Google Professional Machine Learning Engineer
Get started today
Ultimate access to all questions.
Ultimate access to all questions.
You recently developed a deep learning model for a classification task using a large dataset consisting of millions of samples. After training the model for several epochs, you observed that both the training and validation losses remained almost constant and did not decrease. To identify the problem and improve your model, what should you do first?
Explanation:
The correct answer is A. Verifying that your model can obtain a low loss on a small subset of the dataset is a good first step for debugging. This helps to determine whether your model is capable of fitting the data. If your model struggles to learn even on a small dataset, it indicates a more fundamental problem in the model architecture, data preprocessing, or learning algorithm. Debugging with a small dataset is faster, allowing you to iterate through potential solutions more quickly and resolve issues before moving on to larger-scale training.