Testing a Model Using the Data Capture Lab

../_images/dcl-knowledge-pack-test-model.png

The Data Capture Lab has two ways to test a model on your dataset:

  1. Running a Model During Data Collection: Connect to a model during data collection and get the model results in real-time

  2. Running a Model in the Project Explorer: Run a model on any previously collected CSV or WAV files in your project

This lets you see how your model will perform on real data before flashing it to a device. After getting the results from the model you can then save the results to your project and re-train your model using the new results. This is a powerful feature that we will go over in detail below.

Running a Model During Data Collection

The Data Capture Lab has an option to connect to your Knowledge Pack during data collection and see the model results in real-time. You can then save the results to your project and use them for re-training your model. Note - This feature is only supported on devices that implement the Simple Streaming capture protocol

  1. Switch to Capture mode

../_images/dcl-switch-mode-capture.png
  1. Connect to your device

../_images/dcl-sensor-connect1.png
  1. Open the Test Model panel and click Connect

../_images/dcl-knowledge-pack-connect.png
  1. Select a Knowledge Pack

../_images/dcl-knowledge-pack-select.png
  1. Select a Session. This is where the Knowledge Pack results will be saved

../_images/dcl-knowledge-pack-session-select.png
  1. Connect to the Knowledge Pack

../_images/dcl-knowledge-pack-connected-status.png
  1. You will now see your model results in real-time in the graph

../_images/dcl-knowledge-pack-test-model.png
  1. (Optional) You can click Start Recording and the Data Capture Lab will save the Knowledge Pack results to your project. This lets you quickly add additional training data to your project

../_images/dcl-knowledge-pack-recording.png
  1. (Optional) In the Save Confirmation screen you can edit or delete the Knowledge Pack results before saving the results to your project

../_images/dcl-knowledge-pack-results-save-confirmation.png

Running a Model in the Project Explorer

The Project Explorer has an option to connect to a Knowledge Pack and run a model on any previously collected files you have added to your project. You can then save the results to your project and use them for re-training your model.

  1. Open the Project Explorer

../_images/dcl-open-project-explorer.png
  1. Select the files you want to use by holding (Shift + Click) or (Ctrl + Click)

  2. Right-Click and select Segments → Add → From Knowledge Pack

../_images/dcl-project-explorer-knowledge-pack.png
  1. Select a Knowledge Pack

../_images/dcl-knowledge-pack-select.png
  1. Select a Session. This is where the Knowledge Pack results will be saved

../_images/dcl-knowledge-pack-session-select.png
  1. Save the results. (Optional: You can edit or delete the Knowledge Pack results before saving)

../_images/dcl-project-explorer-review-results.png

Automated Labeling Using a Model

By using the model results from either of the two methods above you can re-train your model with new data. By repeating this process you can quickly improve your model with more robust and accurate results.