// Identity function to make the exporter happy
func.func @main(%arg0: tensor<4xi8>) -> tensor<4xi8> {
  func.return %arg0 : tensor<4xi8>
}

//CHECK:func.func @main(%arg0: tensor<4xi8>) -> tensor<4xi8> attributes {tf.entry_function = {inputs = "arg0", outputs = "arg0"}} {
//CHECK: return %arg0 : tensor<4xi8>
//CHECK:}

func.func @custom_tf_op(%arg0: tensor<1x320x1x1xf32>, %arg1: tensor<2xi32>) -> tensor<1x1600x1x1xf32> {

  func.func private @composite_conv2d_with_bias_and_relu6_fn_10(%arg0: tensor<1x3x2x3xf32>, %arg1: tensor<2x3x3x2xf32>, %arg2: tensor<2xf32>) -> tensor<1x2x2x2xf32> attributes {tf.tf_quant.composite_function} {
    %0 = "quantfork.stats"(%arg1) {layerStats = dense<[-3.54062747, 0.54742622]> : tensor<2xf32>} : (tensor<2x3x3x2xf32>) -> tensor<2x3x3x2xf32>

  func.func @value_used_in_multiple_same_scale_ops(%arg0: tensor<*xf32>, %arg1: tensor<*xf32>, %arg2: tensor<*xf32>, %arg3: tensor<*xf32>) -> (tensor<*xf32>, tensor<*xf32>) {
    %cst = "tf.Const"() {device = "", value = dense<2> : tensor<i32>} : () -> tensor<i32>
    %0 = "quantfork.stats"(%arg0) {layerStats = dense<[-0.83811146, 62.4960899]> : tensor<2xf32>} : (tensor<*xf32>) -> tensor<*xf32>

	fakePC = fakePC % retSledSize
	libfuzzerCallTraceIntCmp(&__sanitizer_cov_trace_cmp4, uintptr(arg0), uintptr(arg1), uintptr(fakePC))
}

//go:nosplit
func libfuzzerTraceCmp8(arg0, arg1 uint64, fakePC uint) {
	fakePC = fakePC % retSledSize
	libfuzzerCallTraceIntCmp(&__sanitizer_cov_trace_cmp8, uintptr(arg0), uintptr(arg1), uintptr(fakePC))
}

//go:nosplit

func.func @move_across_multi_operand_op(%arg0: tensor<1x4x4x8xf32>, %arg1: tensor<1x4x4x8xf32>) -> tensor<1x8x4x4xf32> {

  // CHECK: %[[ARG_PERM:.*]] = "tf.Const"() <{value = dense<[0, 3, 1, 2]> : tensor<4xi32>}>
  // CHECK: %[[ARG0_TRANSPOSE:[0-9]*]] = "tf.Transpose"(%arg0, %[[ARG_PERM]])
  // CHECK: %[[ARG1_TRANSPOSE:[0-9]*]] = "tf.Transpose"(%arg1, %[[ARG_PERM]])

func.func @maximum(%arg0: tensor<4xf32>, %arg1: tensor<4xf32>) -> tensor<4xf32> {
  // CHECK-NEXT:  mhlo.maximum %arg0, %arg1 : tensor<4xf32>
  %0 = "tf.Maximum"(%arg0, %arg1) : (tensor<4xf32>, tensor<4xf32>) -> tensor<4xf32>
  func.return %0 : tensor<4xf32>
}

// CHECK-LABEL: func @minimum
func.func @minimum(%arg0: tensor<4xf32>, %arg1: tensor<4xf32>) -> tensor<4xf32> {
  // CHECK-NEXT:  mhlo.minimum %arg0, %arg1 : tensor<4xf32>

    func.return %0, %1 : tensor<*xf32>, tensor<*xf32>
  }
  func.func private @composite_conv2d_with_bias_and_relu6_fn_2(%arg0: tensor<1x2x2x3xf32>, %arg1: tensor<2x2x3x2xf32>, %arg2: tensor<2xf32>) -> tensor<*xf32> attributes {tf_quant.composite_function} {

  // CHECK: func.func @f(%arg0: tensor<?xf32>
  // CHECK: ml_program.global_store [[V]] = %arg0 : tensor<?xf32>
  // CHECK-NEXT: return
  "tf_saved_model.global_tensor"() { is_mutable, sym_name = "v", type = tensor<?xf32>, value = dense<1.> : tensor<1xf32> } : () -> ()
  func.func @f(%arg0: tensor<?xf32> {tf_saved_model.index_path = [0]}, %arg1: tensor<!tf_type.resource<tensor<?xf32>>> {tf_saved_model.bound_input = @v})

// RUN: flatbuffer_translate -mlir-to-tflite-flatbuffer %s -o - | flatbuffer_translate --tflite-flatbuffer-to-mlir - -o - | FileCheck %s

func.func @main(%arg0: tensor<32x4x4x128xf32>, %arg1: tensor<1x32x42x128xf32>, %arg2: tensor<4xi32>) -> tensor<1x64x84x32xf32> {

// CHECK-LABEL: doubleTposeInput
func.func @doubleTposeInput(%arg0: tensor<2x3x4x5xf32>, %arg1: tensor<2x3x4x5xf32>) -> tensor<5x2x3x4xf32> {
  %perm = arith.constant dense<[3, 0, 1, 2]> : tensor<4xi32>
  %0 = "tfl.transpose"(%arg0, %perm) : (tensor<2x3x4x5xf32>, tensor<4xi32>) -> tensor<5x2x3x4xf32>
  %perm1 = arith.constant dense<[3, 0, 1, 2]> : tensor<4xi32>
  %1 = "tfl.transpose"(%arg1, %perm1) : (tensor<2x3x4x5xf32>, tensor<4xi32>) -> tensor<5x2x3x4xf32>