Table of contents

  1. Notice
  2. Summarized environments about the docker with the Nvidia Container Toolkit
  3. Summarized main features of the provided Dockerfile
  4. How to install the docker
  5. How to install the Nvidia Container Toolkit
  6. Preparations to build a Dockerfile
  7. How to build a Dockerfile
  8. How to run a docker container
  9. How to copy files and folders between the host and the docker container
  10. How to save and load a docker image
  11. Basic docker command
  12. Reference

1. Notice

  • A guide for Docker with the Nvidia Container Toolkit (i.e. Nvidia-Docker 2.0)
  • This is an updated post of the Docker with the Nvidia Container Toolkit [1] to cover how to build a Dockerfile for deep learning. Thus, if you are not familar with the docker commands, I highly recommend you should read the Docker with the Nvidia Container Toolkit [1]
  • I recommend that you should ignore the commented instructions with an octothorpe, #.
  • I recommend that you should fill in the appropriate letters between the parentheses, <>.

2. Summarized environments about the docker with the Nvidia Container Toolkit

  • Operating System (OS): Ubuntu MATE 20.04.2 LTS (Focal)
  • Graphics Processing Unit (GPU): NVIDIA TITAN Xp, 1ea
  • GPU driver: Nvidia-460.91.03
  • CUDA toolkit: CUDA-11.1
  • cuDNN: cuDNN v8.2.1
  • Docker: Docker version 20.10.7, build 20.10.7-0ubuntu5~20.04.2
  • Docker image: nvidia/cuda:10.2-cudnn7-devel-ubuntu18.04

3. Summarized main features of the provided Dockerfile

You can check the whole Dockerfile in the 7-2. Dockerfile.

  • TensorRT: 7.0.0.11
  • Torch2TRT: 0.1.0
  • FFmpeg: N-105288-g45e45a6060
  • nv-codec-headers: 10.0.26.2

4. How to install the docker

  1. Docker installation 
    1 $ curl https://get.docker.com | sh
2 $ sudo systemctl start docker && sudo systemctl enable docker
  2. Check the docker version. 
    1 $ sudo docker --version

     Docker version 20.10.7, build 20.10.7-0ubuntu5~20.04.2

5. How to install the Nvidia Container Toolkit

  1. Set up the repository and the Gnu Privacy Guard (GPG) key. 
    1 $ distribution=$(. /etc/os-release;echo $ID$VERSION_ID)
2 $ curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add -
3 $ curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list
4 $ sudo apt-get update
  2. Install the Nvidia-Docker 2.0 for the Nvidia Container Toolkit. 
    1 $ sudo apt-get install -y nvidia-docker2
2 $ sudo systemctl restart docker
  3. Test the Nviida-Docker 2.0. 
    1 $ sudo docker run --rm --gpus all nvidia/cuda:10.2-base nvidia-smi

     Tue Dec 22 08:40:19 2020
 +-----------------------------------------------------------------------------+
 | NVIDIA-SMI 440.100      Driver Version: 440.100      CUDA Version: 10.2     |
 |-------------------------------+----------------------+----------------------+
 | GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
 | Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
 |===============================+======================+======================|
 |   0  TITAN Xp            Off  | 00000000:01:00.0  On |                  N/A |
 | 23%   30C    P8    10W / 250W |    518MiB / 12192MiB |      0%      Default |
 +-------------------------------+----------------------+----------------------+
    
 +-----------------------------------------------------------------------------+
 | Processes:                                                       GPU Memory |
 |  GPU       PID   Type   Process name                             Usage      |
 |=============================================================================|
 +-----------------------------------------------------------------------------+

6. Preparations to build a Dockerfile

  1. Directory structure 
    .
├── Bash
│   └── bash_docker_utility.sh
├── Dockerfile
├── docker_utility
│   ├── awscliv2.zip
│   ├── eksctl
│   ├── ffmpeg.tar.gz
│   ├── goofys
│   ├── kubectl
│   ├── nv-codec-headers-10.0.26.2.tar.gz
│   ├── TensorRT-7.0.0.11.Ubuntu-18.04.x86_64-gnu.cuda-10.2.cudnn7.6.tar.gz
│   └── torch2trt.tar.gz
├── LICENSE
└── README.md

