CPP

Transcribe.cpp
752 points | 158 comments

6.4x faster than llama.cpp, 3.9x faster than MLX
4 points | 2 comments

Hedos – Local models in one native Mac app (llama.cpp, MLX, diffusers)
3 points | 0 comments

Show HN: Run a 120B-parameter MoE on Android mid-range phone CPU-only llama.cpp
1 points | 0 comments

Show HN: Low-latency local LLM runner via OpenJDK Panama FFM (Java 22)
38 points | 10 comments

Show HN: AffineUI a 2 file C++ GPU/HTML5 UI that can replace Electron
6 points | 0 comments

Show HN: Trellis2.c – Local 3D generation with Vulkan and CUDA
2 points | 0 comments