Stars
Implementations of attention with the softpick function, naive and FlashAttention-2
MAGI-1: Autoregressive Video Generation at Scale
Pytorch implementation for the paper titled "SimpleAR: Pushing the Frontier of Autoregressive Visual Generation"
程序员在家做饭方法指南。Programmer's guide about how to cook at home (Simplified Chinese only).
[SIGGRAPH 2025] One Model to Rig Them All: Diverse Skeleton Rigging with UniRig
Lumina-mGPT 2.0: Stand-alone Autoregressive Image Modeling
SparseFlex: High-Resolution and Arbitrary-Topology 3D Shape Modeling
TripoSG: High-Fidelity 3D Shape Synthesis using Large-Scale Rectified Flow Models
Understanding R1-Zero-Like Training: A Critical Perspective
The official implementation of Autoregressive Image Generation using Residual Quantization (CVPR '22)
Roblox Foundation Model for 3D Intelligence
Official code of DeepMesh: Auto-Regressive Artist-mesh Creation with Reinforcement Learning
[CVPR 2025 Best Paper Award Candidate] VGGT: Visual Geometry Grounded Transformer
[CVPR 2025] Code for Deformable Radial Kernel Splatting
[CVPR 2025] MIDI: Multi-Instance Diffusion for Single Image to 3D Scene Generation
Witness the aha moment of VLM with less than $3.
Large Language Model Text Generation Inference
SGLang is a fast serving framework for large language models and vision language models.
Wan: Open and Advanced Large-Scale Video Generative Models
FlashMLA: Efficient MLA decoding kernels
[ICLR 2025] Official PyTorch Implementation of Gated Delta Networks: Improving Mamba2 with Delta Rule
High-Resolution 3D Assets Generation with Large Scale Hunyuan3D Diffusion Models.
[SIGGRAPH Asia 2024, Best Paper Honorable Mention] This is the official implementation of our SIGGRAPH Asia journal artical: TEXGen: a Generative Diffusion Model for Mesh Textures
AnchorAttention: Improved attention for LLMs long-context training