itt

Image-to-text practice repo

Simple front-end graphical user interface (HTML and jQuery) to interact with back-end (Flask) and perform image-to-text captioning with artificial intelligence models from huggingface.

• Running without Triton Inference Server
  • Run file src/app.py
  • Open web browser
  • Navigate to http://localhost:5000/
• Running with Triton Inference Server (Python backend)
  • Create Docker images by running file docker_compose_python.sh
  • Open web browser
  • Navigate to http://localhost:5000/

Using the tool:

1. Upload image
2. Click on "Get Caption!"

Optionally, type a question for executing a Visual Question Answering task.