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

Create Coverage.java #1653

Merged
merged 3 commits into from
Aug 25, 2022
Merged

Create Coverage.java #1653

merged 3 commits into from
Aug 25, 2022

Conversation

gforman44
Copy link
Contributor

Description

For regression problems, you'd like to know how often you're over-estimating the target. Sometimes you want to overestimate, say, 75% of the time, e.g. to make sure you have enough milk in stock to satisfy your customers.

  • Note that this code could be generalized to reshape the predictions to match the labels. It re-uses the AbstractAccuracy class, which requires an axis number in its constructor, but this is ignored here, and generally for regression problems that don't use 1-hot encoding.

zachgk
zachgk previously requested changes May 17, 2022
Copy link
Contributor

@zachgk zachgk left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This looks really good. Can you add a test for it to https://github.com/deepjavalibrary/djl/blob/master/integration/src/main/java/ai/djl/integration/tests/training/EvaluatorTest.java to make sure that it works now and isn't broken by accident later

For regression problems, you'd like to know how often you're over-estimating the target.   Sometimes you want to overestimate, say, 75% of the time, e.g. to make sure you have enough milk in stock to satisfy your customers.
@KexinFeng KexinFeng dismissed zachgk’s stale review August 25, 2022 21:38

Test has been added

@KexinFeng KexinFeng merged commit f9455ac into deepjavalibrary:master Aug 25, 2022
patins1 pushed a commit to patins1/djl that referenced this pull request Aug 26, 2022
For regression problems, you'd like to know how often you're over-estimating the target.   Sometimes you want to overestimate, say, 75% of the time, e.g. to make sure you have enough milk in stock to satisfy your customers.

Co-authored-by: KexinFeng <fengx463@umn.edu>
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

Successfully merging this pull request may close these issues.

4 participants