!1694 Check the input size of BatchNorm before fission in bert

Merge pull request !1694 from YuJianfeng/master

mindspore/ccsrc/pre_activate/ascend/ir_fission/batch_norm_bert_fission.cc

@@ -149,8 +149,17 @@ const BaseRef BatchNormBertFission::DefinePattern() const {
 const AnfNodePtr BatchNormBertFission::Process(const FuncGraphPtr &func_graph, const AnfNodePtr &node,
                                                const EquivPtr &) const {
   MS_EXCEPTION_IF_NULL(func_graph);
+  MS_EXCEPTION_IF_NULL(node);
   std::vector<AnfNodePtr> bn_outputs;
   if (!GetBatchNormOutputs(func_graph, node, &bn_outputs)) {
+    MS_LOG(INFO) << "The BatchNorm node should only have output 0, 3 and 4. The node should not be changed";
+    return nullptr;
+  }
+  auto cnode = node->cast<CNodePtr>();
+  MS_EXCEPTION_IF_NULL(cnode);
+  if (cnode->inputs().size() != kBatchNormRealInputNum + 1) {
+    MS_LOG(INFO) << "The input size of BatchNorm should be " << kBatchNormRealInputNum
+                 << ". The node should not be changed";
     return nullptr;
   }
   AnfNodePtr bn_training_reduce = CreateBNTrainingReduce(func_graph, node);

tests/ut/cpp/pre_activate/ascend/ir_fission/batch_norm_bert_fission_test.cc

@@ -28,7 +28,7 @@ class TestHWBatchNormBertFission : public BackendCommon {
   UT::PyFuncGraphFetcher get_py_fun_;
 };
 
-TEST_F(TestHWBatchNormBertFission, test_fused_batch_norm_fusion) {
+TEST_F(TestHWBatchNormBertFission, test_fused_batch_norm_fission) {
   FuncGraphPtr g = get_py_fun_.CallAndParseRet("test_batch_norm_bert_fission", "before");
   EXPECT_NE(g, nullptr);
   std::vector<int> shp_x{32, 64, 112, 112};
@@ -40,6 +40,23 @@ TEST_F(TestHWBatchNormBertFission, test_fused_batch_norm_fusion) {
     args_spec_list.push_back(y_abstract);
   }
   auto kg = GetKernelGraph(g, args_spec_list);
+  auto ret = kg->get_return();
+  EXPECT_NE(ret, nullptr);
+  auto make_tuple0 = ret->input(1);
+  EXPECT_NE(make_tuple0, nullptr);
+  auto tuple_getitem0 = make_tuple0->cast<CNodePtr>()->input(1);
+  EXPECT_NE(tuple_getitem0, nullptr);
+  auto make_tuple1 = tuple_getitem0->cast<CNodePtr>()->input(1);
+  EXPECT_NE(make_tuple1, nullptr);
+  auto tuple_getitem1 = make_tuple1->cast<CNodePtr>()->input(1);
+  EXPECT_NE(tuple_getitem1, nullptr);
+  auto bn = tuple_getitem1->cast<CNodePtr>()->input(1);
+  EXPECT_NE(bn, nullptr);
+  auto bn_cnode = bn->cast<CNodePtr>();
+  EXPECT_NE(bn_cnode, nullptr);
+  auto inputs = bn_cnode->inputs();
+  std::vector<AnfNodePtr> new_inputs(inputs.begin(), inputs.begin() + 4);
+  bn_cnode->set_inputs(new_inputs);
 
   auto optimizer = std::make_shared<opt::GraphOptimizer>();
   auto pm = std::make_shared<opt::PassManager>();
@@ -50,5 +67,27 @@ TEST_F(TestHWBatchNormBertFission, test_fused_batch_norm_fusion) {
   FuncGraphPtr g_after = get_py_fun_.CallAndParseRet("test_batch_norm_bert_fission", "after");
   EXPECT_TRUE(CheckEqualGraph(g_after, new_graph));
 }
+
+TEST_F(TestHWBatchNormBertFission, test_fused_batch_norm_no_fission) {
+  FuncGraphPtr g = get_py_fun_.CallAndParseRet("test_batch_norm_bert_fission", "before");
+  EXPECT_NE(g, nullptr);
+  std::vector<int> shp_x{32, 64, 112, 112};
+  auto x_abstract = std::make_shared<abstract::AbstractTensor>(kFloat32, shp_x);
+  std::vector<int> shp_y{64};
+  auto y_abstract = std::make_shared<abstract::AbstractTensor>(kFloat32, shp_y);
+  AbstractBasePtrList args_spec_list{x_abstract};
+  for (size_t i = 0; i < 4; ++i) {
+    args_spec_list.push_back(y_abstract);
+  }
+  auto kg = GetKernelGraph(g, args_spec_list);
+
+  auto optimizer = std::make_shared<opt::GraphOptimizer>();
+  auto pm = std::make_shared<opt::PassManager>();
+  pm->AddPass(std::make_shared<opt::BatchNormBertFission>());
+  optimizer->AddPassManager(pm);
+  FuncGraphPtr new_graph = optimizer->Optimize(kg);
+
+  EXPECT_TRUE(CheckEqualGraph(kg, new_graph));
+}
 }  // namespace opt
 }  // namespace mindspore