handson-ml3
handson-ml3 copied to clipboard
[BUG]Chapter 11 of the book contains incorrect content
Describe the bug In the third edition of the book, in Chapter 11 of the Max Norm Regulation, in the last paragraph, you write, 'The max_norm() function has an axis argument that defaults to 0.' A Dense layer usually has weights of shape [number of inputs, number of neurons], so using axis=0 means that the max-norm constraint will apply independently to each neuron’s Weight vector ", which looks similar to 'https://www.tensorflow.org/api_docs/python/tf/keras/constraints/MaxNorm'. The explanation given is different, axis=0 represents the input dimension