Enable op-level dataset determinism configuration for ParseExampleDataset

Users can control determinism at a per-op level by specifying `deterministic` when calling parse_example_dataset(). The `deterministic` argument takes higher priority than the `experimental_deterministic` dataset option.

PiperOrigin-RevId: 294957589
Change-Id: If974e5fdd403be99a214dd352b4095048afb4701

tensorflow/core/api_def/base_api/api_def_ParseExampleDatasetV2.pbtxt

@@ -0,0 +1,80 @@
+op {
+  graph_op_name: "ParseExampleDatasetV2"
+  visibility: HIDDEN
+  in_arg {
+    name: "dense_defaults"
+    description: <<END
+A dict mapping string keys to `Tensor`s.
+The keys of the dict must match the dense_keys of the feature.
+END
+  }
+  attr {
+    name: "sparse_keys"
+    description: <<END
+A list of string keys in the examples features.
+The results for these keys will be returned as `SparseTensor` objects.
+END
+  }
+  attr {
+    name: "dense_keys"
+    description: <<END
+A list of Ndense string Tensors (scalars).
+The keys expected in the Examples features associated with dense values.
+END
+  }
+  attr {
+    name: "sparse_types"
+    description: <<END
+A list of `DTypes` of the same length as `sparse_keys`.
+Only `tf.float32` (`FloatList`), `tf.int64` (`Int64List`),
+and `tf.string` (`BytesList`) are supported.
+END
+  }
+    attr {
+    name: "Tdense"
+    description: <<END
+A list of DTypes of the same length as `dense_keys`.
+Only `tf.float32` (`FloatList`), `tf.int64` (`Int64List`),
+and `tf.string` (`BytesList`) are supported.
+
+END
+  }
+  attr {
+    name: "dense_shapes"
+    description: <<END
+List of tuples with the same length as `dense_keys`.
+The shape of the data for each dense feature referenced by `dense_keys`.
+Required for any input tensors identified by `dense_keys`.  Must be
+either fully defined, or may contain an unknown first dimension.
+An unknown first dimension means the feature is treated as having
+a variable number of blocks, and the output shape along this dimension
+is considered unknown at graph build time.  Padding is applied for
+minibatch elements smaller than the maximum number of blocks for the
+given feature along this dimension.
+END
+  }
+    attr {
+    name: "output_types"
+    description: <<END
+The type list for the return values.
+END
+  }
+    attr {
+    name: "output_shapes"
+    description: <<END
+The list of shapes being produced.
+END
+  }
+    attr {
+    name: "deterministic"
+    description: <<END
+A string indicating the op-level determinism to use. Deterministic controls
+whether the dataset is allowed to return elements out of order if the next
+element to be returned isn't available, but a later element is. Options are
+"true", "false", and "default". "default" indicates that determinism should be
+decided by the `experimental_deterministic` parameter of `tf.data.Options`.
+END
+  }
+   summary: "Transforms `input_dataset` containing `Example` protos as vectors of DT_STRING into a dataset of `Tensor` or `SparseTensor` objects representing the parsed features."
+}
+

tensorflow/core/grappler/optimizers/data/make_sloppy.cc

@@ -32,10 +32,11 @@ constexpr std::array<const char*, 3> kSloppyAttrOps = {
     "ParseExampleDataset",
 };
 
-constexpr std::array<const char*, 3> kDeterministicAttrOps = {
+constexpr std::array<const char*, 4> kDeterministicAttrOps = {
     "ParallelInterleaveDatasetV3",
     "ParallelInterleaveDatasetV4",
     "ParallelMapDatasetV2",
+    "ParseExampleDatasetV2",
 };
 }  // anonymous namespace
 

tensorflow/core/kernels/data/experimental/BUILD

@@ -282,6 +282,8 @@ tf_kernel_library(
         "//tensorflow/core:core_cpu_internal",
         "//tensorflow/core:framework",
         "//tensorflow/core:functional_ops_op_lib",
+        "//tensorflow/core/kernels/data:dataset_utils",
+        "//tensorflow/core/kernels/data:name_utils",
         "//tensorflow/core/kernels/data:parallel_map_dataset_op",
         "//tensorflow/core/kernels/data:stats_utils",
     ],

