grpo repositories

Trending repositories tagged grpo, ranked by trending score. Updated daily.

  1. 1
    datawhalechina/diy-llm

    Covers pre-training data, Tokenizer, Transformer, MoE,distributed training, Scaling Laws, inference & alignment .6 progressive code assignments for full-stack LLM learning | 涵盖预训练数据、分词器、Transformer、MoE、分布式训练、缩放定律、推理与对齐,6 项渐进代码作业,掌握 LLM 全栈知识

    GitHub · Jupyter Notebook · 3,023 stars · 296 forks · +37 stars in 24 hours · Trending score 3.20

  2. 2
    modelscope/ms-swift

    Use PEFT or Full-parameter to CPT/SFT/DPO/GRPO 600+ LLMs (Qwen3.8, DeepSeek-V4, GLM-5.1, InternLM3, Llama4, ...) and 300+ MLLMs (Qwen3-VL, Qwen3-Omni, InternVL3.5, Ovis2.5, GLM5.3, Gemma4, Llava, Phi4, ...) (AAAI 2025).

    GitHub · Python · 16,135 stars · 1,817 forks · +12 stars in 24 hours · Trending score 2.42

  3. 3
    devinxzhang/CycleGRPO

    [ECCV 2026] Official code for CycleGRPO and its video extension Generalized CycleGRPO — unifying region understanding and localization via caption-grounding cycle-consistency RL

    GitHub · Python · 17 stars · 0 forks · +8 stars in 24 hours · Trending score 1.54

  4. 4
    Nicozwy/FastRL

    NeurIPS 2026: Learn from the Gap: Differential-Aware Advantage Pruning with Adaptive Rollout Sampling for GRPO

    GitHub · Python · 58 stars · 2 forks · +2 stars in 24 hours · Trending score 1.43

  5. 5
    whitecircle/halo

    Halo is an open-source framework built by White Circle for training large language and multimodal models

    GitHub · Python · 868 stars · 50 forks · +2 stars in 24 hours · Trending score 1.32

  6. 6
    verl-project/verl-omni

    Multimodal RL training framework for diffusion & omni models

    GitHub · Python · 1,551 stars · 334 forks · 0 stars in 24 hours · Trending score 1.29

  7. 7
    OpenPipe/ART

    Agent Reinforcement Trainer: train multi-step agents for real-world tasks using GRPO. Give your agents on-the-job training. Reinforcement learning for Qwen3.6, GPT-OSS, Llama, and more!

    GitHub · Python · 10,856 stars · 1,016 forks · 0 stars in 24 hours · Trending score 0.93

  8. 8
    Gen-Verse/OpenClaw-RL

    OpenClaw-RL: Train any agent simply by talking

    GitHub · Python · 5,749 stars · 613 forks · +2 stars in 24 hours · Trending score 0.91

  9. 9
    redai-studio/Relax

    An Asynchronous Reinforcement Learning Engine for Omni-Modal Post-Training at Scale

    GitHub · Python · 678 stars · 209 forks · +1 stars in 24 hours · Trending score 0.82

  10. 10
    bcmi/Granular-GRPO

    [CVPR 2026] Fine-Grained GRPO for Precise Preference Alignment in Flow Models

    GitHub · Python · 66 stars · 2 forks · 0 stars in 24 hours · Trending score 0.54

  11. 11
    oumi-ai/oumi

    Easily fine-tune (SFT/RL), evaluate, and deploy Qwen, Gemma, or any open agentic LLM/VLM!

    GitHub · Python · 9,436 stars · 798 forks · 0 stars in 24 hours · Trending score 0.47

  12. 12
    adithya-s-k/FineEnvs

    FineEnvs — RL Environments 101: building and scaling RL environments in the age of LLMs

    GitHub · Python · 563 stars · 78 forks · 0 stars in 24 hours · Trending score 0.43

  13. 13
    hud-evals/hud-python

    RL environments + evals for AI agents. Define once, train anything.

    GitHub · Python · 329 stars · 92 forks · −1 stars in 24 hours · Trending score 0.33

  14. 14
    V01dMur10c/awesome-agent-rl-environments

    A curated list of training & evaluation environments for LLM/VLM agents (SWE-Gym, GEM, RAGEN, AgentGym, WebArena, OSWorld, ToolBench…). Updated weekly.

    GitHub · 12 stars · 5 forks · 0 stars in 24 hours · Trending score 0.32

  15. 15
    inclusionAI/AReno

    An easy-to-use, fast toolkit to scale up RL post-training on a single node.

    GitHub · Python · 389 stars · 141 forks · 0 stars in 24 hours · Trending score 0.30

  16. 16
    Zihao-Wu06/distillops

    Black-box LLM distillation for any verifiable task, with text or image inputs: teacher API -> rejection sampling -> SFT -> eval -> GRPO

    GitHub · Python · 101 stars · 7 forks · 0 stars in 24 hours · Trending score 0.25

  17. 17
    DeepGym/deepgym

    RL environments with verifiable rewards you can trust. Sandboxed coding and computer-use tasks, verifiers kept apart from the answer, a QA gate against reward hacking, and infra failures kept out of training. Works with verl, TRL, OpenRLHF and Axolotl.

    GitHub · Python · 22 stars · 5 forks · −3 stars in 24 hours · Trending score 0.19

  18. 18
    ssddxg/agentic-rl-forge

    Local-first Agent RL workbench for Search-R1 rollouts, reward evaluation, GRPO/verl training-data handoff, and an auxiliary private RAG workspace.

    GitHub · Python · 22 stars · 0 forks · 0 stars in 24 hours · Trending score 0.18

  19. 19
    HodeTech/ForgeLM

    Config-driven LLM fine-tuning with safety evaluation, EU AI Act compliance, 6 alignment methods, and one-command bundled quickstart templates.

    GitHub · Python · 9 stars · 12 forks · 0 stars in 24 hours · Trending score 0.14

  20. 20
    JudgmentLabs/judgeval

    The Continuous-Improvement Stack for Agents. Our environment data and evals power agent improvement and monitoring.

    GitHub · Python · 1,073 stars · 107 forks · −1 stars in 24 hours · Trending score 0.09

  21. 21
    SourceShift/mini-ork

    Prove an AI-written fix actually fixes the bug. mini-ork certify reproduces the bug against your repo's own code, attacks the patch with generated invariants in a Docker sandbox, and returns PROVEN / REFUTED / UNVERIFIED with a JSON certificate. Part of a verifier-gated, multi-model agent runtime.

    GitHub · Python · 35 stars · 11 forks · 0 stars in 24 hours · Trending score 0.09

  22. 22
    mii-llm/palingenesis

    Post-train LLMs and agents from one config: SFT, multi-turn RL in tool environments (MCP, Harbor, OpenEnv), on-policy distillation, DPO and merging, with state-of-the-art defaults.

    GitHub · Python · 7 stars · 2 forks · 0 stars in 24 hours · Trending score 0.09

  23. 23
    basetenlabs/loops-cookbook

    Baseten Loops example recipes and patterns

    GitHub · Python · 7 stars · 0 forks · −1 stars in 24 hours · Trending score 0.08

  24. 24
    tterrasson/retrograd

    Run and Fine-tune GGUF models, with LoRA or full weights, from SFT to agentic GRPO

    GitHub · Rust · 5 stars · 0 forks · 0 stars in 24 hours · Trending score 0.07