Model save
Browse files- README.md +68 -0
- all_results.json +8 -0
- generation_config.json +6 -0
- train_results.json +8 -0
- trainer_state.json +2030 -0
README.md
ADDED
|
@@ -0,0 +1,68 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model: Qwen/Qwen2.5-7B
|
| 3 |
+
library_name: transformers
|
| 4 |
+
model_name: Qwen2.5-7B-Open-R1-GRPO-math-lighteval-log
|
| 5 |
+
tags:
|
| 6 |
+
- generated_from_trainer
|
| 7 |
+
- trl
|
| 8 |
+
- grpo
|
| 9 |
+
licence: license
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# Model Card for Qwen2.5-7B-Open-R1-GRPO-math-lighteval-log
|
| 13 |
+
|
| 14 |
+
This model is a fine-tuned version of [Qwen/Qwen2.5-7B](https://huggingface.co/Qwen/Qwen2.5-7B).
|
| 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="Lansechen/Qwen2.5-7B-Open-R1-GRPO-math-lighteval-log", 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/chenran1995-the-chinese-university-of-hong-kong/huggingface/runs/idxxkq28)
|
| 31 |
+
|
| 32 |
+
|
| 33 |
+
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).
|
| 34 |
+
|
| 35 |
+
### Framework versions
|
| 36 |
+
|
| 37 |
+
- TRL: 0.16.0
|
| 38 |
+
- Transformers: 4.49.0
|
| 39 |
+
- Pytorch: 2.5.1+cu121
|
| 40 |
+
- Datasets: 3.3.1
|
| 41 |
+
- Tokenizers: 0.21.0
|
| 42 |
+
|
| 43 |
+
## Citations
|
| 44 |
+
|
| 45 |
+
Cite GRPO as:
|
| 46 |
+
|
| 47 |
+
```bibtex
|
| 48 |
+
@article{zhihong2024deepseekmath,
|
| 49 |
+
title = {{DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models}},
|
| 50 |
+
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},
|
| 51 |
+
year = 2024,
|
| 52 |
+
eprint = {arXiv:2402.03300},
|
| 53 |
+
}
|
| 54 |
+
|
| 55 |
+
```
|
| 56 |
+
|
| 57 |
+
Cite TRL as:
|
| 58 |
+
|
| 59 |
+
```bibtex
|
| 60 |
+
@misc{vonwerra2022trl,
|
| 61 |
+
title = {{TRL: Transformer Reinforcement Learning}},
|
| 62 |
+
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},
|
| 63 |
+
year = 2020,
|
| 64 |
+
journal = {GitHub repository},
|
| 65 |
+
publisher = {GitHub},
|
| 66 |
+
howpublished = {\url{https://github.com/huggingface/trl}}
|
| 67 |
+
}
|
| 68 |
+
```
|
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"total_flos": 0.0,
|
| 3 |
+
"train_loss": 0.04579740530018937,
|
| 4 |
+
"train_runtime": 16666.94,
|
| 5 |
+
"train_samples": 7500,
|
| 6 |
+
"train_samples_per_second": 0.9,
|
| 7 |
+
"train_steps_per_second": 0.008
|
| 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": "4.49.0"
|
| 6 |
+
}
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"total_flos": 0.0,
|
| 3 |
+
"train_loss": 0.04579740530018937,
|
| 4 |
+
"train_runtime": 16666.94,
|
| 5 |
+
"train_samples": 7500,
|
| 6 |
+
"train_samples_per_second": 0.9,
|
| 7 |
+
"train_steps_per_second": 0.008
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2030 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 1.9850746268656716,
|
| 5 |
+
"eval_steps": 100,
|
| 6 |
+
"global_step": 132,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"clip_ratio": 0.0,
|
| 13 |
+
"completion_length": 823.762321472168,
|
| 14 |
+
"epoch": 0.014925373134328358,
|
| 15 |
+
"grad_norm": 0.17830021679401398,
|
| 16 |
+
"learning_rate": 7.142857142857142e-08,
|
| 17 |
+
"loss": 0.0919,
|
| 18 |
+
"num_tokens": 865267.0,
|
| 19 |
+
"reward": 2.096768334507942,
|
| 20 |
+
"reward_std": 0.32721264474093914,
|
| 21 |
+
"rewards/accuracy_reward": 0.7645089328289032,
|
| 22 |
+
"rewards/format_reward": 0.9899553433060646,
|
| 23 |
+
"rewards/log_scaled_reward": 0.3423039447516203,
|
| 24 |
+
"step": 1
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"clip_ratio": 0.0,
|
| 28 |
+
"completion_length": 465.7957763671875,
|
| 29 |
+
"epoch": 0.029850746268656716,
|
| 30 |
+
"grad_norm": 0.5716150403022766,
|
| 31 |
+
"learning_rate": 1.4285714285714285e-07,
|
| 32 |
+
"loss": 0.1518,
|
| 33 |
+
"num_tokens": 1414092.0,
|
| 34 |
+
"reward": 0.1841081934981048,
|
| 35 |
+
"reward_std": 0.6422095634043217,
|
| 36 |
+
"rewards/accuracy_reward": 0.23437499813735485,
|
| 37 |
+
"rewards/format_reward": 0.05022321571595967,
|
| 38 |
+
"rewards/log_scaled_reward": -0.10049002850428224,
|
| 39 |
+
"step": 2
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"clip_ratio": 0.0,
|
| 43 |
+
"completion_length": 482.07144927978516,
|
| 44 |
+
"epoch": 0.04477611940298507,
|
| 45 |
+
"grad_norm": 0.532157301902771,
|
| 46 |
+
"learning_rate": 2.1428571428571426e-07,
|
| 47 |
+
"loss": 0.1335,
|
| 48 |
+
"num_tokens": 1992932.0,
|
| 49 |
+
"reward": 0.10916334297508001,
|
| 50 |
+
"reward_std": 0.6189606413245201,
|
| 51 |
+
"rewards/accuracy_reward": 0.20647321455180645,
|
| 52 |
+
"rewards/format_reward": 0.053571428754366934,
|
| 53 |
+
"rewards/log_scaled_reward": -0.1508813016116619,
|
| 54 |
+
"step": 3
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"clip_ratio": 0.0,
|
| 58 |
+
"completion_length": 529.5971221923828,
|
| 59 |
+
"epoch": 0.05970149253731343,
|
| 60 |
+
"grad_norm": 0.42384153604507446,
|
| 61 |
+
"learning_rate": 2.857142857142857e-07,
|
| 62 |
+
"loss": 0.2014,
|
| 63 |
+
"num_tokens": 2596963.0,
|
| 64 |
+
"reward": 0.05200400925241411,
|
| 65 |
+
"reward_std": 0.582377951592207,
|
| 66 |
+
"rewards/accuracy_reward": 0.18191964272409678,
|
| 67 |
+
"rewards/format_reward": 0.04464285809081048,
|
| 68 |
+
"rewards/log_scaled_reward": -0.17455849051475525,
|
| 69 |
+
"step": 4
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"clip_ratio": 0.0,
|
| 73 |
+
"completion_length": 490.2690010070801,
|
| 74 |
+
"epoch": 0.07462686567164178,
|
| 75 |
+
"grad_norm": 0.42382651567459106,
|
| 76 |
+
"learning_rate": 3.5714285714285716e-07,
|
| 77 |
+
"loss": 0.1441,
|
| 78 |
+
"num_tokens": 3173748.0,
|
| 79 |
+
"reward": 0.10420671524479985,
|
| 80 |
+
"reward_std": 0.5982032977044582,
|
| 81 |
+
"rewards/accuracy_reward": 0.18750000279396772,
|
| 82 |
+
"rewards/format_reward": 0.07366071455180645,
|
| 83 |
+
"rewards/log_scaled_reward": -0.15695400722324848,
|
| 84 |
+
"step": 5
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"clip_ratio": 0.0,
|
| 88 |
+
"completion_length": 450.69086837768555,
|
| 89 |
+
"epoch": 0.08955223880597014,
|
| 90 |
+
"grad_norm": 0.5930183529853821,
|
| 91 |
+
"learning_rate": 4.285714285714285e-07,
|
| 92 |
+
"loss": 0.12,
|
| 93 |
+
"num_tokens": 3702055.0,
|
| 94 |
+
"reward": 0.16118104895576835,
|
| 95 |
+
"reward_std": 0.6403544321656227,
|
| 96 |
+
"rewards/accuracy_reward": 0.20982143096625805,
|
| 97 |
+
"rewards/format_reward": 0.07142857182770967,
|
| 98 |
+
"rewards/log_scaled_reward": -0.12006895546801388,
|
| 99 |
+
"step": 6
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"clip_ratio": 0.0,
|
| 103 |
+
"completion_length": 469.174129486084,
|
| 104 |
+
"epoch": 0.1044776119402985,
|
| 105 |
+
"grad_norm": 0.5980175137519836,
|
| 106 |
+
"learning_rate": 5e-07,
|
| 107 |
+
"loss": 0.1409,
|
| 108 |
+
"num_tokens": 4251483.0,
|
| 109 |
+
"reward": 0.18134657479822636,
|
| 110 |
+
"reward_std": 0.680236354470253,
|
| 111 |
+
"rewards/accuracy_reward": 0.20982142724096775,
|
| 112 |
+
"rewards/format_reward": 0.09821428637951612,
|
| 113 |
+
"rewards/log_scaled_reward": -0.12668914068490267,
|
| 114 |
+
"step": 7
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"clip_ratio": 0.0,
|
| 118 |
+
"completion_length": 501.63505935668945,
|
| 119 |
+
"epoch": 0.11940298507462686,
|
| 120 |
+
"grad_norm": 0.7299549579620361,
|
| 121 |
+
"learning_rate": 5.714285714285714e-07,
|
| 122 |
+
"loss": 0.1195,
|
| 123 |
+
"num_tokens": 4818596.0,
|
| 124 |
+
"reward": 0.20046980120241642,
|
| 125 |
+
"reward_std": 0.67672199010849,
|
| 126 |
+
"rewards/accuracy_reward": 0.2120535708963871,
|
| 127 |
+
"rewards/format_reward": 0.1183035708963871,
|
| 128 |
+
"rewards/log_scaled_reward": -0.1298873471096158,
|
| 129 |
+
"step": 8
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"clip_ratio": 0.0,
|
| 133 |
+
"completion_length": 495.338191986084,
|
| 134 |
+
"epoch": 0.13432835820895522,
|
| 135 |
+
"grad_norm": 0.4346173405647278,
|
| 136 |
+
"learning_rate": 6.428571428571429e-07,
|
| 137 |
+
"loss": 0.1516,
|
| 138 |
+
"num_tokens": 5387755.0,
|
| 139 |
+
"reward": 0.2097643855959177,
|
| 140 |
+
"reward_std": 0.6800287291407585,
|
| 141 |
+
"rewards/accuracy_reward": 0.2220982164144516,
|
| 142 |
+
"rewards/format_reward": 0.10937500093132257,
|
| 143 |
+
"rewards/log_scaled_reward": -0.12170884059742093,
|
| 144 |
+
"step": 9
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"clip_ratio": 0.0,
|
| 148 |
+
"completion_length": 477.33930587768555,
|
| 149 |
+
"epoch": 0.14925373134328357,
|
| 150 |
+
"grad_norm": 3.780118942260742,
|
| 151 |
+
"learning_rate": 7.142857142857143e-07,
|
| 152 |
+
"loss": 0.0884,
|
| 153 |
+
"num_tokens": 5940091.0,
|
| 154 |
+
"reward": 0.31957440078258514,
|
| 155 |
+
"reward_std": 0.7821567356586456,
|
| 156 |
+
"rewards/accuracy_reward": 0.2500000037252903,
|
| 157 |
+
"rewards/format_reward": 0.16183035727590322,
|
| 158 |
+
"rewards/log_scaled_reward": -0.092255964060314,
|
| 159 |
+
"step": 10
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"clip_ratio": 0.0,
|
| 163 |
+
"completion_length": 529.9196739196777,
|
| 164 |
+
"epoch": 0.16417910447761194,
|
| 165 |
+
"grad_norm": 0.7230793833732605,
|
| 166 |
+
"learning_rate": 7.857142857142856e-07,
|
| 167 |
+
"loss": 0.0843,
|
| 168 |
+
"num_tokens": 6544955.0,
|
| 169 |
+
"reward": 0.2673153153154999,
|
| 170 |
+
"reward_std": 0.7341821119189262,
|
| 171 |
+
"rewards/accuracy_reward": 0.21316964086145163,
|
| 172 |
+
"rewards/format_reward": 0.1897321417927742,
|
| 173 |
+
"rewards/log_scaled_reward": -0.13558648666366935,
|
| 174 |
+
"step": 11
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"clip_ratio": 0.0,
|
| 178 |
+
"completion_length": 498.72323989868164,
|
| 179 |
+
"epoch": 0.1791044776119403,
|
| 180 |
+
"grad_norm": 0.9150818586349487,
|
| 181 |
+
"learning_rate": 8.57142857142857e-07,
|
| 182 |
+
"loss": 0.1105,
|
| 183 |
+
"num_tokens": 7121891.0,
|
| 184 |
+
"reward": 0.34613738395273685,
|
| 185 |
+
"reward_std": 0.7445657253265381,
|
| 186 |
+
"rewards/accuracy_reward": 0.2388392873108387,
|
| 187 |
+
"rewards/format_reward": 0.22544642630964518,
|
| 188 |
+
"rewards/log_scaled_reward": -0.11814834456890821,
|
| 189 |
+
"step": 12
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"clip_ratio": 0.0,
|
| 193 |
+
"completion_length": 483.24221420288086,
|
| 194 |
+
"epoch": 0.19402985074626866,
|
| 195 |
+
"grad_norm": 2.0404834747314453,
|
| 196 |
+
"learning_rate": 9.285714285714285e-07,
|
| 197 |
+
"loss": 0.0535,
|
| 198 |
+
"num_tokens": 7678060.0,
|
| 199 |
+
"reward": 0.5936555862426758,
|
| 200 |
+
"reward_std": 0.7919039279222488,
|
| 201 |
+
"rewards/accuracy_reward": 0.2667410708963871,
|
| 202 |
+
"rewards/format_reward": 0.4084821417927742,
|
| 203 |
+
"rewards/log_scaled_reward": -0.08156766439788043,
|
| 204 |
+
"step": 13
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"clip_ratio": 0.0,
|
| 208 |
+
"completion_length": 416.2042541503906,
|
| 209 |
+
"epoch": 0.208955223880597,
|
| 210 |
+
"grad_norm": 3.1881885528564453,
|
| 211 |
+
"learning_rate": 1e-06,
|
| 212 |
+
"loss": 0.0238,
|
| 213 |
+
"num_tokens": 8174851.0,
|
| 214 |
+
"reward": 0.6864497661590576,
|
| 215 |
+
"reward_std": 0.8334432542324066,
|
| 216 |
+
"rewards/accuracy_reward": 0.2354910708963871,
|
| 217 |
+
"rewards/format_reward": 0.5345982126891613,
|
| 218 |
+
"rewards/log_scaled_reward": -0.0836395358783193,
|
| 219 |
+
"step": 14
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"clip_ratio": 0.0,
|
| 223 |
+
"completion_length": 402.8460006713867,
|
| 224 |
+
"epoch": 0.22388059701492538,
|
| 225 |
+
"grad_norm": 1.9057585000991821,
|
| 226 |
+
"learning_rate": 9.998286624877785e-07,
|
| 227 |
+
"loss": 0.0362,
|
| 228 |
+
"num_tokens": 8650305.0,
|
| 229 |
+
"reward": 0.7661240547895432,
|
| 230 |
+
"reward_std": 0.8178009614348412,
|
| 231 |
+
"rewards/accuracy_reward": 0.22656249813735485,
|
| 232 |
+
"rewards/format_reward": 0.6272321343421936,
|
| 233 |
+
"rewards/log_scaled_reward": -0.08767060440732166,
|
| 234 |
+
"step": 15
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"clip_ratio": 0.0,
|
| 238 |
+
"completion_length": 452.4196662902832,
|
| 239 |
+
"epoch": 0.23880597014925373,
|
| 240 |
+
"grad_norm": 25.345657348632812,
|
| 241 |
+
"learning_rate": 9.99314767377287e-07,
|
| 242 |
+
"loss": 0.0285,
|
| 243 |
+
"num_tokens": 9179041.0,
|
| 244 |
+
"reward": 0.875191293656826,
|
| 245 |
+
"reward_std": 0.7929042428731918,
|
| 246 |
+
"rewards/accuracy_reward": 0.25781249813735485,
|
| 247 |
+
"rewards/format_reward": 0.6964285671710968,
|
| 248 |
+
"rewards/log_scaled_reward": -0.07904981379397213,
|
| 249 |
+
"step": 16
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"clip_ratio": 0.0,
|
| 253 |
+
"completion_length": 416.35605239868164,
|
| 254 |
+
"epoch": 0.2537313432835821,
|
| 255 |
+
"grad_norm": 1.1184340715408325,
|
| 256 |
+
"learning_rate": 9.98458666866564e-07,
|
| 257 |
+
"loss": 0.0563,
|
| 258 |
+
"num_tokens": 9701832.0,
|
| 259 |
+
"reward": 0.9071941375732422,
|
| 260 |
+
"reward_std": 0.8025857880711555,
|
| 261 |
+
"rewards/accuracy_reward": 0.24999999813735485,
|
| 262 |
+
"rewards/format_reward": 0.7265625,
|
| 263 |
+
"rewards/log_scaled_reward": -0.06936839601257816,
|
| 264 |
+
"step": 17
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"clip_ratio": 0.0,
|
| 268 |
+
"completion_length": 415.1797065734863,
|
| 269 |
+
"epoch": 0.26865671641791045,
|
| 270 |
+
"grad_norm": 0.4797329604625702,
|
| 271 |
+
"learning_rate": 9.972609476841365e-07,
|
| 272 |
+
"loss": 0.1162,
|
| 273 |
+
"num_tokens": 10210017.0,
|
| 274 |
+
"reward": 0.9835792705416679,
|
| 275 |
+
"reward_std": 0.7636988162994385,
|
| 276 |
+
"rewards/accuracy_reward": 0.24665178544819355,
|
| 277 |
+
"rewards/format_reward": 0.8147321343421936,
|
| 278 |
+
"rewards/log_scaled_reward": -0.07780469593126327,
|
| 279 |
+
"step": 18
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"clip_ratio": 0.0,
|
| 283 |
+
"completion_length": 441.59042739868164,
|
| 284 |
+
"epoch": 0.2835820895522388,
|
| 285 |
+
"grad_norm": 0.4748988151550293,
|
| 286 |
+
"learning_rate": 9.957224306869053e-07,
|
| 287 |
+
"loss": 0.0578,
|
| 288 |
+
"num_tokens": 10730474.0,
|
| 289 |
+
"reward": 1.0904420465230942,
|
| 290 |
+
"reward_std": 0.80109953135252,
|
| 291 |
+
"rewards/accuracy_reward": 0.300223208963871,
|
| 292 |
+
"rewards/format_reward": 0.8158482164144516,
|
| 293 |
+
"rewards/log_scaled_reward": -0.025629449490224943,
|
| 294 |
+
"step": 19
|
| 295 |
+
},
|
| 296 |
+
{
|
| 297 |
+
"clip_ratio": 0.0,
|
| 298 |
+
"completion_length": 426.41743087768555,
|
| 299 |
+
"epoch": 0.29850746268656714,
|
| 300 |
+
"grad_norm": 0.47011756896972656,
|
| 301 |
+
"learning_rate": 9.938441702975689e-07,
|
| 302 |
+
"loss": 0.0503,
|
| 303 |
+
"num_tokens": 11239824.0,
|
| 304 |
+
"reward": 1.2500263825058937,
|
| 305 |
+
"reward_std": 0.8551982864737511,
|
| 306 |
+
"rewards/accuracy_reward": 0.3671875,
|
| 307 |
+
"rewards/format_reward": 0.8526785746216774,
|
| 308 |
+
"rewards/log_scaled_reward": 0.030160245776642114,
|
| 309 |
+
"step": 20
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"clip_ratio": 0.0,
|
| 313 |
+
"completion_length": 459.67078399658203,
|
| 314 |
+
"epoch": 0.31343283582089554,
|
| 315 |
+
"grad_norm": 0.37291309237480164,
|
| 316 |
+
"learning_rate": 9.916274537819773e-07,
|
| 317 |
+
"loss": 0.0366,
|
| 318 |
+
"num_tokens": 11776161.0,
|
| 319 |
+
"reward": 1.3390378654003143,
|
| 320 |
+
"reward_std": 0.8277674093842506,
|
| 321 |
+
"rewards/accuracy_reward": 0.4196428544819355,
|
| 322 |
+
"rewards/format_reward": 0.8526785746216774,
|
| 323 |
+
"rewards/log_scaled_reward": 0.06671636505052447,
|
| 324 |
+
"step": 21
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"clip_ratio": 0.0,
|
| 328 |
+
"completion_length": 416.93639755249023,
|
| 329 |
+
"epoch": 0.3283582089552239,
|
| 330 |
+
"grad_norm": 0.4895838499069214,
|
| 331 |
+
"learning_rate": 9.890738003669027e-07,
|
| 332 |
+
"loss": 0.0473,
|
| 333 |
+
"num_tokens": 12279200.0,
|
| 334 |
+
"reward": 1.5237962007522583,
|
| 335 |
+
"reward_std": 0.8290813863277435,
|
| 336 |
+
"rewards/accuracy_reward": 0.4810267835855484,
|
| 337 |
+
"rewards/format_reward": 0.8950892835855484,
|
| 338 |
+
"rewards/log_scaled_reward": 0.14768002880737185,
|
| 339 |
+
"step": 22
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"clip_ratio": 0.0,
|
| 343 |
+
"completion_length": 409.1294860839844,
|
| 344 |
+
"epoch": 0.34328358208955223,
|
| 345 |
+
"grad_norm": 0.4449058175086975,
|
| 346 |
+
"learning_rate": 9.861849601988383e-07,
|
| 347 |
+
"loss": 0.0255,
|
| 348 |
+
"num_tokens": 12776356.0,
|
| 349 |
+
"reward": 1.5605345666408539,
|
| 350 |
+
"reward_std": 0.8202421888709068,
|
| 351 |
+
"rewards/accuracy_reward": 0.488839291036129,
|
| 352 |
+
"rewards/format_reward": 0.9151785746216774,
|
| 353 |
+
"rewards/log_scaled_reward": 0.15651662228628993,
|
| 354 |
+
"step": 23
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"clip_ratio": 0.0,
|
| 358 |
+
"completion_length": 487.7277069091797,
|
| 359 |
+
"epoch": 0.3582089552238806,
|
| 360 |
+
"grad_norm": 0.33160600066185,
|
| 361 |
+
"learning_rate": 9.82962913144534e-07,
|
| 362 |
+
"loss": 0.0846,
|
| 363 |
+
"num_tokens": 13349480.0,
|
| 364 |
+
"reward": 1.634689912199974,
|
| 365 |
+
"reward_std": 0.7854569926857948,
|
| 366 |
+
"rewards/accuracy_reward": 0.5301339291036129,
|
| 367 |
+
"rewards/format_reward": 0.9151785746216774,
|
| 368 |
+
"rewards/log_scaled_reward": 0.18937731813639402,
|
| 369 |
+
"step": 24
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"clip_ratio": 0.0,
|
| 373 |
+
"completion_length": 451.1272506713867,
|
| 374 |
+
"epoch": 0.373134328358209,
|
| 375 |
+
"grad_norm": 0.3235984444618225,
|
| 376 |
+
"learning_rate": 9.794098674340966e-07,
|
| 377 |
+
"loss": 0.0424,
|
| 378 |
+
"num_tokens": 13868850.0,
|
| 379 |
+
"reward": 1.9291264861822128,
|
| 380 |
+
"reward_std": 0.6902804151177406,
|
| 381 |
+
"rewards/accuracy_reward": 0.6618303433060646,
|
| 382 |
+
"rewards/format_reward": 0.9441964328289032,
|
| 383 |
+
"rewards/log_scaled_reward": 0.32309958525002,
|
| 384 |
+
"step": 25
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"clip_ratio": 0.0,
|
| 388 |
+
"completion_length": 516.2678871154785,
|
| 389 |
+
"epoch": 0.3880597014925373,
|
| 390 |
+
"grad_norm": 2.3632936477661133,
|
| 391 |
+
"learning_rate": 9.755282581475767e-07,
|
| 392 |
+
"loss": 0.1035,
|
| 393 |
+
"num_tokens": 14469026.0,
|
| 394 |
+
"reward": 1.7230691313743591,
|
| 395 |
+
"reward_std": 0.6944096386432648,
|
| 396 |
+
"rewards/accuracy_reward": 0.5714285708963871,
|
| 397 |
+
"rewards/format_reward": 0.9285714253783226,
|
| 398 |
+
"rewards/log_scaled_reward": 0.22306904755532742,
|
| 399 |
+
"step": 26
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"clip_ratio": 0.0,
|
| 403 |
+
"completion_length": 504.8772506713867,
|
| 404 |
+
"epoch": 0.40298507462686567,
|
| 405 |
+
"grad_norm": 0.28064557909965515,
|
| 406 |
+
"learning_rate": 9.713207455460892e-07,
|
| 407 |
+
"loss": 0.0575,
|
| 408 |
+
"num_tokens": 15048084.0,
|
| 409 |
+
"reward": 1.8045607656240463,
|
| 410 |
+
"reward_std": 0.6639576852321625,
|
| 411 |
+
"rewards/accuracy_reward": 0.6127232238650322,
|
| 412 |
+
"rewards/format_reward": 0.9274553582072258,
|
| 413 |
+
"rewards/log_scaled_reward": 0.2643821220844984,
|
| 414 |
+
"step": 27
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"clip_ratio": 0.0,
|
| 418 |
+
"completion_length": 491.07703018188477,
|
| 419 |
+
"epoch": 0.417910447761194,
|
| 420 |
+
"grad_norm": 0.29867812991142273,
|
| 421 |
+
"learning_rate": 9.667902132486008e-07,
|
| 422 |
+
"loss": 0.0641,
|
| 423 |
+
"num_tokens": 15607481.0,
|
| 424 |
+
"reward": 1.9283190667629242,
|
| 425 |
+
"reward_std": 0.627920113503933,
|
| 426 |
+
"rewards/accuracy_reward": 0.6529017835855484,
|
| 427 |
+
"rewards/format_reward": 0.9575892761349678,
|
| 428 |
+
"rewards/log_scaled_reward": 0.3178279069252312,
|
| 429 |
+
"step": 28
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"clip_ratio": 0.0,
|
| 433 |
+
"completion_length": 559.9810523986816,
|
| 434 |
+
"epoch": 0.43283582089552236,
|
| 435 |
+
"grad_norm": 0.2834532558917999,
|
| 436 |
+
"learning_rate": 9.619397662556433e-07,
|
| 437 |
+
"loss": 0.0735,
|
| 438 |
+
"num_tokens": 16248464.0,
|
| 439 |
+
"reward": 1.7917230874300003,
|
| 440 |
+
"reward_std": 0.527396660298109,
|
| 441 |
+
"rewards/accuracy_reward": 0.6037946455180645,
|
| 442 |
+
"rewards/format_reward": 0.954241082072258,
|
| 443 |
