ryzen-ai-max-395 开源项目推荐（8 个）

hogeheer499-commits /strix-halo-guide

AMD Strix Halo本地LLM指南：为Ryzen AI MAX+ 395 / Radeon 8060S、Ollama、llama.cpp Vulkan/RADV、ROCm、原始证据、直接100 t/s 30B Qwen、140 t/s CHADROCK MTP、120B GGUF。

1bit-MONSTER /engine

One binary to rule them all: model-agnostic local LLM inference for AMD Strix Halo (Ryzen AI MAX+ 395) on Linux. Radeon 8060S iGPU (gfx1151) via HRX + our HIP kernels, XDNA 2 NPU via our NPU runtime, inside Lemonade Server with an OpenAI-compatible API. Qwen3.8, ZAYA1, BitNet b1.58, any GGUF. Apache-2.0.

kingjones30 /strix-halo-quant-lab

Run large LLMs locally on AMD Ryzen AI Max+ 395 (Strix Halo, gfx1151) with ROCmFP4 4-bit quantization. Measured benchmarks, build + serving recipes, and 118 ready-to-run GGUF models.

botAGI /strix-halo-llm-benchmarks

Measured LLM benchmarks for AMD Strix Halo / Ryzen AI Max+ 395 (Radeon 8060S, 128 GB unified): llama.cpp Vulkan & ROCm — decode pace, TTFA, prompt cache, quants, sustained load. Every number links to raw runs.

hec-ovi /laguna-s-2.1-vllm-strix

vLLM serving poolside Laguna S 2.1 (118B MoE, 8B active, INT4) on AMD Strix Halo (Ryzen AI Max+ 395, gfx1151, 128GB unified) via TheRock ROCm nightly. OpenAI-compatible, 256K context, DFlash speculative decoding.

Graphene-Lab /AMD-Ryzen-AI-Max-395-Privacy-Agent

Private, self-hosted AI office assistant for the AMD Ryzen AI Max+ 395 processor (16-core x86, Windows 11 and Linux): local documents, office files, email, Telegram, voice — fully offline with a local model.

kanlas-net /fevm-faex1-bios-flash

FEVM FAEX1 BIOS tooling: extract AGESA/PI from FAEX1 firmware images and build a UEFI Shell USB stick for flashing (AfuEfix64). Tested on FEVM FAEX1, Ryzen AI Max+ 395 (Strix Halo).

1bit-MONSTER /site

Source of https://1bit.gg/, docs and blog of the 1bit engine: model-agnostic local LLM inference on AMD Strix Halo (Ryzen AI MAX+ 395) inside Lemonade Server, on the Radeon 8060S iGPU via HRX and the XDNA 2 NPU. Apache-2.0.