marrlab / med-al-ssl Goto Github PK

There are various clustering methods from K-means to K-medoids. Is there any rationale of applying K-medoids algo? Thank you!

Active learning sampling error:
Traceback (most recent call last): File "train.py", line 347, in <module> main(args=arguments) File "train.py", line 187, in main kwargs, current_labeled) File "/home/zongwei/liangyu/code/Med-AL-SSL/code/utils.py", line 471, in perform_sampling samples_indices) File "/home/zongwei/liangyu/code/Med-AL-SSL/code/utils.py", line 144, in postprocess_indices samples_indices = samples_indices[samples_indices < len(unlabeled_indices)] TypeError: '<' not supported between instances of 'list' and 'int'

I have sufficient memory while not many CPU cores on my server, therefore IO can be the bottleneck of the training trials.

I noticed that you have set pin_memory=False in PyTorch DataLoader, and I didn't see any change of run time from toggling it.

Since this experiment is quite IO heavy, have you tried any speeding up method?

There is possibly a bug at dataset_cls = self.datasets[self.args.dataset](root=self.args.root, add_labeled=self.args.add_labeled, advanced_transforms=True, merged=self.args.merged, remove_classes=self.args.remove_classes, oversampling=self.args.oversampling, unlabeled_subset_ratio=self.args.unlabeled_subset, expand_labeled=self.args.fixmatch_k_img, expand_unlabeled=self.args.fixmatch_k_img*self.args.fixmatch_mu, unlabeled_augmentations=True if self.uncertainty_sampling_method == 'augmentations_based' else False, seed=self.args.seed, start_labeled=self.args.start_labeled) (fixmatch.py line 63).

When passing the parameter to the dataset initializer, the related parameters of K-medoid algo were not passed.

Where is k_medoids_model from?