legraphista

/

Llama-3.2-3B-Instruct-IMat-GGUF

like 1

Few people do this quantization. It is very quick on ARM64 CPU because it benefits from "repack" as Q4_0 does.
Good for running small LLMs on Android phone.