Back: Catalog

Parler TTS

Parler-TTS is a light-based text-sound (TTS) model that converts text to sound by mimicking the style properties of a particular speaker. Presented under an open source license of data sets, training codes and weights, this project houses the infrastructure needed for both inference and model training. SDPA has been accelerated production processes thanks to Flash Attention 2 compatibility and model compilation support.

PA

Parler TTS

55

huggingface · Voice & Text-to-Speech

STARS5.6K
FORKS588
COMMITS199
CONTRIBUTORS13
VERSION—
UPDATED1 yr ago
PythonLLM-BasedMetin PromptKontrolApache-2.0

Common commercial products in this category: ElevenLabs, Amazon Polly, Google Cloud Text-to-Speech

What can it do?
  • You convert text entries into audio with text commands that explain a specific speaker style and tone of voice.
  • Using the provided codebase and model weights, you train your own text-sound models.
  • You make faster audio productions using Flash Attention 2 integration and model compilation features.
Heads up
  • • Last update over 1 year ago
User reviews
  • • NeutralGitHub Issue

    Voice Consistency Working Pretty Well -- Plus Zero-Shot Cloning!

    “I've managed to get a POC with voice consistency working pretty well. Along the way, I've figured out how to do _ok-ish_ zero-shot voice cloning, too. It took drawing on tidbits spread between several issues posted here, the HF repos, the various github sources linked here and…”

  • • NeutralGitHub Issue

    Does not work on macOS with device="mps": "Can't infer missing attention mask on `mps` device"

    “This is my simple test script: ```python import torch from parler_tts import ParlerTTSForConditionalGeneration from transformers import AutoTokenizer import soundfile as sf torch_device = "mps:0" torch_dtype = torch.bfloat16 model_name = "parler-tts/parler-tts-mini-v1"…”

Website GitHub

My note

Alternative pages

ElevenLabs alternativesAmazon Polly alternativesGoogle Cloud Text-to-Speech alternatives

More in: Voice & Text-to-Speech