README.md

---
dataset_info:
- config_name: audio_1000h
  features:
  - name: id
    dtype: string
  - name: audio
    dtype: audio
  - name: audio_language
    dtype: string
  - name: text
    dtype: string
  - name: prompt
    dtype: string
  - name: duration
    dtype: float64
  - name: speaker_id
    dtype: string
  splits:
  - name: train
    num_bytes: 26320870654.0
    num_examples: 184453
  download_size: 26306153220
  dataset_size: 26320870654.0
- config_name: audio_100h
  features:
  - name: id
    dtype: string
  - name: audio
    dtype: audio
  - name: audio_language
    dtype: string
  - name: text
    dtype: string
  - name: prompt
    dtype: string
  - name: duration
    dtype: float64
  - name: speaker_id
    dtype: string
  splits:
  - name: train
    num_bytes: 2628415459.0
    num_examples: 18413
  download_size: 2626967984
  dataset_size: 2628415459.0
- config_name: audio_150h
  features:
  - name: id
    dtype: string
  - name: audio
    dtype: audio
  - name: audio_language
    dtype: string
  - name: text
    dtype: string
  - name: prompt
    dtype: string
  - name: duration
    dtype: float64
  - name: speaker_id
    dtype: string
  splits:
  - name: train
    num_bytes: 3942631432.0
    num_examples: 27651
  download_size: 3940433749
  dataset_size: 3942631432.0
- config_name: audio_1h
  features:
  - name: id
    dtype: string
  - name: audio
    dtype: audio
  - name: audio_language
    dtype: string
  - name: text
    dtype: string
  - name: prompt
    dtype: string
  - name: duration
    dtype: float64
  - name: speaker_id
    dtype: string
  splits:
  - name: train
    num_bytes: 26908030.0
    num_examples: 183
  download_size: 26897235
  dataset_size: 26908030.0
- config_name: audio_200h
  features:
  - name: id
    dtype: string
  - name: audio
    dtype: audio
  - name: audio_language
    dtype: string
  - name: text
    dtype: string
  - name: prompt
    dtype: string
  - name: duration
    dtype: float64
  - name: speaker_id
    dtype: string
  splits:
  - name: train
    num_bytes: 5257900558.0
    num_examples: 36892
  download_size: 5254996729
  dataset_size: 5257900558.0
- config_name: audio_500h
  features:
  - name: id
    dtype: string
  - name: audio
    dtype: audio
  - name: audio_language
    dtype: string
  - name: text
    dtype: string
  - name: prompt
    dtype: string
  - name: duration
    dtype: float64
  - name: speaker_id
    dtype: string
  splits:
  - name: train
    num_bytes: 13153141173.0
    num_examples: 92224
  download_size: 13145835785
  dataset_size: 13153141173.0
- config_name: audio_50h
  features:
  - name: id
    dtype: string
  - name: audio
    dtype: audio
  - name: audio_language
    dtype: string
  - name: text
    dtype: string
  - name: prompt
    dtype: string
  - name: duration
    dtype: float64
  - name: speaker_id
    dtype: string
  splits:
  - name: train
    num_bytes: 1314887921.0
    num_examples: 9215
  download_size: 1314161208
  dataset_size: 1314887921.0
- config_name: dev_test
  features:
  - name: id
    dtype: string
  - name: audio
    dtype: audio
  - name: audio_language
    dtype: string
  - name: text
    dtype: string
  - name: prompt
    dtype: string
  - name: duration
    dtype: float64
  - name: speaker_id
    dtype: string
  splits:
  - name: train
    num_bytes: 1321697853
    num_examples: 9263
  download_size: 1320677962
  dataset_size: 1321697853
- config_name: test
  features:
  - name: id
    dtype: string
  - name: audio
    dtype: audio
  - name: audio_language
    dtype: string
  - name: text
    dtype: string
  - name: prompt
    dtype: string
  - name: duration
    dtype: float64
  - name: speaker_id
    dtype: string
  splits:
  - name: train
    num_bytes: 1325866789
    num_examples: 9265
  download_size: 1325634718
  dataset_size: 1325866789
- config_name: full
  features:
  - name: id
    dtype: string
  - name: audio
    dtype: audio
  - name: audio_language
    dtype: string
  - name: text
    dtype: string
  - name: prompt
    dtype: string
  - name: duration
    dtype: float64
  - name: speaker_id
    dtype: string
  splits:
  - name: train
    num_bytes: 36320484372.0
    num_examples: 253469
  download_size: 36299850732
  dataset_size: 36320484372.0
configs:
- config_name: audio_1000h
  data_files:
  - split: train
    path: audio_1000h/train-*
- config_name: audio_100h
  data_files:
  - split: train
    path: audio_100h/train-*
- config_name: audio_150h
  data_files:
  - split: train
    path: audio_150h/train-*
- config_name: audio_1h
  data_files:
  - split: train
    path: audio_1h/train-*
- config_name: audio_200h
  data_files:
  - split: train
    path: audio_200h/train-*
- config_name: audio_500h
  data_files:
  - split: train
    path: audio_500h/train-*
- config_name: audio_50h
  data_files:
  - split: train
    path: audio_50h/train-*
- config_name: dev_test
  data_files:
  - split: train
    path: dev_test/train-*
- config_name: test
  data_files:
  - split: train
    path: test/train-*
- config_name: full
  data_files:
  - split: train
    path: full/train-*
---
# 📚 Kinyarwanda ASR Dataset

This dataset contains transcribed Kinyarwanda audio, designed to support training and evaluation of Automatic Speech Recognition (ASR) systems. It is part of a study on how varying training data volumes affect model performance using Whisper-large-v3.

## 📂 Data Overview

The full dataset consists of approximately 263,000 audio samples covering 5 key domains:

- 🏥 Health  
- 🏛️ Government  
- 💰 Financial Services  
- 🎓 Education  
- 🌾 Agriculture

To enable analysis of scale effects, the dataset is available in subsets of different durations:

- `audio_1h`, `audio_50h`, `audio_100h`, `audio_150h`, `audio_200h`, `audio_500h`, `audio_1000h`, and `train_cleaned` (~1400h)

Each sample includes:
- `id`: a unique identifier for the sample  
- `audio`: the audio recording  
- `audio_language`: the language code (`kin` for Kinyarwanda)  
- `text`: the corresponding transcription in Kinyarwanda  
- `prompt`: the domain or context of the audio (e.g., Health, Education, etc.)  
- `duration`: the length of the audio clip in seconds  
- `speaker_id`: an anonymized identifier for the speaker  

## 🛠️ Usage Example

You can load and explore the dataset using the 🤗 Datasets library:

```python
from datasets import load_dataset

# Load a specific subset, e.g., 100 hours
dataset = load_dataset("evie-8/kinyarwanda-speech-hackathon", "audio_100h")
```