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.

Sign up for GitHub

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

Jump to bottom

Improve default max length logic for text generation #364

Closed

davidmezzetti opened this issue Oct 17, 2022 · 1 comment

Assignees

Milestone

Member

davidmezzetti commented Oct 17, 2022

Make the following improvements in deriving the max length field for text generation.

Use model.config.max_length if it's not the default value
Else use tokenizer.model_max_length if available

This logic should be added

davidmezzetti added this to the v5.1.0 milestone

davidmezzetti self-assigned this

davidmezzetti closed this as completed in

cd43c47

Member Author

davidmezzetti commented Oct 17, 2022

This logic should be revisited when transformers==5.0.0 is released.

How to derive model_max_length/max_length/max_new_tokens is an open discussion item and likely will lead to txtai changes with transformers V5.

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