+
"rewards/log_scaled_reward": 0.23368733504321426,
|
| 444 |
+
"step": 29
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"clip_ratio": 0.0,
|
| 448 |
+
"completion_length": 510.0111885070801,
|
| 449 |
+
"epoch": 0.44776119402985076,
|
| 450 |
+
"grad_norm": 0.25292134284973145,
|
| 451 |
+
"learning_rate": 9.567727288213004e-07,
|
| 452 |
+
"loss": 0.0778,
|
| 453 |
+
"num_tokens": 16828154.0,
|
| 454 |
+
"reward": 2.033374920487404,
|
| 455 |
+
"reward_std": 0.5300325341522694,
|
| 456 |
+
"rewards/accuracy_reward": 0.7008928507566452,
|
| 457 |
+
"rewards/format_reward": 0.967633917927742,
|
| 458 |
+
"rewards/log_scaled_reward": 0.36484804935753345,
|
| 459 |
+
"step": 30
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"clip_ratio": 0.0,
|
| 463 |
+
"completion_length": 505.8872985839844,
|
| 464 |
+
"epoch": 0.4626865671641791,
|
| 465 |
+
"grad_norm": 0.2629943788051605,
|
| 466 |
+
"learning_rate": 9.512926421749303e-07,
|
| 467 |
+
"loss": 0.0917,
|
| 468 |
+
"num_tokens": 17405221.0,
|
| 469 |
+
"reward": 1.9693890661001205,
|
| 470 |
+
"reward_std": 0.4761252626776695,
|
| 471 |
+
"rewards/accuracy_reward": 0.6595982164144516,
|
| 472 |
+
"rewards/format_reward": 0.9743303507566452,
|
| 473 |
+
"rewards/log_scaled_reward": 0.33546042814850807,
|
| 474 |
+
"step": 31
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"clip_ratio": 0.0,
|
| 478 |
+
"completion_length": 468.31809997558594,
|
| 479 |
+
"epoch": 0.47761194029850745,
|
| 480 |
+
"grad_norm": 0.25987353920936584,
|
| 481 |
+
"learning_rate": 9.455032620941839e-07,
|
| 482 |
+
"loss": 0.116,
|
| 483 |
+
"num_tokens": 17953570.0,
|
| 484 |
+
"reward": 2.0616614371538162,
|
| 485 |
+
"reward_std": 0.49665234982967377,
|
| 486 |
+
"rewards/accuracy_reward": 0.7075892761349678,
|
| 487 |
+
"rewards/format_reward": 0.962053582072258,
|
| 488 |
+
"rewards/log_scaled_reward": 0.39201846718788147,
|
| 489 |
+
"step": 32
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"clip_ratio": 0.0,
|
| 493 |
+
"completion_length": 504.4788246154785,
|
| 494 |
+
"epoch": 0.4925373134328358,
|
| 495 |
+
"grad_norm": 0.27474886178970337,
|
| 496 |
+
"learning_rate": 9.394085563309826e-07,
|
| 497 |
+
"loss": 0.1112,
|
| 498 |
+
"num_tokens": 18531095.0,
|
| 499 |
+
"reward": 1.9522841572761536,
|
| 500 |
+
"reward_std": 0.5375584103167057,
|
| 501 |
+
"rewards/accuracy_reward": 0.6607142835855484,
|
| 502 |
+
"rewards/format_reward": 0.965401791036129,
|
| 503 |
+
"rewards/log_scaled_reward": 0.32616803981363773,
|
| 504 |
+
"step": 33
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"clip_ratio": 0.0,
|
| 508 |
+
"completion_length": 498.3616371154785,
|
| 509 |
+
"epoch": 0.5074626865671642,
|
| 510 |
+
"grad_norm": 2.524815320968628,
|
| 511 |
+
"learning_rate": 9.330127018922193e-07,
|
| 512 |
+
"loss": 0.1055,
|
| 513 |
+
"num_tokens": 19108059.0,
|
| 514 |
+
"reward": 1.9061091989278793,
|
| 515 |
+
"reward_std": 0.5573387667536736,
|
| 516 |
+
"rewards/accuracy_reward": 0.6339285708963871,
|
| 517 |
+
"rewards/format_reward": 0.9642857164144516,
|
| 518 |
+
"rewards/log_scaled_reward": 0.30789486039429903,
|
| 519 |
+
"step": 34
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"clip_ratio": 0.0,
|
| 523 |
+
"completion_length": 444.20873260498047,
|
| 524 |
+
"epoch": 0.5223880597014925,
|
| 525 |
+
"grad_norm": 0.3097524046897888,
|
| 526 |
+
"learning_rate": 9.26320082177046e-07,
|
| 527 |
+
"loss": 0.1342,
|
| 528 |
+
"num_tokens": 19646062.0,
|
| 529 |
+
"reward": 1.9290964603424072,
|
| 530 |
+
"reward_std": 0.5330292023718357,
|
| 531 |
+
"rewards/accuracy_reward": 0.6272321492433548,
|
| 532 |
+
"rewards/format_reward": 0.9765625,
|
| 533 |
+
"rewards/log_scaled_reward": 0.325301731005311,
|
| 534 |
+
"step": 35
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"clip_ratio": 0.0,
|
| 538 |
+
"completion_length": 420.0848388671875,
|
| 539 |
+
"epoch": 0.5373134328358209,
|
| 540 |
+
"grad_norm": 0.3493054211139679,
|
| 541 |
+
"learning_rate": 9.19335283972712e-07,
|
| 542 |
+
"loss": 0.1097,
|
| 543 |
+
"num_tokens": 20165994.0,
|
| 544 |
+
"reward": 1.937290906906128,
|
| 545 |
+
"reward_std": 0.5374783836305141,
|
| 546 |
+
"rewards/accuracy_reward": 0.621651791036129,
|
| 547 |
+
"rewards/format_reward": 0.9732142835855484,
|
| 548 |
+
"rewards/log_scaled_reward": 0.34242471773177385,
|
| 549 |
+
"step": 36
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"clip_ratio": 0.0,
|
| 553 |
+
"completion_length": 367.64064025878906,
|
| 554 |
+
"epoch": 0.5522388059701493,
|
| 555 |
+
"grad_norm": 0.39517220854759216,
|
| 556 |
+
"learning_rate": 9.120630943110077e-07,
|
| 557 |
+
"loss": 0.1309,
|
| 558 |
+
"num_tokens": 20622824.0,
|
| 559 |
+
"reward": 1.9518826305866241,
|
| 560 |
+
"reward_std": 0.5057090371847153,
|
| 561 |
+
"rewards/accuracy_reward": 0.6294642947614193,
|
| 562 |
+
"rewards/format_reward": 0.9765625074505806,
|
| 563 |
+
"rewards/log_scaled_reward": 0.34585576388053596,
|
| 564 |
+
"step": 37
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"clip_ratio": 0.0,
|
| 568 |
+
"completion_length": 344.1038112640381,
|
| 569 |
+
"epoch": 0.5671641791044776,
|
| 570 |
+
"grad_norm": 1.3338432312011719,
|
| 571 |
+
"learning_rate": 9.045084971874737e-07,
|
| 572 |
+
"loss": 0.1333,
|
| 573 |
+
"num_tokens": 21076925.0,
|
| 574 |
+
"reward": 1.9407142996788025,
|
| 575 |
+
"reward_std": 0.5814780332148075,
|
| 576 |
+
"rewards/accuracy_reward": 0.6127232126891613,
|
| 577 |
+
"rewards/format_reward": 0.9709821417927742,
|
| 578 |
+
"rewards/log_scaled_reward": 0.35700881760567427,
|
| 579 |
+
"step": 38
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"clip_ratio": 0.0,
|
| 583 |
+
"completion_length": 277.19309425354004,
|
| 584 |
+
"epoch": 0.582089552238806,
|
| 585 |
+
"grad_norm": 0.4697703421115875,
|
| 586 |
+
"learning_rate": 8.966766701456176e-07,
|
| 587 |
+
"loss": 0.1089,
|
| 588 |
+
"num_tokens": 21451954.0,
|
| 589 |
+
"reward": 1.8939976394176483,
|
| 590 |
+
"reward_std": 0.531686820089817,
|
| 591 |
+
"rewards/accuracy_reward": 0.5703125037252903,
|
| 592 |
+
"rewards/format_reward": 0.984375,
|
| 593 |
+
"rewards/log_scaled_reward": 0.33931003510951996,
|
| 594 |
+
"step": 39
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"clip_ratio": 0.0,
|
| 598 |
+
"completion_length": 292.2109489440918,
|
| 599 |
+
"epoch": 0.5970149253731343,
|
| 600 |
+
"grad_norm": 0.42749401926994324,
|
| 601 |
+
"learning_rate": 8.885729807284854e-07,
|
| 602 |
+
"loss": 0.1051,
|
| 603 |
+
"num_tokens": 21837471.0,
|
| 604 |
+
"reward": 1.7810039222240448,
|
| 605 |
+
"reward_std": 0.5695139020681381,
|
| 606 |
+
"rewards/accuracy_reward": 0.5212053544819355,
|
| 607 |
+
"rewards/format_reward": 0.9832589253783226,
|
| 608 |
+
"rewards/log_scaled_reward": 0.2765395335154608,
|
| 609 |
+
"step": 40
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"clip_ratio": 0.0,
|
| 613 |
+
"completion_length": 227.70425033569336,
|
| 614 |
+
"epoch": 0.6119402985074627,
|
| 615 |
+
"grad_norm": 0.8301727771759033,
|
| 616 |
+
"learning_rate": 8.802029828000155e-07,
|
| 617 |
+
"loss": 0.1365,
|
| 618 |
+
"num_tokens": 22182102.0,
|
| 619 |
+
"reward": 1.7540639638900757,
|
| 620 |
+
"reward_std": 0.504084050655365,
|
| 621 |
+
"rewards/accuracy_reward": 0.4899553582072258,
|
| 622 |
+
"rewards/format_reward": 0.9754464253783226,
|
| 623 |
+
"rewards/log_scaled_reward": 0.2886621206998825,
|
| 624 |
+
"step": 41
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"clip_ratio": 0.0,
|
| 628 |
+
"completion_length": 186.12500953674316,
|
| 629 |
+
"epoch": 0.6268656716417911,
|
| 630 |
+
"grad_norm": 0.6249210238456726,
|
| 631 |
+
"learning_rate": 8.71572412738697e-07,
|
| 632 |
+
"loss": 0.1336,
|
| 633 |
+
"num_tokens": 22471350.0,
|
| 634 |
+
"reward": 1.9936908185482025,
|
| 635 |
+
"reward_std": 0.6450418382883072,
|
| 636 |
+
"rewards/accuracy_reward": 0.5926339402794838,
|
| 637 |
+
"rewards/format_reward": 0.9888392761349678,
|
| 638 |
+
"rewards/log_scaled_reward": 0.41221751645207405,
|
| 639 |
+
"step": 42
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"clip_ratio": 0.0,
|
| 643 |
+
"completion_length": 169.09710693359375,
|
| 644 |
+
"epoch": 0.6417910447761194,
|
| 645 |
+
"grad_norm": 0.7566676139831543,
|
| 646 |
+
"learning_rate": 8.626871855061437e-07,
|
| 647 |
+
"loss": 0.1662,
|
| 648 |
+
"num_tokens": 22758477.0,
|
| 649 |
+
"reward": 1.8480805903673172,
|
| 650 |
+
"reward_std": 0.5597276613116264,
|
| 651 |
+
"rewards/accuracy_reward": 0.511160708963871,
|
| 652 |
+
"rewards/format_reward": 0.9944196343421936,
|
| 653 |
+
"rewards/log_scaled_reward": 0.3425001185387373,
|
| 654 |
+
"step": 43
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"clip_ratio": 0.0,
|
| 658 |
+
"completion_length": 152.94978141784668,
|
| 659 |
+
"epoch": 0.6567164179104478,
|
| 660 |
+
"grad_norm": 0.9348928332328796,
|
| 661 |
+
"learning_rate": 8.535533905932737e-07,
|
| 662 |
+
"loss": 0.1287,
|
| 663 |
+
"num_tokens": 23022432.0,
|
| 664 |
+
"reward": 1.8545437455177307,
|
| 665 |
+
"reward_std": 0.5482046529650688,
|
| 666 |
+
"rewards/accuracy_reward": 0.5078125037252903,
|
| 667 |
+
"rewards/format_reward": 0.9921874925494194,
|
| 668 |
+
"rewards/log_scaled_reward": 0.35454366356134415,
|
| 669 |
+
"step": 44
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"clip_ratio": 0.0,
|
| 673 |
+
"completion_length": 128.6517915725708,
|
| 674 |
+
"epoch": 0.6716417910447762,
|
| 675 |
+
"grad_norm": 1.1114394664764404,
|
| 676 |
+
"learning_rate": 8.441772878468769e-07,
|
| 677 |
+
"loss": 0.1287,
|
| 678 |
+
"num_tokens": 23263912.0,
|
| 679 |
+
"reward": 1.7271955758333206,
|
| 680 |
+
"reward_std": 0.5110182501375675,
|
| 681 |
+
"rewards/accuracy_reward": 0.4330357164144516,
|
| 682 |
+
"rewards/format_reward": 0.9977678507566452,
|
| 683 |
+
"rewards/log_scaled_reward": 0.29639193043112755,
|
| 684 |
+
"step": 45
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"clip_ratio": 0.0,
|
| 688 |
+
"completion_length": 143.97545337677002,
|
| 689 |
+
"epoch": 0.6865671641791045,
|
| 690 |
+
"grad_norm": 1.1403673887252808,
|
| 691 |
+
"learning_rate": 8.34565303179429e-07,
|
| 692 |
+
"loss": 0.1742,
|
| 693 |
+
"num_tokens": 23526706.0,
|
| 694 |
+
"reward": 1.6235045939683914,
|
| 695 |
