Use TensorBoard in Jupyter Workspaces

​

Set up your Environment

1. In the navigation bar, click Environments.
2. Click Create Environment.
3. Give the Environment a name, and then select a base image that has Python 3.8 installed. You can use an Environment as a base image if it uses this Domino standard: quay.io/domino/compute-environment-images:ubuntu20-py3.9-r4.3-domino5.10-gpu quay.io/domino/domino-gpu-environment:develop.3d96fa12
4. After setting the Visibility and entering a description, click Create Environment. The Environment’s Overview page opens.
5. Go to Docker Settings and click Edit Dockerfile.
6. Add the following lines to the Dockerfile Instructions:
   USER root
   
   RUN apt-get update
   RUN apt-get install -y --no-install-recommends libnode-dev npm nodejs
   RUN npm install -g configurable-http-proxy
   RUN pip install -U jupyter jupyter-tensorboard jupyter-server-proxy ipywidgets
   
   USER ubuntu
   
7. Add the following lines to the Pluggable Workspace Tools to add JupyterLab workspace:
   jupyterlab:
     title: "JupyterLab"
     iconUrl: "/assets/images/workspace-logos/jupyterlab.svg"
     start: [  "/opt/domino/workspaces/jupyterlab/start" ]
     httpProxy:
       internalPath: "/{{ownerUsername}}/{{projectName}}/{{sessionPathComponent}}/{{runId}}/{{#if pathToOpen}}tree/{{pathToOpen}}{{/if}}"
       port: 8888
       rewrite: false
       requireSubdomain: false
   
8. Add the following lines to the Pre-run script if the Project you plan to use in this Compute Environment is a Domino Project:
   #!/bin/bash
   set -e
   
   # Ensure required packages are installed (optional safety step)
   pip install --quiet --upgrade jupyter-server-proxy tensorboard "notebook<7"
   
   # Create Jupyter config directory
   CONF_DIR="${HOME}/.jupyter"
   mkdir -p "${CONF_DIR}"
   
   # Write valid server proxy config for TensorBoard
   cat << EOF > "${CONF_DIR}/jupyter_notebook_config.py"
   c.ServerProxy.servers = {
       "tensorboard": {
           "command": [
               "tensorboard",
               "--logdir", "/domino/datasets/local/${DOMINO_PROJECT_NAME}/tensorboard_logs",
               "--host", "0.0.0.0",
               "--port", "{port}"
           ],
           "timeout": 30,
           "launcher_entry": {
               "title": "TensorBoard"
           }
       }
   }
   EOF
   
9. Add the following lines to the Pre-run script if the Project is a Git-based Project:
   #!/bin/bash
   set -e
   
   # Ensure required packages are installed (optional safety step)
   pip install --quiet --upgrade jupyter-server-proxy tensorboard "notebook<7"
   
   # Create Jupyter config directory
   CONF_DIR="${HOME}/.jupyter"
   mkdir -p "${CONF_DIR}"
   
   # Write valid server proxy config for TensorBoard
   cat << EOF > "${CONF_DIR}/jupyter_notebook_config.py"
   c.ServerProxy.servers = {
       "tensorboard": {
           "command": [
               "tensorboard",
               "--logdir", "/mnt/data/${DOMINO_PROJECT_NAME}/tensorboard_logs",
               "--host", "0.0.0.0",
               "--port", "{port}"
           ],
           "timeout": 30,
           "launcher_entry": {
               "title": "TensorBoard"
           }
       }
   }
   EOF
   
10. Click Build. The Revisions page opens. If the new revision builds successfully, you can use this Environment.

​

Use Jupyter-TensorBoard

1. Open the Project you want to use with Jupyter-Tensorboard.
2. In the navigation pane, click Settings.
3. From Compute Environment, select the Environment you created previously. A notification opens to verify that the new Environment is now set.
4. In the navigation pane, click Workspaces, then select Jupyterlab and launch a new workspace.
5. Select Tensorboard from the Launcher.

​

Troubleshooting

1. If you encounter issues loading Tensorboard initially, confirm that Tensorflow is operating properly. Loading the application depends on Tensorflow.
2. By default, Domino’s standard Compute Environments have tensorflow-gpu installed (for example, pip install tensorflow-gpu). Therefore, Tensorboard and Tensorflow will not work on a CPU hardware tier. If you’d like to use Tensorboard on a CPU, make sure that CPU optimized Tensorflow is installed (for example, pip install tensorflow).