post_ocr_correction For the inference, how can we do the alignment?

For the inference, how can we do the alignment?

Open Alexsong111 opened this issue 2 years ago • 2 comments

When we do inference, we do not have aligned input. If we input "Dcmen", the output should be "Documen", which will be longer than the maximum length allowed. Could you please answer this question? Thank!

Aug 04 '23 20:08 Alexsong111

Can you send a code snippet of what you are trying to do? Also, did you FULLY run the second script in the main page? The one titled "Load one of the pre-trained models"?

Aug 07 '23 20:08 jarobyte91

I just ran the code snippet and it works, it doesn't have any trouble outputting sequences longer than the input. Optionally, you can run the script in tests/load_model.py

Aug 07 '23 20:08 jarobyte91

post_ocr_correction post_ocr_correction copied to clipboard

For the inference, how can we do the alignment?

post_ocr_correction
post_ocr_correction copied to clipboard