Pinned Loading

1. TurboQuant TurboQuant Public

   TurboQuant KV Cache Compression for llama.cpp — 5.2x memory reduction with near-lossless quality | Implementation of Google DeepMind's TurboQuant (ICLR 2026)

   C++ 3 2

2. amesianx-proxy amesianx-proxy Public

   General-purpose intercepting proxy for binary protocol inspection (SSV, AMF)

   Python

3. softwire softwire Public

   An instruction assembler that emulates a vertex shader with MMX code.

   C++ 1