C++ API: run shape inference as nodes are constructed

Here's an example of the new generated code:

AddN::AddN(const ::tensorflow::Scope& scope, ::tensorflow::InputList inputs) {
  if (!scope.ok()) return;
  auto _inputs = ::tensorflow::ops::AsNodeOutList(scope, inputs);
  if (!scope.ok()) return;
  ::tensorflow::Node* ret;
  const auto unique_name = scope.GetUniqueNameForOp("AddN");
  auto builder = ::tensorflow::NodeBuilder(unique_name, "AddN")
                     .Input(_inputs)
  ;
  scope.UpdateBuilder(&builder);
  scope.UpdateStatus(builder.Finalize(scope.graph(), &ret));
  if (!scope.ok()) return;
  scope.UpdateStatus(scope.DoShapeInference(ret));
  this->sum = Output(ret, 0);
}

Enabling shape inference unfortunately broke many tests. I fixed some of them, but for others I introduced a Scope::DisabledShapeInferenceScope() static method that returns a scope that doesn't perform shape inference. Eventually we should fix the tests that use this and remove it.

PiperOrigin-RevId: 165378429

tensorflow/cc/framework/cc_op_gen.cc

@@ -812,12 +812,8 @@ string OpInfo::GetConstructorBody() const {
   strings::StrAppend(&body, "  ", scope_str, ".UpdateStatus(builder.Finalize(",
                      scope_str, ".graph(), &ret));\n");
   strings::StrAppend(&body, "  ", return_on_error, "\n");
-
+  strings::StrAppend(&body, "  ", scope_str, ".UpdateStatus(", scope_str,
-  // TODO(b/28152992): Enable this code-path once we have converted
+                     ".DoShapeInference(ret));\n");
-  // all python shape functions to call their C++ versions.
-
-  // strings::StrAppend(&body, "  ", scope_str, ".UpdateStatus(", scope_str,
-  //                    ".refiner()->AddNode(ret));\n");
 
   GetOutput(&body);
   return body;

tensorflow/cc/framework/scope.cc

@@ -37,13 +37,14 @@ Scope& Scope::operator=(const Scope& other) {
 }
 
 Scope::Impl::Impl(Graph* graph, Status* status, NameMap* name_map,
-                  ShapeRefiner* refiner)
+                  ShapeRefiner* refiner, bool disable_shape_inference)
     : graph_(graph),
       status_(status),
       name_map_(name_map),
       refiner_(refiner),
       scope_used_(nullptr),
-      colocation_constraints_() {}
+      colocation_constraints_(),
+      disable_shape_inference_(disable_shape_inference) {}
 
 Scope::Impl::Impl(const std::shared_ptr<Graph>& graph,
                   const std::shared_ptr<Status>& status,
@@ -54,13 +55,23 @@ Scope::Impl::Impl(const std::shared_ptr<Graph>& graph,
       name_map_(name_map),
       refiner_(refiner),
       scope_used_(nullptr),
-      colocation_constraints_() {}
+      colocation_constraints_(),
+      disable_shape_inference_(false) {}
 
 Scope Scope::NewRootScope() {
   Graph* graph = new Graph(OpRegistry::Global());
   ShapeRefiner* refiner =
       new ShapeRefiner(graph->versions(), graph->op_registry());
-  return Scope(new Impl(graph, new Status, new Impl::NameMap, refiner));
+  return Scope(new Impl(graph, new Status, new Impl::NameMap, refiner,
+                        /* disable_shape_inference */ false));
+}
+
+Scope Scope::DisabledShapeInferenceScope() {
+  Graph* graph = new Graph(OpRegistry::Global());
+  ShapeRefiner* refiner =
+      new ShapeRefiner(graph->versions(), graph->op_registry());
+  return Scope(new Impl(graph, new Status, new Impl::NameMap, refiner,
+                        /* disable_shape_inference */ true));
 }
 
 Scope::Impl::Impl(const Scope& other, Tags::ScopeName, const string& name,
@@ -77,7 +88,8 @@ Scope::Impl::Impl(const Scope& other, Tags::ScopeName, const string& name,
       exit_on_error_(other.impl()->exit_on_error_),
       kernel_label_(other.impl()->kernel_label_),
       device_(other.impl()->device_),
-      colocation_constraints_(other.impl()->colocation_constraints_) {}
+      colocation_constraints_(other.impl()->colocation_constraints_),
+      disable_shape_inference_(other.impl()->disable_shape_inference_) {}
 