tensorflow/core/kernels/data/experimental/parse_example_dataset_op.cc

@@ -17,6 +17,8 @@ limitations under the License.
 #include "tensorflow/core/common_runtime/device.h"
 #include "tensorflow/core/common_runtime/metrics.h"
 #include "tensorflow/core/framework/stats_aggregator.h"
+#include "tensorflow/core/kernels/data/dataset_utils.h"
+#include "tensorflow/core/kernels/data/name_utils.h"
 #include "tensorflow/core/kernels/data/parallel_map_dataset_op.h"
 #include "tensorflow/core/kernels/data/stats_utils.h"
 #include "tensorflow/core/util/example_proto_fast_parsing.h"
@@ -30,7 +32,8 @@ class ParseExampleDatasetOp : public UnaryDatasetOpKernel {
  public:
   explicit ParseExampleDatasetOp(OpKernelConstruction* ctx)
       : UnaryDatasetOpKernel(ctx),
-        graph_def_version_(ctx->graph_def_version()) {
+        graph_def_version_(ctx->graph_def_version()),
+        op_version_(ctx->HasAttr("deterministic") ? 2 : 1) {
     OP_REQUIRES_OK(ctx, ctx->GetAttr("sparse_keys", &sparse_keys_));
     OP_REQUIRES_OK(ctx, ctx->GetAttr("dense_keys", &dense_keys_));
     OP_REQUIRES_OK(ctx, ctx->GetAttr("sparse_types", &sparse_types_));
@@ -38,7 +41,24 @@ class ParseExampleDatasetOp : public UnaryDatasetOpKernel {
     OP_REQUIRES_OK(ctx, ctx->GetAttr("dense_shapes", &dense_shapes_));
     OP_REQUIRES_OK(ctx, ctx->GetAttr("output_types", &output_types_));
     OP_REQUIRES_OK(ctx, ctx->GetAttr("output_shapes", &output_shapes_));
-    OP_REQUIRES_OK(ctx, ctx->GetAttr("sloppy", &sloppy_));
+
+    if (op_version_ == 1) {
+      bool sloppy;
+      OP_REQUIRES_OK(ctx, ctx->GetAttr("sloppy", &sloppy));
+      if (sloppy) {
+        deterministic_ =
+            DeterminismPolicy(DeterminismPolicy::Type::kNondeterministic);
+      } else {
+        deterministic_ = DeterminismPolicy(DeterminismPolicy::Type::kDefault);
+      }
+    }
+    if (op_version_ == 2) {
+      std::string deterministic;
+      OP_REQUIRES_OK(ctx, ctx->GetAttr("deterministic", &deterministic));
+      OP_REQUIRES_OK(
+          ctx, DeterminismPolicy::FromString(deterministic, &deterministic_));
+    }
+
     has_ragged_keys_ = ctx->HasAttr("ragged_keys");
     if (has_ragged_keys_) {
       OP_REQUIRES_OK(ctx, ctx->GetAttr("ragged_keys", &ragged_keys_));
@@ -162,12 +182,12 @@ class ParseExampleDatasetOp : public UnaryDatasetOpKernel {
       it->second = i++;
     }
 
-    *output = new Dataset(ctx, input, dense_defaults, sparse_keys_, dense_keys_,
+    *output = new Dataset(
-                          std::move(key_to_output_index), std::move(config),
+        ctx, input, dense_defaults, sparse_keys_, dense_keys_,
-                          num_parallel_calls, sparse_types_, dense_types_,
+        std::move(key_to_output_index), std::move(config), num_parallel_calls,
-                          dense_shapes_, output_types_, output_shapes_, sloppy_,
+        sparse_types_, dense_types_, dense_shapes_, output_types_,
-                          has_ragged_keys_, ragged_keys_, ragged_value_types_,
+        output_shapes_, deterministic_, has_ragged_keys_, ragged_keys_,
-                          ragged_split_types_);
+        ragged_value_types_, ragged_split_types_, op_version_);
   }
 
