Datasets:

Aybee5
/

hausa-tts

language:
  - ha
license: cc-by-4.0
task_categories:
  - text-to-speech
tags:
  - audio
  - hausa
  - tts
  - speech-synthesis
  - multi-speaker
size_categories:
  - 1K<n<10K
dataset_info:
  features:
    - name: text
      dtype: string
    - name: source
      dtype: string
    - name: audio
      dtype: audio
configs:
  - config_name: default
    data_files:
      - split: train
        path: hausa_tts_embedded.parquet
viewer: true

Hausa TTS Dataset

Multi-speaker Hausa text-to-speech dataset.

Usage

from datasets import load_dataset, Audio

# Load the dataset
ds = load_dataset("parquet", data_files="hausa_tts_embedded.parquet", split="train")

# Cast audio column to Audio type with 24kHz sampling rate
ds = ds.cast_column("audio", Audio(sampling_rate=24000))

# Use with Unsloth TTS
from transformers import AutoProcessor
processor = AutoProcessor.from_pretrained("unsloth/csm-1b")

# Your training code here...

Dataset Structure

text: The Hausa text to be spoken
source: Speaker identifier (multi-speaker)
audio: Audio recording (embedded as bytes)