tests: Adding additional test cases for the unowned tensor feature #3993
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
![github-actions[bot]](https://avatars.githubusercontent.com/in/15368?s=60&v=4){kind=small}
There was a problem hiding this comment.
There are some changes that do not conform to Python style guidelines:
--- /home/runner/work/TensorRT/TensorRT/tests/py/dynamo/runtime/test_pre_allocated_outputs.py 2026-01-06 00:40:57.487398+00:00
+++ /home/runner/work/TensorRT/TensorRT/tests/py/dynamo/runtime/test_pre_allocated_outputs.py 2026-01-06 00:41:41.621584+00:00
@@ -163,16 +163,18 @@
self.assertTrue(output_tensors[0] is new_output_tensors[0])
# Run to run, output of output engine is reallocated
self.assertTrue(output_tensors[1] is not new_output_tensors[1])
@parameterized.expand(
- [
- ("python_runtime", True),
- ("cpp_runtime", False),
- ]
- )
- def test_pre_allocated_outputs_unowned_outputs_api_check(self, _, use_python_runtime):
+ [
+ ("python_runtime", True),
+ ("cpp_runtime", False),
+ ]
+ )
+ def test_pre_allocated_outputs_unowned_outputs_api_check(
+ self, _, use_python_runtime
+ ):
class SampleModel(torch.nn.Module):
def forward(self, x):
return torch.softmax(x * 7 + 2, dim=0)
model = SampleModel().eval().cuda()
@@ -191,20 +193,36 @@
)
with torchtrt.runtime.enable_pre_allocated_outputs(optimized_model):
_ = optimized_model(inputs[0])
if use_python_runtime:
- self.assertTrue(all(seen == expected for seen, expected in zip([
- optimized_model._run_on_acc_0.are_output_tensors_unowned(),
- optimized_model._run_on_acc_2.are_output_tensors_unowned()
- ], [False, True])))
+ self.assertTrue(
+ all(
+ seen == expected
+ for seen, expected in zip(
+ [
+ optimized_model._run_on_acc_0.are_output_tensors_unowned(),
+ optimized_model._run_on_acc_2.are_output_tensors_unowned(),
+ ],
+ [False, True],
+ )
+ )
+ )
else:
- self.assertTrue(all(seen == expected for seen, expected in zip([
- optimized_model._run_on_acc_0.engine.are_output_tensors_unowned(),
- optimized_model._run_on_acc_2.engine.are_output_tensors_unowned()
- ], [False, True])))
+ self.assertTrue(
+ all(
+ seen == expected
+ for seen, expected in zip(
+ [
+ optimized_model._run_on_acc_0.engine.are_output_tensors_unowned(),
+ optimized_model._run_on_acc_2.engine.are_output_tensors_unowned(),
+ ],
+ [False, True],
+ )
+ )
+ )
@parameterized.expand(
[
("python_runtime", True),
("cpp_runtime", False),
@@ -256,12 +274,13 @@
check_dtype=True,
)
torch._dynamo.reset()
-
- def test_pre_allocated_outputs_unowned_outputs_multiple_outputs_py_api_check_no_realloc(self):
+ def test_pre_allocated_outputs_unowned_outputs_multiple_outputs_py_api_check_no_realloc(
+ self,
+ ):
class SampleModel(torch.nn.Module):
def forward(self, x):
y = torch.ops.aten.mul(x, 7)
z = torch.ops.aten.add(y, 2)
a = torch.ops.aten.softmax(z, dim=0)
@@ -306,11 +325,13 @@
[
("python_runtime", True),
("cpp_runtime", False),
]
)
- def test_pre_allocated_outputs_unowned_outputs_multiple_outputs_api_check(self, _, use_python_runtime):
+ def test_pre_allocated_outputs_unowned_outputs_multiple_outputs_api_check(
+ self, _, use_python_runtime
+ ):
class SampleModel(torch.nn.Module):
def forward(self, x):
y = torch.ops.aten.mul(x, 7)
z = torch.ops.aten.add(y, 2)
a = torch.ops.aten.softmax(z, dim=0)
@@ -332,28 +353,46 @@
