Dataset Card for "discofuse"
Dataset Summary
DiscoFuse is a large scale dataset for discourse-based sentence fusion.
Dataset Structure
Data Instances
discofuse-sport
Size of downloaded dataset files: 4.33 GB
Size of the generated dataset: 15.04 GB
Total amount of disk used: 19.36 GB
An example of 'train' looks as follows.
{… See the full description on the dataset page:
https://huggingface.co/datasets/google-research-datasets/discofuse.