santiviquez / harry-potter-rnn Goto Github PK

Hey, wonderful work!!

Just one quick question.

Here, we get character level embeddings once we train fully, is there any way we can get word level from it, or need to try another model?

I tried taking character level embeddings for a word and did an average but the results are not satisfactory.
E.g. Harry Potter and Carbon Dioxide are giving 90% cosine similarity.

Any thoughts or any repo can you suggest which learns word embeddings with text generation as a downstream task.
Thank you!!

Asked here since I didn't found any contact details of yours!

Hey,
Nice work!!

I just have one query.
I am working on character-level text generation with data as novels.

In Novels, there are character names, which occur too many times, so will that affect my model while generating text?

Thanks in Advance!