CNN DailyMail Summarization Benchmark for Model Evaluation

Autoevaluate's staging project contains data for evaluating models on the CNN/DailyMail news summarization task. The dataset was uploaded by the user 'autoevaluate' on July 15, 2022. It appears to be a benchmark dataset used for testing automated text summarization systems.

Use Cases

• Train sequence-to-sequence models to generate article_summaries from full news_article text.
• Benchmark extractive summarization models by scoring against reference_summaries.
• Evaluate model performance on specific metrics like ROUGE scores using the provided evaluation splits.

Strengths

• Based on the established CNN/DailyMail dataset, a standard benchmark for summarization.
• Uploaded by a user associated with model evaluation tooling ('autoevaluate').

Limitations

• Specific row count, column details, and modifications from the original dataset are unknown.
• The dataset's recency is limited, last updated in mid-2022.

Provenance

Source

Derived from the CNN/DailyMail dataset, originally created by See et al. (2017).

Collection Method

Likely a processed or staged version for automated evaluation pipelines.

Time Range

null

Freshness

Last updated on 2022-07-15.

Geography

null