Our first fully AI generated deep learning system

High Performance LLM Inference Operator Library

Tile-Based Runtime for Ultra-Low-Latency LLM Inference

Ongoing research training transformer models at scale

[HPCA 2026] A GPU-optimized system for efficient long-context LLMs decoding with low-bit KV cache.

Light Image Video Generation Inference Framework

mHC kernels implemented in CUDA

A compact implementation of SGLang, designed to demystify the complexities of modern LLM serving systems.

CUDA Tile IR is an MLIR-based intermediate representation and compiler infrastructure for CUDA kernel optimization, focusing on tile-based computation patterns and optimizations targeting NVIDIA te…

Accelerating MoE with IO and Tile-aware Optimizations

Helpful kernel tutorials and examples for tile-based GPU programming

cuTile is a programming model for writing parallel kernels for NVIDIA GPUs

NVIDIA cuTile learn

Official inference repo for FLUX.2 models

[NeurIPS'24 Spotlight, ICLR'25, ICML'25] To speed up Long-context LLMs' inference, approximate and dynamic sparse calculate the attention, which reduces inference latency by up to 10x for pre-filli…

Provide with pre-build flash-attention package wheels on Linux and Windows platforms using GitHub Actions

Open Neural Network Exchange to C compiler.

Contexts Optical Compression

Building the Virtuous Cycle for AI-driven LLM Systems

📄 Awesome CV is LaTeX template for your outstanding job application

A Python-embedded DSL that makes it easy to write fast, scalable ML kernels with minimal boilerplate.

Efficient End2End Compiler for Mixed-Precision Deep Learning

a size profiler for cuda binary

🐹 Deep clean and optimize your Mac.

Sparse Attention; Sparse Linear; Diffusion Transformer

Open ABI and FFI for Machine Learning Systems