Formed in 2009, the Archive Team (not to be confused with the archive.org Archive-It Team) is a rogue archivist collective dedicated to saving copies of rapidly dying or deleted websites for the sake of history and digital heritage. The group is 100% composed of volunteers and interested parties, and has expanded into a large amount of related projects for saving online and digital history.
Join GitHub today
GitHub is home to over 50 million developers working together to host and review code, manage projects, and build software together.
Sign upsome mutating tensor operations are not @differentiable #1048
Comments
|
Just starting out with swift for tensorflow. So do we only need to add |
|
Oh yes, I think that is all that you need to do. (If the mutating operations were implemented in terms of other functions that were not differentiable, then you would need to actually specify what the derivative is, but it looks like the mutating operations are defined in terms of functions that are differentiable.) Also some tests that the derivatives work are important. |
|
Is there any way to test the changes without compiling the entire swift for tensorflow toolchain? |
|
Yes, you can use the the toolchain binaries at https://github.com/tensorflow/swift/blob/master/Installation.md if you have one of the supported operating systems. |
|
I meant if I customize the above mentioned file won't I need to recompile the entire thing? |
|
Ah, I see what you mean. The quickest way to test the change is to add a test somewhere in |
|
I would like to work on this. Please let me know if I can pick it. |
+=,-=,*=in https://github.com/tensorflow/swift-apis/blob/master/Sources/TensorFlow/Operators/Math.swift are not differentiable, but it should be pretty straightforward to make them differentiable.