Not understand about how binary cross entropy calculated in tf

Course Q&A Deep Learning Specialization Improving Deep Neural Networks: Hyperparameter tun
1

mean_reduce(max(logits, 0) - logits * labels + log(1 + exp(-abs(logits))), axis=-1)

Is it this is same as L= -y*log(a_L)+(1-y)*log(1-a_L) ?

2

Hi, @LimXiuXian96.

Here’s the answer to your question.

Happy learning :slight_smile:

1 Like