Continue experimenting

Easy experimenting is a key feature of the Peltarion Platform.

Deep learning is an iterative process. The best way to find the best-performing model is to make changes to your experiments and compare the different experiments.

A good tip is to use names that reminds you of what you did in an experiment. Using good names will make it easier to find the experiment you want to continue working on and in the end the experiment you want to deploy.

Several ways to continue with an experiment

  • Tune experiment
    Use model adjustments that the platform suggests to see if you can potentially improve the performance.
    When tuning, you copy the complete experiment without weights. That is, you do not keep what the model has learned previously.

  • Duplicate the experiment and make your own changes to the model.
    When duplicating, you copy the complete project without weights.

  • Iterate on an experiment to resume training from a selected checkpoint with updated settings.
    When iterating, you keep the weights. That is, you keep what the model has learned previously.

    • Use Continue training when an experiment showed really good progress up to a certain checkpoint but then it wasn’t so good anymore.

    • Reuse part of model lets you reuse a part of your model with weights from a specific checkpoint in a new model. For example, when you want to train a model with additional datasets.

Tips on how to improve an experiment

Suppose you want some tips on how to improve your results. In that case, you can probably find some inspiration in our articles on how to improve experiments—both for beginners and for intermediate users.

A duplicated model will have a link to its parent experiment in the status bar.

Link to parent experiment
Was this page helpful?
Yes No