Train Models with SmartML

Train a model with your dataset using SmartML

Start training computer vision models right from Sense using your labeled datasets. With our SmartML model training backend, you can easily begin a training run with a few clicks of your mouse, no coding required.

Starting a Training Run

To begin, you will need a pinned version of your dataset. See Versioning for more information.

Training a model from a labeled dataset

After you've created and pinned a dataset version, you'll be able to train a model using your labeled data.

  1. Go to "Train Models" in the side nav

  2. Click "Add Model" at the top right, or select an existing model and click "New Version"

  3. If this is a new model, enter a name for your model.

  4. Select the desired model options for your training run:

  • Dataset and version. Only locked-in or pinned versions can be used.

  • Labels to be used in training. A model is trained on only one label type.

  • Dataset splits for training/validation/test. Percentages must equal 100%.

  • Training length, between 1-24 hrs.

  • Hyperparameters (optional) for advanced settings

5. Click "Save" or "Save & Start Training" when complete.

Your model version will be created and displayed at the top of your model's version list.

Downloading Your Model

Once training is complete, you will be able to download your model output as a zip file.

  1. At the top right of the model version's details page, click the Download Model button

  2. A SmartML output zip file containing the model output is downloaded.

You can also download the output from the model's versions list:

Learn how to test your model and visualize your detections: