SceneScribe 🖼️🗣️

SceneScribe is a project that aims to generate natural language captions for images. It uses a deep learning model called Blip-image-captioning-base to encode the visual features of an image and decode them into a meaningful sentence. The model is trained on the Flickr30k dataset, which contains 30,000 images and five captions for each image.

Getting Started

To get started with using this tool, you might want to clone or download the repository.

git clone https://https://github.com/JohnPaulGummapu/SceneScribe
cd SceneScribe

Requirements 📋

You can install the dependencies by running:

pip install -r requirements.txt

Usage 🚀

To use this project, you can either run it as a command-line tool or as a web interface.

Command-line tool 💻

To run the project as a command-line tool, you can use the following command:

python caption.py --image <path_to_image>

This will generate a caption for the image and print it to the standard output.

Web interface 🌐

To run the project as a web interface, you can use the following command:

python app.py

This will launch a Flask server and open a web page in your browser. You can then upload an image and see the generated caption.

Contribution 💪

This project is open source and welcomes any contribution to fine-tune the model or improve the user experience. You can fork this repository and submit a pull request with your changes. Please follow the code style and documentation guidelines.