论文代码最后结果

Hanqer / deep-hough-transform

Jittor and Pytorch code for paper "Deep Hough Transform for Semantic Line Detection" (ECCV 2020, PAMI 2021)

344 stars 71 forks source link

论文代码最后结果 #51

Closed zhizhz closed 2 years ago

zhizhz commented 2 years ago

老师您好，我最近跑了一遍您原先的代码发现评价指标如下所示： Mean P: 0.63580007390983 Mean R: 0.7869525443110351 Mean F: 0.7033466363096472 F@0.95: 0.4367617316136169 而这并不能达到您公布的pth文件中所跑出来的 Mean P: 0.6842166184936662 Mean R: 0.766064417762531 Mean F: 0.7228304458943596 F@0.95: 0.46145579498809 请问这样的波动是正常的吗？我的实验环境是2080ti PyTorch 1.7.0 ubuntu18.04 Cuda 11.0

zhizhz commented 2 years ago

其中数据集都是NKL进行的验证

Hanqer commented 2 years ago

论文中的结果是多次实验的平均值，一般不会偏差这么大，我的实验环境是TITAN XP， PyTorch1.2， cuda10.2

zhizhz commented 2 years ago

好的，谢谢老师的解答