metadata
language:
- en
task_categories:
- text-classification
AutoNLP Dataset for project: tweet-sentiment
Table of content
Dataset Descritpion
This dataset has been automatically processed by AutoNLP for project tweet-sentiment.
Languages
The BCP-47 code for the dataset's language is en.
Dataset Structure
Data Instances
A sample from this dataset looks as follows:
[
{
"text": "I am going to see how long I can do this for.",
"target": 8
},
{
"text": "@anitabora yeah, right. What if our politicians start using uploading their pics, lots of inside sto[...]",
"target": 8
}
]
Dataset Fields
The dataset has the following fields (also called "features"):
{
"text": "Value(dtype='string', id=None)",
"target": "ClassLabel(num_classes=13, names=['anger', 'boredom', 'empty', 'enthusiasm', 'fun', 'happiness', 'hate', 'love', 'neutral', 'relief', 'sadness', 'surprise', 'worry'], 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 | 31995 |
| valid | 8005 |