Post-training with Tinker

MMaDA - Open-Sourced Multimodal Large Diffusion Language Models

Solve Visual Understanding with Reinforced VLMs

Mixture-of-Recursions: Learning Dynamic Recursive Depths for Adaptive Token-Level Computation

A programming framework for agentic AI 🤖 PyPi: autogen-agentchat Discord: https://aka.ms/autogen-discord Office Hour: https://aka.ms/autogen-officehour

[EMNLP 2025] TokenSkip: Controllable Chain-of-Thought Compression in LLMs

Code for paper: Optimizing Length Compression in Large Reasoning Models

[EMNLP 2025] LightThinker: Thinking Step-by-Step Compression

Official repository for paper: O1-Pruner: Length-Harmonizing Fine-Tuning for O1-Like Reasoning Pruning

DeepSeek 系列工作解读、扩展和复现。

Verifiers for LLM Reinforcement Learning

Scalable RL solution for the advanced reasoning of language models

OpenR: An Open Source Framework for Advanced Reasoning with Large Language Models

This is a replicate of DeepSeek-R1-Zero and DeepSeek-R1 training on small models with limited data

An Easy-to-use, Scalable and High-performance RLHF Framework (70B+ PPO Full Tuning & Iterative DPO & LoRA & RingAttention & RFT)

Train transformer language models with reinforcement learning.

欢迎来到 LLM-Dojo，这里是一个开源大模型学习场所，使用简洁且易阅读的代码构建模型训练框架(支持各种主流模型如Qwen、Llama、GLM等等)、RLHF框架(DPO/CPO/KTO/PPO)等各种功能。👩‍🎓👨‍🎓

Making LLaVA Tiny via MoE-Knowledge Distillation

This my attempt to create Self-Correcting-LLM based on the paper Training Language Models to Self-Correct via Reinforcement Learning by google

A Framework of Small-scale Large Multimodal Models

SuperCorrect: Supervising and Correcting Language Models with Error-Driven Insights

g1: Using Llama-3.1 70b on Groq to create o1-like reasoning chains

Paper Reproduction Google SCoRE(Training Language Models to Self-Correct via Reinforcement Learning)

Official code implementation of General OCR Theory: Towards OCR-2.0 via a Unified End-to-end Model