I used a small caffe model(16MB)for object detection, when model input size is 450 * 450, every thing is ok; but when I change input size to 800 * 450, the compiler report "not enough resource error".
I 've successfully run two 450 * 450 model parranel(load two graph) on one compute stick, it means that ncs' memory can handle this data. So I wonder if there were a max of flops at one time which causes "not enough resource error".
Waiting for developers' response, Thanks a lot!
