-
Notifications
You must be signed in to change notification settings - Fork 61
Forward AD tests and cleanup for vector interface #436
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
Open
kshyatt
wants to merge
9
commits into
main
Choose a base branch
from
ksh/vi
base: main
Could not load branches
Branch not found: {{ refName }}
Loading
Could not load tags
Nothing to show
Loading
Are you sure you want to change the base?
Some commits from the old base branch may be removed from the timeline,
and old review comments may become outdated.
+25
−109
Open
Changes from all commits
Commits
Show all changes
9 commits
Select commit
Hold shift + click to select a range
333527c
Forward rules for vector interface
kshyatt 00b4c8c
Try falling back to new VI rules
f004820
Fix tangent typo
kshyatt 5f7c771
Some small fixes
kshyatt 2370983
Restore primitive markers
kshyatt bcbc1b3
Update sources
61df722
Test stuff is primitive when it should be
2f79c69
Fix Project.toml now that VI is tagged
cc36f26
Add some adjoint tests for add
File filter
Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -1,89 +1,4 @@ | ||
| @is_primitive DefaultCtx ReverseMode Tuple{typeof(scale!), AbstractTensorMap, Number} | ||
|
|
||
| function Mooncake.rrule!!(::CoDual{typeof(scale!)}, C_ΔC::CoDual{<:AbstractTensorMap}, α_Δα::CoDual{<:Number}) | ||
| # prepare arguments | ||
| C, ΔC = arrayify(C_ΔC) | ||
| α = primal(α_Δα) | ||
|
|
||
| # primal call | ||
| C_cache = copy(C) | ||
| scale!(C, α) | ||
|
|
||
| function scale_pullback(::NoRData) | ||
| copy!(C, C_cache) | ||
| Δαr = _needs_tangent(α) ? project_scalar(α, inner(C, ΔC)) : NoRData() | ||
| scale!(ΔC, conj(α)) | ||
| return NoRData(), NoRData(), Δαr | ||
| end | ||
|
|
||
| return C_ΔC, scale_pullback | ||
| end | ||
|
|
||
| @is_primitive DefaultCtx ReverseMode Tuple{typeof(scale!), AbstractTensorMap, AbstractTensorMap, Number} | ||
|
|
||
| function Mooncake.rrule!!(::CoDual{typeof(scale!)}, C_ΔC::CoDual{<:AbstractTensorMap}, A_ΔA::CoDual{<:AbstractTensorMap}, α_Δα::CoDual{<:Number}) | ||
| # prepare arguments | ||
| C, ΔC = arrayify(C_ΔC) | ||
| A, ΔA = arrayify(A_ΔA) | ||
| α = primal(α_Δα) | ||
|
|
||
| # primal call | ||
| C_cache = copy(C) | ||
| scale!(C, A, α) | ||
|
|
||
| function scale_pullback(::NoRData) | ||
| copy!(C, C_cache) | ||
| add!(ΔA, ΔC, conj(α)) | ||
| Δαr = _needs_tangent(α) ? project_scalar(α, inner(A, ΔC)) : NoRData() | ||
| zerovector!(ΔC) | ||
| return NoRData(), NoRData(), NoRData(), Δαr | ||
| end | ||
|
|
||
| return C_ΔC, scale_pullback | ||
| end | ||
|
|
||
| @is_primitive DefaultCtx ReverseMode Tuple{typeof(add!), AbstractTensorMap, AbstractTensorMap, Number, Number} | ||
|
|
||
| function Mooncake.rrule!!(::CoDual{typeof(add!)}, C_ΔC::CoDual{<:AbstractTensorMap}, A_ΔA::CoDual{<:AbstractTensorMap}, α_Δα::CoDual{<:Number}, β_Δβ::CoDual{<:Number}) | ||
| # prepare arguments | ||
| C, ΔC = arrayify(C_ΔC) | ||
| A, ΔA = arrayify(A_ΔA) | ||
| α = primal(α_Δα) | ||
| β = primal(β_Δβ) | ||
|
|
||
| # primal call | ||
| C_cache = copy(C) | ||
| add!(C, A, α, β) | ||
|
|
||
| function add_pullback(::NoRData) | ||
| copy!(C, C_cache) | ||
|
|
||
| Δαr = _needs_tangent(α) ? project_scalar(α, inner(A, ΔC)) : NoRData() | ||
| Δβr = _needs_tangent(β) ? project_scalar(β, inner(C, ΔC)) : NoRData() | ||
| add!(ΔA, ΔC, conj(α)) | ||
| scale!(ΔC, conj(β)) | ||
|
|
||
| return NoRData(), NoRData(), NoRData(), Δαr, Δβr | ||
| end | ||
|
|
||
| return C_ΔC, add_pullback | ||
| end | ||
|
|
||
| @is_primitive DefaultCtx ReverseMode Tuple{typeof(inner), AbstractTensorMap, AbstractTensorMap} | ||
|
|
||
| function Mooncake.rrule!!(::CoDual{typeof(inner)}, A_ΔA::CoDual{<:AbstractTensorMap}, B_ΔB::CoDual{<:AbstractTensorMap}) | ||
| # prepare arguments | ||
| A, ΔA = arrayify(A_ΔA) | ||
| B, ΔB = arrayify(B_ΔB) | ||
|
|
||
| # primal call | ||
| s = inner(A, B) | ||
|
|
||
| function inner_pullback(Δs) | ||
| add!(ΔA, B, conj(Δs)) | ||
| add!(ΔB, A, Δs) | ||
| return NoRData(), NoRData(), NoRData() | ||
| end | ||
|
|
||
| return CoDual(s, NoFData()), inner_pullback | ||
| end | ||
| @is_primitive DefaultCtx Tuple{typeof(scale!), AbstractTensorMap, Number} | ||
| @is_primitive DefaultCtx Tuple{typeof(scale!), AbstractTensorMap, AbstractTensorMap, Number} | ||
| @is_primitive DefaultCtx Tuple{typeof(add!), AbstractTensorMap, AbstractTensorMap, Number, Number} | ||
| @is_primitive DefaultCtx Tuple{typeof(inner), AbstractTensorMap, AbstractTensorMap} |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Oops, something went wrong.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
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.
Do you know if something like this works:
This would allow us to continue supporting old + new version of VectorInterface.
For the current implementation, I guess this might break the
rruleimplementations that were already there if the correct version of VectorInterface is not loaded. In that case, I would maybe just bite the bullet and actually restrict to VectorInterface v0.6 and just deal with that?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.
Not sure, I can try it tomorrow!
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.
Having to support the old versions again leads to having duplicated rules all over the place, though, one advantage of doing it the current way is being able to cut quite a few lines from the extension.