Building a Default Image Metric Learning Model

This lesson walks through building an image metric learning model in H2O Hydrogen Torch using default hyperparameter values. It covers creating a new experiment, selecting the bicycle_image_metric_learning dataset, naming the experiment, and running it from the Create experiment screen. The lesson explains that the default scorer for this task is Mean Average Precision, where a value closer to 1 indicates the model is better at telling similar and dissimilar bicycle images apart. It closes by noting that results will vary by dataset and modeling choices, setting up the next lesson on hyperparameter tuning.