In datasets/build.py > def build_dataset > should the "anno_path" for mode = 'test' be (args.data_path, 'test.csv') instead of what it is currently (args.data_path, 'val.csv')?

OpenGVLab / VideoMAEv2

[CVPR 2023] VideoMAE V2: Scaling Video Masked Autoencoders with Dual Masking

https://arxiv.org/abs/2303.16727

MIT License

493 stars 56 forks source link

In datasets/build.py > def build_dataset > should the "anno_path" for mode = 'test' be (args.data_path, 'test.csv') instead of what it is currently (args.data_path, 'val.csv')? #27

Closed yerx closed 1 year ago

yerx commented 1 year ago

def build_dataset(is_train, test_mode, args): if is_train: mode = 'train' anno_path = os.path.join(args.data_path, 'train.csv') elif test_mode: mode = 'test' anno_path = os.path.join(args.data_path, 'val.csv') else: mode = 'validation' anno_path = os.path.join(args.data_path, 'val.csv')

congee524 commented 1 year ago

Yes, but generally we use the validation set, and often only at competition will there be additional test datasets. Feel free to change the name of it to what you want.