Chainer implementation of the paper "Mean teachers are better role models: Weight-averaged consistency targets improve semi-supervised deep learning results" (https://arxiv.org/abs/1703.0...