New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

minor fixes in MLP backprop #9

Merged

arthurdouillard merged 2 commits into arthurdouillard:master from navivokaj:master

Dec 7, 2021

Contributor

navivokaj commented Dec 6, 2021

Hi Arthur, nice work on Deep Course! I was going through it to have a refresher on CV and noticed some potential errors on mlp.py:

the softmax in self.fit is redundant since it was already performed in self.forward to produce the probs
grad_bh depends on grad_htilde, not grad_h, since tanh is performed on XW+b

navivokaj added 2 commits

December 7, 2021 02:33


          remove softmax since it was already performed in self.forward to prod…

b4bca7f

…uce probs


          grad_bh depends on grad_htilde since tanh is performed on XW+b

ea14da6

Owner

arthurdouillard commented Dec 7, 2021

Oops, you're right. Thanks for the catch!

arthurdouillard merged commit c5385e7 into arthurdouillard:master

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment