It was really helpful and actually ensured me that I am on a good way with my own project! But I have a problem with conversion of my `inputs` set to tensor - something similar to where you use embedding. How does the embedding in your case work, where does it come from and what should it do?
Awesome tutorial. Can you explain how you can - add the new entries that have been entered by the user to the existing training set and retrain it instantly so that the model gets more and more accurate along the way - How you could add a new type of icon and retrain the model instantly also This would be a great new video :-) Thank you very much