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.

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

Confused by multiply n to both clf_logits and clf_losses? #39

Open
Vimos opened this issue Sep 4, 2018 · 3 comments
Open

Confused by multiply n to both clf_logits and clf_losses? #39

Vimos opened this issue Sep 4, 2018 · 3 comments

Comments

@Vimos
Copy link

Vimos commented Sep 4, 2018

I am confused by the code below.

pytorch-openai-transformer-lm/train.py

Line 52 in eafc28a

clf_logits *= n

pytorch-openai-transformer-lm/train.py

Line 54 in eafc28a

clf_losses *= n

Is this due to any normalization? Thanks!
@thomwolf
Copy link
Member

thomwolf commented Sep 4, 2018

Yes, I don't understand that myself either. It was in the original TensorFlow code so I kept it as is. I have removed it in a fork of this code that I use for another project and it doesn't really change the results. I may try to dig further one day.

@pranoy-k
Copy link

Hi Thomas,

Do you have any insights now for this issue.

@thomwolf
Copy link
Member

I don't

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
3 participants
@Vimos @thomwolf @pranoy-k