uartimcs's picture
Update README.md
59ccba5 verified
metadata
language:
  - en
metrics:
  - accuracy
pipeline_tag: image-to-text

This is the trained DONUT model for document parsing of logistics shipping documents for our FYP. For future applications. To smooth train and run, one should follow the version number of Transformers and timm installed. Inference API in huggingface does not work because I did not use the latest version of transformers.

Demo result: A dummy booking document from Evergreen and its results: image/png

image/png