GGML
Also known as: ggml
GGML is the lightweight tensor library that powers llama.cpp. It’s also the name of the original (pre-GGUF) file format.
- A small, dependency-free math core.
- Optimized for CPU and GPU inference.
- The substrate under most local LLM tooling.
You'll see it: if a local tool mentions "GGML", it's the underlying tensor engine — the muscle under llama.cpp and its GGUF files.