loss": 0.0,
|
| 963 |
+
"reward": 0.42795139411464334,
|
| 964 |
+
"reward_std": 0.19471946358680725,
|
| 965 |
+
"rewards/accuracy_reward": 0.42795139411464334,
|
| 966 |
+
"step": 87
|
| 967 |
+
},
|
| 968 |
+
{
|
| 969 |
+
"completion_length": 505.52084159851074,
|
| 970 |
+
"epoch": 0.700497512437811,
|
| 971 |
+
"grad_norm": 0.0066405292600393295,
|
| 972 |
+
"learning_rate": 1e-06,
|
| 973 |
+
"loss": 0.0,
|
| 974 |
+
"reward": 0.5182291679084301,
|
| 975 |
+
"reward_std": 0.2177837509661913,
|
| 976 |
+
"rewards/accuracy_reward": 0.5182291679084301,
|
| 977 |
+
"step": 88
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"completion_length": 506.095495223999,
|
| 981 |
+
"epoch": 0.708457711442786,
|
| 982 |
+
"grad_norm": 0.005975022446364164,
|
| 983 |
+
"learning_rate": 1e-06,
|
| 984 |
+
"loss": 0.0,
|
| 985 |
+
"reward": 0.5199652845039964,
|
| 986 |
+
"reward_std": 0.17908447259105742,
|
| 987 |
+
"rewards/accuracy_reward": 0.5199652845039964,
|
| 988 |
+
"step": 89
|
| 989 |
+
},
|
| 990 |
+
{
|
| 991 |
+
"completion_length": 510.58246994018555,
|
| 992 |
+
"epoch": 0.7164179104477612,
|
| 993 |
+
"grad_norm": 0.006494089029729366,
|
| 994 |
+
"learning_rate": 1e-06,
|
| 995 |
+
"loss": 0.0,
|
| 996 |
+
"reward": 0.4973958358168602,
|
| 997 |
+
"reward_std": 0.1795428330078721,
|
| 998 |
+
"rewards/accuracy_reward": 0.4973958358168602,
|
| 999 |
+
"step": 90
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"completion_length": 487.8923625946045,
|
| 1003 |
+
"epoch": 0.7243781094527363,
|
| 1004 |
+
"grad_norm": 0.0053886245004832745,
|
| 1005 |
+
"learning_rate": 1e-06,
|
| 1006 |
+
"loss": 0.0,
|
| 1007 |
+
"reward": 0.5616319570690393,
|
| 1008 |
+
"reward_std": 0.16185199399478734,
|
| 1009 |
+
"rewards/accuracy_reward": 0.5616319570690393,
|
| 1010 |
+
"step": 91
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"completion_length": 505.83073234558105,
|
| 1014 |
+
"epoch": 0.7323383084577114,
|
| 1015 |
+
"grad_norm": 0.006866929121315479,
|
| 1016 |
+
"learning_rate": 1e-06,
|
| 1017 |
+
"loss": 0.0,
|
| 1018 |
+
"reward": 0.5407986268401146,
|
| 1019 |
+
"reward_std": 0.22892741695977747,
|
| 1020 |
+
"rewards/accuracy_reward": 0.5407986268401146,
|
| 1021 |
+
"step": 92
|
| 1022 |
+
},
|
| 1023 |
+
{
|
| 1024 |
+
"completion_length": 510.71788787841797,
|
| 1025 |
+
"epoch": 0.7402985074626866,
|
| 1026 |
+
"grad_norm": 0.005727425683289766,
|
| 1027 |
+
"learning_rate": 1e-06,
|
| 1028 |
+
"loss": 0.0,
|
| 1029 |
+
"reward": 0.4331597206182778,
|
| 1030 |
+
"reward_std": 0.16079934942536056,
|
| 1031 |
+
"rewards/accuracy_reward": 0.4331597206182778,
|
| 1032 |
+
"step": 93
|
| 1033 |
+
},
|
| 1034 |
+
{
|
| 1035 |
+
"completion_length": 528.9479293823242,
|
| 1036 |
+
"epoch": 0.7482587064676617,
|
| 1037 |
+
"grad_norm": 0.005505099426954985,
|
| 1038 |
+
