Skip to main content
Version: 0.0.38

Create Custom Model

Building custom models within Lakehousecat allows administrators and builders to tailor advanced AI capabilities to meet specific project needs. This guide will cover the fundamental steps involved in creating Custom Models, while referencing the Provider Models being utilized.

Overview​

Custom Models are user-defined AI models created to perform specific functions within a project. They leverage existing Provider Models from providers such as Anthropic, Google, and OpenAI to enhance utility and customization.

Note: For creating Provider Models, refer to the Models section where instructions for tasks specific to each provider are detailed.

Role Requirements​

To create Custom Models, users must have one of the following roles within Lakehousecat:

  • admin
  • builder

Step-by-Step Guide to Creating Custom Models​

Step 1: Access the Workspace​

  1. Navigate to the Workspace

    • Access your Lakehousecat workspace where model management occurs.

  2. Go to the Models Section

    • Click on the 'Models' menu to explore available modeling options.

Step 2: Initiate Custom Model Creation​

  1. Select "Custom Models"

    • Click on "Custom Models" to start the creation process.

  2. Begin Creation

    • Click the Plus (+) icon to create a new Custom Model.

Step 3: Define the Custom Model​

  • Model Name: Choose and enter a clear, informative name for your custom model. A consistent naming convention helps maintain clarity across different models.
  • Link Foundation Model: In the Combo-Box, select the Foundation Model with which your model will be associated. This selection determines where the requests will be directed.
  • Description: Optionally, enter or dictate a description to convey the model's purpose or scope.

Step 4: Validate and Finalize​

  1. Complete the Creation Process
    • After inputting the necessary information, proceed with the model creation.
  2. Internal Validation
    • Lakehousecat will perform an internal validation. Ensure all parameters are correct for successful validation.
  3. Finalize Custom Model
    • Once validation is successful, your custom model will be created and ready for implementation.

Best Practices​

  • Follow Naming Conventions: Ensure the names of custom models are distinct and informative to facilitate smooth model recognition and management.
  • Utilize Descriptions: Providing detailed descriptions aids in understanding the model's function, especially when managing multiple custom models.

Troubleshooting​

If issues arise during model creation:

  • Validation Errors: Confirm you have selected the correct Provider Models and that all input details are complete.
  • Access Permissions: Verify that you have the necessary roles and permissions within Lakehousecat.