In this quickstart, we'll explore how to harness the power of Snowpark Container Services (SPCS) using GPUs to execute a notebook directly within Snowflake. Specifically, we'll train an XGBoost model and walk through a workflow that involves inspecting GPU resources, loading data from a Snowflake table, and setting up that data for modeling. In the notebook, we will train two XGBoost models—one on CPUs and the other using a GPU cluster—and then compare their runtimes and results.
For prerequisites, environment setup, step-by-step guide and instructions, please refer to the QuickStart Guide.
