Hugging Face's logo
Olivier's picture
7 1 18

Olivier

oliviermills
Itisissac's profile picture avishetty's profile picture denisfitz's profile picture
Β·
https://oliviermills.com

AI & ML interests

LLMs, Data, AI for non-profits

Recent Activity

commented on a paper 7 days ago
Beyond Release: Access Considerations for Generative AI Systems
reacted to lewtun's post with πŸ”₯ 3 months ago
We are reproducing the full DeepSeek R1 data and training pipeline so everybody can use their recipe. Instead of doing it in secret we can do it together in the open! πŸ§ͺ Step 1: replicate the R1-Distill models by distilling a high-quality reasoning corpus from DeepSeek-R1. 🧠 Step 2: replicate the pure RL pipeline that DeepSeek used to create R1-Zero. This will involve curating new, large-scale datasets for math, reasoning, and code. πŸ”₯ Step 3: show we can go from base model -> SFT -> RL via multi-stage training. Follow along: https://github.com/huggingface/open-r1
replied to lewtun's post 3 months ago
We are reproducing the full DeepSeek R1 data and training pipeline so everybody can use their recipe. Instead of doing it in secret we can do it together in the open! πŸ§ͺ Step 1: replicate the R1-Distill models by distilling a high-quality reasoning corpus from DeepSeek-R1. 🧠 Step 2: replicate the pure RL pipeline that DeepSeek used to create R1-Zero. This will involve curating new, large-scale datasets for math, reasoning, and code. πŸ”₯ Step 3: show we can go from base model -> SFT -> RL via multi-stage training. Follow along: https://github.com/huggingface/open-r1
View all activity

Organizations

Baobab Tech's profile picture

oliviermills's activity

commented a paper 7 days ago
reacted to lewtun's post with πŸ”₯ 3 months ago
view post
Post
10335
We are reproducing the full DeepSeek R1 data and training pipeline so everybody can use their recipe. Instead of doing it in secret we can do it together in the open!

πŸ§ͺ Step 1: replicate the R1-Distill models by distilling a high-quality reasoning corpus from DeepSeek-R1.

🧠 Step 2: replicate the pure RL pipeline that DeepSeek used to create R1-Zero. This will involve curating new, large-scale datasets for math, reasoning, and code.

πŸ”₯ Step 3: show we can go from base model -> SFT -> RL via multi-stage training.

Follow along: https://github.com/huggingface/open-r1
Β·
replied to lewtun's post 3 months ago
liked a dataset 10 months ago
liked a model 10 months ago
liked a model 11 months ago
upvoted a paper 12 months ago
New activity in BAAI/bge-reranker-base 12 months ago

Languages covered?

1
#16 opened 12 months ago by
oliviermills
liked a model about 1 year ago
commented a paper about 1 year ago
liked 4 models about 1 year ago
commented a paper over 1 year ago
liked 5 models over 1 year ago