PKGBUILDs to build stable AMDVLK-releases from source (amdvlk-git, lib32-amdvlk-git, amdvlk-bin)
在基于TypeScript的LayaAir HTML5游戏开发中使用AMD
Rise a lite AMD application in minutes
🚚 On-demand AMD npm package bundling service
Threejs Controls as modules (from examples folder)
Observed during normal image generation. In our case, ComfyUI-N-Nodes / RIFE appears to keep GPU memory outside ComfyUI's model manager. Under heavier...
AMD frequency and community NVIDIA metrics patches for Stats 3.0.23, plus a local JSON feed for VK03 Panel.
A multimodal Retrieval-Augmented Generation (RAG) pipeline for the AMD AI League. Features exact-set citations, VLM image parsing, and a persistent Un...
Real-time full-duplex voice AI (NVIDIA PersonaPlex-7B) on AMD Radeon under native Windows: ROCm PyTorch + custom HIP kernels, 2.5x real time on an RX...
Save/load MiniMax H3 audio+video (nested) latents in ComfyUI: the bridge for a two-pass H3 render on low-VRAM GPUs.
AMD 歌词汉化下载器(液态玻璃风味)
Frontend for Qwen-Image-2.1 inference on AMD ROCm: text-to-image and image editing via headless ComfyUI, one-step installer for Linux and Windows
Demo DLSS 5 neural rendering on a live webcam, comparing skin, hair, and fabric quality side by side in real time.
AMD ROCm云端光学–SAR建筑损毁分类与复核优先级研究|5476栋|跨事件评估、理想复核模拟与真实负结果
Ternary LLMs on llama.cpp, any GPU — run a 27B model in 5.5–6.7 GB. Prebuilt Linux/Windows binaries, Vulkan and CUDA, OpenAI-compatible API.
Windows memory-allocation fix for AMD Strix Halo / Ryzen AI Max+ 395 with 128 GB RAM: why gpt-oss-120b class models fail to load at a 96 GB VRAM carve...
A HIP/ROCm build toolchain for consumer Radeon cards on Windows, assembled without installing anything and without touching the gaming driver. Verifie...
Adaptive AMD RDNA inference runtime: benchmark Atomic + current llama.cpp, then promote the best backend safely.
Run JAX and MuJoCo MJX on AMD Ryzen AI MAX (Strix Halo, gfx1151) from pip wheels - no system ROCm - with a hang-safe check that tells you which wheel...
Local-first private AI agent with deterministic permission and audit boundaries.
Minimal AITER LDS and ROCr idle-CPU fixes for Radeon AI PRO R9700 on vLLM 0.29.0 / ROCm 7.2.3
Community ROCm 7.2 / RDNA4 gfx1201 patch stack for nvdiffrast. Non-commercial use restrictions may apply; see the NVIDIA Source Code License.
A lightweight utility to toggle AMD CPU Boost on Linux — cooler thermals and quieter fans by default, full clocks on demand.
Human-governed visitor-risk operations with deterministic controls, bounded Fireworks AI review, audit trails and portable proof packets.
Circuit OS: AI hardware compiler. Plain English to validated schematic, Arduino firmware, SPICE simulation and BOM. Runs on AMD ROCm via vLLM.
Guia completo em PT-BR: AMD BC-250 (APU do PS5) para games com CachyOS - BIOS, ACPI, governor, unlock de 8 núcleos e 40 CUs, overclock, scripts e FAQ
Architecture analysis: why modern gaming still stutters — the missing overlay coordination layer in Windows (DXGI swapchain, MPO, DLL injection, TDR)....
AMD 平台黑苹果(Hackintosh)OpenCore EFI 主线工作流技能:八步流程 + 8 份参考文档 + 6 个零第三方依赖的 Python 脚本
Auditable hybrid ML and agentic trading research with evidence verification, portfolio risk controls, and AMD ROCm receipts.
RDNA-first OpenAI-compatible local embedding engine for llama.cpp
非官方实测:在 Windows 上让 AMD RX 6800 XT (gfx1030/RDNA2) 原生跑 PyTorch GPU 加速。含路线排除、8 条踩坑清单与自检脚本。
OpenAI-compatible LLM server for AMD Radeon RX 9070 / 9070 XT (RDNA4). Native fp8/int4 WMMA — nothing dequantizes to fp16. One docker compose up.
BuffAdvisor (HackCU 2025): CU Boulder campus assistant running a Quark-quantized DeepSeek-R1-Distill-Llama-8B locally on an AMD AI PC, with RAG.
Run Qwen3.8-Flash-Next EXL3 (turboderp pack) on one AMD Strix Halo / Framework Desktop (gfx1151) via the exllamav3-amd runtime fork. 41 tok/s mean, 47...
AMD Ryzen 7900X / Raphael iGPU VFIO passthrough on Arch Linux: Windows 11 Code 43 reset workaround, Looking Glass, libvirt XML and VirtIO tuning
Real-time energy monitoring. Tracks CPU package, cores, uncore, DRAM, and platform total (PSYS) power at configurable intervals. Logs to SQLite, gener...
A curated list of AI/LLM inference tools, builds, and guides for AMD RDNA GPUs (RX 7000/9000, Strix Halo, R9700)
An Agent-Skill that helps users deploy different models, including asr, tts, llm on AMD ai max 395
Qwen3.8-Flash-Next (qwen4exp) local deployment on AMD AI Max 395 — self-built llama.cpp/ROCm, OpenAI-compatible API, LM Studio-style console
Source forked from official AMD FSR 4.0.2
Single-file Windows hardware monitor in Python — live CPU, RAM, per-GPU utilisation/VRAM/watts, disk I/O and PSU headroom. Built for multi-GPU AMD rig...
Reproducible LLM KV-cache tiered-storage benchmark on AMD MI308X: LMCache parallel-read patch, load clients, orchestration, measured results
Private, local-first financial research agents on AMD Radeon and ROCm.
Own inference stack on AMD gfx1100: Mojo GPU kernels, self-describing GGUF engine
A Context-Native Creative Operating System that transforms persistent context into production-ready creative artifacts through intelligent agent orche...
Perseus Vault x AMD Instinct — encrypted, local-first agent memory kept off the GPU. AMD Developer Hackathon Act II (Unicorn Track).
Decky Loader plugin for turning AMD x86 SteamOS handhelds into a more Steam Deck-like experience
Experimental ROCm 10 compatibility build for Qwen3.8-Flash-Next on 4x AMD R9700
DirectML Image-to-3D workflow for Windows AMD GPUs, targeting Radeon RX 5000/6000/7000. End-to-end validated on RX 6750 GRE.
This repository documents how to build and run LAMMPS with the native PKG_ML-MACE interface for MACE machine learning force field simulations on the L...