Intel® Distribution of OpenVINO™ Toolkit
Community assistance about the Intel® Distribution of OpenVINO™ toolkit, OpenCV, and all aspects of computer vision-related on Intel® platforms.
6392 Discussions

Model Optimizer works but demo application failed with ssd

Chao__Stone
Beginner
348 Views

Hi! all:

I was able to transfer a tensorflow ssd model to IR files but fail to run the demo program with Openvino 2020.2.117.

I use the following to generate IR files successfully.

python mo_tf.py --input_model C:\temp\openvino_test\mymodel_ssd\frozen_inference_graph.pb --output_dir C:\temp\openvino_test\mymodel_ssd\FP16 --data_type FP16 --batch 1 --tensorflow_object_detection_api_pipeline_config "C:\temp\openvino_test\mymodel_ssd\pipeline.config" --tensorflow_use_custom_operations_config "C:\Program Files (x86)\IntelSWTools\openvino_2020.2.117\deployment_tools\model_optimizer\extensions\front\tf\ssd_v2_support.json"

and run the demo program with this:

object_detection_demo_ssd_async -m "C:\temp\openvino_test\mymodel_ssd\FP16\frozen_inference_graph.xml"  -d CPU -i cam

However, it crashes on line 179  ( for (const auto op : ngraphFunction->get_ops()) { ).
Exception: 0x00007FFB3912FC5E (ngraph.dll) (於 object_detection_demo_ssd_async.exe) exception: 0xC0000005: access viloate 0xFFFFFFFFFFFFFFFF.

Since I use tensorflow 1.14, according to this document (https://docs.openvinotoolkit.org/latest/_docs_MO_DG_prepare_model_convert_model_tf_specific_Convert_Object_Detection_API_Models.html ),
I tried to generate IR files with ssd_support_api_v1.14.json.
But it cause errors as the following.


[ ERROR ] Failed to match nodes from custom replacement description with id 'ObjectDetectionAPISSDPostprocessorReplacement': It means model and custom replacement description are incompatible. Try to correct custom replacement description according to documentation with respect to model node names [ ERROR ] Cannot infer shapes or values for node "Postprocessor/ToFloat". [ ERROR ] 0 [ ERROR ] [ ERROR ] It can happen due to bug in custom shape infer function <function Cast.infer at 0x000001A0730B7D08>. [ ERROR ] Or because the node inputs have incorrect values/shapes. [ ERROR ] Or because input shapes are incorrect (embedded to the model or passed via --input_shape). [ ERROR ] Run Model Optimizer with --log_level=DEBUG for more information. [ ANALYSIS INFO ] Your model looks like TensorFlow Object Detection API Model. Check if all parameters are specified: --tensorflow_use_custom_operations_config --tensorflow_object_detection_api_pipeline_config --input_shape (optional) --reverse_input_channels (if you convert a model to use with the Inference Engine sample applications) Detailed information about conversion of this model can be found at https://docs.openvinotoolkit.org/latest/_docs_MO_DG_prepare_model_convert_model_tf_specific_Convert_Object_Detection_API_Models.html [ ERROR ] Exception occurred during running replacer "REPLACEMENT_ID" (<class 'extensions.middle.PartialInfer.PartialInfer'>): Stopped shape/value propagation at "Postprocessor/ToFloat" node. For more information please refer to Model Optimizer FAQ (https://docs.openvinotoolkit.org/latest/_docs_MO_DG_prepare_model_Model_Optimizer_FAQ.html), question #38.

The customized model should be ok since I can transfer it and run demo application with openvino 2019.2.275 on another machine.

Any idea to fix this?

Thanks a lot!

0 Kudos
1 Reply
JAIVIN_J_Intel
Employee
348 Views

Hi Stone,

Are you able to run the object_detection_demo_ssd_async demo(from OpenVINO 2020.2) with any OpenVINO pre-trained models, for eg: person-detection-retail-0013

If yes, please share the TF model and necessary files, so that we could reproduce the issue.

If required, you can share the model privately using PM.

Regards,

Jaivin

 

0 Kudos
Reply