Adding fid calculation at the end of epochs (at train.py)#1192
Open
OmriKaduri wants to merge 1 commit intojunyanz:masterfrom
Open
Adding fid calculation at the end of epochs (at train.py)#1192OmriKaduri wants to merge 1 commit intojunyanz:masterfrom
OmriKaduri wants to merge 1 commit intojunyanz:masterfrom
Conversation
|
While we still don't have the implemented metrics I prepared simple way to get FID after training using test data real images on domain B and generated images from domain B --> input domain is simply ignored |
Owner
|
We cannot accept this PR at this time, due to two reasons:
It might be great to first add the FID evaluation code for pre-trained checkpoints (after training is done). We will see if we could do that in the near future. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Hi,
Thanks for a great repo for experimenting with CycleGAN and pix2pix!
I wanted to compute FID scores during training (between the validation set and the generated images on validation), so I added this as an option at train.py.
Let me know if any changes needed. I believe it might be useful for others.