 Scope::Impl::Impl(const Scope& other, Tags::OpName, const string& name,
                   const string& op_name)
@@ -92,7 +104,8 @@ Scope::Impl::Impl(const Scope& other, Tags::OpName, const string& name,
       exit_on_error_(other.impl()->exit_on_error_),
       kernel_label_(other.impl()->kernel_label_),
       device_(other.impl()->device_),
-      colocation_constraints_(other.impl()->colocation_constraints_) {}
+      colocation_constraints_(other.impl()->colocation_constraints_),
+      disable_shape_inference_(other.impl()->disable_shape_inference_) {}
 
 Scope::Impl::Impl(const Scope& other, Tags::ControlDeps,
                   std::vector<Operation> control_deps, bool clear_control_deps)
@@ -113,7 +126,8 @@ Scope::Impl::Impl(const Scope& other, Tags::ControlDeps,
       exit_on_error_(other.impl()->exit_on_error_),
       kernel_label_(other.impl()->kernel_label_),
       device_(other.impl()->device_),
-      colocation_constraints_(other.impl()->colocation_constraints_) {}
+      colocation_constraints_(other.impl()->colocation_constraints_),
+      disable_shape_inference_(other.impl()->disable_shape_inference_) {}
 
 Scope::Impl::Impl(const Scope& other, Tags::Device, const string& device)
     : graph_(other.impl()->graph_),
@@ -127,7 +141,8 @@ Scope::Impl::Impl(const Scope& other, Tags::Device, const string& device)
       exit_on_error_(other.impl()->exit_on_error_),
       kernel_label_(other.impl()->kernel_label_),
       device_(device),
-      colocation_constraints_(other.impl()->colocation_constraints_) {}
+      colocation_constraints_(other.impl()->colocation_constraints_),
+      disable_shape_inference_(other.impl()->disable_shape_inference_) {}
 
 Scope::Impl::Impl(const Scope& other, Tags::SingleUseScope,
                   const string& op_name)
@@ -142,7 +157,8 @@ Scope::Impl::Impl(const Scope& other, Tags::SingleUseScope,
       exit_on_error_(other.impl()->exit_on_error_),
       kernel_label_(other.impl()->kernel_label_),
       device_(other.impl()->device_),
-      colocation_constraints_(other.impl()->colocation_constraints_) {}
+      colocation_constraints_(other.impl()->colocation_constraints_),
+      disable_shape_inference_(other.impl()->disable_shape_inference_) {}
 
 Scope::Impl::Impl(const Scope& other, Tags::ExitOnError)
     : graph_(other.impl()->graph_),
@@ -156,7 +172,8 @@ Scope::Impl::Impl(const Scope& other, Tags::ExitOnError)
       exit_on_error_(true),
       kernel_label_(other.impl()->kernel_label_),
       device_(other.impl()->device_),
-      colocation_constraints_(other.impl()->colocation_constraints_) {}
+      colocation_constraints_(other.impl()->colocation_constraints_),
+      disable_shape_inference_(other.impl()->disable_shape_inference_) {}
 
 Scope::Impl::Impl(const Scope& other, Tags::KernelLabel,
                   const string& kernel_label)
@@ -171,7 +188,8 @@ Scope::Impl::Impl(const Scope& other, Tags::KernelLabel,
       exit_on_error_(other.impl()->exit_on_error_),
       kernel_label_(kernel_label),
       device_(other.impl()->device_),
-      colocation_constraints_(other.impl()->colocation_constraints_) {}
+      colocation_constraints_(other.impl()->colocation_constraints_),
+      disable_shape_inference_(other.impl()->disable_shape_inference_) {}
 
 Scope::Impl::Impl(const Scope& other, Tags::Colocate,
                   const Operation& colocate_with_op, bool clear_colocations)
@@ -189,7 +207,8 @@ Scope::Impl::Impl(const Scope& other, Tags::Colocate,
       colocation_constraints_(
           clear_colocations
               ? std::unordered_set<string>()
-              : other.impl()->GetColocationConstraints(colocate_with_op)) {}
+              : other.impl()->GetColocationConstraints(colocate_with_op)),
+      disable_shape_inference_(other.impl()->disable_shape_inference_) {}
 
 std::unordered_set<string> Scope::Impl::GetColocationConstraints(
     const Operation& colocate_with_op) const {
@@ -404,6 +423,11 @@ CompositeOpScopes Scope::GetCompositeOpScopes(
   }
 }
 
+Status Scope::DoShapeInference(Node* node) const {
+  if (impl_->disable_shape_inference_) return Status::OK();
+  return impl_->refiner_->AddNode(node);
+}
+
 class InternalScope {
  public:
   // NewScope doesn't take ownership of the inputs.

