diff --git a/lib/Conversion/TorchToLinalg/TensorConstructors.cpp b/lib/Conversion/TorchToLinalg/TensorConstructors.cpp
index 02853b14072a..4f57263077c3 100644
--- a/lib/Conversion/TorchToLinalg/TensorConstructors.cpp
+++ b/lib/Conversion/TorchToLinalg/TensorConstructors.cpp
@@ -497,8 +497,7 @@ class ConvertAtenEmptyMemoryFormatOp
         cast<RankedTensorType>(typeConverter->convertType(op.getType()));
     Type resultElementType;
     if (isa<Torch::NoneType>(op.getDtype().getType())) {
-      resultElementType = getDefaultDtypeForTorchScalar(
-          Torch::FloatType::get(op->getContext()));
+      resultElementType = resultType.getElementType();
     } else {
       int64_t dtypeInt;
       if (!matchPattern(op.getDtype(), m_TorchConstantInt(&dtypeInt)))
diff --git a/test/Conversion/TorchToLinalg/basic.mlir b/test/Conversion/TorchToLinalg/basic.mlir
index 1b61f75703f6..cef844658258 100644
--- a/test/Conversion/TorchToLinalg/basic.mlir
+++ b/test/Conversion/TorchToLinalg/basic.mlir
@@ -354,3 +354,20 @@ func.func @torch.aten.transpose$basic(%arg0: !torch.vtensor<[4,3],f32>) -> !torc
   %0 = torch.aten.transpose.int %arg0, %int0, %int1 : !torch.vtensor<[4,3],f32>, !torch.int, !torch.int -> !torch.vtensor<[3,4],f32>
   return %0 : !torch.vtensor<[3,4],f32>
 }
+
+
+// -----
+// CHECK-LABEL:   func.func @torch.aten.empty.memory_format$noneDtype()
+// CHECK:           %[[EMPTY:.*]] = tensor.empty() : tensor<200x200x26xf64>
+// CHECK:           %[[CAST:.*]] = tensor.cast %[[EMPTY]] : tensor<200x200x26xf64> to tensor<200x200x26xf64>
+// CHECK:           %[[RES:.*]] = torch_c.from_builtin_tensor %[[CAST]] : tensor<200x200x26xf64> -> !torch.vtensor<[200,200,26],f64>
+// CHECK:           return %[[RES]] : !torch.vtensor<[200,200,26],f64>
+func.func @torch.aten.empty.memory_format$noneDtype() -> !torch.vtensor<[200,200,26],f64> attributes {torch.assume_strict_symbolic_shapes} {
+    %int200 = torch.constant.int 200
+    %int26 = torch.constant.int 26
+    %false = torch.constant.bool false
+    %none = torch.constant.none
+    %0 = torch.prim.ListConstruct %int200, %int200, %int26 : (!torch.int, !torch.int, !torch.int) -> !torch.list<int>
+    %1 = torch.aten.empty.memory_format %0, %none, %none, %none, %false, %none : !torch.list<int>, !torch.none, !torch.none, !torch.none, !torch.bool, !torch.none -> !torch.vtensor<[200,200,26],f64>
+    return %1 : !torch.vtensor<[200,200,26],f64>
+}