error when training lpo for bounding box performance

[vuptran]

When I tried to train LPO on the COCO dataset using the command:
python train_lpo.py ../models/lpo_COCO_0.02.dat -b -t -f0 0.02
I get the following error:
Traceback (most recent call last):
File "train_lpo.py", line 121, in
boxes = [proposals.Proposals(s,np.eye(np.max(s)+1).astype(bool)).toBoxes() for s in segmentations]
File "/home/anaconda/lib/python2.7/site-packages/numpy/core/fromnumeric.py", line 2135, in amax
out=out, keepdims=keepdims)
File "/home/anaconda/lib/python2.7/site-packages/numpy/core/_methods.py", line 26, in _amax
return umr_maximum(a, axis, None, out, keepdims)
ValueError: operands could not be broadcast together with shapes (10,2) (5,2)

Without the -b argument, I was able to successfully train LPO on the COCO dataset. I'm using Mac OS X 10.10.3.