This is a text summarizer web app that is attached to a large language model, specifically flan-t5-base, fine-tuned to generate short TL;DR summaries of any text.
Huggingface🤗 Repository
The flan-t5-base model was trained on a dataset of 100 thousand Reddit posts that contained a tl;dr summary in the comments, obtained from this dataset. The model was trained for 5 epochs using the Huggingface Transformers library. The model was trained on AWS Sagemaker using a g4dn.xlarge instance for ~12 hours.
- Open your browser and go to toolongdontread.live, or run frontend/index.html locally.
- Paste the text you want to summarize in the input box.
- Click on the "TL;DR" button.
- The summary will be displayed below the input box.
This project was heavily inspired by and follows the training setup from this tutorial