"learning_rate": 1e-06,
|
| 1039 |
+
"loss": 0.0,
|
| 1040 |
+
"reward": 0.5173611119389534,
|
| 1041 |
+
"reward_std": 0.16449494729749858,
|
| 1042 |
+
"rewards/accuracy_reward": 0.5173611119389534,
|
| 1043 |
+
"step": 94
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"completion_length": 467.2135429382324,
|
| 1047 |
+
"epoch": 0.7562189054726368,
|
| 1048 |
+
"grad_norm": 0.005926135461777449,
|
| 1049 |
+
"learning_rate": 1e-06,
|
| 1050 |
+
"loss": 0.0,
|
| 1051 |
+
"reward": 0.5173611175268888,
|
| 1052 |
+
"reward_std": 0.16620213235728443,
|
| 1053 |
+
"rewards/accuracy_reward": 0.5173611175268888,
|
| 1054 |
+
"step": 95
|
| 1055 |
+
},
|
| 1056 |
+
{
|
| 1057 |
+
"completion_length": 490.4392375946045,
|
| 1058 |
+
"epoch": 0.764179104477612,
|
| 1059 |
+
"grad_norm": 0.0067347330041229725,
|
| 1060 |
+
"learning_rate": 1e-06,
|
| 1061 |
+
"loss": 0.0,
|
| 1062 |
+
"reward": 0.4739583395421505,
|
| 1063 |
+
"reward_std": 0.15511562791652977,
|
| 1064 |
+
"rewards/accuracy_reward": 0.4739583395421505,
|
| 1065 |
+
"step": 96
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"completion_length": 540.5928859710693,
|
| 1069 |
+
"epoch": 0.7721393034825871,
|
| 1070 |
+
"grad_norm": 0.005511025432497263,
|
| 1071 |
+
"learning_rate": 1e-06,
|
| 1072 |
+
"loss": 0.0,
|
| 1073 |
+
"reward": 0.41319444216787815,
|
| 1074 |
+
"reward_std": 0.14969537244178355,
|
| 1075 |
+
"rewards/accuracy_reward": 0.41319444216787815,
|
| 1076 |
+
"step": 97
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"completion_length": 494.7031364440918,
|
| 1080 |
+
"epoch": 0.7800995024875622,
|
| 1081 |
+
"grad_norm": 0.005637906491756439,
|
| 1082 |
+
"learning_rate": 1e-06,
|
| 1083 |
+
"loss": 0.0,
|
| 1084 |
+
"reward": 0.5633680630126037,
|
| 1085 |
+
"reward_std": 0.13407049677334726,
|
| 1086 |
+
"rewards/accuracy_reward": 0.5633680630126037,
|
| 1087 |
+
"step": 98
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"completion_length": 478.24132537841797,
|
| 1091 |
+
"epoch": 0.7880597014925373,
|
| 1092 |
+
"grad_norm": 0.005772520788013935,
|
| 1093 |
+
"learning_rate": 1e-06,
|
| 1094 |
+
"loss": 0.0,
|
| 1095 |
+
"reward": 0.5373263955116272,
|
| 1096 |
+
"reward_std": 0.1734980179462582,
|
| 1097 |
+
"rewards/accuracy_reward": 0.5373263955116272,
|
| 1098 |
+
"step": 99
|
| 1099 |
+
},
|
| 1100 |
+
{
|
| 1101 |
+
"completion_length": 513.6310844421387,
|
| 1102 |
+
"epoch": 0.7960199004975125,
|
| 1103 |
+
"grad_norm": 0.006371782626956701,
|
| 1104 |
+
"learning_rate": 1e-06,
|
| 1105 |
+
"loss": 0.0,
|
| 1106 |
+
"reward": 0.42968750186264515,
|
| 1107 |
+
"reward_std": 0.20266994088888168,
|
| 1108 |
+
"rewards/accuracy_reward": 0.42968750186264515,
|
| 1109 |
+
"step": 100
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"completion_length": 488.7838592529297,
|
| 1113 |
+
