Whisper ASR Webservice

Whisper is a general-purpose speech recognition model. It is trained on a large dataset of diverse audio and is also a multitask model that can perform multilingual speech recognition as well as speech translation and language identification. For more details: github.com/openai/whisper

Features

Current release (v1.2.0) supports following whisper models:

• openai/whisper@v20230918
• guillaumekln/faster-whisper@0.9.0
• whisperX@v3.1.1

Quick Usage

CPU

docker run -d -p 9000:9000 -e ASR_MODEL=base -e ASR_ENGINE=openai_whisper onerahmet/openai-whisper-asr-webservice:latest

GPU

docker run -d --gpus all -p 9000:9000 -e ASR_MODEL=base -e ASR_ENGINE=openai_whisper onerahmet/openai-whisper-asr-webservice:latest-gpu

for more information:

• Documentation/Run
• Docker Hub

Documentation

Explore the documentation by clicking here.