metadata
language:
- hi
- en
license: mit
task_categories:
- automatic-speech-recognition
pretty_name: Hindi-English ASR Dataset
viewer: true
Dataset Card for harshk04/h
This dataset contains bilingual Hindi and English audio-text pairs formatted for automatic speech recognition (ASR) fine-tuning.
Dataset Overview
- Languages: Hindi and English
- Samples: 400
- Split: train
- License: MIT
Columns
| Column | Type | Description |
|---|---|---|
| audio | Audio | Path to the audio file |
| language | string | Primary language of speech |
| languagesKnown | string | List of other known languages |
| gender | string | Speaker gender |
| state | string | Indian state name |
| district | string | District name |
| pincode | int64 | Pincode of speaker location |
| stay(years) | string | Duration of stay at location |
| isTranscriptionAvailable | string | Whether transcription is available |
| transcript | string | Text transcript |
| referenceImage | string | Path to reference image (if any) |
Maintainer: Harsh K (@harshk04)