mgail
mgail copied to clipboard
Model-Based Generative Adversarial Imitation Learning
Hello Caspi! After reading your paper MGAIL , I'm very inspired! But I'm little confused about the output of Discriminator: why there be 'out_dim=2' in Discriminator, shouldn't it be a...
Hi, I have a question regarding this line of code. https://github.com/itaicaspi/mgail/blob/b3b91aa5e0bd47923f726a27522f45146721940d/mgail.py#L109 It seems you are using a fixed std gaussian policy? I am wondering if I am getting it correctly.
How many expert trajectories have you considered while creating the file _hopper_er.bin_ ?