python Experiment.py -c train -r RUN_NAME -d ./datasets/mvtec -ds mvtec --mode rgb --dtd-path ./datasets/dtd/images/ --category zipper --fg-mask-path ./fg_masks/
training mvtec data error expect category bottle, but if i delete some bottle good pcs from mvtec/bottle/train/good folder and fg_masks/mvtec/bottle folder ,it appers same error.
And i found it wrong in eval period , it became nan!
predicted_noise, predicted_mean, predicted_mask = self.model(input, t)
what's the problem.
python Experiment.py -c train -r RUN_NAME -d ./datasets/mvtec -ds mvtec --mode rgb --dtd-path ./datasets/dtd/images/ --category zipper --fg-mask-path ./fg_masks/training mvtec data error expect category bottle, but if i delete some bottle good pcs from mvtec/bottle/train/good folder and fg_masks/mvtec/bottle folder ,it appers same error.
And i found it wrong in eval period , it became nan!
predicted_noise, predicted_mean, predicted_mask = self.model(input, t)what's the problem.