Ludwig: Visualization API AssertionError

Created on 21 Apr 2020  路  3Comments  路  Source: ludwig-ai/ludwig

Describe the bug
I try to create the compare_performance chart with the visualization API and it throws an AssertionError.

To Reproduce
from ludwig.visualize import compare_performance

stats = [{'combined': {'accuracy': 0.0, 'loss': 21.042979235191247}, 'target_category': {'jaccard': 0.14934778770208643, 'loss': 21.042979235191247}}, {'combined': {'accuracy': 0.0, 'loss': 20.444052153835703}, 'target_category': {'jaccard': 0.1627610707413676, 'loss': 20.444052153835703}}]

model_names = ['Model1', 'Model2']

compare_performance( test_stats_per_model=stats, output_feature_name='target_category', model_names=model_names )

That's the Error:

image

It also doesn't change when I set a output_directory for the method.

Expected behavior
Show chart with the comparisation.

Environment (please complete the following information):

  • OS: Ubuntu
  • Version 19.10

    • Python version 3.7.5

    • Ludwig version 0.2.1

bug looking into it

All 3 comments

@donfour10 thank you for letting us know. That visualization was originally implemented for categorical output features, but yours is a set output feature. Let me fix the doce to work with any output feature type. Will keep you posted.

The most recent commit fixes this. please install from master and confirm it this soves your problem.
To install from master:
pip uninstall ludwig and thenpip install git+http://github.com/uber/ludwig.git

thanks for the quickly fix!

Was this page helpful?
0 / 5 - 0 ratings