deepmask-pytorch
deepmask-pytorch copied to clipboard
A confusion about your code
I just can not understand the following code:
**
**
I wonder why you set d1 < maxDim <= d2? Is it for selecting the positive samples?
Why the item satisfied this constrain is the positive sample?
Looking forward to your reply, thx.
Another confusion is that the on line 109 of coco_loader.py, why is the ‘ss’ has a negative symbol of (s*scale)? This is incompatible with line 181 in the jitterbox method.