A little Helper for all your Impish needs, Excellent at math and coding for it's size. If you enable bad hehavior it will behave badly.

This model Is meant to give Android devices a functional AI with 70B style capabilities. It is not censored and has simple instruction base template. It was quantized using a custom importance matrix to GGUF format from nvidia/AceMath-1.5B-Instruct

Install llama.cpp through brew (works on Mac and Linux)

Invoke the llama.cpp server or the CLI.