Skip to content

[FR] unload llama model #21

Description

@ImagineerNL

First of all, a wonderful tool, it works great in getting what i want in a good structure.

I'm using a local llama.cpp docker server on the same computer. For some reason youve kept the 'local llama' to be 'localhost' only, but that doesn't get accepted. (also, you cant enter an api key in that field)
So i am using the 'custom api' and that works great.

It would be great if you could add an 'unload model' call to the same server (and apikey). For llama cpp that is http://localhost:8080/models/unload (notice it is without the /v1) as that saves people from manually needing to login to llama and unload the model (or set the sleep timeout extremely low)

implementation wise, it could be an input field where people can submit the 'unload' call, so it is flexible for different toolings.
And then in the generation tab a button to unload, or even a checkbox 'auto unload model after generation'

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions