CODEWITHSUNDEEP
tensorflowconcatenationkeras
Tavakoli
Tavakoli

Reputation: 1375

How can implement keras's merge whith mode of sum

I want implement merge([a, b], mode='sum') in tensorflow. What is difference between mode='sum' and mode='concat' in keras?

and how can implement merge([a, b], mode='sum') in tensorlfow?

Upvotes: 2

Views: 1541

Answers (1)

chasep255
chasep255

Reputation: 12175

Say you have two tensors like [1, 2] and [2, 3]. If you use mode = 'concat' it will concatenate them and give [1, 2, 3, 4]. If you use mode = 'sum' it will add them element wise and give a result [1 + 2, 3 + 2] = [3, 5]. Sum can be implemented using the + operator in tensorflow or the tf.add function. The concat can be used in tensorflow using the tf.concat function.

Upvotes: 2

Related Questions