This project helps you build a search and generation pipeline for Easy Language Images (Leichte Sprache Bilder).
Our mission at SUMM AI is to make information accessible for everyone at the click of a button. This includes not only the translation of complex texts into Easy Language (Leichte Sprache) but also the use of easy illustrations to support understanding.
This project focuses on the generation and search of images for Easy Language texts by providing comprehensive pipelines for both functionalities. The image generation pipeline allows you to fine-tune an image generation model using your own image dataset. The image search pipeline enables you to search for specific images within an existing image database using text queries. Both pipelines are integrated with Gradio apps to simplify access for endusers.
This repository is organized into two main modules:
-
Image Generation:
This module includes a comprehensive pipeline for fine-tuning an image generation model using your own image dataset. Additionally, it provides an inference pipeline integrated in a Gradio app. The app allows users to convert Easy Language (Leichte Sprache) text into image prompts and generate images using the fine-tuned model. -
Image Search:
This module includes a pipeline to enable image search within a dataset. Users can input a dataset of images and search for specific images using text queries. A Gradio app is included to simplify access for endusers.
If you want to contribute to this open source project, please create an issue (if not already existing) or raise a pull request.
SUMM AI has developed an AI-powered tool to make complex text easy to understand and accessible for everyone at the click of a button. More info: summ-ai.com
If you have any questions, feedback, or inquiries, please feel free to reach out to us. You can contact us through the following channels:
- Website: https://summ-ai.com
- Linkedin: https://www.linkedin.com/company/summ-ai/
- Instagram: https://www.instagram.com/summ_ai_leichtesprache/
This project is funded by the Bundesministerium für Arbeit und Soziales (BMAS) via the „Civic Innovation – Förderung von gemeinwohlorientierten KI-Projekten“ funding guideline and implemented in collaboration with the Senatskanzlei of Hamburg. More info about the project.