Since the data-set is scrapped from the internet you need to remove the starting and ending
tags and clean it before feeding to model…
I did not find any tags and my model is also generating new lines,
Getting 0% accuracy
Make sure you use np.random.seed(11), and there are no tags to be removed, in updated training file.
Try to avoid any randomness in your code.
If you still do not find anything, share your ipynb by saving it on google drive and sharing the link.
Happy Learning.