In the pipeline, I integrate the entity extraction, and my inputs are PDF documents. However, I know that the PDF should have this information. I receive null in my entity extraction using LLM for some of the PDFs. Does anyone know what happens here?
Are you receiving nulls in the trail runs or in the actual build? Are there any errors that come up you can check by checking off this box:
Also, is your output type a struct? Do you think you could share the details of your LLM panel here? There was a case earlier where a null date within a struct was found to null out the entire struct so I’m curious if you’re able to extract some fields and it only becomes a problem when you started extracting more