)
with torchtrt.runtime.enable_pre_allocated_outputs(optimized_model):
_ = optimized_model(inputs[0])
if use_python_runtime:
- self.assertTrue(all(seen == expected for seen, expected in zip([
- optimized_model._run_on_acc_0.are_output_tensors_unowned(),
- optimized_model._run_on_acc_2.are_output_tensors_unowned()
- ], [True, True])))
+ self.assertTrue(
+ all(
+ seen == expected
+ for seen, expected in zip(
+ [
+ optimized_model._run_on_acc_0.are_output_tensors_unowned(),
+ optimized_model._run_on_acc_2.are_output_tensors_unowned(),
+ ],
+ [True, True],
+ )
+ )
+ )
else:
- self.assertTrue(all(seen == expected for seen, expected in zip([
- optimized_model._run_on_acc_0.engine.are_output_tensors_unowned(),
- optimized_model._run_on_acc_2.engine.are_output_tensors_unowned()
- ], [True, True])))
-
- @parameterized.expand(
- [
- ("python_runtime", True),
- ("cpp_runtime", False),
- ]
- )
- def test_pre_allocated_outputs_unowned_outputs_multi_outputs(self, _, use_python_runtime):
+ self.assertTrue(
+ all(
+ seen == expected
+ for seen, expected in zip(
+ [
+ optimized_model._run_on_acc_0.engine.are_output_tensors_unowned(),
+ optimized_model._run_on_acc_2.engine.are_output_tensors_unowned(),
+ ],
+ [True, True],
+ )
+ )
+ )
+
+ @parameterized.expand(
+ [
+ ("python_runtime", True),
+ ("cpp_runtime", False),
+ ]
+ )
+ def test_pre_allocated_outputs_unowned_outputs_multi_outputs(
+ self, _, use_python_runtime
+ ):
class SampleModel(torch.nn.Module):
def forward(self, x):
y = torch.ops.aten.mul(x, 7)
z = torch.ops.aten.add(y, 2)
a = torch.ops.aten.softmax(z, dim=0)
There was a problem hiding this comment.
There are some changes that do not conform to Python style guidelines:
--- /home/runner/work/TensorRT/TensorRT/tests/py/dynamo/runtime/test_pre_allocated_outputs.py 2026-01-06 00:41:47.593434+00:00
+++ /home/runner/work/TensorRT/TensorRT/tests/py/dynamo/runtime/test_pre_allocated_outputs.py 2026-01-06 00:42:31.114551+00:00
@@ -163,16 +163,18 @@
self.assertTrue(output_tensors[0] is new_output_tensors[0])
# Run to run, output of output engine is reallocated
self.assertTrue(output_tensors[1] is not new_output_tensors[1])
@parameterized.expand(
- [
- ("python_runtime", True),
- ("cpp_runtime", False),
- ]
- )
- def test_pre_allocated_outputs_unowned_outputs_api_check(self, _, use_python_runtime):
+ [
+ ("python_runtime", True),
+ ("cpp_runtime", False),
+ ]
+ )
+ def test_pre_allocated_outputs_unowned_outputs_api_check(
+ self, _, use_python_runtime
+ ):
class SampleModel(torch.nn.Module):
def forward(self, x):
return torch.softmax(x * 7 + 2, dim=0)
model = SampleModel().eval().cuda()
@@ -191,20 +193,36 @@
)
with torchtrt.runtime.enable_pre_allocated_outputs(optimized_model):
_ = optimized_model(inputs[0])
if use_python_runtime:
- self.assertTrue(all(seen == expected for seen, expected in zip([
- optimized_model._run_on_acc_0.are_output_tensors_unowned(),
- optimized_model._run_on_acc_2.are_output_tensors_unowned()
- ], [False, True])))
+ self.assertTrue(
+ all(
+ seen == expected
+ for seen, expected in zip(
+ [
+ optimized_model._run_on_acc_0.are_output_tensors_unowned(),
+ optimized_model._run_on_acc_2.are_output_tensors_unowned(),
+ ],
+ [False, True],
+ )
+ )
+ )
else:
- self.assertTrue(all(seen == expected for seen, expected in zip([
