xdna

Close-to-metal Python API for AMD Ryzen™ AI NPUs, built on MLIR-AIE: declare operators, compile them into graphs, autotune them, and run models entirely on the NPU.

Run LLMs and Ollama Qwen across AMD XDNA1 NPU, CPU, and GPU on Ryzen AI Linux systems.

Practical applications for the AMD XDNA NPU on Linux — webcam effects (working) + embeddings/audio/ggml roadmap. Built on open-xdna.

Stable Diffusion image generation on AMD Ryzen AI NPUs for Linux

AI Engine RunTime, lite: language models and image generation on the NPU of AMD Ryzen AI processors

LLM and speech-to-text on the AMD Ryzen AI NPU with custom AI Engine kernels: LFM2.5-230M at 58 tok/s, Parakeet TDT 0.6B encoder at 185 ms per 10 s audio with FP32 accuracy. IRON/MLIR-AIE, XRT.

A lightweight TUI monitor for AMD Ryzen AI NPUs

Detect, validate, enable and assess the AMD XDNA NPU on Linux. Dependency-free. Gives an honest, machine-specific LLM-feasibility verdict for XDNA 1 (Phoenix) vs XDNA 2 (Strix).

Unlocking the power of the Halo NPU: NPU + iGPU live verification, compression, and routing research on AMD Strix Halo

Whisper encoder on the AMD XDNA1 NPU (Ryzen AI, Phoenix) under Linux — open toolchain only: XRT, MLIR-AIE/IRON, peano. Includes an HTTP service speaking whisper.cpp's contract.

Bazarr subtitles with XDNA1: Whisper on the AMD XDNA1 NPU (Ryzen 7 8845HS) with word timestamps, VAD, lyrics, speaker diarization and the whisper-asr-webservice API

Everyday local AI on the AMD Ryzen AI NPU under Linux: OpenAI-compatible gateway, dictation, OCR, semantic search, waybar status (FastFlowLM)

Ten CNNs benchmarked on AMD XDNA1 (Phoenix) NPU vs CPU vs iGPU, with an interactive web studio

Local AI on the NPU of AMD Ryzen AI laptops, on Linux: chat, vision, voice and generation with its own NPU engines; OpenAI/Ollama-compatible API

Text embeddings on AMD XDNA NPU (Strix Halo) — OpenAI-compatible /v1/embeddings endpoint powered by all-MiniLM-L6-v2 running on the NPU tile array

Research-grade DSP, SDR, transforms, and the retained M32/M33 PQC foundation on AMD Phoenix XDNA1 using MLIR-AIE, IRON, Peano, and XRT.

Talk to a 27B model on one Ryzen AI mini PC: Open WebUI + Qwen3.8-27B on the iGPU, speech in and out on the NPU. Nothing leaves the machine.

CosyVoice3 text-to-speech and Whisper speech-to-text on the AMD XDNA1 NPU (Ryzen AI 7040/8040), int8, next to an LLM on the same chip.

Experiments with the AMD XDNA1 NPU (Ryzen 7040/8040) on Linux: custom mlir-aie matmul kernels, benchmarks and MiniLM embeddings via XRT

Whisper ASR on the AMD Ryzen AI NPU — zero-dependency Bash wrapper over FastFlowLM, plus a RAPL power/latency benchmark harness. RTF ≈ 0.17.