Hi all,
On the last page of the “Generating Text with Deep Learning” lesson it says that: “You could also convert the one-hot vectors into word embeddings during training to improve the model.” Since I’m trying to make a generative chatbot independently with different data, can anyone tell me which part of the lesson this conversion would be done in? Better yet, exactly how would you convert the vectors if you were to modify the lesson yourself?
Thank you for your time,
mews_mochi