tensorflow/cc/framework/scope.h

@@ -199,6 +199,18 @@ class Scope {
   // edges from the source and to the sink node, resolves back edges
   // by name), and makes sure the resulting graph is valid.
   Status ToGraph(Graph* g) const;
+
+  // Calls AddNode() using this scope's ShapeRefiner. This exists in the public
+  // API to prevent custom op wrappers from needing access to shape_refiner.h or
+  // scope_internal.h.
+  // TODO(skyewm): remove this from public API
+  Status DoShapeInference(Node* node) const;
+
+  // Creates a new root scope that causes all DoShapeInference() calls to return
+  // Status::OK() (on the returned scope and any subscopes). Used for testing.
+  // TODO(skyewm): fix tests that still require this and eventually remove, or
+  // at least remove from public API
+  static Scope DisabledShapeInferenceScope();
   // END_SKIP_DOXYGEN
 
   const std::vector<Operation>& control_deps() const;

tensorflow/cc/framework/scope_internal.h

@@ -58,7 +58,8 @@ class Scope::Impl {
     enum class Colocate;
   };
 
-  Impl(Graph* graph, Status* status, NameMap* name_map, ShapeRefiner* refiner);
+  Impl(Graph* graph, Status* status, NameMap* name_map, ShapeRefiner* refiner,
+       bool disable_shape_inference);
   Impl(const Scope& other, Tags::ScopeName, const string& name,
        bool copy_names);
   Impl(const Scope& other, Tags::OpName, const string& name,
@@ -103,6 +104,10 @@ class Scope::Impl {
   const string kernel_label_ = "";
   const string device_ = "";
   const std::unordered_set<string> colocation_constraints_;
+
+  // If true, Scope::DoShapeInference() always returns Status:OK().
+  // TODO(skyewm): remove this when possible
+  const bool disable_shape_inference_;
 };
 
 }  // namespace tensorflow

tensorflow/cc/framework/test_op.cc

@@ -13,6 +13,7 @@ See the License for the specific language governing permissions and
 limitations under the License.
 ==============================================================================*/
 
+#include "tensorflow/core/framework/common_shape_fns.h"
 #include "tensorflow/core/framework/op.h"
 
 namespace tensorflow {
@@ -24,6 +25,7 @@ REGISTER_OP("ThrowAway1")
     .Attr("scope: int")
     .Attr("builder: int = 1")
     .Attr("while: int")
+    .SetShapeFn(shape_inference::UnknownShape)
     .Doc(R"doc(
 Op to test keywords and reserved words in input and attr names.
 
@@ -36,12 +38,20 @@ REGISTER_OP("ThrowAway2")
     .Attr("scope: int = 2")
     .Attr("throw_away2: int = 2")
     .Attr("attrs: int = 4")
-    .Attr("node: int = 4");
+    .Attr("node: int = 4")
+    .SetShapeFn(shape_inference::UnknownShape);
 
-REGISTER_OP("ThrowAway3").Output("node: int32");
+REGISTER_OP("ThrowAway3")
+    .Output("node: int32")
+    .SetShapeFn(shape_inference::UnknownShape);
 
-REGISTER_OP("ThrowAway4").Input("node: int32");
+REGISTER_OP("ThrowAway4")
+    .Input("node: int32")
+    .SetShapeFn(shape_inference::UnknownShape);
 
-REGISTER_OP("ThrowAway5").Output("foo: int32").Attr("node: int = 4");
+REGISTER_OP("ThrowAway5")
+    .Output("foo: int32")
+    .Attr("node: int = 4")
+    .SetShapeFn(shape_inference::UnknownShape);
 
 }  // namespace tensorflow

tensorflow/cc/ops/const_op.cc

@@ -34,7 +34,9 @@ Output Const(const Scope& scope, const Input::Initializer& val) {
                      .Attr("dtype", val.tensor.dtype());
   scope.UpdateBuilder(&builder);
   scope.UpdateStatus(builder.Finalize(graph, &ret));
+  if (!scope.ok()) return Output();
 
+  scope.UpdateStatus(scope.DoShapeInference(ret));
   if (!scope.ok()) return Output();
 
   return Output(ret);

tensorflow/cc/ops/const_op.h

@@ -56,6 +56,8 @@ Output Const(const Scope& scope, const Input::Initializer& val) {
   scope.UpdateBuilder(&cast_builder);
   Node* ret;
   scope.UpdateStatus(cast_builder.Finalize(scope.graph(), &ret));
+  if (!scope.ok()) return Output();
+  scope.UpdateStatus(scope.DoShapeInference(ret));
   return Output(ret, 0);
 }
 

tensorflow/compiler/tf2xla/ops/functional_ops.cc

@@ -13,6 +13,7 @@ See the License for the specific language governing permissions and
 limitations under the License.
 ==============================================================================*/
 
