

In [2]:

    
car = ddpg(environment = 'Reacher-v1', noise_scale= 1.0, maxstep= 2e5, ql2= 0.01, learning_rates= (1e-4,1e-3))
car.main()









    



[2016-07-24 13:13:46,599] Making new env: Reacher-v1






    



state dim 11
action dim 2
result after minibatch no. 0 : mean squared error: 484.962677002
result after minibatch no. 1000 : mean squared error: 308.146331787
result after minibatch no. 2000 : mean squared error: 180.791763306
result after minibatch no. 3000 : mean squared error: 98.7425079346






    












    












    



result after minibatch no. 4000 : mean squared error: 46.9946746826
result after minibatch no. 5000 : mean squared error: 19.1027622223






    



---------------------------------------------------------------------------
KeyboardInterrupt                         Traceback (most recent call last)
<ipython-input-2-b81e7fe0bc0c> in <module>()
      1 car = ddpg(environment = 'Reacher-v1', noise_scale= 1.0, maxstep= 2e5, ql2= 0.01, learning_rates= (1e-4,1e-3))
----> 2 car.main()

/home/frederik/Dokumente/DeepRL/dpg/ddpg3.py in main(self)
    471 
    472         self.initialize_training(self.sess)
--> 473         return self.start_training()
    474 
    475     def plot_replay_memory_2d_state_histogramm(self):

/home/frederik/Dokumente/DeepRL/dpg/ddpg3.py in start_training(self, dataname, save)
    169 
    170             # run episode
--> 171             episode_length = self.run_episode(test_run= False, enable_render=False, limit= 10000)
    172 
    173             self.train_lengths.append(episode_length)

/home/frederik/Dokumente/DeepRL/dpg/ddpg3.py in run_episode(self, enable_render, limit, test_run)
    150 
    151                 if (len(self.replay_memory) > self.warmup) and (self.samples_count % (self.batch_size/2) == 0):
--> 152                     self.train_networks()
    153 
    154             state = state_prime

/home/frederik/Dokumente/DeepRL/dpg/ddpg3.py in train_networks(self)
    418 
    419         if self.tensorboard_logs:
--> 420             summary, _, _, mse_val = self.sess.run([self.merged, self.q_train_step, self.mu_train_step, self.q_loss], feed_dict= dict_)
    421             if self.step % 10 == 0:
    422                 self.train_writer.add_summary(summary, self.step)

/home/frederik/anaconda2/lib/python2.7/site-packages/tensorflow/python/client/session.pyc in run(self, fetches, feed_dict, options, run_metadata)
    370     try:
    371       result = self._run(None, fetches, feed_dict, options_ptr,
--> 372                          run_metadata_ptr)
    373       if run_metadata:
    374         proto_data = tf_session.TF_GetBuffer(run_metadata_ptr)

/home/frederik/anaconda2/lib/python2.7/site-packages/tensorflow/python/client/session.pyc in _run(self, handle, fetches, feed_dict, options, run_metadata)
    634     try:
    635       results = self._do_run(handle, target_list, unique_fetches,
--> 636                              feed_dict_string, options, run_metadata)
    637     finally:
    638       # The movers are no longer used. Delete them.

/home/frederik/anaconda2/lib/python2.7/site-packages/tensorflow/python/client/session.pyc in _do_run(self, handle, target_list, fetch_list, feed_dict, options, run_metadata)
    706     if handle is None:
    707       return self._do_call(_run_fn, self._session, feed_dict, fetch_list,
--> 708                            target_list, options, run_metadata)
    709     else:
    710       return self._do_call(_prun_fn, self._session, handle, feed_dict,

/home/frederik/anaconda2/lib/python2.7/site-packages/tensorflow/python/client/session.pyc in _do_call(self, fn, *args)
    713   def _do_call(self, fn, *args):
    714     try:
--> 715       return fn(*args)
    716     except errors.OpError as e:
    717       message = compat.as_text(e.message)

/home/frederik/anaconda2/lib/python2.7/site-packages/tensorflow/python/client/session.pyc in _run_fn(session, feed_dict, fetch_list, target_list, options, run_metadata)
    695         return tf_session.TF_Run(session, options,
    696                                  feed_dict, fetch_list, target_list,
--> 697                                  status, run_metadata)
    698 
    699     def _prun_fn(session, handle, feed_dict, fetch_list):

KeyboardInterrupt: