Ingested pdf document is not recognized

[JannikSchneider12]

Hey,

I have run the ingest.py and I says that my provided pdfs are ingested (no errors). However if I ask my model something about these documents (e.g. even if it knows the "name.pdf" file), it says no and isnt aware of the information. Does anybody have an idea how I could solve this or where the error could be?

Thanks for your help

[TechInnovate01]

Hey,

I have run the ingest.py and I says that my provided pdfs are ingested (no errors). However if I ask my model something about these documents (e.g. even if it knows the "name.pdf" file), it says no and isnt aware of the information. Does anybody have an idea how I could solve this or where the error could be?

Thanks for your help

Hi , Before run the python run_localgpt.py command you have run the python ingest.py command. If not , then try it once, if yes, what is error message you get while running the run_localgpt.py file.

[neural-oracle]

but what I do not get: sure... after I ingest ... it can answer questions.... however, after I remove the doc from the ingest folder it cannot answer the question anymore. I do not understand this behaviour. I expect that after ingesting the document contents are split, and then via embeddings stored in the vector store. So even after removing the document the content should still be known. But they are not ...