Modelpublic

vaishnavanand/911-voice-to-text-20epochs

sourceHugging Faceupdated 1y agoView on Hugging Face

2likes7downloads

Model Card

911 Voice-to-Text Transcription Model (20 Epochs)

This is a specialized Whisper-based model trained on 911 emergency call recordings for accurate real-time voice-to-text transcription.

Performance

—Training Epochs: 20
—Coherence Score: 82%
—Model Size: Base Whisper
—Real-time Capable: Yes

Usage

python

from transformers import WhisperProcessor, WhisperForConditionalGeneration

processor = WhisperProcessor.from_pretrained("vaishnavanand/911-voice-to-text-20epochs")
model = WhisperForConditionalGeneration.from_pretrained("vaishnavanand/911-voice-to-text-20epochs")

Emergency Keywords

Help, Emergency, 911, Urgent, Critical, Accident, Fire, Police, Ambulance, Medical, Shooting, Robbery, Assault, Heart Attack, Stroke, Bleeding, Unconscious, Breathing, Chest Pain, Gun, Knife, Weapon, Crash, Trapped

License

MIT License - Free for commercial and research use.