hf_spaces/README.md

STOP Sign Identification App

This Gradio application identifies whether an image contains a STOP sign or not using an AutoGluon Multimodal predictor.

How to Use

1. Upload an Image: You can upload an image from your computer or use your webcam.
2. Select an Example: Choose one of the provided example images to quickly test the application.
3. Apply Preprocessing: The "Apply Preprocessing" checkbox controls whether the input image is resized and converted to RGB before being fed to the model. By default, preprocessing is applied.
4. View Results: The application will display:
   • The original uploaded image.
   • The preprocessed image (if preprocessing is applied).
   • The predicted class probabilities ("Not a STOP sign" and "STOP sign").

Model

The application uses a pre-trained AutoGluon Multimodal predictor fine-tuned for this binary classification task.

Deployment on Hugging Face Spaces

This application is designed to be easily deployed on Hugging Face Spaces. The app.py and requirements.txt files generated in the hf_spaces directory contain all the necessary code and dependencies. Simply create a new Space on Hugging Face, select the "Gradio" SDK, and upload the contents of the hf_spaces directory.