Ref: https://github.com/jonesckevin/whisper-webapp

Whisper WebApp Example

Whisper WebApp Dockerized version for running OpenAI’s Whisper speech-to-text model. It provides an easy-to-use web interface for transcribing audio/video files with support for multiple languages and models.

Deploying Whisper WebApp in a Docker container simplifies setup and ensures consistent environments for reliable transcription.

yaml 33 lines
---

services:
    whisper-webapp:
        image: 'jonesckevin/whisper-webapp:latest'
        ports:
        - "${WHISPERPORT:-8000}:5000"
        volumes:
        # Mapped directories for external file access via network share
        - ./data/uploads:/data/uploads
        - ./data/completed:/data/completed
        # Whisper models cache - persists models outside container
        - ./data/models:/root/.cache/whisper
        environment:
        - NVIDIA_VISIBLE_DEVICES=all
        - CUDA_VISIBLE_DEVICES=0
        - FLASK_ENV=production
        - MAX_UPLOAD_SIZE_GB=5
        - PRELOAD_WHISPER_MODELS=false  # Set to 'true' to download all models on first run
        deploy:
        resources:
            reservations:
            devices:
                - driver: nvidia
                count: 1
                capabilities: [gpu]
        restart: unless-stopped
        healthcheck:
        test: ["CMD", "curl", "-f", "http://localhost:5000/api/health"]
        interval: 30s
        timeout: 10s
        retries: 3
        start_period: 60s

Docker Run

GPU

bash 15 lines
# GPU Version
docker run -d \
  --name whisper-webapp \
  -p 8000:5000 \
  -v "$(pwd)/data/uploads:/data/uploads" \
  -v "$(pwd)/data:/data/db" \
  -v "$(pwd)/data/models:/root/.cache/whisper" \
  -e NVIDIA_VISIBLE_DEVICES=all \
  -e CUDA_VISIBLE_DEVICES=0 \
  -e FLASK_ENV=production \
  -e MAX_UPLOAD_SIZE_GB=5 \
  -e PRELOAD_WHISPER_MODELS=false \
  --gpus '"device=0"' \
  --restart unless-stopped \
  jonesckevin/whisper-webapp:gpu