search

uni-medical / SAM-Med3D

SAM-Med3D: An Efficient General-purpose Promptable Segmentation Model for 3D Volumetric Medical Image
Apache License 2.0
427 stars 56 forks source link

请问为什么碰到了零除错误 #36

Closed 111678900 closed 4 months ago

111678900 commented 5 months ago

Traceback (most recent call last): File "/content/SAM-Med3D/train.py", line 507, in main() File "/content/SAM-Med3D/train.py", line 466, in main trainer.train() File "/content/SAM-Med3D/train.py", line 361, in train epoch_loss, epoch_iou, epoch_dice, pred_list = self.train_epoch(epoch, num_clicks) File "/content/SAM-Med3D/train.py", line 335, in train_epoch epoch_loss /= step ZeroDivisionError: float division by zero请问这个到底为什么呀

111678900 commented 5 months ago

而且请问怎么测试呀

blueyo0 commented 5 months ago

一般就是 step 为 0 所以除以0就出错了，建议debug查一下数据数量和路径。

测试的话目前只有validation.py 的 benchmarking 功能，教程在readme里，还没有在无标签数据直接test的功能。

对于自己的数据或者无标签数据的inference 参见Hint PR #22的说明，目前没写这个内容，后面会提供的。这个项目我只是业余在维护，这段时间一直有更高优先级的事挤在前面，我还发烧了，如有需要请稍微多等等吧 😿。