braindecode 0.4.1__tar.gz → 0.4.4__tar.gz

{Braindecode-0.4.1 → Braindecode-0.4.4}/Braindecode.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 1.1
 Name: Braindecode
-Version: 0.4.1
+Version: 0.4.4
 Summary: A deep learning toolbox to decode raw time-domain EEG.
 Home-page: https://github.com/robintibor/braindecode
 Author: Robin Tibor Schirrmeister

{Braindecode-0.4.1 → Braindecode-0.4.4}/Braindecode.egg-info/SOURCES.txt

@@ -63,7 +63,6 @@ docs/notebooks/Cropped_Trialwise_Explanation_Figures.ipynb
 docs/notebooks/Trialwise_Decoding.ipynb
 docs/notebooks/Trialwise_Manual_Training_Loop.ipynb
 docs/notebooks/visualization/Perturbation.ipynb
-docs/notebooks/visualization/Perturbation_Old.ipynb
 docs/source/braindecode.datasets.rst
 docs/source/braindecode.datautil.rst
 docs/source/braindecode.experiments.rst

{Braindecode-0.4.1 → Braindecode-0.4.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 1.1
 Name: Braindecode
-Version: 0.4.1
+Version: 0.4.4
 Summary: A deep learning toolbox to decode raw time-domain EEG.
 Home-page: https://github.com/robintibor/braindecode
 Author: Robin Tibor Schirrmeister

{Braindecode-0.4.1 → Braindecode-0.4.4}/braindecode/experiments/experiment.py

@@ -146,8 +146,6 @@ class Experiment(object):
         start of training.
     loggers: list of :class:`.Logger`
         How to show computed metrics.
-    seed: int
-        Random seed for python random module numpy.random and torch.
         
     Attributes
     ----------
@@ -163,8 +161,7 @@ class Experiment(object):
                  do_early_stop=True,
                  reset_after_second_run=False,
                  log_0_epoch=True,
-                 loggers=('print',),
-                 seed=2382938):
+                 loggers=('print',)):
         if run_after_early_stop or reset_after_second_run:
             assert do_early_stop == True, ("Can only run after early stop or "
             "reset after second run if doing an early stop")
@@ -199,7 +196,6 @@ class Experiment(object):
         self.do_early_stop = do_early_stop
         self.reset_after_second_run = reset_after_second_run
         self.log_0_epoch = log_0_epoch
-        self.seed = seed
         self.loggers = loggers
 
     def run(self):
@@ -230,7 +226,7 @@ class Experiment(object):
 
     def setup_training(self):
         """
-        Setup training, i.e. set random seeds, transform model to cuda,
+        Setup training, i.e. transform model to cuda,
         initialize monitoring.
         """
         # reset remember best extension in case you rerun some experiment
@@ -239,7 +235,6 @@ class Experiment(object):
         if self.loggers == ('print',):
             self.loggers = [Printer()]
         self.epochs_df = pd.DataFrame()
-        set_random_seeds(seed=self.seed, cuda=self.cuda)
         if self.cuda:
             assert th.cuda.is_available(), "Cuda not available"
             self.model.cuda()

{Braindecode-0.4.1 → Braindecode-0.4.4}/braindecode/models/base.py

@@ -57,7 +57,7 @@ class BaseModel(object):
             self.cuda = False
             self.compiled = False
 
-    def compile(self, loss, optimizer, monitors=None,  cropped=False, iterator_seed=0):
+    def compile(self, loss, optimizer, extra_monitors=None,  cropped=False, iterator_seed=0):
         """
         Setup training for this model.
         
@@ -66,7 +66,7 @@ class BaseModel(object):
         loss: function (predictions, targets) -> torch scalar
         optimizer: `torch.optim.Optimizer` or string
             Either supply an optimizer or the name of the class (e.g. 'adam')
-        monitors: List of Braindecode monitors, optional
+        extra_monitors: List of Braindecode monitors, optional
             In case you want to monitor additional values except for loss, misclass and runtime.
         cropped: bool
             Whether to perform cropped decoding, see cropped decoding tutorial.
@@ -84,7 +84,7 @@ class BaseModel(object):
             optimizer_class = find_optimizer(optimizer)
             optimizer = optimizer_class(self.network.parameters())
         self.optimizer = optimizer
-        self.extra_monitors = monitors
+        self.extra_monitors = extra_monitors
         # Already setting it here, so multiple calls to fit
         # will lead to different batches being drawn
         self.seed_rng = RandomState(iterator_seed)
@@ -93,7 +93,8 @@ class BaseModel(object):
 
     def fit(self, train_X, train_y, epochs, batch_size, input_time_length=None,
             validation_data=None, model_constraint=None,
-            remember_best_column=None, scheduler=None):
+            remember_best_column=None, scheduler=None,
+            log_0_epoch=True):
         """
         Fit the model using the given training data.
         
@@ -122,6 +123,8 @@ class BaseModel(object):
             determines the best epoch.
         scheduler: 'cosine' or None, optional
             Whether to use cosine annealing (:class:`.CosineAnnealing`).
+        log_0_epoch: bool
+            Whether to compute the metrics once before training as well.
 
         Returns
         -------
@@ -148,12 +151,15 @@ class BaseModel(object):
             self.iterator = CropsFromTrialsIterator(
                 batch_size=batch_size, input_time_length=input_time_length,
                 n_preds_per_input=n_preds_per_input,
-                seed=self.seed_rng.randint(0, 4294967295))
+                seed=self.seed_rng.randint(0, np.iinfo(np.int32).max-1))
         else:
             self.iterator = BalancedBatchSizeIterator(
                 batch_size=batch_size,
-                seed=self.seed_rng.randint(0, 4294967295))
-        stop_criterion = MaxEpochs(epochs - 1)# -1 since we dont print 0 epoch, which matters for this stop criterion
+                seed=self.seed_rng.randint(0, np.iinfo(np.int32).max-1))
+        if log_0_epoch:
+            stop_criterion = MaxEpochs(epochs)
+        else:
+            stop_criterion = MaxEpochs(epochs - 1)
         train_set = SignalAndTarget(train_X, train_y)
         optimizer = self.optimizer
         if scheduler is not None:
@@ -193,7 +199,8 @@ class BaseModel(object):
                          monitors=self.monitors,
                          stop_criterion=stop_criterion,
                          remember_best_column=remember_best_column,
-                         run_after_early_stop=False, cuda=self.cuda, log_0_epoch=False,
+                         run_after_early_stop=False, cuda=self.cuda,
+                         log_0_epoch=log_0_epoch,
                          do_early_stop=(remember_best_column is not None))
         exp.run()
         self.epochs_df = exp.epochs_df
@@ -238,7 +245,7 @@ class BaseModel(object):
                          stop_criterion=stop_criterion,
                          remember_best_column=None,
                          run_after_early_stop=False, cuda=self.cuda,
-                         log_0_epoch=False,
+                         log_0_epoch=True,
                          do_early_stop=False)
 
         exp.monitor_epoch({'train': train_set})

{Braindecode-0.4.1 → Braindecode-0.4.4}/braindecode/models/deep4.py

@@ -17,10 +17,10 @@ class Deep4Net(BaseModel):
     ----------
 
     .. [1] Schirrmeister, R. T., Springenberg, J. T., Fiederer, L. D. J., 
-       Glasstetter, M., Eggensperger, K., Tangermann, M., ... & Ball, T. (2017).
+       Glasstetter, M., Eggensperger, K., Tangermann, M., Hutter, F. & Ball, T. (2017).
        Deep learning with convolutional neural networks for EEG decoding and
        visualization.
-       arXiv preprint arXiv:1703.05051.
+       Human Brain Mapping , Aug. 2017. Online: http://dx.doi.org/10.1002/hbm.23730
     """
     def __init__(self, in_chans,
                  n_classes,
@@ -103,7 +103,7 @@ class Deep4Net(BaseModel):
             suffix = '_{:d}'.format(block_nr)
             model.add_module('drop' + suffix,
                              nn.Dropout(p=self.drop_prob))
-            model.add_module('conv' + suffix.format(block_nr),
+            model.add_module('conv' + suffix,
                              nn.Conv2d(n_filters_before, n_filters,
                                        (filter_length, 1),
                                        stride=(conv_stride, 1),

