Skip to content

Latest commit

 

History

History
67 lines (52 loc) · 2.49 KB

INSTALLATION-DOCKER.md

File metadata and controls

67 lines (52 loc) · 2.49 KB

To use the dockerized version of Tensorflow, first follow the official Nvidia installation guide: https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/install-guide.html#docker

DockerHub Image for Linux

The ready-to-use image is accessible on Docker Hub.

Pull the image

docker pull mortentabaka/landcover_semantic_segmentation_with_deeplabv3plus:latest

Run the image in interactive mode

docker run --gpus all -it -p 8888:8888 mortentabaka/landcover_semantic_segmentation_with_deeplabv3plus:latest

Run the image and create files locally

export PROJECT_PATH_LOCALLY="/path/to/local/code/directory" &&
git clone https://github.com/MortenTabaka/Semantic-segmentation-of-LandCover.ai-dataset.git "$PROJECT_PATH_LOCALLY" &&
docker run --gpus all -it -p 8888:8888 -v $PROJECT_PATH_LOCALLY:/app/ mortentabaka/landcover_semantic_segmentation_with_deeplabv3plus:latest

Dockerfile - Tensorflow GPU

Clone the repository:

git clone git@github.com:MortenTabaka/Semantic-segmentation-of-LandCover.ai-dataset.git && cd Semantic-segmentation-of-LandCover.ai-dataset/

Build the docker image with the project's Dockerfile:

docker build -t landcover_semantic_segmentation .

An official image was used as a base: https://hub.docker.com/layers/tensorflow/tensorflow/2.5.1-gpu-jupyter/images/sha256-5cdcd4446fc110817e5f6c5784eba6254258632036b426b9f194087e200f8a96?context=explore

Run the Jupyter Notebook with:

docker run --gpus all -it --rm -p 8888:8888 -v $(pwd):/app landcover_semantic_segmentation

Dockerfile - Tensorflow CPU (not tested)

Clone the repository

git clone git@github.com:MortenTabaka/Semantic-segmentation-of-LandCover.ai-dataset.git && cd Semantic-segmentation-of-LandCover.ai-dataset/

In Dockerfile change tensorflow image name to tensorflow/tensorflow:2.5.1-jupyter.

Build a image.

docker build -t landcover_semantic_segmentation .

To run the image, do not use flag --gpus all:

docker run -it --rm -p 8888:8888 -v $(pwd):/app landcover_semantic_segmentation

If port is already in use

If port 8888 is already in use, then change its value, e.g. -p 5555:8888. Remember to manually replace port in a link to the chosen value:

Would be: http://127.0.0.1:8888/?token=...

Should be: http://127.0.0.1:5555/?token=...