// CHECK-SAME: %[[ARG3:[^:[:space:]]+]]
  func.func private @main(%arg0: tensor<1x4xf32>, %arg1: tensor<4x3xf32>) -> tensor<1x3xf32> attributes {_from_xla_call_module} {
    %0 = stablehlo.dot_general %arg0, %arg1, contracting_dims = [1] x [0] : (tensor<1x4xf32>, tensor<4x3xf32>) -> tensor<1x3xf32>
    return %0 : tensor<1x3xf32>
    // CHECK: %[[DOT:.+]] = stablehlo.dot_general %[[ARG2]], %[[ARG3]]
    // CHECK: return %[[DOT]]
  }

// CHECK: return

  func.func private @composite_dot_general_fn(%arg0: tensor<1x4xf32>, %arg1: tensor<4x3xf32>) -> tensor<1x3xf32> attributes {_from_xla_call_module} {
      %0 = stablehlo.dot_general %arg0, %arg1, contracting_dims = [1] x [0] : (tensor<1x4xf32>, tensor<4x3xf32>) -> tensor<1x3xf32>
      return %0 : tensor<1x3xf32>
  }
}

// -----

func.func @main(%arg0: tensor<1x4xf32>) -> tensor<1x3xf32> {
  %0 = stablehlo.constant dense<1.0> : tensor<4x3xf32>
  %1 = stablehlo.dot_general %arg0, %0, contracting_dims = [1] x [0], precision = [DEFAULT, DEFAULT] : (tensor<1x4xf32>, tensor<4x3xf32>) -> tensor<1x3xf32>
  return %1 : tensor<1x3xf32>
}
// CHECK: @main(%[[ARG_0:.+]]: tensor<1x4xf32>) -> tensor<1x3xf32>

  %0 = "tf.FakeQuantWithMinMaxVars"(%arg0, %cst, %cst_0) {
    device = "", narrow_range = false, num_bits = 8 : i64
  } : (tensor<3xf32>, tensor<f32>, tensor<f32>) -> tensor<3xf32>
  %1 = "tf.FakeQuantWithMinMaxVarsPerChannel"(%arg1, %cst_1, %cst_2) {
    device = "", narrow_range = true, num_bits = 8 : i64
  } : (tensor<4x3xf32>, tensor<3xf32>, tensor<3xf32>) -> tensor<4x3xf32>
  func.return %0, %1 : tensor<3xf32>, tensor<4x3xf32>
}

func.func @test_batch_matmul_to_einsum(%arg0: tensor<1x2x3xf32>, %arg1: tensor<3x4xf32>) -> tensor<1x2x4xf32> {
  // CHECK-LABEL: test_batch_matmul_to_einsum
  // CHECK: "tf.Einsum"(%arg0, %arg1) <{equation = "...mk,...kn->...mn"}> : (tensor<1x2x3xf32>, tensor<3x4xf32>) -> tensor<1x2x4xf32>
  %0 = "tf.BatchMatMul"(%arg0, %arg1) {adj_x = false, adj_y = false} : (tensor<1x2x3xf32>, tensor<3x4xf32>) -> tensor<1x2x4xf32>
  func.return %0: tensor<1x2x4xf32>
}

    module attributes {} {
      func.func @main(%arg0: tensor<1x4xf32>) -> tensor<1x3xf32> attributes {} {
        %0 = stablehlo.constant dense<1.0> : tensor<4x3xf32>
        %1 = stablehlo.dot_general %arg0, %0, contracting_dims = [1] x [0], precision = [DEFAULT, DEFAULT] : (tensor<1x4xf32>, tensor<4x3xf32>) -> tensor<1x3xf32>
        return %1 : tensor<1x3xf32>
      }
    }
  )mlir");
  ASSERT_TRUE(module_op);

      %2 = "quantfork.stats"(%1) {layerStats = dense<[5.00000000e-6, 7.00000000e-1]> : tensor<2xf32>} : (tensor<1x3xf32>) -> tensor<1x3xf32>
      return %2 : tensor<1x3xf32>
    }

    func.func private @composite_dot_general_fn(%arg0: tensor<1x2xf32>, %arg1: tensor<2x3xf32>) -> tensor<1x3xf32> attributes {_from_xla_call_module} {

  // CHECK: device = "cpu"
  %2 = "tf.Relu"(%1) {T = f32, _output_shapes = ["tfshape$dim { size: 3 } dim { size: 3 }"], device = "cpu"} : (tensor<3x3xf32>) -> tensor<3x3xf32>
  // CHECK: device = "xpu"
  %3 = "tf.Relu"(%2) {T = f32, _output_shapes = ["tfshape$dim { size: 3 } dim { size: 3 }"]} : (tensor<3x3xf32>) -> tensor<3x3xf32>
  func.return %3 : tensor<3x3xf32>
}

# CHECK-NEXT: [[R0:%.*]] = "tf.ReadVariableOp"([[ARG1]]) {{{.*}}} : (tensor<!tf_type.resource<tensor<1x3xf32>>>) -> tensor<1x3xf32>
# CHECK-NEXT: [[R1:%.*]] = "tf.MatMul"([[ARG0]], [[R0]]) <{{{.*}}}> {device = ""} : (tensor<3x1xf32>, tensor<1x3xf32>) -> tensor<3x3xf32>
# CHECK-NEXT: return [[R1]] : tensor<3x3xf32>


def Test():

  x = tf.constant([[1.0], [1.0], [1.0]])
  y = tf.compat.v1.get_variable(
      name='y',

    return %2 : tensor<1x3xf32>
  }

  func.func private @composite_dot_general_fn(%arg0: tensor<1x2xf32>, %arg1: tensor<2x3xf32>) -> tensor<1x3xf32> attributes {_from_xla_call_module} {
    %0 = stablehlo.dot_general %arg0, %arg1, contracting_dims = [1] x [0] : (tensor<1x2xf32>, tensor<2x3xf32>) -> tensor<1x3xf32>
    return %0 : tensor<1x3xf32>
  }
}