README.md

---
language:
- en
license: apache-2.0
tags:
- gaming
annotation_creators:
- crowdsourced
language_creators:
- crowdsourced
pretty_name: pale
size_categories:
- 10K<n<100K
task_categories:
- text-generation
- text-classification
- automatic-speech-recognition
configs:
- config_name: quotes
  data_files:
  - split: train
    path: quotes/*.parquet
  default: true
- config_name: vanilla
  data_files:
  - split: train
    path: vanilla/*.parquet
  default: false
- config_name: annotated
  data_files:
  - split: train
    path: annotated/*.parquet
  default: false
- config_name: pulled
  data_files:
  - split: train
    path: pulled/*.parquet
  default: false
dataset_info:
  features:
  - name: header
    dtype: string
  - name: subheader
    dtype: string
  - name: text
    dtype: string
  - name: source
    dtype: string
  - name: champion
    dtype: string
  splits:
  - name: train
    num_bytes: 14430202
    num_examples: 67575
  download_size: 2675223
  dataset_size: 14430202
---

# Dataset card for pale

## Table of contents

- [Dataset description](#dataset-description)
    - [Dataset summary](#dataset-summary)
- [Dataset structure](#dataset-structure)
    - [Dataset instance](#dataset-instance)
    - [Dataset fields](#dataset-fields)

## Dataset description

- **Homepage:** [pale homepage](https://huggingface.co/datasets/zeio/pale)
- **Repository:** [pale repository](https://huggingface.co/datasets/zeio/pale)
- **Point of contact:** [Zeio Nara](mailto:zeionara@gmail.com)
- **Dataset version:** `30.10.2023`

### Dataset summary

This dataset contains league of legends champions' quotes parsed from [fandom](https://leagueoflegends.fandom.com).
See dataset usage example [at google colab](https://cutt.ly/3wEKDUI9).
The dataset is available in the following configurations:

1. `vanilla` - all data pulled from the website without significant modifications apart from the web page structure parsing;
1. `quotes` - truncated version of the corpus, which does't contain sound effects;
1. `annotated` - an extended version of the full configuration with a couple of additional columns with labels;
1. `pulled` - same as vanilla, but sound files have been pulled from the website, and `source` column is replaced with `sound`.  


## Dataset structure

### Data instance

An example of an entry from the dataset is given below:

```json
{
  "header": "Attack",
  "subheader": "Attacking",
  "text": "Kindred: \"The masks of the Kindred seek you!\"",
  "source": "https://static.wikia.nocookie.net/leagueoflegends/images/1/12/Kindred_Original_Passive_Mark_Enemy_6.ogg/revision/latest?cb=20221204121356",
  "champion": "kindred"
}
```

### Data fields

Each dataset entry therefore consists of the following fields:

- `header` - main category of the text;
- `subheader` - secondary category of the text (none in some cases);
- `text` - text said by the champion or description of sound made by the champion;
- `source` - link to the audio file (only `vanilla` configuration);
- `champion` - name of the champion in lowercase;
- `quote` - binary field displaying whether corresponding text contains quote or not (only `annotated` configuration);
- `sound` - audio data for the entry (only `pulled` configuration).