fastapi
uvicorn
python-multipart
torch
datasets
pdf2image
accelerate
pytesseract
transformers[torch,sentencepiece]
haystack-ai
qdrant-haystack 
fastembed-haystack
scikit-learn
pdfplumber
unstructured
pdfminer.six
pillow_heif
unstructured_inference
unstructured_pytesseract