Use only recognition model

[lucas2lab]

Hello,

I have a very simple issue, I did some research but did not find any pointers beyond reading the codebase of doctr.

On very long and complex documents I need to extract a single word which is a serial number. I trained a model that is able to locate these serials very accurately and would like to feed them to the doctr. My issue is that the ocr_predictor seems to be always coming with its detection_predictor which is trained to on whole documents, not single words.

I have difficulties finding how could I send directly my pictures to the recognition_predictor so that I can get the text for my single-word pictures.

Any pointers?

Thank you very much!

[lucas2lab]

Found it :)

from doctr.models import recognition_predictor
from doctr.io import DocumentFile

img = DocumentFile.from_images('/path/to/img.jpg')

model = recognition_predictor(pretrained=True)
model(img)