If several models with different random seed per as of date are inside a model group, is important to check the stability of the metric regarding the fluctuations caused by the random seed.
One easy way to solve this, is ignoring and just get the avg or another aggregation in the query before using audition.
If several models with different random seed per as of date are inside a model group, is important to check the stability of the metric regarding the fluctuations caused by the random seed.
One easy way to solve this, is ignoring and just get the
avg
or another aggregation in the query before using audition.