Ir para o conteúdo

facebookseamless-m4t-v2-large

SeamlessM4Tv2

Selects SeamlessM4T v2 transcription rather than speech translation.

Automatic speech recognitionVoiceHub-nativeseamless-m4t-v2-s2tParameters: 1.5BLanguages: afr, amh +96Training: nativeLicense: CC-BY-NC-4.0

Parameter metadata: Exact parameter total for the speech-to-text subset loaded from VoiceHub's audited unified default checkpoint.

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: Selects SeamlessM4T v2 transcription rather than speech translation.

Inputs and controls: The native complete-waveform path is greedy and does not claim timestamp alignment.

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(
    'facebook/seamless-m4t-v2-large',
    model_type='asr_seamless_m4t_v2',
    device="cuda",
    lazy_load=True,
)
output = model.transcribe(
    AUDIO_FILE,
    task="transcribe",
    num_beams=1,
    max_new_tokens=256,
)
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_seamless_m4t_v2 is a VoiceHub automatic speech recognition integration. This page is generated from its registry contract. Open the asr_seamless_m4t_v2 Colab notebook.

Property Value
Task Automatic speech recognition
Architecture seamless-m4t-v2-s2t
Runtime VoiceHub-native
Languages afr, amh, arb, ary, … complete audited list below
Capabilities automatic-speech-recognition, multilingual, safetensors, fine-tuning, voicehub-native, native-runtime, greedy-decoding, full-model-training
Reusable components —
Normalized output ASROutput

Language support

Supported language abbreviations

afr, amh, arb, ary, arz, asm, azj, bel, ben, bos, bul, cat, ceb, ces, ckb, cmn, cmn_Hant, cym, dan, deu, ell, eng, est, eus, fin, fra, fuv, gaz, gle, glg, guj, heb, hin, hrv, hun, hye, ibo, ind, isl, ita, jav, jpn, kan, kat, kaz, khk, khm, kir, kor, lao, lit, lug, luo, lvs, mai, mal, mar, mkd, mlt, mni, mya, nld, nno, nob, npi, nya, ory, pan, pbt, pes, pol, por, ron, rus, sat, slk, slv, sna, snd, som, spa, srp, swe, swh, tam, tel, tgk, tgl, tha, tur, ukr, urd, uzn, vie, yor, yue, zlm, zul

These are output-language prompts supported by the audited S2T checkpoint.

Paper and GitHub

Configuration

Load configuration without constructing the model:

from voicehub import AutoConfig

config = AutoConfig.for_model('asr_seamless_m4t_v2')
print(config.model_type)
Property Value
Canonical model type asr_seamless_m4t_v2
Configuration class SeamlessM4Tv2ASRConfig
Architecture class SeamlessM4Tv2ForSpeechRecognition

Processing

Create the registered processor without allocating model weights:

from voicehub import AutoProcessor

processor = AutoProcessor.from_pretrained(
    'facebook/seamless-m4t-v2-large',
    model_type='asr_seamless_m4t_v2',
)
print(type(processor).__name__)

Inference

The Usage example returns ASROutput through AutoModelForSpeechRecognition.

Input and output contract

Property Value
Readiness integrated-raw
Data architecture speech-sequence-to-sequence
Sample rate 16,000 Hz
Contract getter get_asr_dataset_spec('asr_seamless_m4t_v2')
Variant Required fields One of Boundary Other rules
raw-audio audio text / transcription / transcript Source at most one: text / transcription / transcript
seamless-model-ready input_features, attention_mask, labels — Prepared —

SeamlessM4T-v2 multilingual speech-to-text records. 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 facebook/seamless-m4t-v2-large
Native training graph yes
Phase Kind Components Required inputs Loss keys
speech_recognition objective model.speech_encoder, model.text_decoder, model.shared, model.lm_head input_features, 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 facebook/seamless-m4t-v2-large
Hugging Face ID facebook/seamless-m4t-v2-large
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_seamless_m4t_v2.modeling_asr_seamless_m4t_v2.SeamlessM4Tv2ForSpeechRecognition
Configuration voicehub.models.asr_seamless_m4t_v2.configuration_asr_seamless_m4t_v2.SeamlessM4Tv2ASRConfig
Source provenance voicehub/architectures/seamless_m4t_v2/SOURCE.json
License CC-BY-NC-4.0

The pinned SeamlessM4T-v2 Large checkpoint and fine-tuned derivatives are non-commercial under CC-BY-NC-4.0. The VoiceHub-native S2T architecture port is audited against Apache-2.0 Transformers source. Commercial use: not allowed.

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

SeamlessM4Tv2ASRConfig

View source

SeamlessM4Tv2ASRConfig(**config_kwargs)

Parameters

  • **config_kwargs — Configuration fields validated by SeamlessM4Tv2ASRConfig.

Model

SeamlessM4Tv2ForSpeechRecognition

View source

AutoModelForSpeechRecognition.from_pretrained(
    pretrained_model_name_or_path,
    *,
    model_type='asr_seamless_m4t_v2',
    config=None,
    **model_kwargs,
)

Parameters

  • pretrained_model_name_or_path — Hub ID or compatible local directory.
  • model_type — Canonical model type; use 'asr_seamless_m4t_v2'.
  • config — Optional preloaded SeamlessM4Tv2ASRConfig instance.
  • **model_kwargs — Model-specific loading arguments.
from voicehub import get_model_spec

spec = get_model_spec('asr_seamless_m4t_v2')
print(spec.display_name, spec.task.value)
Purpose Public object
Discover get_model_spec('asr_seamless_m4t_v2')
Load and run AutoModelForSpeechRecognition
Configure SeamlessM4Tv2ASRConfig
Process AutoProcessor
Model implementation SeamlessM4Tv2ForSpeechRecognition
Normalized output ASROutput
Training contract get_training_spec('asr_seamless_m4t_v2')
Optimization lifecycle available_optimization_passes, apply_optimization_plan, optimization_manifest, restore_optimization_plan

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