Targs = [tensor<32xf32>, tensor<64xf32>, tensor<8xf32>, tensor<2xf32>],
      Tresults = [tensor<32xf32>, tensor<64xf32>, tensor<8xf32>]}
      : (tensor<32xf32>, tensor<64xf32>, tensor<8xf32>, tensor<2xf32>, tensor<2x!tf_type.string>)
        -> (tensor<32xf32>, tensor<64xf32>, tensor<8xf32>)
    tf_device.return %0#0, %0#1, %0#2 : tensor<32xf32>, tensor<64xf32>, tensor<8xf32>

         data_format = "NCHW",
         epsilon = 1.001000e-05 : f32,
         is_training = false
       } : (tensor<?x64x112x112xf32>, tensor<64xf32>, tensor<64xf32>, tensor<64xf32>, tensor<64xf32>)
        -> (tensor<?x64x112x112xf32>, tensor<64xf32>, tensor<64xf32>, tensor<64xf32>, tensor<64xf32>, tensor<*xf32>)

  // CHECK: "tf.FusedBatchNormV3"
  // CHECK-SAME: data_format = "NHWC"

         exponential_avg_factor = 1.0 : f32,
         is_training = true
       }
        : (tensor<1x28x28x64xf32>, tensor<64xf32>, tensor<64xf32>,
           tensor<64xf32>, tensor<64xf32>)
       -> (tensor<1x28x28x64xf32>, tensor<64xf32>, tensor<64xf32>,
           tensor<64xf32>, tensor<64xf32>, tensor<64xf32>)

  func.return %y : tensor<1x28x28x64xf32>
}

// CHECK-LABEL: func @transposeFusedBatchNormGradV3

         exponential_avg_factor = 1.0 : f32,
         is_training = true
       }
        : (tensor<1x64x28x28xf32>, tensor<64xf32>, tensor<64xf32>,
           tensor<64xf32>, tensor<64xf32>)
       -> (tensor<1x64x28x28xf32>, tensor<64xf32>, tensor<64xf32>,
           tensor<64xf32>, tensor<64xf32>, tensor<64xf32>)

  func.return %y : tensor<1x64x28x28xf32>
}

// CHECK-LABEL: bias_add_nchw

         exponential_avg_factor = 1.0 : f32,
         is_training = false
       }
        : (tensor<1x112x112x64xf32>, tensor<64xf32>, tensor<64xf32>, tensor<64xf32>, tensor<64xf32>)
       -> (tensor<1x112x112x64xf32>, tensor<64xf32>, tensor<64xf32>, tensor<64xf32>, tensor<64xf32>, tensor<64xf32>)

  func.return %2#0 : tensor<1x112x112x64xf32>
}

// CHECK-LABEL: func @fold_into_pad_with_extra_uses

  %b = arith.constant dense<-1.23697901> : tensor<64xf32>
  %conv = "tfl.conv_2d"(%arg0, %w, %b) {dilation_h_factor = 1 : i32, dilation_w_factor = 1 : i32, fused_activation_function = "NONE", padding = "SAME", stride_h = 2 : i32, stride_w = 2 : i32} : (tensor<1x224x224x3xf32>, tensor<64x3x3x3xf32>, tensor<64xf32>) -> tensor<1x112x112x64xf32>
  func.return %conv : tensor<1x112x112x64xf32>

  %0 = "tfl.reshape"(%arg0, %shape0) : (tensor<4x4x4xf32>, tensor<2xi32>) -> tensor<16x4xf32>
  %1 = "tfl.reshape"(%0, %shape1) : (tensor<16x4xf32>, tensor<1xi32>) -> tensor<64xf32>
  %2 = "tfl.reshape"(%0, %shape1) : (tensor<16x4xf32>, tensor<1xi32>) -> tensor<64xf32>
  %3 = arith.addf %1, %2 : tensor<64xf32>
  func.return %3 : tensor<64xf32>

// CHECK-LABEL: func @reshape_removeAdjacentWithMultipleUse

  %9 = mhlo.floor %8 : tensor<64xf32>
  %10 = mhlo.convert %9 : (tensor<64xf32>) -> tensor<64xi32>
  %11 = mhlo.compare  LT, %10, %1,  SIGNED : (tensor<64xi32>, tensor<64xi32>) -> tensor<64xi1>
  %12 = mhlo.add %10, %0 : tensor<64xi32>
  %13 = mhlo.select %11, %12, %10 : tensor<64xi1>, tensor<64xi32>
  %14 = mhlo.reshape %13 : (tensor<64xi32>) -> tensor<64x1xi32>

        tensor<20xf32>, tensor<20xf32>, tensor<20xf32>,
        tensor<20xf32>, tensor<20xf32>, tensor<20xf32>, tensor<20xf32>,
        tensor<20x20xf32>, none,
        tensor<1x20xf32>, tensor<1x20xf32>,
        none, none, none, none) -> tensor<1x28x20xf32>
    %1 = "quantfork.stats"(%0) {layerStats = dense<[-1.0, 2.0]> : tensor<2xf32>} : (tensor<1x28x20xf32>) -> tensor<1x28x20xf32>

  // CHECK: %[[VAL_4:.*]] = "tf.Identity"(%[[VAL_3]]) : (tensor<*xf32>) -> tensor<*xf32>
  // CHECK: return %[[VAL_4]]
  %3 = "tf.MatMul"(%arg1, %arg2) {transpose_a = false, transpose_b = false} : (tensor<8x32xf32>, tensor<32x64xf32>) -> tensor<*xf32>
  %4 = "tf.BiasAdd"(%3, %arg0) {data_format = "NHWC"} : (tensor<*xf32>, tensor<64xf32>) -> tensor<*xf32>
  %5 = "tf.Elu"(%4) : (tensor<*xf32>) -> tensor<*xf32>