convolutional auto encoder train and test functions and scripts

[tyiannak]

No description provided.

[rodo-K]

Regarding "convolutional_autoencoder" branch, I added 3 files:
bin/basic_ae_train.py lib/ae_training.py models/ae_cnn.py

When I'm trying to run bin/basic_ae_train.py ( as the bin/basic_training.py runs) I'm getting wrong dimensions from the decoder part as you can see below:

Using a target size (torch.Size([16, 1, 128, 201])) that is different to the input size (torch.Size([16, 1, 129, 201])). This will likely lead to incorrect results due to broadcasting. Please ensure they have the same size.

Please let me know what I should check and If the initial try of the autoencoder is ok besides this error.

[rodo-K]

after calculating the right padding, stride etc. Got output 201 x 128. Somehow still not getting the result I want from the loss function. Can't figure out why.