-
Notifications
You must be signed in to change notification settings - Fork 9
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
Stabilize DoubleSym #42
Conversation
|
||
dth1 = JthetaTmv(this.K,dAZ,zeros(T,0),Y) | ||
KopZ = Amv(this.K, th1, Z) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@eldadHaber @lruthotto How attached are you two to the idea of have getOp
return a LinearOperator
? We find that the LinearOperator
does not return a type stable result, even though the prod
function it calls is stable.
I'm not sure what is confusing the compiler, but just skipping the LinearOperator
and calling the function as done here on line 201 fixed the issue. Do you have any problem with us doing this?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I'm not too attached to it although it makes the code more readable. If this is a problem with the LinearOperator package and you can reproduce this easily it might be worth reporting it there.
Probably just need to add Juno to the require file to make the tests pass |
I'll just remove the line, Juno is IDE specific so I don't want to make it
required
…On Feb 5, 2018 1:42 PM, "DavidBegert" ***@***.***> wrote:
Probably just need to add Juno to the require file to make the tests pass
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
<#42 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/ATplBbFpkSIBapugxxkAKXse7RdVakiqks5tR3WqgaJpZM4R3-In>
.
|
src/activations/tanhActivation.jl
Outdated
if doDerivative | ||
dA .= one(T) .- A.^2 | ||
else | ||
fill!(dA, zero(T)) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Dunno if theres a performance diff between this and just creating it as a 2-D array of zeros as on the left. We know that the return will be 2 dimensional so unsure if this is needed xo
No description provided.