search

nanase1025 / M-SAM

👏Mask-Enhanced Segment Anything Model for Tumor Lesion Semantic Segmentation[MICCAI 2024]
Apache License 2.0
9 stars 0 forks source link

can not apply transform on several samples of LiTS #2

Open Devil-Ideal opened 2 months ago

Devil-Ideal commented 2 months ago

Hi ! When I try to run train_msam.py on LiTS datasst, an error accurs (RuntimeError: Trying to resize storage that is not resizable). I find that the dataset class will report this error as below: if self.transform: try: subject = self.transform(subject) except: print(self.image_paths[index]) Did you just delete these data or there are some methods to handle this?

Devil-Ideal commented 2 months ago

There also some samples without tumor region in LiTS dataset, should I remove these samples ? Can you provide the details of training test set partitioning, since different partition results in dfferent results.