mirror of
https://github.com/NousResearch/hermes-agent.git
synced 2026-04-29 07:21:37 +08:00
Comprehensive guide covering: - llama.cpp and MLX (omlx) setup on Apple Silicon - Model selection and memory optimization (quantized KV cache) - Real benchmarks on M5 Max comparing both backends - Hermes connection instructions Cherry-picked from PR #2590.