Skip to content

Inference graph export  #15

@JL2718

Description

@JL2718

Hi. I'm trying to get an inference graph from the baseline models, but the tensorflow and/or object detection toolkit versions are not compatible:

python ~/code/models/research/object_detection/export_inference_graph.py --input_type image_tensor --pipeline_config_path ./pipleine.config --trained_checkpoint_prefix ./multires_aug-ckpt --output_directory ./export Traceback (most recent call last): File "/home/jl/code/models/research/object_detection/export_inference_graph.py", line 156, in <module> tf.app.run() File "/home/jl/anaconda3/envs/xview/lib/python3.7/site-packages/tensorflow/python/platform/app.py", line 125, in run _sys.exit(main(argv)) File "/home/jl/code/models/research/object_detection/export_inference_graph.py", line 152, in main write_inference_graph=FLAGS.write_inference_graph) File "/home/jl/code/models/research/object_detection/exporter.py", line 449, in export_inference_graph is_training=False) File "/home/jl/code/models/research/object_detection/builders/model_builder.py", line 121, in build return _build_ssd_model(model_config.ssd, is_training, add_summaries) File "/home/jl/code/models/research/object_detection/builders/model_builder.py", line 244, in _build_ssd_model is_training=is_training) File "/home/jl/code/models/research/object_detection/builders/model_builder.py", line 220, in _build_ssd_feature_extractor return feature_extractor_class(**kwargs) File "/home/jl/code/models/research/object_detection/models/ssd_inception_v2_feature_extractor.py", line 75, in __init__ raise ValueError('SSD Inception V2 feature extractor always uses' ValueError: SSD Inception V2 feature extractor always usesscope returned by conv_hyperparams_fnfor both the base feature extractor and the additional layers added since there is no arg_scope defined for the base feature extractor.

Is it possible that you could use your version to export the inference graph (inference_graph.pbtxt) with the above command, or let me know the tensorflow version and HEAD version of your object detection toolbox?

Thanks!

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions