the pre-train model can't achieve 80% F1-score. about east HOT 6 CLOSED

I met some problem when I test the pre-train model. Could I send you e-mail to ask question?

from east.

@facetohard of course, what is the problem with you, first i think you should use python3 not python2.

from east.

Do you make any changes?

from east.

@argman I download the real icdar2015 dataset from http://rrc.cvc.uab.es/?ch=4&com=downloads, but these are some warning when I train the model with icdar2015_data_set, can I ignore this warn?
File "/home/yxzh/mount_2/project/ocr/east_yxzh/icdar.py", line 533, in generate_rbox
if point_dist_to_line(p1, new_p2, p0) > point_dist_to_line(p1, new_p2, p3):
File "/home/yxzh/mount_2/project/ocr/east_yxzh/icdar.py", line 248, in point_dist_to_line
return np.linalg.norm(np.cross(p2 - p1, p1 - p3)) / np.linalg.norm(p2 - p1)
TypeError: unsupported operand type(s) for -: 'NoneType' and 'float'

from east.

yes

from east.

I met the same question ,can anybody solve it now?

from east.