  2. The version of the required utilities for the docker

    • If you want to refer to the official site for each utility, please click the hyperlink in the below table.
    No. Utility Version SHA-1 Git commit  
    1 awscliv2.zip 2 80153150f54149ca2ce5da1248130487fd791e92 -  
    2 eksctl 0.79.0 b8c03f8706d4be4c7cc8a94aa316582809326964 -  
    3 ffmpeg.tar.gz N-105288-g45e45a6060 dd3c209e500afac258db3a2fa1077d06b9650663 45e45a60  
    4 goofys 0.24.0 438a43f737772bf117753c3b40f524ca39814c12 -  
    5 kubectl 1.17.11 0790bcc0d98c8f19320359b4537359ea058cfadb -  
    6 nv-codec-headers-10.0.26.2.tar.gz 10.0.26.2 1006a328ba4387ae46e4b1e4b208bd064223370d -  
    7 TensorRT-7.0.0.11.Ubuntu-18.04.x86_64-gnu.cuda-10.2.cudnn7.6.tar.gz 7.0.0.11 1f17eb8c0b2f1ea0c7bf935cd209a6d6f0d23f93 -  
    8 torch2trt.tar.gz 0.1.0 f9fa0032a7b3e5a1b434fdab4a4d5a859f01d553 75637700  


    • How to get the SHA-1 in the Ubuntu 
      1 $ sha1sum NAME_FILE 

          CHARACTER_SHA1_40 NAME_FILE
  3. Preparations for the utilities
    There are two options to prepare the utilities: i) 6-4. bash_docker_utility.sh; ii) Google Request. I recommend that you should select the option, 6-4. bash_docker_utility.sh because of downloading the latest version of the utilities. However, If you have any difficulties, please fill out the Google Request form. After reviewing the completed request, I will give you via email that you provided a compressed file, docker_utility.zip (SHA-1: 4f77bf2bd59e8af8c7b31bc4d5137b6f6ac6484a) including all utilities. Thus, you should fill in the correct email address. After receiving the email, you may be able to get the compressed file using the gdown that is a python3 package for downloading a item from the Google Drive.
    Unlike the compressed file, docker_utility.zip, the 6-4. bash_docker_utility.sh does not have codes to download the TensorRT. Thus, you should download the TensorRT from the official site when using the 6-4. bash_docker_utility.sh.
  4. bash_docker_utility.sh 


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64



 #!/bin/bash
 #
 #
 # Dveloper: vujadeyoon
 # Email: vujadeyoon@gmail.com
 # Github: https://github.com/vujadeyoon
 # Personal website: https://vujadeyoon.github.io
 #
 # Title: bash_docker_utility.sh
 # Description: Preparation for the required utilities.
 #
 #
 path_curr=$(pwd)
 path_parent=$(dirname ${path_curr})
 #
 #
 path_docker_utility=${path_curr}/docker_utility
 #
 #
 # Make a directory, docker_utility.
 mkdir -p ${path_docker_utility}
 #
 #
 # AWSCLI version 2
 curl --silent https://awscli.amazonaws.com/awscli-exe-linux-x86_64.zip -o ${path_docker_utility}/awscliv2.zip
 #
 #
 # eksctl
 curl --silent --location "https://github.com/weaveworks/eksctl/releases/latest/download/eksctl_$(uname -s)_amd64.tar.gz" | tar xz -C ${path_docker_utility}/
 #
 #
 # FFmpeg
 git clone https://git.ffmpeg.org/ffmpeg.git ${path_docker_utility}/ffmpeg
 cd ${path_docker_utility}/ffmpeg && git checkout 45e45a606077ccd0aab7eaffb8697e633b876fb2
 cd ${path_docker_utility} && tar -czvf ffmpeg.tar.gz ffmpeg
 rm -rf ${path_docker_utility}/ffmpeg
 #
 #
 # goofys
 wget --quiet https://github.com/kahing/goofys/releases/download/v0.24.0/goofys -P ${path_docker_utility}
 #
 #
 # kubectl
 curl --silent --location -o ${path_docker_utility}/kubectl https://amazon-eks.s3.us-west-2.amazonaws.com/1.17.11/2020-09-18/bin/linux/amd64/kubectl
 chmod +x ${path_docker_utility}/kubectl
 #
 #
 # nv-codec-headers
 wget --quiet https://github.com/FFmpeg/nv-codec-headers/releases/download/n10.0.26.2/nv-codec-headers-10.0.26.2.tar.gz -P ${path_docker_utility}
 #
 #
 # TensorRT
 # You should download the TensorRT from the official site, https://developer.nvidia.com/tensorrt.
 #
 #
 # Torch2TRT
 git clone https://github.com/NVIDIA-AI-IOT/torch2trt ${path_docker_utility}/torch2trt
 cd ${path_docker_utility}/torch2trt && git checkout 756377002708121d43a3e66ba72f56af65696402
 cd ${path_docker_utility} && tar -czvf torch2trt.tar.gz torch2trt
 rm -rf ${path_docker_utility}/torch2trt
 #
 #
 cd ${path_curr} || return

7. How to build a Dockerfile

  1. Build a Dockerfile shown in the 7-2. Dockerfile. 
    1 $ sudo docker build -t <REPOSITORY>:<TAG> .
  2. Dockerfile 


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282



 # Dveloper: vujadeyoon
 # Email: vujadeyoon@gmail.com
 # Github: https://github.com/vujadeyoon
 # Personal website: https://vujadeyoon.github.io
 #
 # Title: Dockerfile
 # Description: A Dockerfile for the NVIDIA Container Toolkit for Deep Learning
 #
 #
 FROM nvidia/cuda:10.2-cudnn7-devel-ubuntu18.04
 
 
 LABEL maintainer="vujadeyoon"
 LABEL email="vujadeyoon@gmial.com"
 LABEL version="1.0"
 LABEL description="Docker with NVIDIA Container Toolkit for Deep Learning"
 
 
 ENV DEBIAN_FRONTEND=noninteractive
 
 
 # 1. Install the essential ubuntu packages.
 RUN apt-get update &&  \
     apt-get upgrade -y &&  \
     apt-get install -y --no-install-recommends \
         build-essential \
         apt-utils \
         cmake \
         git \
         curl \
         vim \
         ssh \
         sudo \
         tar \
         libcurl3-dev \
         libfreetype6-dev \
         pkg-config \
         ca-certificates \
         libjpeg-dev \
         libpng-dev && \
     apt-get clean && \
     rm -rf /var/lib/apt/lists/*
 
 
 # 2. Install the useful ubuntu packages.
 RUN apt-get update &&  \
     apt-get install -y \
         ffmpeg \
         libgtk2.0-dev \
         python3-matplotlib \
         wget \
         tmux \
         zsh \
         locales \
         ncdu \
         htop \
         zip \
         unzip \
         rsync && \
     apt-get clean && \
     rm -rf /var/lib/apt/lists/*
 
 
 # 3. Set the locale
 RUN locale-gen en_US.UTF-8
 ENV LANG en_US.UTF-8
 ENV LANGUAGE en_US:en
 ENV LC_ALL en_US.UTF-8
 
 
 # 4. Install python.
 RUN apt-get update && \
     apt-get install software-properties-common -y && \
     add-apt-repository ppa:deadsnakes/ppa -y && \
     apt-get install -y \
          python-pip \
          python3-pip && \
     apt-get clean && \
     rm -rf /var/lib/apt/lists/*
 
