The number of unlabeled data for CIFAR-10.

[omg777]

"For CIFAR-10 [34], 4,000 labeled examples are kept as labeled
data while 41,000 examples are used as unlabeled data"

Why did you use whole data for the unlabeled data?

def x_u_split(args, labels):
    label_per_class = args.num_labeled // args.num_classes
    labels = np.array(labels)
    labeled_idx = []
    # unlabeled data: all training data
     ...