Unifying Vision, Text, and Layout for Universal Document Processing
Paper
•
2212.02623
•
Published
•
11
UDOP
UDOP is a general multimodal model for document AI
-
-
microsoft/udop-large
Image-Text-to-Text • 0.7B • Updated • 3.09k • 116 -
microsoft/udop-large-512
Image-Text-to-Text • 0.7B • Updated • 283 • 5
microsoft/udop-large-512-300k
Image-Text-to-Text • 0.7B • Updated • 166 • 32Note This is the best performing model, as it uses the highest image resolution (512x512) and is trained the longest.