W3, assignment one_hot_matrix function

Course Q&A Deep Learning Specialization Improving Deep Neural Networks: Hyperparameter tun
1

Dear Instructor,

In one_hot_matrix function, what would be the second dimension of tf.reshape(). One is the depth that is 6.

Thank you so much for your help.

Best regards,
Sunita

2

There is one example in the same notebook. Please see “def normalize(image):” function.
This routine convert a (64,64,3) tensor to a (64x64x3,) tensor. See what tf.reshape() does.
“-1” is a special value. Tensor flow automatically calculates the size to keep the total size.
In your case, you can specify “depth” or “-1”. Either should be fine. Then, use the same syntax to specify the shape.

1 Like