codemaster
codemaster
look at the lines and issue: https://github.com/argman/EAST/blob/master/model.py#L79 https://github.com/argman/EAST/blob/master/eval.py#L159 https://github.com/argman/EAST/issues/210
Hello, I have the same question, and I find this line, may be the answer. if a>45, then predict a, else predict b: https://github.com/argman/EAST/blob/master/icdar.py#L371 
https://github.com/taki0112/AdaBound-Tensorflow
I also met this problem
L_g, y_true_cls, training_mask with the same shape: 64*1*128*128
At beginning , the loss from about 16 , gradually, about 200 epoch, the loss get about 0.6, then not lower any more
While with latest pytorch, use https://github.com/t-vi/warp-ctc/, there's some error, t-vi said, he will fix it some. While with latest pytorch, but with pytorch's own ctc-loss, use batch size 64, back...
@MiZhangWhuer I train on icdar2015. @xxlxx1 Can you print you error message?
somewhere should be changed, but I forgot it.