Documented exponential moving average

Removed comments

Updating vgg16

Updated inceptionV2

Added VGG16

Updated VGG16

Updated VGG16

Updated VGG16

Updated inceptionV1

Inceptionv2 batch norm

Updated batch norm

Added moving average

Added batch norm to InceptionV2 and InceptionV1

Applied moving averages

Changed the default value for the l2 regularizer

Fixed the loss collection

Fixed issue with the mode

pep8 formatting done by yapf -ri .

See merge request !42

Created mechanism that allows us to train only parts of the graph during the graph adaptation

See merge request !35

# Conflicts:
#   bob/learn/tensorflow/estimators/Logits.py
#   bob/learn/tensorflow/estimators/Siamese.py

Fixed the contrastive loss equation in the comments

Created mechanism that shuts down parts of the graph with respect to backpropagation

Fixed small bug

Reverted from non_trainable_variables to trainable_variables

Pylint the code

pylinting

Reorganized transfer learning mechanism

remove exectuable permission from files

See merge request !41

remove the models after the tests

See merge request !40

Use tf.train.get_or_create_global_step (introduced in 1.2) instead of the contrib one

See merge request !39

Add a script to call tf.estimator.train_and_evaluate

See merge request !37

Early stopping hook

See merge request !38

Use bob.extension to load config files

See merge request !36

biopredict: Handle number of parallel jobs correctly

See merge request !34

Changes to the biogenerator

See merge request !33

# Conflicts:
#   bob/learn/tensorflow/estimators/Logits.py

Organizing transfer learning

See merge request !32

Renamed the keyword argumente is_training_mode to mode and use the tf.estimator.Modekeys to switch between training/validation/prediction modes