  private:
@@ -182,10 +202,11 @@ class ParseExampleDatasetOp : public UnaryDatasetOpKernel {
             const DataTypeVector& dense_types,
             const std::vector<PartialTensorShape>& dense_shapes,
             const DataTypeVector& output_types,
-            const std::vector<PartialTensorShape>& output_shapes, bool sloppy,
+            const std::vector<PartialTensorShape>& output_shapes,
-            bool has_ragged_keys, std::vector<string> ragged_keys,
+            const DeterminismPolicy& deterministic, bool has_ragged_keys,
+            std::vector<string> ragged_keys,
             const DataTypeVector& ragged_value_types,
-            const DataTypeVector& ragged_split_types)
+            const DataTypeVector& ragged_split_types, int op_version)
         : DatasetBase(DatasetContext(ctx)),
           input_(input),
           dense_defaults_(std::move(dense_defaults)),
@@ -202,8 +223,9 @@ class ParseExampleDatasetOp : public UnaryDatasetOpKernel {
           dense_shapes_(dense_shapes),
           output_types_(output_types),
           output_shapes_(output_shapes),
-          sloppy_(sloppy),
+          deterministic_(deterministic),
-          has_ragged_keys_(has_ragged_keys) {
+          has_ragged_keys_(has_ragged_keys),
+          op_version_(op_version) {
       input_->Ref();
     }
 
@@ -213,9 +235,14 @@ class ParseExampleDatasetOp : public UnaryDatasetOpKernel {
         const string& prefix) const override {
       std::unique_ptr<ParallelMapFunctor> parse_example_functor =
           absl::make_unique<ParseExampleFunctor>(this);
+      name_utils::IteratorPrefixParams params;
+      params.op_version = op_version_;
+      bool deterministic =
+          deterministic_.IsDeterministic() || deterministic_.IsDefault();
       return NewParallelMapIterator(
-          {this, strings::StrCat(prefix, "::ParseExample")}, input_,
+          {this, name_utils::IteratorPrefix("ParseExample", prefix, params)},
-          std::move(parse_example_functor), num_parallel_calls_, !sloppy_,
+          input_, std::move(parse_example_functor), num_parallel_calls_,
+          deterministic,
           /*preserve_cardinality=*/true);
     }
 
@@ -228,7 +255,9 @@ class ParseExampleDatasetOp : public UnaryDatasetOpKernel {
     }
 
     string DebugString() const override {
-      return "ParseExampleDatasetOp::Dataset";
+      name_utils::DatasetDebugStringParams params;
+      params.op_version = op_version_;
+      return name_utils::DatasetDebugString("ParseExampleDataset", params);
     }
 
     int64 Cardinality() const override { return input_->Cardinality(); }
@@ -257,60 +286,60 @@ class ParseExampleDatasetOp : public UnaryDatasetOpKernel {
         dense_defaults_nodes.emplace_back(node);
       }
 
-      AttrValue sparse_keys_attr;
+      std::vector<std::pair<StringPiece, AttrValue>> attrs;
-      AttrValue dense_keys_attr;
-      AttrValue sparse_types_attr;
-      AttrValue dense_attr;
-      AttrValue dense_shapes_attr;
-      AttrValue sloppy_attr;
 
+      AttrValue sparse_keys_attr;
       b->BuildAttrValue(sparse_keys_, &sparse_keys_attr);
+      attrs.emplace_back("sparse_keys", sparse_keys_attr);
+
+      AttrValue dense_keys_attr;
       b->BuildAttrValue(dense_keys_, &dense_keys_attr);
+      attrs.emplace_back("dense_keys", dense_keys_attr);
+
+      AttrValue sparse_types_attr;
       b->BuildAttrValue(sparse_types_, &sparse_types_attr);
+      attrs.emplace_back("sparse_types", sparse_types_attr);
+
+      AttrValue dense_attr;
       b->BuildAttrValue(dense_types_, &dense_attr);
+      attrs.emplace_back("Tdense", dense_attr);
+
+      AttrValue dense_shapes_attr;
       b->BuildAttrValue(dense_shapes_, &dense_shapes_attr);
-      b->BuildAttrValue(sloppy_, &sloppy_attr);
+      attrs.emplace_back("dense_shapes", dense_shapes_attr);
+
+      if (op_version_ == 1) {
+        AttrValue sloppy_attr;
+        b->BuildAttrValue(deterministic_.IsNondeterministic(), &sloppy_attr);
+        attrs.emplace_back("sloppy", sloppy_attr);
+      }
+      if (op_version_ == 2) {
+        AttrValue deterministic_attr;
+        b->BuildAttrValue(deterministic_.String(), &deterministic_attr);
+        attrs.emplace_back("deterministic", deterministic_attr);
+      }
 
       if (has_ragged_keys_) {
         AttrValue ragged_keys_attr;
-        AttrValue ragged_value_types_attr;
-        AttrValue ragged_split_types_attr;
         b->BuildAttrValue(ragged_keys_, &ragged_keys_attr);
-        b->BuildAttrValue(ragged_value_types_, &ragged_value_types_attr);
+        attrs.emplace_back("ragged_keys", ragged_keys_attr);
-        b->BuildAttrValue(ragged_split_types_, &ragged_split_types_attr);
 
