Docling Optimize

Default docling project have performance issue.
This repo change some processing logic to save RAM and GPU/CPU.
(Still updating)

Testing

0. Model setup

Put offline models inside models folder or change import to model2 in document.py to use online model.

1. pymanager (Windows):

py exec -V:3.11 -m venv .venv
.\.venv\Scripts\activate
pip install -r .\requirements.txt

2. Test

python -m src.docling