14EXPERIMENTS WITH CIFAR-10
In this chapter, we’ll perform a series of experiments with the CIFAR-10 dataset we built in Chapter 5. First, we’ll see how two models, one shallow, the other deeper, perform on the full dataset. After that, we’ll work with grouped subsets of the entire dataset to see if we can tell the difference between animals and vehicles. Next, we’ll answer the question of what’s better for the CIFAR-10 dataset, a single multiclass model or a set of binary models, one per class.
We’ll close the chapter by introducing transfer learning and fine-tuning. These are important concepts, often confounded, that are widely used in the ...
Get Practical Deep Learning now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