+#include "tensorflow/core/framework/common_shape_fns.h"
 #include "tensorflow/core/framework/op.h"
 
 namespace tensorflow {
@@ -26,6 +27,7 @@ REGISTER_OP("XlaWhile")
     .Attr("cond: func")
     .Attr("body: func")
     .SetIsStateful()
+    .SetShapeFn(shape_inference::UnknownShape)
     .Doc(R"doc(
 output = input; While (Cond(output)) { output = Body(output) }
 

tensorflow/compiler/tf2xla/ops/sendrecv_ops.cc

@@ -13,6 +13,7 @@ See the License for the specific language governing permissions and
 limitations under the License.
 ==============================================================================*/
 
+#include "tensorflow/core/framework/common_shape_fns.h"
 #include "tensorflow/core/framework/op.h"
 
 namespace tensorflow {
@@ -22,6 +23,7 @@ REGISTER_OP("_XLASend")
     .Attr("T: type")
     .Attr("tensor_name: string")
     .SetIsStateful()
+    .SetShapeFn(shape_inference::UnknownShape)
     .Doc(R"doc(
 Sends the named tensor to another XLA computation.
 
@@ -35,6 +37,7 @@ REGISTER_OP("_XLARecv")
     .Attr("tensor_name: string")
     .Attr("shape: shape")
     .SetIsStateful()
+    .SetShapeFn(shape_inference::UnknownShape)
     .Doc(R"doc(
 Receives the named tensor from another XLA computation.
 

tensorflow/compiler/tf2xla/xla_compiler_test.cc

@@ -25,6 +25,7 @@ limitations under the License.
 #include "tensorflow/compiler/xla/shape_util.h"
 #include "tensorflow/compiler/xla/tests/literal_test_util.h"
 #include "tensorflow/core/common_runtime/function.h"
+#include "tensorflow/core/framework/common_shape_fns.h"
 #include "tensorflow/core/framework/resource_mgr.h"
 #include "tensorflow/core/framework/tensor_testutil.h"
 #include "tensorflow/core/graph/graph.h"
@@ -76,6 +77,8 @@ class DummyReadResourceCC {
     scope.UpdateBuilder(&builder);
     scope.UpdateStatus(builder.Finalize(scope.graph(), &ret));
     if (!scope.ok()) return;
+    scope.UpdateStatus(scope.DoShapeInference(ret));
+    if (!scope.ok()) return;
     this->output_ = Output(ret, 0);
   }
   Node* node() const { return output_.node(); }
@@ -86,6 +89,7 @@ class DummyReadResourceCC {
 REGISTER_OP("DummyReadResource")
     .Input("input: int32")
     .Output("output: int32")
+    .SetShapeFn(shape_inference::UnknownShape)
     .Doc(R"doc(
 A dummy Op.
 

tensorflow/contrib/tpu/ops/replication_ops.cc

@@ -13,6 +13,7 @@ See the License for the specific language governing permissions and
 limitations under the License.
 ==============================================================================*/
 
+#include "tensorflow/core/framework/common_shape_fns.h"
 #include "tensorflow/core/framework/op.h"
 #include "tensorflow/core/framework/shape_inference.h"
 
@@ -67,6 +68,7 @@ REGISTER_OP("TPUReplicate")
     .Input("broadcast_inputs: Tbroadcast_inputs")
     .Input("variables: NumVariables * resource")
     .Output("outputs: output_types")
+    .SetShapeFn(shape_inference::UnknownShape)
     .Doc(R"doc(
 Runs replicated computations on a distributed TPU system.
 

tensorflow/contrib/tpu/ops/tpu_configuration_ops.cc

@@ -13,6 +13,7 @@ See the License for the specific language governing permissions and
 limitations under the License.
 ==============================================================================*/
 
+#include "tensorflow/core/framework/common_shape_fns.h"
 #include "tensorflow/core/framework/op.h"
 #include "tensorflow/core/framework/shape_inference.h"
 #include "tensorflow/core/lib/core/status.h"
@@ -154,7 +155,10 @@ global_tpu_array: A two-dimensional array. For each host (the outer
 dimension) the array lists the global ids of the TPUs on that host.
 )doc");
 
-REGISTER_OP("_ShutdownDistributedTPU").SetIsStateful().Doc(R"doc(
+REGISTER_OP("_ShutdownDistributedTPU")
+    .SetIsStateful()
+    .SetShapeFn(shape_inference::UnknownShape)
+    .Doc(R"doc(
 An op that shuts down a running distributed TPU system. The Op returns
 an error if no system is running. This Op must be run on the same
 TPU_SYSTEM device as the corresponding _ConfigureDistributedTPU was run
@@ -184,6 +188,7 @@ tpu_ids: A vector containing the global TPU id of each TPU on the host.
 REGISTER_OP("_DisconnectHostFromDistributedTPUSystem")
     .Output("number_of_tpu_chips: int32")
     .SetIsStateful()
+    .SetShapeFn(shape_inference::UnknownShape)
     .Doc(R"doc(
 An op that disconnects the TPUs on a host from a running distributed
 TPU system.
@@ -196,6 +201,7 @@ REGISTER_OP("ConfigureDistributedTPU")
     .Output("global_tpu_array: int32")
     .Attr("embedding_config: string = ''")
     .SetIsStateful()
+    .SetShapeFn(shape_inference::UnknownShape)
     .Doc(R"doc(
 An op that sets up the centralized structures for a distributed TPU
 system.
@@ -205,7 +211,10 @@ dimension) the array lists the global ids of the TPUs on that host.
 embedding_config: Internal use.
 )doc");
 
-REGISTER_OP("ShutdownDistributedTPU").SetIsStateful().Doc(R"doc(
+REGISTER_OP("ShutdownDistributedTPU")
+    .SetIsStateful()
+    .SetShapeFn(shape_inference::UnknownShape)
+    .Doc(R"doc(
 An op that shuts down a running distributed TPU system. The Op returns
 an error if no system is running.
 )doc");

tensorflow/core/common_runtime/constant_folding_test.cc

@@ -282,6 +282,7 @@ TEST_F(ConstantFoldingTest, TestNoReplaceFunctionCall) {
     Status status;
     Node* times_two = s.graph()->AddNode(def, &status);
     TF_ASSERT_OK(status);
+    TF_ASSERT_OK(s.DoShapeInference(times_two));
     s.graph()->AddEdge(c.node(), 0, times_two, 0);
 
     auto times_two_send =
@@ -297,7 +298,10 @@ TEST_F(ConstantFoldingTest, TestNoReplaceFunctionCall) {
   EXPECT_FALSE(was_mutated);
 }
 
-REGISTER_OP("ConstantFoldingTestOp").Input("a: int64").Output("b: int64");
+REGISTER_OP("ConstantFoldingTestOp")
+    .Input("a: int64")
+    .Output("b: int64")
+    .SetShapeFn(shape_inference::UnknownShape);
 
 TEST_F(ConstantFoldingTest, TestNoReplaceNonCPUOp) {
   Graph g(OpRegistry::Global());
@@ -312,6 +316,7 @@ TEST_F(ConstantFoldingTest, TestNoReplaceNonCPUOp) {
     Status status;
     Node* non_cpu = s.graph()->AddNode(def, &status);
     TF_ASSERT_OK(status);
+    TF_ASSERT_OK(s.DoShapeInference(non_cpu));
 
     auto non_cpu_send =
         ops::_Send(s.WithOpName("non_cpu_send"), Output(non_cpu),

tensorflow/core/common_runtime/function_test.cc

@@ -284,6 +284,7 @@ Output Call(Scope* scope, const string& op_name, const string& fn_name,
   Status status;
   Node* n = scope->graph()->AddNode(def, &status);
   TF_CHECK_OK(status);
+  TF_CHECK_OK(scope->DoShapeInference(n));
   for (int i = 0; i < inputs.size(); ++i) {
     scope->graph()->AddEdge(inputs[i].node(), inputs[i].index(), n, i);
   }
@@ -989,7 +990,7 @@ TEST(OptimizationTest, RemoveDeadNodes) {
 
   GraphDef expected;
   {
-    Scope s = Scope::NewRootScope();
+    Scope s = Scope::DisabledShapeInferenceScope();
     auto x = ops::_Arg(s.WithOpName("x"), DT_INT32, 0);
     auto o = ops::Const(s.WithOpName("o"), 1);
     auto keep_me = ops::RandomUniform(s.WithOpName("keep_me"), {o}, DT_FLOAT);
@@ -1070,7 +1071,7 @@ TEST(OptimizationTest, RemoveIdentityNodes) {
        {{"y"}, "Add", {"a", "o"}, {{"T", T}}}});
 
   {
-    Scope s = Scope::NewRootScope();
+    Scope s = Scope::DisabledShapeInferenceScope();
     auto x = ops::_Arg(s.WithOpName("x"), DT_INT32, 0);
     auto o = ops::Const(s.WithOpName("o"), 1);
     auto a = ops::Square(s.WithOpName("a"), x);
@@ -1087,7 +1088,7 @@ TEST(OptimizationTest, RemoveIdentityNodes) {
   }
 
