%1 = "tfl.reshape"(%0, %cst) : (tensor<200xf32>, tensor<2xi64>) -> tensor<2x100xf32>
    func.return %1 : tensor<2x100xf32>
  }
  func.func @func_0_CPU_FLOAT(%arg0: tensor<100xf32>, %arg1: tensor<100xf32>, %arg2: tensor<100xf32>) -> tensor<100xf32> attributes {tac.cost = 2.000000e+02 : f32, tac.device = "CPU", tac.inference_type = "FLOAT", tac.interface_name = "func_0"} {

  %arg7: tensor<2048x640xf32>,
  %arg8: tensor<2048x640xf32>,
  %arg9: tensor<2048xf32>,
  %arg10: tensor<2048xf32>,
  %arg11: tensor<2048xf32>,
  %arg12: tensor<2048xf32>,
  %arg13: tensor<640x2048xf32>,
  %arg14: tensor<640xf32>,
  %arg15: tensor<2048xf32>,
  %arg16: tensor<2048xf32>,
  %arg17: tensor<2048xf32>,
  %arg18: tensor<2048xf32>,
  %arg19: tensor<1x640xf32>,
  %arg20: tensor<1x2048xf32>

  %3 = "tfl.dequantize"(%2) : (tensor<2x3x!quant.uniform<i16:f32, 1.0>>) -> (tensor<2x3xf32>)
  %4 = "tfl.concatenation"(%1, %3) {axis = -1 : i32, fused_activation_function = "NONE"} : (tensor<2x1xf32>, tensor<2x3xf32>) -> tensor<2x4xf32>
  %5 = "tfl.add"(%4, %arg2) {fused_activation_function = "NONE"} : (tensor<2x4xf32>, tensor<2x4xf32>) -> tensor<2x4xf32>
  func.return %5: tensor<2x4xf32>

// CHECK-LABEL: hardcode_all
func.func @hardcode_all(%arg0: tensor<2x2xf32>, %arg1: tensor<2x1xf32>) -> tensor<2x2xf32> {
  %0 = "tfl.add"(%arg0, %arg1) {fused_activation_function="NONE"}: (tensor<2x2xf32>, tensor<2x1xf32>) -> tensor<2x2xf32>
  func.return %0 : tensor<2x2xf32>

// CHECK: %[[q0:.*]] = "tfl.quantize"(%arg1) <{qtype = tensor<2x1x!quant.uniform<u8:f32, 0.0078431372549019607:128>>}>

        tensor<1x1x5xf32>,
        tensor<2x5xf32>, tensor<2x5xf32>, tensor<2x5xf32>, tensor<2x5xf32>,
        tensor<2x4xf32>, tensor<2x4xf32>, tensor<2x4xf32>, tensor<2x4xf32>,
        tensor<2xf32>, tensor<2xf32>, tensor<2xf32>,
        tensor<2xf32>, tensor<2xf32>, tensor<2xf32>, tensor<2xf32>,
        tensor<4x2xf32>, tensor<4xf32>,
        tensor<1x4xf32>, tensor<1x2xf32>,

  %1 = "tfl.sum"(%0, %cst) {keep_dims = false} : (tensor<2x2xf32>, tensor<1xi32>) -> tensor<f32>
  %2 = "tfl.sqrt"(%1) : (tensor<f32>) -> tensor<f32>
  %3 = "tfl.div"(%arg0, %2) {fused_activation_function = "NONE"} : (tensor<2x2xf32>, tensor<f32>) -> tensor<2x2xf32>
  func.return %3: tensor<2x2xf32>

    : ( tensor<1x28x28xf32>,
        tensor<20x20xf32>, tensor<20x20xf32>, tensor<20x20xf32>, tensor<20x20xf32>,
        tensor<20x20xf32>, tensor<20x20xf32>, tensor<20x20xf32>, tensor<20x20xf32>,
        tensor<20xf32>, tensor<20xf32>, tensor<20xf32>,
        tensor<20xf32>, tensor<20xf32>, tensor<20xf32>, tensor<20xf32>,
        tensor<20x20xf32>, none,
        tensor<1x20xf32>, tensor<1x20xf32>,

  // CHECK-SAME: %[[ARG1:.*]]: tensor<2x4xf32>
  func.func private @composite_and_slice(%arg0: tensor<3x2xf32>, %arg1: tensor<2x4xf32>) -> tensor<2x2xf32> {
    // CHECK: %[[Q1:.*]] = "quantfork.qcast"(%[[ARG0]]) {volatile} : (tensor<3x2xf32>) -> tensor<3x2x!quant.uniform<i8:f32, 5.000000e-03>>

    version = 5 : i64
  } : (tensor<2x2xf32>, tensor<2x3xf32>) -> tensor<2x3xf32>
  %2 = stablehlo.add %cst, %0 : tensor<2x3xf32>
  return %2 : tensor<2x3xf32>
}

// CHECK-LABEL: func.func @no_qdq_for_multiple_users
// CHECK-NOT: quantfork.qcast

// CHECK:           %[[VAL_5:.*]] = "tf.Tanh"(%[[VAL_4]]) : (tensor<2xf32>) -> tensor<2xf32>
// CHECK:           %[[VAL_6:.*]] = "tf.Mul"(%[[VAL_5]], %[[VAL_3]]) : (tensor<2xf32>, tensor<2xf32>) -> tensor<2xf32>
// CHECK:           %[[VAL_7:.*]] = "tf.AddV2"(%[[VAL_6]], %[[VAL_3]]) : (tensor<2xf32>, tensor<2xf32>) -> tensor<2xf32>
// CHECK:           return %[[VAL_7]] : tensor<2xf32>