@pratik_domadiya Please don’t share your notebook with a learner who hasn’t completed the course.
@geerishajain1 Don’t ask for the notebook 
1 Like
The starter code for windowed_dataset predicts 1 timestep. Your code predicts for multiple timesteps. This is why the grader complains of incompatible shapes.
Please code at places only where you have to. Don’t change the notebook structure like adding / removing cells / print statements.
1 Like
@Huseyin_kocak
Please start with print(model.summary()) and look at the output shape.
1 Like
I have the same issue. Does anyone help me?
1 Like
I just gave you a hint.
2 Likes
(None, None, 1) is my output shape. how can i fix
1 Like
i figured it out. thanks…
1 Like
Model output should have shape (None, 1). Your model has output shape (None, None, 1). Please fix this.
Can you help me how to fix this problem?
1 Like
What did you learn from print(model.summary()) and the return value of return_sequences in a recurrent layer?
Read this reply as well.
1 Like
@Ved_Patel Have you seen this ?
2 Likes
hi, guys. if someone has a problem with this make sure to have only a LSTM layer, this is gonna solve the problem. and use input_shape=[None, 1] in the first and only convolutional layer Conv1D
2 Likes