[bug-fix] Fix issue with initialize not resetting step count (#3962)

com.unity.ml-agents/CHANGELOG.md

 - Unity Player logs are now written out to the results directory. (#3877)
 - Run configuration YAML files are written out to the results directory at the end of the run. (#3815)
 ### Bug Fixes
+- An issue was fixed where using `--initialize-from` would resume from the past step count. (#3962)
 #### com.unity.ml-agents (C#)
 #### ml-agents / ml-agents-envs / gym-unity (Python)
 

ml-agents/mlagents/trainers/policy/tf_policy.py

                     )
                 )
             if reset_global_steps:
+                self._set_step(0)
                 logger.info(
                     "Starting training from step 0 and saving to {}.".format(
                         self.model_path

ml-agents/mlagents/trainers/tests/test_nn_policy.py

     trainer_params["output_path"] = path1
     policy = create_policy_mock(trainer_params)
     policy.initialize_or_load()
+    policy._set_step(2000)
     policy.save_model(2000)
 
     assert len(os.listdir(tmp_path)) > 0
     policy2.initialize_or_load()
     _compare_two_policies(policy, policy2)
+    assert policy2.get_current_step() == 2000
 
     # Try initialize from path 1
     trainer_params["model_path"] = path2
 
     _compare_two_policies(policy2, policy3)
+    # Assert that the steps are 0.
+    assert policy3.get_current_step() == 0
 
 
 def _compare_two_policies(policy1: NNPolicy, policy2: NNPolicy) -> None: