DYJNG

Same question ! Look forward to your reply...

+1， 把后处理 [def get_mini_boxes()](https://github.com/WenmuZhou/DBNet.pytorch/blob/master/post_processing/seg_detector_representer.py)函数里的box四点排序搬过来就可以了~ ``` def order_points_clockwise(pts): points = sorted(list(pts), key=lambda x: x[0]) index_1, index_2, index_3, index_4 = 0, 1, 2, 3 if points[1][1] > points[0][1]: index_1 = 0 index_4...

mark

Hello, have you achieved it?

> 已解决 您好，我也遇到了类似的问题，请问您是如何解决的呢？

Hi, this project is so awesome. I need this dataset for research. Could you share me the dataset link for me doing my research? Thank you! My email is [email protected]...

You can try increase the size of shared memory(shm). The size of shm can see by commond "df -h". Or you can set "num_works=0" in DataLoder, but it will make...

@skyler1253 @xieydd @Raneee @donghoonsagong @TheLittleBee @Wallace00V @jaybdub @crazybob23 @zbz-cool same problem, do you solve this problem?

@ceci3 @cryoco 感谢回复，其实我的本意是希望能够把量化后的paddle模型转化为独立的trt模型，脱离paddle框架，仅使用trt去做推理，但是目前我还没有想到实现的途径。我先尝试您们提供的建议测试下性能，再次感谢！