This code trains a simple Bag-of-Words model for sentiment classification of reviews with PyTorch.
I used the Yelp Review Polarity Dataset available at Xiang Zhang's Google Drive dir here and used in
Xiang Zhang, et al, Character-level Convolutional Networks for Text Classification. Advances in Neural Information Processing Systems 28 (NIPS 2015).
Pretty surprising, that it is possible to achive 80%+ accuracy on this task with such a simple model and without any text preprocessing.