func.func @quantized_types(%arg0: tensor<!tf_type.resource<tensor<1x3x!tf_type.quint8>>>,
                      %arg1: tensor<!tf_type.resource<tensor<1x3x!tf_type.quint16>>>,
                      %arg2: tensor<!tf_type.resource<tensor<1x3x!tf_type.qint8>>>,
                      %arg3: tensor<!tf_type.resource<tensor<1x3x!tf_type.qint16>>>,
                      %arg4: tensor<!tf_type.resource<tensor<1x3x!tf_type.qint32>>>) {
  // CHECK: tf.ReadVariableOp

  %5 = "tf.Identity"(%4) {__op_key = 10: i32, device = "/device:CPU:0"} : (tensor<!tf_type.variant>) -> tensor<!tf_type.variant>
  func.return %5 : tensor<!tf_type.variant>
}
// CHECK-LABEL: __inference_Dataset_map_lambda_160

    // CHECK-NEXT: %[[UDQ:.*]] = "tf.UniformDequantize"(%[[UQ]], %[[CST_0]], %[[CST_1]]) <{quantization_axis = -1 : i64, quantization_max_val = 127 : i64, quantization_min_val = -128 : i64}> : (tensor<1x!tf_type.qint8>, tensor<f32>, tensor<i32>) -> tensor<1xf32>
    // CHECK-NEXT: return %[[UDQ]] : tensor<1xf32>

  "tfl.assign_variable"(%1, %5) : (tensor<!tf_type.resource>, tensor<1x2x1x3xf32>) -> ()
  func.return %arg0 : tensor<1x2x1x3x!quant.uniform<i8:f64, 1.0>>

// CHECK-NEXT:  %[[vh:.*]] = "tfl.var_handle"() <{container = "", shared_name = ""}> : () -> tensor<*x!tf_type.resource<tensor<1x2x1x3x!quant.uniform<i8:f32, 1.000000e+00>>>>

%outputs_18, %outputs, %outputs_14, %outputs_28, %arg0) {T = [i32, i32, i32, !tf_type.variant, !tf_type.variant, f32], _lower_using_switch_merge = true, _num_original_outputs = 6 : i64, _read_only_resource_inputs = [], body = @map_while_body_180, cond = @map_while_cond_170, device = "", is_stateless = true, output_shapes = [#tf_type.shape<>, #tf_type.shape<>, #tf_type.shape<>, #tf_type.shape<>, #tf_type.shape<>, #tf_type.shape<?>], parallel_iterations = 4 : i64, shape_invariant} : (tensor<i32>, tensor<i32>,...

    %0 = "tf.HashTableV2"() {container = "", device = "", key_dtype = i64, shared_name = "hash_table_dba2ccaa-f1b1-46d6-b276-98008f69da71", use_node_name_sharing = false, value_dtype = !tf_type.string} : () -> tensor<!tf_type.resource>
    "tf.LookupTableImportV2"(%0, %cst, %cst_0) {device = ""} : (tensor<!tf_type.resource>, tensor<4xi64>, tensor<4x!tf_type.string>) -> ()

// CHECK-NEXT:    return
//
 func.func @serving_default(%arg0: tensor<1x3xf32>) -> tensor<1x1xf32> {
    %0 = "tf.VarHandleOp"() <{container = "", shared_name = "y"}> : () -> tensor<!tf_type.resource<tensor<3x1xf32>>>

    %5 = "tf.ReadVariableOp"(%arg5) : (tensor<*x!tf_type.resource<tensor<1024x1xf32>>>) -> tensor<1024x1xf32>
    %6 = "tf.ReadVariableOp"(%arg20) : (tensor<*x!tf_type.resource<tensor<1024xf32>>>) -> tensor<1024xf32>
    %7 = "tf.ReadVariableOp"(%arg21) : (tensor<*x!tf_type.resource<tensor<1024x1xf32>>>) -> tensor<1024x1xf32>

  %ragged_keys = "tf.Const"() {device = "/device:CPU:0", dtype = !tf_type.string, value = dense<""> : tensor<0x!tf_type.string>} : () -> tensor<0x!tf_type.string>
  %sparse_keys = "tf.Const"() {device = "/device:CPU:0", dtype = !tf_type.string, value = dense<""> : tensor<2x!tf_type.string>} : () -> tensor<2x!tf_type.string>

// CHECK-SAME: (%[[ARG_0:.*]]: tensor<!tf_type.resource<tensor<32xf32>>>, %[[ARG_1:.*]]: tensor<!tf_type.resource<tensor<32xf32>>>, %[[ARG_2:.*]]: tensor<!tf_type.resource<tensor<32xf32>>>, %[[ARG_3:.*]]: tensor<!tf_type.resource<tensor<32xf32>>>, %[[ARG_4:.*]]: tensor<!tf_type.resource<tensor<32xf32>>>)