shenava
Datasets
All datasets matching “shenava”shenava-koochik-number-stress-36k
Shenava Koochik Number Stress 36K
This repository contains the 36,000-row Persian number-stress text curriculum and
an audio-backed subset of 6,669 short clips generated with Gemini TTS.
The Dataset Viewer default configuration is the 6,669-clip audio subset. It
exposes audio, text, category, and duration_s; the metadata links each
row to its WAV using a relative file_name such as audio/num-000000.wav.
The complete text-only source remains available at
data/train.jsonl (36,000… See the full description on the dataset page: https://huggingface.co/datasets/Reza2kn/shenava-koochik-number-stress-36k.shenava-phaseb-processedshenava-jobs-scratch
shenava-jobs-scratch