   {
-    Scope s = Scope::NewRootScope();
+    Scope s = Scope::DisabledShapeInferenceScope();
     auto x = ops::_Arg(s.WithOpName("x"), DT_INT32, 0);
     auto o = ops::Const(s.WithOpName("o"), 1);
     auto a = ops::Square(s.WithOpName("a"), x);
@@ -1137,7 +1138,7 @@ TEST(OptimizationTest, RemoveListArrayConverter) {
       {{"o", "o:sum"}});
 
   {
-    Scope scope = Scope::NewRootScope();
+    Scope scope = Scope::DisabledShapeInferenceScope();
     auto i = ops::_Arg(scope.WithOpName("i"), DT_FLOAT, 0);
     auto zero = ops::Const(scope.WithOpName("zero"), 0);
     auto s = ops::Split(scope.WithOpName("s"), zero, i, 4);
@@ -1222,7 +1223,7 @@ TEST(OptimizationTest, RemoveListArrayConverter_WithContolDeps) {
       {{"o", "o:sum"}});
 
   {
-    Scope s = Scope::NewRootScope();
+    Scope s = Scope::DisabledShapeInferenceScope();
     auto i = ops::_Arg(s.WithOpName("i"), DT_FLOAT, 0);
     auto dummy = ops::Const(s.WithOpName("dummy"), 0);
     auto x = ops::_ListToArray(s.WithOpName("x").WithControlDependencies(dummy),

tensorflow/core/common_runtime/gpu/gpu_stream_util_test.cc

@@ -64,7 +64,7 @@ TEST_F(GpuStreamUtilTest, EmptyGraph) {
 }
 
 TEST_F(GpuStreamUtilTest, SimpleGraphOneStream) {
-  auto root = Scope::NewRootScope().ExitOnError();
+  auto root = Scope::DisabledShapeInferenceScope().ExitOnError();
   ops::MatMul(root, {}, {});
   Graph g(OpRegistry::Global());
   TF_ASSERT_OK(root.ToGraph(&g));
@@ -83,7 +83,7 @@ TEST_F(GpuStreamUtilTest, SimpleGraphOneStream) {
 }
 
 TEST_F(GpuStreamUtilTest, SimpleGraphManyStreams) {
-  auto root = Scope::NewRootScope().ExitOnError();
+  auto root = Scope::DisabledShapeInferenceScope().ExitOnError();
   ops::MatMul(root, {}, {});
   Graph g(OpRegistry::Global());
   TF_ASSERT_OK(root.ToGraph(&g));
@@ -104,7 +104,7 @@ TEST_F(GpuStreamUtilTest, SimpleGraphManyStreams) {
 }
 
 TEST_F(GpuStreamUtilTest, StreamOverrides) {
-  auto root = Scope::NewRootScope().ExitOnError();
+  auto root = Scope::DisabledShapeInferenceScope().ExitOnError();
   ops::_Recv(root.WithOpName("input"), DT_FLOAT, "input", "/cpu:0", 0,
              "/device:GPU:0");
   Output n = ops::MatMul(root, {}, {});

tensorflow/core/common_runtime/shape_refiner_test.cc

@@ -882,7 +882,7 @@ TEST_F(ShapeRefinerTest, ConstantValueAsShape_Shape) {
 }
 
 TEST_F(ShapeRefinerTest, ConstantValueAsShape_PackInt32) {
-  Scope root = Scope::NewRootScope();
+  Scope root = Scope::DisabledShapeInferenceScope();
   Node* scalar_non_const;
   TF_ASSERT_OK(NodeBuilder("in", "NonConstScalarInt32")
                    .Finalize(root.graph(), &scalar_non_const));
@@ -914,7 +914,7 @@ TEST_F(ShapeRefinerTest, ConstantValueAsShape_PackInt32) {
 }
 
 TEST_F(ShapeRefinerTest, ConstantValueAsShape_PackInt64) {
-  Scope root = Scope::NewRootScope();
+  Scope root = Scope::DisabledShapeInferenceScope();
   Node* scalar_non_const;
   TF_ASSERT_OK(NodeBuilder("in", "NonConstScalarInt64")
                    .Finalize(root.graph(), &scalar_non_const));
@@ -997,7 +997,7 @@ TEST_F(ShapeRefinerTest, ConstantValueAsShape_PackInvalidInput) {
 }
 
