ekacare/vistaar_small_asr_eval

Vistaar Small ASR Eval

Dataset Description

The Vistaar Small ASR Eval dataset is a multilingual automatic speech recognition evaluation dataset containing 9,486 audio samples across 12 Indian languages. This dataset represents a subset of the larger Vistaar dataset published by AI4Bharat, designed specifically for evaluating ASR model performance on diverse Indian language speech data. A smaller evaluation dataset was created for the use-cases where a quick benchmarking of models is needed. Currently, Vistaar can be accessed through github only, we are redistributing this subset of the dataset through Huggingface for easier usage under the same MIT Liscence.

Original Vistaar dataset and models can be found here - https://github.com/AI4Bharat/vistaar

Dataset Statistics

• Total samples: 9,486
• Languages: 12 Indian languages
• Total duration: ~18.6 hours
• Use case: ASR model evaluation and benchmarking

Dataset Structure

Data Subsets

The dataset comprises six distinct subsets, each contributing unique characteristics to the overall evaluation framework:

Subset	Samples	Duration
Kathbath	1,962	3.7 hrs
Kathbath Hard	1,962	3.7 hrs
MUCS	1,786	2.5 hrs
Common Voice	1,775	2.8 hrs
FLEURS	1,101	3.7 hrs
IndicTTS	900	2.2 hrs

Supported Languages

The dataset covers 12 Indian languages, providing comprehensive coverage for multilingual ASR evaluation across the Indian subcontinent.

Usage

Loading the Dataset

from datasets import load_dataset

# Load a specific language subset and split
dataset = load_dataset('ekacare/vistaar_small_asr_eval', 'bn', split='test')

# Load all splits for a specific language
dataset = load_dataset('ekacare/vistaar_small_asr_eval', 'bn')

# Load the complete dataset (all languages and subsets)
dataset = load_dataset('ekacare/vistaar_small_asr_eval')