I'm trying to mvNCCompile my custom network.
When I run mvNCCompile the following error is shown:
TypeError: unsupported operand type(s) for /: "NoneType' and 'int'
I think this error occurs because NCS currently supports batch size = 1 only.
If I was using Caffe, it looks easy to feed the deploy network with single batch.
But in Tensorflow, I set the input placeholder to have shape like [None, w, h, c] to train, and the network graph is saved right after training.
So is there any method to train the network with large batch and save the single-batch model ?
