Query about data preparation for finetuning for nyuv2-depth

[AntiLibrary5]

Hi, I think the following correction holds: Line 357 in https://github.com/EPFL-VILAB/MultiMAE/blob/main/run_finetuning_depth.py should be dataset_train = build_regression_dataset(args, data_path=args.train_data_path, transform=train_transform) instead of dataset_train = build_regression_dataset(args, data_path=args.data_path, transform=train_transform) or the argument train_data_path should be changed to data_path.

Apart from that, I am trying to recreate your results on NYUv2 for depth. but the dataset preparation instructions are not clear from the instructions in SETUP. As explained about the folder structure , where should the GT be when finetuning for depth and evaluating. Apart from that, mask_valid for fine-tuning? RuntimeError: Found 0 logs in subfolders of: /tmp-network/user/varora/multimae/multimae_data/train/mask_valid

[roman-bachmann]

Hi!

Thanks for noticing! We just pushed the fix to rename the argument to data_path.

Regarding the dataset preparation: We structured all our datasets in the style of the usual ImageNet folder structure (train/val folders, and class folders), with the difference that we add another level of folders that specifies the modality. Therefore, try to add another folder inside mask_valid, that acts as the "class folder". For example, you can just call it all. The final structure should be train/mask_valid/all/{images}.

Best, Roman