+
"reward_std": 0.5148132182657719,
|
| 696 |
+
"rewards/accuracy_reward": 0.386160708963871,
|
| 697 |
+
"rewards/format_reward": 0.987723208963871,
|
| 698 |
+
"rewards/log_scaled_reward": 0.24962060060352087,
|
| 699 |
+
"step": 46
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"clip_ratio": 0.0,
|
| 703 |
+
"completion_length": 104.42187976837158,
|
| 704 |
+
"epoch": 0.7014925373134329,
|
| 705 |
+
"grad_norm": 1.6832914352416992,
|
| 706 |
+
"learning_rate": 8.247240241650917e-07,
|
| 707 |
+
"loss": 0.1296,
|
| 708 |
+
"num_tokens": 23736324.0,
|
| 709 |
+
"reward": 1.690138816833496,
|
| 710 |
+
"reward_std": 0.5133109800517559,
|
| 711 |
+
"rewards/accuracy_reward": 0.4051339328289032,
|
| 712 |
+
"rewards/format_reward": 0.9933035671710968,
|
| 713 |
+
"rewards/log_scaled_reward": 0.29170125164091587,
|
| 714 |
+
"step": 47
|
| 715 |
+
},
|
| 716 |
+
{
|
| 717 |
+
"clip_ratio": 0.0,
|
| 718 |
+
"completion_length": 92.65848731994629,
|
| 719 |
+
"epoch": 0.7164179104477612,
|
| 720 |
+
"grad_norm": 1.6501548290252686,
|
| 721 |
+
"learning_rate": 8.146601955249187e-07,
|
| 722 |
+
"loss": 0.1411,
|
| 723 |
+
"num_tokens": 23946162.0,
|
| 724 |
+
"reward": 1.6304273456335068,
|
| 725 |
+
"reward_std": 0.49451132118701935,
|
| 726 |
+
"rewards/accuracy_reward": 0.36941964365541935,
|
| 727 |
+
"rewards/format_reward": 0.995535708963871,
|
| 728 |
+
"rewards/log_scaled_reward": 0.26547193340957165,
|
| 729 |
+
"step": 48
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"clip_ratio": 0.0,
|
| 733 |
+
"completion_length": 78.72098636627197,
|
| 734 |
+
"epoch": 0.7313432835820896,
|
| 735 |
+
"grad_norm": 2.571906566619873,
|
| 736 |
+
"learning_rate": 8.043807145043603e-07,
|
| 737 |
+
"loss": 0.1793,
|
| 738 |
+
"num_tokens": 24153096.0,
|
| 739 |
+
"reward": 1.581173524260521,
|
| 740 |
+
"reward_std": 0.43424950167536736,
|
| 741 |
+
"rewards/accuracy_reward": 0.33748282864689827,
|
| 742 |
+
"rewards/format_reward": 0.9966517835855484,
|
| 743 |
+
"rewards/log_scaled_reward": 0.24970022030174732,
|
| 744 |
+
"step": 49
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"clip_ratio": 0.0,
|
| 748 |
+
"completion_length": 64.70759201049805,
|
| 749 |
+
"epoch": 0.746268656716418,
|
| 750 |
+
"grad_norm": 2.7052981853485107,
|
| 751 |
+
"learning_rate": 7.938926261462365e-07,
|
| 752 |
+
"loss": 0.1791,
|
| 753 |
+
"num_tokens": 24349226.0,
|
| 754 |
+
"reward": 1.5854334235191345,
|
| 755 |
+
"reward_std": 0.3977060168981552,
|
| 756 |
+
"rewards/accuracy_reward": 0.33593750186264515,
|
| 757 |
+
"rewards/format_reward": 0.9955357015132904,
|
| 758 |
+
"rewards/log_scaled_reward": 0.2539601270109415,
|
| 759 |
+
"step": 50
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"clip_ratio": 0.0,
|
| 763 |
+
"completion_length": 46.40067148208618,
|
| 764 |
+
"epoch": 0.7611940298507462,
|
| 765 |
+
"grad_norm": 6.834611415863037,
|
| 766 |
+
"learning_rate": 7.832031184624164e-07,
|
| 767 |
+
"loss": 0.1055,
|
| 768 |
+
"num_tokens": 24528585.0,
|
| 769 |
+
"reward": 1.5368833392858505,
|
| 770 |
+
"reward_std": 0.47740813344717026,
|
| 771 |
+
"rewards/accuracy_reward": 0.30133928544819355,
|
| 772 |
+
"rewards/format_reward": 0.9988839253783226,
|
| 773 |
+
"rewards/log_scaled_reward": 0.2366600539535284,
|
| 774 |
+
"step": 51
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"clip_ratio": 0.0,
|
| 778 |
+
"completion_length": 38.36384057998657,
|
| 779 |
+
"epoch": 0.7761194029850746,
|
| 780 |
+
"grad_norm": 4.969581127166748,
|
| 781 |
+
"learning_rate": 7.723195175075135e-07,
|
| 782 |
+
"loss": 0.1206,
|
| 783 |
+
"num_tokens": 24691607.0,
|
| 784 |
+
"reward": 1.415985830128193,
|
| 785 |
+
"reward_std": 0.29450324457138777,
|
| 786 |
+
"rewards/accuracy_reward": 0.2377232169965282,
|
| 787 |
+
"rewards/format_reward": 0.9977678507566452,
|
| 788 |
+
"rewards/log_scaled_reward": 0.18049467215314507,
|
| 789 |
+
"step": 52
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"clip_ratio": 0.0,
|
| 793 |
+
"completion_length": 33.99218940734863,
|
| 794 |
+
"epoch": 0.7910447761194029,
|
| 795 |
+
"grad_norm": 4.195903778076172,
|
| 796 |
+
"learning_rate": 7.612492823579744e-07,
|
| 797 |
+
"loss": 0.145,
|
| 798 |
+
"num_tokens": 24849736.0,
|
| 799 |
+
"reward": 1.5035328567028046,
|
| 800 |
+
"reward_std": 0.34667503647506237,
|
| 801 |
+
"rewards/accuracy_reward": 0.27790178544819355,
|
| 802 |
+
"rewards/format_reward": 0.9988839253783226,
|
| 803 |
+
"rewards/log_scaled_reward": 0.2267470918595791,
|
| 804 |
+
"step": 53
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"clip_ratio": 0.0,
|
| 808 |
+
"completion_length": 26.474331378936768,
|
| 809 |
+
"epoch": 0.8059701492537313,
|
| 810 |
+
"grad_norm": 24.98171615600586,
|
| 811 |
+
"learning_rate": 7.5e-07,
|
| 812 |
+
"loss": 0.0613,
|
| 813 |
+
"num_tokens": 25008289.0,
|
| 814 |
+
"reward": 1.5673803389072418,
|
| 815 |
+
"reward_std": 0.28740744665265083,
|
| 816 |
+
"rewards/accuracy_reward": 0.3069196417927742,
|
| 817 |
+
"rewards/format_reward": 1.0,
|
| 818 |
+
"rewards/log_scaled_reward": 0.26046060863882303,
|
| 819 |
+
"step": 54
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"clip_ratio": 0.0,
|
| 823 |
+
"completion_length": 25.195313692092896,
|
| 824 |
+
"epoch": 0.8208955223880597,
|
| 825 |
+
"grad_norm": 5.9815216064453125,
|
| 826 |
+
"learning_rate": 7.385793801298042e-07,
|
| 827 |
+
"loss": 0.0376,
|
| 828 |
+
"num_tokens": 25169984.0,
|
| 829 |
+
"reward": 1.4668240398168564,
|
| 830 |
+
"reward_std": 0.3536365833133459,
|
| 831 |
+
"rewards/accuracy_reward": 0.25369161926209927,
|
| 832 |
+
"rewards/format_reward": 1.0,
|
| 833 |
+
"rewards/log_scaled_reward": 0.2157079027965665,
|
| 834 |
+
"step": 55
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"clip_ratio": 0.0,
|
| 838 |
+
"completion_length": 23.91294765472412,
|
| 839 |
+
"epoch": 0.835820895522388,
|
| 840 |
+
"grad_norm": 7.0327372550964355,
|
| 841 |
+
"learning_rate": 7.269952498697734e-07,
|
| 842 |
+
"loss": 0.0386,
|
| 843 |
+
"num_tokens": 25322466.0,
|
| 844 |
+
"reward": 1.4718168079853058,
|
| 845 |
+
"reward_std": 0.3211175389587879,
|
| 846 |
+
"rewards/accuracy_reward": 0.2578124962747097,
|
| 847 |
+
"rewards/format_reward": 1.0,
|
| 848 |
+
"rewards/log_scaled_reward": 0.21400425024330616,
|
| 849 |
+
"step": 56
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"clip_ratio": 0.0,
|
| 853 |
+
"completion_length": 23.36049222946167,
|
| 854 |
+
"epoch": 0.8507462686567164,
|
| 855 |
+
"grad_norm": 5.700937271118164,
|
| 856 |
+
"learning_rate": 7.152555484041475e-07,
|
| 857 |
+
"loss": 0.0265,
|
| 858 |
+
"num_tokens": 25466397.0,
|
| 859 |
+
"reward": 1.5502240508794785,
|
| 860 |
+
"reward_std": 0.2749287262558937,
|
| 861 |
+
"rewards/accuracy_reward": 0.296875,
|
| 862 |
+
"rewards/format_reward": 0.9988839253783226,
|
| 863 |
+
"rewards/log_scaled_reward": 0.25446509197354317,
|
| 864 |
+
"step": 57
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"clip_ratio": 0.0,
|
| 868 |
+
"completion_length": 23.949777364730835,
|
| 869 |
+
"epoch": 0.8656716417910447,
|
| 870 |
+
"grad_norm": 5.249198913574219,
|
| 871 |
+
"learning_rate": 7.033683215379002e-07,
|
| 872 |
+
"loss": 0.023,
|
| 873 |
+
"num_tokens": 25613824.0,
|
| 874 |
+
"reward": 1.5308541655540466,
|
| 875 |
+
"reward_std": 0.26417338382452726,
|
| 876 |
+
"rewards/accuracy_reward": 0.28794642724096775,
|
| 877 |
+
"rewards/format_reward": 0.9988839253783226,
|
| 878 |
+
"rewards/log_scaled_reward": 0.2440237421542406,
|
| 879 |
+
"step": 58
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"clip_ratio": 0.0,
|
| 883 |
+
"completion_length": 23.375000953674316,
|
| 884 |
+
"epoch": 0.8805970149253731,
|
| 885 |
+
"grad_norm": 8.684850692749023,
|
| 886 |
+
"learning_rate": 6.913417161825449e-07,
|
| 887 |
+
"loss": 0.0289,
|
| 888 |
+
"num_tokens": 25760176.0,
|
| 889 |
+
"reward": 1.566646233201027,
|
| 890 |
+
"reward_std": 0.22551130689680576,
|
| 891 |
+
"rewards/accuracy_reward": 0.30691963620483875,
|
| 892 |
+
"rewards/format_reward": 0.9966517761349678,
|
| 893 |
+
"rewards/log_scaled_reward": 0.26307476311922073,
|
| 894 |
+
"step": 59
|
| 895 |
+
},
|
| 896 |
+
{
|
| 897 |
+
"clip_ratio": 0.0,
|
| 898 |
+
"completion_length": 21.989956378936768,
|
| 899 |
+
"epoch": 0.8955223880597015,
|
| 900 |
+
"grad_norm": 34.08607864379883,
|
| 901 |
+
"learning_rate": 6.7918397477265e-07,
|
| 902 |
+
"loss": 0.0144,
|
| 903 |
+
"num_tokens": 25914911.0,
|
| 904 |
+
"reward": 1.628076210618019,
|
| 905 |
+
"reward_std": 0.3200679961591959,
|
| 906 |
+
"rewards/accuracy_reward": 0.33705356903374195,
|
| 907 |
+
"rewards/format_reward": 0.9966517761349678,
|
| 908 |
+
"rewards/log_scaled_reward": 0.2943707942031324,
|
| 909 |
+
"step": 60
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"clip_ratio": 0.0,
|
| 913 |
+
"completion_length": 22.574777841567993,
|
| 914 |
+
"epoch": 0.9104477611940298,
|
| 915 |
+
"grad_norm": 5.27999210357666,
|
| 916 |
+
"learning_rate": 6.669034296168854e-07,
|
| 917 |
+
"loss": 0.0217,
|
| 918 |
+
"num_tokens": 26077314.0,
|
| 919 |
+
"reward": 1.6822472661733627,
|
| 920 |
+
"reward_std": 0.2845423389226198,
|
| 921 |
+
"rewards/accuracy_reward": 0.3649553582072258,
|
| 922 |
+
"rewards/format_reward": 0.995535708963871,
|
| 923 |
+
"rewards/log_scaled_reward": 0.3217560853809118,
|
| 924 |
+
"step": 61
|
| 925 |
+
},
|
| 926 |
+
{
|
| 927 |
+
"clip_ratio": 0.0,
|
| 928 |
+
"completion_length": 22.198662042617798,
|
| 929 |
+
"epoch": 0.9253731343283582,
|
| 930 |
+
"grad_norm": 4.5566277503967285,
|
| 931 |
+
"learning_rate": 6.545084971874736e-07,
|
| 932 |
+
"loss": 0.0064,
|
| 933 |
+
"num_tokens": 26222620.0,
|
| 934 |
+
"reward": 1.5729791224002838,
|
| 935 |
+
"reward_std": 0.22709419997408986,
|
| 936 |
+
"rewards/accuracy_reward": 0.31138393096625805,
|
| 937 |
+
"rewards/format_reward": 0.9921874925494194,
|
| 938 |
+
"rewards/log_scaled_reward": 0.2694076579064131,
|
| 939 |
+
"step": 62
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"clip_ratio": 0.0,
|
| 943 |
+
"completion_length": 21.32924246788025,
|
| 944 |
+
"epoch": 0.9402985074626866,
|
| 945 |
+
"grad_norm": 3.140645980834961,
|
| 946 |
+
"learning_rate": 6.420076723519614e-07,
|
| 947 |
+
"loss": 0.0068,
|
| 948 |
+
