subformer/meta-omnilingual-asr-7b 🖼️❓ → 📝

▶️ 337 runs 📅 Nov 2025 ⚙️ Cog 0.16.7 🔗 GitHub 📄 Paper ⚖️ License
language-detection multilingual speech-to-text

About

Omnilingual ASR 7B by Meta (Unofficial) - Automatic speech recognition supporting 1,693 languages with best-in-class accuracy. Meta's recommended variant for mission-critical transcription requiring maximum quality.

Example Output

Output

this is the micromachine man presenting the most midget miniature modicate of micromachines each one has dramatic details terrific trimmed precision pain jobs plus incredible micromachine pocket playsets there is a police station fire station restaurant service station and more perfect pocket portables to take any place and there are many miniature playsets to play within each one comes with its own special addition micromachine vehicle and fun fantastic features that miraculously move raise the boatlift at the airport marina man the gun turret at the army base clean your car at the car wash raise the toll bridge and these playsets fit together to form a micromachine world micromachine pocket playsets so tremendously tiny so perfectly precise so dazzlingly detailed you'll want to pocket them all micromachines are micromachine pocket playsets sold seperately from galoob the smaller they are the better they are

Performance Metrics

10.88s Prediction Time
10.89s Total Time
All Input Parameters
{
  "audio": "https://replicate.delivery/pbxt/O8bnN3bgiFoQWiC293bSbW0hom0ikfrxuZ3CvM4hTCjUy7kX/micro-machines.wav",
  "language": "eng_Latn"
}
Input Parameters
audio (required) Type: string
Audio file to transcribe. Supports most audio formats (MP3, WAV, FLAC, etc). Maximum 40 seconds.
language Default: auto
Select language for transcription. Choose 'auto' for automatic language detection, or select a specific language code (ISO 639-3 + ISO 15924 format: e.g., 'eng_Latn' for English, 'cmn_Hans' for Simplified Chinese).
Output Schema

Output

Type: string

Example Execution Logs
2025-11-28 00:34:54 - INFO - ============================================================
2025-11-28 00:34:54 - INFO - Starting new transcription request
2025-11-28 00:34:54 - INFO - ============================================================
2025-11-28 00:34:54 - INFO - Input parameters:
2025-11-28 00:34:54 - INFO -   - Audio file: /tmp/tmp7aiz1agumicro-machines.wav
2025-11-28 00:34:54 - INFO -   - Language: eng_Latn
2025-11-28 00:34:54 - INFO - Audio file found - Size: 5148.82 KB
2025-11-28 00:34:54 - INFO - Converting audio to WAV format...
2025-11-28 00:34:54 - INFO - Audio duration: 29.89 seconds
2025-11-28 00:34:54 - INFO - Audio converted to WAV: /tmp/tmpsmt24016.wav (sample rate: 44100 Hz)
2025-11-28 00:34:54 - INFO - Using specified language: eng_Latn
2025-11-28 00:34:54 - INFO - Starting transcription...
2025-11-28 00:35:04 - INFO - Transcription completed in 10.25 seconds
2025-11-28 00:35:04 - INFO - Transcription result length: 922 characters
2025-11-28 00:35:04 - INFO - Preview: this is the micromachine man presenting the most midget miniature modicate of micromachines each one...
2025-11-28 00:35:04 - INFO - ============================================================
2025-11-28 00:35:04 - INFO - Transcription request completed successfully
2025-11-28 00:35:04 - INFO - ============================================================
2025-11-28 00:35:04 - INFO - Cleaned up temporary file: /tmp/tmpsmt24016.wav
Version Details
Version ID
c1b60d827c17233c21429fc4b572d34b34251f503c145f43984361119586e1b1
Version Created
November 27, 2025
Run on Replicate →