Dropout Vector

[nelsoonc]

Hello @mvpcom, I want to ask about the dropout vector. In the code, the dropout vector seems to have 1.0 and 0.7 value when in the imitation-learning paper stated 50% dropout after fully connected layers and 20% dropout after convolutional layer. Or am I missing something here? Thank you

[mvpcom]

@nelsoonc Hi, I am suggesting using the official repo. This was only an initial attempt to reproduce their results before they publish the code.