Skip to content

Dataset · MiniMaxAI

Multilingual test set for voice cloning

MiniMaxAI/TTS-Multilingual-Test-Set

Speaker audio samples and text phrases in 24 languages for zero-shot speech synthesis evaluation by MiniMaxAI.

What's inside

The test set, sized under 1 thousand records, includes audio and texts across 24 languages. For each language, 100 unique validation sentences and reference recordings of two speakers (male and female) selected from Mozilla Common Voice are prepared.

What you can do with it

  • Evaluating zero-shot voice cloning quality in TTS models
  • Comparing multilingual speech synthesis systems on a unified benchmark
Tasks
Text → speech
Size
up to 1K rows
Downloads this month
2K
Likes
47

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("MiniMaxAI/TTS-Multilingual-Test-Set")