42 lines
1.2 KiB
Markdown
42 lines
1.2 KiB
Markdown
---
|
|
license: apache-2.0
|
|
base_model: Qwen/Qwen2.5-1.5B
|
|
datasets: [openai/gsm8k]
|
|
pipeline_tag: text-generation
|
|
tags: [grpo, reinforcement-learning, nemo-rl, math]
|
|
---
|
|
|
|
# Qwen2.5-1.5B GRPO GSM8K
|
|
|
|
Qwen2.5-1.5B trained with GRPO using NVIDIA NeMo RL v0.7.0.
|
|
|
|
## Results
|
|
|
|
| Model | GSM8K test (pass@1) | Correct |
|
|
|---|---|---|
|
|
| Qwen/Qwen2.5-1.5B (base) | 35.03% | 462 / 1319 |
|
|
| This model | 73.84% | 974 / 1319 |
|
|
|
|
+38.8 points, 2.1x relative. Both models evaluated identically on the full
|
|
1319-problem GSM8K test split, greedy decoding, same prompt template.
|
|
|
|
## Training
|
|
|
|
- GRPO, 130 steps, 16 prompts x 8 generations (2,080 samples)
|
|
- Reward: binary exact-match on the answer inside boxed tags
|
|
- DTensor v2 trainer + vLLM generation, colocated
|
|
- 1x H100 80GB, ~3 hours
|
|
- lr 1e-6 AdamW, KL penalty 0.01, clip 0.2/0.2, seq len 1024
|
|
|
|
No human-written solutions. The model learned from a programmatic grader.
|
|
|
|
## Prompt format
|
|
|
|
Trained with this template and expects it at inference. A bare question without
|
|
the wrapper degrades results substantially.
|
|
|
|
Think step-by-step to solve the following problem. Output your answer inside of \boxed{} tags.:
|
|
{question}
|
|
|
|
Let's think step-by-step
|