-
Notifications
You must be signed in to change notification settings - Fork 27.7k
Support torch.linalg.outer #63293
Copy link
Copy link
Open
Labels
featureA request for a proper, new feature.A request for a proper, new feature.module: linear algebraIssues related to specialized linear algebra operations in PyTorch; includes matrix multiply matmulIssues related to specialized linear algebra operations in PyTorch; includes matrix multiply matmulmodule: python array apiIssues related to the Python Array APIIssues related to the Python Array APItriagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module
Metadata
Metadata
Assignees
Labels
featureA request for a proper, new feature.A request for a proper, new feature.module: linear algebraIssues related to specialized linear algebra operations in PyTorch; includes matrix multiply matmulIssues related to specialized linear algebra operations in PyTorch; includes matrix multiply matmulmodule: python array apiIssues related to the Python Array APIIssues related to the Python Array APItriagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module
🚀 Feature
Support
torch.linalg.outer.cc @jianyuh @nikitaved @pearu @mruberry @heitorschueroff @walterddr @IvanYashchuk @xwang233 @lezcano @rgommers @pmeier @asmeurer @leofang @AnirudhDagar @asi1024 @emcastillo @kmaehashi
Motivation & Pitch
Python Array API proposes the use of
linalg.outeramong other linear algebra functions. Currently, PyTorch supports the same functionality withtorch.outer.torch.ger(to be deprecated in 1.10) is also an alias totorch.outerat the moment.We can either make
torch.linalg.outerthe main function withtorch.outerandtorch.geras aliases to that. Or we can keeptorch.outeras is and just aliastorch.linalg.outerto that.The only difference is that Array API doesn't define a keyword-only
out=argument which should be fine since PyTorch's API will just be a superset of Array API here.Additional context