Pytorch implementation for AAAI21 paper: Deep Partial Rank Aggregation for Personalized Attributes

Run

Command: python mle_train.py [train_file] [test_file] [save_path] [log_file] [gpu_id] [backbone]

Parameters:

• train_file: training annotation file, e.g., 'lfw10/lfw10_train.txt'
• test_file: test annotation file, e.g., 'lfw10/lfw10_test.txt'
• save_path: model save path
• log_file: output log file path
• gpu_id: single id supported
• backbone: 'alexnet', 'vgg', 'resnet'

Competitor: naive_train.py