README.md

metadata

language:
  - en
license: mit
task_categories:
  - automatic-speech-recognition
size_categories:
  - 10K<n<100K
dataset_info:
  features:
    - name: audio
      dtype:
        audio:
          sampling_rate: 16000
    - name: text
      dtype: string
    - name: chunk_id
      dtype: string
    - name: original_file
      dtype: string
    - name: start_time
      dtype: float64
    - name: end_time
      dtype: float64
    - name: speaker_id
      dtype: string
    - name: domain
      dtype: string
    - name: noise_conditions
      dtype: string
    - name: language
      dtype: string
  splits:
    - name: train_0
      num_bytes: 14241773830.618
      num_examples: 36239
    - name: train_1
      num_bytes: 14241773830.618
      num_examples: 36239
    - name: train_2
      num_bytes: 14241773830.618
      num_examples: 36239
    - name: train_3
      num_bytes: 14241773830.618
      num_examples: 36239
    - name: train_4
      num_bytes: 14241773830.618
      num_examples: 36239
    - name: train_5
      num_bytes: 14241773830.618
      num_examples: 36239
    - name: train_6
      num_bytes: 14241773830.618
      num_examples: 36239
    - name: train_7
      num_bytes: 14241773830.618
      num_examples: 36239
    - name: train_8
      num_bytes: 14241773830.618
      num_examples: 36239
    - name: train_9
      num_bytes: 14241773830.618
      num_examples: 36239
    - name: train_10
      num_bytes: 14241773830.618
      num_examples: 36239
    - name: train_11
      num_bytes: 14241773830.618
      num_examples: 36239
  download_size: 161139820068
  dataset_size: 170901285967.41602
configs:
  - config_name: default
    data_files:
      - split: train_0
        path: data/train_0-*
      - split: train_1
        path: data/train_1-*
      - split: train_2
        path: data/train_2-*
      - split: train_3
        path: data/train_3-*
      - split: train_4
        path: data/train_4-*
      - split: train_5
        path: data/train_5-*
      - split: train_6
        path: data/train_6-*
      - split: train_7
        path: data/train_7-*
      - split: train_8
        path: data/train_8-*
      - split: train_9
        path: data/train_9-*
      - split: train_10
        path: data/train_10-*
      - split: train_11
        path: data/train_11-*

Streaming ASR Dataset

This dataset is designed for training real-time (streaming) ASR models, with a focus on handling chunk-based audio processing. It contains standardized audio segments from LibriSpeech dev-clean, processed for streaming ASR applications.

Dataset Description

Dataset Summary

• Source: LibriSpeech dev-clean
• Total chunks: 2,703
• Total duration: ~20 hours (1,212.26 seconds)
• Unique speakers: 40
• Audio format: 16 kHz mono WAV
• Language: English
• Domain: Audiobooks (clean speech)

Dataset Structure

openwhisper/
├── chunks/          # Audio files (16kHz mono WAV)
├── transcripts/     # Text transcriptions
└── metadata/        # JSON files with detailed information

Data Fields

Each sample consists of:

1. Audio file (WAV)

   • 16 kHz sampling rate
   • Mono channel
   • 16-bit PCM format

2. Transcript file (TXT)

   • Clean text transcription
   • Includes punctuation and casing
   • Aligned with audio chunks

3. Metadata file (JSON)

   • speaker_id: Unique speaker identifier
   • chunk_id: Unique chunk identifier
   • start_time: Start time in original audio
   • end_time: End time in original audio
   • duration: Chunk duration in seconds
   • language: Language code (en)
   • noise_conditions: Audio quality label (clean)
   • original_file: Source file reference

Data Splits

This dataset contains only the dev-clean portion of LibriSpeech, processed into overlapping chunks suitable for streaming ASR training.

Dataset Creation

Preprocessing

1. Audio standardization

   • Resampling to 16 kHz
   • Conversion to mono channel
   • Format conversion to WAV

2. Chunking strategy

   • Fixed chunk duration with overlap
   • Natural pause boundary detection
   • Consistent chunk size for training stability

3. Transcript processing

   • Alignment with audio chunks
   • Preservation of punctuation and casing
   • Clean text normalization

Usage

Loading the Dataset

from datasets import load_dataset

dataset = load_dataset("orgho98/openwhisper")

Training Example

# Example code for loading audio and transcript pairs
for sample in dataset:
    audio = sample['audio']
    transcript = sample['text']
    metadata = sample['metadata']
    
    # Process for streaming ASR training
    # ...

License

This dataset is released under the MIT License, following LibriSpeech's licensing terms.

Citation

If you use this dataset, please cite:

@misc{openwhisper2024,
  title={Streaming ASR Dataset},
  author={Automagically AI},
  year={2024},
  publisher={Hugging Face},
  howpublished={\url{https://huggingface.co/datasets/orgho98/openwhisper}}
}

Limitations

• Limited to clean speech from audiobooks
• Single language (English)
• May not represent real-world streaming conditions perfectly

Additional Information

• Curated by: Automagically AI
• License: MIT
• Version: 1.0.0