Error response from daemon: unknown or invalid runtime name: nvidia

[xiaotupansy]

docker-compose up -d Error response from daemon: unknown or invalid runtime name: nvidia

Here's the yml:

networks: ai_network: driver: bridge

services: heygem-tts: image: guiji2025/fish-speech-ziming:1.0.39 container_name: heygem-tts restart: always runtime: nvidia environment: - NVIDIA_VISIBLE_DEVICES=0 - NVIDIA_DRIVER_CAPABILITIES=compute,graphics,utility,video,display ports: - '18180:8080' volumes: - e:/heygem/data/voice/data:/code/data command: /bin/bash -c "/opt/conda/envs/python310/bin/python3 -m tools.api --listen 0.0.0.0:8080" networks: - ai_network heygem-f2f: image: guiji2025/heygem.ai:0.0.7_sdk_slim container_name: heygem-f2f restart: always privileged: true volumes: - e:/heygem/data/face2face:/code/data environment: - PYTORCH_CUDA_ALLOC_CONF=max_split_size_mb:512 deploy: resources: reservations: devices: - capabilities: [gpu] shm_size: '8g' ports: - '8383:8383' command: python /code/app_local.py networks: - ai_network heygem-asr: image: guiji2025/fun-asr:1.0.1 container_name: heygem-asr restart: always runtime: nvidia privileged: true working_dir: /workspace/FunASR/runtime ports: - '10095:10095' volumes: - e:/heygem/data/asr/runtime-resources/models:/workspace/models command: sh /run.sh deploy: resources: reservations: devices: - driver: nvidia count: all capabilities: [gpu] networks: - ai_network

[whl88]

Make sure you have an NVIDIA graphics card installed on your computer and that the drivers are correctly installed.

[xiaotupansy]

Yes. I have RTX 3060 in my computer.

[maxadc]

同样的问题，win11

[likely-cloud]

我也遇到了，也是win11

[whl88]

We haven't encountered this kind of problem. But I consulted a large - language model, and the solution it provided seems quite reliable. How about giving it a try?

1. 首先确保您已经安装了： - 最新的 NVIDIA 显卡驱动

   • Windows 版本的 Docker Desktop
   • NVIDIA Container Toolkit for Windows

2. 在 Windows 上安装 NVIDIA Container Toolkit 的步骤： 安装 NVIDIA CUDA Toolkit（如果还没安装的话）： https://developer.nvidia.com/cuda-downloads

3. 安装完成后，需要在 Docker Desktop 中启用 NVIDIA Runtime：

   • 打开 Docker Desktop
   • 点击右上角的设置图标（齿轮）
   • 转到 "Docker Engine" 配置
   • 在配置文件中添加或修改以下内容：

{
  "runtimes": {
    "nvidia": {
      "path": "nvidia-container-runtime"
    }
  }
}

4. 点击 "Apply & Restart" 重启 Docker Desktop

5. 验证安装： 打开 PowerShell 或命令提示符，运行：

docker run --gpus all nvidia/cuda:11.0-base nvidia-smi

如果一切正常，您应该能看到 GPU 信息输出。

[Frank-1995]

我的是11.6的cuda，我按照你说的，不行，docker启动TTS服务启动不起来，真是奇怪，报错：Cannot start Docker Compose application. Reason: compose [start] exit status 1. Container duix-avatar-asr Starting Container duix-avatar-gen-video Starting Container duix-avatar-tts Starting Container duix-avatar-asr Started Container duix-avatar-gen-video Started Error response from daemon: failed to create task for container: failed to create shim task: OCI runtime create failed: runc create failed: unable to start container process: error during container init: error running prestart hook #0: exit status 1, stdout: , stderr: Auto-detected mode as 'legacy' nvidia-container-cli: requirement error: unsatisfied condition: cuda>=12.1, please update your driver to a newer version, or use an earlier cuda container: unknown

14 hours ago