"epoch": 0.8039800995024876,
|
| 1114 |
+
"grad_norm": 0.006290055345743895,
|
| 1115 |
+
"learning_rate": 1e-06,
|
| 1116 |
+
"loss": 0.0,
|
| 1117 |
+
"reward": 0.5668402779847383,
|
| 1118 |
+
"reward_std": 0.20122395176440477,
|
| 1119 |
+
"rewards/accuracy_reward": 0.5668402779847383,
|
| 1120 |
+
"step": 101
|
| 1121 |
+
},
|
| 1122 |
+
{
|
| 1123 |
+
"completion_length": 504.9401111602783,
|
| 1124 |
+
"epoch": 0.8119402985074626,
|
| 1125 |
+
"grad_norm": 0.005973901599645615,
|
| 1126 |
+
"learning_rate": 1e-06,
|
| 1127 |
+
"loss": 0.0,
|
| 1128 |
+
"reward": 0.5269097238779068,
|
| 1129 |
+
"reward_std": 0.1675855671055615,
|
| 1130 |
+
"rewards/accuracy_reward": 0.5269097238779068,
|
| 1131 |
+
"step": 102
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"completion_length": 514.7942790985107,
|
| 1135 |
+
"epoch": 0.8199004975124378,
|
| 1136 |
+
"grad_norm": 0.006263998337090015,
|
| 1137 |
+
"learning_rate": 1e-06,
|
| 1138 |
+
"loss": 0.0,
|
| 1139 |
+
"reward": 0.511284725740552,
|
| 1140 |
+
"reward_std": 0.17091166111640632,
|
| 1141 |
+
"rewards/accuracy_reward": 0.511284725740552,
|
| 1142 |
+
"step": 103
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"completion_length": 497.96963119506836,
|
| 1146 |
+
"epoch": 0.8278606965174129,
|
| 1147 |
+
"grad_norm": 0.006026262417435646,
|
| 1148 |
+
"learning_rate": 1e-06,
|
| 1149 |
+
"loss": 0.0,
|
| 1150 |
+
"reward": 0.5069444456603378,
|
| 1151 |
+
"reward_std": 0.18491672072559595,
|
| 1152 |
+
"rewards/accuracy_reward": 0.5069444456603378,
|
| 1153 |
+
"step": 104
|
| 1154 |
+
},
|
| 1155 |
+
{
|
| 1156 |
+
"completion_length": 513.2239627838135,
|
| 1157 |
+
"epoch": 0.835820895522388,
|
| 1158 |
+
"grad_norm": 0.005511322058737278,
|
| 1159 |
+
"learning_rate": 1e-06,
|
| 1160 |
+
"loss": 0.0,
|
| 1161 |
+
"reward": 0.5008680615574121,
|
| 1162 |
+
"reward_std": 0.155679541407153,
|
| 1163 |
+
"rewards/accuracy_reward": 0.5008680615574121,
|
| 1164 |
+
"step": 105
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"completion_length": 506.9704837799072,
|
| 1168 |
+
"epoch": 0.8437810945273632,
|
| 1169 |
+
"grad_norm": 0.005759282503277063,
|
| 1170 |
+
"learning_rate": 1e-06,
|
| 1171 |
+
"loss": 0.0,
|
| 1172 |
+
"reward": 0.5190972294658422,
|
| 1173 |
+
"reward_std": 0.15153015381656587,
|
| 1174 |
+
"rewards/accuracy_reward": 0.5190972294658422,
|
| 1175 |
+
"step": 106
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"completion_length": 508.5946216583252,
|
| 1179 |
+
"epoch": 0.8517412935323383,
|
| 1180 |
+
"grad_norm": 0.005872826091945171,
|
| 1181 |
+
"learning_rate": 1e-06,
|
| 1182 |
+
"loss": 0.0,
|
| 1183 |
+
"reward": 0.4861111119389534,
|
| 1184 |
+
"reward_std": 0.17534176167100668,
|
| 1185 |
+
"rewards/accuracy_reward": 0.4861111119389534,
|
| 1186 |
+
"step": 107
