In week 1 assignment Building_a_Recurrent_Neural_Network_Step_by_Step. I got this error.
1 Like
Error:
The error indicates that the shapes of variables do not match in the following lines:
a[:,:,t] = a_next
y_pred[:,:,t] = yt_pred
What to do?
Ensure that a, y_pred, a_next, and yt_pred are initialized with the correct shapes. Also, make sure that a_next and yt_pred returned by rnn_cell_forward have the expected shapes.
Hope this helps!
2 Likes
in my opinion, if you understand how broadcasting works, you will be able to solve this issue.
Here is a quick reference-> Broadcasting — NumPy v2.0 Manual
1 Like
If you look at the test case there, you see that n_x = 3 and n_a = 5, so this is some kind of a dimension mismatch problem. E.g. that you are using some version of a where you should be using x or vice versa.
