Starred repositories
Claude Code 源码逆向恢复项目 | Source Map 逆向 · 架构分析 · 可运行版本 | 1900+ 文件 · 51万行代码 · 12章节课程
[CVPR 2026 Highlight] Unified Generation and Self-Verification for Vision-Language Models via Advantage Decoupled Preference Optimization
An Asynchronous Reinforcement Learning Engine for Omni-Modal Post-Training at Scale
App for Claude Code / Codex / Gemini / OpenCode, vibe coding anytime, anywhere
ARIS ⚔️ (Auto-Research-In-Sleep) — Lightweight Markdown-only skills for autonomous ML research: cross-model review loops, idea discovery, and experiment automation. No framework, no lock-in — works…
Tempo: Small Vision-Language Models are Smart Compressors for Long Video Understanding
A LaTeX class for producing presentations and slides
Towards Scalable Pre-training of Visual Tokenizers for Generation
Official Code of "Distribution Matching Distillation Meets Reinforcement Learning"
Harness engineering official style beginner tutorial, from 0 to 1
Conversational literature research workflow for agents
Official implementation of "Fast-dLLM: Training-free Acceleration of Diffusion LLM by Enabling KV Cache and Parallel Decoding"
Optimize prompts, code, and more with AI-powered Reflective Text Evolution
The repo is finally unlocked. enjoy the party! The fastest repo in history to surpass 100K stars ⭐. Join Discord: https://discord.gg/5TUQKqFWd Built in Rust using oh-my-codex.
🤖 The analysis of Claude Code
轻量级大语言模型MiniMind的源码解读，包含tokenizer、RoPE、MoE、KV Cache、pretraining、SFT、LoRA、DPO等完整流程
🎓从0开始训练一个大模型Minimind项目的超详细解析，包括但不限于用到的架构，算法，以及大模型面试经验
The Cowork Agent for Everything — trainable advertising AI + 14 platform MCP servers + agent skills. Based on minimind (42k stars). Train from zero in 2 hours.
Video PreTraining (VPT): Learning to Act by Watching Unlabeled Online Videos
[arXiv 2025] SAGE: Training Smart Any-Horizon Agents for Long Video Reasoning with Reinforcement Learning
Designing Multi-Agent Systems with Zero Supervision
Latent Collaboration in Multi-Agent Systems