-        TF_RETURN_IF_ERROR(
+        AttrValue ragged_value_types_attr;
-            b->AddDataset(this,
+        b->BuildAttrValue(ragged_value_types_, &ragged_value_types_attr);
-                          {
+        attrs.emplace_back("ragged_value_types", ragged_value_types_attr);
-                              {0, input_graph_node},
+
-                              {1, num_parallle_calls_node},
+        AttrValue ragged_split_types_attr;
-                          },
+        b->BuildAttrValue(ragged_split_types_, &ragged_split_types_attr);
-                          {{2, dense_defaults_nodes}},
+        attrs.emplace_back("ragged_split_types", ragged_split_types_attr);
-                          {{"sparse_keys", sparse_keys_attr},
-                           {"dense_keys", dense_keys_attr},
-                           {"sparse_types", sparse_types_attr},
-                           {"Tdense", dense_attr},
-                           {"dense_shapes", dense_shapes_attr},
-                           {"sloppy", sloppy_attr},
-                           {"ragged_keys", ragged_keys_attr},
-                           {"ragged_value_types", ragged_value_types_attr},
-                           {"ragged_split_types", ragged_split_types_attr}},
-                          output));
-      } else {
-        TF_RETURN_IF_ERROR(b->AddDataset(this,
-                                         {
-                                             {0, input_graph_node},
-                                             {1, num_parallle_calls_node},
-                                         },
-                                         {{2, dense_defaults_nodes}},
-                                         {{"sparse_keys", sparse_keys_attr},
-                                          {"dense_keys", dense_keys_attr},
-                                          {"sparse_types", sparse_types_attr},
-                                          {"Tdense", dense_attr},
-                                          {"dense_shapes", dense_shapes_attr},
-                                          {"sloppy", sloppy_attr}},
-                                         output));
       }
+
+      TF_RETURN_IF_ERROR(b->AddDataset(this,
+                                       {
+                                           {0, input_graph_node},
+                                           {1, num_parallle_calls_node},
+                                       },
+                                       {{2, dense_defaults_nodes}}, attrs,
+                                       output));
       return Status::OK();
     }
 
@@ -443,14 +472,15 @@ class ParseExampleDatasetOp : public UnaryDatasetOpKernel {
     const std::vector<PartialTensorShape> dense_shapes_;
     const DataTypeVector output_types_;
     const std::vector<PartialTensorShape> output_shapes_;
-    const bool sloppy_;
+    const DeterminismPolicy deterministic_;
     const bool has_ragged_keys_;
+    const int op_version_;
   };
 
   const int graph_def_version_;
   DataTypeVector output_types_;
   std::vector<PartialTensorShape> output_shapes_;
-  bool sloppy_;
+  DeterminismPolicy deterministic_;
   std::vector<string> sparse_keys_;
   std::vector<string> dense_keys_;
   std::vector<string> ragged_keys_;
@@ -462,10 +492,13 @@ class ParseExampleDatasetOp : public UnaryDatasetOpKernel {
   std::vector<bool> variable_length_;
   std::vector<std::size_t> elements_per_stride_;
   bool has_ragged_keys_;
+  const int op_version_;
 };
 
 REGISTER_KERNEL_BUILDER(Name("ParseExampleDataset").Device(DEVICE_CPU),
                         ParseExampleDatasetOp);
+REGISTER_KERNEL_BUILDER(Name("ParseExampleDatasetV2").Device(DEVICE_CPU),
+                        ParseExampleDatasetOp);
 REGISTER_KERNEL_BUILDER(
     Name("ExperimentalParseExampleDataset").Device(DEVICE_CPU),
     ParseExampleDatasetOp);

tensorflow/core/ops/experimental_dataset_ops.cc

@@ -596,6 +596,27 @@ REGISTER_OP("ParseExampleDataset")
     .Attr("ragged_split_types: list({int32,int64}) >= 0 = []")
     .SetShapeFn(shape_inference::ScalarShape);
 
