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

ImageDataGenerator does not standartize validation_data #1950

Closed
eyaler opened this issue Mar 11, 2016 · 4 comments
Closed

ImageDataGenerator does not standartize validation_data #1950

eyaler opened this issue Mar 11, 2016 · 4 comments

Comments

@eyaler
Copy link

eyaler commented Mar 11, 2016

ImageDataGenerator should apply standartize() on validation_data parameter or at least the docs should mention the the validation data is assumed to already be standartized

@fchollet
Copy link
Collaborator

There seems to be some confusion between ImageDataGenerator and fit_generator here. What do you mean?

@eyaler
Copy link
Author

eyaler commented Mar 12, 2016

e.g. in the cifar10_cnn.py example, if you add featurewise normalization, in ImageDataGenerator, the fit_generator does not apply this normalization to validation_data=X_test

@asmith26
Copy link
Contributor

asmith26 commented Apr 25, 2016

Regarding the doc, I came across this also - it looks like the ImageDataGenerator doc is not being automatically generated, and hence the actual signature is not being pulled and updating the doc (with the default of featurewise_std_normalization=False).

I was about to add the ImageDataGenerator class to be automatically updated, however it looks like that class may be soon re-written (see #2446), so I'll hold off for now.

@stale stale bot added the stale label May 23, 2017
@stale stale bot closed this as completed Jun 22, 2017
@eyaler
Copy link
Author

eyaler commented Jun 23, 2017

please reopen. this is still an issue. i missed the stale tag.

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

3 participants