Skip to main content
GitHub source

​
Usage

wandb docker [OPTIONS] [DOCKER_RUN_ARGS]... [DOCKER_IMAGE]

​
Description

Run your code in a docker container. Run your program inside a Docker image with W&B configured. Set the WANDB_DOCKER and WANDB_API_KEY environment variables in the container and mount the current working directory at /app by default. Pass additional arguments to insert them into docker run before the image name. If you do not specify an image, select a default image automatically. 
wandb docker -v /mnt/dataset:/app/data
wandb docker gcr.io/kubeflow-images-public/tensorflow-1.12.0-notebook-cpu:v0.4.0 --jupyter
wandb docker wandb/deepo:keras-gpu --no-tty --cmd "python train.py --epochs=5"
Override the container entrypoint by default to ensure wandb is installed. If you pass --jupyter, ensure Jupyter is installed and start JupyterLab on port 8888. If NVIDIA Docker is available, use the NVIDIA runtime automatically. To set W&B environment variables for an existing docker run command without modifying the entrypoint, use wandb docker-run.

​
Arguments

NameDefaultType
docker_run_argsNoneSTR
docker_imageNoneSTR

​
Options

FlagTypeDescription
--nvidia / --no-nvidiaBOOL FlagUse the nvidia runtime, defaults to nvidia if nvidia-docker is present Default: False
--digestBOOL FlagOutput the image digest and exit Default: False
--jupyter / --no-jupyterBOOL FlagRun jupyter lab in the container Default: False
--dirSTRWhich directory to mount the code in the container Default: /app
--no-dirBOOL FlagDon’t mount the current directory Default: False
--shellSTRThe shell to start the container with Default: /bin/bash
--portSTRThe host port to bind jupyter on Default: 8888
--cmdSTRThe command to run in the container Default: None
--no-ttyBOOL FlagRun the command without a tty Default: False