fix error analysis showing float values with many zeros for some classification task nodes #1991
Conversation
This file contains 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
imatiach-msft
requested review from
gaugup,
RubyZ10,
vinuthakaranth,
tongyu-microsoft,
xuke444 and
hawestra
as code owners
gaugup
reviewed
Mar 2, 2023
tongyu-microsoft
approved these changes
Mar 2, 2023
RubyZ10
approved these changes
Mar 2, 2023
The validation code is for both regression and classification tasks |
…sification task nodes
imatiach-msft
force-pushed
the
ilmat/fix-error-float
branch
from
797ff36 to
a5a6950
Compare
3 tasks
3 tasks
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Outdated suggestions cannot be applied.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.
Description
I noticed in several customer meetings that sometimes the nodes in error analysis would have values with many zeros. I was actually able to reproduce this locally for one of our notebooks (census classification scenario). It seems this occurs only for classification tasks, as we don't use precision in front-end then, and we cast the error variable passed from backend to frontend cast as float. The fix is to conditionally cast the error variable to frontend or backend depending on whether a regression metric was used.
Before fix:
After fix:
Checklist