Deep Learning with PyTorch

Underfitting

There are times when our model may fail to learn any patterns from our training data, which will be quite evident when the model fails to perform well even on the dataset it is trained on. One common thing to try when your model underfits is to acquire more data for the algorithm to train on. Another approach is to increase the complexity of the model by increasing the number of layers or by increasing the number of weights or parameters used by the model. It is often a good practice not to use any of the aforementioned regularization techniques until we actually overfit the dataset.