[RL-baseline] Model v5, experiment #2

[ziritrion]

Action set #1 was chosen for this experiment:
[0.0, 0.0, 0.0], # no action
[0.0, 0.8, 0.0], # throttle
[0.0, 0.0, 0.6], # break
[-0.9, 0.0, 0.0], # left
[-0.5, 0.0, 0.0], # left
[-0.2, 0.0, 0.0], # left
[0.9, 0.0, 0.0], # right
[0.5, 0.0, 0.0], # right
[0.2, 0.0, 0.0], # right

Entropy and Running Reward dropped to very low levels between the 8k and the 20k episode mark, but seemed to start a growing trend right before the 20k episode mark. It's likely that the running reward could grow to positive values if the model is trained further. The max RR value is 282 right before the 7k episode mark, with a final RR of -45.

Tensorboard captures below:

Sample video below. The car never turns.
https://user-images.githubusercontent.com/1465235/113405881-d6a82800-93aa-11eb-8a8a-4e32dd27b275.mp4