Tensorflow Object Detection API / ImportError: cannot import name 'keypoint_box_coder_pb2'

Question

I followed instructions (https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/installation.md) to install the Tensorflow Object Detection API. I use Anaconda on Windows 10 and all of the steps in the instructions seemed to complete OK.

But when trying to run python object_detection/builders/model_builder_test.py to test the installation, I get: ImportError: cannot import name 'keypoint_box_coder_pb2'

What am I missing?

(tensorflow) C:\Users\User\Documents\GitHub\models\research>python object_detection/builders/model_builder_test.py Traceback (most recent call last): File "object_detection/builders/model_builder_test.py", line 21, in from object_detection.builders import model_builder File "C:\Users\User\New\Anaconda3\envs\tensorflow\lib\site-packages\object_detection-0.1-py3.5.egg\object_detection\builders\model_builder.py", line 18, in from object_detection.builders import box_coder_builder File "C:\Users\User\New\Anaconda3\envs\tensorflow\lib\site-packages\object_detection-0.1-py3.5.egg\object_detection\builders\box_coder_builder.py", line 21, in from object_detection.protos import box_coder_pb2 File "C:\Users\User\New\Anaconda3\envs\tensorflow\lib\site-packages\object_detection-0.1-py3.5.egg\object_detection\protos\box_coder_pb2.py", line 17, in ImportError: cannot import name 'keypoint_box_coder_pb2'

Answer 1

I got it to work. What I did was:

I ran:

protoc --python_out=. .\object_detection\protos\anchor_generator.proto .\object_detection\protos\argmax_matcher.proto .\object_detection\protos\bipartite_matcher.proto .\object_detection\protos\box_coder.proto .\object_detection\protos\box_predictor.proto .\object_detection\protos\eval.proto .\object_detection\protos\faster_rcnn.proto .\object_detection\protos\faster_rcnn_box_coder.proto .\object_detection\protos\grid_anchor_generator.proto .\object_detection\protos\hyperparams.proto .\object_detection\protos\image_resizer.proto .\object_detection\protos\input_reader.proto .\object_detection\protos\keypoint_box_coder.proto .\object_detection\protos\losses.proto .\object_detection\protos\matcher.proto .\object_detection\protos\mean_stddev_box_coder.proto .\object_detection\protos\model.proto .\object_detection\protos\optimizer.proto .\object_detection\protos\pipeline.proto .\object_detection\protos\post_processing.proto .\object_detection\protos\preprocessor.proto .\object_detection\protos\region_similarity_calculator.proto .\object_detection\protos\square_box_coder.proto .\object_detection\protos\ssd.proto .\object_detection\protos\ssd_anchor_generator.proto .\object_detection\protos\string_int_label_map.proto .\object_detection\protos\train.proto

What was missing was: .\object_detection\protos\keypoint_box_coder.proto

Then ran:

python setup.py build
python setup.py install

Then changed: nets to research.slim.net in faster_rcnn_inception_resnet_v2_feature_extractor.py and faster_rcnn_resnet_v1_feature_extractor.py (in the models folder)

Ran:

python object_detection/builders/model_builder_test.py

Got:

Ran 11 tests in 0.195s

OK

Answer 2

It looks like you have not run

export PYTHONPATH=$PYTHONPATH:`pwd`:`pwd`/slim     # (from tensorflow/models/research folder) 

in the same terminal before launching the test, have you ? If you've chosen the .bashrc option, you probably need to reload your bshrc filewith . ~/.bashrc