- Fork this repo
- Clone it and open it on your editor
- Download this text dataset : https://drive.google.com/file/d/1tgHdgYLq5eP9gjYUEietqy8tNEmaARig/view?usp=sharing (or use your own text dataset)
- Save it to the directory with name "./src/full_text_dataset.txt"
- Make sure your current directory is "src"
- Run "python .\train.py"
Number of parameters = 356M