Development repository for the Triton language and compiler

FlagTree is a unified compiler supporting multiple AI chip backends for custom Deep Learning operations, which is forked from triton-lang/triton.

Lightweight Image Video Action Generation Inference Framework

Domain-specific language designed to streamline the development of high-performance GPU/CPU/Accelerators kernels

FlagGems is an operator library for large language models implemented in the Triton Language.

FlagGems-sglang is part of FlagOS. It is a high-performance operator library designed for multiple hardware backends.

Next-Generation AI-Assisted Kernel Engineering for Multi-Chip Systems

A collection of memory efficient attention operators implemented in the Triton language.

Native C++/HIP attention kernels, shape specializations, and validated performance practices across AMD GPUs.

super repo for rocm libraries

FlyDSL is the Python front‑end of the project: a Flexible Layout Python DSL for expressing tiling, partitioning, data movement, and kernel structure at a high level.

An Efficient Self-Improving Agent Harness That Learns from Experience

Generating Efficient AI-Centric Kernels

An agentic system that auto-optimizes LLM workloads on AMD GPUs.

Repository for OpenCV's extra modules

LLM inference in C/C++

Open Source Computer Vision Library

An easy-to-use framework for large scale recommendation algorithms.

Examples for Recommenders - easy to train and deploy on accelerated infrastructure.

Repository for OpenVINO's extra modules

A unified inference runtime for VLA models.