Popular repositories Loading

1. llama.cpp-turboquant-hip llama.cpp-turboquant-hip Public

   TurboQuant KV cache compression for llama.cpp — HIP/ROCm port for AMD RDNA3 (gfx1100)

   C++ 17 6

2. hermes-skills hermes-skills Public

   Reusable skills and dependency graph system for Hermes Agent — bitwarden secrets, email composition, autoresearch loop

   Python 6

3. turboquant-hip turboquant-hip Public

   TurboQuant KV cache compression for AMD GPUs (HIP/ROCm). First working HIP port. 4.9x compression, zero accuracy loss.

   C++ 2

4. agentsearch agentsearch Public

   CLI tool to index and search AI agent sessions across Hermes, Moltis, Nanobot, and markdown notes. BM25 via tantivy.

   Rust 1

5. triattention-ggml triattention-ggml Public

   Frequency-based KV cache pruning for llama.cpp — 25% cache reduction, improved PPL at long context. GPU compaction kernel for HIP/ROCm.

   Python 1

6. moltis moltis Public

   Forked from dr34ming/moltis

   A Rust-native claw you can trust. One binary — sandboxed, secure, auditable. Voice, memory, MCP tools, and multi-channel access built-in.

   Rust