Skip to content
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

[Relay] Fix Fuse #3035

Merged
merged 4 commits into from
Apr 17, 2019
Merged

[Relay] Fix Fuse #3035

merged 4 commits into from
Apr 17, 2019

Conversation

MarisaKirisame
Copy link
Contributor

Thanks for contributing to TVM! Please refer to guideline https://docs.tvm.ai/contribute/ for useful information and tips. After the pull request is submitted, please request code reviews from Reviewers.

@MarisaKirisame
Copy link
Contributor Author

@tqchen @masahi this fix #3026 .

@@ -865,9 +865,17 @@ class FuseMutator : private ExprMutator {
}

Expr MakeNewFunction(GraphPartitioner::Group* group, Type ret_type, Expr body) {
// If the function has no call, it is not a primitive function.
struct HasCallVisitor : ExprVisitor {
bool HasCall = false;
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

should be snake case

Copy link
Member

@masahi masahi left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Great thanks. Just a minor nit

@masahi masahi merged commit 8d50312 into apache:master Apr 17, 2019
@tqchen
Copy link
Member

tqchen commented Apr 17, 2019

I believe that we will need a bit more discussion about what is happening here. In particular, the usage of fusor and primitive functions. There might be a better fix to this problem e.g. (not fuse these ops)

@MarisaKirisame
Copy link
Contributor Author

@tqchen I believe the fusion algorithm need some update to handle tuple correctly - see #2931 . I think we should chat a bit as I am confused about the code. In particluar sometimes information flow forward and sometimes backward (but in dataflow analysis it is always one directional AFAICT).

@tqchen
Copy link
Member

tqchen commented Apr 17, 2019

Move discussion to #3039

wweic pushed a commit to wweic/tvm that referenced this pull request May 13, 2019
* save

* fix

* Update fuse_ops.cc
wweic pushed a commit to neo-ai/tvm that referenced this pull request May 13, 2019
* save

* fix

* Update fuse_ops.cc
@MarisaKirisame MarisaKirisame deleted the fuse branch July 13, 2019 20:59
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants