[RL-baseline] Model v2, experiment #2

[ziritrion]

Third of the new experiments with the new action sets (5k episodes) and Learning Rate 1e-3-

The model managed to get a max Running Reward of ~364. Last episode managed 254.

Action set #2 copied below:
[0.0, 0.8, 0.0], # throttle
[0.0, 0.0, 0.6], # break
[-0.8, 0.0, 0.0], # left
[0.8, 0.0, 0.0], # right

Results are copied below:

Sample video below (somehow it managed to get an extraordinarily good result).
https://user-images.githubusercontent.com/1465235/111882632-d8d7b300-89b6-11eb-97bc-1114db39d397.mp4

[ziritrion]

Updated with 20k episodes. Running reward is now 240 with an achieved max of 565.

Sample video below. In this sample, the car skids out of the circuit and attempts to go back, but starts going backwards.

openaigym.video.0.187057.video000000.mp4