Add 3d autoencoderkl tutorial (#78)

* Add 3d autoencoder tutorial

* Add perceptual loss and autocast #78

* [WIP] Add discriminator loss

Note: there is one assertion error when training the model

* Train autoencoderkl 3D for 30 epochs (#78)

* Remove changes made to the perceptual loss (#78)

* Improve the image transformation and add zero_grad (#78)

Add zero_grad to the generator optimizer

* Use MONAI's activation layer and update import of the PatchDiscriminator (#78)

* Run model training for more epochs (#78)

* Re-run all cells (#78)

* Address comments (#78)

* Increase network capacity (#78)

Author: JessyD