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"completion_length": 513.0425434112549,
|
| 1190 |
+
"epoch": 0.8597014925373134,
|
| 1191 |
+
"grad_norm": 0.006046103313565254,
|
| 1192 |
+
"learning_rate": 1e-06,
|
| 1193 |
+
"loss": 0.0,
|
| 1194 |
+
"reward": 0.507812506519258,
|
| 1195 |
+
"reward_std": 0.18332828022539616,
|
| 1196 |
+
"rewards/accuracy_reward": 0.507812506519258,
|
| 1197 |
+
"step": 108
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"completion_length": 515.2300395965576,
|
| 1201 |
+
"epoch": 0.8676616915422886,
|
| 1202 |
+
"grad_norm": 0.0062822867184877396,
|
| 1203 |
+
"learning_rate": 1e-06,
|
| 1204 |
+
"loss": 0.0,
|
| 1205 |
+
"reward": 0.4765625046566129,
|
| 1206 |
+
"reward_std": 0.19052056316286325,
|
| 1207 |
+
"rewards/accuracy_reward": 0.4765625046566129,
|
| 1208 |
+
"step": 109
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"completion_length": 500.3810749053955,
|
| 1212 |
+
"epoch": 0.8756218905472637,
|
| 1213 |
+
"grad_norm": 0.005929249804466963,
|
| 1214 |
+
"learning_rate": 1e-06,
|
| 1215 |
+
"loss": 0.0,
|
| 1216 |
+
"reward": 0.5277777798473835,
|
| 1217 |
+
"reward_std": 0.17764607537537813,
|
| 1218 |
+
"rewards/accuracy_reward": 0.5277777798473835,
|
| 1219 |
+
"step": 110
|
| 1220 |
+
},
|
| 1221 |
+
{
|
| 1222 |
+
"completion_length": 519.638017654419,
|
| 1223 |
+
"epoch": 0.8835820895522388,
|
| 1224 |
+
"grad_norm": 0.0055696116760373116,
|
| 1225 |
+
"learning_rate": 1e-06,
|
| 1226 |
+
"loss": 0.0,
|
| 1227 |
+
"reward": 0.5164930606260896,
|
| 1228 |
+
"reward_std": 0.2034295415505767,
|
| 1229 |
+
"rewards/accuracy_reward": 0.5164930606260896,
|
| 1230 |
+
"step": 111
|
| 1231 |
+
},
|
| 1232 |
+
{
|
| 1233 |
+
"completion_length": 507.8932342529297,
|
| 1234 |
+
"epoch": 0.891542288557214,
|
| 1235 |
+
"grad_norm": 0.00611339695751667,
|
| 1236 |
+
"learning_rate": 1e-06,
|
| 1237 |
+
"loss": 0.0,
|
| 1238 |
+
"reward": 0.5225694514811039,
|
| 1239 |
+
"reward_std": 0.18155742809176445,
|
| 1240 |
+
"rewards/accuracy_reward": 0.5225694514811039,
|
| 1241 |
+
"step": 112
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"completion_length": 478.3941020965576,
|
| 1245 |
+
"epoch": 0.8995024875621891,
|
| 1246 |
+
"grad_norm": 0.005754369776695967,
|
| 1247 |
+
"learning_rate": 1e-06,
|
| 1248 |
+
"loss": 0.0,
|
| 1249 |
+
"reward": 0.5572916716337204,
|
| 1250 |
+
"reward_std": 0.15763491089455783,
|
| 1251 |
+
"rewards/accuracy_reward": 0.5572916716337204,
|
| 1252 |
+
"step": 113
|
| 1253 |
+
},
|
| 1254 |
+
{
|
| 1255 |
+
"completion_length": 503.7673645019531,
|
| 1256 |
+
"epoch": 0.9074626865671642,
|
| 1257 |
+
"grad_norm": 0.005771811120212078,
|
| 1258 |
+
"learning_rate": 1e-06,
|
| 1259 |
+
"loss": 0.0,
|
| 1260 |
+
"reward": 0.5703125055879354,
|
| 1261 |
+
"reward_std": 0.17914820974692702,
|
| 1262 |
+
