A high-throughput and memory-efficient inference and serving engine for LLMs

Open deep learning compiler stack for cpu, gpu and specialized accelerators

NumPy & SciPy for GPU

Stable Diffusion web UI

Large-scale LLM inference engine

The most powerful and modular stable diffusion GUI, api and backend with a graph/nodes interface. Now ZLUDA enhanced for better AMD GPU performance.

Tensors and Dynamic neural networks in Python with strong GPU acceleration

vLLM: A high-throughput and memory-efficient inference and serving engine for LLMs

Simple monkeypatch to boost AMD Navi 3 GPUs

Example on how to use pytorch/yolov8 object detection on computers with AMD integrated GPUs

ROCm Install Utilities: rocminstall.py script to install a specific ROCm release version/revision.

Python Interface to HIP and hiprtc Library

TOML-annotated C header file format for packaging binary files, from Microsoft Research

HIP backend patch for Numba, the NumPy aware dynamic Python compiler using LLVM.

Voice-to-voice personal assistant, Full-local

Complete Guide how to run Pytorch with AMD rx460,470,480 (gfx803) GPUs

Benchmark for both NVIDIA and AMD GPU

This project has scripts to set up, build and test installation of AMD ROCm MIVisionX

Instructions on how to use PyTorch on AMD GPU with Linux

同步AI开发常用的docker镜像到阿里云镜像仓库，便于在国内快速拉取镜像。如：pytorch