Hugging Face BlogJan 28, 2025, 12:00 AMimportant 90
Open-R1：Hugging Face 推出完全開源的 DeepSeek-R1 重現計劃
Original: Open-R1: a fully open reproduction of DeepSeek-R1
### Project Background: Recreating the Open-Source Miracle of DeepSeek-R1 The emergence of DeepSeek-R1 sent shockwaves through the global…
Hugging Face 宣布啟動 Open-R1 專案，目標是完全重現 DeepSeek-R1 的推理能力。該專案將開源所有訓練配方、資料集與模型權重，並利用 TRL 框架中的 GRPO 演算法進行強化學習。這項計畫將降低開發者構建高效推理模型的門檻，推動開源 AI 社群的推理模型民主化。
### Project Background: Recreating the Open-Source Miracle of DeepSeek-R1
Full summary
Free shows the 3-line summary; Pro unlocks the full deep summary (~300 words) so you never have to click through.
See Pro plans →Want the original English / full article?
Read on Hugging Face Blog →Summaries are AI-generated; the original article is authoritative.