ResNet Fixes

[Simsso]

This PR fixes several issues which we ran into when testing out model in our last working group meeting (#56). They were partly listed in this comment.

Changes

• Decrease RAM to 4GB
• Decrease #CPU cores to 2
• Disable shuffling for validation data
• Rename batch_size to physical_batch_size
• Make the replacement conditional, depending on is_training
• Fix iterations bug, which led to very large training epochs which were in fact virtual_batch_size_factor many epochs.
• Fix learning_rate to 1 (learning adjustment with alpha, beta, and gamma).
• Extend logging significantly

Validation

Started a couple of validation runs for testing purposes.

[Simsso]

Concerning the memory requirements, I had to upscale for one of the training runs because Python ran out of RAM. It looks like 12 GB should suffice.