Remove boilerplate examples and add a text at the dialect level to describe
what kind of operands the operations accept (i.e., scalar, tensor or vector).
Left a shorter sentence describing the input operands for each operation as
this redundancy is convenient when browsing the documentation using the
website.
Details
Details
- Reviewers
ftynse - Commits
- rG850dbff708f1: [MLIR][Math] Improve docs (NFC)
Diff Detail
Diff Detail
- Repository
- rG LLVM Github Monorepo
Event Timeline
Comment Actions
I wonder if it would be better to just add a blob to the top of dialect documentation mentioning that all operations apply elementwise to vectors and tensors unless explicitly specified otherwise, rather than adding more boilerplate examples.
Comment Actions
Right, perhaps it is better. I just did this to be consistent with other operations, but we can avoid boilerplate. I will do it.
Comment Actions
Remove boilerplate examples and add a text at the dialect level to describe
what kind of operands the operations accept (i.e., scalar, tensor or vector).
Left a shorter sentence describing the input operands for each operation as
this redundancy is convenient when browsing the documentation using the
website.
mlir/include/mlir/Dialect/Math/IR/MathBase.td | ||
---|---|---|
18 | One example here still makes sense :) |
One example here still makes sense :)