Repositories list

• LightLLM

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

  nlpdeep-learningllama+ 4gptmodel-servingllmopenai-triton

  Python

  •

  Apache License 2.0

  •277•3.6k•79•22•Updated Sep 6, 2025Sep 6, 2025

• LightX2V

  Public
  Light Video Generation Inference Framework

  video-generationdiffusion-modelshunyuan-video+ 1wan-video

  Python

  •30•541•22•1•Updated Sep 5, 2025Sep 5, 2025

• mtc-token-healing

  Public
  Token healing implementation in Rust

  Rust

  •

  Apache License 2.0

  •0•4•0•4•Updated Sep 5, 2025Sep 5, 2025

• ComfyUI-Lightx2vWrapper

  Public
  ComfyUI custom node for lightx2v

  comfyuicomfyui-nodes

  Python

  •

  MIT License

  •5•36•0•0•Updated Sep 5, 2025Sep 5, 2025

• LightKernel

  Public
  HTML

  •

  Apache License 2.0

  •0•1•0•0•Updated Sep 5, 2025Sep 5, 2025

• lightllm-blog

  Public
  SCSS

  •

  MIT License

  •0•1•0•0•Updated Sep 4, 2025Sep 4, 2025

• general-sam-py

  Public
  Python bindings for general-sam and some utilities

  Python

  •

  Apache License 2.0

  •0•4•0•3•Updated Sep 2, 2025Sep 2, 2025

• Qwen-Image-Lightning

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

  Python

  •

  Apache License 2.0

  •24•575•5•0•Updated Aug 29, 2025Aug 29, 2025

• Wan2.2-Lightning

  Public
  Wan2.2-Lightning: Speed up wan2.2 model with distillation

  Python

  •

  Apache License 2.0

  •371•130•14•0•Updated Aug 28, 2025Aug 28, 2025

• LightCompress

  Public
  A powerful toolkit for compressing large models including LLM, VLM, and video generation models.

  benchmarkdeploymenttool+ 15evaluationpruningquantizationwanawqlarge-language-modelsllm+ 8

  Python

  •

  Apache License 2.0

  •61•556•38•0•Updated Aug 22, 2025Aug 22, 2025

• general-sam

  Public
  A general suffix automaton implementation in Rust with Python bindings

  Rust

  •

  Apache License 2.0

  •0•7•0•3•Updated Aug 18, 2025Aug 18, 2025

• greedy-tokenizer

  Public
  Greedily tokenize strings with the longest tokens iteratively.

  Python

  •

  Apache License 2.0

  •0•0•0•2•Updated Aug 18, 2025Aug 18, 2025

• SageAttention

  Public
  Quantized Attention achieves speedup of 2-5x and 3-11x compared to FlashAttention and xformers, without lossing end-to-end metrics across language, image, and video models.

  Cuda

  •

  Apache License 2.0

  •209•2•0•0•Updated Aug 16, 2025Aug 16, 2025

• fa3

  Public
  Python

  •

  BSD 3-Clause "New" or "Revised" License

  •1•0•0•0•Updated Aug 7, 2025Aug 7, 2025

• flash-attn-3-build

  Public
  Dockerfile

  •2•0•0•0•Updated Jul 24, 2025Jul 24, 2025

• HarmoniCa

  Public
  [ICML 2025] This is the official PyTorch implementation of "🎵 HarmoniCa: Harmonizing Training and Inference for Better Feature Caching in Diffusion Transformer Acceleration".

  accelerationicmldit+ 6pixartdiffusion-modelsdiffusion-transformerpixart-sigmafeature-cachingicml-2025

  Python

  •

  Apache License 2.0

  •1•42•2•0•Updated Jul 10, 2025Jul 10, 2025

• TFMQ-DM

  Public
  [CVPR 2024 Highlight & TPAMI 2025] This is the official PyTorch implementation of "TFMQ-DM: Temporal Feature Maintenance Quantization for Diffusion Models".

  highlightquantizationcvpr+ 8ldmdiffusion-modelstpamipost-training-quantizationddimstable-diffusioncvpr2024+ 1

  Jupyter Notebook

  •

  Apache License 2.0

  •4•103•0•0•Updated Jul 10, 2025Jul 10, 2025

• LightTTS

  Public
  Light-tts is a lightweight TTS inference framework optimized for CosyVoice2, enabling fast and scalable speech synthesis in Python.

  Python

  •

  Apache License 2.0

  •0•7•0•0•Updated Jun 24, 2025Jun 24, 2025

• OmniBal

  Public
  [ICML 2025] This is the official PyTorch implementation of "OmniBal: Towards Fast Instruction-Tuning for Vision-Language Models via Omniverse Computation Balance".

  vlmicml-2025

  Python

  •

  Apache License 2.0

  •3•24•3•0•Updated Jun 16, 2025Jun 16, 2025

• lightx2v_comfyui_node

  Public
  0•0•0•0•Updated Apr 28, 2025Apr 28, 2025

• MQBench

  Public
  Model Quantization Benchmark

  Python

  •

  Apache License 2.0

  •141•833•10•5•Updated Apr 20, 2025Apr 20, 2025

• flash-attention

  Public
  Fast and memory-efficient exact attention

  Python

  •

  BSD 3-Clause "New" or "Revised" License

  •2k•0•0•0•Updated Apr 17, 2025Apr 17, 2025

• verl

  Public
  verl: Volcano Engine Reinforcement Learning for LLMs

  Python

  •

  Apache License 2.0

  •2.3k•1•0•0•Updated Mar 17, 2025Mar 17, 2025

• LLM_QAT

  Public
  Python

  •0•0•0•0•Updated Feb 19, 2025Feb 19, 2025

• modeltc.github.io

  Public
  HTML

  •0•0•0•0•Updated Jan 25, 2025Jan 25, 2025

• quant_horizon

  Public
  Cuda

  •

  Apache License 2.0

  •1•11•0•0•Updated Jan 10, 2025Jan 10, 2025

• EasyLLM

  Public
  Built upon Megatron-Deepspeed and HuggingFace Trainer, EasyLLM has reorganized the code logic with a focus on usability. While enhancing usability, it also ensures training efficiency.

  Python

  •

  Apache License 2.0

  •8•49•0•0•Updated Sep 18, 2024Sep 18, 2024

• DeepSpeed

  Public
  DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.

  Python

  •

  Apache License 2.0

  •4.5k•0•0•0•Updated Sep 13, 2024Sep 13, 2024

• opencompass

  Public
  OpenCompass is an LLM evaluation platform, supporting a wide range of models (Llama3, Mistral, InternLM2,GPT-4,LLaMa2, Qwen,GLM, Claude, etc) over 100+ datasets.

  Python

  •

  Apache License 2.0

  •656•1•0•0•Updated Sep 6, 2024Sep 6, 2024

• xtuner

  Public
  An efficient, flexible and full-featured toolkit for fine-tuning LLM (InternLM2, Llama3, Phi3, Qwen, Mistral, ...)

  Python

  •

  Apache License 2.0

  •354•0•0•0•Updated Aug 22, 2024Aug 22, 2024