EtienneAb3d/WhisperHallu
Experimental code: sound file preprocessing to optimize Whisper transcriptions without hallucinated texts
This tool helps you get more accurate transcriptions from audio by cleaning up the sound first. It takes an audio file with spoken content and removes background noise, silences, and other distractions. The output is a clearer audio file that, when transcribed by Whisper, significantly reduces irrelevant or 'hallucinated' text. This is perfect for anyone who relies on automated speech-to-text for interviews, lectures, or content analysis and needs high precision.
348 stars. No commits in the last 6 months.
Use this if you are experiencing inaccurate or nonsensical text in your automated audio transcriptions and need a cleaner, more reliable output.
Not ideal if your primary goal is real-time transcription, as this involves a preprocessing step that adds latency.
Stars
348
Forks
25
Language
Python
License
—
Category
Last pushed
Nov 12, 2024
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/voice-ai/EtienneAb3d/WhisperHallu"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
YuanGongND/whisper-at
Code and Pretrained Models for Interspeech 2023 Paper "Whisper-AT: Noise-Robust Automatic Speech...
adi-gov-tw/Taiwan-Tongues-ASR-CE
Taiwan Tongues ASR CE 是一個開源語音辨識（Automatic Speech Recognition, ASR）模型專案，專為台灣多元語言環境設計。 本模型支援...
huggingface/distil-whisper
Distilled variant of Whisper for speech recognition. 6x faster, 50% smaller, within 1% word error rate.
phineas-pta/fine-tune-whisper-vi
jupyter notebooks to fine tune whisper models on Vietnamese using Colab and/or Kaggle and/or AWS EC2
KevKibe/African-Whisper
🚀 Framework for seamless fine-tuning of Whisper model on a multi-lingual dataset and deployment to prod.