AI Repository Expansion and Updates | Generated by AI
Done. Here’s the summary:
Added 37 repos to repos.json (62 → 99 total). All 99 repos are cloned and ready for ww projects update.
New repos by category:
ai-models (+6): Qwen2-Audio, Qwen2.5-Math, llama, gemma, transformers, whisper.cpp
ai-training (+6): ColossalAI, flash-attention, LlamaFactory, LLMs-from-scratch, minbpe, mamba
ai-inference (+5): vllm, ollama, mlx, ggml, mlx-examples
ai-apps (+5): LibreChat, litellm, anything-llm, Whisper, facefusion
ai-agents (+6): OpenHands, autogen, markitdown, browser-use, opencode, gemini-cli
dev-tools (+3): warp, iclaw, mcp
infra-network (+6): mihomo, chroma, qdrant, localGPT, quivr, llm-app
Key picks:
- vllm (★81k) — high-throughput inference, essential for serving
- ollama (★173k) — local model runner, most popular in its class
- transformers (★161k) — HuggingFace model framework, fundamental
- LlamaFactory (★72k) — unified fine-tuning for 100+ LLMs
- browser-use (★96k) — web automation for AI agents
- OpenHands (★75k) — AI-driven software development
- litellm (★49k) — unified LLM API proxy
- markitdown (★134k) — file-to-Markdown conversion
- mihomo — your proxy core, already in ~/projects
- flash-attention (★24k) — fast attention implementation