About the extra actions in parallel_runner.py

[Mohan-Zhang-u]

In line 96 of parallel_runner.py, it uses a "While True" to collect pre_transition_data and post_transition_data. However, this will collect and update actions_chosen even when all the envrionments are terminated and none of them are going to be executed. So why bother perform the action_selection operations, when you would even remove them later on in the learners? Thanks for your time and reply!

[JianyuSu]

line 120 of parallel_runner will make sure that once all of your environments terminate, it stops to collect and update batch data. I don't know if I answered your question.