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

Support torch.export.Dim.AUTO in ONNX conversion pass #1586

Closed
wants to merge 1 commit into from

support torch.export.Dim.AUTO

63bf1c2
Select commit
Loading
Failed to load commit list.
Closed

Support torch.export.Dim.AUTO in ONNX conversion pass #1586

support torch.export.Dim.AUTO
63bf1c2
Select commit
Loading
Failed to load commit list.
Azure Pipelines / Olive CI failed Jan 30, 2025 in 57m 48s

Build #20250129.4 had test failures

8 errors / 2 warnings

Details

Tests

  • Failed: 64 (2.75%)
  • Passed: 2,137 (91.76%)
  • Other: 128 (5.50%)
  • Total: 2,329

Annotations

Check failure on line 10663 in Build log

See this annotation in the file changed.

@azure-pipelines azure-pipelines / Olive CI

Build log #L10663 

Bash exited with code '1'.

Check failure on line 19 in Build log

See this annotation in the file changed.

@azure-pipelines azure-pipelines / Olive CI

Build log #L19 

There are one or more test failures detected in result files. Detailed summary of published test results can be viewed in the Tests tab.

Check failure on line 20 in Build log

See this annotation in the file changed.

@azure-pipelines azure-pipelines / Olive CI

Build log #L20 

There are one or more test failures detected in result files. Detailed summary of published test results can be viewed in the Tests tab.

Check failure on line 19 in Build log

See this annotation in the file changed.

@azure-pipelines azure-pipelines / Olive CI

Build log #L19 

There are one or more test failures detected in result files. Detailed summary of published test results can be viewed in the Tests tab.

Check failure on line 1 in test_resnet[resnet_ptq_cpu.json-local_system-pass-by-pass-random]

See this annotation in the file changed.

@azure-pipelines azure-pipelines / Olive CI

test_resnet[resnet_ptq_cpu.json-local_system-pass-by-pass-random] 

AssertionError: footprints is empty. The search must have failed for all accelerator specs.
Raw output 
search_algorithm = 'random', execution_order = 'pass-by-pass'
system = 'local_system', olive_json = 'resnet_ptq_cpu.json'

    @pytest.mark.parametrize("search_algorithm", ["random"])
    @pytest.mark.parametrize("execution_order", ["pass-by-pass"])
    @pytest.mark.parametrize("system", ["local_system"])
    @pytest.mark.parametrize(
        "olive_json",
        [
            "resnet_ptq_cpu.json",
            # TODO(trajep): remove skip once the bug of azureml-fsspec is fixed
            pytest.param(
                "resnet_ptq_cpu_aml_dataset.json", marks=pytest.mark.skip(reason="credential bug in azureml-fsspec")
            ),
        ],
    )
    @pytest.mark.skipif(
        version.parse(OrtVersion) == version.parse("1.16.0"),
        reason="resnet is not supported in ORT 1.16.0 caused by https://github.com/microsoft/onnxruntime/issues/17627",
    )
    def test_resnet(search_algorithm, execution_order, system, olive_json):
        from olive.workflows import run as olive_run
    
        olive_config = patch_config(olive_json, search_algorithm, execution_order, system)
    
        footprint = olive_run(olive_config, tempdir=os.environ.get("OLIVE_TEMPDIR", None))
>       check_output(footprint)

D:\a\_work\1\s\examples\test\local\test_resnet_ptq_cpu.py:49: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
D:\a\_work\1\s\examples\test\utils.py:20: in check_output
    assert_nodes(footprints)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

footprints = {}

    def assert_nodes(footprints):
>       assert footprints, "footprints is empty. The search must have failed for all accelerator specs."
E       AssertionError: footprints is empty. The search must have failed for all accelerator specs.

D:\a\_work\1\s\examples\test\utils.py:25: AssertionError

Check failure on line 1 in test_resnet[resnet_ptq_cpu.json-local_system-pass-by-pass-random]

See this annotation in the file changed.