"num_tokens": 26371243.0,
|
| 949 |
+
"reward": 1.6745910048484802,
|
| 950 |
+
"reward_std": 0.20502757839858532,
|
| 951 |
+
"rewards/accuracy_reward": 0.3582589291036129,
|
| 952 |
+
"rewards/format_reward": 1.0,
|
| 953 |
+
"rewards/log_scaled_reward": 0.3163319919258356,
|
| 954 |
+
"step": 63
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"clip_ratio": 0.0,
|
| 958 |
+
"completion_length": 21.34709930419922,
|
| 959 |
+
"epoch": 0.9552238805970149,
|
| 960 |
+
"grad_norm": 4.071537494659424,
|
| 961 |
+
"learning_rate": 6.294095225512604e-07,
|
| 962 |
+
"loss": 0.0055,
|
| 963 |
+
"num_tokens": 26527050.0,
|
| 964 |
+
"reward": 1.5537814646959305,
|
| 965 |
+
"reward_std": 0.30234235525131226,
|
| 966 |
+
"rewards/accuracy_reward": 0.2979910681024194,
|
| 967 |
+
"rewards/format_reward": 1.0,
|
| 968 |
+
"rewards/log_scaled_reward": 0.25579037982970476,
|
| 969 |
+
"step": 64
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"clip_ratio": 0.0,
|
| 973 |
+
"completion_length": 21.24776864051819,
|
| 974 |
+
"epoch": 0.9701492537313433,
|
| 975 |
+
"grad_norm": 4.297541618347168,
|
| 976 |
+
"learning_rate": 6.167226819279527e-07,
|
| 977 |
+
"loss": 0.0114,
|
| 978 |
+
"num_tokens": 26679280.0,
|
| 979 |
+
"reward": 1.5483618080615997,
|
| 980 |
+
"reward_std": 0.1530774086713791,
|
| 981 |
+
"rewards/accuracy_reward": 0.2968749953433871,
|
| 982 |
+
"rewards/format_reward": 0.9966517835855484,
|
| 983 |
+
"rewards/log_scaled_reward": 0.25483495742082596,
|
| 984 |
+
"step": 65
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"clip_ratio": 0.0,
|
| 988 |
+
"completion_length": 21.180555820465088,
|
| 989 |
+
"epoch": 0.9850746268656716,
|
| 990 |
+
"grad_norm": 3.6126675605773926,
|
| 991 |
+
"learning_rate": 6.039558454088795e-07,
|
| 992 |
+
"loss": 0.009,
|
| 993 |
+
"num_tokens": 26828211.0,
|
| 994 |
+
"reward": 1.6490006893873215,
|
| 995 |
+
"reward_std": 0.29192496836185455,
|
| 996 |
+
"rewards/accuracy_reward": 0.3459821417927742,
|
| 997 |
+
"rewards/format_reward": 0.9988839253783226,
|
| 998 |
+
"rewards/log_scaled_reward": 0.304134588688612,
|
| 999 |
+
"step": 66
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"clip_ratio": 0.0,
|
| 1003 |
+
"completion_length": 20.77455449104309,
|
| 1004 |
+
"epoch": 1.0149253731343284,
|
| 1005 |
+
"grad_norm": 4.067701816558838,
|
| 1006 |
+
"learning_rate": 5.911177627460738e-07,
|
| 1007 |
+
"loss": 0.0094,
|
| 1008 |
+
"num_tokens": 26965209.0,
|
| 1009 |
+
"reward": 1.7030873149633408,
|
| 1010 |
+
"reward_std": 0.2512226551771164,
|
| 1011 |
+
"rewards/accuracy_reward": 0.37388391979038715,
|
| 1012 |
+
"rewards/format_reward": 0.9966517761349678,
|
| 1013 |
+
"rewards/log_scaled_reward": 0.3325514607131481,
|
| 1014 |
+
"step": 67
|
| 1015 |
+
},
|
| 1016 |
+
{
|
| 1017 |
+
"clip_ratio": 0.0,
|
| 1018 |
+
"completion_length": 20.35267925262451,
|
| 1019 |
+
"epoch": 1.0298507462686568,
|
| 1020 |
+
"grad_norm": 3.4311537742614746,
|
| 1021 |
+
"learning_rate": 5.782172325201155e-07,
|
| 1022 |
+
"loss": 0.0126,
|
| 1023 |
+
"num_tokens": 27116397.0,
|
| 1024 |
+
"reward": 1.5380910784006119,
|
| 1025 |
+
"reward_std": 0.20529233757406473,
|
| 1026 |
+
"rewards/accuracy_reward": 0.2901785708963871,
|
| 1027 |
+
"rewards/format_reward": 0.9988839253783226,
|
| 1028 |
+
"rewards/log_scaled_reward": 0.24902847222983837,
|
| 1029 |
+
"step": 68
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"clip_ratio": 0.0,
|
| 1033 |
+
"completion_length": 20.400670528411865,
|
| 1034 |
+
"epoch": 1.044776119402985,
|
| 1035 |
+
"grad_norm": 4.965053081512451,
|
| 1036 |
+
"learning_rate": 5.652630961100258e-07,
|
| 1037 |
+
"loss": 0.0182,
|
| 1038 |
+
"num_tokens": 27260012.0,
|
| 1039 |
+
"reward": 1.4891109764575958,
|
| 1040 |
+
"reward_std": 0.2253081511007622,
|
| 1041 |
+
"rewards/accuracy_reward": 0.26562499813735485,
|
| 1042 |
+
"rewards/format_reward": 0.9988839253783226,
|
| 1043 |
+
"rewards/log_scaled_reward": 0.22460200637578964,
|
| 1044 |
+
"step": 69
|
| 1045 |
+
},
|
| 1046 |
+
{
|
| 1047 |
+
"clip_ratio": 0.0,
|
| 1048 |
+
"completion_length": 19.54799175262451,
|
| 1049 |
+
"epoch": 1.0597014925373134,
|
| 1050 |
+
"grad_norm": 9.666987419128418,
|
| 1051 |
+
"learning_rate": 5.522642316338268e-07,
|
| 1052 |
+
"loss": 0.0249,
|
| 1053 |
+
"num_tokens": 27404823.0,
|
| 1054 |
+
"reward": 1.5430727303028107,
|
| 1055 |
+
"reward_std": 0.2664187829941511,
|
| 1056 |
+
"rewards/accuracy_reward": 0.2924107164144516,
|
| 1057 |
+
"rewards/format_reward": 0.9977678507566452,
|
| 1058 |
+
"rewards/log_scaled_reward": 0.25289412308484316,
|
| 1059 |
+
"step": 70
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"clip_ratio": 0.0,
|
| 1063 |
+
"completion_length": 18.02120590209961,
|
| 1064 |
+
"epoch": 1.0746268656716418,
|
| 1065 |
+
"grad_norm": 15.713820457458496,
|
| 1066 |
+
"learning_rate": 5.392295478639225e-07,
|
| 1067 |
+
"loss": 0.0167,
|
| 1068 |
+
"num_tokens": 27555962.0,
|
| 1069 |
+
"reward": 1.1339266449213028,
|
| 1070 |
+
"reward_std": 0.34662946686148643,
|
| 1071 |
+
"rewards/accuracy_reward": 0.1026785708963871,
|
| 1072 |
+
"rewards/format_reward": 0.9676339328289032,
|
| 1073 |
+
"rewards/log_scaled_reward": 0.06361408122756984,
|
| 1074 |
+
"step": 71
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"clip_ratio": 0.0,
|
| 1078 |
+
"completion_length": 19.39732265472412,
|
| 1079 |
+
"epoch": 1.0895522388059702,
|
| 1080 |
+
"grad_norm": 21.441844940185547,
|
| 1081 |
+
"learning_rate": 5.26167978121472e-07,
|
| 1082 |
+
"loss": 0.0653,
|
| 1083 |
+
"num_tokens": 27696134.0,
|
| 1084 |
+
"reward": 1.0387937128543854,
|
| 1085 |
+
"reward_std": 0.14305981155484915,
|
| 1086 |
+
"rewards/accuracy_reward": 0.04464285704307258,
|
| 1087 |
+
"rewards/format_reward": 0.9866071417927742,
|
| 1088 |
+
"rewards/log_scaled_reward": 0.0075436777296999935,
|
| 1089 |
+
"step": 72
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"clip_ratio": 0.0,
|
| 1093 |
+
"completion_length": 15.95089340209961,
|
| 1094 |
+
"epoch": 1.1044776119402986,
|
| 1095 |
+
"grad_norm": 12.856225967407227,
|
| 1096 |
+
"learning_rate": 5.130884741539366e-07,
|
| 1097 |
+
"loss": 0.018,
|
| 1098 |
+
"num_tokens": 27843106.0,
|
| 1099 |
+
"reward": 0.9690398126840591,
|
| 1100 |
+
"reward_std": 0.01615892370318761,
|
| 1101 |
+
"rewards/accuracy_reward": 0.0022321429569274187,
|
| 1102 |
+
"rewards/format_reward": 0.9988839253783226,
|
| 1103 |
+
"rewards/log_scaled_reward": -0.03207633784040809,
|
| 1104 |
+
"step": 73
|
| 1105 |
+
},
|
| 1106 |
+
{
|
| 1107 |
+
"clip_ratio": 0.0,
|
| 1108 |
+
"completion_length": 15.233259677886963,
|
| 1109 |
+
"epoch": 1.1194029850746268,
|
| 1110 |
+
"grad_norm": 1.9851570129394531,
|
| 1111 |
+
"learning_rate": 5e-07,
|
| 1112 |
+
"loss": 0.0015,
|
| 1113 |
+
"num_tokens": 27969987.0,
|
| 1114 |
+
"reward": 0.9649922177195549,
|
| 1115 |
+
"reward_std": 0.0031741062664423225,
|
| 1116 |
+
"rewards/accuracy_reward": 0.0,
|
| 1117 |
+
"rewards/format_reward": 0.9988839253783226,
|
| 1118 |
+
"rewards/log_scaled_reward": -0.03389178216457367,
|
| 1119 |
+
"step": 74
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"clip_ratio": 0.0,
|
| 1123 |
+
"completion_length": 15.01897418498993,
|
| 1124 |
+
"epoch": 1.1343283582089552,
|
| 1125 |
+
"grad_norm": 0.0,
|
| 1126 |
+
"learning_rate": 4.869115258460634e-07,
|
| 1127 |
+
"loss": 0.0,
|
| 1128 |
+
"num_tokens": 28107828.0,
|
| 1129 |
+
"reward": 0.9661169648170471,
|
| 1130 |
+
"reward_std": 0.0,
|
| 1131 |
+
"rewards/accuracy_reward": 0.0,
|
| 1132 |
+
"rewards/format_reward": 1.0,
|
| 1133 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1134 |
+
"step": 75
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"clip_ratio": 0.0,
|
| 1138 |
+
"completion_length": 14.996652722358704,
|
| 1139 |
+
"epoch": 1.1492537313432836,
|
| 1140 |
+
"grad_norm": 1.0442914962768555,
|
| 1141 |
+
"learning_rate": 4.7383202187852804e-07,
|
| 1142 |
+
"loss": -0.0006,
|
| 1143 |
+
"num_tokens": 28248081.0,
|
| 1144 |
+
"reward": 0.9638938158750534,
|
| 1145 |
+
"reward_std": 0.006288029253482819,
|
| 1146 |
+
"rewards/accuracy_reward": 0.0,
|
| 1147 |
+
"rewards/format_reward": 0.9977678507566452,
|
| 1148 |
+
"rewards/log_scaled_reward": -0.03387411683797836,
|
| 1149 |
+
"step": 76
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"clip_ratio": 0.0,
|
| 1153 |
+
"completion_length": 15.00334918498993,
|
| 1154 |
+
"epoch": 1.164179104477612,
|
| 1155 |
+
"grad_norm": 1.5838154554367065,
|
| 1156 |
+
"learning_rate": 4.6077045213607755e-07,
|
| 1157 |
+
"loss": 0.0005,
|
| 1158 |
+
"num_tokens": 28386956.0,
|
| 1159 |
+
"reward": 0.9627697318792343,
|
| 1160 |
+
"reward_std": 0.009467384777963161,
|
| 1161 |
+
"rewards/accuracy_reward": 0.0,
|
| 1162 |
+
"rewards/format_reward": 0.9966517835855484,
|
| 1163 |
+
"rewards/log_scaled_reward": -0.0338821173645556,
|
| 1164 |
+
"step": 77
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"clip_ratio": 0.0,
|
| 1168 |
+
"completion_length": 15.000000953674316,
|
| 1169 |
+
"epoch": 1.1791044776119404,
|
| 1170 |
+
"grad_norm": 1.337672472000122,
|
| 1171 |
+
"learning_rate": 4.477357683661733e-07,
|
| 1172 |
+
"loss": -0.0,
|
| 1173 |
+
"num_tokens": 28527204.0,
|
| 1174 |
+
"reward": 0.9638830795884132,
|
| 1175 |
+
"reward_std": 0.006318369880318642,
|
| 1176 |
+
"rewards/accuracy_reward": 0.0,
|
| 1177 |
+
"rewards/format_reward": 0.9977678582072258,
|
| 1178 |
+
"rewards/log_scaled_reward": -0.033884843811392784,
|
| 1179 |
+
"step": 78
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"clip_ratio": 0.0,
|
| 1183 |
+
"completion_length": 15.000000953674316,
|
| 1184 |
+
"epoch": 1.1940298507462686,
|
| 1185 |
+
"grad_norm": 0.0,
|
| 1186 |
+
"learning_rate": 4.347369038899743e-07,
|
| 1187 |
+
"loss": 0.0,
|
| 1188 |
+
"num_tokens": 28671788.0,
|
| 1189 |
+
"reward": 0.9661169648170471,
|
| 1190 |
+
"reward_std": 0.0,
|
| 1191 |
+
"rewards/accuracy_reward": 0.0,
|
| 1192 |
+
"rewards/format_reward": 1.0,
|
| 1193 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1194 |
+
"step": 79
|
| 1195 |
+
},
|
| 1196 |
+
{
|
| 1197 |
+
"clip_ratio": 0.0,
|
| 1198 |
+
"completion_length": 15.000000953674316,
|
| 1199 |
+
"epoch": 1.208955223880597,
|
| 1200 |
+
"grad_norm": 0.0,
|
| 1201 |
+
"learning_rate": 4.2178276747988444e-07,
|
| 1202 |
+
"loss": 0.0,
|
| 1203 |
+
"num_tokens": 28811644.0,
|
| 1204 |
+
"reward": 0.9661169648170471,
|
| 1205 |
+
"reward_std": 0.0,
|
| 1206 |
+
"rewards/accuracy_reward": 0.0,
|
| 1207 |
+
"rewards/format_reward": 1.0,
|
| 1208 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1209 |
+
"step": 80
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"clip_ratio": 0.0,
|
| 1213 |
+
"completion_length": 15.000000953674316,
|
| 1214 |
+
"epoch": 1.2238805970149254,
|
| 1215 |
+
"grad_norm": 0.0,
|
| 1216 |
+
"learning_rate": 4.0888223725392624e-07,
|
| 1217 |
+
"loss": 0.0,
|
| 1218 |
+
"num_tokens": 28949996.0,
|
| 1219 |
+
"reward": 0.9661169648170471,
|
| 1220 |
+
"reward_std": 0.0,
|
| 1221 |
+
"rewards/accuracy_reward": 0.0,
|
| 1222 |
+
"rewards/format_reward": 1.0,
|
| 1223 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1224 |
+
"step": 81
|
| 1225 |
+
},
|
| 1226 |
+
{
|
| 1227 |
+
"clip_ratio": 0.0,
|
| 1228 |
+
"completion_length": 15.000000953674316,
|
| 1229 |
+
"epoch": 1.2388059701492538,
|
| 1230 |
+
"grad_norm": 0.0,
|
| 1231 |
+
"learning_rate": 3.960441545911204e-07,
|
| 1232 |
+
"loss": 0.0,
|
| 1233 |
+
"num_tokens": 29100540.0,
|
| 1234 |
+
"reward": 0.9661169648170471,
|
| 1235 |
+
"reward_std": 0.0,
|
| 1236 |
+
"rewards/accuracy_reward": 0.0,
|
| 1237 |
+
"rewards/format_reward": 1.0,
|
| 1238 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1239 |
+
"step": 82
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"clip_ratio": 0.0,
|
| 1243 |
+
"completion_length": 15.000000953674316,
|
| 1244 |
+
"epoch": 1.2537313432835822,
|
| 1245 |
+
"grad_norm": 0.9363570213317871,
|
| 1246 |
+
"learning_rate": 3.8327731807204744e-07,
|
| 1247 |
+
"loss": -0.0,
|
| 1248 |
+
"num_tokens": 29242748.0,
|
| 1249 |
+
"reward": 0.9650017619132996,
|
| 1250 |
+
"reward_std": 0.0031542566139250994,
|
| 1251 |
+
"rewards/accuracy_reward": 0.0,
|
| 1252 |
+
"rewards/format_reward": 0.9988839253783226,
|
| 1253 |
+
"rewards/log_scaled_reward": -0.03388223238289356,
|
| 1254 |
+
"step": 83
|
| 1255 |
+
},
|
| 1256 |
+
{
|
| 1257 |
+
"clip_ratio": 0.0,
|
| 1258 |
+
"completion_length": 15.000000953674316,
|
| 1259 |
+
"epoch": 1.2686567164179103,
|
| 1260 |
+
"grad_norm": 0.0,
|
| 1261 |
+
"learning_rate": 3.7059047744873955e-07,
|
| 1262 |
+
"loss": 0.0,
|
| 1263 |
+
"num_tokens": 29394812.0,
|
| 1264 |
+
"reward": 0.9661169648170471,
|
| 1265 |
+
"reward_std": 0.0,
|
| 1266 |
+
"rewards/accuracy_reward": 0.0,
|
| 1267 |
+
"rewards/format_reward": 1.0,
|
| 1268 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1269 |
+
"step": 84
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"clip_ratio": 0.0,
|
| 1273 |
+
"completion_length": 15.000000953674316,
|
| 1274 |
+
"epoch": 1.2835820895522387,
|
| 1275 |
+
"grad_norm": 0.0,
|
| 1276 |
+
"learning_rate": 3.5799232764803867e-07,
|
| 1277 |
+
"loss": 0.0,
|
| 1278 |
+
"num_tokens": 29526756.0,
|
| 1279 |
+
"reward": 0.9661169648170471,
|
| 1280 |
+
"reward_std": 0.0,
|
| 1281 |
+
"rewards/accuracy_reward": 0.0,
|
| 1282 |
+
"rewards/format_reward": 1.0,
|
| 1283 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1284 |
+
"step": 85
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"clip_ratio": 0.0,
|
| 1288 |
+
"completion_length": 15.000000953674316,
|
| 1289 |
+
"epoch": 1.2985074626865671,
|
| 1290 |
+
"grad_norm": 0.0,
|
| 1291 |
+
"learning_rate": 3.454915028125263e-07,
|
| 1292 |
+
"loss": 0.0,
|
| 1293 |
+
"num_tokens": 29662892.0,
|
| 1294 |
+
"reward": 0.9661169648170471,
|
| 1295 |
+
"reward_std": 0.0,
|
| 1296 |
+
"rewards/accuracy_reward": 0.0,
|
| 1297 |
+
"rewards/format_reward": 1.0,
|
| 1298 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1299 |
+
"step": 86
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"clip_ratio": 0.0,
|
| 1303 |
+
"completion_length": 15.000000953674316,
|
| 1304 |
+
"epoch": 1.3134328358208955,
|
| 1305 |
+
"grad_norm": 0.0,
|
| 1306 |
+
"learning_rate": 3.330965703831146e-07,
|
| 1307 |
+
"loss": 0.0,
|
| 1308 |
+
"num_tokens": 29808492.0,
|
| 1309 |
+
"reward": 0.975348062813282,
|
| 1310 |
+
"reward_std": 0.0,
|
| 1311 |
+
"rewards/accuracy_reward": 0.0,
|
| 1312 |
+
"rewards/format_reward": 1.0,
|
| 1313 |
+
"rewards/log_scaled_reward": -0.024652006570249796,
|
| 1314 |
+
"step": 87
|
| 1315 |
+
},
|
| 1316 |
+
{
|
| 1317 |
+
"clip_ratio": 0.0,
|
| 1318 |
+
"completion_length": 15.000000953674316,
|
| 1319 |
+
"epoch": 1.328358208955224,
|
| 1320 |
+
"grad_norm": 0.0,
|
| 1321 |
+
"learning_rate": 3.2081602522734985e-07,
|
| 1322 |
+
"loss": 0.0,
|
| 1323 |
+
"num_tokens": 29960316.0,
|
| 1324 |
+
"reward": 0.9661169648170471,
|
| 1325 |
+
"reward_std": 0.0,
|
| 1326 |
+
"rewards/accuracy_reward": 0.0,
|
| 1327 |
+
"rewards/format_reward": 1.0,
|
| 1328 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1329 |
+
"step": 88
|
| 1330 |
+
},
|
| 1331 |
+
{
|
| 1332 |
+
"clip_ratio": 0.0,
|
| 1333 |
+
"completion_length": 15.000000953674316,
|
| 1334 |
+
"epoch": 1.3432835820895521,
|
| 1335 |
+
"grad_norm": 0.0,
|
| 1336 |
+
"learning_rate": 3.086582838174551e-07,
|
| 1337 |
+
"loss": 0.0,
|
| 1338 |
+
"num_tokens": 30099644.0,
|
| 1339 |
+
"reward": 0.9661169648170471,
|
| 1340 |
+
"reward_std": 0.0,
|
| 1341 |
+
"rewards/accuracy_reward": 0.0,
|
| 1342 |
+
"rewards/format_reward": 1.0,
|
| 1343 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1344 |
+
"step": 89
|
| 1345 |
+
},
|
| 1346 |
+
{
|
| 1347 |
+
"clip_ratio": 0.0,
|
| 1348 |
+
"completion_length": 15.005581259727478,
|
| 1349 |
+
"epoch": 1.3582089552238805,
|
| 1350 |
+
"grad_norm": 0.0,
|
| 1351 |
+
"learning_rate": 2.9663167846209996e-07,
|
| 1352 |
+
"loss": 0.0,
|
| 1353 |
+
"num_tokens": 30246217.0,
|
| 1354 |
+
"reward": 0.9661169648170471,
|
| 1355 |
+
"reward_std": 0.0,
|
| 1356 |
+
"rewards/accuracy_reward": 0.0,
|
| 1357 |
+
"rewards/format_reward": 1.0,
|
| 1358 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1359 |
+
"step": 90
|
| 1360 |
+
},
|
| 1361 |
+
{
|
| 1362 |
+
"clip_ratio": 0.0,
|
| 1363 |
+
"completion_length": 15.000000953674316,
|
| 1364 |
+
"epoch": 1.373134328358209,
|
| 1365 |
+
"grad_norm": 0.0,
|
| 1366 |
+
"learning_rate": 2.847444515958523e-07,
|
| 1367 |
+
"loss": 0.0,
|
| 1368 |
+
"num_tokens": 30391785.0,
|
| 1369 |
+
"reward": 0.9661169648170471,
|
| 1370 |
+
"reward_std": 0.0,
|
| 1371 |
+
"rewards/accuracy_reward": 0.0,
|
| 1372 |
+
"rewards/format_reward": 1.0,
|
| 1373 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1374 |
+
"step": 91
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"clip_ratio": 0.0,
|
| 1378 |
+
"completion_length": 15.000000953674316,
|
| 1379 |
+
"epoch": 1.3880597014925373,
|
| 1380 |
+
"grad_norm": 0.0,
|
| 1381 |
+
"learning_rate": 2.730047501302266e-07,
|
| 1382 |
+
"loss": 0.0,
|
| 1383 |
+
"num_tokens": 30531289.0,
|
| 1384 |
+
"reward": 0.9661169648170471,
|
| 1385 |
+
"reward_std": 0.0,
|
| 1386 |
+
"rewards/accuracy_reward": 0.0,
|
| 1387 |
+
"rewards/format_reward": 1.0,
|
| 1388 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1389 |
+
"step": 92
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"clip_ratio": 0.0,
|
| 1393 |
+
"completion_length": 15.000000953674316,
|
| 1394 |
+
"epoch": 1.4029850746268657,
|
| 1395 |
+
"grad_norm": 3.6358273029327393,
|
| 1396 |
+
"learning_rate": 2.6142061987019574e-07,
|
| 1397 |
+
"loss": -0.0,
|
| 1398 |
+
"num_tokens": 30664633.0,
|
| 1399 |
+
"reward": 0.966103158891201,
|
| 1400 |
+
"reward_std": 1.4753467439732049e-05,
|
| 1401 |
+
"rewards/accuracy_reward": 0.0,
|
| 1402 |
+
"rewards/format_reward": 1.0,
|
| 1403 |
+
"rewards/log_scaled_reward": -0.033896906301379204,
|
| 1404 |
+
"step": 93
|
| 1405 |
+
},
|
| 1406 |
+
{
|
| 1407 |
+
"clip_ratio": 0.0,
|
| 1408 |
+
"completion_length": 15.000000953674316,
|
| 1409 |
+
"epoch": 1.417910447761194,
|
| 1410 |
+
"grad_norm": 0.0,
|
| 1411 |
+
"learning_rate": 2.500000000000001e-07,
|
| 1412 |
+
"loss": 0.0,
|
| 1413 |
+
"num_tokens": 30820657.0,
|
| 1414 |
+
"reward": 0.9661169648170471,
|
| 1415 |
+
"reward_std": 0.0,
|
| 1416 |
+
"rewards/accuracy_reward": 0.0,
|
| 1417 |
+
"rewards/format_reward": 1.0,
|
| 1418 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1419 |
+
"step": 94
|
| 1420 |
+
},
|
| 1421 |
+
{
|
| 1422 |
+
"clip_ratio": 0.0,
|
| 1423 |
+
"completion_length": 15.000000953674316,
|
| 1424 |
+
"epoch": 1.4328358208955223,
|
| 1425 |
+
"grad_norm": 0.0,
|
| 1426 |
+
"learning_rate": 2.387507176420256e-07,
|
| 1427 |
+
"loss": 0.0,
|
| 1428 |
+
"num_tokens": 30965737.0,
|
| 1429 |
+
"reward": 0.9661169648170471,
|
| 1430 |
+
"reward_std": 0.0,
|
| 1431 |
+
"rewards/accuracy_reward": 0.0,
|
| 1432 |
+
"rewards/format_reward": 1.0,
|
| 1433 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1434 |
+
"step": 95
|
| 1435 |
+
},
|
| 1436 |
+
{
|
| 1437 |
+
"clip_ratio": 0.0,
|
| 1438 |
+
"completion_length": 15.00111699104309,
|
| 1439 |
+
"epoch": 1.4477611940298507,
|
| 1440 |
+
"grad_norm": 32.66771697998047,
|
| 1441 |
+
"learning_rate": 2.2768048249248644e-07,
|
| 1442 |
+
"loss": -0.0002,
|
| 1443 |
+
"num_tokens": 31108402.0,
|
| 1444 |
+
"reward": 0.9649775922298431,
|
| 1445 |
+
"reward_std": 0.0031557143665850163,
|
| 1446 |
+
"rewards/accuracy_reward": 0.0,
|
| 1447 |
+
"rewards/format_reward": 0.9988839253783226,
|
| 1448 |
+
"rewards/log_scaled_reward": -0.03390640066936612,
|
| 1449 |
+
"step": 96
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"clip_ratio": 0.0,
|
| 1453 |
+
"completion_length": 15.000000953674316,
|
| 1454 |
+
"epoch": 1.462686567164179,
|
| 1455 |
+
"grad_norm": 0.0,
|
| 1456 |
+
"learning_rate": 2.167968815375837e-07,
|
| 1457 |
+
"loss": 0.0,
|
| 1458 |
+
"num_tokens": 31257634.0,
|
| 1459 |
+
"reward": 0.9661169648170471,
|
| 1460 |
+
"reward_std": 0.0,
|
| 1461 |
+
"rewards/accuracy_reward": 0.0,
|
| 1462 |
+
"rewards/format_reward": 1.0,
|
| 1463 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1464 |
+
"step": 97
|
| 1465 |
+
},
|
| 1466 |
+
{
|
| 1467 |
+
"clip_ratio": 0.0,
|
| 1468 |
+
"completion_length": 15.000000953674316,
|
| 1469 |
+
"epoch": 1.4776119402985075,
|
| 1470 |
+
"grad_norm": 0.0,
|
| 1471 |
+
"learning_rate": 2.0610737385376348e-07,
|
| 1472 |
+
"loss": 0.0,
|
| 1473 |
+
"num_tokens": 31412970.0,
|
| 1474 |
+