 
 ## 5. Set python3.7 as default
 #RUN apt-get update && \
 #    apt-get install -y \
 #        python3.7 \
 #        python3.7-dev && \
 #    apt-get clean && \
 #    rm -rf /var/lib/apt/lists/* && \
 #    rm /usr/bin/python3 && \
 #    ln -s /usr/bin/python3.7 /usr/bin/python3
 
 
 # 6. Install more python packages
 RUN pip install --upgrade pip && \
     pip3 install --upgrade pip
 
 
 # 7. Python3 packages for mathematical functions and plotting
 RUN pip3 install \
     opencv-python==4.5.1.48 \
     opencv-contrib-python==4.5.1.48 \
     ffmpeg-python \
     Pillow==7.2.0 \
     imageio \
     kornia==0.2.2 \
     matplotlib==3.3.1 \
     scikit-image \
     scikit-learn \
     pandas \
     openpyxl \
     plotly \
     seaborn \
     shapely
 
 
 # 8. Python3 packages for monitoring and debugging
 RUN pip3 install \
     jupyter \
     wandb \
     gpustat \
     getgpu \
     tqdm \
     ipdb \
     icecream
 
 
 # 9. Other python3 packages
 RUN pip3 install \
     scipy \
     Cython \
     prettyprinter \
     colorlog \
     randomcolor \
     future==0.18.2 \
     imutils \
     psutil \
     PyYAML \
     pycrypto
 
 
 # 10. Install python3 packages for the deep learning research.
 RUN pip3 install \
     PyWavelets \
     pycuda==2020.1 \
     torch==1.4.0 \
     torchvision==0.5.0 \
     torchsummary==1.5.1
 
 
 # 11. Install python3 packages related to the server.
 RUN pip3 install \
     flask==1.1.2 \
     Flask-RESTful==0.3.8 \
     gevent==21.1.2 \
     boto3==1.17.9 \
     kubernetes==12.0.1
 
 
 # 12. Make required directories.
 RUN mkdir -p /home/Vujadeyoon/
 RUN mkdir -p /home/Vujadeyoon/vdisk/
 
 
 # 13. Install a python3 package, TensorRT 7.0.0.11.
 #     i) Preparations.
 RUN mkdir -p /home/Vujadeyoon/pip3_packages/
 COPY ./docker_utility/TensorRT-7.0.0.11.Ubuntu-18.04.x86_64-gnu.cuda-10.2.cudnn7.6.tar.gz /home/Vujadeyoon/pip3_packages/
 WORKDIR /home/Vujadeyoon/pip3_packages/
 #     ii) Uncompress the downloaded tar.gz file.
 RUN tar -xzvf TensorRT-7.0.0.11.Ubuntu-18.04.x86_64-gnu.cuda-10.2.cudnn7.6.tar.gz
 RUN mkdir -p ./TensorRT-7.0.0.11/_arxiv/
 RUN mv ./TensorRT-7.0.0.11.Ubuntu-18.04.x86_64-gnu.cuda-10.2.cudnn7.6.tar.gz TensorRT-7.0.0.11/_arxiv/
 #     iii) Register an environmental variable.
 ENV LD_LIBRARY_PATH $LD_LIBRARY_PATH:/home/Vujadeyoon/pip3_packages/TensorRT-7.0.0.11/lib
 #     iv) Copy plugins of the TensorRT to the CUDA toolkit directory.
 RUN mkdir -p /usr/local/cuda-originals/cuda-10.2/
 RUN cp -r /usr/local/cuda-10.2/targets/ /usr/local/cuda-originals/cuda-10.2/
 RUN cp -r /home/Vujadeyoon/pip3_packages/TensorRT-7.0.0.11/include/* /usr/local/cuda-10.2/targets/x86_64-linux/include/
 RUN cp -r /home/Vujadeyoon/pip3_packages/TensorRT-7.0.0.11/targets/x86_64-linux-gnu/lib/* /usr/local/cuda-10.2/targets/x86_64-linux/lib/
 #     v) Install python3 packages related to the TensorRT.
 WORKDIR /home/Vujadeyoon/pip3_packages/TensorRT-7.0.0.11/
 RUN pip3 install \
     ./python/tensorrt-7.0.0.11-cp36-none-linux_x86_64.whl \
     ./uff/uff-0.6.5-py2.py3-none-any.whl \
     ./graphsurgeon/graphsurgeon-0.4.1-py2.py3-none-any.whl
 #     vi) Go to the base directory.
 WORKDIR /home/Vujadeyoon/
 
