whereas the respective dataset is referred to astraining set. Here, the model makes predictions and subsequently adjusts its parameters based on a metric that quantifies how good or bad the predictions are (loss function). It is typical that during training, the model will be applied to an ...