---
language:
- en
task_categories:
- text-classification
---
# AutoNLP Dataset for project: tweet-sentiment

## Table of content
- [Dataset Description](#dataset-description)
    - [Languages](#languages)
- [Dataset Structure](#dataset-structure)
  - [Data Instances](#data-instances)
  - [Data Fields](#data-fields)
  - [Data Splits](#data-splits)

## 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:

```json
[
  {
    "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"):

```json
{
  "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 |