Question about manipulation of documents in doctr

[Raul220]

Hello!
I want to know if when I use the doctr models on a document, as the following example:

model = ocr_predictor(det_arch='db_resnet50', reco_arch='crnn_vgg16_bn', pretrained=True)
pdf_doc = DocumentFile.from_pdf("path/to/your/doc.pdf")
result = predictor(doc)

and the models db_resnet50 and crnn_vgg16_bn are already in cache, does the predictor or from_pdf functions performs the processing locally?
Is the pdf document uploaded to the internet?

Thanks

[felixT2K]

Hey @Raul220 👋🏼 ,

everything runs on your machine. No files are uploaded anywhere. :)

All downoaded files (models & datasets) can be found in ~/.cache/doctr you can also specify a custom location by setting the DOCTR_CACHE_DIR environment variable :)

ref.: str(os.environ.get("DOCTR_CACHE_DIR", os.path.join(os.path.expanduser("~"), ".cache", "doctr")))