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

Fix a couple SameDiff training issues #253

Merged
merged 5 commits into from Sep 11, 2019
Merged

Fix a couple SameDiff training issues #253

merged 5 commits into from Sep 11, 2019

Conversation

@rnett
Copy link

rnett commented Sep 11, 2019

Fix an issue with SameDiff training where an NPE would happen if a just a history listener with an evaluation was used (wasn't added to activeListeners properly, loss gradient wasn't calculated but was used for loss curve).

Also fix an issue where validation didn't specify outputs properly, causing errors whenever validation was used.

Added tests for both, as both really should have been caught by tests instead of by writing examples.

Ryan Nett added 4 commits Sep 10, 2019
Ryan Nett
Signed-off-by: Ryan Nett <rnett@skymind.io>
Ryan Nett
Signed-off-by: Ryan Nett <rnett@skymind.io>
Ryan Nett
Signed-off-by: Ryan Nett <rnett@skymind.io>
Ryan Nett
tests
Signed-off-by: Ryan Nett <rnett@skymind.io>
@rnett rnett requested a review from AlexDBlack Sep 11, 2019
@rnett rnett self-assigned this Sep 11, 2019
Ryan Nett
Signed-off-by: Ryan Nett <rnett@skymind.io>
@rnett

This comment has been minimized.

Copy link
Author

rnett commented Sep 11, 2019

Also adds single batch (DataSet/MultiDataSet) methods for SameDiff.output(). Unrelated to the rest of the changes.

Copy link

AlexDBlack left a comment

LGTM, thanks 👍

@rnett rnett merged commit 8a05ec2 into master Sep 11, 2019
@rnett rnett deleted the rn_sd_training_fix branch Sep 11, 2019
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
2 participants
You can’t perform that action at this time.