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

Validation freq #46

Open
RAMIhamza opened this issue Oct 14, 2021 · 3 comments
Open

Validation freq #46

RAMIhamza opened this issue Oct 14, 2021 · 3 comments

Comments

@RAMIhamza
Copy link

Hey,

When I run two trainings (strong baseline) with the same config file, except for val_freq which is the frequency of validation (2 and 5 in my case). I surprisingly get two different curves of the training cross entropy loss.
I believe that even though the model is in eval mode, it still changes some of its settings. Am I wrong ?

Thanks in advance.

@RAMIhamza
Copy link
Author

Please find bellow some specifications of my setup.
I'm working with the following versions:

  • Python 3.6.9
  • Torch '1.7.1+cu110'

Also the two plots of the cross entropy loss that correspond to the two trainings (strong baseline) with different val frequencies.

image
Grey line : val_freq = 5
Orange line: val_freq = 2

@chongyangwang-song
Copy link

@RAMIhamza Hi, I study this code recently, could we have some Communication?

@RAMIhamza
Copy link
Author

Yeah sure !
Here is my mailing address : [email protected]

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

2 participants