python - warning during py-faster-rcnn training on custom datasets -

while training py-faster-rcnn on custom dataset following instructions @ https://github.com/deboc/py-faster-rcnn/blob/master/help/readme.md

i encountered errors

attributeerror: 'numpy.ndarray' object has no attribute 'toarray' in py-faster-rcnn 

which managed bypass editing https://github.com/rbgirshick/py-faster-rcnn/blob/master/lib/roi_data_layer/roidb.py

gt_overlaps = roidb[i]['gt_overlaps'] gt_overlaps = sp.sparse.csr_matrix(gt_overlaps).toarray() 

however, during training process, received warning twice

runtimewarning: invalid value encountered in log targets_dw = np.log(gt_widths / ex_widths) 

in file https://github.com/rbgirshick/py-faster-rcnn/blob/master/lib/fast_rcnn/bbox_transform.py

are results going affected ?

do need different ?

maybe u should try modify “lib/datasets/pascal_voc.py”

in function “_load_pascal_annotation(,)” right 1 should be:

x1 = float(bbox.find('xmin').text)                                                                  y1 = float(bbox.find('ymin').text)                                       x2 = float(bbox.find('xmax').text)                                                                  y2 = float(bbox.find('ymax').text)  

the reason in own data, x1 or y1 maybe equals 1, if minus 1,then number negative, caused error