The following repository contains code for executing summarization on the CNN/Dailymail Datasets. I have used extractive summarization here by using the Hugging Face Transformer since it's metric was easy for me to understand and explain than abstractive summarization. But, I had done abstractive summarization parallely with this and can explain it if required.
SETUP
- Download the following dataset (train.csv) and upload it to Google Colab https://drive.google.com/file/d/1pk7yOaO7KtMKK7g8C3RcfXafPSkiqWDR/view?usp=share_link
- Run each section sequentially