README.md

---
dataset_info:
  features:
  - name: index
    dtype: int64
  - name: tokens
    sequence: int64
  - name: text
    dtype: string
  splits:
  - name: train
    num_bytes: 19886519752
    num_examples: 2420047
  download_size: 3660752702
  dataset_size: 19886519752
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
license: mit
task_categories:
- automatic-speech-recognition
language:
- en
tags:
- audio2text
- multimodal model
size_categories:
- 1M<n<10M
---

## Dataset Overview 

This dataset contains over 2,4M English ASR samples, using:
- The a training set of [parler-tts/mls_eng_10k](https://huggingface.co/datasets/parler-tts/mls_eng_10k)
- Tokenized using [WhisperVQ](https://huggingface.co/WhisperSpeech/WhisperSpeech/blob/main/whisper-vq-stoks-medium-en%2Bpl.model).

## Usage

```python
from datasets import load_dataset, Audio
# Load Instruction Speech dataset

dataset = load_dataset("homebrewltd/raw-speech-whispervq-v1",split='train')
```

## Dataset Fields

 Field             | Type       | Description                                      |
|------------------|------------|--------------------------------------------------|
| `tokens`         | sequence   | Tokenized using Encodec                          |
| `text`           | sequence   | Converted audio tokens                           |

## Bias, Risks, and Limitations

- Dataset may reflect biases inherent in its source.
- Current version lacks quality control for prompts and responses.
- The usage of Encodec may compromise sound tokens quality.
- Users should consider these limitations when applying the dataset.

## Licensing Information

The dataset is released under the [MIT license](https://opensource.org/license/MIT).

## Citation Information

```
@article{Instruction Speech 2024,
  title={Instruction Speech},
  author={JanAI},
  year=2024,
  month=June},
  url={https://huggingface.co/datasets/jan-hq/instruction-speech}
```