Awesome MoE Diffusion Models

FlashInfer: Kernel Library for LLM Serving

AI agents running research on single-GPU nanochat training automatically

LLM驱动的 A/H/美股智能分析器：多数据源行情 + 实时新闻 + LLM决策仪表盘 + 多渠道推送，零成本定时运行，纯白嫖. LLM-powered stock analysis system for A/H/US markets.

Gen-Searcher: Reinforcing Agentic Search for Image Generation

Wan: Open and Advanced Large-Scale Video Generative Models

Train speculative decoding models effortlessly and port them smoothly to SGLang serving.

DFlash: Block Diffusion for Flash Speculative Decoding

Your own personal AI assistant. Any OS. Any Platform. The lobster way. 🦞

Model compression toolkit engineered for enhanced usability, comprehensiveness, and efficiency.

WeDLM: The fastest diffusion language model with standard causal attention and native KV cache compatibility, delivering real speedups over vLLM-optimized baselines.

记录量化LLM中的总结。

TurboDiffusion: 100–200× Acceleration for Video Diffusion Models

[ICLR2025, ICML2025, NeurIPS2025 Spotlight] Quantized Attention achieves speedup of 2-5x compared to FlashAttention, without losing end-to-end metrics across language, image, and video models.

PyTorch implementation of JiT https://arxiv.org/abs/2511.13720

SGLang is a high-performance serving framework for large language models and multimodal models.

Implementation of "Live Avatar: Streaming Real-time Audio-Driven Avatar Generation with Infinite Length"

[ICLR 2026] Taming large-scale few-step training with self-adversarial flows! 👏🏻

Nano vLLM

(arXiv) MixGRPO: Unlocking Flow-based GRPO Efficiency with Mixed ODE-SDE

An official implementation of DanceGRPO: Unleashing GRPO on Visual Generation

Light Image Video Generation Inference Framework

LightLLM is a Python-based LLM (Large Language Model) inference and serving framework, notable for its lightweight design, easy scalability, and high-speed performance.

Qwen-Image-Lightning: Speed up Qwen-Image model with distillation

Dimple, the first Discrete Diffusion Multimodal Large Language Model

[NeurIPS 2025] An official implementation of Flow-GRPO: Training Flow Matching Models via Online RL

dInfer: An Efficient Inference Framework for Diffusion Language Models

DeepGEMM: clean and efficient FP8 GEMM kernels with fine-grained scaling

This project is the official implementation of 'DreamOmni2: Multimodal Instruction-based Editing and Generation (CVPR2026 Highlight)''