Trouble results delivered by rnn_backward

Course Q&A Deep Learning Specialization Sequence Models
1

Hi,

“My” rnn_backward function runs well.
Nevertheless, I don’t find the attempted result.
Here is the result :slight_smile:

gradients[“dx”][1][2] = [-0.15028183 -0.34554547 0.02071758 0.01483317]
gradients[“dx”].shape = (3, 10, 4)
gradients[“da0”][2][3] = -0.17268893183890754
gradients[“da0”].shape = (5, 10)
gradients[“dWax”][3][1] = 4.081485734449453
gradients[“dWax”].shape = (5, 3)
gradients[“dWaa”][1][2] = 1.056012342849445
gradients[“dWaa”].shape = (5, 5)
gradients[“dba”][4] = [-0.12427391]
gradients[“dba”].shape = (5, 1)

I will be very pleased to read your thought about it.
Thank you

2

If 1 and 2 don’t help, please click my name and message your notebook as an attachment.

1 Like
3

@CyrilleB Please read the 1st link in my previous reply and fix rnn_cell_backward.

4

Thank you.
I fhave found my mistakes… 2 :

  • I added da_prevt and so weel-understood why.
  • I had a bad initialization of da_prevt because I added it one more dimension…

I thank you very much for your help !
Have a nice day.
Cyrille