Manage models
Predictive models are fundamental components of workflows.
NL Flow installations can have a built-in inventory of models called Knowledge Models.
Custom models, either symbolic or ML, can be added to the inventory by publishing them from the authoring application1 or by importing them from package files exported in another Platform application or in Studio.
Manage models through the Models view of the dashboard.
The following articles describe all the operations for managing models.
-
Pay attention to the terminology: to publish a model means copying the model from a project of the authoring application to the model inventory of NL Flow, so that it can be used in workflows.
To publish a workflow, instead, is something you do inside NL Flow, to create, inside a runtime, a program corresponding to the workflow diagram.
See for example this article:that is about publishing models from categorization projects to NL Flow. ↩