facebook
CrossValidationPlot object
Summary:
CrossValidation is one of the most frequent plots used accross Ax. This change imports it to the ax.analysis module, simplifies its logic and provides a easy to use access point, which takes in the common experiment and model data types.
- "get_df()" exports a basic dataframe with the data used to generate the plot
- "get_fig()" exports the classic cross validation plots, with the same functionality as used today.
New Usage
from ax.analysis.cross_validation_plot import CrossValidationPlot
cv_plot = CrossValidationPlot(experiment=scheduler.experiment, model=model)
plot = cv_plot.get_fig()
Old Usage
from ax.plot.diagnostic import interact_cross_validation_plotly
from ax.modelbridge.cross_validation import cross_validate
cv_results = cross_validate(model=model)
plot = interact_cross_validation_plotly(cv_results)
Differential Revision: D54432757
Codecov Report
Attention: Patch coverage is 96.30996% with 10 lines in your changes are missing coverage. Please review.
Project coverage is 94.87%. Comparing base (
e8018a2) to head (dd06d1b).
:exclamation: Current head dd06d1b differs from pull request most recent head 3803c9a. Consider uploading reports for the commit 3803c9a to get more accurate results
Additional details and impacted files
@@ Coverage Diff @@
## main #2252 +/- ##
========================================
Coverage 94.86% 94.87%
========================================
Files 467 476 +9
Lines 46455 46726 +271
========================================
+ Hits 44068 44329 +261
- Misses 2387 2397 +10
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
