Intel® Distribution of OpenVINO™ Toolkit
Community support and discussions about the Intel® Distribution of OpenVINO™ toolkit, OpenCV, and all things computer vision-related on Intel® platforms.
5962 Discussions

How to create json and config file for custom model?

RAHUL
Beginner
386 Views

I have .pb model file, now how can I generate config and json file so that I can convert it to xml and bin file. I am using below command.

python3 mo_tf.py --input_model=model.pb --transformations_config *** --tensorflow_object_detection_api_pipeline_config **** --reverse_input_channels

Thanks in advance

0 Kudos
3 Replies
SuryaPSC_Intel
Employee
386 Views

Hi Rahul,

Please follow the steps for freezing the custom model and then Convert Tensorflow model to IR.

Best Regards,

Surya

RAHUL
Beginner
386 Views
Hi Surya, I am already following the link you mentioned. Actually example models can be successfully converted to xml using above command but problem is with custom model for that I don't have json and config file. Without these files I cannot optimize my tf object detection model.
SuryaPSC_Intel
Employee
386 Views

Hi Rahul,

OpenVINO provides .json files for some standard models which you may refer to generate your own .json file. You may find the .json files at the below directory in windows.

C:\Program Files (x86)\IntelSWTools\openvino_2020.1.033\deployment_tools\model_optimizer\extensions\front\tf 

Best Regards,

Surya

Reply