Model
Inference & Serving
Parler-TTS
1Text-to-speech 2model 3controlled 4by 5text 6prompts
the six Ws · specification
W1
Who
Hugging Face, building on research by Dan Lyth and Simon King, released Parler-TTS.
W2
What
Parler-TTS is a text-to-speech model whose voice style, pitch, and recording quality can be steered with a natural language description prompt.
W3
Where
Hosted on GitHub at huggingface/parler-tts and on the Hugging Face Hub.
W4
When
Released in 2024, with Parler-TTS Mini and Large versions following the original paper.
W5
Why
It gives developers fine-grained, prompt-based control over synthetic voice characteristics without recording new training data per voice.
W6
With
Built on PyTorch and integrates with the Hugging Face transformers and datasets libraries.
W7
Watch
Released under the Apache 2.0 license by Hugging Face; fully open weights.
alternatives
works with