Skip to content

README.md

Latest commit

 

History

History
18 lines (13 loc) · 782 Bytes

README.md

File metadata and controls

18 lines (13 loc) · 782 Bytes

Image Captioning with BLIP

Demo

This project uses Salesforce's BLIP (Bootstrapping Language-Image Pre-training) model to automatically generate text descriptions for uploaded images. The interface is built using the Gradio library.

Features

  • Generate natural language captions for any image.
  • Simple web interface (drag and drop files or browse).
  • Optimized generation parameters to avoid repetitions (repetition_penalty, no_repeat_ngram_size).
  • Automatic conversion of images to RGB format (support for PNG with transparency).

Installation

  1. Clone the repository:
git clone https://github.com/MNJMARIA/blip-image-captioning.git
cd blip-image-captioning