10.2. Getting Started (GUI)

This page outlines how to train a first model with Edge AI Studio Model Composer.

Note

Step names follow the 1.4.0 guide and were not re-verified against the current web UI. If a menu or button is named differently, follow the live application.

10.2.1. Prerequisites

  1. A TI account (https://www.ti.com)

  2. A current browser (Chrome, Firefox or Edge)

  3. A dataset, or one of the built-in example datasets

10.2.2. Step 1: Open Model Composer

  1. Go to https://dev.ti.com/edgeaistudio/ and log in.

  2. Select Model Composer and create a new project.

10.2.3. Step 2: Create the project

  1. Enter a project name (for example “Arc Fault Detection”).

  2. Select the task type (for example time series classification).

  3. Select the target device (for example F28P55).

10.2.4. Step 3: Add a dataset

Use an example dataset, or upload your own as a ZIP file. The expected layout of a custom classification dataset is described in Classification Dataset Format (other tasks: Bring Your Own Data). Preview the data to verify it after upload.

10.2.5. Step 4: Feature extraction

Pick a feature extraction preset, or customize the parameters. Start with a preset. The presets correspond to the feature_extraction_name values of the command line flow, see Feature Extraction.

10.2.6. Step 5: Training settings

Select a model and set the number of epochs, batch size and learning rate. For NPU devices (F28P55, AM13E2, MSPM0G5187) choose an NPU-compliant model and enable quantization. Model names and their target devices are listed in Model Zoo Reference.

10.2.7. Step 6: Train

Start training and follow the progress and the loss graph. When it finishes, the GUI shows a results summary.

10.2.8. Step 7: Analyze and test

Review the accuracy, per-class results and confusion matrix, then evaluate on the test set. If the test results are much lower than training, the model may be overfitting.

10.2.9. Step 8: Export

Export the compiled artifacts or the configuration. See Exporting Models.

10.2.10. Next steps