%2 = stablehlo.add %1, %arg1 : tensor<1x4x5x5xf32>
  return %2 : tensor<1x4x5x5xf32>
}
// CHECK-DAG: %[[WEIGHT_CONST:.+]] = stablehlo.constant {{.*}} : tensor<3x3x2x4xf32>
// CHECK-DAG: %[[TRANSPOSE_0:.+]] = stablehlo.transpose %[[ARG_0]], dims = [0, 2, 3, 1] : (tensor<1x2x5x5xf32>) -> tensor<1x5x5x2xf32>

// CHECK: return %6 : tensor<1x5x5x3xf32>
}

// CHECK-LABEL: bias_adjust_duplicate_filter
func.func @bias_adjust_duplicate_filter(%arg0: tensor<1x5x5x2xf32>) -> (tensor<1x5x5x3xf32>, tensor<1x5x5x3xf32>) {
  %0 = "quantfork.stats"(%arg0) {
    layerStats = dense<[-1.28e-5, 1.27e-5]> : tensor<2xf32>
  } : (tensor<1x5x5x2xf32>) -> tensor<1x5x5x2xf32>

func.func @CheckNumericVerifyWholeModelNoQuantizeOps(%arg0: tensor<?x5x5x2xf32>) -> (tensor<?x1x1x3xf32>) {
  %0 = "quantfork.stats"(%arg0) {
    layerStats = dense<[0.0, 1.0]> : tensor<2xf32>
  } : (tensor<?x5x5x2xf32>) -> tensor<?x5x5x2xf32>
  %1 = "tfl.pseudo_const"() {value = dense<1.000000e+00> : tensor<3x5x5x2xf32>} : () -> tensor<3x5x5x2xf32>

  %2 = "tf.BatchToSpaceND"(%1, %cst, %cst_0) : (tensor<4x64x64x8xf32>, tensor<2xi32>, tensor<2x2xi32>) -> tensor<1x120x120x8xf32>
  func.return %2 : tensor<1x120x120x8xf32>

  // CHECK-LABEL: testDilatedConv
  // CHECK-SAME: ([[INPUT:%.*]]: tensor<1x128x128x3xf32>, [[FILTER:%.*]]: tensor<5x5x3x8xf32>)

  func.return %47 : tensor<*xf32>

  // CHECK-NOT: %[[slice:.*]] = "tfl.slice"
}

// CHECK-LABEL: @StridedSliceToSliceBeginNeg
func.func @StridedSliceToSliceBeginNeg(%arg0: tensor<5x5x5x5xf32>) -> tensor<*xf32> {
  %44 = arith.constant dense<[-5, 0, 0, 0]> : tensor<4xi32>

module {
func.func @main(%arg0: tensor<4x68x68x3xf32>, %arg1: tensor<5x5x3x8xf32>) -> tensor<4x64x64x8xf32> {
  %0 = "tf.Conv2D"(%arg0, %arg1) {padding = "VALID", strides = [1, 1, 1, 1]} : (tensor<4x68x68x3xf32>, tensor<5x5x3x8xf32>) -> tensor<4x64x64x8xf32>
  func.return %0 : tensor<4x64x64x8xf32>
}