← 최신 프로덕트

NanoRL

RL training for LLMs in ~1,800 lines. One RL training loop that runs CartPole on a laptop and disaggregated async RLVR on a GPU cluster. About 1,800 lines across 7 files. No Ray, no TRL, no DeepSpeed. It's small enough to read in an afternoon, and it's meant to be forked, not imported. The smallest async RL trainer I could write: one loop that runs REINFORCE on CartPole on a laptop and async GRPO on a cluster (e.g.

오픈소스

13 스타 · GitHub외부 사이트

github.com 방문외부 사이트

반응 (1)

비슷한 프로덕트

설명이 비슷한 순

  1. Minimal LLM Post-Training Experiments on an 8GB GPU (SFT

    github.com

    DPO, GRPO)

    Hacker News21점

  2. I built a tiny LLM to demystify how language models work

    github.com

    A ~9M parameter LLM that talks like a small fish.

    Hacker News915점

  3. Tiny-vLLM

    github.com

    high performance LLM inference engine in C++ and CUDA

    Hacker News205점

  4. Terminal-Bench-RL

    github.com

    Training long-horizon terminal agents with RL

    Hacker News125점

  5. LLM App

    github.com

    build a realtime LLM app in 30 lines, with no vector database

    Hacker News11점