-
Notifications
You must be signed in to change notification settings - Fork 112
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.
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
Difficulties with nn.sparse_softmax_cross_entropy_with_logits
#124
Comments
I will take a look soonish |
Nope, I'm dead wrong, there's no strict relationship defined between the number of valid label values and the number of output nodes in the net. I was confused between the docs of |
If you alter the shape of your weights from |
I checked; there is a genuine issue with taking gradients of SparseSoftmaxCrossEntryWithLogits operations. Not sure of the cause yet though. |
This is fixed by JuliaIO/ProtoBuf.jl#87 |
Closed by JuliaLang/METADATA.jl#7811 (comment) |
I have been having trouble working out how to use
nn.sparse_softmax_cross_entropy_with_logits
.See this StackOverflow question
I always seem to get
Tensorflow error: Status: Incompatible shapes:
when I run an optimizer over it.
I feel like it would benefit from a test proving it works (and giving an example),
and maybe from some docs if it is different from the python one.
Here is a MWE (smaller than on SO):
Gives me:
The text was updated successfully, but these errors were encountered: