This is a simple implementation of transformer models uisng nothing but some of the pytorch libraries(mostly releated to tensor multiplication and adam optimizer)
this code currently will spell out stories but could be trained on any text input.
git clone 'https://github.com/KrishnaAgarwal1308/Story-former.git'
cd 'Story-former'
python run full_paper_made.py
this by iteself will install the dataset and will train on your gpu or cpu as available.
All issues are invited