diff --git a/onnxruntime/python/onnxruntime_pybind_state_common.h b/onnxruntime/python/onnxruntime_pybind_state_common.h
index dbc96096cb..0d541dbfa5 100644
--- a/onnxruntime/python/onnxruntime_pybind_state_common.h
+++ b/onnxruntime/python/onnxruntime_pybind_state_common.h
@@ -42,9 +42,6 @@ struct PySessionOptions : public SessionOptions {
 
 // Thin wrapper over internal C++ InferenceSession to accommodate custom op library management for the Python user
 struct PyInferenceSession {
-  // Default ctor is present only to be invoked by the PyTrainingSession class
-  PyInferenceSession() {}
-
   PyInferenceSession(Environment& env, const PySessionOptions& so, const std::string& arg, bool is_arg_file_name) {
     if (is_arg_file_name) {
       // Given arg is the file path. Invoke the corresponding ctor().
@@ -70,6 +67,10 @@ struct PyInferenceSession {
   virtual ~PyInferenceSession() {}
 
  protected:
+  PyInferenceSession(std::unique_ptr<InferenceSession> sess) {
+    sess_ = std::move(sess);
+  }
+
   // Hold CustomOpLibrary resources so as to tie it to the life cycle of the InferenceSession needing it.
   // NOTE: Declare this above `sess_` so that this is destructed AFTER the InferenceSession instance -
   // this is so that the custom ops held by the InferenceSession gets destroyed prior to the library getting unloaded
diff --git a/orttraining/orttraining/python/orttraining_pybind_state.cc b/orttraining/orttraining/python/orttraining_pybind_state.cc
index e110e6b950..a818d37564 100644
--- a/orttraining/orttraining/python/orttraining_pybind_state.cc
+++ b/orttraining/orttraining/python/orttraining_pybind_state.cc
@@ -206,9 +206,8 @@ void addObjectMethodsForTraining(py::module& m) {
 
   // Thin wrapper over internal C++ InferenceSession to accommodate custom op library management for the Python user
   struct PyTrainingSession : public PyInferenceSession {
-    PyTrainingSession(Environment& env, const PySessionOptions& so) {
-      // `sess_` is inherited from PyinferenceSession
-      sess_ = onnxruntime::make_unique<onnxruntime::training::TrainingSession>(so, env);
+    PyTrainingSession(Environment& env, const PySessionOptions& so)
+        : PyInferenceSession(onnxruntime::make_unique<TrainingSession>(so, env)) {
     }
   };