backtime92

@ThisIsIsaac I've changed the mistake. Em... I think you can train for MLT. Maybe IC15 train has some mistakes. I will refine it when I am no busy.

@ThisIsIsaac In the trainic15data.py, code in line 114 should be freeze=False. Sorry, I do not know what's the meaning of PR.

@nbhupendra My method merge the boxes with one line, it can help you?

@ThisIsIsaac thanks for you advice I will merge it

@hanish3464 Yes,thanks for your advice, I noticed the subtle differences，but i do not know how much distance should be reduced.And the main reason is afiinity map which impact the performance.

@yts19871111 https://github.com/backtime92/CRAFT-Reimplementation/issues/6#issuecomment-530716570 the reason is that boxes' width or height smaller than 1 so that can not do the warpaffinity.

@goodluckcwl No. In the [issues #34](https://github.com/clovaai/CRAFT-pytorch/issues/34#issuecomment-524525473) the author said the perforamnce difference was not significant with Resnet or PVAnet. But I plan to train with EfficientNet. I will release the...

@YoungminBaek Thank you for your information. Could you share your lr change details and batch size?

@YoungminBaek How many high loss pixel in the negative pixels are selected? Or selects all high loss pixel in the negative pixels (region where the region score is smaller than...

@YoungminBaek I select all positive pixels and triple negative high loss pixels to train the model, but I found that loss are very high about 4000 when I train for...