Add support for higher-order derivatives #233
Conversation
…cessary `return` in rrule
yitan1
changed the title
|
Your PR no longer requires formatting changes. Thank you for your contribution! |
Codecov Report✅ All modified and coverable lines are covered by tests.
... and 24 files with indirect coverage changes 🚀 New features to boost your workflow:
|
lkdvos
left a comment
lkdvos
left a comment
There was a problem hiding this comment.
I have only one minor comment, otherwise happy to merge this.
| @non_differentiable TensorOperations.tensorcontract_structure(args...) | ||
| @non_differentiable TensorOperations.tensorcontract_type(args...) | ||
| @non_differentiable TensorOperations.tensoralloc_contract(args...) | ||
| @non_differentiable Base.promote_op(args...) |
There was a problem hiding this comment.
Do you know if your example also works if this isn't included? We can't really do this here since that is type piracy, so we would have to find an alternative way around this...
| @non_differentiable Base.promote_op(args...) |
There was a problem hiding this comment.
Unfortunately, the example fails if this is removed. To avoid type piracy, would this alternative be acceptable?
@non_differentiable TensorOperations.promote_contract(args...)
@non_differentiable TensorOperations.promote_add(args...)There was a problem hiding this comment.
Yes, that is completely fine, thank you!
There was a problem hiding this comment.
Thanks! Updated in the latest commit.
2 participants
This closes #227 by following changes
@non_differentiableforBase.promote_optensorscalarto avoid inplace functionreturnfromrrule