+REGISTER_OP("ParseExampleDatasetV2")
+    .Input("input_dataset: variant")
+    .Input("num_parallel_calls: int64")
+    .Input("dense_defaults: Tdense")
+    .Output("handle: variant")
+    .Attr("sparse_keys: list(string) >= 0")
+    .Attr("dense_keys: list(string) >= 0")
+    .Attr("sparse_types: list({float,int64,string}) >= 0")
+    .Attr("Tdense: list({float,int64,string}) >= 0")
+    .Attr("dense_shapes: list(shape) >= 0")
+    .Attr("output_types: list(type) >= 1")
+    .Attr("output_shapes: list(shape) >= 1")  // Output components will be
+                                              // sorted by key (dense_keys and
+                                              // sparse_keys combined) here.
+    // "true", "false", or "default".
+    .Attr("deterministic: string = 'default'")
+    .Attr("ragged_keys: list(string) >= 0 = []")
+    .Attr("ragged_value_types: list({float,int64,string}) >= 0 = []")
+    .Attr("ragged_split_types: list({int32,int64}) >= 0 = []")
+    .SetShapeFn(shape_inference::ScalarShape);
+
 REGISTER_OP("ExperimentalParseExampleDataset")
     .Input("input_dataset: variant")
     .Input("num_parallel_calls: int64")

tensorflow/python/data/experimental/kernel_tests/BUILD

@@ -476,6 +476,7 @@ tf_py_test(
     name = "parse_example_dataset_test",
     size = "small",
     srcs = ["parse_example_dataset_test.py"],
+    shard_count = 4,
     deps = [
         "//tensorflow/core:protos_all_py",
         "//tensorflow/python:client_testlib",

tensorflow/python/data/experimental/kernel_tests/parse_example_dataset_test.py

@@ -1110,6 +1110,43 @@ class ParseExampleDatasetTest(test_base.DatasetTestBase,
         expected_values=expected_output,
         create_iterator_twice=True)
 
+  @combinations.generate(
+      combinations.times(
+          test_base.default_test_combinations(),
+          combinations.combine(
+              local_determinism=[None, True, False],
+              global_determinism=[True, False])))
+  def testDeterminism(self, local_determinism, global_determinism):
+    num_elements = 1000
+    batches = []
+    for i in range(num_elements):
+      example_i = example(features=features({
+          "a": int64_feature([i]),
+      }))
+      batches.append([example_i.SerializeToString()])
+
+    test_features = {"a": parsing_ops.FixedLenFeature((), dtype=dtypes.int64)}
+    dataset = dataset_ops.Dataset.from_tensor_slices(batches)
+    dataset = dataset.apply(
+        contrib_parsing_ops.parse_example_dataset(
+            test_features,
+            num_parallel_calls=10,
+            deterministic=local_determinism))
+
+    opts = dataset_ops.Options()
+    opts.experimental_deterministic = global_determinism
+    dataset = dataset.with_options(opts)
+
+    expected = list(range(num_elements))
+    actual = [elem["a"][0] for elem in self.getDatasetOutput(dataset)]
+
+    require_order = local_determinism or (local_determinism is None and
+                                          global_determinism)
+    if require_order:
+      self.assertAllEqual(expected, actual)
+    else:
+      self.assertCountEqual(expected, actual)
+
 
 if __name__ == "__main__":
   test.main()

tensorflow/python/data/experimental/ops/parsing_ops.py

@@ -17,6 +17,7 @@ from __future__ import absolute_import
 from __future__ import division
 from __future__ import print_function
 
+from tensorflow.python.compat import compat
 from tensorflow.python.data.ops import dataset_ops
 from tensorflow.python.data.util import structure
 from tensorflow.python.framework import dtypes
@@ -31,13 +32,20 @@ from tensorflow.python.util.tf_export import tf_export
 class _ParseExampleDataset(dataset_ops.UnaryDataset):
   """A `Dataset` that parses `example` dataset into a `dict` dataset."""
 
