Facing "graph Contains A Cycle" Error While Using Tensorflow Object Detection Models

I am using SSD Inception V2 to generate The .pb file, and Inturn using Open Vinos Model Optimizer for Converting TensorFlow Object Detection API Models for better Inference results

Solution 1:

We will get this errror if you have used tensoflow version <1.6 .

Please refer to the below documentation page on the same error

Doc path:

computer_vision_sdk_2018./deployment_tools/documentation/docs/TensorFlowObjectDetectionSSD.html

In order to fix this issue, it is necessary to specify output nodes of the graph manually using the --output command line parameter.

Command to convert:

Example Model Optimizer Command-Line for TensorFlow* SSD The final command line to convert SSDs from the TensorFlow Object Detection API Zoo is:

./mo_tf.py --input_model="path_to_frozen.pb" --tensorflow_use_custom_operations_config extensions/front/tf/legacy_ssd_support.json --output="detection_boxes,detection_scores,num_detections"

Thanks, Rajeswari Ponnuru.