17 Convnets in Practice
In the last chapter we discussed convolution, and we wrapped up with a simplified example of a convolutional network, or convnet.
In this chapter, we look at two real convnets designed for image classification. The first identifies grayscale handwritten digits, and the second identifies what object is dominant in a color photograph, choosing from 1,000 different categories.
Categorizing Handwritten Digits
Categorizing handwritten digits is a famous problem in machine learning (LeCun et al. 1989), thanks to a freely available dataset called MNIST (pronounced em´-nist). It contains 60,000 hand-drawn digits from 0 to ...
Get 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.
