gaia  by amd

Local LLM agent execution on Ryzen AI PCs

Created 1 year ago
867 stars

Top 41.4% on SourcePulse

GitHubView on GitHub
1 Expert Loves This Project
Project Summary

GAIA by AMD is an open-source solution for running generative AI applications locally on Windows PCs, targeting users who want to leverage their hardware for LLM-based tasks. It offers a hybrid hardware approach, utilizing AMD's NPU and iGPU for optimized performance, alongside a generic mode for broader compatibility.

How It Works

GAIA employs a hybrid acceleration strategy, combining the AMD Neural Processing Unit (NPU) and Integrated Graphics Processing Unit (iGPU) found in Ryzen AI PCs for efficient LLM execution. For broader compatibility, a generic mode uses Ollama as a backend. This approach aims to deliver fast and efficient AI processing without cloud dependencies, supporting various use cases from basic chat to RAG-enhanced applications.

Quick Start & Requirements

  • Install: Download and run gaia-windows-setup.exe. Supports silent installation with /S and mode selection (/MODE=HYBRID, /MODE=GENERIC).
  • Prerequisites:
    • Hybrid Mode: AMD Ryzen AI 9 HX 300 series or newer processor, Windows 11 (Home/Pro), minimum 16GB RAM (32GB recommended). Specific NPU (32.0.203.237/240) and iGPU (32.0.12010.8007 or 32.0.12033.1030) drivers are recommended.
    • Generic Mode: Any Windows PC meeting Ollama's requirements.
  • Setup Time: Approximately 5-10 minutes.
  • Links: GAIA Blog Article, FAQ, Developer Guide.

Highlighted Details

  • Hybrid acceleration leveraging AMD NPU and iGPU for optimized performance.
  • Provides both a GUI and CLI for interaction.
  • Extensible architecture for custom agents and use cases.
  • Supports multiple installation modes: Hybrid, NPU (coming soon), and Generic.

Maintenance & Community

  • Actively developed project. Contributions via Pull Requests are encouraged.
  • Contact information provided for questions and issues.

Licensing & Compatibility

  • MIT License. Permissive for commercial use and closed-source linking.

Limitations & Caveats

  • Hybrid mode is strictly limited to AMD Ryzen AI HX 300 series or newer processors.
  • No automatic uninstaller is currently available; manual removal is required.
  • The installer is not digitally signed, triggering Windows Security warnings.
  • NPU Driver 32.0.203.242 may cause issues; reversion to 32.0.203.240 is recommended.
  • Currently supports only Windows 11; macOS and Linux are not supported.
Health Check
Last Commit

1 day ago

Responsiveness

1 day

Pull Requests (30d)
28
Issues (30d)
18
Star History
73 stars in the last 30 days

Explore Similar Projects

Starred by Chip Huyen Chip Huyen(Author of "AI Engineering", "Designing Machine Learning Systems").

JittorLLMs by Jittor

0%
2k
Low-resource LLM inference library
Created 2 years ago
Updated 10 months ago
Starred by Andrej Karpathy Andrej Karpathy(Founder of Eureka Labs; Formerly at Tesla, OpenAI; Author of CS 231n), Gabriel Almeida Gabriel Almeida(Cofounder of Langflow), and
2 more.

torchchat by pytorch

0.1%
4k
PyTorch-native SDK for local LLM inference across diverse platforms
Created 1 year ago
Updated 4 months ago
Feedback? Help us improve.