Add regression tests for previously-failing CoreML lowerings

backends/apple/coreml/test/test_coreml_partitioner.py

@@ -338,6 +338,123 @@ def forward(self, x):
                 torch.allclose(et_outputs, eager_outputs, atol=1e-02, rtol=1e-02)
             )
 
+    def test_previously_failing_ops_lower_successfully(self):
+        """
+        Each of these snippets used to crash the CoreML partitioner / lowering
+        pipeline.  They are kept here as regression tests so any future change
+        that re-breaks one of them surfaces in CI.
+
+        - Conv1d with stride>kernel and groups>1 (#11688)
+        - int32 matmul with a constant weight (#11691)
+        - BatchNorm3d / InstanceNorm3d on a rank-5 input (#11701, #11702)
+        - ReflectionPad3d / ReplicationPad3d (#11708, #11709)
+        """
+
+        cases = []
+
+        class Conv1dCase(torch.nn.Module):
+            def __init__(self):
+                super().__init__()
+                self.conv = torch.nn.Conv1d(
+                    16, 4, 6, stride=8, padding=0, dilation=2, groups=2, bias=False
+                )
+
+            def forward(self, x):
+                return self.conv(x)
+
+        cases.append(("issue_11688_conv1d", Conv1dCase().eval(), (torch.randn(2, 16, 11),)))
+
+        class Int32MmCase(torch.nn.Module):
+            def __init__(self):
+                super().__init__()
+                self.weight = torch.randint(0, 100, (8, 8)).to(torch.int32)
+
+            def forward(self, x):
+                return torch.mm(x, self.weight)
+
+        cases.append(
+            (
+                "issue_11691_int32_mm",
+                Int32MmCase().eval(),
+                (torch.randn(8, 8).to(torch.int32),),
+            )
+        )
+
+        class BatchNorm3dCase(torch.nn.Module):
+            def __init__(self):
+                super().__init__()
+                self.norm = torch.nn.BatchNorm3d(3)
+
+            def forward(self, x):
+                return self.norm(x)
+
+        cases.append(
+            (
+                "issue_11701_batchnorm3d",
+                BatchNorm3dCase().eval(),
+                (torch.randn(1, 3, 4, 4, 4),),
+            )
+        )
+
+        class InstanceNorm3dCase(torch.nn.Module):
+            def __init__(self):
+                super().__init__()
+                self.norm = torch.nn.InstanceNorm3d(3)
+
+            def forward(self, x):
+                return self.norm(x)
+
+        cases.append(
+            (
+                "issue_11702_instancenorm3d",
+                InstanceNorm3dCase().eval(),
+                (torch.randn(1, 3, 4, 4, 4),),
+            )
+        )
+
+        class ReflectionPad3dCase(torch.nn.Module):
+            def __init__(self):
+                super().__init__()
+                self.pad = torch.nn.ReflectionPad3d(2)
+
+            def forward(self, x):
+                return self.pad(x)
+
+        cases.append(
+            (
+                "issue_11708_reflection_pad3d",
+                ReflectionPad3dCase().eval(),
+                (torch.randn(1, 6, 6, 6, 6),),
+            )
+        )
+
+        class ReplicationPad3dCase(torch.nn.Module):
+            def __init__(self):
+                super().__init__()
+                self.pad = torch.nn.ReplicationPad3d(2)
+
+            def forward(self, x):
+                return self.pad(x)
+
+        cases.append(
+            (
+                "issue_11709_replication_pad3d",
+                ReplicationPad3dCase().eval(),
+                (torch.randn(1, 6, 6, 6, 6),),
+            )
+        )
+
+        for name, model, example_inputs in cases:
+            with self.subTest(name=name):
+                ep = torch.export.export(model, example_inputs, strict=True)
+                executorch.exir.to_edge_transform_and_lower(
+                    ep,
+                    partitioner=[CoreMLPartitioner()],
+                    compile_config=executorch.exir.EdgeCompileConfig(
+                        _check_ir_validity=False
+                    ),
+                )
+
     def test_deprecation_warning_for_to_backend_workflow(self):
         """
         Test that the deprecated to_edge + to_backend workflow shows a deprecation warning.