README.md

---
task_categories:
- summarization

---
# AutoTrain Dataset for project: musicprompt

## Dataset Description

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

### 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
[
  {
    "feat_ytid": "XILyHZyyCik",
    "feat_start_s": 140,
    "feat_end_s": 150,
    "feat_audioset_positive_labels": "/m/015lz1,/m/028sqc,/m/04rlf",
    "text": "['pop', 'low quality', 'live performance', 'flat male vocal', 'passionate female vocal', 'wide harmonizing vocals', 'punchy kick', 'punchy snare', 'shimmering hi hats', 'groovy bass', 'crowd cheering', 'crowd singing', 'energetic', 'groovy', 'emotional', 'passionate']",
    "target": "The low quality recording features a live performance of a pop song that consists of flat male vocal talking, passionate female vocal, alongside harmonizing wide female vocals, singing over punchy kick and snare hits, shimmering hi hats and groovy bass. There are crowd singing and cheering sounds in the background. It sounds groovy, emotional, energetic and passionate.",
    "feat_author_id": 4,
    "feat_is_balanced_subset": false,
    "feat_is_audioset_eval": false
  },
  {
    "feat_ytid": "t9aSL2MwEDM",
    "feat_start_s": 30,
    "feat_end_s": 40,
    "feat_audioset_positive_labels": "/m/015lz1,/m/04rlf,/m/04wptg,/m/05w3f,/m/064t9,/m/06rqw",
    "text": "['pop', 'amateur recording', 'e-bass', 'e-guitar', 'acoustic drums', 'synth', 'female voice singing', 'mid range', 'mideum to uptempo']",
    "target": "This song contains a female singer singing with a lower voice. An acoustic drum is playing a simple groove with a catchy bassline. An e-guitar is playing the notes along with the bass. A keyboard is emulating an e-guitar sound by playing a little lick. This song may be playing at a live concert.",
    "feat_author_id": 6,
    "feat_is_balanced_subset": false,
    "feat_is_audioset_eval": false
  }
]
```

### Dataset Fields

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

```json
{
  "feat_ytid": "Value(dtype='string', id=None)",
  "feat_start_s": "Value(dtype='int64', id=None)",
  "feat_end_s": "Value(dtype='int64', id=None)",
  "feat_audioset_positive_labels": "Value(dtype='string', id=None)",
  "text": "Value(dtype='string', id=None)",
  "target": "Value(dtype='string', id=None)",
  "feat_author_id": "Value(dtype='int64', id=None)",
  "feat_is_balanced_subset": "Value(dtype='bool', id=None)",
  "feat_is_audioset_eval": "Value(dtype='bool', 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        | 4416 |
| valid        | 1105 |