[Week1] Excercise 2 - Convolutional_model- AttributeError: 'Flatten' object has no attribute 'shape'

Yeah I’ve been thinking for awhile and I’m pretty sure that I’m follow these steps to make Functional API works properly . But I’m still getting this error.

Where this error comes from and how can I solve it?
Thanks

2 Likes

Solved it. Magic happened.

Hi, i have the same problem… How did you solve it?
Thanks very much :slight_smile:

3 Likes

hey still have that prob?

No, I’ve fixed it…thanks :slight_smile:

I have the same problem. How did you guys fix it?

hey consider the input in each statement Z1, A1, P1, Z2, A2, P2 and F, you need to set proper input

2 Likes

Im also facing the same error, How did you solve it.

the solution is to always include the previous output (i.e., Z1, A1, P1…) at the end of the next operation, for all the layers in the CNN. E.g.,
A1 = tf.keras.layers.ReLU()(Z1)

2 Likes