-  def __init__(self, input_dataset, features, num_parallel_calls):
+  def __init__(self, input_dataset, features, num_parallel_calls,
+               deterministic):
     self._input_dataset = input_dataset
     if not structure.are_compatible(
         input_dataset.element_spec,
         tensor_spec.TensorSpec([None], dtypes.string)):
       raise TypeError("Input dataset should be a dataset of vectors of strings")
     self._num_parallel_calls = num_parallel_calls
+    if deterministic is None:
+      self._deterministic = "default"
+    elif deterministic:
+      self._deterministic = "true"
+    else:
+      self._deterministic = "false"
     # pylint: disable=protected-access
     self._features = parsing_ops._prepend_none_dimension(features)
     # TODO(b/112859642): Pass sparse_index and sparse_values for SparseFeature
@@ -77,19 +85,35 @@ class _ParseExampleDataset(dataset_ops.UnaryDataset):
       self._element_spec[key] = ragged_tensor.RaggedTensorSpec(
           input_dataset_shape.concatenate([None]), value_type, 1, splits_type)
 
-    variant_tensor = (
+    if deterministic is not None or compat.forward_compatible(2020, 3, 6):
-        gen_experimental_dataset_ops.parse_example_dataset(
+      variant_tensor = (
-            self._input_dataset._variant_tensor,  # pylint: disable=protected-access
+          gen_experimental_dataset_ops.parse_example_dataset_v2(
-            self._num_parallel_calls,
+              self._input_dataset._variant_tensor,  # pylint: disable=protected-access
-            self._dense_defaults,
+              self._num_parallel_calls,
-            self._sparse_keys,
+              self._dense_defaults,
-            self._dense_keys,
+              self._sparse_keys,
-            self._sparse_types,
+              self._dense_keys,
-            self._dense_shapes,
+              self._sparse_types,
-            ragged_keys=self._ragged_keys,
+              self._dense_shapes,
-            ragged_value_types=self._ragged_value_types,
+              deterministic=self._deterministic,
-            ragged_split_types=self._ragged_split_types,
+              ragged_keys=self._ragged_keys,
-            **self._flat_structure))
+              ragged_value_types=self._ragged_value_types,
+              ragged_split_types=self._ragged_split_types,
+              **self._flat_structure))
+    else:
+      variant_tensor = (
+          gen_experimental_dataset_ops.parse_example_dataset(
+              self._input_dataset._variant_tensor,  # pylint: disable=protected-access
+              self._num_parallel_calls,
+              self._dense_defaults,
+              self._sparse_keys,
+              self._dense_keys,
+              self._sparse_types,
+              self._dense_shapes,
+              ragged_keys=self._ragged_keys,
+              ragged_value_types=self._ragged_value_types,
+              ragged_split_types=self._ragged_split_types,
+              **self._flat_structure))
     super(_ParseExampleDataset, self).__init__(input_dataset, variant_tensor)
 
   @property
@@ -99,7 +123,7 @@ class _ParseExampleDataset(dataset_ops.UnaryDataset):
 
 # TODO(b/111553342): add arguments names and example names as well.
 @tf_export("data.experimental.parse_example_dataset")
-def parse_example_dataset(features, num_parallel_calls=1):
+def parse_example_dataset(features, num_parallel_calls=1, deterministic=None):
   """A transformation that parses `Example` protos into a `dict` of tensors.
 
   Parses a number of serialized `Example` protos given in `serialized`. We refer
@@ -119,6 +143,13 @@ def parse_example_dataset(features, num_parallel_calls=1):
      `VarLenFeature`, `RaggedFeature`, and `SparseFeature` values.
    num_parallel_calls: (Optional.) A `tf.int32` scalar `tf.Tensor`,
       representing the number of parsing processes to call in parallel.
+   deterministic: (Optional.) A boolean controlling whether determinism
+      should be traded for performance by allowing elements to be produced out
+      of order if some parsing calls complete faster than others. If
+      `deterministic` is `None`, the
+      `tf.data.Options.experimental_deterministic` dataset option (`True` by
+      default) is used to decide whether to produce elements
+      deterministically.
 
