google / dopamine

Dopamine is a research framework for fast prototyping of reinforcement learning algorithms.

Home Page:https://github.com/google/dopamine

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

bundle_and_checkpoint bottleneck

mattgallivan opened this issue · comments

When subclassing DQNAgent, my code seems to run slowest in the bundle_and_checkpoint and unbundle functions. If I overload these functions to simply do nothing, does this affect the DQN agent significantly other than not saving and loading checkpoints?

Thanks!