I am trying to use MO on a re-trained model (I followed this tutorial:https://github.com/EdjeElectronics/TensorFlow-Object-Detection-API-Tutorial-Train-Multiple-Objects-W...)
There was an error using my model with raspberry pi3 + ncs2 after converting from faster_rcnn_inception_v2_coco to IR
"terminate called after throwing an instance of 'InferenceEngine::details::InferenceEngineException'
what(): Dims and format are inconsistent."
I using Openvino 2019 R2 for pc and Openvino 4.1.2 for raspberry.
What should I do?
Hi Thien Tran,
Thank you posting your issue.
Could you please answer the following:
As the example you are trying to run is not an officially tested example, try to run our Object Detection Faster R-CNN C++ Demo and let us if it works.
I am having the same issue. I'm trying to use different versions of OpenVINO, and different ways to convert my model to IR, however, I have the same answer: Dims and format are inconsistent. I am using Raspberry Pi 3 with other .xml and .bin files, and it is working fine.
Thank you very much in advance for your help.
I am having the same issue.
- I tried different ways to optimized my model to use it with Movidus in the Raspberry Pi 3. In general, this is the code used: sudo python3 ./mo_tf.py --input_model [my_path]/frozen_inference_graph.pb --tensorflow_object_detection_api_pipeline_config [my_path]/object_detection/training/pipeline.config --output=detection_boxes,detection_scores,num_detections --output_dir [output_path] --tensorflow_use_custom_operations_config extensions/front/tf/faster_rcnn_support_api_v1.10.json
- I am using OpenCV 4.2.0-openvino
- I was using OpenVino 2020 but it was not possible to use the samples, so I changed to OpenVino 2019.
- Yes, I can run different samples made by Caffe and Tensorflow. It is working corretly. Not for mine.
Thank you very much in advance.