Features
-
Multi-GPU processing
for faster and more efficient processing. -
Text extraction with Docling
from PDF and other file types -
Greek OCR support
recognize Greek text from images and PDFs using DeepSeek OCR or RapidOCR. -
Formula recognition
in LaTeX with Docling's math enhancement model or DeepSeek OCR -
Fast CPU-only Text Extraction
with self.batch_policy = "safe" using pypdfium backend