Datasets:
100 Speech-to-Text / ASR Shruk Dataset
Dataset Description
Traditional Kashmiri poetic verses (Shruks) paired with their Kashmiri script transcriptions. Designed for training and evaluating Automatic Speech Recognition (ASR) / Speech-to-Text (STT) models for the Kashmiri language.
Dataset Structure
| Field | Type | Description |
|---|---|---|
audio |
Audio | WAV recording of the Shruk |
transcription |
string | Kashmiri script transcription |
shruk_number |
int | Original shruk number (1–100) |
category |
string | Thematic category |
tags |
string | Pipe-separated semantic tags |
Statistics
- Total samples: 99
- Audio format: WAV
- Language: Kashmiri (ks)
Usage
from datasets import load_dataset
# Load locally
ds = load_dataset("audiofolder", data_dir="Omarrran/100_shruk-speech_to_Text__ASR_dataset")
print(ds["train"][0])
- Downloads last month
- 8