Dataset · laion
4M Cutscene Voice Acting Prompts
laion/voice-acting-cutscene-prompts
English and German text scripts with sudden emotional shifts for a single character to train expressive TTS models.
What's inside
The collection contains 4,057,000 generated text prompts in English and German without audio files. Each example describes a single character in two contrasting emotional scenes separated by a CUT TO: transition, with lines in quotes and delivery directions in parentheses.
What you can do with it
- Training and testing emotional and expressive text-to-speech (TTS) models
- Fine-tuning LLMs to generate script stage directions and dialogue for voice acting
- Tasks
- Text generation
- Languages
- English, German
- Size
- 1M–10M rows
- Downloads this month
- 273.7K
- Likes
- 2
How to load it
from datasets import load_dataset
ds = load_dataset("laion/voice-acting-cutscene-prompts")