Skip to content

Dataset · laion

Synthetic voiceover with emotions and vocal bursts

laion/laions_got_talent

Generated audio-text dataset of acting voiceovers with various emotions for training and evaluating speech models.

What's inside

The dataset contains synthetic audio recordings of acting voiceovers and associated text, ranging from 100 thousand to 1 million examples. The recordings cover a wide range of emotions, vocal bursts, and topics. The data was created as part of the BUD-E project by LAION with support from Intel.

What you can do with it

  • Training speech emotion recognition models
  • Training emotional text-to-speech (TTS)
  • Research and SFT of voice assistants
Size
100K–1M rows
Downloads this month
5.3K
Likes
42

How to load it

from datasets import load_dataset

ds = load_dataset("laion/laions_got_talent")