Upstream Google autoencoder models to CompressAI

[mmuckley]

At the moment we have several model implementations that are already implemented in CompressAI (e.g., Scale Hyperprior, Mean-Scale Hyperprior). At this point CompressAI has pretty good adoption, so we should be able to remove these from our repository and upstream the dependency.

By default CompressAI doesn't handle reflective image padding for users, so if desired we could include wrappers like that in PR #185 to handle this for users unfamiliar with the functionality of these models.

[desi-ivanova]

Happy to help with that at some point.

Padding and image standardization (to 0-1 range) -- both of which are supposed to be handled by the user (described in the second to last paragraph in compressai documentation here) -- can be handled by wrappers instead, as you suggest.

[mmuckley]

Addressed in PR #196.