Concise

- Complete re-write. Now moved completely to Keras from pure TensorFlow.

* fix: Update tensorflow function (tf.op_scope -> tf.name_scope, initialize_all_variables -> tf.global_variables_initializer)

* fix: When X_feat had 0 columns, loading its weights from file was failing.
* feat: When training the global model in ConciseCV, use the average number of epochs yielding the best validation-set accuracy.

* feat: added early_stop_patience argument

* same as 0.4.1 (pypi upload failed for 0.4.1)

* bugfix: multi-task learning