"rewards/accuracy_reward": 0.5703125055879354,
|
| 1263 |
+
"step": 114
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"completion_length": 484.3923645019531,
|
| 1267 |
+
"epoch": 0.9154228855721394,
|
| 1268 |
+
"grad_norm": 0.0070279622450470924,
|
| 1269 |
+
"learning_rate": 1e-06,
|
| 1270 |
+
"loss": 0.0,
|
| 1271 |
+
"reward": 0.5347222220152617,
|
| 1272 |
+
"reward_std": 0.19823500025086105,
|
| 1273 |
+
"rewards/accuracy_reward": 0.5347222220152617,
|
| 1274 |
+
"step": 115
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"completion_length": 516.4401035308838,
|
| 1278 |
+
"epoch": 0.9233830845771144,
|
| 1279 |
+
"grad_norm": 0.006754903122782707,
|
| 1280 |
+
"learning_rate": 1e-06,
|
| 1281 |
+
"loss": 0.0,
|
| 1282 |
+
"reward": 0.49045139644294977,
|
| 1283 |
+
"reward_std": 0.2149109251331538,
|
| 1284 |
+
"rewards/accuracy_reward": 0.49045139644294977,
|
| 1285 |
+
"step": 116
|
| 1286 |
+
},
|
| 1287 |
+
{
|
| 1288 |
+
"completion_length": 501.6666736602783,
|
| 1289 |
+
"epoch": 0.9313432835820895,
|
| 1290 |
+
"grad_norm": 0.006792979780584574,
|
| 1291 |
+
"learning_rate": 1e-06,
|
| 1292 |
+
"loss": 0.0,
|
| 1293 |
+
"reward": 0.49392362032085657,
|
| 1294 |
+
"reward_std": 0.19255853188224137,
|
| 1295 |
+
"rewards/accuracy_reward": 0.49392362032085657,
|
| 1296 |
+
"step": 117
|
| 1297 |
+
},
|
| 1298 |
+
{
|
| 1299 |
+
"completion_length": 483.9678840637207,
|
| 1300 |
+
"epoch": 0.9393034825870646,
|
| 1301 |
+
"grad_norm": 0.005810786038637161,
|
| 1302 |
+
"learning_rate": 1e-06,
|
| 1303 |
+
"loss": 0.0,
|
| 1304 |
+
"reward": 0.5112847303971648,
|
| 1305 |
+
"reward_std": 0.18844515248201787,
|
| 1306 |
+
"rewards/accuracy_reward": 0.5112847303971648,
|
| 1307 |
+
"step": 118
|
| 1308 |
+
},
|
| 1309 |
+
{
|
| 1310 |
+
"completion_length": 510.7786464691162,
|
| 1311 |
+
"epoch": 0.9472636815920398,
|
| 1312 |
+
"grad_norm": 0.006106381770223379,
|
| 1313 |
+
"learning_rate": 1e-06,
|
| 1314 |
+
"loss": 0.0,
|
| 1315 |
+
"reward": 0.42881944589316845,
|
| 1316 |
+
"reward_std": 0.17682828847318888,
|
| 1317 |
+
"rewards/accuracy_reward": 0.42881944589316845,
|
| 1318 |
+
"step": 119
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"completion_length": 495.50608253479004,
|
| 1322 |
+
"epoch": 0.9552238805970149,
|
| 1323 |
+
"grad_norm": 0.00595560297369957,
|
| 1324 |
+
"learning_rate": 1e-06,
|
| 1325 |
+
"loss": 0.0,
|
| 1326 |
+
"reward": 0.5217013895162381,
|
| 1327 |
+
"reward_std": 0.16389768570661545,
|
| 1328 |
+
"rewards/accuracy_reward": 0.5217013895162381,
|
| 1329 |
+
"step": 120
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"completion_length": 535.8029499053955,
|
| 1333 |
+
"epoch": 0.96318407960199,
|
| 1334 |
+
"grad_norm": 0.0052213650196790695,
|
| 1335 |
+
"learning_rate": 1e-06,
|
| 1336 |
+
"loss": 0.0,
|
| 1337 |
+
