Skip to content
This repository was archived by the owner on Apr 1, 2021. It is now read-only.

Map batch matmul from PT to TVM.#95

Open
kimishpatel wants to merge 1 commit intopytorch:masterfrom
kimishpatel:lower_batch_matmul
Open

Map batch matmul from PT to TVM.#95
kimishpatel wants to merge 1 commit intopytorch:masterfrom
kimishpatel:lower_batch_matmul

Conversation

@kimishpatel
Copy link
Contributor

Also added test.

Summary:

Test Plan:

Reviewers:

Subscribers:

Tasks:

Tags:


auto transpose_attrs = tvm::make_node<tvm::relay::TransposeAttrs>();
auto& axes = transpose_attrs->axes;
axes.push_back(0);
Copy link

Choose a reason for hiding this comment

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

Why this specific order?

@yinghai
Copy link

yinghai commented Aug 20, 2019

@kimishpatel
Copy link
Contributor Author

kimishpatel commented Aug 20, 2019 via email

@yinghai
Copy link

yinghai commented Aug 20, 2019

Hmm, try rebasing and it should work.

Also added test.

Summary:

Test Plan:

Reviewers:

Subscribers:

Tasks:

Tags:
@yinghai
Copy link

yinghai commented Aug 22, 2019

Rebase to master and we should expect CI to pass.

@kimishpatel
Copy link
Contributor Author

This one is causing some correctness issues when run with bert model, even though unit tests pass. So I need to do some investigation before pushing it in.

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants