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.

Sign up for GitHub

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

add elementwise max/min conversation and test case #92

Merged
merged 5 commits into from
Sep 16, 2022
Merged

add elementwise max/min conversation and test case #92

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
87486e6
add elementwise max/min conversation and test case
doombeaker Sep 14, 2022
6f49eb8
add cpu test
doombeaker Sep 15, 2022
ab9a5fa
Merge branch 'main' into support_op_elementwise_minimum
doombeaker Sep 15, 2022
06f2035
Merge branch 'main' into support_op_elementwise_minimum
doombeaker Sep 16, 2022
2238c2a
Merge branch 'main' into support_op_elementwise_minimum
doombeaker Sep 16, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions examples/oneflow2onnx/nodes/CPU/test_math.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -36,6 +36,7 @@ def forward(self, x: flow.Tensor) -> flow.Tensor:
y3 = flow.acos(y3)
y3 = flow.pow(y3, 2.0)
y2 = y2 + y3
flow.minimum(y2, y3)

B = x # shape: (1, 3, 224, 224)
flow.matmul(self.A, B)
Expand Down
1 change: 1 addition & 0 deletions examples/oneflow2onnx/nodes/GPU/test_math.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -36,6 +36,7 @@ def forward(self, x: flow.Tensor) -> flow.Tensor:
y3 = flow.acos(y3)
y3 = flow.pow(y3, 2.0)
y2 = y2 + y3
flow.minimum(y2, y3)
Copy link
Contributor

Choose a reason for hiding this comment

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

cpu那里的测试也对应加一下吧

Copy link
Contributor Author

Choose a reason for hiding this comment

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

done


B = x # shape: (1, 3, 224, 224)
flow.matmul(self.A, B) # (128, 224) x (1, 3, 224, 224)
Expand Down
2 changes: 2 additions & 0 deletions oneflow_onnx/oneflow2onnx/handlers/math.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -211,6 +211,8 @@ def _MakeMinOrMaxOp(ctx, op_type, inputs, outputs, output_shapes=None, output_dt

@flow_op("broadcast_minimum", onnx_op="Min")
@flow_op("broadcast_maximum", onnx_op="Max")
@flow_op("elementwise_minimum", onnx_op="Min")
@flow_op("elementwise_maximum", onnx_op="Max")
class MinMaxOp:
@classmethod
def Version_1(cls, ctx, node, **kwargs):
Copy link
Contributor Author

Choose a reason for hiding this comment

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

这里我有个疑问，我看 https://github.com/onnx/onnx/blob/main/docs/Changelog.md#Min-1 提到 Version 1 的 Min，是不支持 broadcast 的，所以在下面 _MakeMinOrMaxOp 里做了一些处理。

然而后面较高版本的 onnx op （比如 12 https://github.com/onnx/onnx/blob/main/docs/Changelog.md#Min-12）又支持 broadcast 了。

不过自定义的 trick 好像又够用了，这个是不是暂时就不用添加更高版本的映射了？

Copy link
Contributor

Choose a reason for hiding this comment

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

是的

Expand Down