Hi, i have a problem in exercise 10.
I passed all the previous tests and i just cant find any solution to finish this assignment. Can you help me?
Hi, i have a problem in exercise 10.
I passed all the previous tests and i just cant find any solution to finish this assignment. Can you help me?
Hi @merveyubogluu it seems something is not correct in the rule you used to update W and b given the larning rate and the derivatives dW and db. Are you sure you have taken the strategy to move in the direction of the gradient with an amplitude modulated by the learning rate? As state in the excercise 10 introduction for b (similar also for W):
b[l] = b[l] - alpha*db[l]
Hope this can help you
Thanks for the response!
I was trying to do something like b[l] = db[l] - alpha*db[l]. Now i fixed it with your help. Thanks again.