Skip to content

This guide walks you through the end-to-end workflow of training your first Foundation Model, building a Downstream Model for predictions, and generating inference results. We use the H&M Personalized Fashion Recommendations Kaggle dataset as a working example so you can follow along with real data.

By the end of this Quick Start, you will have a fully trained model producing predictions stored in a Snowflake table.

Prerequisites

  • Snowflake Native Apps - BaseModel installed and configured.
  • BaseModel license
  • Data on Snowflake
  • Dataset that meets minimum data requirements

In this section, you will learn the basics of how to train your first Foundation Model and then Downstream Model.

Upload Data to Your Snowflake Account

Follow these steps to upload the H&M data:

  1. Log into your Snowflake account.

  2. Log into Kaggle and download the CSV files here - This will require you to have/create a free Kaggle account

  3. Convert CSV files to Parquet files

  4. Upload via SnowSql:

  5. Please refer to official documentation on how to load the data into snowflake. For example, loading the data from local system is described here.