File size: 472 Bytes
8fae830 3b0e2c5 8fae830 3b0e2c5 8fae830 3b0e2c5 5d3772d 3b0e2c5 5d3772d 3b0e2c5 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 |
---
base_model:
- deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B
datasets:
- knoveleng/open-rs
language:
- en
license: apache-2.0
metrics:
- accuracy
library_name: transformers
pipeline_tag: text-generation
---
GPG: A Simple and Strong Reinforcement Learning Baseline for Model Reasoning
https://arxiv.org/abs/2504.02546
The RL model trained on the Open-r1 dataset based on GPG, using DeepSeek-R1-Distill-Qwen-1.5B as the baseline model.
Code: https://github.com/AMAP-ML/GPG |