MobiLlama icon indicating copy to clipboard operation
MobiLlama copied to clipboard

Android App load local model

Open AnswerZhao opened this issue 1 year ago • 3 comments

Could you modify the home activity to load the local model instead? The network speed for downloading the model is quite slow. Thank you for sharing this amazing AI project and Android app.

AnswerZhao avatar Feb 29 '24 10:02 AnswerZhao

Hi @AnswerZhao,

Thank you for interest in our work.

Please refer to llama.cpp (https://github.com/ggerganov/llama.cpp) and llama_cpp_dart (https://github.com/netdur/llama_cpp_dart) for loading local model.

Thanks,

OmkarThawakar avatar Mar 01 '24 17:03 OmkarThawakar

You could likely get faster inference using the native android APIs to use NPU but would be a undertaking for sure

nonetrix avatar Mar 02 '24 23:03 nonetrix

@OmkarThawakar Thank you for the references. I will give it a try.

AnswerZhao avatar Mar 08 '24 05:03 AnswerZhao