@azure-pipelines azure-pipelines / Olive CI

test_resnet[resnet_ptq_cpu.json-local_system-pass-by-pass-random] 

AssertionError: footprints is empty. The search must have failed for all accelerator specs.
Raw output 
search_algorithm = 'random', execution_order = 'pass-by-pass'
system = 'local_system', olive_json = 'resnet_ptq_cpu.json'

    @pytest.mark.parametrize("search_algorithm", ["random"])
    @pytest.mark.parametrize("execution_order", ["pass-by-pass"])
    @pytest.mark.parametrize("system", ["local_system"])
    @pytest.mark.parametrize(
        "olive_json",
        [
            "resnet_ptq_cpu.json",
            # TODO(trajep): remove skip once the bug of azureml-fsspec is fixed
            pytest.param(
                "resnet_ptq_cpu_aml_dataset.json", marks=pytest.mark.skip(reason="credential bug in azureml-fsspec")
            ),
        ],
    )
    @pytest.mark.skipif(
        version.parse(OrtVersion) == version.parse("1.16.0"),
        reason="resnet is not supported in ORT 1.16.0 caused by https://github.com/microsoft/onnxruntime/issues/17627",
    )
    def test_resnet(search_algorithm, execution_order, system, olive_json):
        from olive.workflows import run as olive_run
    
        olive_config = patch_config(olive_json, search_algorithm, execution_order, system)
    
        footprint = olive_run(olive_config, tempdir=os.environ.get("OLIVE_TEMPDIR", None))
>       check_output(footprint)

/olive/examples/test/local/test_resnet_ptq_cpu.py:49: 
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
/olive/examples/test/utils.py:20: in check_output
    assert_nodes(footprints)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

footprints = {}

    def assert_nodes(footprints):
>       assert footprints, "footprints is empty. The search must have failed for all accelerator specs."
E       AssertionError: footprints is empty. The search must have failed for all accelerator specs.

/olive/examples/test/utils.py:25: AssertionError

Check failure on line 1 in test_get_model_io_config[False]

See this annotation in the file changed.

@azure-pipelines azure-pipelines / Olive CI

test_get_model_io_config[False] 

AssertionError: assert {'dynamic_axe...output_names'} == {'dynamic_axe...output_names'}
  
  Extra items in the left set:
  'dynamic_shapes'
  
  Full diff:
    {
        'dynamic_axes',...
  
  ...Full output truncated (4 lines hidden), use '-vv' to show
Raw output 
with_past = False

    @pytest.mark.parametrize("with_past", [True, False])
    def test_get_model_io_config(with_past):
        model_name, task = get_model_name_task(with_past)
        model = load_model_from_task(task, model_name)
        io_config = get_model_io_config(model_name, task, model)
        expected_keys = ["input_names", "output_names", "dynamic_axes"]
>       assert set(io_config.keys()) == set(expected_keys)
E       AssertionError: assert {'dynamic_axe...output_names'} == {'dynamic_axe...output_names'}
E         
E         Extra items in the left set:
E         'dynamic_shapes'
E         
E         Full diff:
E           {
E               'dynamic_axes',...
E         
E         ...Full output truncated (4 lines hidden), use '-vv' to show

test/unit_test/common/test_hf.py:102: AssertionError

Check failure on line 1 in test_get_model_io_config[True]

See this annotation in the file changed.

@azure-pipelines azure-pipelines / Olive CI

test_get_model_io_config[True] 

AssertionError: assert {'dynamic_axe...output_names'} == {'dynamic_axe...output_names'}
  
  Extra items in the left set:
  'dynamic_shapes'
  
  Full diff:
    {
        'dynamic_axes',...
  
  ...Full output truncated (4 lines hidden), use '-vv' to show
Raw output 
with_past = True

    @pytest.mark.parametrize("with_past", [True, False])
    def test_get_model_io_config(with_past):
        model_name, task = get_model_name_task(with_past)
        model = load_model_from_task(task, model_name)
        io_config = get_model_io_config(model_name, task, model)
        expected_keys = ["input_names", "output_names", "dynamic_axes"]
>       assert set(io_config.keys()) == set(expected_keys)
E       AssertionError: assert {'dynamic_axe...output_names'} == {'dynamic_axe...output_names'}
E         
E         Extra items in the left set:
E         'dynamic_shapes'
E         
E         Full diff:
E           {
E               'dynamic_axes',...
E         
E         ...Full output truncated (4 lines hidden), use '-vv' to show

test/unit_test/common/test_hf.py:102: AssertionError
View more details on Azure Pipelines