What kind of text extraction process can this "ibm-granite/granite-3.1-8b-instruct" perform?
#4
by
dhivakarsa
- opened
- text content from readable PDF, 2) text content from non readable PDF, and 3) text content from image files
Hi @dhivakarsa ! The Granite 3.1 models are pure language models, so they don't handle any modalities beyond text. They have not been trained to extract text from PDFs or images at this time. We are, however, actively working on multi-modality for the upcoming Granite 3.2 models, so stay tuned!
Thanks for your response @gabegoodhart