Train labelled/unlabeled split for semi-supervised setup

HUANGLIZI / LViT

[IEEE Transactions on Medical Imaging/TMI] This repo is the official implementation of "LViT: Language meets Vision Transformer in Medical Image Segmentation"

MIT License

298 stars 26 forks source link

Train labelled/unlabeled split for semi-supervised setup #49

Closed muqeemmm closed 2 months ago

muqeemmm commented 2 months ago

I have been following your work and I have been unable to find the labeled/unlabeled split for QaTa-COV19 and MosMedData+ used in the semi-supervised setup. To replicate the results proposed in the paper what splitting strategy should I use? If there is a split already available, please do let me know.

HUANGLIZI commented 2 months ago

Please use train_dataset, test_dataset = torch.utils.data.random_split(full_dataset, [train_size, test_size]) to split the dataset, and the random seed is the same as others.

muqeemmm commented 2 months ago

Thank you. I am closing the issue.