/document-classification-api

Primary LanguagePython

api for doc classification

fastapi app for documents classificatio

model used

ocr

  • textract

language detection

  • XLM-Roberta-base-language-detection

zero shot classification

  • Bart-MNLI-CNN_news

install

 sudo apt-get install tesseract-ocr
pip install -r requirements.txt

Usage

create a config.py

create config.py in ./apis containe

labels = [] #labels list ["news","sport"]
huggingface_key = "<huggingface key>"
deta_key = "<deta key>"

run the app

uvicorn main:app --reload

you can check /docs endpoint to get the documentation