'Segmentation fault' seen while using ESRGAN model for super resolution

Niranjan__Udipi · ‎12-12-2019

I converted the ESRGAN model available from https://github.com/WolframRhodium/Super-Resolution-Zoo/tree/master/ESRGAN using the following model converter command:

./mo_mxnet.py --input_model ~/Downloads/Zoom-Experiments/Openvino-SR/ESRGAN_4x-0000.params --output_dir /tmp/ --input_shape [1,3,64,64]

No issues, Openvino format model was created successfully as ESRGAN_4x-0000.xml and ESRGAN_4x-0000.bin files.

Then, I used the Openvino super resolution example with this new model as:

./super_resolution_demo -m /tmp/ESRGAN_4x-0000.xml -i ~/Pictures/image640x640.jpeg

I get the following error, what is wrong?

[ INFO ] InferenceEngine:
   API version ............ 2.1
   Build .................. custom_releases/2019/R3_ac8584cb714a697a12f1f30b7a3b78a5b9ac5e05
   Description ....... API
[ INFO ] Parsing input parameters
[ INFO ] Files were added: 1
[ INFO ]     /home/niranjan/Pictures/image640x640.jpeg
[ INFO ] Loading Inference Engine
[ INFO ] Device info:
   CPU
   MKLDNNPlugin version ......... 2.1
   Build ........... 32974

[ INFO ] Loading network files
[ INFO ] Preparing input blobs
Segmentation fault (core dumped)