How to run a Jupyter Notebook inside Tensorflow Container for Accelerated Machine Learning E2E GPU Wizard?

Launching an Instance based on Tensorflow container with Jupyter-notebook Running

Using myaccount portal

1. First sign into the myaccount portal.

2. Navigate to GPU Wizard:

   • Under the compute menu, click on GPU.
   • Then click on GPU Wizard.
   • For NGC Container Tensorflow, click on Next under actions column.
   • Choose the card according to requirement. A100 is recommended.

   

3. Choose the plan as per the requirement:

   • Optionally you can add SSH key (recommended) or subscribe to CDP backup.
   • Click on “Create my node”.
   • Wait for a few minutes and confirm that node is in running state.

   

4. Open terminal on your local PC and type the following command:

   ssh -NL localhost:1234:localhost:8888 root@<your_node_ip>
   

• The command usually will not show any output, which indicates that the command has run without any error.
• Go to a web browser on your local PC and hit the URL: http://localhost:1234.

• Congratulations! Now you can run your Python code inside this Jupyter notebook, which has TensorFlow and all the libraries frequently used in machine learning preconfigured.
• To get the most out of GPU acceleration, use RAPIDS and DALI, which are already installed inside this container.
• RAPIDS and DALI accelerate tasks in machine learning, including data loading and preprocessing, beyond just the learning process.

How to Run

• MNIST dataset is a kind of “Hello world” implementation in machine learning.
• Here we are about to use Fashion MNIST (clothing classification) as given in the NGC link here.
• Copy the wget command as given below.

• In your Jupyter Notebook UI, create a new notebook.

• Enter ‘!’ in the cell and paste the wget command.

• Once you run the command, you can find the downloaded file on the Jupyter home page.

• Now open the notebook FashionMNIST+Notebook.ipynb.
  • You can skip up to step 4 (which was already done) and directly go to Step 5.
  • You can run cell by cell from Step 5 and observe each output, or you can go to the menu bar and select “Cell -> Run All” to run all of them.

• You can see that after 15 epochs, the accuracy will be around 92% on training data and 88% on testing data.

How to troubleshoot if the notebook isn’t accessible

1. SSH into your node with the public IP: ssh username@<public-ip>.
2. Check the output of the docker ps command.

3. If it shows a container is running using the TensorFlow image, you can use the following command in the terminal:

         cnid=$(docker ps |grep tensorflow|head -1|awk '{print $1}');/usr/bin/nohup /usr/bin/docker exec $cnid /usr/local/bin/jupyter-notebook --no-browser --port=8888 --allow-root --NotebookApp.token='' --NotebookApp.password='' --NotebookApp.quit_button=False --NotebookApp.notebook_dir=/home/tensorflow &

• The command usually will not show any output, which indicates that the command has run without any error.
• Go to a web browser on your local PC and hit the URL: http://localhost:1234.

• Congratulations! Now you can run your Python code inside this Jupyter notebook, which has TensorFlow and all the libraries frequently used in machine learning preconfigured.
• To get the most out of GPU acceleration, use RAPIDS and DALI, which are already installed inside this container.
• RAPIDS and DALI accelerate tasks in machine learning, including data loading and preprocessing, beyond just the learning process.

How to Run

• MNIST dataset is a kind of “Hello world” implementation in machine learning.
• Here we are about to use Fashion MNIST (clothing classification) as given in the NGC link here.
• Copy the wget command as given below.

• In your Jupyter Notebook UI, create a new notebook.

• Enter ‘!’ in the cell and paste the wget command.

• Once you run the command, you can find the downloaded file on the Jupyter home page.

• Now open the notebook FashionMNIST+Notebook.ipynb.
  • You can skip up to step 4 (which was already done) and directly go to Step 5.
  • You can run cell by cell from Step 5 and observe each output, or you can go to the menu bar and select “Cell -> Run All” to run all of them.

• You can see that after 15 epochs, the accuracy will be around 92% on training data and 88% on testing data.

How to troubleshoot if the notebook isn’t accessible

1. SSH into your node with the public IP: ssh username@<public-ip>.
2. Check the output of the docker ps command.

3. If it shows a container is running using the TensorFlow image, you can use the following command in the terminal:

         cnid=$(docker ps |grep tensorflow|head -1|awk '{print $1}');/usr/bin/nohup /usr/bin/docker exec $cnid /usr/local/bin/jupyter-notebook --no-browser --port=8888 --allow-root --NotebookApp.token='' --NotebookApp.password='' --NotebookApp.quit_button=False --NotebookApp.notebook_dir=/home/tensorflow &

• Now create an SSH tunnel again on your local PC with the following command:
  ssh -NL localhost:1234:localhost:8888 root@<your_node_ip>
  and attempt to access the Jupyter notebook.

Why doesn't my notebook have a password?

The notebook is accessed via an SSH tunnel, which creates a secure encrypted connection between the node and the PC, making it accessible only for those who have SSH access. However, if you want to set a password, you can do so by following these steps:

• Stop any existing Jupyter notebook processes inside the container.
• Edit the field –NotebookApp.password=’Your password’ in the above commands.
• To set it on the next boot, open the file /etc/rc.local in your server and set the password as shown below.