When using the Neuton framework, only a single training session can be created.
If you selected Axon NPU as the technology, the LiteRT framework is automatically applied. In this case, you can create multiple training sessions with different model architectures, allowing you to compare results and choose the most suitable model for your use case.
All sessions are displayed in the Sessions tab, where you can monitor training progress and evaluate model performance. Each session shows the following information:
- Sessions — A unique identifier assigned to the training session.
- Name — The name of the training session.
- Axon version — The Axon compiler version the model is now compiled with (only available for the LiteRT framework).
- Created — The date and time when the session was created.
- Framework — The framework used for training (Neuton or LiteRT).
- Status — The current status of the training session (In Progress, Completed, or Stopped).
- Evaluation metric — The best value of the evaluation metric achieved during training (for example, accuracy for classification tasks or RMSE for regression tasks).