 TEST_F(ShapeRefinerTest, ConstantValueAsShape_Concat) {
-  Scope root = Scope::NewRootScope();
+  Scope root = Scope::DisabledShapeInferenceScope();
   Graph* g = root.graph();
   Node* partial_1;
   Node* partial_2;
@@ -1034,7 +1034,7 @@ TEST_F(ShapeRefinerTest, ConstantValueAsShape_Concat) {
 }
 
 TEST_F(ShapeRefinerTest, ConstantValueAsShape_ConcatWithUnknown) {
-  Scope root = Scope::NewRootScope();
+  Scope root = Scope::DisabledShapeInferenceScope();
   Graph* g = root.graph();
   Node* scalar_non_const;
   TF_ASSERT_OK(NodeBuilder("in", "NonConstScalarInt32")
@@ -1077,7 +1077,7 @@ TEST_F(ShapeRefinerTest, ConstantValueAsShape_ConcatWithUnknown) {
 }
 
 TEST_F(ShapeRefinerTest, ConstantValueAsShape_ConcatInvalidDimValue) {
-  Scope root = Scope::NewRootScope();
+  Scope root = Scope::DisabledShapeInferenceScope();
   Graph* g = root.graph();
   Node* scalar_non_const;
   TF_ASSERT_OK(NodeBuilder("in", "NonConstScalarInt32")

tensorflow/core/graph/graph_partition_test.cc

@@ -25,6 +25,7 @@ limitations under the License.
 #include "tensorflow/cc/ops/math_ops.h"
 #include "tensorflow/cc/ops/random_ops.h"
 #include "tensorflow/cc/ops/sendrecv_ops.h"
+#include "tensorflow/core/framework/common_shape_fns.h"
 #include "tensorflow/core/framework/function_testlib.h"
 #include "tensorflow/core/framework/op.h"
 #include "tensorflow/core/framework/versions.pb.h"
@@ -141,9 +142,17 @@ void CheckLoopConstruction(const GraphDef& graph_def) {
   }
 }
 
-REGISTER_OP("FloatInput").Output("o: float");
+REGISTER_OP("FloatInput")
-REGISTER_OP("BoolInput").Output("o: bool");
+    .Output("o: float")
-REGISTER_OP("Combine").Input("a: float").Input("b: float").Output("o: float");
+    .SetShapeFn(shape_inference::UnknownShape);
+REGISTER_OP("BoolInput")
+    .Output("o: bool")
+    .SetShapeFn(shape_inference::UnknownShape);
+REGISTER_OP("Combine")
+    .Input("a: float")
+    .Input("b: float")
+    .Output("o: float")
+    .SetShapeFn(shape_inference::UnknownShape);
 
 Output ConstructOp(const Scope& scope, const string& op_type,
                    const gtl::ArraySlice<Input>& inputs) {
@@ -158,6 +167,8 @@ Output ConstructOp(const Scope& scope, const string& op_type,
   Node* ret;
   scope.UpdateStatus(builder.Finalize(scope.graph(), &ret));
   if (!scope.ok()) return Output();
+  scope.UpdateStatus(scope.DoShapeInference(ret));
+  if (!scope.ok()) return Output();
   return Output(ret);
 }
 

tensorflow/core/grappler/optimizers/auto_parallel_test.cc

@@ -28,7 +28,7 @@ namespace {
 class AutoParallelTest : public ::testing::Test {};
 
 TEST_F(AutoParallelTest, SimpleParallel) {
-  tensorflow::Scope s = tensorflow::Scope::NewRootScope();
+  tensorflow::Scope s = tensorflow::Scope::DisabledShapeInferenceScope();
   Output constant_a = ops::Const(s.WithOpName("constant_a"), 1.0f, {1});
   Output constant_b = ops::Const(s.WithOpName("constant_b"), 1, {1});
   Output var = ops::Variable(s.WithOpName("var"), {1}, DT_FLOAT);

tensorflow/core/kernels/encode_wav_op_test.cc

@@ -35,7 +35,7 @@ namespace tensorflow {
 using namespace ops;  // NOLINT(build/namespaces)
 
 TEST(EncodeWavOpTest, EncodeWavTest) {
-  Scope root = Scope::NewRootScope();
+  Scope root = Scope::DisabledShapeInferenceScope();
 
   Tensor audio_tensor(DT_FLOAT, {4, 2});
   test::FillValues<float>(

tensorflow/core/kernels/fuzzing/fuzz_session.h

@@ -88,7 +88,7 @@ class FuzzSession {
     }
     initialized_ = true;
 
