Read the Train/Val Gap

Overfitting happens when your model memorizes the training data instead of learning patterns that work on new, unseen examples. 🧠

You track two losses: the training loss on data the model learns from, and the validation loss on data it never updates on.