Leverage the ability of the mPLUG-Owl doc understanding mannequin to ask questions on your paperwork
This text will focus on the Alibaba doc understanding mannequin, just lately launched with mannequin weights and datasets. It’s a highly effective mannequin able to performing varied duties corresponding to doc query answering, extracting data, and doc embedding, making it a useful instrument when working with paperwork. This text will implement the mannequin regionally and check it out on completely different duties to provide an opinion on its efficiency and usefulness.
· Motivation
· Duties
· Operating the mannequin regionally
· Testing of the mannequin
∘ Knowledge
∘ Testing the primary, leftmost receipt:
∘ Testing the second, rightmost receipt:
∘ Testing the primary, leftmost lecture observe:
∘ Testing the second, rightmost lecture observe
· My ideas on the mannequin
· Conclusion
My motivation for this text is to check out the newest machine-learning fashions which can be publicly accessible. This mannequin caught my consideration since I’ve labored and am nonetheless engaged on machine studying utilized to paperwork. I’ve additionally beforehand written an article on my work with an identical mannequin known as Donut that does OCR-free doc understanding. I feel the idea of getting a doc and asking visible and textual questions on it’s superior, so I spend time working with paperwork, understanding fashions, and testing their efficiency. This text is the second article in my collection on testing out the newest machine-learning fashions, and you’ll learn my first article on time collection forecasting with Chronos beneath:
