metadata
task_categories:
- summarization
AutoTrain Dataset for project: text2hashtag
Dataset Description
This dataset has been automatically processed by AutoTrain for project text2hashtag.
Languages
The BCP-47 code for the dataset's language is unk.
Dataset Structure
Data Instances
A sample from this dataset looks as follows:
[
{
"text": "An intense battle between moderates and progressives has already spilled into public view.",
"target": "['Presidential Election of 2020', 'Police Reform', 'United States Politics and Government', 'Black People', 'Blacks', 'Elections, House of Representatives', 'George Floyd Protests (2020)', 'Democratic Party', 'Democratic Socialists of America', 'Justice Democrats', 'Sunrise Movement']"
},
{
"text": "The blaze is setting off mines and other ordnance littering the war zone in eastern Ukraine, hampering already dangerous firefighting and evacuation efforts.",
"target": "['Wildfires', 'Explosions (Accidental)', 'Ukraine']"
}
]
Dataset Fields
The dataset has the following fields (also called "features"):
{
"text": "Value(dtype='string', id=None)",
"target": "Value(dtype='string', id=None)"
}
Dataset Splits
This dataset is split into a train and validation split. The split sizes are as follow:
| Split name | Num samples |
|---|---|
| train | 2399 |
| valid | 600 |