Skip to content

Latest commit

 

History

History
35 lines (33 loc) · 1022 Bytes

README.md

File metadata and controls

35 lines (33 loc) · 1022 Bytes

AI for Cyber purposes

This repository includes the following:

  1. NER - extract cyber entities. Evaluation of different models
  2. Image-Captioning - caption desktop, screenshots and regular images. Evaluation of different models
  3. Playground - Export the selected NER and Image-Captioning models via streamlit
  4. Triton - Incomplete efforts to improve models inference times using Nvidia Triton

Run the playground

  1. In NER folder:
./build/export.sh
  1. In Image-Captioning folder:
./build/download_lavis_model.sh
  1. In Image-Captioning/build/ folder:
git clone [email protected]:huggingface/optimum.git
  1. In Image-Captioning/build/optimum folder:
python3 setup.py develop
cp /usr/local/bin/optimum-cli .
  1. In Image-Captioning/build folder:
./optimum/optimum-cli export onnx -m google/pix2struct-screen2words-base --optimize O3 onnx/pix2struct_base
  1. In root folder:
docker-compose build && docker-compose up
  1. Access the playground via http://0.0.0.0:8501/