Skip to content

feat(atenlib): add ops (new_empty, new_empty_strided) #436

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

Merged
xiaowuhu merged 10 commits into from
Feb 16, 2023
Merged

feat(atenlib): add ops (new_empty, new_empty_strided) #436

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
173d8dd
Update core.py
xiaowuhu Feb 14, 2023
f50cbdd
add new ops
xiaowuhu Feb 14, 2023
a09fc77
Update core.py
xiaowuhu Feb 14, 2023
1fad593
Merge branch 'main' into xiaowu/addOps(0214)
xiaowuhu Feb 15, 2023
1366519
fix comment
xiaowuhu Feb 15, 2023
9e73133
Update core.py
xiaowuhu Feb 15, 2023
d555ff1
Update core.py
xiaowuhu Feb 15, 2023
01fd43f
Update core.py
xiaowuhu Feb 15, 2023
eaf6ce3
Update core.py
xiaowuhu Feb 15, 2023
0c98ee8
Merge branch 'main' into xiaowu/addOps(0214)
xiaowuhu Feb 16, 2023
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
12 changes: 8 additions & 4 deletions onnxscript/function_libs/torch_aten/ops/core.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3930,16 +3930,20 @@ def aten_negative(self: TensorType) -> TensorType:
raise NotImplementedError()


def aten_new_empty(self: TensorType, size: INT64) -> TensorType:
@torch_op("aten::new_empty")
def aten_new_empty(self: TTensor, size: INT64) -> TTensor:
# new_empty(Tensor self, SymInt[] size, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor

raise NotImplementedError()
zero = op.Constant(value_float=0.0)
return op.Expand(zero, size)


def aten_new_empty_strided(self: TensorType, size: INT64, stride: INT64) -> TensorType:
@torch_op("aten::new_empty_strided")
Copy link
Copy Markdown
Collaborator

@justinchuby justinchuby Feb 15, 2023

Choose a reason for hiding this comment

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

potentially trace only

def aten_new_empty_strided(self: TTensor, size: INT64, stride: INT64) -> TTensor:
# new_empty_strided(Tensor self, SymInt[] size, SymInt[] stride, *, ScalarType? dtype=None, Layout? layout=None, Device? device=None, bool? pin_memory=None) -> Tensor

raise NotImplementedError()
zero = op.Constant(value_float=0.0)
return op.ConstantOfShape(size)


@torch_op("aten::new_full")
Expand Down
4 changes: 4 additions & 0 deletions onnxscript/tests/function_libs/torch_aten/ops_correctness_test.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -302,6 +302,8 @@ def _where_input_wrangler(
"mul": core_ops.aten_mul,
"ne": core_ops.aten_ne,
"neg": core_ops.aten_neg,
"new_empty": core_ops.aten_new_empty,
"new_empty_strided": core_ops.aten_new_empty_strided,
"new_full": core_ops.aten_new_full,
"nn.functional.adaptive_avg_pool1d": nn_ops.aten_adaptive_avg_pool1d,
"nn.functional.adaptive_avg_pool2d": nn_ops.aten_adaptive_avg_pool2d,
Expand Down Expand Up @@ -396,6 +398,8 @@ def _where_input_wrangler(
skip("empty_like", reason="Using zeros_like to simulate empty_like"),
xfail("logcumsumexp", reason="naive implementation not numerically stable"),
xfail("logsumexp", reason="ONNX Runtime 1.13 does not support ReduceLogSumExp-18"),
skip("new_empty", reason="Using zeros to simulate empty"),
skip("new_empty_strided", reason="Using zeros to simulate empty"),
xfail(
"nn.functional.upsample_nearest2d",
reason="enable when ONNX Runtime does support opset18",
Expand Down