Recipes
Last updated
Last updated
A collection of recipes is used to build a machine learning model. Each recipe consists of a series of transformations that are applied to a dataset. These transformations perform specific actions at different stages of the model-building process. You can run these recipes on an uploaded dataset to achieve the desired transformations. After running a recipe, the transformed output is generated, which could be a dataset or a chart. You can then apply additional recipes to this output, continuing the process until you achieve the desired result.
The platform provides users with four distinct types of recipes to build and transform machine learning models:
This type of recipe leverages the power of artificial intelligence to assist users in creating code-based solutions. Users can either allow the AI to generate the code automatically or write their own custom logic using Python. This feature is particularly useful for those who want to develop complex transformations or model-building steps with the help of AI.
This option is designed for users who may not have coding expertise. Without the need for programming or using advanced recipe types, users can quickly build machine learning models with just a few clicks. The Rapid Model recipe simplifies the process, allowing even non-technical users to develop data pipeline and thereby models efficiently.
The Template recipe allows users to apply predefined templates for data transformation and model-building tasks. These templates are pre-configured to handle common data transformation operations, streamlining the process for users who need ready-made solutions for their datasets.
This recipe type is designed for users who wish to write their own Python code. By writing Python code, users can create or manipulate datasets by connecting to external services, running recipes on the dataset, and generating the desired outcomes.
You can easily duplicate a recipe within the canvas, creating an exact copy of the selected recipe. The duplicated recipe inherits all properties of the original, which you can update as needed.
To duplicate a recipe:
Open your project and navigate to the Canvas view.
Right-click on the recipe node you want to duplicate.
A confirmation box appears, indicating that the duplicate recipe will:
Retain all properties of the original recipe.
Be connected to the same input.
Generate the same outputs (you have to run the duplicate recipe to see the outputs)
Click Yes, Proceed to create the duplicate recipe on the canvas. The duplicated recipe will be in an unbuilt state.
Behavior of Duplicated Recipes:
AskAI Recipe: After duplication, refresh the chat to re-run all previously executed queries and add the recipe code to the pipeline for output generation.
Rapid Model Recipe: Run the duplicated recipe manually to generate the output.
Code Recipe: The entire code in the Code tab is copied along with the recipe structure.
Template Recipe: The transformation and configured values remain the same, but you must update the output dataset name to execute the transformation and generate the output.
See also: