Ana içeriğe geç

facebookmms-tts-eng

Vits

Controls MMS-VITS speaking rate, stochastic duration, and output-frame guardrails.

Text to speechVoiceHub-nativevitsParameters: 36.3MLanguage: enTraining: preprocessedLicense: Checkpoint-specific

Parameter metadata: Exact Safetensors total reported by the Hugging Face model API for the registered default checkpoint, retrieved 2026-08-13.

Use this model Checkpoint
Resources
Paper Upstream GitHub VoiceHub source Open in Colab

Usage

Complete the VoiceHub installation once, then run this repository-authored example. Model pages intentionally contain no package-install command.

This example is maintained against VoiceHub's public API; it is not copied from an upstream demo or package README.

Model-specific path: Controls MMS-VITS speaking rate, stochastic duration, and output-frame guardrails.

Inputs and controls: The registered English checkpoint is single-speaker; choose a different HF ID for another MMS language.

from pathlib import Path

from voicehub import AutoModelForTextToSpeech, TTSGenerationConfig

model = AutoModelForTextToSpeech.from_pretrained(
    'facebook/mms-tts-eng',
    model_type='vits',
    device="cuda",
    lazy_load=True,
)
output = model.generate(
    'VoiceHub keeps model integrations explicit and reproducible.',
    generation_config=TTSGenerationConfig(
        seed=42,
        output_file=Path("output.wav"),
    ),
    speaking_rate=1.0,
    noise_scale=0.667,
    max_output_frames=240_000,
)
print(output.file_path, output.sample_rate, output.metadata)

Use authorized recordings. Verify hardware needs and pin a revision in production.

Overview

vits is a VoiceHub text to speech integration. This page is generated from its registry contract. Open the vits Colab notebook.

Property Value
Task Text to speech
Architecture vits
Runtime VoiceHub-native
Languages en
Capabilities text-to-speech, multilingual, mms-tts, safetensors, fine-tuning, voicehub-native, native-runtime, raw-audio-training, preprocessed-training, adversarial-training, generator-warm-start, explicit-acoustic-training-config
Reusable components —
Normalized output TTSOutput

Language support

Supported language abbreviations

en

Paper and GitHub

Configuration

Load configuration without constructing the model:

from voicehub import AutoConfig

config = AutoConfig.for_model('vits')
print(config.model_type)
Property Value
Canonical model type vits
Configuration class VitsConfig
Architecture class VitsForTextToSpeech

Processing

Create the registered processor without allocating model weights:

from voicehub import AutoProcessor

processor = AutoProcessor.from_pretrained(
    'facebook/mms-tts-eng',
    model_type='vits',
)
print(type(processor).__name__)

Inference

The Usage example returns TTSOutput through AutoModelForTextToSpeech.

Input and output contract

Property Value
Readiness integrated-raw
Data architecture vits
Sample rate Model/checkpoint specific
Contract getter get_tts_dataset_spec('vits')
Variant Required fields One of Boundary Other rules
raw-adversarial text audio / audio_values Source —
tokenized-raw-adversarial input_ids audio / audio_values Source —
precomputed-spectrogram spectrogram text / input_ids; audio / audio_values Prepared —

VITS/GAN text, waveform, spectrogram, and adversarial data. See the data workflow.

Training and optimization

Use available_optimization_passes() to discover reversible public passes. Unsupported runtime or hardware fails closed before mutation.

Training contract

Property Value
Support preprocessed
Family vits
Recipe adversarial
Default phase generator
Training checkpoint facebook/mms-tts-eng
Native training graph yes
Phase Kind Components Required inputs Loss keys
discriminator discriminator training_model.discriminator input_ids, audio_values loss
generator generator training_model.native_model input_ids, audio_values loss

Prepare the exact tensors listed in the data contract before this step. Call model.validate_training_support() first, then follow the training workflow.

Checkpoints, provenance, license, and limitations

Property Value
Default checkpoint facebook/mms-tts-eng
Hugging Face ID facebook/mms-tts-eng
Repository availability verified through the Hugging Face model API on 2026-08-11; pin a revision before production use.
Checkpoint status Registry default; pin an immutable revision for production and reproducible evidence
Optional dependency extra Core package
Hardware and runtime Usage selects cuda; verify checkpoint-specific requirements
Real-checkpoint evidence Release evidence; a registry default alone is not execution evidence
Implementation voicehub.models.vits.modeling_vits.VitsForTextToSpeech
Configuration voicehub.models.vits.configuration_vits.VitsConfig
Source provenance voicehub/architectures/vits/SOURCE.json
License Checkpoint-specific

No VoiceHub-specific license override is registered. Verify the checkpoint and upstream source terms before use.

Confirm the checkpoint revision, access terms, provenance, and license.

Limitations

  • No integration-specific checkpoint limitation is registered. Verify the selected checkpoint revision and its documented runtime requirements.
  • Validate memory, precision, and optional dependencies on the target system.
  • Public optimizations fail closed when the runtime or hardware cannot satisfy their validation contract; an unavailable pass is not reported as applied.
  • Contract tests do not replace the linked released-checkpoint evidence.

Public API

Use the stable configuration, processor, and task-model facades below.

Configuration

VitsConfig

View source

VitsConfig(**config_kwargs)

Parameters

  • **config_kwargs — Configuration fields validated by VitsConfig.

Model

VitsForTextToSpeech

View source

AutoModelForTextToSpeech.from_pretrained(
    pretrained_model_name_or_path,
    *,
    model_type='vits',
    config=None,
    **model_kwargs,
)

Parameters

  • pretrained_model_name_or_path — Hub ID or compatible local directory.
  • model_type — Canonical model type; use 'vits'.
  • config — Optional preloaded VitsConfig instance.
  • **model_kwargs — Model-specific loading arguments.
from voicehub import get_model_spec

spec = get_model_spec('vits')
print(spec.display_name, spec.task.value)
Purpose Public object
Discover get_model_spec('vits')
Load and run AutoModelForTextToSpeech
Configure VitsConfig
Process AutoProcessor
Model implementation VitsForTextToSpeech
Normalized output TTSOutput
Training contract get_training_spec('vits')
Optimization lifecycle available_optimization_passes, apply_optimization_plan, optimization_manifest, restore_optimization_plan

See all model guides, inference, and the training matrix.