embedding

[ddy-ddy]

Hi author, I was very inspired by your article, but I have some doubts in the process of reproduction and I would like to ask you for advice.

1. Is the code embedding you mentioned generated during the training process, or can it be generated in advance?
2. Are the code embeddings and word embeddings in the same distribution space?
3. If I want to take out the embedding data of the code embedding and the question embedding separately, how can I do that?
   Thank you very much for your article which has given me a lot of inspiration.