microsoftVibeVoice-ASR-HF

VibeVoice¶

Requests VibeVoice-ASR timestamps with a concise transcription prompt.

Automatic speech recognitionVoiceHub-nativevibevoice-asrParameters: 8.3BLanguages: en, zh +49Training: nativeLicense: Checkpoint-specific

Parameter metadata: Exact serialized tensor-element total from VoiceHub's audited native primary checkpoint; a distinct learned-parameter total is not available.

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: Requests VibeVoice-ASR timestamps with a concise transcription prompt.

Inputs and controls: Keep the prompt task-focused and verify timestamp granularity for the selected checkpoint revision.

from pathlib import Path

from voicehub import AutoModelForSpeechRecognition

AUDIO_FILE = Path("speech.wav")
if not AUDIO_FILE.is_file():
    raise FileNotFoundError(AUDIO_FILE)

model = AutoModelForSpeechRecognition.from_pretrained(
    'microsoft/VibeVoice-ASR-HF',
    model_type='asr_vibevoice',
    device="cuda",
    lazy_load=True,
)
output = model.transcribe(
    AUDIO_FILE,
    return_timestamps=True,
    prompt="Transcribe every spoken turn.",
)
print(output.text)
for segment in output.segments:
    print(segment.start, segment.end, segment.text, segment.confidence)

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

Overview¶

asr_vibevoice is a VoiceHub automatic speech recognition integration. This page is generated from its registry contract. Open the asr_vibevoice Colab notebook.

Property	Value
Task	Automatic speech recognition
Architecture	`vibevoice-asr`
Runtime	`VoiceHub-native`
Languages	`en`, `zh`, `es`, `pt`, … complete audited list below
Capabilities	`automatic-speech-recognition`, `multilingual`, `speaker-attribution`, `timestamps`, `hotwords`, `long-form`, `safetensors`, `fine-tuning`, `voicehub-native`, `native-runtime`
Reusable components	—
Normalized output	`ASROutput`

Language support¶

Supported language abbreviations

en, zh, es, pt, de, ja, ko, fr, ru, id, sv, it, he, nl, pl, no, tr, th, ar, hu, ca, cs, da, fa, af, hi, fi, et, aa, el, ro, vi, bg, is, sl, sk, lt, sw, uk, kl, lv, hr, ne, sr, tl, yi, ms, ur, mn, hy, jv

Paper and GitHub¶

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

Configuration¶

Load configuration without constructing the model:

from voicehub import AutoConfig

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

Property	Value
Canonical model type	`asr_vibevoice`
Configuration class	`VibeVoiceASRConfig`
Architecture class	`VibeVoiceForSpeechRecognition`

Processing¶

Create the registered processor without allocating model weights:

from voicehub import AutoProcessor

processor = AutoProcessor.from_pretrained(
    'microsoft/VibeVoice-ASR-HF',
    model_type='asr_vibevoice',
)
print(type(processor).__name__)

Inference¶

The Usage example returns ASROutput through AutoModelForSpeechRecognition.

Input and output contract¶

Property	Value
Readiness	`integrated-raw`
Data architecture	`prompted-multimodal`
Sample rate	24,000 Hz
Contract getter	`get_asr_dataset_spec('asr_vibevoice')`

Variant	Required fields	One of	Boundary	Other rules
`segmented-audio`	`audio`, `segments`	—	Source	forbidden: text, transcription, transcript
`serialized-audio`	`audio`	text / transcription / transcript	Source	at most one: text / transcription / transcript; forbidden: segments
`vibevoice-model-ready`	`input_ids`, `attention_mask`, `input_values`, `padding_mask`, `labels`	—	Prepared	—

VibeVoice structured long-form ASR targets and multimodal prompt inputs. 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	`speech-sequence-to-sequence`
Recipe	`single-phase`
Default phase	`speech_recognition`
Training checkpoint	`microsoft/VibeVoice-ASR-HF`
Native training graph	`yes`

Phase	Kind	Components	Required inputs	Loss keys
`speech_recognition`	objective	`model.model.multi_modal_projector`, `model.model.language_model`, `model.lm_head`	`input_ids`, `attention_mask`, `input_values`, `padding_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	`microsoft/VibeVoice-ASR-HF`
Hugging Face ID	`microsoft/VibeVoice-ASR-HF` 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.asr_vibevoice.modeling_asr_vibevoice.VibeVoiceForSpeechRecognition`
Configuration	`voicehub.models.asr_vibevoice.configuration_asr_vibevoice.VibeVoiceASRConfig`
Source provenance	`voicehub/architectures/vibevoice/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

`VibeVoiceASRConfig`¶

View source

VibeVoiceASRConfig(**config_kwargs)

Parameters¶

**config_kwargs — Configuration fields validated by VibeVoiceASRConfig.

Model

`VibeVoiceForSpeechRecognition`¶

View source

AutoModelForSpeechRecognition.from_pretrained(
    pretrained_model_name_or_path,
    *,
    model_type='asr_vibevoice',
    config=None,
    **model_kwargs,
)

Parameters¶

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

from voicehub import get_model_spec

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

Purpose	Public object
Discover	`get_model_spec('asr_vibevoice')`
Load and run	`AutoModelForSpeechRecognition`
Configure	`VibeVoiceASRConfig`
Process	`AutoProcessor`
Model implementation	`VibeVoiceForSpeechRecognition`
Normalized output	`ASROutput`
Training contract	`get_training_spec('asr_vibevoice')`
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¶

VibeVoiceASRConfig¶

Parameters¶

VibeVoiceForSpeechRecognition¶

Parameters¶

`VibeVoiceASRConfig`¶

`VibeVoiceForSpeechRecognition`¶