 
 # 14. Install a python3 package, Torch2TRT.
 #     i) Preparations.
 RUN mkdir -p /home/Vujadeyoon/pip3_packages/
 COPY ./docker_utility/torch2trt.tar.gz /home/Vujadeyoon/pip3_packages/
 WORKDIR /home/Vujadeyoon/pip3_packages/
 #     ii) Uncompress the downloaded tar.gz file.
 RUN tar -xzvf /home/Vujadeyoon/pip3_packages/torch2trt.tar.gz
 RUN mkdir -p ./torch2trt/_arxiv/
 RUN mv ./torch2trt.tar.gz torch2trt/_arxiv/
 #     iii) Install the python3 package, Torch2TRT.
 WORKDIR /home/Vujadeyoon/pip3_packages/torch2trt/
 RUN python3 setup.py install --plugins
 #     iv) Go to the base directory.
 WORKDIR /home/Vujadeyoon/
 
 
 # 15. Install the AWS CLI version 2 on Linux.
 RUN mkdir -p /home/Vujadeyoon/AWS/
 COPY ./docker_utility/awscliv2.zip /home/Vujadeyoon/AWS/
 RUN unzip /home/Vujadeyoon/AWS/awscliv2.zip -d /home/Vujadeyoon/AWS/
 RUN /home/Vujadeyoon/AWS/aws/install
 RUN rm -rf /home/Vujadeyoon/AWS/
 
 
 # 16. Enroll the AWS credentials and config for development.
 # RUN mkdir /root/.aws
 # RUN echo '[default]' >> /root/.aws/credentials
 # RUN echo 'aws_access_key_id = <SECRET>' >> /root/.aws/credentials
 # RUN echo 'aws_secret_access_key = <SECRET>' >> /root/.aws/credentials
 # RUN echo '[default]' >> /root/.aws/config
 # RUN echo 'region = ap-northeast-2' >> /root/.aws/config
 # RUN echo 'output = json' >> /root/.aws/config
 
 
 # 17. Install the Goofys
 COPY ./docker_utility/goofys /usr/local/bin/
 RUN chmod +x /usr/local/bin/goofys
 
 
 # 18. Install the EKSCTL.
 COPY ./docker_utility/eksctl /usr/local/bin/
 RUN chmod +x /usr/local/bin/eksctl
 RUN eksctl completion bash >> /root/.bash_completion
 
 
 # 19. Install the KUBECTL.
 COPY ./docker_utility/kubectl /usr/local/bin/
 RUN chmod +x /usr/local/bin/kubectl
 RUN kubectl completion bash >> /root/.bash_completion
 
