// CHECK-NEXT:  %[[quant:.*]] = "tfl.quantize"(%[[split]]#0) <{qtype = tensor<2x!quant.uniform<u8:f32, 1.000000e+00>>}> {volatile} : (tensor<2xf32>) -> tensor<2x!quant.uniform<u8:f32, 1.000000e+00>>
// CHECK-NEXT:  return %[[quant]] : tensor<2x!quant.uniform<u8:f32, 1.000000e+00>>

  %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>

  %1 = stablehlo.constant dense<2.000000e+00> : tensor<4xf32>
  %2 = stablehlo.convolution(%arg0, %0) dim_numbers = [b, 0, 1, f]x[0, 1, i, o]->[b, 0, 1, f], window = {pad = [[1, 1], [1, 1]]} {batch_group_count = 1 : i64, feature_group_count = 1 : i64} : (tensor<1x3x3x4xf32>, tensor<3x3x4x4xf32>) -> tensor<1x3x3x4xf32>
  %3 = stablehlo.broadcast_in_dim %1, dims = [3] : (tensor<4xf32>) -> tensor<1x3x3x4xf32>
  %4 = stablehlo.add %2, %3 : tensor<1x3x3x4xf32>

        tensor<2xf32>, tensor<2xf32>, tensor<2xf32>,
        tensor<2xf32>, tensor<2xf32>, tensor<2xf32>, tensor<2xf32>,
        tensor<4x2xf32>, tensor<4xf32>,
        tensor<1x4xf32>, tensor<1x2xf32>,
        tensor<2xf32>, tensor<2xf32>, tensor<2xf32>, tensor<2xf32>) -> tensor<*xf32>
  %24 = "quantfork.stats"(%23) {layerStats = dense<[-1.0, 2.0]> : tensor<2xf32>} : (tensor<*xf32>) -> tensor<*xf32>

  %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>

"TANH", input_to_cell_intermediate = tensor<0xf32>, input_to_forget_intermediate = tensor<0xf32>, input_to_input_intermediate = tensor<0xf32>, input_to_output_intermediate = tensor<0xf32>, proj_clip = 0.000000e+00 : f32, time_major = false} : (tensor<?x?x?xf32>, tensor<?x?xf32>, tensor<?x?xf32>, tensor<?x?xf32>, tensor<?x?xf32>, tensor<?x?xf32>, tensor<?x?xf32>, tensor<?x?xf32>, tensor<?x?xf32>, tensor<?xf32>, tensor<?xf32>, tensor<?xf32>, tensor<?xf32>, tensor<?xf32>, tensor<?xf32>, tensor<?xf32>,...

func.func @multiply(%arg0: tensor<1xf32>, %arg1: tensor<1xf32>) -> tensor<1xf32> {
  %0 = "vhlo.multiply_v1"(%arg0, %arg1) : (tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>
  func.return %0 : tensor<1xf32>
}

// CHECK:func.func private @multiply(%arg0: tensor<1xf32>, %arg1: tensor<1xf32>) -> tensor<1xf32> {
// CHECK: %0 = "vhlo.multiply_v1"(%arg0, %arg1) : (tensor<1xf32>, tensor<1xf32>) -> tensor<1xf32>
// CHECK: return %0 : tensor<1xf32>
// CHECK:}

// CHECK-LABEL: func @maximum
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>

  } : (tensor<2x!tf_type.qint8>, tensor<f32>, tensor<i32>) -> tensor<2xf32>
  func.return %1 : tensor<2xf32>
}

// -----

// CHECK-LABEL: func @uniform_quantize_and_dequantize_per_axis
func.func @uniform_quantize_and_dequantize_per_axis(%arg0 : tensor<2x2xf32>) -> tensor<2x2xf32> {
  %scales = "tf.Const"() { value = dense<[1.0, 2.0]> : tensor<2xf32> } : () -> tensor<2xf32>

        none, none, none,
        tensor<3xf32>, tensor<3xf32>, tensor<3xf32>, tensor<3xf32>,
        none, none,
        tensor<1x3xf32>, tensor<1x3xf32>,
        none, none, none, none) -> tensor<1x2x3xf32>
  %17 = "quantfork.stats"(%16) {layerStats = dense<[-0.1, 0.1]> : tensor<2xf32>} : (tensor<1x2x3xf32>) -> tensor<1x2x3xf32>
  func.return %17 : tensor<1x2x3xf32>