Train your first Foundation Model!

In this section we will jumpstart you with the right knowledge to get your firs Foundation Model and Downstream Model ready.

For the purpose of this exercise, we will be using the HM Kaggle dataset, described in detail here .

Step 1 - Creating a model

After you log into the BaseModel, you will see a listing of all models created so far. More details about this screen and navigation can be found here.

1. Navigate to Create Model button on the upper right hand side corner, and click it

2. You will now see an interface with 3 separate blocks that require providing additional info.

   1. Click onUntitled Model to edit the name of your model, for example "Quickstart"
   2. Then navigate to first block - Data sources and click the Add data source button.

Step 2 - Define data sources

1. Click the Add data source button in the new window that appeared. Your screen should look like this at this stage:

   

2. Here we need to provide all necessary data sources for Foundation Model to work. You can read more about this in the main body of documentation here.

   1. Let's start with events. We can provide any name that we choose - in this example transactions as it is the same as the name of the table that we have created in snowflake.

   2. Let's add the rest of data sources before we continue with data sources setup. We will add customers and articles. Your data sources panel should look like this:

      

   3. Next, click on Transactions so we can define the exact data parameters for this data source.

      1. You need to provide Warehouse, Database, Schema and Source Table - pointing BaseModel precisely at the location that data resides in your Snowflake DB

      2. Select eventfrom the Table Type and provide date and time column as well as Entites unique idenfifier. Don't forget to include columns from your data that you want the model to use. The whole looks like this in our example:

         

      3. Now we move to Customers and do the same thing, the only difference is that we select Main entity attibute as a table type:

         

      4. Next, the same thing for attributes - which will enrich our transaction data:

         

      5. Finally, we need to tell the model how to join the articles table to the Transactions. Navigate to Transactions and click on Joins and define the join like this:

         

      6. Now click Next button in upper hand right side to close this window and move back to model configuration

      7. Click Apply on the block to save and validate your changes.

         

Step 3 - Define Audience and Schedule

1. Next we can refine the audience in the Audience filter block - for this tutorial we skip this option as we want to train our model on all existing data.
2. Finally we need to define schedule. Let's select One-time training and Start immediately in the options and then click Apply
3. Now you are ready to hit Run Model button and start training your first Foundation Model!