GitLab

* Added export for SavedModel format.

* Fixed some pylint errors.

* research/attention_ocr: Minor changes to make it compatible with python 3.

* research/attention_ocr: Script to create a smaller test file.

Before the fix the demo_inference.py used batch_norm and it did the
normalization of input image implicitly. If at inference time the
batch_norm was disabled the inference produced incorrect results.
This fix does the proper input image normalization and disables the batch_norm
at inference time.

1. Update README.md
2. argmax, use axis instead of deminsion
3. use tf.profiler.profile instead of model_analyzer.print_model_analysis

Changes:
- A working version
- Make the predicted_text to be a part of the model.

The spatial attention described in the paper was ported from the internal
implementation and is disabled by default.