---
task_categories:
- summarization

---
# AutoTrain Dataset for project: facebook2

## Dataset Description

This dataset has been automatically processed by AutoTrain for project facebook2.

### Languages

The BCP-47 code for the dataset's language is unk.

## Dataset Structure

### Data Instances

A sample from this dataset looks as follows:

```json
[
  {
    "text": "\u201cWhen the vessel is anchored or docked, the sonar is set up with continuous surveillance 360\u00b0 around the vessel for detection of underwater threats; such as divers, diver delivery vehicles, or submersibles.\u201d",
    "target": "identify underwater threats"
  },
  {
    "text": "\u201cLimiter: Useful to reduce external interference, such as acoustic interference from other sonar/sounder equipment on board and electrical interference from power supplies and power cables.\u201d",
    "target": "identify the limiter settings"
  }
]
```

### Dataset Fields

The dataset has the following fields (also called "features"):

```json
{
  "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        | 264 |
| valid        | 66 |