If your data is still very large, make sure you are using the right validation option. When you open a new session in the Regression Learner app and select the data, cross-validation is selected by default.
Cross-validation helps you partition your data into some number of folds (k), trains the model, and calculates the average test error over all folds. This method provides better protection against overfitting compared with other options, but requires multiple fits, so it works well for small and medium-sized data sets.
Holdout validation helps you select a percentage of the data to use as a testing set using the slider control. The app will train a model on the training set and assess its performance with the testing set. The model used for testing is based on only a portion of the data, so holdout validation is especially appropriate for large data sets.
You can of course also opt not to validate your model, but that opens you up to overfitting to the training data. Learn more about validation options for regression problems.