Hugging Face's logo
Kevin Wu's picture
11

Kevin Wu

Flagpoles00
yinanrock's profile picture mariglentroka's profile picture CaramelDelight8491's profile picture
·

AI & ML interests

Always ready to learn!

Recent Activity

liked a model about 2 months ago
zed-industries/zeta
liked a Space about 2 months ago
fantaxy/adult-novel
liked a model about 2 months ago
AlejandroOlmedo/DeepSeek-R1-Distill-Qwen-7B-GRPO_Math-8bit-mlx
View all activity

Organizations

MLX Community's profile picture

Flagpoles00's activity

liked a model about 2 months ago
liked a Space about 2 months ago
liked a model about 2 months ago
reacted to Jaward's post with 👍 about 2 months ago
view post
Post
3896
Finally here it is: a faster, custom, scalable GRPO trainer for smaller models with < 500M params, can train on 8gb ram cpu, also supports gpu for sanity sake (includes support for vllm + flash attention). Using smolLM2-135M/360M-instructs as ref & base models. Experience your own “aha” moment 🐳 on 8gb ram.
Code: https://github.com/Jaykef/ai-algorithms/blob/main/smollm2_360M_135M_grpo_gsm8k.ipynb
  • 2 replies
·
liked 8 models about 2 months ago