跳转至

ZyphraZONOS2

Zonos2

Uses ZONOS2's language, speed, accurate-mode, and speaker-conditioning controls.

Text to speechVoiceHub-nativezonos2Parameters: 7.7BLanguages: en, zh +32Training: nativeLicense: Checkpoint-specific

Parameter metadata: Exact learned-parameter total for VoiceHub's audited native primary graph at the registered default selection; separately loaded auxiliary models are excluded.

Use this model Checkpoint
Resources
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: Uses ZONOS2's language, speed, accurate-mode, and speaker-conditioning controls.

Inputs and controls: Do not pass both a speaker waveform and a precomputed speaker embedding.

from pathlib import Path

from voicehub import AutoModelForTextToSpeech, TTSGenerationConfig

REFERENCE_AUDIO = Path("reference.wav")
REFERENCE_TEXT = "The reference transcript must exactly match the authorized audio."
if not REFERENCE_AUDIO.is_file():
    raise FileNotFoundError(REFERENCE_AUDIO)

model = AutoModelForTextToSpeech.from_pretrained(
    'Zyphra/ZONOS2',
    model_type='zonos2',
    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"),
    ),
    speaker_audio_path=str(REFERENCE_AUDIO),
    language="en",
    speed=1.0,
    accurate_mode=True,
)
print(output.file_path, output.sample_rate, output.metadata)

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

Overview

zonos2 is a VoiceHub text to speech integration. This page is generated from its registry contract. Open the zonos2 Colab notebook.

Property Value
Task Text to speech
Architecture zonos2
Runtime VoiceHub-native
Languages en, zh, ja, ko, … complete audited list below
Capabilities text-to-speech, voice-cloning, multilingual, fine-tuning, safetensors, voicehub-native, native-runtime
Reusable components dac
Normalized output TTSOutput

Language support

Supported language abbreviations

en, zh, ja, ko, ru, it, pt, fr, es, vi, de, he, nl, sv, hi, ta, te, th, no, bn, tl, ar, da, id, pl, uk, ro, fi, hu, lt, et, sk, hr, lv

Paper and GitHub

Configuration

Load configuration without constructing the model:

from voicehub import AutoConfig

config = AutoConfig.for_model('zonos2')
print(config.model_type)
Property Value
Canonical model type zonos2
Configuration class Zonos2Config
Architecture class Zonos2ForTextToSpeech

Processing

Create the registered processor without allocating model weights:

from voicehub import AutoProcessor

processor = AutoProcessor.from_pretrained(
    'Zyphra/ZONOS2',
    model_type='zonos2',
)
print(type(processor).__name__)

Inference

The Usage example returns TTSOutput through AutoModelForTextToSpeech.

Input and output contract

Property Value
Readiness integrated-raw
Data architecture codec-lm
Sample rate 44,100 Hz
Contract getter get_tts_dataset_spec('zonos2')
Variant Required fields One of Boundary Other rules
raw-audio — text / texts; audio / audio_values Source at most one: text / texts; audio / audio_values; forbidden: audio_codes, input_ids, labels
cached-dac audio_codes text / texts Prepared at most one: text / texts; forbidden: audio, audio_values, input_ids, labels
model-ready input_ids, labels — Prepared forbidden: text, texts, audio, audio_values, audio_codes

Autoregressive text/audio-token or codec-language-model 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 native
Family causal-lm
Recipe single-phase
Default phase reconstructed_codec_language_model
Training checkpoint Zyphra/ZONOS2
Native training graph yes
Phase Kind Components Required inputs Loss keys
reconstructed_codec_language_model objective model input_ids, labels loss

The integration accepts its declared source or prepared contract directly. Call model.validate_training_support() first, then follow the training workflow.

Checkpoints, provenance, license, and limitations

Property Value
Default checkpoint Zyphra/ZONOS2
Hugging Face ID Zyphra/ZONOS2
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.zonos2.modeling_zonos2.Zonos2ForTextToSpeech
Configuration voicehub.models.zonos2.configuration_zonos2.Zonos2Config
Source provenance voicehub/models/zonos2/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

Zonos2Config

View source

Zonos2Config(**config_kwargs)

Parameters

  • **config_kwargs — Configuration fields validated by Zonos2Config.

Model

Zonos2ForTextToSpeech

View source

AutoModelForTextToSpeech.from_pretrained(
    pretrained_model_name_or_path,
    *,
    model_type='zonos2',
    config=None,
    **model_kwargs,
)

Parameters

  • pretrained_model_name_or_path — Hub ID or compatible local directory.
  • model_type — Canonical model type; use 'zonos2'.
  • config — Optional preloaded Zonos2Config instance.
  • **model_kwargs — Model-specific loading arguments.
from voicehub import get_model_spec

spec = get_model_spec('zonos2')
print(spec.display_name, spec.task.value)
Purpose Public object
Discover get_model_spec('zonos2')
Load and run AutoModelForTextToSpeech
Configure Zonos2Config
Process AutoProcessor
Model implementation Zonos2ForTextToSpeech
Normalized output TTSOutput
Training contract get_training_spec('zonos2')
Optimization lifecycle available_optimization_passes, apply_optimization_plan, optimization_manifest, restore_optimization_plan

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