Fast text pytorch implementation

I am unable to understand the fast text paper, like the exact mathematical operation done in the classification task. Is there any PyTorch implementation out there which I can read through to understand the model and how word emeddings are derived?