select_scatter decomp

Author: apbose

py/torch_tensorrt/dynamo/lowering/_decompositions.py

@@ -162,6 +162,21 @@ def var_decomposition(
     return variance
 
 
+@register_torch_trt_decomposition(
+    torch.ops.select_scatter, registry=TORCH_TRT_DECOMPOSITIONS
+)
+def select_scatter_decomposition(
+    input_tensor: torch.Tensor,
+    src_tensor: torch.Tensor,
+    dim: int,
+    index: int,
+) -> torch.Tensor:
+    input_tensor.shape[dim] = torch.le(index, input_tensor.shape[dim])
+    src_tensor = torch.expand(torch.unsqueeze(src_tensor, dim), input_tensor.shape)
+    input_tensor_shape = input_tensor.shape
+    return torch.where(torch.eq((input_tensor_shape[dim]), index)), src_tensor, input_tensor)
+
+
 def get_decompositions(
     enable_experimental_decompositions: bool = False,
 ) -> Dict[OpOverload, Callable[[Any], Any]]:

tests/py/dynamo/lowering/test_decompositions.py

@@ -420,6 +420,70 @@ def forward(self, x):
             f"MaxPool3d TRT outputs don't match with the original model.",
         )
 
+    def test_lowering_select_scatter_module(self):
+        class selectScatter(torch.nn.Module):
+            def __init__(self, *args, **kwargs) -> None:
+                super().__init__(*args, **kwargs)
+
+            def forward(self, x, src, dim, index):
+                y = self.select_scatter(x, src, dim, index)
+                return y
+
+        # Operations expected to be removed in the traced graph after decompositions
+        expected_ops = {
+            torch.ops.aten.lt.default,
+            torch.ops.aten.expand.default,
+            torch.ops.aten.unsqueeze.default,
+            torch.ops.aten.where.default,
+        }
+        unexpected_ops = {torch.ops.aten.select_scatter}
+
+        inputs = [torch.randn(2, 2), torch.ones(2)]
+
+        fx_graph = torch.fx.symbolic_trace(selectScatter())
+        unexpected_ops_seen, expected_ops_unseen = lower_graph_testing(
+            fx_graph,
+            inputs,
+            expected_ops=expected_ops,
+            unexpected_ops=unexpected_ops,
+            min_block_size=1,
+        )
+
+        self.assertEquals(
+            len(unexpected_ops_seen),
+            0,
+            f"The following unexpected ops were encountered: {unexpected_ops_seen}",
+        )
+
+        self.assertEquals(
+            len(expected_ops_unseen),
+            0,
+            f"The following expected ops were not encountered: {expected_ops_unseen}",
+        )
+
+        torch._dynamo.reset()
+
+        # Validate that the results between Torch and Torch-TRT are similar
+        optimized_model = torch_tensorrt.compile(
+            fx_graph,
+            "torch_compile",
+            inputs,
+            min_block_size=1,
+            pass_through_build_failures=True,
+        )
+        optimized_model_results = optimized_model(*inputs).detach().cpu()
+        torch_model_results = fx_graph(*inputs).detach().cpu()
+
+        max_diff = float(
+            torch.max(torch.abs(optimized_model_results - torch_model_results))
+        )
+        self.assertAlmostEqual(
+            max_diff,
+            0,
+            DECIMALS_OF_AGREEMENT,
+            f"Select_scatter TRT outputs don't match with the original model.",
+        )
+
 
 if __name__ == "__main__":
     run_tests()