Once you’ve selected the data and the model, the next step in your Foundry process is to define your model run parameters and (optionally) verify results.
Specific settings vary according to the selected model and what you’re attempting to do. For example, some models require ontologies while others provide advanced settings to refine results.
Here, you:
The following sections discuss typical settings that do not necessarily apply to all situations.
The Model panel displays the name of the selected model and lets you use settings from other model runs using the same model.
Use Load model run config to use settings from earlier model runs. When you do so, a menu displays model runs based on the same model. You can search for specific model runs (case-insensitive).
Some models display a Save as App button, which lets you save the current model run as a Foundry app.
The Ontology panel previews the current ontology and its features.
When Select an ontology is displayed, you need to choose the ontology for the model run. Use Select an ontology to do so.
When you choose this option, the Select ontology view lists available ontologies:
You can only select ontologies compatible with your data and your model.
You can filter the ontology list. Use:
To choose an ontology, select it.
Once you’ve selected an ontology, you can use:
When displayed, the Prompt and Examples options lets you refine how the model interacts with the data. Syntax and options vary according to the model.
Use displayed, you can use Generate example to see a simple example for your model.
To learn more about valid prompts and example syntax, consult the appropriate docs for your model.
Depending on the model, the task, and your data, additional settings may appear. For example, Confidence threshold appears for models that support that option.
Additional settings may also be displayed.
Advanced model settings show additional configuration options supported by your model. These often include hints that describe purpose and acceptable value ranges.
Optionally, you can use Generate preview to verify the results of your current model run settings. This processes your model run against a limited number of data rows (up to five).
Depending on capacity, this can take several minutes to complete, as it involves setting up and running an instance of your selected model.
Select Stop to cancel preview generation.
The Submit button appears only when your settings are valid for the selected model, ontology, and underlying data types. For example, you cannot submit a model run for models that require an ontology until you select an ontology.
When satisfied with the settings and generated previews, name and submit your model run.
At this point, you need to wait until you model run completes. You can track progress or do other things.