[AutoDiff upstream] Add derivative function type calculation. #29218
Conversation
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
Add `AnyFunctionType::getAutoDiffDerivativeFunctionType`, which returns the derivative `AnyFunctionType` for an "original" `AnyFunctionType`, given: - Differentiability parameter indices - Differentiability result index - Derivative function kind - Derivative function generic signature (optional) - Other auxiliary parameters Add doc comments explaining typing rules, preconditions, and other details. Progress towards TF-828: upstream `@differentiable` attribute type-checking.
dan-zheng
force-pushed
the
autodiff-upstream-derivative-type
branch
from
ba4a714 to
732e0f4
Compare
|
@swift-ci Please smoke test |
|
macOS flaky test failed (passes locally) Rerunning CI now. |
|
@swift-ci Please smoke test macOS platform |
|
Merging to unblock progress. Happy to address review feedback later. |
dan-zheng
commented
Jan 15, 2020
| resultTupleEltType->getAutoDiffTangentSpace(lookupConformance) | ||
| ->getType()); | ||
| } else { | ||
| assert(resultIndex == 0 && "resultIndex out of bounds"); |
There was a problem hiding this comment.
Minor: this should match the assertion below:
assert(resultIndex == 0 &&
"Expected result index 0 for non-tuple result");
| SmallVector<TupleTypeElt, 2> retElts; | ||
| retElts.push_back(originalResult); | ||
| retElts.push_back(linearMapType); | ||
| auto retTy = TupleType::get(retElts, ctx); |
There was a problem hiding this comment.
Minor: I prefer using the name "result type" instead of "return type". retTy -> resultTy
rxwei
reviewed
Jan 15, 2020
| // Creates an `AnyFunctionType` from the given parameters, result type, | ||
| // generic signature, and `ExtInfo`. | ||
| static AnyFunctionType * | ||
| makeFunctionType(ArrayRef<AnyFunctionType::Param> parameters, Type resultType, |
There was a problem hiding this comment.
This is a general utility for AST function types. Consider adding it to AnyFunctionType, or just making it a closure inside getAutoDiffDerivativeFunctionType.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Add
AnyFunctionType::getAutoDiffDerivativeFunctionType, which returns thederivative
AnyFunctionTypefor an "original"AnyFunctionType, given:Add doc comments explaining typing rules, preconditions, and other details.
Progress towards TF-828: upstream
@differentiableattribute type-checking.@differentiableattribute type-checking mega-patch: #29091Tests exist on
tensorflowbranch and will be upstreamed later.Currently, not enough code has been upstreamed for meaningful testing.