Extract all information?
#2
by
altozachmo
- opened
Is it possible to use this model to extract all of the text from an image? I would like to use this generically for OCR
Yes, you could do this. You should just play around with the template to see what works best for your specific use case. E.g. starting with something simple like {"text_content": "string"} could be enough.
Hi, I'm trying to run Transformers example (from 'Use this model') and always receiving error: Image features and image tokens do not match: tokens: 0, features 2976. Tried to change prompt to {"text_content": "string"}, but still the same error. Any ideas what I'm doing wrong?