Specific Settings of the ToMe Model

[PhuTran1005]

Could you share the right settings in the 'extract_tome14.py' file?
Is it right if we use the default settings as follows:
"
parser.add_argument("--gpu", dest='gpu', type=str, default='0',
help='Set CUDA_VISIBLE_DEVICES environment variable, optional')
parser.add_argument("--dir_path", type=str, default='/data/users/guangyao_li/AVQA_Tsinghua/avqa_frame_1fps',
help='sec path')
parser.add_argument("--dst_path", type=str, default='/data/users/guangyao_li/AVQA_Tsinghua/ToMe_feat/visual_tome14',
help='sec save path')
parser.add_argument("--sample_frames", type=int, default=10, help='sample frames')
parser.add_argument("--tokens", type=int, default=14, help='merge tokens numbers')
parser.add_argument("--layers", type=int, default=23, help='merge tokens layers, 25, 23, 22, 20, 16, 8')
parser.add_argument("--model_name", type=str, default="vit_large_patch16_384", help='model_name')
parser.add_argument("--global_pool", type=str, default="None", help='global_pool')
"

Thank you so much.
Regards,
PhuTD