Lists (3)
Starred repositories
Forge: Swarm Agents That Turn Slow PyTorch Into Fast CUDA/Triton Kernels
Minimal TPU implementation with 8x8 systolic array and PyTorch integration
Pure Triton kernels for Qwen3.5-27B inference on NVIDIA B200
Run a 1-billion parameter LLM on a $10 board with 256MB RAM
Autoresearch for GPU kernels. Give it any PyTorch model, go to sleep, wake up to optimized Triton kernels.
Development stack for AI-assisted multi-repo work. Persistent memory, quality gates, deployment automation. CLI + Claude Code plugin.
🔷 Zig implementation of TOON (Token-Oriented Object Notation) — Compact, human-readable JSON encoding for LLM prompts with 30-60% token reduction. Spec-compliant v3.0 implementation.
Zig implementation of TOON (Token-Oriented Object Notation).
Efficient Universal Perception Encoder: a single on-device vision encoder with versatile representations that match or exceed specialized experts across multiple task domains.
The highest-scoring AI memory system ever benchmarked. And it's free.
Understand any codebase instantly. System intelligence for codebases, built for humans and AI.
Using LLMs for iteratively exploring the solution search space at scale.
Customize Claude Code's system prompts, create custom toolsets, input pattern highlighters, themes/thinking verbs/spinners, customize input box & user message styling, support AGENTS.md, unlock pri…
All parts of Claude Code's system prompt, 24 builtin tool descriptions, sub agent prompts (Plan/Explore/Task), utility prompts (CLAUDE.md, compact, statusline, magic docs, WebFetch, Bash cmd, secur…
Model parallel transformers in JAX and Haiku
A modern model graph visualizer and debugger
A gallery that showcases on-device ML/GenAI use cases and allows people to try and use models locally.
Cross-platform, customizable ML solutions for live and streaming media.
LiteRT, successor to TensorFlow Lite. is Google's On-device framework for high-performance ML & GenAI deployment on edge platforms, via efficient conversion, runtime, and optimization
The open-source memory operating system for AI agents. Persistent memory, semantic search, loop detection, agent messaging, crash recovery, and real-time observability.
Asynchronous Zoned Unified Runtime Environment for Agentic LLMs
PokeClaw (PocketClaw) — first on-device AI that controls your Android phone. Gemma 4, no cloud, no API key. Poke is short for Pocket.
LakeSail's computation framework with a mission to unify batch processing, stream processing, and compute-intensive AI workloads.
[MLsys2026]: RAG on Everything with LEANN. Enjoy 97% storage savings while running a fast, accurate, and 100% private RAG application on your personal device.
A new chunking strategy developed by ZeroEntropy for general semantic chunking using Llama-70B.
AI-powered penetration testing assistant using local LLM on linux (Parrot OS)
JiuwenClaw is an intelligent AI Agent built on openJiuwen. It extends the powerful capabilities of large language models directly to your fingertips through various communication apps you use daily.