"reward": 0.9661169648170471,
|
| 1475 |
+
"reward_std": 0.0,
|
| 1476 |
+
"rewards/accuracy_reward": 0.0,
|
| 1477 |
+
"rewards/format_reward": 1.0,
|
| 1478 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1479 |
+
"step": 98
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"clip_ratio": 0.0,
|
| 1483 |
+
"completion_length": 15.000000953674316,
|
| 1484 |
+
"epoch": 1.4925373134328357,
|
| 1485 |
+
"grad_norm": 0.0,
|
| 1486 |
+
"learning_rate": 1.9561928549563966e-07,
|
| 1487 |
+
"loss": 0.0,
|
| 1488 |
+
"num_tokens": 31564498.0,
|
| 1489 |
+
"reward": 0.9661169648170471,
|
| 1490 |
+
"reward_std": 0.0,
|
| 1491 |
+
"rewards/accuracy_reward": 0.0,
|
| 1492 |
+
"rewards/format_reward": 1.0,
|
| 1493 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1494 |
+
"step": 99
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 1.5074626865671643,
|
| 1498 |
+
"grad_norm": 0.0,
|
| 1499 |
+
"learning_rate": 1.8533980447508135e-07,
|
| 1500 |
+
"loss": 0.0,
|
| 1501 |
+
"step": 100
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 1.5074626865671643,
|
| 1505 |
+
"eval_clip_ratio": 0.0,
|
| 1506 |
+
"eval_completion_length": 15.000150587305676,
|
| 1507 |
+
"eval_loss": 0.0,
|
| 1508 |
+
"eval_num_tokens": 31697802.0,
|
| 1509 |
+
"eval_reward": 0.9661169648170471,
|
| 1510 |
+
"eval_reward_std": 0.0,
|
| 1511 |
+
"eval_rewards/accuracy_reward": 0.0,
|
| 1512 |
+
"eval_rewards/format_reward": 1.0,
|
| 1513 |
+
"eval_rewards/log_scaled_reward": -0.03388310596346855,
|
| 1514 |
+
"eval_runtime": 724.0402,
|
| 1515 |
+
"eval_samples_per_second": 6.906,
|
| 1516 |
+
"eval_steps_per_second": 0.062,
|
| 1517 |
+
"step": 100
|
| 1518 |
+
},
|
| 1519 |
+
{
|
| 1520 |
+
"clip_ratio": 0.0,
|
| 1521 |
+
"completion_length": 15.000000953674316,
|
| 1522 |
+
"epoch": 1.5223880597014925,
|
| 1523 |
+
"grad_norm": 0.0,
|
| 1524 |
+
"learning_rate": 1.7527597583490823e-07,
|
| 1525 |
+
"loss": 0.0,
|
| 1526 |
+
"num_tokens": 31837018.0,
|
| 1527 |
+
"reward": 0.9707325138151646,
|
| 1528 |
+
"reward_std": 0.0,
|
| 1529 |
+
"rewards/accuracy_reward": 0.0,
|
| 1530 |
+
"rewards/format_reward": 1.0,
|
| 1531 |
+
"rewards/log_scaled_reward": -0.029267556266859174,
|
| 1532 |
+
"step": 101
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"clip_ratio": 0.0,
|
| 1536 |
+
"completion_length": 15.000000953674316,
|
| 1537 |
+
"epoch": 1.537313432835821,
|
| 1538 |
+
"grad_norm": 0.0,
|
| 1539 |
+
"learning_rate": 1.6543469682057104e-07,
|
| 1540 |
+
"loss": 0.0,
|
| 1541 |
+
"num_tokens": 31986642.0,
|
| 1542 |
+
"reward": 0.9661169648170471,
|
| 1543 |
+
"reward_std": 0.0,
|
| 1544 |
+
"rewards/accuracy_reward": 0.0,
|
| 1545 |
+
"rewards/format_reward": 1.0,
|
| 1546 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1547 |
+
"step": 102
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"clip_ratio": 0.0,
|
| 1551 |
+
"completion_length": 15.000000953674316,
|
| 1552 |
+
"epoch": 1.5522388059701493,
|
| 1553 |
+
"grad_norm": 0.0,
|
| 1554 |
+
"learning_rate": 1.5582271215312293e-07,
|
| 1555 |
+
"loss": 0.0,
|
| 1556 |
+
"num_tokens": 32126642.0,
|
| 1557 |
+
"reward": 0.9661169648170471,
|
| 1558 |
+
"reward_std": 0.0,
|
| 1559 |
+
"rewards/accuracy_reward": 0.0,
|
| 1560 |
+
"rewards/format_reward": 1.0,
|
| 1561 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1562 |
+
"step": 103
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"clip_ratio": 0.0,
|
| 1566 |
+
"completion_length": 15.000000953674316,
|
| 1567 |
+
"epoch": 1.5671641791044775,
|
| 1568 |
+
"grad_norm": 0.0,
|
| 1569 |
+
"learning_rate": 1.4644660940672627e-07,
|
| 1570 |
+
"loss": 0.0,
|
| 1571 |
+
"num_tokens": 32276154.0,
|
| 1572 |
+
"reward": 0.9661169648170471,
|
| 1573 |
+
"reward_std": 0.0,
|
| 1574 |
+
"rewards/accuracy_reward": 0.0,
|
| 1575 |
+
"rewards/format_reward": 1.0,
|
| 1576 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1577 |
+
"step": 104
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"clip_ratio": 0.0,
|
| 1581 |
+
"completion_length": 15.000000953674316,
|
| 1582 |
+
"epoch": 1.582089552238806,
|
| 1583 |
+
"grad_norm": 0.0,
|
| 1584 |
+
"learning_rate": 1.3731281449385628e-07,
|
| 1585 |
+
"loss": 0.0,
|
| 1586 |
+
"num_tokens": 32426898.0,
|
| 1587 |
+
"reward": 0.9661169648170471,
|
| 1588 |
+
"reward_std": 0.0,
|
| 1589 |
+
"rewards/accuracy_reward": 0.0,
|
| 1590 |
+
"rewards/format_reward": 1.0,
|
| 1591 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1592 |
+
"step": 105
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"clip_ratio": 0.0,
|
| 1596 |
+
"completion_length": 15.000000953674316,
|
| 1597 |
+
"epoch": 1.5970149253731343,
|
| 1598 |
+
"grad_norm": 0.0,
|
| 1599 |
+
"learning_rate": 1.284275872613028e-07,
|
| 1600 |
+
"loss": 0.0,
|
| 1601 |
+
"num_tokens": 32571226.0,
|
| 1602 |
+
"reward": 0.9661169648170471,
|
| 1603 |
+
"reward_std": 0.0,
|
| 1604 |
+
"rewards/accuracy_reward": 0.0,
|
| 1605 |
+
"rewards/format_reward": 1.0,
|
| 1606 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1607 |
+
"step": 106
|
| 1608 |
+
},
|
| 1609 |
+
{
|
| 1610 |
+
"clip_ratio": 0.0,
|
| 1611 |
+
"completion_length": 15.000000953674316,
|
| 1612 |
+
"epoch": 1.6119402985074627,
|
| 1613 |
+
"grad_norm": 0.0,
|
| 1614 |
+
"learning_rate": 1.1979701719998454e-07,
|
| 1615 |
+
"loss": 0.0,
|
| 1616 |
+
"num_tokens": 32719426.0,
|
| 1617 |
+
"reward": 0.9661169648170471,
|
| 1618 |
+
"reward_std": 0.0,
|
| 1619 |
+
"rewards/accuracy_reward": 0.0,
|
| 1620 |
+
"rewards/format_reward": 1.0,
|
| 1621 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1622 |
+
"step": 107
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"clip_ratio": 0.0,
|
| 1626 |
+
"completion_length": 15.000000953674316,
|
| 1627 |
+
"epoch": 1.626865671641791,
|
| 1628 |
+
"grad_norm": 0.0,
|
| 1629 |
+
"learning_rate": 1.1142701927151454e-07,
|
| 1630 |
+
"loss": 0.0,
|
| 1631 |
+
"num_tokens": 32852794.0,
|
| 1632 |
+
"reward": 0.9661169648170471,
|
| 1633 |
+
"reward_std": 0.0,
|
| 1634 |
+
"rewards/accuracy_reward": 0.0,
|
| 1635 |
+
"rewards/format_reward": 1.0,
|
| 1636 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1637 |
+
"step": 108
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"clip_ratio": 0.0,
|
| 1641 |
+
"completion_length": 15.000000953674316,
|
| 1642 |
+
"epoch": 1.6417910447761193,
|
| 1643 |
+
"grad_norm": 0.0,
|
| 1644 |
+
"learning_rate": 1.0332332985438247e-07,
|
| 1645 |
+
"loss": 0.0,
|
| 1646 |
+
"num_tokens": 33000498.0,
|
| 1647 |
+
"reward": 0.9661169648170471,
|
| 1648 |
+
"reward_std": 0.0,
|
| 1649 |
+
"rewards/accuracy_reward": 0.0,
|
| 1650 |
+
"rewards/format_reward": 1.0,
|
| 1651 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1652 |
+
"step": 109
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"clip_ratio": 0.0,
|
| 1656 |
+
"completion_length": 15.000000953674316,
|
| 1657 |
+
"epoch": 1.6567164179104479,
|
| 1658 |
+
"grad_norm": 0.0,
|
| 1659 |
+
"learning_rate": 9.549150281252632e-08,
|
| 1660 |
+
"loss": 0.0,
|
| 1661 |
+
"num_tokens": 33139850.0,
|
| 1662 |
+
"reward": 0.9661169648170471,
|
| 1663 |
+
"reward_std": 0.0,
|
| 1664 |
+
"rewards/accuracy_reward": 0.0,
|
| 1665 |
+
"rewards/format_reward": 1.0,
|
| 1666 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1667 |
+
"step": 110
|
| 1668 |
+
},
|
| 1669 |
+
{
|
| 1670 |
+
"clip_ratio": 0.0,
|
| 1671 |
+
"completion_length": 15.000000953674316,
|
| 1672 |
+
"epoch": 1.671641791044776,
|
| 1673 |
+
"grad_norm": 0.0,
|
| 1674 |
+
"learning_rate": 8.793690568899215e-08,
|
| 1675 |
+
"loss": 0.0,
|
| 1676 |
+
"num_tokens": 33276866.0,
|
| 1677 |
+
"reward": 0.9661169648170471,
|
| 1678 |
+
"reward_std": 0.0,
|
| 1679 |
+
"rewards/accuracy_reward": 0.0,
|
| 1680 |
+
"rewards/format_reward": 1.0,
|
| 1681 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1682 |
+
"step": 111
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"clip_ratio": 0.0,
|
| 1686 |
+
"completion_length": 15.000000953674316,
|
| 1687 |
+
"epoch": 1.6865671641791045,
|
| 1688 |
+
"grad_norm": 0.0,
|
| 1689 |
+
"learning_rate": 8.066471602728803e-08,
|
| 1690 |
+
"loss": 0.0,
|
| 1691 |
+
"num_tokens": 33407970.0,
|
| 1692 |
+
"reward": 0.9661169648170471,
|
| 1693 |
+
"reward_std": 0.0,
|
| 1694 |
+
"rewards/accuracy_reward": 0.0,
|
| 1695 |
+
"rewards/format_reward": 1.0,
|
| 1696 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1697 |
+
"step": 112
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"clip_ratio": 0.0,
|
| 1701 |
+
"completion_length": 15.000000953674316,
|
| 1702 |
+
"epoch": 1.7014925373134329,
|
| 1703 |
+
"grad_norm": 0.0,
|
| 1704 |
+
"learning_rate": 7.36799178229539e-08,
|
| 1705 |
+
"loss": 0.0,
|
| 1706 |
+
"num_tokens": 33549218.0,
|
| 1707 |
+
"reward": 0.9661169648170471,
|
| 1708 |
+
"reward_std": 0.0,
|
| 1709 |
+
"rewards/accuracy_reward": 0.0,
|
| 1710 |
+
"rewards/format_reward": 1.0,
|
| 1711 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1712 |
+
"step": 113
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"clip_ratio": 0.0,
|
| 1716 |
+
"completion_length": 15.000000953674316,
|
| 1717 |
+
"epoch": 1.716417910447761,
|
| 1718 |
+
"grad_norm": 0.0,
|
| 1719 |
+
"learning_rate": 6.698729810778064e-08,
|
| 1720 |
+
"loss": 0.0,
|
| 1721 |
+
"num_tokens": 33694498.0,
|
| 1722 |
+
"reward": 0.9661169648170471,
|
| 1723 |
+
"reward_std": 0.0,
|
| 1724 |
+
"rewards/accuracy_reward": 0.0,
|
| 1725 |
+
"rewards/format_reward": 1.0,
|
| 1726 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1727 |
+
"step": 114
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"clip_ratio": 0.0,
|
| 1731 |
+
"completion_length": 15.000000953674316,
|
| 1732 |
+
"epoch": 1.7313432835820897,
|
| 1733 |
+
"grad_norm": 0.0,
|
| 1734 |
+
"learning_rate": 6.059144366901736e-08,
|
| 1735 |
+
"loss": 0.0,
|
| 1736 |
+
"num_tokens": 33836482.0,
|
| 1737 |
+
"reward": 0.9661169648170471,
|
| 1738 |
+
"reward_std": 0.0,
|
| 1739 |
+
"rewards/accuracy_reward": 0.0,
|
| 1740 |
+
"rewards/format_reward": 1.0,
|
| 1741 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1742 |
+
"step": 115
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"clip_ratio": 0.0,
|
| 1746 |
+
"completion_length": 15.000000953674316,
|
| 1747 |
+
"epoch": 1.7462686567164178,
|
| 1748 |
+
"grad_norm": 0.0,
|
| 1749 |
+
