You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
It would be better to have an option to create chunks by page for PDF in data_preparation.py. Currently, chunking is done by tokens(1024 tokens). This is only option.
I think index quality might be better to separate PDF into pages to have 1 page has 1 topic. My customer is cooperating to prepare for PDFs that 1 page has 1 topic. (1 topic not across pages etc..)
The text was updated successfully, but these errors were encountered:
It would be better to have an option to create chunks by page for PDF in data_preparation.py. Currently, chunking is done by tokens(1024 tokens). This is only option.
I think index quality might be better to separate PDF into pages to have 1 page has 1 topic. My customer is cooperating to prepare for PDFs that 1 page has 1 topic. (1 topic not across pages etc..)
The text was updated successfully, but these errors were encountered: