Let's use the multi feature flag that Alex implemented to turn off the model analyzer and reward validation when multi=true
957 will show where the reward function validation happens.
and ModelAnalyzerApiClient is what is used to communicate with the model analyzer.
This PR can just shortcut the checks and return values that let these steps pass.
Original comment:
We can just disable model analyzer and reward validation for test environment only. The objective is to just learn more about multi-agent simulations since it might become important later.
Let's use the
multi
feature flag that Alex implemented to turn off the model analyzer and reward validation whenmulti=true
957 will show where the reward function validation happens.
and
ModelAnalyzerApiClient
is what is used to communicate with the model analyzer.This PR can just shortcut the checks and return values that let these steps pass.
Original comment:
We can just disable model analyzer and reward validation for test environment only. The objective is to just learn more about multi-agent simulations since it might become important later.
Originally posted by @ejunprung in https://github.com/SkymindIO/pathmind-webapp/issues/919#issuecomment-587959198