Enhance Table Parsing by Utilizing GPT-4o

[0x01111]

Issue Description:

We are encountering issues with the Marker tool when parsing tables from PDF documents, resulting in format errors.

Proposed Solution:

Enhance the table extraction process by integrating GPT-4o to improve accuracy and reliability in parsing table content from PDFs.

[conjuncts]

Good suggestion. Does the newly added --use_llm argument suffice?