Create a standard Databricks cluster and manually install TensorFlow and PyTorch in each notebook session.

Create a cluster with the Databricks Runtime for Machine Learning, then use the cluster's library management feature to install TensorFlow and PyTorch globally.

Create a cluster with the Databricks Runtime for Machine Learning, but only install TensorFlow globally and expect users to install PyTorch individually.

Create a cluster with the Databricks Runtime for Machine Learning, but do not install any additional libraries; users should install their own libraries as needed.