This project is an automated system for generating captions and tags for images using a pre-trained BLIP (Bootstrapped Language-Image Pretraining) model. It processes images from a database, generates captions and tags, and saves the results back to the database.
- Image Captioning: Generates descriptive captions for images using the BLIP model.
- Tag Generation: Extracts tags from captions for easier categorization.
- Database Integration: Fetches images from a database and saves generated tags back to it.
- Dockerized Deployment: Supports multi-architecture Docker builds for easy deployment.
- Logging: Provides detailed logs for debugging and monitoring.
- Python: Core programming language.
- Transformers: For using the BLIP model.
- Pillow: For image processing.
- PyTorch: For running the BLIP model.
- Docker: For containerization.
- GitHub Actions: For CI/CD workflows.
- Python 3.11
- Docker (optional for containerized deployment)
- A database for storing and retrieving images and tags
- Required Python dependencies (see
requirements.txt)
- Clone the repository:
git clone https://github.com/leet5/image-captioning-system.git cd image-captioning-system