Add NOUGAT model for pdf conversion

[panta-123]

There a opensource model from facebook: (NOUGAT) Neural Optical Understanding for Academic Documents.
This converts PDF to markdown and good for scientific article like EIC-RAG use..
May a thing to try and compare with PyPDFToDocument algorithm.

Github : https://github.com/facebookresearch/nougat/tree/main

[panta-123]

Another option will be: Marker
https://github.com/VikParuchuri/marker

[panta-123]

Seems marker has better accuracy and is faster.

[karthik18495]

I apologize for the naive question but May I ask how is the accuracy computed here?

[panta-123]

Decription here:
https://github.com/VikParuchuri/marker/tree/master?tab=readme-ov-file#benchmarks

Benchmark is calculated here:
https://github.com/VikParuchuri/marker/blob/master/benchmarks/overall.py

Just gazing through quickly , I think accuracy is defined as scores.
https://github.com/VikParuchuri/marker/blob/master/marker/benchmark/scoring.py#L35