PeterPatrick/swahili-tts-rose

language:
- sw
license: apache-2.0
task_categories:
- text-to-speech
task_ids:
- speech-synthesis
tags:
- swahili
- tts
- speech-synthesis
- audio
- kiswahili
- rose
- female-voice
---

# Swahili TTS Dataset - Rose

A Swahili text-to-speech dataset featuring 500 high-quality audio recordings by female speaker Rose.

## Dataset Description

A comprehensive Swahili text-to-speech dataset featuring **500 high-quality audio recordings** by a female speaker named Rose.

### Dataset Summary

- **Total Samples**: 500
- **Train Samples**: 450
- **Validation Samples**: 50
- **Language**: Swahili (sw)
- **Speaker**: Female (Rose)

### Text Statistics

- **Average text length**: 22.9 characters
- **Shortest text**: 12 characters
- **Longest text**: 63 characters
- **Average words per sentence**: 3.4 words

### Sample Texts

- "Hujambo Mimi ni Rose Ninaweza kukusaidiaje leo"

• "Kutafuta mikahawa iliyo karibu"

• "Kufuta noti ya mwisho"

• "Kuzima alamu"

• "Kusoma vichwa vya habari"

• "Kukagua muunganisho wa internet"

• "Kufuta barua pepe ya mwisho"

• "Kusinzia alamu"

• "Kuanza video tena"

• "Kutengeneza noti mpya"

• "Kuweka picha kwenye instagramu"

• "Anza kurekodi video"

• "Mode ya giza imewashwa"

• "Ninasoma ujumbe wako wa mwisho"

• "Hujambo Mimi ni Rose Ninaweza kukusaidiaje leo"

  Usage

  from datasets import load_dataset
  
  dataset = load_dataset("PeterPatrick/swahili-tts-rose")
  train_dataset = dataset['train']
  val_dataset = dataset['validation']