Iill-defined empirical covariance in build diverge model task

[holyketzer]

Sometimes build divergence model fails with error

Fitting the mixture model failed because some components have ill-defined empirical covariance (for instance caused by singleton or collapsed samples). Try to decrease the number of components, or increase reg_covar.

It's not clear why it happens

https://app.auger.ai/admin/cluster_tasks/280647

dataset: s3://auger-options-1sunr2/temp/options-a2ml/data_temp/options-2020-03-06_4208.csv.parquet_review_B26364B24FF94E7.parquet

sources: https://github.com/augerai/a2ml/blob/master/a2ml/api/stats/feature_divergence.py

cc @skatedplrn

[holyketzer]

For Vlad Khizanov