Model configuration
In this section we assume you have access to an edge device with the Scailable AI manager installed and that you have registered your device.
This section describes how to select AI models that are available to you from the AI manager. You can also deploy models directly to a device from the Scailable platform. Furthermore, if you are looking to create your own models for deployment, please see our documentation for data scientist.
Selecting an AI model to to run on your respective edge device is surprisingly simple. On the "Model" tab, simply click the "Select model" button.
When clicking the select model button you will be redirected to the Scailable platform. You might need to log in with the account credentials associated with your user account. Also, the edge device will need to have an internet connection at this point.
After clicking the "Select model" button you will end up at the Scailable platform and you will be presented with a list of models suitable for your device. You can simply browse our library models (and see their documentation) and browse your own user-generated models.
You can select a model for deployment to the device that you are configuring.
At this point you can select only a single model to run on the device. Please note that it is possible to run multiple models in parallel through the advanced settings.
Once you have selected your model you will see the model name and model id as presented below:
Please note that model names don't need to be unique, and hence make sure you are selecting the model with the correct model id.
Finally, note that the table of key's and values specifies the model input dimensions and model normalization; these settings are used by the AI manager to properly pre-process the image coming from the camera and these setting are created when creating a new model.
You can learn more about individual models, their intended use, and their output, by browsing the model library within the Scailable platform.
Last modified 6d ago