diff --git a/onnxruntime/core/optimizer/noop_elimination.cc b/onnxruntime/core/optimizer/noop_elimination.cc
index 1421ea7416..c01bdc42bf 100644
--- a/onnxruntime/core/optimizer/noop_elimination.cc
+++ b/onnxruntime/core/optimizer/noop_elimination.cc
@@ -55,6 +55,10 @@ bool NoopElimination::SatisfyCondition(const Graph& graph, const Node& node, con
   if (add_init.size() > 1) {
     return false;
   }
+  // handle edge case where the total size of the initializer is 0
+  if (add_init.size() == 0) {
+    return true;
+  }
   switch (data_type) {
     case ONNX_NAMESPACE::TensorProto_DataType_FLOAT:
       if (*add_init.data<float>() != 0.f) {
diff --git a/onnxruntime/test/optimizer/graph_transform_test.cc b/onnxruntime/test/optimizer/graph_transform_test.cc
index 3cfc34492d..70256732ba 100644
--- a/onnxruntime/test/optimizer/graph_transform_test.cc
+++ b/onnxruntime/test/optimizer/graph_transform_test.cc
@@ -195,7 +195,7 @@ TEST_F(GraphTransformationTests, NoopElimination) {
   ASSERT_STATUS_OK(Model::Load(model_uri, model, nullptr, *logger_));
   Graph& graph = model->MainGraph();
   std::map<std::string, int> op_to_count = CountOpsInGraph(graph);
-  ASSERT_TRUE(op_to_count["Add"] == 4);
+  ASSERT_TRUE(op_to_count["Add"] == 5);
 
   auto rule_transformer_L1 = std::make_unique<RuleBasedGraphTransformer>("RuleTransformer1");
   ASSERT_STATUS_OK(rule_transformer_L1->Register(std::make_unique<NoopElimination>()));
diff --git a/onnxruntime/test/testdata/transform/noop-add.onnx b/onnxruntime/test/testdata/transform/noop-add.onnx
index e5793f5fda..b10a87283a 100644
Binary files a/onnxruntime/test/testdata/transform/noop-add.onnx and b/onnxruntime/test/testdata/transform/noop-add.onnx differ
diff --git a/onnxruntime/test/testdata/transform/noop-add.py b/onnxruntime/test/testdata/transform/noop-add.py
index 0d1bb9801b..3e996d42e6 100644
--- a/onnxruntime/test/testdata/transform/noop-add.py
+++ b/onnxruntime/test/testdata/transform/noop-add.py
@@ -34,6 +34,10 @@ def GenerateModel(model_name):
         helper.make_node("Identity", ["X4"], ["id_7"], "id_7"),
         helper.make_node("Add", ["id_7", "int64_2"], ["add_4"], "add_4"),
         helper.make_node("Identity", ["add_4"], ["Y4"], "id_8"),
+        #float
+        helper.make_node("Identity", ["X5"], ["id_9"], "id_9"),
+        helper.make_node("Add", ["float_2", "id_9"], ["add_5"], "add_5"),     
+        helper.make_node("Identity", ["add_5"], ["Y5"], "id_10"),
     ]
 
     inputs = [  # inputs
@@ -41,6 +45,7 @@ def GenerateModel(model_name):
             helper.make_tensor_value_info('X2', TensorProto.FLOAT16, ['M', 'K']),
             helper.make_tensor_value_info('X3', TensorProto.INT64, ['M', 'K']),
             helper.make_tensor_value_info('X4', TensorProto.INT64, ['M', 'K']),
+            helper.make_tensor_value_info('X5', TensorProto.FLOAT, ['M', 'K']),
         ]
 
     initializers = [
@@ -50,6 +55,8 @@ def GenerateModel(model_name):
             helper.make_tensor('int64_1', TensorProto.INT64, (), [0]),
             # higher rank
             helper.make_tensor('int64_2', TensorProto.INT64, [1,1,1], [0]),
+            #float - set initializer size = 0
+            helper.make_tensor('float_2', TensorProto.FLOAT, [0], []),
         ]
 
     graph = helper.make_graph(
@@ -61,6 +68,7 @@ def GenerateModel(model_name):
             helper.make_tensor_value_info('Y2', TensorProto.FLOAT16, ['M', 'K']),
             helper.make_tensor_value_info('Y3', TensorProto.INT64, ['M', 'K']),
             helper.make_tensor_value_info('Y4', TensorProto.INT64, ['M', 'K', 1]),
+            helper.make_tensor_value_info('Y5', TensorProto.FLOAT, ['M', 'K']),
         ],
         initializers)