Learning rates do not match paper

[bsaund]

Thanks for the work and for releasing the code! We hope to use this on our robot.
I also appreciate how well all of the learning parameters are documented.

In the paper, Section "3.4 Training" lists the following learning rates for the Adam optimizer:
discriminator: 0.0001
generator: 0.0005

The code provided uses:
discriminator: 0.00005
generator: 0.0001

Is this a typo (both switching dis and gen, and missing a '0')? Did you change the learning rates after writing the paper?

[Yang7879]

@bsaund Great thanks for reporting the mismatch. The learning rate in the release code is correct. We will update it in next paper version.