v1.2.0

Flattening

Notes

• Removed unnecessary code
• Added more comments to better describe the code

V1.1.0

Device Control

Notes

• Added device control to PPO.py
• Added a test file to show an example of how testing should be done.
• Set Mountain Car Continuous to the example environment ¹.

Footnotes

1. In the rollout function, you'll notice that when a new episode is started we set our observation to be the first element only. This is because the reset function of the environment returns a tuple of our observation and null so we have to set it to only be the first element otherwise we cannot use our data with the neural network. ↩

v1.0.0

Initial release

Notes

This is the first and probably last release. There is a zip file with just the files so you can copy them into a folder without having to rename the folder after cloning using git.