priyank-m

/

m_OCR

like 12

?

apologies for the late response, the model does not have it's own preprocessor
it uses tokenizer from "xlm-roberta-base" and the feature_extractor from "facebook/vit-mae-large"
the code for this will look something like below (you can look into the app.py file in the model demo Space as well to see how to use it):

from transformers import VisionEncoderDecoderModel, AutoTokenizer, AutoFeatureExtractor
model = VisionEncoderDecoderModel.from_pretrained("priyank-m/m_OCR")
tokenizer = AutoTokenizer.from_pretrained("xlm-roberta-base")
feature_extractor = AutoFeatureExtractor.from_pretrained("facebook/vit-mae-large")