Add save_metrics flag to disable loggings

d3rlpy/base.py

@@ -244,6 +244,7 @@ def load_model(self, fname):
 
     def fit(self,
             episodes,
+            save_metrics=True,
             experiment_name=None,
             with_timestamp=True,
             logdir='d3rlpy_logs',
@@ -261,6 +262,9 @@ def fit(self,
 
         Args:
             episodes (list(d3rlpy.dataset.Episode)): list of episodes to train.
+            save_metrics (bool): flag to record metrics in files. If False,
+                the log directory is not created and the model parameters are
+                not saved during training.
             experiment_name (str): experiment name for logging. If not passed,
                 the directory name will be `{class name}_{timestamp}`.
             with_timestamp (bool): flag to add timestamp string to the last of
@@ -298,8 +302,9 @@ def fit(self,
             self.create_impl(observation_shape, action_size)
 
         # setup logger
-        logger = self._prepare_logger(experiment_name, with_timestamp, logdir,
-                                      verbose, tensorboard)
+        logger = self._prepare_logger(save_metrics, experiment_name,
+                                      with_timestamp, logdir, verbose,
+                                      tensorboard)
 
         # save hyperparameters
         self._save_params(logger)
@@ -392,12 +397,13 @@ def _generate_new_data(self, transitions):
     def _get_loss_labels(self):
         raise NotImplementedError
 
-    def _prepare_logger(self, experiment_name, with_timestamp, logdir, verbose,
-                        tensorboard):
+    def _prepare_logger(self, save_metrics, experiment_name, with_timestamp,
+                        logdir, verbose, tensorboard):
         if experiment_name is None:
             experiment_name = self.__class__.__name__
 
         logger = D3RLPyLogger(experiment_name,
+                              save_metrics=save_metrics,
                               root_dir=logdir,
                               verbose=verbose,
                               tensorboard=tensorboard,

d3rlpy/logger.py

@@ -20,10 +20,12 @@ def default_json_encoder(obj):
 class D3RLPyLogger:
     def __init__(self,
                  experiment_name,
+                 save_metrics=True,
                  root_dir='logs',
                  verbose=True,
                  tensorboard=True,
                  with_timestamp=True):
+        self.save_metrics = save_metrics
         self.verbose = verbose
 
         # add timestamp to prevent unintentional overwrites
@@ -34,16 +36,18 @@ def __init__(self,
             else:
                 self.experiment_name = experiment_name
 
-            self.logdir = os.path.join(root_dir, self.experiment_name)
-
-            if not os.path.exists(self.logdir):
-                os.makedirs(self.logdir)
-                break
-            else:
-                if with_timestamp:
-                    time.sleep(1.0)
+            if self.save_metrics:
+                self.logdir = os.path.join(root_dir, self.experiment_name)
+                if not os.path.exists(self.logdir):
+                    os.makedirs(self.logdir)
+                    break
                 else:
-                    raise ValueError('%s already exists.' % self.logdir)
+                    if with_timestamp:
+                        time.sleep(1.0)
+                    else:
+                        raise ValueError('%s already exists.' % self.logdir)
+            else:
+                break
 
         self.metrics_buffer = {}
 
@@ -59,9 +63,10 @@ def __init__(self,
     def add_params(self, params):
         assert self.params is None, 'add_params can be called only once.'
 
-        # save dictionary as json file
-        with open(os.path.join(self.logdir, 'params.json'), 'w') as f:
-            f.write(json.dumps(params, default=default_json_encoder))
+        if self.save_metrics:
+            # save dictionary as json file
+            with open(os.path.join(self.logdir, 'params.json'), 'w') as f:
+                f.write(json.dumps(params, default=default_json_encoder))
 
         if self.verbose:
             for key, val in params.items():
@@ -80,8 +85,9 @@ def commit(self, epoch, step):
         for name, buffer in self.metrics_buffer.items():
             metric = sum(buffer) / len(buffer)
 
-            with open(os.path.join(self.logdir, name + '.csv'), 'a') as f:
-                print('%d,%d,%f' % (epoch, step, metric), file=f)
+            if self.save_metrics:
+                with open(os.path.join(self.logdir, name + '.csv'), 'a') as f:
+                    print('%d,%d,%f' % (epoch, step, metric), file=f)
 
             if self.verbose:
                 print('epoch=%d step=%d %s=%f' % (epoch, step, name, metric))
@@ -99,6 +105,7 @@ def commit(self, epoch, step):
                                     global_step=epoch)
 
     def save_model(self, epoch, algo):
-        # save entire model
-        model_path = os.path.join(self.logdir, 'model_%d.pt' % epoch)
-        algo.save_model(model_path)
+        if self.save_metrics:
+            # save entire model
+            model_path = os.path.join(self.logdir, 'model_%d.pt' % epoch)
+            algo.save_model(model_path)

d3rlpy/online/iterators.py

@@ -15,6 +15,7 @@ def train(env,
           update_start_step=0,
           eval_env=None,
           eval_epsilon=0.05,
+          save_metrics=True,
           experiment_name=None,
           with_timestamp=True,
           logdir='d3rlpy_logs',
@@ -38,6 +39,8 @@ def train(env,
             skipped.
         eval_epsilon (float): :math:`\\epsilon`-greedy factor during
             evaluation.
+        save_metrics (bool): flag to record metrics. If False, the log
+            directory is not created and the model parameters are not saved.
         experiment_name (str): experiment name for logging. If not passed,
             the directory name will be `{class name}_online_{timestamp}`.
         with_timestamp (bool): flag to add timestamp string to the last of
@@ -55,6 +58,7 @@ def train(env,
     if experiment_name is None:
         experiment_name = algo.__class__.__name__ + '_online'
     logger = D3RLPyLogger(experiment_name,
+                          save_metrics=save_metrics,
                           root_dir=logdir,
                           verbose=verbose,
                           tensorboard=tensorboard,

docs/references/logging.rst

@@ -23,10 +23,20 @@ You can designate the directory.
     # the directory will be `custom_logs/custom_YYYYMMDDHHmmss`
     dqn.fit(dataset.episodes, logdir='custom_logs', experiment_name='custom')
 
+If you want to disable all loggings, you can pass `save_metrics=False`.
+
+.. code-block:: python
+
+   dqn.fit(dataset.episodes, save_metrics=False)
+
+TensorBoard
+-----------
+
 The same information is also automatically saved for tensorboard under `runs`
 directory.
 You can interactively visualize training metrics easily.
 
+
 .. code-block:: shell
 
     $ pip install tensorboard