Example with plot to see learning rate.

[kalee]

This is a question not a problem. The libraries are working well.

Question: If I want to see the progress in a plot, as sometime the learning seems to get really good, then can drop off to being really bad depending on the model. Is this there in the code. I don't want to create something new to a library that has a lot of thought built into it.

Silly question: If I want to have the example dqn.py learn how to do cartpole or lunar lander. Once this has been learned, is there a quick way to see this graphically. There is the render='human' that can be passed in. I didn't quite understand what I would do once the model was trained successfully to actually run it. The use case would be to train pong, then have the trained model show a game being played where the training model demonstrates how well it can now play. I'd find this interesting to see how a model performs visually at different learned rates.

Thank you, Ken