Issue about export_inference_graph.py

[zxwxz]

Hi @rky0930 , I trying to using this framework but not fully support correctly I can training by your config But it fail on export inference graph Traceback (most recent call last): File "object_detection/export_inference_graph.py", line 119, in <module> tf.app.run() File "/home/brian.lin/anaconda3/envs/work_py36/lib/python3.6/site-packages/tensorflow/python/platform/app.py", line 126, in run _sys.exit(main(argv)) File "object_detection/export_inference_graph.py", line 115, in main FLAGS.output_directory, input_shape) File "/home/brian.lin/YOLOV2_OD/yolo_v2/research/object_detection/exporter.py", line 427, in export_inference_graph input_shape, optimize_graph, output_collection_name) File "/home/brian.lin/YOLOV2_OD/yolo_v2/research/object_detection/exporter.py", line 391, in _export_inference_graph initializer_nodes='') File "/home/brian.lin/YOLOV2_OD/yolo_v2/research/object_detection/exporter.py", line 72, in freeze_graph_with_def_protos optimize_tensor_layout=True) ValueError: Protocol message RewriterConfig has no "optimize_tensor_layout" field.

Can you take a look? I will try to compare with other model's config first

After check the code, I temporarily disable 'optimize_graph' for avoid this issue

[rky0930]

Hi, Thanks for using my code. I think this issue is relevant with the issue. https://github.com/tensorflow/models/issues/2861 it's depends on what version of TF use. What TF version do you use?

[minda163]

change ‘’yolo_v2/research/object_detection/exporter.py", line 72 optimize_tensor_layout=True to layout_optimizer=1