- optimized_model._run_on_acc_0.engine.are_output_tensors_unowned(),
- optimized_model._run_on_acc_2.engine.are_output_tensors_unowned()
- ], [False, True])))
+ self.assertTrue(
+ all(
+ seen == expected
+ for seen, expected in zip(
+ [
+ optimized_model._run_on_acc_0.engine.are_output_tensors_unowned(),
+ optimized_model._run_on_acc_2.engine.are_output_tensors_unowned(),
+ ],
+ [False, True],
+ )
+ )
+ )
@parameterized.expand(
[
("python_runtime", True),
("cpp_runtime", False),
@@ -256,12 +274,13 @@
check_dtype=True,
)
torch._dynamo.reset()
-
- def test_pre_allocated_outputs_unowned_outputs_multiple_outputs_py_api_check_no_realloc(self):
+ def test_pre_allocated_outputs_unowned_outputs_multiple_outputs_py_api_check_no_realloc(
+ self,
+ ):
class SampleModel(torch.nn.Module):
def forward(self, x):
y = torch.ops.aten.mul(x, 7)
z = torch.ops.aten.add(y, 2)
a = torch.ops.aten.softmax(z, dim=0)
@@ -306,11 +325,13 @@
[
("python_runtime", True),
("cpp_runtime", False),
]
)
- def test_pre_allocated_outputs_unowned_outputs_multiple_outputs_api_check(self, _, use_python_runtime):
+ def test_pre_allocated_outputs_unowned_outputs_multiple_outputs_api_check(
+ self, _, use_python_runtime
+ ):
class SampleModel(torch.nn.Module):
def forward(self, x):
y = torch.ops.aten.mul(x, 7)
z = torch.ops.aten.add(y, 2)
a = torch.ops.aten.softmax(z, dim=0)
@@ -332,28 +353,46 @@
)
with torchtrt.runtime.enable_pre_allocated_outputs(optimized_model):
_ = optimized_model(inputs[0])
if use_python_runtime:
- self.assertTrue(all(seen == expected for seen, expected in zip([
- optimized_model._run_on_acc_0.are_output_tensors_unowned(),
- optimized_model._run_on_acc_2.are_output_tensors_unowned()
- ], [True, True])))
+ self.assertTrue(
+ all(
+ seen == expected
+ for seen, expected in zip(
+ [
+ optimized_model._run_on_acc_0.are_output_tensors_unowned(),
+ optimized_model._run_on_acc_2.are_output_tensors_unowned(),
+ ],
+ [True, True],
+ )
+ )
+ )
else:
- self.assertTrue(all(seen == expected for seen, expected in zip([
- optimized_model._run_on_acc_0.engine.are_output_tensors_unowned(),
- optimized_model._run_on_acc_2.engine.are_output_tensors_unowned()
- ], [True, True])))
-
- @parameterized.expand(
- [
- ("python_runtime", True),
- ("cpp_runtime", False),
- ]
- )
- def test_pre_allocated_outputs_unowned_outputs_multi_outputs(self, _, use_python_runtime):
+ self.assertTrue(
+ all(
+ seen == expected
+ for seen, expected in zip(
+ [
+ optimized_model._run_on_acc_0.engine.are_output_tensors_unowned(),
+ optimized_model._run_on_acc_2.engine.are_output_tensors_unowned(),
+ ],
+ [True, True],
+ )
+ )
+ )
+
+ @parameterized.expand(
+ [
+ ("python_runtime", True),
+ ("cpp_runtime", False),
+ ]
+ )
+ def test_pre_allocated_outputs_unowned_outputs_multi_outputs(
+ self, _, use_python_runtime
+ ):
class SampleModel(torch.nn.Module):
def forward(self, x):
y = torch.ops.aten.mul(x, 7)
z = torch.ops.aten.add(y, 2)
a = torch.ops.aten.softmax(z, dim=0)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
Adds a number of test cases for the unowned tensor feature for pre allocated tensors including C++ and API tests
@cehongwang Feel free to change names or tests as you see fi t
Fixes # (issue)
Type of change
Please delete options that are not relevant and/or add your own.
Checklist: