Flexible Python configuration system. The last one you will ever need.

Main Web Site (Online Books)

An extremely fast Python package and project manager, written in Rust.

An extremely fast Python linter and code formatter, written in Rust.

Bring projects, wikis, and teams together with AI. AppFlowy is the AI collaborative workspace where you achieve more without losing control of your data. The leading open source Notion alternative.

A Telegram bot to recommend arXiv papers

Puzzles for learning Triton, play it with minimal environment configuration!

veRL: Volcano Engine Reinforcement Learning for LLM

An efficient video loader for deep learning with smart shuffling that's super easy to digest

A generative world for general-purpose robotics & embodied AI learning.

Implementation of Classifier Free Guidance in Pytorch, with emphasis on text conditioning, and flexibility to include multiple text embedding models

Implementation of 💍 Ring Attention, from Liu et al. at Berkeley AI, in Pytorch

Implementation of RLHF (Reinforcement Learning with Human Feedback) on top of the PaLM architecture. Basically ChatGPT but with PaLM

Memory-optimized training scripts for video models based on Diffusers

Implementation of Denoising Diffusion Probabilistic Model in Pytorch

Efficient, Flexible and Portable Structured Generation

Code repository of all OpenGL chapters from the book and its accompanying website https://learnopengl.com

Materials for learning SGLang

My learning notes/codes for ML SYS.

《Python Cookbook》 3rd Edition Translation

FastAPI framework, high performance, easy to learn, fast to code, ready for production

Quantized Attention that achieves speedups of 2.1-3.1x and 2.7-5.1x compared to FlashAttention2 and xformers, respectively, without lossing end-to-end metrics across various models.

Scalene: a high-performance, high-precision CPU, GPU, and memory profiler for Python with AI-powered optimization proposals

Official inference framework for 1-bit LLMs

中文版 llm-numbers

Flutter makes it easy and fast to build beautiful apps for mobile and beyond

Code of Pyramidal Flow Matching for Efficient Video Generative Modeling

📚200+ Tensor/CUDA Cores Kernels, ⚡️flash-attn-mma, ⚡️hgemm with WMMA, MMA and CuTe (98%~100% TFLOPS of cuBLAS/FA2 🎉🎉).

Vchitect-2.0: Parallel Transformer for Scaling Up Video Diffusion Models