Skip to content

Pipeline Assistant

model-pipeline-assistant

Supports the use of Models and Model Runtimes in a pipeline

This assistant can be configured to hold multiple models in a sequence of your choice. This allows several models to work one after the other to enable a combination of readily available models or even simple models of your choice.

You can list a component in the marketplace, and define if you want it to be a template.

✅ Available in Marketplace

✅ Can be used as a template to create a new component

What is an Assistant?

Assistants are the building blocks of the automation system. They are the components that decides the actual work of the automation system. They can be used to perform a wide variety of tasks, from sending notifications, building pipelines and organizing workflow.

Options

Below are the options for this assistant.

Option Name Default Required? Type Description
pipeline None False pipeline The pipeline to use
complete_label MODEL-APPLIED False string The label name to apply after processing
training_store None False documentStore The store to use for training models in this pipeline
taxonomies None False list Extraction Data Structure
data_store None False tableStore An instance of a table store that we will use
apply_labels_from_related None False boolean Using source metadata from the taxonomy store, apply labels to the extracted data
write_back_to_store True False boolean Write back the updated document to the data store
## Reactive

This assistant can be triggered by content based events on stores which it is monitoring.



## Custom Events

This assistant also supports custom events that you can use to trigger automations. These events are listed below.



    ### Test


        Run the pipeline against the current document
Option Name Default Required? Type Description
test_only False True boolean Skip training the model and just run the inference
model_options None True pipelineModelOptions The options for the models in the pipeline
enable_logging False True boolean Enable logging of the training process
    ### Train


        Use documents in the training store to train the models in the pipeline
Option Name Default Required? Type Description
training_name {'helper': 'randomName'} True string The name of the training to use
make_active True True boolean The name of the training to use
reprocess True True boolean Reprocess the operational stores after training
model_options None True pipelineModelOptions The options for the models in the pipeline