func.func @simple_folding(%arg0: tensor<1x1x1x1xi32>, %arg1: tensor<1x1x1x1xf32>) -> tensor<?x?x?x?xf32> {
    // CHECK: %[[SHAPE:.*]] = "tf.Shape"
    // CHECK: %[[CONV:.*]] = "tf.Conv2DBackpropInput"(%[[SHAPE]]
    // CHECK-SAME: (tensor<4xi32>, tensor<1x1x1x1xf32>, tensor<1x1x1x1xf32>) -> tensor<1x1x1x1xf32>
    // CHECK: return %[[CONV]] : tensor<1x1x1x1xf32>
    %0 = "tf.Shape"(%arg0) : (tensor<1x1x1x1xi32>) -> tensor<4xi32>

    } : (tensor<1x6x6x16xf32>) -> tensor<1x1x1x16xf32>
  func.return %1 : tensor<1x1x1x16xf32>

// CHECK: %0 = "tfl.dequantize"(%arg0)
// CHECK: %1 = "tfl.average_pool_2d"(%0)
// CHECK: %2 = "tfl.quantize"(%1)
// CHECK: %3 = "tfl.dequantize"(%2)
// CHECK: return %3 : tensor<1x1x1x16xf32>
}

// CHECK-LABEL: QuantizeMaximum

    quantization_axis = -1 : i64, quantization_min_val = -128 : i64, quantization_max_val = 127 : i64
  } : (tensor<1x2x2x1x!tf_type.qint32>, tensor<f32>, tensor<i32>) -> tensor<1x2x2x1xf32>

  return %output : tensor<1x2x2x1xi32>

// CHECK:           %[[VAL_7:.*]] = "tf.Const"() <{value = dense<[3, 1, 2, 0]> : tensor<4xi64>}> : () -> tensor<4xi64>
// CHECK:           %[[VAL_8:.*]] = "tf.Transpose"(%[[VAL_6]], %[[VAL_7]]) : (tensor<1x8x8x16xf32>, tensor<4xi64>) -> tensor<16x8x8x1xf32>
// CHECK:           return %[[VAL_8]] : tensor<16x8x8x1xf32>
// CHECK:         }

// CHECK-LABEL: func @tfAssertTrue
func.func @tfAssertTrue(%arg0: tensor<1x1x6x2xf32>) {
  %t = arith.constant dense<true> : tensor<i1>
  // CHECK-NOT: tf.Assert
  "tf.Assert"(%t, %arg0) {summarize = 3} : (tensor<i1>, tensor<1x1x6x2xf32>) -> ()
  func.return
}

// CHECK-LABEL: func @tfAssertFalse
func.func @tfAssertFalse(%arg0: tensor<1x1x6x2xf32>) {
  %f = arith.constant dense<false> : tensor<i1>
  // CHECK: tf.Assert

func.func @conv_valid_padding(%arg0: tensor<1x4x5x1xf32>, %arg1: tensor<3x3x1x1xf32>) -> tensor<1x2x3x1xf32> {
  // CHECK: mhlo.convolution(%arg0, %arg1)

  %0 = "tf.Conv2D"(%arg0, %arg1) {data_format = "NHWC", dilations = [1, 1, 1, 1], padding = "VALID", strides = [1, 1, 1, 1]} : (tensor<1x4x5x1xf32>, tensor<3x3x1x1xf32>) -> tensor<1x2x3x1xf32>
  func.return %0 : tensor<1x2x3x1xf32>
}

// -----

  // Move binary op batched RHS before reshape:
  // binary(reshape(lhs), rhs) => reshape(binary(lhs, flatten(rhs)))
  // Pattern targetted here is as follows-
  // [input, lhr, rhs] == [<1x1024x128>, <1x1024x8x16>, <1x1x8x16xf32>]
  // This is valid only when the-
  // 1.last dimension of lhs is equal to the number of elements in constant rhs.
  // 2.Reduded shape of rhs, here <8x16> is equal to last dimensions of lhs.