Supertonic¶

Selects a Supertonic style ID, language, diffusion-step count, and speaking speed.

Text to speechVoiceHub-nativesupertonicParameters: Not reportedLanguages: en, ko +29Training: preprocessedLicense: Checkpoint-specific

Parameter metadata: Not reported: the audited metadata available for the registered default does not provide an exact parameter total.

Use this model Checkpoint

Resources

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: Selects a Supertonic style ID, language, diffusion-step count, and speaking speed.

Inputs and controls: Voice/style IDs and languages are validated against files published by the checkpoint.

from pathlib import Path

from voicehub import AutoModelForTextToSpeech, TTSGenerationConfig

model = AutoModelForTextToSpeech.from_pretrained(
    'Supertone/supertonic-3',
    model_type='supertonic',
    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"),
    ),
    voice="F1",
    language="en",
    total_steps=5,
    speed=1.05,
)
print(output.file_path, output.sample_rate, output.metadata)

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

Overview¶

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

Property	Value
Task	Text to speech
Architecture	`supertonic`
Runtime	`VoiceHub-native`
Languages	`en`, `ko`, `ja`, `ar`, … complete audited list below
Capabilities	`text-to-speech`, `multilingual`, `fine-tuning`, `safetensors`, `voicehub-native`, `native-runtime`, `preprocessed-training`
Reusable components	—
Normalized output	`TTSOutput`

Language support¶

Supported language abbreviations

en, ko, ja, ar, bg, cs, da, de, el, es, et, fi, fr, hi, hr, hu, id, it, lt, lv, nl, pl, pt, ro, ru, sk, sl, sv, tr, uk, vi

Paper and GitHub¶

Paper: No dedicated upstream research paper is published for this integration.
Upstream GitHub: Supertonic
VoiceHub source: VoiceHub model implementation

Configuration¶

Load configuration without constructing the model:

from voicehub import AutoConfig

config = AutoConfig.for_model('supertonic')
print(config.model_type)

Property	Value
Canonical model type	`supertonic`
Configuration class	`SupertonicConfig`
Architecture class	`SupertonicForTextToSpeech`

Processing¶

Create the registered processor without allocating model weights:

from voicehub import AutoProcessor

processor = AutoProcessor.from_pretrained(
    'Supertone/supertonic-3',
    model_type='supertonic',
)
print(type(processor).__name__)

Inference¶

The Usage example returns TTSOutput through AutoModelForTextToSpeech.

Input and output contract¶

Property	Value
Readiness	`preprocessed`
Data architecture	`diffusion`
Sample rate	44,100 Hz
Contract getter	`get_tts_dataset_spec('supertonic')`

Variant	Required fields	One of	Boundary	Other rules
`text-style-object`	`text`, `style`	target_duration / duration / duration_seconds / target_latent / latent / latents	Prepared	—
`text-style-tensors`	`text`, `style_ttl`, `style_dp`	target_duration / duration / duration_seconds / target_latent / latent / latents	Prepared	—
`tokenized-style-object`	`text_ids`, `style`	text_mask / text_lengths; target_duration / duration / duration_seconds / target_latent / latent / latents	Prepared	—
`tokenized-style-tensors`	`text_ids`, `style_ttl`, `style_dp`	text_mask / text_lengths; target_duration / duration / duration_seconds / target_latent / latent / latents	Prepared	—

Conditional flow-matching, rectified-flow, or diffusion 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	`flow-matching`
Recipe	`single-phase`
Default phase	`published_graph`
Training checkpoint	`Supertone/supertonic-3`
Native training graph	`yes`

Phase	Kind	Components	Required inputs	Loss keys
`published_graph`	objective	`model`	`text_ids`, `text_mask`, `style_ttl`, `style_dp`	`loss`, `duration_loss`, `flow_step_loss`, `vocoder_l1_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	`Supertone/supertonic-3`
Hugging Face ID	`Supertone/supertonic-3` 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.supertonic.modeling_supertonic.SupertonicForTextToSpeech`
Configuration	`voicehub.models.supertonic.configuration_supertonic.SupertonicConfig`
Source provenance	`voicehub/models/supertonic/source/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

`SupertonicConfig`¶

View source

SupertonicConfig(**config_kwargs)

Parameters¶

**config_kwargs — Configuration fields validated by SupertonicConfig.

Model

`SupertonicForTextToSpeech`¶

View source

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

Parameters¶

pretrained_model_name_or_path — Hub ID or compatible local directory.
model_type — Canonical model type; use 'supertonic'.
config — Optional preloaded SupertonicConfig instance.
**model_kwargs — Model-specific loading arguments.

from voicehub import get_model_spec

spec = get_model_spec('supertonic')
print(spec.display_name, spec.task.value)

Purpose	Public object
Discover	`get_model_spec('supertonic')`
Load and run	`AutoModelForTextToSpeech`
Configure	`SupertonicConfig`
Process	`AutoProcessor`
Model implementation	`SupertonicForTextToSpeech`
Normalized output	`TTSOutput`
Training contract	`get_training_spec('supertonic')`
Optimization lifecycle	`available_optimization_passes`, `apply_optimization_plan`, `optimization_manifest`, `restore_optimization_plan`

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

Usage¶

Overview¶

Language support¶

Paper and GitHub¶

Configuration¶

Processing¶

Inference¶

Input and output contract¶

Training and optimization¶

Training contract¶

Checkpoints, provenance, license, and limitations¶

Limitations¶

Public API¶

SupertonicConfig¶

Parameters¶

SupertonicForTextToSpeech¶

Parameters¶

`SupertonicConfig`¶

`SupertonicForTextToSpeech`¶