[onnx] Fix grad op domain

Author: twata

pytorch_pfn_extras/onnx/_grad.py

@@ -85,7 +85,7 @@ def _grad(  # type: ignore
             @staticmethod
             def symbolic(g, output, grad_output, *inputs):  # type: ignore
                 return g.op(
-                    "ai.onnx.preview::Gradient",
+                    "ai.onnx.preview.training::Gradient",
                     *inputs,
                     xs_s=input_names,
                     zs_s=[],

tests/pytorch_pfn_extras_tests/onnx_tests/test_grad.py

@@ -62,7 +62,7 @@ def forward(self, x):
 
 
 @pytest.mark.parametrize("use_pfto", [False, True])
-@pytest.mark.filterwarnings("ignore:The shape inference of ai.onnx.preview..Gradient type is missing:UserWarning")
+@pytest.mark.filterwarnings("ignore:The shape inference of ai.onnx.preview.training..Gradient type is missing:UserWarning")
 @pytest.mark.filterwarnings("ignore:Specified output_names .*:UserWarning")
 @pytest.mark.filterwarnings("ignore::DeprecationWarning")
 def test_grad(use_pfto: bool):
@@ -103,6 +103,7 @@ def forward(self, x):
     )
 
     actual_onnx = onnx.load(os.path.join(output_dir, 'model.onnx'))
+    print(actual_onnx)
     named_nodes = {n.name: n for n in actual_onnx.graph.node}
     if pytorch_pfn_extras.requires("1.13") and not use_pfto:
         assert '/_ppe_as_out_module/conv/Conv' in named_nodes
@@ -136,7 +137,7 @@ def forward(self, x):
 
 
 @pytest.mark.parametrize("use_pfto", [False, True])
-@pytest.mark.filterwarnings("ignore:The shape inference of ai.onnx.preview..Gradient type is missing:UserWarning")
+@pytest.mark.filterwarnings("ignore:The shape inference of ai.onnx.preview.training..Gradient type is missing:UserWarning")
 @pytest.mark.filterwarnings("ignore:Specified output_names .*:UserWarning")
 @pytest.mark.filterwarnings("ignore::DeprecationWarning")
 def test_grad_multiple_times(use_pfto: bool):
@@ -218,7 +219,7 @@ def forward(self, x):
 
 
 @pytest.mark.parametrize("use_pfto", [False, True])
-@pytest.mark.filterwarnings("ignore:The shape inference of ai.onnx.preview..Gradient type is missing:UserWarning")
+@pytest.mark.filterwarnings("ignore:The shape inference of ai.onnx.preview.training..Gradient type is missing:UserWarning")
 @pytest.mark.filterwarnings("ignore:Specified output_names .*:UserWarning")
 @pytest.mark.filterwarnings("ignore::DeprecationWarning")
 def test_grad_with_multiple_inputs(use_pfto: bool):