CosyVoice Docker & Self-Hosted Deployment

GitHub

Multi-lingual large voice generation model, providing inference, training and deployment full-stack ability.

20,951 stars Python Python 1 file · ~3,296 tokens #audio-generation#cantonese#chatbot#chatgpt#chinese#cosyvoice#cross-lingual#english
RAW Compose
1. Pull Image (DOCKERHUB) docker pull soar97/triton-cosyvoice
2. Docker Run (1-Click Launch) docker run -d --name cosyvoice -p 8000:8000 --restart always soar97/triton-cosyvoice:25.06
3. Podman Run podman run -d --name cosyvoice -p 8000:8000 --restart always soar97/triton-cosyvoice:25.06
4. Skopeo Inspect (Zero-Pull Manifest) skopeo inspect docker://docker.io/soar97/triton-cosyvoice:25.06
5. ORAS Pull (OCI Artifact) oras pull soar97/triton-cosyvoice:25.06
6. Compose Deployment docker compose up -d
runtime/triton_trtllm/docker-compose.cosyvoice2.dit.yml
Download .yml 
services:
  tts:
    image: soar97/triton-cosyvoice:25.06
    shm_size: '1gb'
    ports:
      - "8000:8000"
      - "8001:8001"
      - "8002:8002"
    environment:
      - PYTHONIOENCODING=utf-8
      - MODEL_ID=${MODEL_ID}
    deploy:
      resources:
        reservations:
          devices:
            - driver: nvidia
              device_ids: ['0']
              capabilities: [gpu]
    command: >
      /bin/bash -c "pip install modelscope && cd /workspace && git clone https://github.com/yuekaizhang/Step-Audio2.git -b trt && git clone https://github.com/yuekaizhang/CosyVoice.git -b streaming && cd CosyVoice && git submodule update --init --recursive && cd runtime/triton_trtllm && bash run_stepaudio2_dit_token2wav.sh 0 3"