 
 # 20. FFmpeg-GPU
 RUN apt-get update &&  \
     apt-get install -y \
         yasm \
         libtool \
         libc6 \
         libc6-dev \
         libnuma1 \
         libnuma-dev \
         libmp3lame-dev \
         psmisc \
         libass-dev \
         libgl1-mesa-glx && \
     apt-get clean && \
     rm -rf /var/lib/apt/lists/*
 RUN mkdir -p /home/Vujadeyoon/FFmpeg_GPU/
 COPY ./docker_utility/nv-codec-headers-10.0.26.2.tar.gz /home/Vujadeyoon/FFmpeg_GPU/
 COPY ./docker_utility/ffmpeg.tar.gz /home/Vujadeyoon/FFmpeg_GPU/
 WORKDIR /home/Vujadeyoon/FFmpeg_GPU/
 RUN tar -xzvf nv-codec-headers-10.0.26.2.tar.gz
 RUN tar -xzvf ffmpeg.tar.gz
 WORKDIR /home/Vujadeyoon/FFmpeg_GPU/nv-codec-headers-10.0.26.2/
 RUN make install
 # RUN git clone https://git.videolan.org/git/ffmpeg/nv-codec-headers.git
 # RUN cd nv-codec-headers && make install && cd /data
 WORKDIR /home/Vujadeyoon/FFmpeg_GPU/ffmpeg/
 RUN ./configure --enable-cuda-sdk --enable-cuvid --enable-nvenc --enable-nonfree --enable-libnpp  --enable-libass --enable-libmp3lame --extra-cflags=-I/usr/local/cuda/include --extra-ldflags=-L/usr/local/cuda/lib64 && make -j 8 && make install
 
 
 # 21. Update and clean the Ubuntu packages.
 RUN apt-get update &&  \
     apt-get clean && \
     rm -rf /var/lib/apt/lists/*
 
 
 # 22. Go to the base directory.
 WORKDIR /home/Vujadeyoon/
 
 
 # 23. Run the command.
 CMD [ "/bin/bash" ]

8. How to run a docker container

  1. Run a docker container corresponding to the docker image, <REPOSITORY>:<TAG> with some useful options. 
    1 $ sudo docker run -it --rm --privileged --runtime nvidia -p <PORT_HOST>:<PORT_CONTAINER> -v <PATH_HOST>:<PATH_CONTAINER> <REPOSITORY>:<TAG> "/bin/bash"

9. How to copy files and folders between the host and a docker container

You can get the <CONTAINER_ID> using the below command. If you are not familar with the command, please refer to the 11. Basic docker command. The items can be both folders or files.

1 $ sudo docker ps --all

  CONTAINER ID   IMAGE          COMMAND                  CREATED        STATUS                      PORTS     NAMES
  1. Copy itmes from the host to the docker container. 
    1 $ docker cp <PATH_HOST> <CONTAINER_ID>:<PATH_CONTAINER>
  2. Copy itmes from the docker container to the host. 
    1 $ docker cp <CONTAINER_ID>:<PATH_CONTAINER> <PATH_HOST>

10. How to save and load a docker image

  1. Save a docker image. 
    1 $ sudo docker save <REPOSITORY:TAG> -o <NAME_TAR>.tar
  2. Load a pre-built docker image. 
    1 $ sudo docker load -i <NAME_TAR>.tar

11. Basic docker command

  1. Exit the docker container. 
    1 $ ctrl + d
  2. List docker images. 
    1 $ sudo docker images
  3. List all docker containers. 
    1 $ sudo docker ps --all
  4. Stop a docker container. 
    1 $ sudo docker stop <CONTAINER_ID>
  5. Start the stopped docker container. 
    1 $ sudo docker start <CONTAINER_ID>
  6. Attach a working docker container. 
    1 $ sudo docker attach <CONTAINER_ID>
  7. Remove docker containers.
    1. Remove docker containers corresponding to the container ID list. 
       1 $ sudo docker rm <CONTAINER_ID_1>, <CONTAINER_ID_2> ,..., <CONTAINER_ID_N>
    2. Remove all docker containers. 
       1 $ sudo docker rm `sudo docker ps -a -q`
  8. Remove a docker images.
    1. Remove docker images corresponding to the docker image list. 
       1 $ sudo docker rmi <IMAGE_1>, <IMAGE_2> ,..., <IMAGE_N>
    2. Remove docker images with the corresponding docker containers from the given docker image list. 
       1 $ sudo docker rmi -f <IMAGE_1>, <IMAGE_2> ,..., <IMAGE_N>

12. Reference

  1. Docker-Nvidia-Container-Toolkit
  2. Docker
  3. Github for the Nvidia-Docker
  4. Document for the Nvidia-Docker