Hyperparameter optimization using Ray Tune
Neural networks have hyperparameters that define their structure and learning process. Hyperparameters include the learning rate or the number of hidden layers and units. Different hyperparameter values can affect the learning process and the accuracy of models. Incorrectly chosen values can result in underfitting or overfitting, which decreases the model’s performance. So, it’s important to optimize the value of hyperparameters to get the most out of deep learning models. In this recipe, we’ll explore how to do hyperparameter optimization using Ray Tune, including learning rate, regularization parameters, the number of hidden layers, and so on. The optimization of these parameters is very important to the performance of our models. More often than not, we face poor results in fitting neural network models simply due to poor selection of hyperparameters, which can lead to underfitting or overfitting unseen data.
