Channel recovery

[TheLemonPig]

Channel Recovery training:

1. Adding a recover_channels parameter to the autoencoder training
2. Adding a recover_channels parameter to the gan training
3. Train AE and GAN with this same parameter adjustment

[TheLemonPig]

Is there some way to check the Autoencoder's configuration while you are inside the GAN training? The only reason to add any mention of channel recovery in the GAN training is just to make sure that the user knows that the AE they are using was trained to recover channels (i.e. to make sure they choose the right AE for their GAN training)

[chadcwilliams]

I am delaying the PR for two reasons.

1. The dev branch is currently EEG-GAN v2.0 as will be deployed, so at this point it should not be modified. I will work on deploying it soon so that the dev branch is again open for changes.

2. You mentioned in slack that these changes are not yet tested. I believe it would be best to ensure that this works on your branch before we merge into the code. This would include making sure that all past functionality still works (i.e., your changes did not effect the package when not doing channel recovery).

[TheLemonPig]

Hi @chadcwilliams sounds good. I completely understand. I wanted to create this as a draft PR so that you guys could see the work as it was being done. I wouldn't want to make it a proper PR until I can confirm it works. I think that is a great idea to also make sure that all previous functionality is not broken by the added features.