A data scientist is working on training a deep learning model with a large dataset in Databricks, facing challenges due to the resource-intensive and time-consuming nature of the process. What strategy can the data scientist employ to enhance the training performance within the Databricks environment?