Popular repositories Loading

1. kernel-anvil kernel-anvil Public

   Profile-guided GPU kernel optimizer for AMD/RDNA3. Auto-tunes llama.cpp MMVQ kernels per model shape. 2x decode speedup on 7900 XTX.

   Python 52 1

2. llama-cpp-turboquant llama-cpp-turboquant Public

   Forked from TheTom/llama-cpp-turboquant

   LLM inference in C/C++

   C++ 1

3. hamm-r hamm-r Public

   Headless Agent Mobile Management Relay -- Run Claude Code from your phone

   Kotlin

4. moirai moirai Public

   Three-model local coding daemon: planner + coder + reviewer-orchestrator on a single GPU. Swap-on-demand llama-server, RO loop with tool-call discipline, turboquant KV compression on 7900 XTX.

   Go