AMD Logo

AMD AI DEMOS

Fooocus

Fooocus

A powerful AI image generation tool optimized for AMD GPUs, delivering unparalleled efficiency and high-quality rendering.

Kotaemon

Kotaemon

Discover Kotaemon: the next-gen LLM powered by Ollama, enhanced with RAG, and accelerated by AMD MI210 GPUs. Get precise, context-aware responses beyond static training data, with lightning-fast performance and full local control. Test Kotaemon now and experience the future of AI!

LivePortrait

LivePortrait

Real-time AI-powered facial animation, utilizing AMD GPUs to transform static images into dynamic and lifelike portraits.

Flux

Flux

AI-powered video style transfer, bringing artistic transformations to your footage with AMD-optimized deep learning models.

XTTS

XTTS

Cutting-edge text-to-speech AI, delivering natural and expressive voices with AMD GPU acceleration for real-time synthesis.

LLaVA-Med

LLaVA-Med

Large Language and Vision Assistant for Medical Research Demo power by AMD MI210 and ROCM !!! let's have a try.

XTTS

WAN 2.2

Experience the magic of WAN 2.2, a cutting-edge AI video generator running entirely on AMD hardware. Powered by ROCm and accelerated by the Radeon Instinct MI210, WAN 2.2 transforms short prompts into stunning, high-quality videos — fast, fluid, and locally hosted.

This demo showcases how AMD GPUs deliver real performance and creativity without compromise, bringing generative video AI to life on open, efficient, and scalable technology.

YOLO

YOLO

Identify what is on your image

StarterPack

Sarter Pack

Create your starter pack fom a photo or the webcam with the latest Flux2.klein 9B running on MI210

FaceFusion

Face Fusion

FaceFusion Demo FaceFusion is a real-time face swapping and face enhancement demo designed to showcase modern AI inference on high-performance hardware. It combines face detection, alignment, and identity transfer to seamlessly replace a reference face while preserving expressions, pose, and image quality. This demo runs on AMD GPUs with ROCm acceleration, highlighting efficient inference using HIP, ORT, and optimized deep learning pipelines. It is intended for technical demonstrations, performance evaluation, and AI experimentation, not for production or misuse. FaceFusion demonstrates how advanced computer vision models can be deployed as interactive services, integrated with web interfaces, and accelerated on AMD platforms.