Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[User privacy] Select the model utilized in the backend. #1

Open
ravindra-ut opened this issue Jun 29, 2023 · 5 comments
Open

[User privacy] Select the model utilized in the backend. #1

ravindra-ut opened this issue Jun 29, 2023 · 5 comments

Comments

@ravindra-ut
Copy link

Hi,

Gorilla-CLI fuses the capabilities of various Language Learning Models (LLMs) like Gorilla LLM, OpenAI's GPT-4, Claude v1, and others to present a user-friendly command-line interface.

Does it mean user queries are submitted to GPT4(OpenAI) or Claudev1(Anthropic) ?
Can I select or opt for Gorilla only inference ?
Can user be provided option with setting up Gorilla model on local setup ? Is there plan to release Gorilla model checkpoints ?

@ravindra-ut ravindra-ut changed the title Which model is being used in the backend [User privacy] ? [User privacy] Choose the model is being used in the backend Jun 29, 2023
@ravindra-ut ravindra-ut changed the title [User privacy] Choose the model is being used in the backend [User privacy] Select the model utilized in the backend. Jun 29, 2023
@ShishirPatil
Copy link
Member

Thanks for trying out Gorilla-cli @ravindra-ut . This is a great suggestion. I was wondering perhaps user can pass an argument --model and include that in the RESTful call? Do you want to maybe raise a PR for this - if you are interested - would love to have you join this project as a contributor! I can modify the back end to respect the model choice.

@ShishirPatil
Copy link
Member

Yup always! The models are evolving rapidly rite now. So, we will release the model checkpoints once it stabilizes to a certain extent.

@Noppapon
Copy link

Noppapon commented Sep 14, 2023

@ShishirPatil An issue that may arise when passing the --model argument is that some users may find it cumbersome to specify it with every command they run. I think it would be great to have a config.json file that saves user settings, and we can potentially remove USER_ID_FILE and UPDATE_CHECK_FILE as well. All the information could then be stored within the config.json file. The file will also be useful and make the code cleaner as the system grows and has more complex features that require individual user information to be saved.

With this approach, we would check the config.json if models are specified, and if so, pass it into the RESTFUL call. I can make a PR for this.

@ShishirPatil
Copy link
Member

Hey @Noppapon thanks for taking this on! Yeah, that is a good suggestion - we can have a config.json that can have all this information in one place. Maybe we can still allow the users to over-ride the config with the --model or -m flag, in those cases where they want to use a model one-off? Thoughts!
Otherwise, this sounds good - let me know when you raise the PR and I can review it. Thanks man!

@Noppapon
Copy link

Thank you, @ShishirPatil ! I've made several additions in the same PR to clean up the code related to argument passing and the UI, using the Click framework. Along with that, users can now execute essential commands like --help, --version., and some more, which I think would be useful. The PR is now ready for review.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants