eval_retrieve_knn_pred.py [25:43]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
    args = parser.parse_args()

    for i in range(args.num_replica):
        os.path.exists(os.path.join(args.output_dir, 'feature_{}_{}.npy'.format(args.trainsplit, i)))
        os.path.exists(os.path.join(args.output_dir, 'feature_{}_cls_{}.npy'.format(args.trainsplit, i)))
        os.path.exists(os.path.join(args.output_dir, 'feature_{}_{}.npy'.format(args.valsplit, i)))
        os.path.exists(os.path.join(args.output_dir, 'feature_{}_cls_{}.npy'.format(args.valsplit, i)))
        os.path.exists(os.path.join(args.output_dir, 'vid_num_{}.npy'.format(args.trainsplit)))
        os.path.exists(os.path.join(args.output_dir, 'vid_num_{}.npy'.format(args.valsplit)))

    vid_num_train = np.load(os.path.join(args.output_dir, 'vid_num_{}.npy'.format(args.trainsplit)))
    train_padding_num = vid_num_train[0] % args.num_replica
    vid_num_val = np.load(os.path.join(args.output_dir, 'vid_num_{}.npy'.format(args.valsplit)))
    val_padding_num = vid_num_val[0] % args.num_replica

    feat_train = []
    feat_train_cls = []
    for i in range(args.num_replica):
        feat_train.append(np.load(os.path.join(args.output_dir, 'feature_{}_{}.npy'.format(args.trainsplit, i))))
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -



eval_svm_feature_perf.py [23:41]:
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
    args = parser.parse_args()

    for i in range(args.num_replica):
        os.path.exists(os.path.join(args.output_dir, 'feature_{}_{}.npy'.format(args.trainsplit, i)))
        os.path.exists(os.path.join(args.output_dir, 'feature_{}_cls_{}.npy'.format(args.trainsplit, i)))
        os.path.exists(os.path.join(args.output_dir, 'feature_{}_{}.npy'.format(args.valsplit, i)))
        os.path.exists(os.path.join(args.output_dir, 'feature_{}_cls_{}.npy'.format(args.valsplit, i)))
        os.path.exists(os.path.join(args.output_dir, 'vid_num_{}.npy'.format(args.trainsplit)))
        os.path.exists(os.path.join(args.output_dir, 'vid_num_{}.npy'.format(args.valsplit)))

    vid_num_train = np.load(os.path.join(args.output_dir, 'vid_num_{}.npy'.format(args.trainsplit)))
    train_padding_num = vid_num_train[0] % args.num_replica
    vid_num_val = np.load(os.path.join(args.output_dir, 'vid_num_{}.npy'.format(args.valsplit)))
    val_padding_num = vid_num_val[0] % args.num_replica

    feat_train = []
    feat_train_cls = []
    for i in range(args.num_replica):
        feat_train.append(np.load(os.path.join(args.output_dir, 'feature_{}_{}.npy'.format(args.trainsplit, i))))
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -