Integrate permissively licensed open_llama model

[isaiahsherfick]

https://github.com/openlm-research/open_llama is a fully permissively licensed recreation of Facebook's original LLaMA base model. We should experiment with using this model so that this utility could potentially be usable in a production setting.

I will see if I am able to put together a pull request myself for this, but I am unfamiliar with this codebase and am unsure how easy that will be and make promises about neither the timeliness nor quality of said hypothetical PR.

I am also unsure if this change alone will make the utility production-ready, so please correct me if that is not the case and I will happily amend or remove this issue. Thank you!

[ashokrs]

@isaiahsherfick Are you able to succeed in running the 13-b model in localGPT? what is the machine spec? what is the memory size of the GPU needed?

[isaiahsherfick]

haven't messed with it, been busy. might get around to it eventually but it's on the back burner right now.

On Thu, Jun 8, 2023, 10:03 AM Ashok RS @.***> wrote:

@isaiahsherfick https://github.com/isaiahsherfick Are you able to succeed in running the 13-b model in localGPT?

— Reply to this email directly, view it on GitHub https://github.com/PromtEngineer/localGPT/issues/71#issuecomment-1582760817, or unsubscribe https://github.com/notifications/unsubscribe-auth/AI3QOGC3RNKG6HIGBTQNWDLXKHSTPANCNFSM6AAAAAAYZNG5WI . You are receiving this because you were mentioned.Message ID: @.***>