CODEWITHSUNDEEP
neural-networktensorflow
Wayne
Wayne

Reputation: 123

Construct a layer with tensorflow to normalize a matrix/tensor

Using conv2d_transpose() function, a matrix could be generated with size m*n (only 1 channel/filter). I just wonder how to normalize this matrix by column to make sure the sum of each column is one. It is very similar as softmax activation, but I don't know how to apply softmax activation to each column either.

Upvotes: 0

Views: 379

Answers (1)

Wayne
Wayne

Reputation: 123

I just figure this out using tf.reduce_sum() and tf_div(). If you have the same question, please let me know.

Upvotes: 1

Related Questions