// CHECK-SAME:        rhs_contracting_dimensions = [1]
  %0 = "tf.BatchMatMulV2"(%arg0, %arg1) {adj_x = true, adj_y = true, device = ""} : (tensor<2x5xf32>, tensor<4x2xf32>) -> tensor<5x4xf32>
  func.return %0 : tensor<5x4xf32>
}

func.func @batchmatmulv2_adj_complex(%arg0: tensor<2x5xcomplex<f32>>, %arg1: tensor<4x2xcomplex<f32>>) -> tensor<5x4xcomplex<f32>> {
// CHECK-LABEL:   func @batchmatmulv2_adj_complex(

  func.return %2 : tensor<3x3xf32>

func.func @squaredDifference(%arg0: tensor<4xf32>, %arg1: tensor<4xf32>) -> tensor<4xf32> {
  %0 = "tfl.squared_difference"(%arg0, %arg1) : (tensor<4xf32>, tensor<4xf32>) -> tensor<4xf32>
  func.return %0 : tensor<4xf32>
}

// CHECK:       func @squaredDifference(%arg0: tensor<4xf32>, %arg1: tensor<4xf32>) -> tensor<4xf32> {
// CHECK:         %0 = "tf.Sub"(%arg0, %arg1) : (tensor<4xf32>, tensor<4xf32>) -> tensor<4xf32>

  %2 = "tfl.add"(%arg0, %arg3) {fused_activation_function = "RELU6", per_device_costs = {CPU = 5.0 : f32, GPU = 1.0 : f32}, tac.device = "GPU"} : (tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>
  %3 = "tfl.pack"(%1, %2) {axis = 0 : i32, per_device_costs = {CPU = 2.0 : f32, GPU = -1.0 : f32}, values_count = 2 : i32, tac.device = "CPU"} : (tensor<1xf32>, tensor<1xf32>) -> tensor<2x1xf32>

  module {
    func.func @constant_add() -> (tensor<3x2xf32>) {
      %cst1 = stablehlo.constant dense<2.4> : tensor<3x2xf32>
      %cst2 = stablehlo.constant dense<5.7> : tensor<3x2xf32>
      %add = stablehlo.add %cst1, %cst2 : (tensor<3x2xf32>, tensor<3x2xf32>) -> tensor<3x2xf32>
      func.return %add : tensor<3x2xf32>
    }
  }
)mlir";

module {
  func.func @simpleTest(%arg0: tensor<1xf32>, %arg1: tensor<1xf32>, %arg2: tensor<1xf32>, %arg3: tensor<1xf32>) -> tensor<2x1xf32> {
    %0 = func.call @func_0_GPU_FLOAT(%arg0, %arg1, %arg2) {tac.interface_name = "func_0"} : (tensor<1xf32>, tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>
    %1 = func.call @func_1_GPU_FLOAT(%arg0, %arg3) {tac.interface_name = "func_1"} : (tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>

func.func private @gelu_decomp_2(%arg0: tensor<5x10xf32>) -> tensor<5x10xf32>
func.func @gelu_aten_approximate(%arg0: tensor<5x10xf32>) -> (tensor<*xf32>) {
  %0 = mhlo.composite "aten.gelu.default" %arg0 {composite_attributes = {approximate = "tanh"}, decomposition = @gelu_decomp_2} : (tensor<5x10xf32>) -> tensor<5x10xf32>

// CHECK-NEXT:        buffer: 4
// CHECK-NEXT:      } ],
// CHECK-NEXT:      signature_defs: [  ]
// CHECK-NEXT:    }

  %0 = "tfl.cast" (%arg0) : (tensor<4x5xbf16>) -> tensor<4x5xf32> loc("cast1")
  %1 = "tfl.cast" (%0) : (tensor<4x5xf32>) -> tensor<4x5xbf16> loc("cast2")
  func.return %1 : tensor<4x5xbf16>

}

// CHECK-LABEL: prepareAdd
func.func @prepareAdd(%arg0: tensor<2x2xf32>) -> tensor<2x2xf32> {
  %cst = arith.constant dense<[[0.0, 1.0], [2.0, 255.0]]> : tensor<2x2xf32>
  %add = "tfl.add"(%arg0, %cst) {fused_activation_function="NONE"} : (tensor<2x2xf32>, tensor<2x2xf32>) -> tensor<2x2xf32>
  func.return %add : tensor<2x2xf32>

  %outputs =  "tf.Const"() {device = "/job:localhost/replica:0/task:0/device:CPU:0", value = dense<0> : tensor<i32>} : () -> tensor<i32>...