%8 = "tfl.concatenation"(%2, %0) {axis = -1 : i32, fused_activation_function = "NONE"} : (tensor<1x1xf32>, tensor<1x1xf32>) -> tensor<1x2xf32>
  %9 = "quantfork.stats"(%8) {layerStats = dense<[-0.488159984, 0.189515018]> : tensor<2xf32>} : (tensor<1x2xf32>) -> tensor<1x2xf32>
  %10 = "tfl.concatenation"(%9, %7) {axis = -1 : i32, fused_activation_function = "NONE"} : (tensor<1x2xf32>, tensor<1x2xf32>) -> tensor<1x4xf32>

}

func.func @FusingaddRelu(%arg0: tensor<1xf32>, %arg1: tensor<1xf32>) -> tensor<1xf32> {
  %0 = "tf.Add"(%arg0, %arg1) : (tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>
  %1 = "tf.Add"(%arg0, %0) : (tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>
  %2 = "tf.Relu"(%1) : (tensor<1xf32>) -> tensor<1xf32>
  %3 = "tf.Relu"(%arg0) : (tensor<1xf32>) -> tensor<1xf32>
  %4 = "tf.Add"(%3, %2) : (tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>

    %12 = stablehlo.subtract %10, %11 : tensor<1x3xf32>  // q1 * q2 - z1 * q2
    %13 = stablehlo.multiply %12, %4 : tensor<1x3xf32>  // s1 * s2
    %14 = call @uniform_quantize_1(%13, %5, %6) : (tensor<1x3xf32>, tensor<1x1xf32>, tensor<1x1xi8>) -> tensor<1x3xi8>
    %15 = call @uniform_dequantize_0(%14, %5, %6) : (tensor<1x3xi8>, tensor<1x1xf32>, tensor<1x1xi8>) -> tensor<1x3xf32>
    return %15 : tensor<1x3xf32>
  }
// CHECK-NOT: stablehlo.uniform_quantize

// 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>
  }
}

// -----

// CHECK:           }) : (tensor<1x?xf32>, tensor<1x0xf32>, tensor<1x0xf32>, tensor<1x0xf32>, tensor<1x0xf32>, tensor<1x3xf32>, tensor<1x3xf32>, tensor<1x3xf32>, tensor<1x3xf32>, none, none, none, tensor<1xf32>, tensor<1xf32>, tensor<1xf32>, tensor<1xf32>, tensor<3x1xf32>, tensor<3xf32>, tensor<1x3xf32>, tensor<1x1xf32>, tensor<1xf32>, tensor<1xf32>, tensor<1xf32>, tensor<1xf32>) -> tensor<1x3xf32>

  %5 = "quantfork.stats"(%4) {layerStats = dense<[-56.2916565, 122.922478]> : tensor<2xf32>} : (tensor<1x4xf32>) -> tensor<1x4xf32>
  %6 = "tfl.svdf"(%0, %1, %2, %3, %5) {fused_activation_function = "RELU", rank = 1 : i32} : (tensor<1x3xf32>, tensor<2x3xf32>, tensor<2x1xf32>, tensor<2xf32>, tensor<1x4xf32>) -> tensor<1x2xf32>

  func.return %6#2 : tensor<28x1x16xf32>
}

  %0 = mhlo.custom_call @tf.ResizeBilinear(%arg0) {
    backend_config = "this_is_a_test_string"
  } : (tensor<1x4xf32>) -> (tensor<1x8xf32>)
  func.return %0 : tensor<1x8xf32>

  //       CHECK: %0 = mhlo.custom_call @tf.ResizeBilinear(%arg0) {
  //  CHECK-SAME:   backend_config = "this_is_a_test_string"
  //  CHECK-SAME: } : (tensor<1x4xf32>) -> tensor<1x8xf32>

      %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} {

    %9 = stablehlo.select %arg2, %6, %8 : (tensor<1x3xi1>, tensor<1x3xf32>, tensor<1x3xf32>) -> tensor<1x3xf32>
    %10 = "quantfork.qcast"(%9) {volatile} : (tensor<1x3xf32>) -> tensor<1x3x!quant.uniform<i8:f32, 0.13170163023705575:-1>>
    %11 = "quantfork.dcast"(%10) : (tensor<1x3x!quant.uniform<i8:f32, 0.13170163023705575:-1>>) -> tensor<1x3xf32>
    return %11 : tensor<1x3xf32>
  }