-
Notifications
You must be signed in to change notification settings - Fork 106
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
IdentityOperator should take a range
argument
#1069
Comments
Duplicate of #59? |
Yes, to some degree. But the implementation here would be |
Or, as opposed to the implementation in #95, we wouldn't have to worry about mappings from, say, |
Well, the adjoint would not be as simple so its not trivial to implement. Anyway I think that issue had some work on this as well, as you mention about #95. |
Anyway my point is that I'm not at all sure I want to add this to I'd probably prefer a new class for this, or some other construction anyway. |
That's true, but I think that part will be very widely applicable and can be factored out to some helper function. When you have an operator
Then you can move all the logic to a function that can also distinguish cases, e.g., merge weighting factors when one of them is a constant, or ignore constant 1.0 etc. Then operators just need to use that function, and there's no weird dependency whatsoever. |
Well, currently the whole You could make a PR (most important code will be copy-pastable anyway) and we'll see how it looks. The functionality is needed regardless so this is only a question of where we want it. |
True, but what we call
I'll see what I can do. |
We should allow the identity operator (and together with it
ScalingOperator
and a couple more) to map to a different space. For instance fromrn(5)
touniform_discr(0, 1, 5)
.The text was updated successfully, but these errors were encountered: