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

Better gemini model names support #37

Merged
merged 2 commits into from
Sep 12, 2024

Conversation

megatron6000
Copy link
Collaborator

@megatron6000 megatron6000 commented Sep 12, 2024

Feature Description

Currently model names like gemini-1.5-pro-preview-0514 raise errors. It's because the library that is used to calculate the tokens offline lack support for names like that. To fix that, only the "core" of the name can be used to init the tokenizer (i.e. gemini-1.5-pro ).

Changed

  • if initializing tokenizer with a model name provided in config doesn't work, initialization only with the "core" of the model name is tried, and only if this won't work the error is raised.

@megatron6000 megatron6000 merged commit 4e3c893 into main Sep 12, 2024
4 checks passed
@megatron6000 megatron6000 deleted the better-gemini-model-names-support branch September 12, 2024 11:44
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

Successfully merging this pull request may close these issues.

2 participants