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

bug: Bart speedup only 1.6x #327

Open
2 tasks done
sinking-point opened this issue Jul 13, 2023 · 0 comments
Open
2 tasks done

bug: Bart speedup only 1.6x #327

sinking-point opened this issue Jul 13, 2023 · 0 comments

Comments

@sinking-point
Copy link

Description

I was immensely impressed by the 7.3x speedup demonstrated in the t5 tutorial (though I was only able to reproduce 4.0x on my machine, still pretty good).

However, I can only speed up Bart by 1.6x in the same way.

I have determined this is not due to the difference in model size, as t5-base (which is larger than bart-base) is sped up 3.8x.

Steps to reproduce

Use the t5 notebook but replace:

't5-small' with 'facebook/bart-base'

optimize_model(model.encoder) with optimize_model(model.model.encoder)

optimize_model(model.decoder) with optimize_model(model.model.decoder)

Expected Behavior

Speedup comparable to that seen in T5.

Actual Behavior

A comparatively small 1.6x speedup.

Your environment

The docker container from the README.

Self-service

  • I would be willing to help fix this bug myself.

Code of Conduct

  • I agree to follow this project's Code of Conduct
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

1 participant