GitLab

* Create convert_to_tflite.py

Script to convert quantized deeplab models to tflite

* Update quantize.md

- Make it consistent with '/bin/bash' shebang
- Fix "Bad substitution" error in BASH_SOURCE[0]

* DeepLab: Cityscapes splits refactoring

* DeepLab: Process 'test_fine' split in Cityscapes

* convert_cityscapes.sh: Update PYTHONPATH

Fix error `ModuleNotFoundError: No module named 'cityscapesscripts'`

* Check the number of labels and images

* DeepLab: Fix generation of tfrecord for Cityscapes

To fix the following error:
File "/usr/local/lib/python3.6/dist-packages/absl/flags/_flagvalues.py", line 698, in get_value raise _exceptions.Error('Missing value for flag ' + arg) # pylint: disable=undefined-loop-variable absl.flags._exceptions.Error: Missing value for flag -v

Thanks amangu@ for fixing this script.

Co-authored-by: yukun <yukun@google.com>

Python 2 is deprecated. Tested by running the notebook in a Python 3 runtime.

Colab will soon update the default version of tensorflow to 2.1.0. In order for this notebook to continue to work, I'm adding a line magic that will ensure this notebook continues to use tensorflow 1.x and execute without errors.

* Make this file TensorFlow 2.0 compatible

* Update remove_gt_colormap.py

Thanks to issue raised at https://github.com/tensorflow/models/issues/7799

* Update for py3 and some internal changes

* \nInternal refactor\n

PiperOrigin-RevId: 279809827

* clean up

* deeplab quantize

* Fix bug in train.py

* Create quantize.md

* Update colab demo

* fix typo

* Internal changes

PiperOrigin-RevId: 237183552

* update readme

PiperOrigin-RevId: 237184584

* Fixed incorrect tensor

* Replaced xrange with range for python3 compatibility

* \nRefactor deeplab to use MonitoredTrainingSession\n

PiperOrigin-RevId: 234237190

* Update export_model.py

* Update nas_cell.py

* Update nas_network.py

* Update train.py

* Update deeplab_demo.ipynb

* Update nas_cell.py

Open source checkpoints for mobilenetv2_dm05_coco_voc_trainaug and mobilenetv2_dm05_coco_voc_trainval

Open-sourcing the checkpoint so that users could reproduce our PASCAL VOC 2012 validation set result when training on train_aug set.