Jiayi-Pan/TinyZero: Minimal reproduction of DeepSeek R1-Zero
github.com · 351 words · saved by 1 readers
Minimal reproduction of DeepSeek R1-Zero
⚠️ Deprecation Notice: This repo is no longer actively maintained. For running RL experiments, please directly use the latest veRL library. For the archived original documentation, see OLD_README.md. TinyZero is a reproduction of DeepSeek R1 Zero in countdown and multiplication tasks. We built upon veRL. Through RL, the 3B base LM develops self-verification and search abilities all on its own. You can experience the Aha moment yourself for < $30. Twitter thread: https://x.com/jiayi_pirate/status/1882839370505621655 Full experiment log: https://wandb.ai/jiayipan/TinyZero 📢: We release…
saved by
related reading
- DeepSeek-R1arxiv.org
- PiTorch: ML on Baremetal Raspberry Pis | projectsmasonjwang.com
- GitHub - karpathy/autoresearch: AI agents running research on single-GPU nanochat training automaticallygithub.com
- GitHub - deepseek-ai/DeepSeek-R1 · GitHubgithub.com
- Understanding R1-Zero-Like Training: A Critical Perspectivearxiv.org
- GitHub - karpathy/nanochat: The best ChatGPT that $100 can buy.github.com
- Tiny Interaction Models - Rajan Agarwalrajan.sh
- DeepSeek R1's recipe to replicate o1 and the future of reasoning LMsinterconnects.ai
- Qwen3: Think Deeper, Act Faster | Qwenqwenlm.github.io
- 2502.18449arxiv.org
- Let's reproduce GPT-2 (124M)youtube.com
- Tinkerthinkingmachines.ai