콘텐츠로 이동

OpenMOSS-TeamMOSS-TTS-v1.5

MossTTS

Combines MOSS-TTS language, instruction, and quality controls without importing upstream demo code.

Text to speechVoiceHub-nativemoss-ttsParameters: 8.5BLanguages: zh, yue +29Training: 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: Combines MOSS-TTS language, instruction, and quality controls without importing upstream demo code.

Inputs and controls: Keep instructions descriptive and validate the requested language against the selected checkpoint.

from pathlib import Path

from voicehub import AutoModelForTextToSpeech, TTSGenerationConfig

model = AutoModelForTextToSpeech.from_pretrained(
    'OpenMOSS-Team/MOSS-TTS-v1.5',
    model_type='mosstts',
    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"),
    ),
    language="en",
    instruction="Calm, clear studio speech",
    quality="high",
)
print(output.file_path, output.sample_rate, output.metadata)

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

Overview

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

Property Value
Task Text to speech
Architecture moss-tts
Runtime VoiceHub-native
Languages zh, yue, en, ar, … complete audited list below
Capabilities text-to-speech, voice-cloning, multilingual, fine-tuning, safetensors, voicehub-native, native-runtime, delay-variant, local-variant, local-v1.5-variant, realtime-variant, raw-audio-fine-tuning, preencoded-rvq-fine-tuning, native-codec-v1, native-codec-v2, buffered-generation
Reusable components —
Normalized output TTSOutput

Language support

Supported language abbreviations

zh, yue, en, ar, cs, da, de, nl, es, fr, fi, el, he, hi, hu, ja, it, ko, mk, ms, ru, fa, pl, pt, sv, ro, sw, tl, th, tr, vi

Paper and GitHub

Configuration

Load configuration without constructing the model:

from voicehub import AutoConfig

config = AutoConfig.for_model('mosstts')
print(config.model_type)
Property Value
Canonical model type mosstts
Configuration class MossTTSConfig
Architecture class MossTTSForTextToSpeech

Processing

Create the registered processor without allocating model weights:

from voicehub import AutoProcessor

processor = AutoProcessor.from_pretrained(
    'OpenMOSS-Team/MOSS-TTS-v1.5',
    model_type='mosstts',
)
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 Model/checkpoint specific
Contract getter get_tts_dataset_spec('mosstts')
Variant Required fields One of Boundary Other rules
raw-audio text audio / waveform / audio_path Source at most one: audio / waveform / audio_path; forbidden: speech_tokens
preencoded-rvq text, speech_tokens — Prepared forbidden: audio, waveform, audio_path

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 semantic_language_model
Training checkpoint OpenMOSS-Team/MOSS-TTS-v1.5
Native training graph yes
Phase Kind Components Required inputs Loss keys
semantic_language_model objective model input_ids, attention_mask, 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 OpenMOSS-Team/MOSS-TTS-v1.5
Hugging Face ID OpenMOSS-Team/MOSS-TTS-v1.5
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.mosstts.modeling_mosstts.MossTTSForTextToSpeech
Configuration voicehub.models.mosstts.configuration_mosstts.MossTTSConfig
Source provenance voicehub/models/mosstts/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

MossTTSConfig

View source

MossTTSConfig(**config_kwargs)

Parameters

  • **config_kwargs — Configuration fields validated by MossTTSConfig.

Model

MossTTSForTextToSpeech

View source

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

Parameters

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

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

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