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

small_model_filename cannot be assigned during fine-tuning #149

Open
INF800 opened this issue Aug 25, 2023 · 1 comment · May be fixed by #325
Open

small_model_filename cannot be assigned during fine-tuning #149

INF800 opened this issue Aug 25, 2023 · 1 comment · May be fixed by #325
Labels
bug Something isn't working help wanted Extra attention is needed

Comments

@INF800
Copy link
Contributor

INF800 commented Aug 25, 2023

The below code

small_model_filename = kwargs.get("small_model_filename", None)

tries to extract small_model_filename from kwargs which is passed down from client.complete(...) -> optimizer.complete(...) -> optimizer.complete_delay_train(...)

I think the main purpose of this is to make sure we can save finetuned model to any location via llm_vm.client, but it will cause error because the same kwargs with small_model_filename will be passed to chat_gpt model:

best_completion = self.call_big(prompt, **kwargs)

Resulting in

openai.error.InvalidRequestError: Unrecognized request argument supplied: small_model_filename
@INF800
Copy link
Contributor Author

INF800 commented Aug 25, 2023

The apt name would be small_model_filepath instead of small_model_filename.

@mmirman mmirman added bug Something isn't working $50 help wanted Extra attention is needed labels Aug 28, 2023
@kshitiz305 kshitiz305 linked a pull request Oct 12, 2023 that will close this issue
@mmirman mmirman removed the $50 label Mar 12, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug Something isn't working help wanted Extra attention is needed
Projects
None yet
Development

Successfully merging a pull request may close this issue.

2 participants