%1 = "quantfork.stats"(%0) {
    layerStats = dense<[-1.0, 1.0]> : tensor<2xf32>,
    axisStats = dense<[
      [-1.0, 1.0],
      [-8.0, 8.0],
      [-0.5, 0.5]
    ]> : tensor<3x2xf32>, axis = 2 : i64
  } : (tensor<8x4x3xf32>) -> tensor<8x4x3xf32>
  func.return %1 : tensor<8x4x3xf32>

// CHECK-LABEL: main
// CHECK: return %arg0 : tensor<5x7xf32>
}

// - transpose
//
func.func @transpose_2d(%arg0: tensor<2x3xf32>) -> tensor<3x2xf32> {
  %0 = "mhlo.transpose"(%arg0) <{permutation = dense<[1, 0]> : tensor<2xi64>}> : (tensor<2x3xf32>) -> tensor<3x2xf32>
  func.return %0 : tensor<3x2xf32>

// CHECK-LABEL:   transpose_2d
// CHECK-NEXT:    %0 = "tfl.pseudo_const"() <{value = dense<[1, 0]> : tensor<2xi64>}> : () -> tensor<2xi64>

  %value0 = "tfl.pseudo_const"() {value = dense<[[1.0, 2.0], [3.0, 4.0], [5.0, 6.0]]> : tensor<3x2xf32>} : () -> tensor<3x2xf32>
  %value1 = "tfl.pseudo_const"() {value = dense<[[[1.0, 2.0], [3.0, 4.0]], [[5.0, 6.0], [7.0, 8.0]]]> : tensor<2x2x2xf32>} : () -> tensor<2x2x2xf32>
  %lookup0 = "tfl.embedding_lookup"(%index0, %value0) : (tensor<5xi32>, tensor<3x2xf32>) -> tensor<5x2xf32>

  %input = "quantfork.stats"(%arg0) {
    layerStats = dense<[0.0, 1.0]> : tensor<2xf32>,
    axisStats = dense<[
      [-1.0, 1.0],
      [-8.0, 8.0],
      [-0.5, 0.5]
    ]> : tensor<3x2xf32>, axis = 2 : i64
  } : (tensor<1x2x3xf32>) -> tensor<1x2x3xf32>
  %1 = "tfl.pseudo_const"() {value = dense<[[0.1]]> : tensor<1x1xf32>} : () -> tensor<1x1xf32>

  func.func @cluster(%arg0: tensor<!tf_type.resource<tensor<3x2xf32>>>, %arg1: tensor<!tf_type.resource<tensor<3x2xf32>>>) {
    // CHECK: %[[READ_VAR_0:[0-9]*]] = "tf.ReadVariableOp"(%arg0)
    %read0 = "tf.ReadVariableOp"(%arg0) : (tensor<!tf_type.resource<tensor<3x2xf32>>>) -> tensor<3x2xf32>
    // CHECK: %[[READ_VAR_1:[0-9]*]] = "tf.ReadVariableOp"(%arg1)

  %0 = "tf.BroadcastTo"(%input, %shape) : (tensor<3xi16>, tensor<2xi32>) -> tensor<3x3xi16>
  func.return %0: tensor<3x3xi16>

// CHECK-LABEL: broadcast_to_i16_low_dim
// CHECK:    %0 = "tf.BroadcastTo"(%arg0, %arg1) : (tensor<3xi16>, tensor<2xi32>) -> tensor<3x3xi16>
// CHECK:    return %0 : tensor<3x3xi16>
}

  // CHECK-DAG: %[[OPERAND:.*]] = mhlo.uniform_quantize %arg0 : (tensor<3x2xf32>) -> tensor<3x2x!quant.uniform<i32:f32:1, {2.000000e+00:4,2.000000e+00:4}>>
  %0 = "tf.UniformQuantize"(%input, %scales, %zps) {
    quantization_axis = 1 : i64, quantization_min_val = -2147483648 : i64, quantization_max_val = 2147483647 : i64
  } : (tensor<3x2xf32>, tensor<2xf32>, tensor<2xi32>) -> tensor<3x2x!tf_type.qint32>

// `tfl.broadcast_to`.

func.func @broadcast_in_dim_float(%arg0: tensor<1x2xf32>) -> tensor<3x2xf32> {
  %0 = "stablehlo.broadcast_in_dim"(%arg0) {
    broadcast_dimensions = array<i64: 0, 1>
  } : (tensor<1x2xf32>) -> tensor<3x2xf32>
  return %0 : tensor<3x2xf32>
}
// CHECK-LABEL: broadcast_in_dim_float
// CHECK-NOT: tfl.broadcast_to
// CHECK-NOT: tfl.transpose

				for j := 0; j < hi*vi; j++ {
					// The blocks are traversed one MCU at a time. For 4:2:0 chroma
					// subsampling, there are four Y 8x8 blocks in every 16x16 MCU.
					//
					// For a sequential 32x16 pixel image, the Y blocks visiting order is:
					//	0 1 4 5
					//	2 3 6 7
					//
					// For progressive images, the interleaved scans (those with nComp > 1)

  %0 = "tfl.broadcast_to"(%arg0, %arg1) : (tensor<3xi16>, tensor<2xi32>) -> tensor<3x3xi16>
  return %0 : tensor<3x3xi16>
  // CHECK:  %cst = arith.constant dense<1> : tensor<3x3xi16>
  // CHECK:  %0 = tfl.mul(%arg0, %cst) <{fused_activation_function = "NONE"}> : (tensor<3xi16>, tensor<3x3xi16>) -> tensor<3x3xi16>
  // CHECK:  return %0 : tensor<3x3xi16>
}