Skip to main content

6. Train Model

Training your model process uses machine learning to teach the model to recognize different vegetation types based on your ground-truth points.

Model Training Overview​

Training a classification model is a critical step in generating accurate classification results using TytonAI. The Train Model workflow wizard allows you to fine-tune an existing model or train one from scratch using your labeled training data. This process uses powerful AI infrastructure (MLaaS) and TytonAI's MegaModel backbone to deliver high-performing, custom vegetation and erosion classifiers.

Key Considerations​

Before starting the training process, it's important to understand that model training:

  • Training consumes significant computational resources.
  • Incurs cloud computing costs (based on the number of epochs/area).
  • Fine-tuning an existing model is more cost- and time-efficient than training from scratch.

Training Workflow​

1. Select Training Type​

Use the top toggle bar to choose your training method:

  • Fine Tune (Recommended): Improves an existing Mega Model using new training data.
  • Scratch Train: Builds a model from scratch without using a Mega Model.

2. Enter Model Details​

  • Model Name: Enter a descriptive name for your new model.
  • Select Mega Model: Choose the Mega Model to use for fine-tuning. This option is hidden if using Scratch Train.

3. Select Input Data​

  • Select Assessment Data: Choose the Assessment points to be used as source of truth for training your model.
  • Select Training Areas: Choose the training areas you have created to help train your new model.
  • Select Training Datasets: Include published training datasets from other projects you've created, if applicable.

4. Choose Classes for Training​

Use the checklist to include the vegetation classes the model should learn. Examples are:

  • Ground.
  • Herb.
  • Shrub.
  • Tree.
  • Grass.
  • Sedge.

At least one class must be selected to proceed with training.

5. Choose Training Preset​

  • Number of Epochs: Define how many epochs to be trained for the model. Each epoch is one full pass through your dataset.
  • Training Preset: Choose a preset configuration (Safe, Turbo, Experimental ) or select from predefined options. (Safe Recommended)
  • Enable Show preset values to view and modify the default configuration parameters.

6. (Optional) Additional Training Settings​

Expand Show additional options to access advanced model tuning parameters such as:

  • Learning rate.
  • Batch size.
  • Class weighting strategies.
  • Image augmentations.

7. Review Epoch Pricing​

  • Pricing is shown at the bottom of the Train Model panel.
  • Each epoch typically costs 10 credits (e.g. 10 epochs = 100 credits).

8. Train your Model​

  • Click Train Model to queue your model for training.
  • A confirmation message will appear, and training progress will be visible under Job History.

After Training Starts​

  • The training progress appears in the Job History panel.
  • Once training is complete, your model becomes available in your model library under Library > Models.
  • You can use the model for classification once training finishes.
Improving Training Results

For optimal model performance:

  • Ensure good distribution of accuracy points across your study area.
  • Include examples of each vegetation type in different landforms.
  • Consider seasonal variations in vegetation appearance.
  • Include examples of vegetation at different growth stages.
Cost Management

To optimize cloud computing costs:

  • Plan your training data carefully before starting.
  • Combine multiple areas into a single training run.
  • Use fine-tuning instead of training from scratch.
  • Consider sharing trained models across similar projects.