// CHECK-SAME: (%[[ARG_0:[a-z0-9]*]]: tensor<*x!tf_type.resource>, %[[ARG_1:[a-z0-9]*]]: tensor<*x!tf_type.resource>)
func.func @replicate_resource_var_arg_shape(%arg0: tensor<*x!tf_type.resource>, %arg1: tensor<*x!tf_type.resource>) {
  %0:6 = tf_device.replicate([%arg0, %arg1] as %ri: tensor<*x!tf_type.resource>) {n = 2: i32} {
    %1 = "tf.ReadVariableOp"(%ri) {dtype = "tfdtype$DT_FLOAT"} : (tensor<*x!tf_type.resource>) -> tensor<*xf32>

      %outputs_5, %control_6 = tf_executor.island wraps "tf.TPUReplicatedInput"(%outputs_3) {N = 1 : i64, T = "tfdtype$DT_FLOAT", device = "", name = "input0"} : (tensor<0xf32>) -> tensor<0xf32>

    %0:2 = tf_executor.island wraps "tf.Const"() {dtype = "tfdtype$DT_INT32", value = dense<0> : tensor<i32>} : () -> (tensor<i32>) loc("foo")
    // CHECK: name: "foo{{_.*_1}}"
    %1:2 = tf_executor.island wraps "tf.Const"() {dtype = "tfdtype$DT_INT32", value = dense<1> : tensor<i32>} : () -> (tensor<i32>) loc("foo")
    // CHECK: name: "foo1"

  %0 = "tf.Einsum"(%arg0, %arg1) {T = "tfdtype$DT_FLOAT", equation = "ijk,j->i"}: (tensor<4x5x6xf32>, tensor<5xf32>) -> tensor<4xf32>
  func.return %0 : tensor<4xf32>
// CHECK-LABEL: einsum_no_match
// CHECK: %[[v0:.*]] = "tf.Einsum"(%arg0, %arg1) <{equation = "ijk,j->i"}> {T = "tfdtype$DT_FLOAT"} : (tensor<4x5x6xf32>, tensor<5xf32>) -> tensor<4xf32>
// CHECK: return %[[v0]]
}

        attr_map = ""
      } : (tensor<*x!tf_type.qint8>, tensor<*x!tf_type.qint8>, tensor<*xf32>, tensor<*xi32>, tensor<*xf32>, tensor<*xi32>, tensor<*xf32>, tensor<*xi32>) -> tensor<*x!tf_type.qint32>
    func.return %conv_out : tensor<*x!tf_type.qint32>
  }

  // MatMul.
  func.func private @internal_matmul_fn(%input : tensor<*x!tf_type.qint8>, %filter : tensor<*x!tf_type.qint8>,

    // The following node would be renamed to bar2
    %3:2 = tf_executor.island wraps "tf.Identity"(%2) {device = "", dtype = "tfdtype$DT_INT32"} : (tensor<10xi32>) -> tensor<10xi32> loc ("bar")

                                %weight_scale, %weight_zp) {
        Tlhs = "tfdtype$DT_FLOAT",
        Trhs = "tfdtype$DT_QINT8",
        Tout = "tfdtype$DT_FLOAT",
        rhs_quantization_axis = -1,
        rhs_quantization_min_val = -128,
        rhs_quantization_max_val = 127,
        attr_map = ""
      } : (tensor<*xf32>, tensor<*x!tf_type.qint8>, tensor<*xf32>, tensor<*xi32>) -> tensor<*xf32>

    tensor<1x9x9x9x!tf_type.qint8>
  %quant_filter = "tf.Cast"(%filter) {} : (tensor<3x3x9x10xi8>) ->
    tensor<3x3x9x10x!tf_type.qint8>
  %0 = "tf.UniformQuantizedConvolution"(
    %quant_input, %quant_filter, %input_scale, %input_zp,
    %filter_scale, %filter_zp, %accum_scale, %accum_zp
  ) {
    Tin = "tfdtype$DT_QINT8", Tout = "tfdtype$DT_QINT32",
    attr_map = "", batch_group_count = 1 : i64,

    %0 = "tf.VarHandleOp"() {shared_name = "vocab_file"} : () -> tensor<*x!tf_type.resource<tensor<!tf_type.string>>>
    %1 = "tf.Const"() {value = dense<"assets/vocab_file.txt"> : tensor<!tf_type.string>} : () -> tensor<!tf_type.string>
    "tf.AssignVariableOp"(%0, %1) : (tensor<*x!tf_type.resource<tensor<!tf_type.string>>>, tensor<!tf_type.string>) -> ()

    %outputs_1, %control_1 = tf_executor.island wraps "tf.TPUReplicatedInput"(%outputs_0) {N = 1 : i64, T = "tfdtype$DT_FLOAT", device = "", name = "input1"} : (tensor<0xf32>) -> tensor<0xf32>