%cst = stablehlo.constant dense<1.0> : tensor<1x3xf32>
      %0 = stablehlo.dot_general %arg0, %arg1, contracting_dims = [1] x [0], precision = [] : (tensor<1x1024xf32>, tensor<1024x3xf32>) -> tensor<1x3xf32>
      %1 = stablehlo.subtract %cst, %0 : tensor<1x3xf32>
      %2 = stablehlo.add %0, %cst : tensor<1x3xf32>
      return %2 : tensor<1x3xf32>
    }
  }

}

// CHECK-LABEL: func @testLeakyRelu
func.func @testLeakyRelu(%arg0 : tensor<16xf32>) -> (tensor<16xf32>, tensor<f32>, tensor<f32>, tensor<16xf32>) {
  %pos = arith.constant dense<5.0> : tensor<f32>
  %neg = arith.constant dense<-5.0> : tensor<f32>
  %no = "tf.LeakyRelu"(%arg0) {alpha = 0.2 : f32} : (tensor<16xf32>) -> tensor<16xf32>
  %0 = "tf.LeakyRelu"(%pos) {alpha = 0.3 : f32} : (tensor<f32>) -> tensor<f32>

}

// CHECK-LABEL: QuantizeConcat
func.func @QuantizeConcat(tensor<1x2xf32>, tensor<1x2xf32>) -> tensor<2x2x!quant.uniform<u8:f32, 1.000000e-01:128>> {
^bb0(%arg0: tensor<1x2xf32>, %arg1: tensor<1x2xf32>):
  %0 = "tfl.concatenation"(%arg0, %arg1) {axis = 0 : i32, fused_activation_function = "NONE"} : (tensor<1x2xf32>, tensor<1x2xf32>) -> tensor<2x2xf32>

        %3 = "tf.Mul"(%arg0, %0) : (tensor<16xf32>, tensor<f32>) -> tensor<16xf32>
        %4 = "tf.Mul"(%3, %0) : (tensor<16xf32>, tensor<f32>) -> tensor<16xf32>
        %5 = "tf.Mul"(%4, %1) : (tensor<16xf32>, tensor<f32>) -> tensor<16xf32>
        %6 = "tf.Mul"(%5, %2) : (tensor<16xf32>, tensor<f32>) -> tensor<16xf32>
        tf_device.return %6 : tensor<16xf32>
      }) {} : () -> tensor<16xf32>
      tf_executor.yield %3 : tensor<16xf32>
    }

  %2 = "tf.Mul"(%1, %value) {T = "tfdtype$DT_FLOAT"} : (tensor<256x8x7x16xf32>, tensor<16xf32>) -> tensor<256x8x7x16xf32>
  func.return %2 : tensor<256x8x7x16xf32>

// CHECK-DAG: %[[cst:.*]] = "tf.Const{{.*}} dense<8.000000e+00> : tensor<3x3x3x16xf32>
// CHECK-DAG: %[[cst_0:.*]] = "tf.Const{{.*}} dense<1.200000e+01> : tensor<16xf32>

// CHECK: Placeholder
// CHECK: key: "type"
// CHECK: type: DT_INT8

func.func @main(%arg0 : tensor<16xf32>) {
  tf_executor.graph {
    %1:2 = tf_executor.island wraps "tf.MlirPassthroughOp"(%arg0) {extra_type_attr = [tensor<5xi32>, tensor<16xf32>], Tinputs = [tensor<16xf32>], Toutputs = [tensor<16xf32>], mlir_module = ""} : (tensor<16xf32>) -> tensor<16xf32>
    tf_executor.fetch
  }
  func.return
}

func.func @plain() {
  tf_executor.graph {

func.func @addNSkipFoldingIfBroadcasting(%arg0: tensor<1xf32>) -> tensor<10xf32> {
  %0 = "tf.Const"() {value = dense<0.000000e+00> : tensor<10xf32>} : () -> tensor<10xf32>
  // CHECK: [[ZERO:%.*]] = "tf.Const"() <{value = dense<0.000000e+00> : tensor<10xf32>}>
  // CHECK: [[ADD_N:%.*]] = "tf.AddN"(%arg0, [[ZERO]])
  // CHECK: return [[ADD_N]]
  %1 = "tf.AddN"(%arg0, %0) : (tensor<1xf32>, tensor<10xf32>) -> tensor<10xf32>
  func.return %1: tensor<10xf32>
}

  %min = arith.constant dense<0.0> : tensor<16xf32>
  %max = arith.constant dense<255.0> : tensor<16xf32>
  %mini = "tf.Identity"(%min) : (tensor<16xf32>) -> tensor<16xf32>
  %maxi = "tf.Identity"(%max) : (tensor<16xf32>) -> tensor<16xf32>
  %fq = "tf.FakeQuantWithMinMaxVarsPerChannel"(%in, %mini, %maxi) {num_bits = 5, narrow_range = false} : (tensor<3x3x3x16xf32>, tensor<16xf32>, tensor<16xf32>) -> tensor<3x3x3x16xf32>

  %min = arith.constant dense<0.0> : tensor<16xf32>
  %max = arith.constant dense<15.0> : tensor<16xf32>
  %mini = "tf.Identity"(%min) : (tensor<16xf32>) -> tensor<16xf32>
  %maxi = "tf.Identity"(%max) : (tensor<16xf32>) -> tensor<16xf32>
  %fq = "tf.FakeQuantWithMinMaxVarsPerChannel"(%in, %mini, %maxi) {num_bits = 3, narrow_range = false} : (tensor<3x3x3x16xf32>, tensor<16xf32>, tensor<16xf32>) -> tensor<3x3x3x16xf32>

  %2 = func.call @fold_all_test(%arg0, %0, %1) : (tensor<256x32x32x3xf32>, tensor<16x3x3x3xf32>, tensor<16xf32>) -> tensor<256x30x30x16xf32>
  func.return %2 : tensor<256x30x30x16xf32>
}

// ALL-LABEL: @fold_all_test
func.func @fold_all_test(%arg0: tensor<256x32x32x3xf32>, %arg1: tensor<16x3x3x3xf32>, %arg2: tensor<16xf32>) -> tensor<256x30x30x16xf32> {