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

evaluate.py compute_losses? #70

Open
MathildaHwang opened this issue May 8, 2021 · 0 comments
Open

evaluate.py compute_losses? #70

MathildaHwang opened this issue May 8, 2021 · 0 comments

Comments

@MathildaHwang
Copy link

MathildaHwang commented May 8, 2021

Hi! Thank u for this wonderful work!
In evaluate.py-->compute_losses(self, dataloader), it seems that only one batch of data is used for evaluation.
But when it comes to loss computation,

losses = {k: sum(v) / len(dataloader) for k, v in storer.items()}

it uses len(dataloader) to average the loss. Should that be the length of element v?
I wonder if I misunderstand the above computation.
Any help will be appreciated!

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