imgcaption

Simple Python script to generate descriptive texts (captions) for images, using the Hugging Face Transformers library with the Salesforce/blip-image-captioning-base model. It is basically just a wrapper around such library.

Inspired by:

• Building an Image Captioning Model Using Salesforce's BLIP Model - by Pranav Kumar - Medium
• What is Image Captioning and How to use Python to Generate Caption from an Image - by Jim Wang - Medium
• An Introduction to Image Captioning with BLIP - Aggregata

Usage

Important: this has been tested with Python 3.12.4 on Windows 10.

Set up a Python venv (virtual environment) and install some packages inside it:

1
2

python3 -mvenv venv
venv/bin/python3 -mpip install -r requirements.txt

Then you can use the script like this:

1

time printf '%s\n' *.jpg | venv/bin/python3 main.py