From the PDF file, I want to know what is wrong on the funtion of to compute cost? the function is :-

#
GRADED FUNCTION: compute_cost

def compute_cost(logits, labels):

cost=tf.reduce_mean(tf.keras.losses.categorical_crossentropy(y_true=labels,y_pred=logits))

#
YOUR CODE ENDS HERE

`return cost`

Welcome to our community !

I think you are very close. You need to add one more parameter for your “tf.keras.losses.categorical_crossentropy()”.

**from_logits** : Whether y_pred is expected to be a logits tensor. By default, we assume that y_pred encodes a probability distribution.

By the way, it is better not to attach any solution in here. Please remove PDF file. Thank you.

Noted with thanks,

How pdf file is deleted? I tried a couple of options but it was’nt deleted