Make create losses private

ml-agents/mlagents/trainers/ppo/optimizer.py

                 self.learning_rate = LearningModel.create_learning_rate(
                     lr_schedule, lr, self.policy.global_step, int(max_step)
                 )
-                self.create_losses(
+                self._create_losses(
                     self.policy.log_probs,
                     self.old_log_probs,
                     self.value_heads,
             keepdims=True,
         )
 
-    def create_losses(
+    def _create_losses(
         self, probs, old_probs, value_heads, entropy, beta, epsilon, lr, max_step
     ):
         """

ml-agents/mlagents/trainers/sac/optimizer.py

                 self.learning_rate = LearningModel.create_learning_rate(
                     lr_schedule, lr, self.policy.global_step, int(max_step)
                 )
-                self.create_losses(
+                self._create_losses(
                     self.policy_network.q1_heads,
                     self.policy_network.q2_heads,
                     lr,
                 self.prev_action = self.policy_network.prev_action
             self.next_memory_in = self.target_network.memory_in
 
-    def create_losses(
+    def _create_losses(
         self,
         q1_streams: Dict[str, tf.Tensor],
         q2_streams: Dict[str, tf.Tensor],