curiousily/dissertation

dissertation's Issues

eps_threshold verification before enter DQN

Hello Venelin Valkov,
I saw your post https://www.curiousily.com/posts/android-reinforcement-learning-environment/. It is a very helpful post. Thank you.
I just checked your code on github automate.py in (https://github.com/curiousily/dissertation/blob/master/code/automate.py)
I would like to understand in line 177 max_idx = vals[:len(actions)].max(0)[1][0]
Why the selection of len(actions).max(0)(1)(0) ?
what represent the last (0)?

Other doubt:

• The reward is the code coverage but when the app crashes it will be incremented?

Thank you