Datasets:

abidanoaman
/

unified-urdu-speech-asr

Tasks:

Automatic Speech Recognition

Languages:

Urdu

Size:

10K<n<100K

License:

Dataset card Files Files and versions

xet

Community

Dataset Viewer

The dataset viewer is not available for this subset.

Cannot get the split names for the config 'default' of the dataset.

Exception:    SplitsNotFoundError
Message:      The split names could not be parsed from the dataset config.
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.12/site-packages/datasets/inspect.py", line 286, in get_dataset_config_info
                  for split_generator in builder._split_generators(
                                         ^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/packaged_modules/folder_based_builder/folder_based_builder.py", line 249, in _split_generators
                  raise ValueError(
              ValueError: `file_name` or `*_file_name` must be present as dictionary key (with type string) in metadata files
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/split_names.py", line 65, in compute_split_names_from_streaming_response
                  for split in get_dataset_split_names(
                               ^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/inspect.py", line 340, in get_dataset_split_names
                  info = get_dataset_config_info(
                         ^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/inspect.py", line 291, in get_dataset_config_info
                  raise SplitsNotFoundError("The split names could not be parsed from the dataset config.") from err
              datasets.inspect.SplitsNotFoundError: The split names could not be parsed from the dataset config.

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

Unified Urdu Speech Dataset for ASR

A balanced Urdu speech dataset combining three high-quality sources for Automatic Speech Recognition.

Dataset Description

This dataset combines:

HuggingFace UrduSpeech (filtered for Urdu only)
Mozilla Common Voice Urdu
Urdu Emotion Voice Dataset

Dataset Statistics

Total Samples: 10,458
Train: 8,366 (80%)
Validation: 1,046 (10%)
Test: 1,046 (10%)
Sample Rate: 16,000 Hz
Format: WAV (mono, normalized)
Balance: 33:33:33 ratio across sources

Dataset Structure

unified-urdu-speech-asr/
├── train/
│   ├── metadata.csv
│   └── [8,366 WAV files]
├── val/
│   ├── metadata.csv
│   └── [1,046 WAV files]
└── test/
    ├── metadata.csv
    └── [1,046 WAV files]

Loading the Dataset

Option 1: Using datasets library

from datasets import load_dataset

# Load from local files
dataset = load_dataset("audiofolder", data_dir="path/to/dataset")

# Or download from HuggingFace Hub
dataset = load_dataset("abidanoaman/unified-urdu-speech-asr")

Option 2: Manual loading

import pandas as pd
import torchaudio

# Load metadata
metadata = pd.read_csv("train/metadata.csv")

# Load audio
for idx, row in metadata.iterrows():
    audio_path = f"train/{row['audio']}"
    waveform, sr = torchaudio.load(audio_path)
    transcript = row['transcript']
    # Your code here...

Metadata Fields

audio: Audio filename (WAV, 16kHz, mono)
transcript: Urdu text transcription
source: Data source (HuggingFace, CommonVoice, EmotionVoice)
emotion: Emotion label (only for EmotionVoice samples)

Source Distribution

Source	Samples	Percentage
HuggingFace	3,486	33.3%
CommonVoice	3,486	33.3%
EmotionVoice	3,486	33.3%

Use Cases

Automatic Speech Recognition (ASR)
Speech-to-Text
Urdu language modeling
Multi-speaker recognition
Emotion recognition (subset)

Citation

If you use this dataset, please cite the original sources:

HuggingFace UrduSpeech: humairawan/UrduSpeech
Mozilla Common Voice
Urdu Emotion Voice Dataset

License

Please check individual source licenses for terms of use.

Dataset Creation

This dataset was created by:

Filtering Urdu samples from HuggingFace UrduSpeech (26,102 samples)
Loading Mozilla Common Voice Urdu (3,486 samples)
Adding Urdu Emotion Voice for diversity (3,486 samples)
Balancing to 33:33:33 ratio (3,486 from each)
Normalizing to 16kHz mono WAV format
Splitting into 80/10/10 train/val/test

Contact

For questions or issues, please open an issue on the dataset repository.

Downloads last month: 23

Total file size:

1.59 GB