Stars
Distributed Compiler based on Triton for Parallel Systems
Lightweight coding agent that runs in your terminal
High performance inference engine for diffusion models
FlexAttention based, minimal vllm-style inference engine for fast Gemma 2 inference.
Qwen-Image is a powerful image generation foundation model capable of complex text rendering and precise image editing.
User-friendly AI Interface (Supports Ollama, OpenAI API, ...)
An educational resource to help anyone learn deep reinforcement learning.
✂️ Clipboard access for local and remote tmux sessions
Replace 'hub' with 'ingest' in any GitHub URL to get a prompt-friendly extract of a codebase
Code and model for the paper "Improving Language Understanding by Generative Pre-Training"
Official codebase for "Self Forcing: Bridging Training and Inference in Autoregressive Video Diffusion" (NeurIPS 2025 Spotlight)
An implementation of model parallel autoregressive transformers on GPUs, based on the Megatron and DeepSpeed libraries
Build and share delightful machine learning apps, all in Python. 🌟 Star to support our work!
[NeurIPS'23 Oral] Visual Instruction Tuning (LLaVA) built towards GPT-4V level capabilities and beyond.
✨✨Latest Advances on Multimodal Large Language Models
Block Diffusion: Interpolating Between Autoregressive and Diffusion Language Models
NVIDIA Isaac GR00T N1.5 - A Foundation Model for Generalist Robots.
xDiT: A Scalable Inference Engine for Diffusion Transformers (DiTs) with Massive Parallelism
A unified inference and post-training framework for accelerated video generation.
Open-sourced codes for MiniGPT-4 and MiniGPT-v2 (https://minigpt-4.github.io, https://minigpt-v2.github.io/)
Implement a ChatGPT-like LLM in PyTorch from scratch, step by step
Modeling, training, eval, and inference code for OLMo
🤗 PEFT: State-of-the-art Parameter-Efficient Fine-Tuning.
[NeurIPS 2024] Official code for PuLID: Pure and Lightning ID Customization via Contrastive Alignment
[SIGGRAPH Asia 2025] DreamO: A Unified Framework for Image Customization
Official implementations for paper: VACE: All-in-One Video Creation and Editing