Upload cnn_dailymail dataset and README

README.md

@@ -0,0 +1,48 @@
+---
+pretty_name: "CNN_Dailymail"
+license: "mit"
+language:
+  - en
+tags:
+  - summarization
+  - news
+task_categories:
+  - summarization
+---
+
+
+
+# CNN_Dailymail
+
+This repository hosts a copy of the **CNN_Dailymail** dataset, a large-scale dataset designed for evaluating abstractive text summarization systems.
+
+CNN_Dailymail consists of news articles paired with human-written summaries, commonly used for training and evaluating models on summarization tasks. It contains articles from CNN and Daily Mail, covering a wide range of topics.
+
+## Contents
+
+- `cnn_dailymail.jsonl` (or your actual filename): the standard set of news articles and summaries.
+
+Each entry contains:
+
+```
+{
+  "id": "...",
+  "article": "...",
+  "highlights": "..."
+}
+```
+
+## Usage
+
+```
+from datasets import load_dataset
+ds = load_dataset("S3IC/cnn_dailymail")
+```
+
+## Source
+
+This dataset is taken from the public CNN_Dailymail release: https://github.com/abisee/cnn-dailymail
+
+## License
+
+CNN_Dailymail is released under the original license provided by the authors. See the LICENSE file for more details.