-    Scope root = Scope::NewRootScope().ExitOnError();
+    Scope root = Scope::DisabledShapeInferenceScope().ExitOnError();
     SessionOptions options;
     session_ = std::unique_ptr<Session>(NewSession(options));
 

tensorflow/core/kernels/immutable_constant_op_test.cc

@@ -121,7 +121,7 @@ TEST(ImmutableConstantOpTest, ExecutionError) {
   const TensorShape kBadTensorShape({40, 100});
   const TensorShape kTestTensorShapeT({1, 4});
 
-  auto root = Scope::NewRootScope().ExitOnError();
+  auto root = Scope::DisabledShapeInferenceScope().ExitOnError();
   auto node1 =
       ops::ImmutableConst(root, DT_FLOAT, kBadTensorShape, "test:///2");
   auto node2 =

tensorflow/core/kernels/mfcc_op_test.cc

@@ -35,7 +35,7 @@ namespace tensorflow {
 using namespace ops;  // NOLINT(build/namespaces)
 
 TEST(MfccOpTest, SimpleTest) {
-  Scope root = Scope::NewRootScope();
+  Scope root = Scope::DisabledShapeInferenceScope();
 
   Tensor spectrogram_tensor(DT_FLOAT, TensorShape({1, 1, 513}));
   test::FillIota<float>(&spectrogram_tensor, 1.0f);

tensorflow/core/ops/sendrecv_ops.cc

@@ -13,6 +13,7 @@ See the License for the specific language governing permissions and
 limitations under the License.
 ==============================================================================*/
 
+#include "tensorflow/core/framework/common_shape_fns.h"
 #include "tensorflow/core/framework/op.h"
 
 namespace tensorflow {
@@ -26,6 +27,7 @@ REGISTER_OP("_Send")
     .Attr("recv_device: string")
     .Attr("client_terminated: bool = false")
     .SetIsStateful()
+    .SetShapeFn(shape_inference::UnknownShape)
     .Doc(R"doc(
 Sends the named tensor from send_device to recv_device.
 
@@ -49,6 +51,7 @@ REGISTER_OP("_Recv")
     .Attr("recv_device: string")
     .Attr("client_terminated: bool = false")
     .SetIsStateful()
+    .SetShapeFn(shape_inference::UnknownShape)
     .Doc(R"doc(
 Receives the named tensor from send_device on recv_device.
 
@@ -72,6 +75,7 @@ REGISTER_OP("_HostSend")
     .Attr("recv_device: string")
     .Attr("client_terminated: bool = false")
     .SetIsStateful()
+    .SetShapeFn(shape_inference::UnknownShape)
     .Doc(R"doc(
 Sends the named tensor from send_device to recv_device.
 
@@ -98,6 +102,7 @@ REGISTER_OP("_HostRecv")
     .Attr("recv_device: string")
     .Attr("client_terminated: bool = false")
     .SetIsStateful()
+    .SetShapeFn(shape_inference::UnknownShape)
     .Doc(R"doc(
 Receives the named tensor from send_device on recv_device.
 

tensorflow/tools/graph_transforms/fake_quantize_training_test.cc

@@ -36,7 +36,7 @@ class FakeQuantizeTrainingTest : public ::testing::Test {};
 // TODO(suharshs): Once we implement the fake_quantize_training transform
 // using the GTT, write proper tests of the transform here.
 TEST_F(FakeQuantizeTrainingTest, TransformOccurred) {
-  auto root = tensorflow::Scope::NewRootScope();
+  auto root = tensorflow::Scope::DisabledShapeInferenceScope();
   using namespace ::tensorflow::ops;  // NOLINT(build/namespaces)
 
   Tensor a_data(DT_FLOAT, TensorShape());

tensorflow/tools/graph_transforms/quantize_weights_test.cc

@@ -40,7 +40,7 @@ class QuantizeWeightsTest : public ::testing::Test {
                      const TensorShape& weight_shape,
                      std::initializer_list<float> weight_values,
                      GraphDef* original_graph_def) {
-    auto root = tensorflow::Scope::NewRootScope();
+    auto root = tensorflow::Scope::DisabledShapeInferenceScope();
 
     Tensor input_data(DT_FLOAT, input_shape);
     test::FillValues<float>(&input_data, input_values);

tensorflow/tools/graph_transforms/transform_utils_test.cc

@@ -622,7 +622,7 @@ class TransformUtilsTest : public ::testing::Test {
   }
 
   void TestRenameNodeInputsWithWildcard() {
-    auto root = tensorflow::Scope::NewRootScope();
+    auto root = tensorflow::Scope::DisabledShapeInferenceScope();
     using namespace ::tensorflow::ops;  // NOLINT(build/namespaces)
 
     const int width = 10;