I encountered problem using my tiny-YOLO model on Neural Stick 2. I get WinUsb_ReadPipe: System err 31, and the Neural Stick 2 cannot be pinged (see screenshot).
Strangely enough, when I use full YOLO model, everything is working fine.
I am using Windows 10. The models were converted according to the YOLO tutorial on Intel website (https://software.intel.com/en-us/articles/OpenVINO-Using-TensorFlow). Please see the minimal example in attachment.
Any advice, please?
Solved, I made some mistake during the conversion from Darknet to Tensorflow. When I send the model to the GPU insted to the NS2, I got an error message that there is a problem with shapes of some tensors. It's still slightly strange that the NS2 just crashed without any reasonable error message, but w/e.