// CHECK-NOT: tf.BatchFunction
// CHECK: %[[ADD0:.*]] = "tf.AddV2"(%[[ARG0]], %[[ARG1]])
// CHECK: return %[[ADD0]] : tensor<1xf32>

func.func private @batched_func(%arg0: tensor<1xf32>, %arg1: tensor<1xf32>) -> tensor<1xf32> {
  %0 = "tf.Identity"(%arg0) : (tensor<1xf32>) -> tensor<1xf32>
  %1 = "tf.Identity"(%arg1) : (tensor<1xf32>) -> tensor<1xf32>

  func.func @testFunctionSkiped(%arg0: tensor<1xf32>, %arg1: tensor<1xf32>) {
    // CHECK: tfl.add
    // CHECK-SAME: tac.skip_target_annotation
    %0 = "tfl.add"(%arg0, %arg1) {fused_activation_function = "RELU6"} : (tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>
    // CHECK: tfl.add
    // CHECK-SAME: tac.skip_target_annotation
    %1 = "tfl.add"(%arg0, %0) {fused_activation_function = "RELU"} : (tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>

func.func @sharding_custom_call_removed(%arg0: tensor<3xf32>) -> tensor<3xf32> {
  %1 = stablehlo.custom_call @Sharding(%arg0) {mhlo.sharding = ""} : (tensor<3xf32>) -> tensor<3xf32>
  return %1 : tensor<3xf32>
}
// CHECK-NOT: custom_call

// -----

// Tests that a custom_call that is not @Sharding is not removed.

// CHECK-LABEL: custom_call_not_removed
func.func @custom_call_not_removed(%arg0: tensor<3xf32>) -> tensor<3xf32> {

  // CHECK-DAG: %[[CST_0:.+]] = mhlo.constant dense<[1.000000e-01, 2.000000e-01]> : tensor<2xf32>
  // CHECK-DAG: %[[CST_1:.+]] = mhlo.constant dense<[3.000000e-01, 4.000000e-01]> : tensor<2xf32>
  // CHECK: %[[CST_BCAST:.+]] = "mhlo.broadcast_in_dim"(%[[CST_0]]) <{broadcast_dimensions = dense<3> : tensor<1xi64>}> : (tensor<2xf32>) -> tensor<1x1x3x2xf32>

//RUN: tf_tfl_translate --enable-stablehlo-conversion --input-mlir %s -o /tmp/temp.stablehlo; [ -f /tmp/temp.stablehlo ]


module {
func.func @main(%arg0: tensor<2xf32>) -> tensor<2xf32> {
  %0 = "tf.Sigmoid"(%arg0) : (tensor<2xf32>) -> tensor<2xf32>
  func.return %0 : tensor<2xf32>
}

    %4 = "quantfork.stats"(%2) {layerStats = dense<[-1.36523, 3.57373]> : tensor<2xf32>} : (tensor<1x2x2x2xf32>) -> tensor<1x2x2x2xf32>
    %5 = "chlo.broadcast_add"(%4, %3) : (tensor<1x2x2x2xf32>, tensor<2xf32>) -> tensor<1x2x2x2xf32>
    %6 = "quantfork.stats"(%5) {layerStats = dense<[-1.31055, 2.62842]> : tensor<2xf32>} : (tensor<1x2x2x2xf32>) -> tensor<1x2x2x2xf32>

module {
func.func @main(%arg0 : tensor<5xf32>, %arg1 : tensor<f32>, %arg2 : tensor<f32>) -> tensor<5xf32> {
  %0 = "tf.ClipByValue"(%arg0, %arg1, %arg2) : (tensor<5xf32>, tensor<f32>, tensor<f32>) -> tensor<5xf32>
  func.return %0 : tensor<5xf32>
}

  %fill = "tf.Fill"(%cst_0, %cst) : (tensor<1xi64>, tensor<f32>) -> tensor<1xf32>
  tf_device.replicate([%0, %fill] as %arg_r0: tensor<1xf32>) {n = 2 : i32} {
    %1 = "tf_device.launch"() <{device = "TPU_REPLICATED_HOST_0"}> ({
      %2 = "tf.Identity"(%arg_r0) : (tensor<1xf32>) -> tensor<1xf32>
      tf_device.return %2 : tensor<1xf32>
    }) : () -> tensor<1xf32>

  %2:2 = "tf.RecvTPUEmbeddingActivations"() {_tpu_embedding_layer = "call1", config = "\0A\0B\0C\0D"} : () -> (tensor<2x2xf32>, tensor<4x4xf32>)
  func.return
}

// -----

  %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>