"reward": 0.48871528543531895,
|
| 1338 |
+
"reward_std": 0.17809172347187996,
|
| 1339 |
+
"rewards/accuracy_reward": 0.48871528543531895,
|
| 1340 |
+
"step": 121
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"completion_length": 516.939245223999,
|
| 1344 |
+
"epoch": 0.9711442786069652,
|
| 1345 |
+
"grad_norm": 0.00579698896035552,
|
| 1346 |
+
"learning_rate": 1e-06,
|
| 1347 |
+
"loss": 0.0,
|
| 1348 |
+
"reward": 0.4652777789160609,
|
| 1349 |
+
"reward_std": 0.2078009396791458,
|
| 1350 |
+
"rewards/accuracy_reward": 0.4652777789160609,
|
| 1351 |
+
"step": 122
|
| 1352 |
+
},
|
| 1353 |
+
{
|
| 1354 |
+
"completion_length": 497.8142395019531,
|
| 1355 |
+
"epoch": 0.9791044776119403,
|
| 1356 |
+
"grad_norm": 0.005309565458446741,
|
| 1357 |
+
"learning_rate": 1e-06,
|
| 1358 |
+
"loss": 0.0,
|
| 1359 |
+
"reward": 0.49652778543531895,
|
| 1360 |
+
"reward_std": 0.1809900659136474,
|
| 1361 |
+
"rewards/accuracy_reward": 0.49652778543531895,
|
| 1362 |
+
"step": 123
|
| 1363 |
+
},
|
| 1364 |
+
{
|
| 1365 |
+
"completion_length": 486.0590305328369,
|
| 1366 |
+
"epoch": 0.9870646766169154,
|
| 1367 |
+
"grad_norm": 0.005974112078547478,
|
| 1368 |
+
"learning_rate": 1e-06,
|
| 1369 |
+
"loss": 0.0,
|
| 1370 |
+
"reward": 0.5312500074505806,
|
| 1371 |
+
"reward_std": 0.16324150539003313,
|
| 1372 |
+
"rewards/accuracy_reward": 0.5312500074505806,
|
| 1373 |
+
"step": 124
|
| 1374 |
+
},
|
| 1375 |
+
{
|
| 1376 |
+
"completion_length": 469.81510734558105,
|
| 1377 |
+
"epoch": 0.9950248756218906,
|
| 1378 |
+
"grad_norm": 0.005214582197368145,
|
| 1379 |
+
"learning_rate": 1e-06,
|
| 1380 |
+
"loss": 0.0,
|
| 1381 |
+
"reward": 0.6154513941146433,
|
| 1382 |
+
"reward_std": 0.13105975766666234,
|
| 1383 |
+
"rewards/accuracy_reward": 0.6154513941146433,
|
| 1384 |
+
"step": 125
|
| 1385 |
+
},
|
| 1386 |
+
{
|
| 1387 |
+
"epoch": 0.9950248756218906,
|
| 1388 |
+
"step": 125,
|
| 1389 |
+
"total_flos": 0.0,
|
| 1390 |
+
"train_loss": 2.3229669920965534e-08,
|
| 1391 |
+
"train_runtime": 35950.9286,
|
| 1392 |
+
"train_samples_per_second": 0.335,
|
| 1393 |
+
"train_steps_per_second": 0.003
|
| 1394 |
+
}
|
| 1395 |
+
],
|
| 1396 |
+
"logging_steps": 1,
|
| 1397 |
+
"max_steps": 125,
|
| 1398 |
+
"num_input_tokens_seen": 0,
|
| 1399 |
+
"num_train_epochs": 1,
|
| 1400 |
+
"save_steps": 10,
|
| 1401 |
+
"stateful_callbacks": {
|
| 1402 |
+
"TrainerControl": {
|
| 1403 |
+
"args": {
|
| 1404 |
+
"should_epoch_stop": false,
|
| 1405 |
+
"should_evaluate": false,
|
| 1406 |
+
"should_log": false,
|
| 1407 |
+
"should_save": true,
|
| 1408 |
+
"should_training_stop": true
|
| 1409 |
+
},
|
| 1410 |
+
"attributes": {}
|
| 1411 |
+
}
|
| 1412 |
+
},
|
| 1413 |
+
"total_flos": 0.0,
|
| 1414 |
+
"train_batch_size": 1,
|
| 1415 |
+
"trial_name": null,
|
| 1416 |
+
"trial_params": null
|
| 1417 |
+
}
|