---
dataset_info:
- config_name: default
  features:
  - name: text
    dtype: string
  - name: summary
    dtype: string
  - name: id
    dtype: string
  splits:
  - name: train
    num_bytes: 1260614391
    num_examples: 287113
  - name: validation
    num_bytes: 57672089
    num_examples: 13368
  - name: test
    num_bytes: 49877597
    num_examples: 11490
  download_size: 836881022
  dataset_size: 1368164077
- config_name: mini
  features:
  - name: text
    dtype: string
  - name: summary
    dtype: string
  - name: id
    dtype: string
  - name: cluster_id
    dtype: int64
  splits:
  - name: train
    num_bytes: 169001211
    num_examples: 38500
  download_size: 103090954
  dataset_size: 169001211
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
  - split: test
    path: data/test-*
- config_name: mini
  data_files:
  - split: train
    path: mini/train-*
license: odc-by
task_categories:
- summarization
language:
- en
---

# cnn_dailymail: cleaned

Original [cnn_dailymail](https://huggingface.co/datasets/abisee/cnn_dailymail) config 3.0.0 with the following changes:

- renamed columns to `text` and `summary`
- cleaning applied to `summary` column to correct punctuation, etc.