Correct the predict in AutoEncoder to consider non-vector layer outputs

[sylvaticus]

Change the loop order in cache/predict of the AutoEncoder to allow convolutional layers with non-array output in layers