Model pipeline selection and configuration

Nx Witness, Nx Go and Nx Meta will work in the same way for the configuration.

The device detail page is the central place to manage the model pipelines that the device should run. Pipelines are chains of one or more models that can run with the given video input from the device. Usually a device has a default pipeline configured after the plugin is enabled for the device.

Add a new pipeline

When no pipelines are set, the only available option is to add a new pipeline:

A device detail page without any pipelines

Clicking the "Add A New Pipeline" button redirects you to the model catalogue.

Selecting a model

In the model catalogue you can select a model to use in the new (or current) pipeline.

The top of the page will show a message that you are "Assigning a model to...".

If you have uploaded custom models, these will also be available in the catalogue.

The model catalogue — select a model to add to the pipeline

Note that the model visibility is context sensitive. When you are browsing models you can see all models that are available to all organisations that you are a member of. If you are assigning a model to a device, only the models that are available for the organisation that the device belongs to are visible.

When a model is selected it will be assigned and downloaded to the server for the current device.

The selected model is being deployed to the server

You will be redirected to the device details page again, with the newly assigned model.

The device detail page with a single pipeline containing one model

Add a chained model

Chained models are models that use the input of a parent model instead of processing the video from the device directly.

You add a chained model by clicking the "Add Chained Model" button next to the parent model name. You will be redirected to the model catalogue to select a chained model.

Model pipeline — click Add Chained Model to chain a model

The way a chained model will use the input can be chosen by setting a chaining method, currently there are three modes and not all of them may be available at the same time:

• Direct - the chained model gets the output from the parent model as input

• Conditional - the chained model will only run if a field with a given name outputs "true". This is only available if the model itself returns a boolean value.

• Feature extraction - the chained model will get the contents of the bounding boxes that have a certain label. This is only available if the model returns one or more named classes.

A model pipeline with chaining options and feature extraction class selection

Choose the appropriate method, and when you have entered new data the pipeline form will change to indicate that settings need to be saved.

A model pipeline with changes that need to be saved

The settings will be applied to the device when you click the "Save pipelines" button. The changes will be lingering for an hour, unless you either save or cancel the changes.

Change a model in a pipeline

A parent model or a chained model can be replaced by clicking "Switch" button next to the model title. This will redirect you to the normal model selection process where you can select a replacement model.

The rest of the settings in the model pipeline are not affected, unless they are directly related to the model.

Remove a chained model from a pipeline

Removing a chained model from a pipeline can be done by clicking the "Remove" button next to a chained model and the model will be removed directly.

Undoing is not possible, to undo this you will have to select the same model again and set the same chaining options and pre- and post-processing.

Remove a pipeline

Removing a pipeline with all the models and settings can be done by clicking the "Remove" button next to a parent model in a pipeline and the full model pipeline with all associated settings and chained models will be removed.

Undoing is not possible, to restore the settings the same model pipeline must be rebuilt completely.