[Relay] Remove overwriting of matmul shapes when they are static#13615
[Relay] Remove overwriting of matmul shapes when they are static#13615slyubomirsky merged 3 commits intoapache:mainfrom
Conversation
|
Thanks for contributing to TVM! Please refer to the contributing guidelines https://tvm.apache.org/docs/contribute/ for useful information and tips. Please request code reviews from Reviewers by @-ing them in a comment.
Generated by tvm-bot |
1 similar comment
|
Thanks for contributing to TVM! Please refer to the contributing guidelines https://tvm.apache.org/docs/contribute/ for useful information and tips. Please request code reviews from Reviewers by @-ing them in a comment.
Generated by tvm-bot |
slyubomirsky
left a comment
slyubomirsky
left a comment
There was a problem hiding this comment.
The changes seem reasonable, but I think we should add at least some test cases that demonstrate the difference in logic to avoid possible regression.
|
I've added a check to |
slyubomirsky
left a comment
slyubomirsky
left a comment
There was a problem hiding this comment.
Thanks for adding in the assertion. I'm surprised this quirk was lurking in an existing test case.
…che#13615) In the Relay Matmul shape relation, we are a little over enthusiastic about unifying dynamic shapes. If one of the shapes is static, it does not need to be unified. This change only rewrites dynamic shapes to required static constraints. * Remove overwriting of matmul shapes when they are static * Simplify nesting * Add shape check to dense tests.
…che#13615) In the Relay Matmul shape relation, we are a little over enthusiastic about unifying dynamic shapes. If one of the shapes is static, it does not need to be unified. This change only rewrites dynamic shapes to required static constraints. * Remove overwriting of matmul shapes when they are static * Simplify nesting * Add shape check to dense tests.
5 participants
In the Relay Matmul shape relation, we are a little over enthusiastic about unifying dynamic shapes. If one of the shapes is static, it does not need to be unified. This change only rewrites dynamic shapes to required static constraints.