initialize trainer step count (#2498) (#2505)

* initialize trainer step count

* remove step init from RLTrainer

ml-agents/mlagents/trainers/rl_trainer.py

 
     def __init__(self, *args, **kwargs):
         super(RLTrainer, self).__init__(*args, **kwargs)
-        self.step = 0
         # Make sure we have at least one reward_signal
         if not self.trainer_parameters["reward_signals"]:
             raise UnityTrainerException(

ml-agents/mlagents/trainers/trainer.py

         self.summary_writer = tf.summary.FileWriter(self.summary_path)
         self._reward_buffer: Deque[float] = deque(maxlen=reward_buff_cap)
         self.policy: Policy = None
+        self.step: int = 0
 
     def check_param_keys(self):
         for k in self.param_keys: