Is there a loader that uses OCR as

reader = PDFNougatOCR()
pdf_path = Path(directory_path + 'my_pdf.pdf')
documents = reader.load_data(pdf_path)

gets the error: ERROR:root:An error occurred while processing the PDF: [Errno 2] No such file or directory: 'nougat'

am i supposed to just only have a single string with the full path? or is this some kind of issue with the import

yeah whats your directory path? seems like it doesn't exist

i was missing a requirement which solved that error

but wow this takes forever to run on 1 pdf if you are using CPU

yes need gpu!

its pretty good tho

any suggestions for MAC ppl?

most OCR models that are worth it are going to use GPU 😅

Although I don't think theres a loader for tesseract, so you'd have to create the document objects yourself (it's really not that hard though 🙏 )

@velocitybolt @Logan M use colab? is there a way to make nougat use mac's gpu's (m2/m1)?

I think logan once told me model.to("mps:0")

nougat tries to use mps automatically under the hood, but I couldn't get it to work on my M2 🤔 (its all in some external library)

colab is a good option

hello, do you have any code to help me with, i am using colab and this is not really working for me from llama_index import PDFNougatOCR

Initialize the PDFNougatOCR reader

pdf_reader = PDFNougatOCR()

as i want to convert the non modifiable pdf to modifiable onces to use them in my RAG