(Hands-On) Monitor the Distributed Training Process

In this section, we will view the pipeline run and monitor the distributed training process.

1. Navigate to the Runs

Click on the Runs tab:

2. Find the pipeline run

Find the pipeline run and click on it:

3. View the progress

View the progress of the pipeline run:

4. View the submit_training_job step

In this step, Kale submits a PyTorch distributed training job:

5. View the logs of the distributed training job

To view the logs of the distributed training job, click on the monitor step:

Go to the Logs tab:

6. View the pipeline run

Wait for the pipeline run to complete:

Congratulations! You have successfully trained an artificial neural network in a distributed manner using Kubeflow, Kale, PyTorch, and Rok.