Semi-supervised_MNIST

Semi-supervised Learning for MNIST Dataset

1. I use 3000 labeled data and 47000 unlabeled data for this learning task.
2. I've tried feature extraction and pseudo-label methods.
3. Several common techniques are used in this task.

How to run the code

1. Make sure you have installed pytorch.
2. Make sure you have downloaded every file.
3. You can download the MNIST data here.