Error when training the FRL model by: $ python main.py

[bailingbird]

Error is shown as follows:
tensorflow.python.framework.errors_impl.InvalidArgumentError: You must feed a value for placeholder tensor 'beta_q_input' with dtype float and shape [?,16]

I have no changes to the codes and errors happen when I run the main.py of gridworld.

[xkw1]

I have the same problem. Have you solved it?

[xkw1]

I know the problem. Because the default value of "--preset_lambda" in the code given by the author is False, the corresponding "--train_mode" should be changed from 'frl_lambda' to 'frl_separate'.

[zxwblublu]

Thank you! This answer really solves my problem. By the way, to other people's convenience, the "--train_mode" to be revised is in line71 in main.py