   Returns:
     A dataset transformation function, which can be passed to
@@ -132,7 +163,8 @@ def parse_example_dataset(features, num_parallel_calls=1):
 
   def _apply_fn(dataset):
     """Function from `Dataset` to `Dataset` that applies the transformation."""
-    out_dataset = _ParseExampleDataset(dataset, features, num_parallel_calls)
+    out_dataset = _ParseExampleDataset(dataset, features, num_parallel_calls,
+                                       deterministic)
     if any(
         isinstance(feature, parsing_ops.SparseFeature) or
         (isinstance(feature, parsing_ops.RaggedFeature) and feature.partitions)

tensorflow/tools/api/golden/v1/tensorflow.data.experimental.pbtxt

@@ -190,7 +190,7 @@ tf_module {
   }
   member_method {
     name: "parse_example_dataset"
-    argspec: "args=[\'features\', \'num_parallel_calls\'], varargs=None, keywords=None, defaults=[\'1\'], "
+    argspec: "args=[\'features\', \'num_parallel_calls\', \'deterministic\'], varargs=None, keywords=None, defaults=[\'1\', \'None\'], "
   }
   member_method {
     name: "prefetch_to_device"

tensorflow/tools/api/golden/v1/tensorflow.raw_ops.pbtxt

@@ -2660,6 +2660,10 @@ tf_module {
     name: "ParseExampleDataset"
     argspec: "args=[\'input_dataset\', \'num_parallel_calls\', \'dense_defaults\', \'sparse_keys\', \'dense_keys\', \'sparse_types\', \'dense_shapes\', \'output_types\', \'output_shapes\', \'sloppy\', \'ragged_keys\', \'ragged_value_types\', \'ragged_split_types\', \'name\'], varargs=None, keywords=None, defaults=[\'False\', \'[]\', \'[]\', \'[]\', \'None\'], "
   }
+  member_method {
+    name: "ParseExampleDatasetV2"
+    argspec: "args=[\'input_dataset\', \'num_parallel_calls\', \'dense_defaults\', \'sparse_keys\', \'dense_keys\', \'sparse_types\', \'dense_shapes\', \'output_types\', \'output_shapes\', \'deterministic\', \'ragged_keys\', \'ragged_value_types\', \'ragged_split_types\', \'name\'], varargs=None, keywords=None, defaults=[\'default\', \'[]\', \'[]\', \'[]\', \'None\'], "
+  }
   member_method {
     name: "ParseExampleV2"
     argspec: "args=[\'serialized\', \'names\', \'sparse_keys\', \'dense_keys\', \'ragged_keys\', \'dense_defaults\', \'num_sparse\', \'sparse_types\', \'ragged_value_types\', \'ragged_split_types\', \'dense_shapes\', \'name\'], varargs=None, keywords=None, defaults=[\'None\'], "

tensorflow/tools/api/golden/v2/tensorflow.data.experimental.pbtxt

@@ -158,7 +158,7 @@ tf_module {
   }
   member_method {
     name: "parse_example_dataset"
-    argspec: "args=[\'features\', \'num_parallel_calls\'], varargs=None, keywords=None, defaults=[\'1\'], "
+    argspec: "args=[\'features\', \'num_parallel_calls\', \'deterministic\'], varargs=None, keywords=None, defaults=[\'1\', \'None\'], "
   }
   member_method {
     name: "prefetch_to_device"

tensorflow/tools/api/golden/v2/tensorflow.raw_ops.pbtxt

@@ -2660,6 +2660,10 @@ tf_module {
     name: "ParseExampleDataset"
     argspec: "args=[\'input_dataset\', \'num_parallel_calls\', \'dense_defaults\', \'sparse_keys\', \'dense_keys\', \'sparse_types\', \'dense_shapes\', \'output_types\', \'output_shapes\', \'sloppy\', \'ragged_keys\', \'ragged_value_types\', \'ragged_split_types\', \'name\'], varargs=None, keywords=None, defaults=[\'False\', \'[]\', \'[]\', \'[]\', \'None\'], "
   }
+  member_method {
+    name: "ParseExampleDatasetV2"
+    argspec: "args=[\'input_dataset\', \'num_parallel_calls\', \'dense_defaults\', \'sparse_keys\', \'dense_keys\', \'sparse_types\', \'dense_shapes\', \'output_types\', \'output_shapes\', \'deterministic\', \'ragged_keys\', \'ragged_value_types\', \'ragged_split_types\', \'name\'], varargs=None, keywords=None, defaults=[\'default\', \'[]\', \'[]\', \'[]\', \'None\'], "
+  }
   member_method {
     name: "ParseExampleV2"
     argspec: "args=[\'serialized\', \'names\', \'sparse_keys\', \'dense_keys\', \'ragged_keys\', \'dense_defaults\', \'num_sparse\', \'sparse_types\', \'ragged_value_types\', \'ragged_split_types\', \'dense_shapes\', \'name\'], varargs=None, keywords=None, defaults=[\'None\'], "