Closed zhicong01 closed 10 months ago
Thanks for your comment, but this is not a mistake, its on purpose.
Since we train with the InfoNCE loss we have to make sure that the same building is not multiple times in the batch. We do this with our specialized shuffle method in the dataset. The pairs we generate in the dataset init are dependent on the position of the query_folder_train and gallery_folder_train.
Thanks to the author for the well-commented code. I find a mistake in code. In the train_university.py (line94-line95):
config.query_folder_train = './data/U1652/train/satellite' config.gallery_folder_train = './data/U1652/train/drone'
which should change to :config.query_folder_train = './data/U1652/train/drone' config.gallery_folder_train = './data/U1652/train/satellite'