C5_W1_Assigment_Build_RNN_Step_By_Step

Hello it seems like I fail to pass this function cause there is some difference between the args “parameters” used in the function and the one used in the test. I passed it one time and after that I try to open the run_utils.py in another tab but it duplicate the assignment page. After that I can’t get the test pass anymore. Have you some idea to recommend? I already try to restart the kernel and even delete all my files and load one from the last update. The problem occurs with the function ‘run_cell_forward’
Here is the message error:
a_next[4] =
[ 0.59584544 0.18141802 0.61311866 0.99808218 0.85016201 0.99980978
-0.18887155 0.99815551 0.6531151 0.82872037]
a_next.shape =
(5, 10)
yt_pred[1] =
[0.88978316 0.02620746 0.08013124 0.09478513 0.64180858 0.27232029
0.30926225 0.90529271 0.95200778 0.24373739]
yt_pred.shape =
(2, 10)

AssertionError Traceback (most recent call last)
in
16
17 # UNIT TESTS
—> 18 rnn_cell_forward_tests(rnn_cell_forward)

~/work/W1A1/public_tests.py in rnn_cell_forward_tests(target)
25
26 assert np.allclose(np.tanh(parameters_tmp[‘ba’]), a_next_tmp), “Problem 1 in a_next expression. Related to ba?”
—> 27 assert np.allclose(softmax(parameters_tmp[‘by’]), yt_pred_tmp), “Problem 1 in yt_pred expression. Related to by?”
28
29 # Only xt in expression

AssertionError: Problem 1 in yt_pred expression. Related to by?

I noticed that your values of a_next[4] is same as the expect but yt_pred[1] is different. So, check how you are implementing the yt_pred. You have to use dot and softmax functions.

Thanks for your help. It was indeed a softmax. I don’t know why I implemented a sigmoid.

I am glad you figured it out…