Mechanically replace Model.*_generator with Model.* in Python files

[huonw]

Tensorflow Keras's Model type has deprecated its ..._generator methods in Tensorflow 2.1.0.

https://www.tensorflow.org/api_docs/python/tf/keras/Model#fit_generator:

Warning: THIS FUNCTION IS DEPRECATED. It will be removed in a future version. Instructions for updating: Please use Model.fit, which supports generators.

https://github.com/tensorflow/tensorflow/releases/tag/v2.1.0-rc1:

Model.fit_generator, Model.evaluate_generator, and Model.predict_generator are deprecated endpoints. They are subsumed by Model.fit, Model.evaluate, and Model.predict which now support generators and Sequences.

This patch does an automatic replacement of the following, in *.py files:

• fit_generator → fit
• predict_generator → predict
• evaluate_generator → evaluate

It does not make any changes to ensemble.py/test_ensemble.py (#1065), because we define our own fit_generator (etc.) methods there, and thus manual work is required. Nor does it do any of the notebooks (#1066), in order to break the work up and reduce merge conflicts (notebooks are more likely to have merge conflicts, and they're more annoying to resolve).

There's a few places where manual changes might be good (e.g. mistakes in comments), but I'd prefer not to do them here: keeping this fully automated makes it easy to resolve any merge conflicts (revert the file to the other form, and then redo the automated replacement). As you review, feel free to do some manual changes as your own separate branch/PR.

See: #803

[codeclimate]

Code Climate has analyzed commit 82f9d7d and detected 0 issues on this pull request.

View more on Code Climate.

[kjun9]

🎉