Dataset Viewer
The dataset viewer is not available for this dataset.
Unexpected token '<', "<html> <h"... is not valid JSON

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.

IndicVoices Cleaned

IndicVoices Cleaned is a curated repository containing cleaned transcripts derived from the IndicVoices dataset. We start with normalized verbatim text from IndicVoices and further process it using Google's Gemini model to produce grammatically correct and fluent sentences.

These cleaned sentences are intended for use in creating high-quality datasets for:

  • Machine translation
  • Transliteration
  • Other multilingual NLP tasks

Supported Languages

Primary Languages

We offer robust support for the following 15 major Indian languages:

Assamese, Bengali, Gujarati, Hindi, Kannada, Maithili, Malayalam, Marathi, Nepali, Odia, Punjabi, Sanskrit, Tamil, Telugu, Urdu

Low-Resource Languages

We also provide preliminary support for the following 7 low-resource Indian languages:

Bodo, Dogri, Kashmiri, Konkani, Manipuri, Santali, Sindhi

Downloads last month
28