https://d2l.ai/chapter_recommender-systems/autorec.html
How can the training loss be so low?
Thank you for this awesome book!