Unable to Convert Custom YOLOv4* Model to Intermediate Representation (IR) Format

• Converted custom YOLOv4* model into Intermediate Representation (IR) with the command:

  python mo_tf.py --input_model frozen_darknet_yolov4_model.pb --tensorflow_use_custom_operations_config yolo_v4_tiny.json --batch 1 --reverse_input_channels

• Received error: Cannot infer shapes or values for node "detector/yolo-v4-tiny/Tile/YoloRegion".

Add scale value and input shape parameters in command line for the Model Optimizer to convert YOLOv4 model into IR:

python mo_tf.py --input_model=<Installed_directory>\Intel\openvino_2021.3.394\deployment_tools\open_model_zoo\tools\downloader\public\yolo-v4-tf\yolo-v4.pb --output_dir="<path/to/save/>" --data_type=FP32 --model_name=yolo-v4-tf --input_shape=[1,608,608,3] --input=image_input --scale_values=image_input[255] --reverse_input_channels

The conversion of the YOLOv4 model is now available in the OpenVINO™ toolkit 2021.4.