I am trying to fine-tune TrOCR on Arabic, I followed TrOCR Fine-Tune Tutorials by @NielsRogge, choose a suitable encoder/decoder option suggested by @NielsRogge on discussion board but I get very high WER score and the generated text is gibberish, I tried the pytorch implementation and it generated text prediction to continue the sentence instead of writing the sentence in the image.
1 Like
how should
I fix it?
did you arrive to a solution ?
check mine and tell me : (successful run)HandWritten-OCR-TrOCR (kaggle.com)