Comprehensive documentation for the IPFS Accelerate Python framework - a complete solution for hardware-accelerated machine learning inference with IPFS network-based distribution.
- 🚀 Getting Started - Complete beginner's guide (5 minutes to first inference!)
- 📖 Installation & Setup - Detailed installation instructions
- 📚 Usage Guide - Learn how to use all framework features
- 🔧 API Reference - Complete API documentation with examples
- 🏗️ Architecture Overview - System design and components
- ❓ FAQ - Frequently asked questions and troubleshooting
- 📋 Changelog - Version history and release notes
- 🔐 Security Policy - Security reporting and best practices
- 🤝 Contributing Guide - How to contribute to the project
- 📄 License - AGPLv3+ license details
- 🗂️ Project Docs - production project status, summaries, dashboard, router, and migration docs
- 🧠 MCP++ Docs - conformance, cutover, and unification records
- ⚙️ Hardware Optimization - Maximize performance across different hardware
- 🌐 IPFS Integration - Leverage distributed inference and content addressing
- 🌍 WebNN/WebGPU Integration - Browser-based acceleration
- 🔗 P2P Architecture - P2P workflow scheduling and distributed computing
- 🧪 Testing Guide - Comprehensive testing framework and best practices
- 📊 Performance Tuning - Advanced optimization techniques
IPFS Accelerate Python is a comprehensive, enterprise-grade framework that combines:
- ✨ Hardware Acceleration: Support for CPU, CUDA, ROCm, OpenVINO, Apple MPS, WebNN, and WebGPU
- 🌐 IPFS Integration: Distributed model storage, caching, and peer-to-peer inference
- 🌍 Browser Support: Client-side acceleration using WebNN and WebGPU
- 🤖 300+ Models: Compatible with HuggingFace Transformers and custom models
- 🔒 Enterprise Security: Zero-trust architecture with compliance validation
- ⚡ High Performance: Optimized inference pipelines with intelligent caching
- 🚀 Cross-Platform: Works on Linux, macOS, and Windows
| Feature | Benefit |
|---|---|
| Multi-Hardware Support | Run on any device - from servers to browsers |
| Distributed Architecture | Scale horizontally with P2P networking |
| Zero Configuration | Sensible defaults, works out of the box |
| Production Ready | Battle-tested, comprehensive monitoring |
| Open Source | AGPLv3+ license, community-driven |
- Installation & Setup - Complete installation guide with hardware setup
- Usage Guide - Basic to advanced usage patterns with examples
- Examples - Practical examples and demos
- API Reference - Complete API documentation with all methods and parameters
- Architecture Overview - System design, components, and data flow
- Testing Guide - Testing framework, benchmarks, and quality assurance
- Hardware Optimization - Platform-specific optimization strategies
- IPFS Integration - Distributed inference and content addressing
- P2P Architecture - P2P workflow scheduling and distributed computing
- WebNN/WebGPU Integration - Browser-based acceleration
- GitHub Guides - GitHub Actions, autoscaling, authentication, P2P cache
- Docker Guides - Container deployment, caching, security
- P2P Guides - Distributed computing, libp2p, workflow scheduling
- Deployment Guides - Production deployment, cross-platform
- Project Documentation - permanent location for project status, summary, and migration records
- MCP++ Records - cutover evidence, conformance, and migration backlog
- CPU Optimization: x86/x64, ARM with SIMD acceleration
- NVIDIA CUDA: GPU acceleration with TensorRT optimization
- AMD ROCm: AMD GPU support with HIP/ROCm
- Intel OpenVINO: CPU and Intel GPU optimization
- Apple Silicon: Metal Performance Shaders (MPS) for M1/M2/M3
- WebNN/WebGPU: Browser-based hardware acceleration
- Qualcomm: Mobile and edge device acceleration
- Content Addressing: Cryptographically secure model storage
- Distributed Inference: Peer-to-peer model sharing and processing
- Intelligent Caching: Multi-level caching with LRU eviction
- Provider Discovery: Automatic network peer discovery and selection
- Fault Tolerance: Robust error handling and fallback mechanisms
- Text Models: BERT, GPT, T5, RoBERTa, DistilBERT, ALBERT, etc.
- Vision Models: ViT, ResNet, EfficientNet, CLIP, DETR, etc.
- Audio Models: Whisper, Wav2Vec2, WavLM, etc.
- Multimodal: CLIP, BLIP, LLaVA, etc.
- Custom Models: Support for custom model architectures
- Cross-Browser: Chrome, Firefox, Edge, Safari support
- WebNN API: Native neural network acceleration
- WebGPU: High-performance GPU compute in browsers
- Precision Control: fp16, fp32, mixed precision support
- Real-time Performance: Optimized for interactive applications
- 📖 Getting Started Guide - Complete beginner's tutorial
- ❓ FAQ - Frequently asked questions and quick answers
- 📚 Full Documentation Index - Comprehensive guide listing
- Use the Table of Contents in each document for quick navigation
- Look for 🔗 Cross-references between related sections
- Check 💡 Tips and Examples throughout the documentation
- Reference
⚠️ Troubleshooting sections when needed
- GitHub Issues: Report bugs and request features
- Discussions: Community questions and sharing
- Examples: Browse the examples directory for inspiration
- Contributing Guide - Detailed contribution guidelines
- Security Policy - Security reporting and best practices
- Code of Conduct - Community guidelines
- Development Setup - Follow the Testing Guide
All active, maintained documentation is organized in this directory:
- Core Docs: Installation, Usage, API, Architecture, Testing
- Guides: Topic-specific guides (GitHub, Docker, P2P, Deployment)
- Architecture: System architecture and design docs
- Project: project execution history, dashboard workstreams, router summaries, SDK-utilization records, and migration guides
- MCP++: MCP++ cutover, conformance, and server-unification records
- Archive: Historical session summaries and implementation reports
- Development History: Major milestones and phase completions
- Exports: HTML, PDF, and other non-markdown exports
A comprehensive audit was completed in January 2026:
- Audit Report: Complete findings and recommendations
- 200+ files reviewed, duplicates removed, links fixed
- Archive organized and documented
This documentation was comprehensively updated to reflect the current state of the IPFS Accelerate Python framework, including recent additions such as:
- P2P Workflow Scheduler: Distributed task execution with merkle clocks and fibonacci heaps
- MCP Server: Model Context Protocol server with 14+ tools
- CLI Endpoint Adapters: Direct integration with Claude, OpenAI, Gemini, VSCode CLIs
- Enhanced Inference: Multi-backend routing (local, distributed, API, CLI modes)
- GitHub Integration: P2P cache, autoscaler, workflow discovery
All examples, APIs, and features have been verified and updated for accuracy.
Last Updated: January 2026
Last Audit: January 31, 2026
Framework Version: 0.0.45+
Documentation Coverage: Comprehensive (Core + Recent Features)
Start with the Installation Guide to begin using IPFS Accelerate Python! 🚀