Official Implementation of Rectified Flow (ICLR2023 Spotlight)

Official implementation for the paper "DoLa: Decoding by Contrasting Layers Improves Factuality in Large Language Models"

Official repository of ’Visual-RFT: Visual Reinforcement Fine-Tuning’

Lightning Attention-2: A Free Lunch for Handling Unlimited Sequence Lengths in Large Language Models

HunyuanVideo-I2V: A Customizable Image-to-Video Model based on HunyuanVideo

Large World Model -- Modeling Text and Video with Millions Context

MM-EUREKA: Exploring Visual Aha Moment with Rule-based Large-scale Reinforcement Learning

Paper List of Inference/Test Time Scaling/Computing

[ICLR 2025] Autoregressive Video Generation without Vector Quantization

DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model

FlashMLA: Efficient MLA decoding kernels

DeepEP: an efficient expert-parallel communication library

A very simple GRPO implement for reproducing r1-like LLM thinking.

📖 This is a repository for organizing papers, codes and other resources related to unified multimodal models.

s1: Simple test-time scaling

🚀 Efficient implementations of state-of-the-art linear attention models in Pytorch and Triton

code for "Diffusion Forcing: Next-token Prediction Meets Full-Sequence Diffusion"

Fully open reproduction of DeepSeek-R1

[ICLR 2025] Reconstructive Visual Instruction Tuning

Infinity ∞ : Scaling Bitwise AutoRegressive Modeling for High-Resolution Image Synthesis

Model Compression Toolbox for Large Language Models and Diffusion Models

Code for Neurips24 paper: QuaRot, an end-to-end 4-bit inference of large language models.

Code repo for the paper "SpinQuant LLM quantization with learned rotations"

SANA: Efficient High-Resolution Image Synthesis with Linear Diffusion Transformer

[CVPR 2025] CoDe: Collaborative Decoding Makes Visual Auto-Regressive Modeling Efficient