"learning_rate": 5.44967379058161e-08,
|
| 1750 |
+
"loss": 0.0,
|
| 1751 |
+
"num_tokens": 33970506.0,
|
| 1752 |
+
"reward": 0.9661169648170471,
|
| 1753 |
+
"reward_std": 0.0,
|
| 1754 |
+
"rewards/accuracy_reward": 0.0,
|
| 1755 |
+
"rewards/format_reward": 1.0,
|
| 1756 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1757 |
+
"step": 116
|
| 1758 |
+
},
|
| 1759 |
+
{
|
| 1760 |
+
"clip_ratio": 0.0,
|
| 1761 |
+
"completion_length": 15.002233028411865,
|
| 1762 |
+
"epoch": 1.7611940298507462,
|
| 1763 |
+
"grad_norm": 0.0,
|
| 1764 |
+
"learning_rate": 4.870735782506979e-08,
|
| 1765 |
+
"loss": 0.0,
|
| 1766 |
+
"num_tokens": 34139396.0,
|
| 1767 |
+
"reward": 0.9661169648170471,
|
| 1768 |
+
"reward_std": 0.0,
|
| 1769 |
+
"rewards/accuracy_reward": 0.0,
|
| 1770 |
+
"rewards/format_reward": 1.0,
|
| 1771 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1772 |
+
"step": 117
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"clip_ratio": 0.0,
|
| 1776 |
+
"completion_length": 15.000000953674316,
|
| 1777 |
+
"epoch": 1.7761194029850746,
|
| 1778 |
+
"grad_norm": 0.0,
|
| 1779 |
+
"learning_rate": 4.322727117869951e-08,
|
| 1780 |
+
"loss": 0.0,
|
| 1781 |
+
"num_tokens": 34279244.0,
|
| 1782 |
+
"reward": 0.9661169648170471,
|
| 1783 |
+
"reward_std": 0.0,
|
| 1784 |
+
"rewards/accuracy_reward": 0.0,
|
| 1785 |
+
"rewards/format_reward": 1.0,
|
| 1786 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1787 |
+
"step": 118
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"clip_ratio": 0.0,
|
| 1791 |
+
"completion_length": 15.000000953674316,
|
| 1792 |
+
"epoch": 1.7910447761194028,
|
| 1793 |
+
"grad_norm": 0.0,
|
| 1794 |
+
"learning_rate": 3.806023374435663e-08,
|
| 1795 |
+
"loss": 0.0,
|
| 1796 |
+
"num_tokens": 34418092.0,
|
| 1797 |
+
"reward": 0.9661169648170471,
|
| 1798 |
+
"reward_std": 0.0,
|
| 1799 |
+
"rewards/accuracy_reward": 0.0,
|
| 1800 |
+
"rewards/format_reward": 1.0,
|
| 1801 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1802 |
+
"step": 119
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"clip_ratio": 0.0,
|
| 1806 |
+
"completion_length": 15.000000953674316,
|
| 1807 |
+
"epoch": 1.8059701492537314,
|
| 1808 |
+
"grad_norm": 0.0,
|
| 1809 |
+
"learning_rate": 3.3209786751399184e-08,
|
| 1810 |
+
"loss": 0.0,
|
| 1811 |
+
"num_tokens": 34570476.0,
|
| 1812 |
+
"reward": 0.9661169648170471,
|
| 1813 |
+
"reward_std": 0.0,
|
| 1814 |
+
"rewards/accuracy_reward": 0.0,
|
| 1815 |
+
"rewards/format_reward": 1.0,
|
| 1816 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1817 |
+
"step": 120
|
| 1818 |
+
},
|
| 1819 |
+
{
|
| 1820 |
+
"clip_ratio": 0.0,
|
| 1821 |
+
"completion_length": 15.000000953674316,
|
| 1822 |
+
"epoch": 1.8208955223880596,
|
| 1823 |
+
"grad_norm": 0.0,
|
| 1824 |
+
"learning_rate": 2.8679254453910785e-08,
|
| 1825 |
+
"loss": 0.0,
|
| 1826 |
+
"num_tokens": 34714220.0,
|
| 1827 |
+
"reward": 0.9661169648170471,
|
| 1828 |
+
"reward_std": 0.0,
|
| 1829 |
+
"rewards/accuracy_reward": 0.0,
|
| 1830 |
+
"rewards/format_reward": 1.0,
|
| 1831 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1832 |
+
"step": 121
|
| 1833 |
+
},
|
| 1834 |
+
{
|
| 1835 |
+
"clip_ratio": 0.0,
|
| 1836 |
+
"completion_length": 15.000000953674316,
|
| 1837 |
+
"epoch": 1.835820895522388,
|
| 1838 |
+
"grad_norm": 0.0,
|
| 1839 |
+
"learning_rate": 2.4471741852423233e-08,
|
| 1840 |
+
"loss": 0.0,
|
| 1841 |
+
"num_tokens": 34874452.0,
|
| 1842 |
+
"reward": 0.9661169648170471,
|
| 1843 |
+
"reward_std": 0.0,
|
| 1844 |
+
"rewards/accuracy_reward": 0.0,
|
| 1845 |
+
"rewards/format_reward": 1.0,
|
| 1846 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1847 |
+
"step": 122
|
| 1848 |
+
},
|
| 1849 |
+
{
|
| 1850 |
+
"clip_ratio": 0.0,
|
| 1851 |
+
"completion_length": 15.000000953674316,
|
| 1852 |
+
"epoch": 1.8507462686567164,
|
| 1853 |
+
"grad_norm": 0.0,
|
| 1854 |
+
"learning_rate": 2.0590132565903473e-08,
|
| 1855 |
+
"loss": 0.0,
|
| 1856 |
+
"num_tokens": 35020932.0,
|
| 1857 |
+
"reward": 0.9661169648170471,
|
| 1858 |
+
"reward_std": 0.0,
|
| 1859 |
+
"rewards/accuracy_reward": 0.0,
|
| 1860 |
+
"rewards/format_reward": 1.0,
|
| 1861 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1862 |
+
"step": 123
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"clip_ratio": 0.0,
|
| 1866 |
+
"completion_length": 15.000000953674316,
|
| 1867 |
+
"epoch": 1.8656716417910446,
|
| 1868 |
+
"grad_norm": 0.0,
|
| 1869 |
+
"learning_rate": 1.7037086855465898e-08,
|
| 1870 |
+
"loss": 0.0,
|
| 1871 |
+
"num_tokens": 35158836.0,
|
| 1872 |
+
"reward": 0.9661169648170471,
|
| 1873 |
+
"reward_std": 0.0,
|
| 1874 |
+
"rewards/accuracy_reward": 0.0,
|
| 1875 |
+
"rewards/format_reward": 1.0,
|
| 1876 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1877 |
+
"step": 124
|
| 1878 |
+
},
|
| 1879 |
+
{
|
| 1880 |
+
"clip_ratio": 0.0,
|
| 1881 |
+
"completion_length": 15.000000953674316,
|
| 1882 |
+
"epoch": 1.8805970149253732,
|
| 1883 |
+
"grad_norm": 0.0,
|
| 1884 |
+
"learning_rate": 1.3815039801161722e-08,
|
| 1885 |
+
"loss": 0.0,
|
| 1886 |
+
"num_tokens": 35299044.0,
|
| 1887 |
+
"reward": 0.9661169648170471,
|
| 1888 |
+
"reward_std": 0.0,
|
| 1889 |
+
"rewards/accuracy_reward": 0.0,
|
| 1890 |
+
"rewards/format_reward": 1.0,
|
| 1891 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1892 |
+
"step": 125
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"clip_ratio": 0.0,
|
| 1896 |
+
"completion_length": 15.000000953674316,
|
| 1897 |
+
"epoch": 1.8955223880597014,
|
| 1898 |
+
"grad_norm": 0.0,
|
| 1899 |
+
"learning_rate": 1.0926199633097154e-08,
|
| 1900 |
+
"loss": 0.0,
|
| 1901 |
+
"num_tokens": 35438660.0,
|
| 1902 |
+
"reward": 0.9661169648170471,
|
| 1903 |
+
"reward_std": 0.0,
|
| 1904 |
+
"rewards/accuracy_reward": 0.0,
|
| 1905 |
+
"rewards/format_reward": 1.0,
|
| 1906 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1907 |
+
"step": 126
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"clip_ratio": 0.0,
|
| 1911 |
+
"completion_length": 15.000000953674316,
|
| 1912 |
+
"epoch": 1.9104477611940298,
|
| 1913 |
+
"grad_norm": 0.0,
|
| 1914 |
+
"learning_rate": 8.372546218022746e-09,
|
| 1915 |
+
"loss": 0.0,
|
| 1916 |
+
"num_tokens": 35589036.0,
|
| 1917 |
+
"reward": 0.9661169648170471,
|
| 1918 |
+
"reward_std": 0.0,
|
| 1919 |
+
"rewards/accuracy_reward": 0.0,
|
| 1920 |
+
"rewards/format_reward": 1.0,
|
| 1921 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1922 |
+
"step": 127
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"clip_ratio": 0.0,
|
| 1926 |
+
"completion_length": 15.000000953674316,
|
| 1927 |
+
"epoch": 1.9253731343283582,
|
| 1928 |
+
"grad_norm": 0.0,
|
| 1929 |
+
"learning_rate": 6.15582970243117e-09,
|
| 1930 |
+
"loss": 0.0,
|
| 1931 |
+
"num_tokens": 35732844.0,
|
| 1932 |
+
"reward": 0.9661169648170471,
|
| 1933 |
+
"reward_std": 0.0,
|
| 1934 |
+
"rewards/accuracy_reward": 0.0,
|
| 1935 |
+
"rewards/format_reward": 1.0,
|
| 1936 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1937 |
+
"step": 128
|
| 1938 |
+
},
|
| 1939 |
+
{
|
| 1940 |
+
"clip_ratio": 0.0,
|
| 1941 |
+
"completion_length": 15.000000953674316,
|
| 1942 |
+
"epoch": 1.9402985074626866,
|
| 1943 |
+
"grad_norm": 0.0,
|
| 1944 |
+
"learning_rate": 4.277569313094809e-09,
|
| 1945 |
+
"loss": 0.0,
|
| 1946 |
+
"num_tokens": 35869612.0,
|
| 1947 |
+
"reward": 0.9661169648170471,
|
| 1948 |
+
"reward_std": 0.0,
|
| 1949 |
+
"rewards/accuracy_reward": 0.0,
|
| 1950 |
+
"rewards/format_reward": 1.0,
|
| 1951 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1952 |
+
"step": 129
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"clip_ratio": 0.0,
|
| 1956 |
+
"completion_length": 15.000000953674316,
|
| 1957 |
+
"epoch": 1.955223880597015,
|
| 1958 |
+
"grad_norm": 0.0,
|
| 1959 |
+
"learning_rate": 2.739052315863355e-09,
|
| 1960 |
+
"loss": 0.0,
|
| 1961 |
+
"num_tokens": 36024796.0,
|
| 1962 |
+
"reward": 0.9661169648170471,
|
| 1963 |
+
"reward_std": 0.0,
|
| 1964 |
+
"rewards/accuracy_reward": 0.0,
|
| 1965 |
+
"rewards/format_reward": 1.0,
|
| 1966 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1967 |
+
"step": 130
|
| 1968 |
+
},
|
| 1969 |
+
{
|
| 1970 |
+
"clip_ratio": 0.0,
|
| 1971 |
+
"completion_length": 15.000000953674316,
|
| 1972 |
+
"epoch": 1.9701492537313432,
|
| 1973 |
+
"grad_norm": 0.0,
|
| 1974 |
+
"learning_rate": 1.541333133436018e-09,
|
| 1975 |
+
"loss": 0.0,
|
| 1976 |
+
"num_tokens": 36165404.0,
|
| 1977 |
+
"reward": 0.9661169648170471,
|
| 1978 |
+
"reward_std": 0.0,
|
| 1979 |
+
"rewards/accuracy_reward": 0.0,
|
| 1980 |
+
"rewards/format_reward": 1.0,
|
| 1981 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1982 |
+
"step": 131
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"clip_ratio": 0.0,
|
| 1986 |
+
"completion_length": 15.0,
|
| 1987 |
+
"epoch": 1.9850746268656716,
|
| 1988 |
+
"grad_norm": 0.0,
|
| 1989 |
+
"learning_rate": 6.852326227130833e-10,
|
| 1990 |
+
"loss": 0.0,
|
| 1991 |
+
"num_tokens": 36316348.0,
|
| 1992 |
+
"reward": 0.9661169648170471,
|
| 1993 |
+
"reward_std": 0.0,
|
| 1994 |
+
"rewards/accuracy_reward": 0.0,
|
| 1995 |
+
"rewards/format_reward": 1.0,
|
| 1996 |
+
"rewards/log_scaled_reward": -0.03388310596346855,
|
| 1997 |
+
"step": 132
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 1.9850746268656716,
|
| 2001 |
+
"step": 132,
|
| 2002 |
+
"total_flos": 0.0,
|
| 2003 |
+
"train_loss": 0.04579740530018937,
|
| 2004 |
+
"train_runtime": 16666.94,
|
| 2005 |
+
"train_samples_per_second": 0.9,
|
| 2006 |
+
"train_steps_per_second": 0.008
|
| 2007 |
+
}
|
| 2008 |
+
],
|
| 2009 |
+
"logging_steps": 1,
|
| 2010 |
+
"max_steps": 134,
|
| 2011 |
+
"num_input_tokens_seen": 0,
|
| 2012 |
+
"num_train_epochs": 2,
|
| 2013 |
+
"save_steps": 500,
|
| 2014 |
+
"stateful_callbacks": {
|
| 2015 |
+
"TrainerControl": {
|
| 2016 |
+
"args": {
|
| 2017 |
+
"should_epoch_stop": false,
|
| 2018 |
+
"should_evaluate": false,
|
| 2019 |
+
"should_log": false,
|
| 2020 |
+
"should_save": true,
|
| 2021 |
+
"should_training_stop": false
|
| 2022 |
+
},
|
| 2023 |
+
"attributes": {}
|
| 2024 |
+
}
|
| 2025 |
+
},
|
| 2026 |
+
"total_flos": 0.0,
|
| 2027 |
+
"train_batch_size": 16,
|
| 2028 |
+
"trial_name": null,
|
| 2029 |
+
"trial_params": null
|
| 2030 |
+
}
|