{Braindecode-0.4.1 → Braindecode-0.4.4}/braindecode/models/hybrid.py

@@ -10,18 +10,18 @@ from braindecode.models.util import to_dense_prediction_model
 
 class HybridNet(nn.Module, BaseModel):
     """
-       Hybrid ConvNet model from [3]_.
+    Hybrid ConvNet model from [3]_.
        
-       Very hardcoded at the moment.
+    Very hardcoded at the moment.
 
-       References
-       ----------
+    References
+    ----------
 
-       .. [3] Schirrmeister, R. T., Springenberg, J. T., Fiederer, L. D. J., 
-          Glasstetter, M., Eggensperger, K., Tangermann, M., ... & Ball, T. (2017).
-          Deep learning with convolutional neural networks for EEG decoding and
-          visualization.
-          arXiv preprint arXiv:1703.05051.
+    .. [3] Schirrmeister, R. T., Springenberg, J. T., Fiederer, L. D. J.,
+       Glasstetter, M., Eggensperger, K., Tangermann, M., Hutter, F. & Ball, T. (2017).
+       Deep learning with convolutional neural networks for EEG decoding and
+       visualization.
+       Human Brain Mapping , Aug. 2017. Online: http://dx.doi.org/10.1002/hbm.23730
     """
     def __init__(self, n_chans, n_classes, input_time_length):
         super(HybridNet, self).__init__()

{Braindecode-0.4.1 → Braindecode-0.4.4}/braindecode/models/shallow_fbcsp.py

@@ -15,11 +15,11 @@ class ShallowFBCSPNet(BaseModel):
     References
     ----------
 
-    .. [2] Schirrmeister, R. T., Springenberg, J. T., Fiederer, L. D. J., 
-       Glasstetter, M., Eggensperger, K., Tangermann, M., ... & Ball, T. (2017).
+    .. [2] Schirrmeister, R. T., Springenberg, J. T., Fiederer, L. D. J.,
+       Glasstetter, M., Eggensperger, K., Tangermann, M., Hutter, F. & Ball, T. (2017).
        Deep learning with convolutional neural networks for EEG decoding and
        visualization.
-       arXiv preprint arXiv:1703.05051.
+       Human Brain Mapping , Aug. 2017. Online: http://dx.doi.org/10.1002/hbm.23730
     """
     def __init__(self, in_chans,
                  n_classes,

{Braindecode-0.4.1 → Braindecode-0.4.4}/braindecode/torch_ext/constraints.py

@@ -10,10 +10,10 @@ class MaxNormDefaultConstraint(object):
     ----------
 
     .. [1] Schirrmeister, R. T., Springenberg, J. T., Fiederer, L. D. J., 
-       Glasstetter, M., Eggensperger, K., Tangermann, M., ... & Ball, T. (2017).
+       Glasstetter, M., Eggensperger, K., Tangermann, M., Hutter, F. & Ball, T. (2017).
        Deep learning with convolutional neural networks for EEG decoding and
        visualization.
-       arXiv preprint arXiv:1703.05051.
+       Human Brain Mapping , Aug. 2017. Online: http://dx.doi.org/10.1002/hbm.23730
     
     """
     def apply(self, model):

Braindecode-0.4.4/braindecode/version.py

@@ -0,0 +1 @@
+__version__ = "0.4.4"

{Braindecode-0.4.1 → Braindecode-0.4.4}/braindecode/visualization/perturbation.py

@@ -278,10 +278,10 @@ def compute_amplitude_prediction_correlations(pred_fn, examples, n_iterations,
     ----------
 
     .. [EEGDeepLearning] Schirrmeister, R. T., Springenberg, J. T., Fiederer, L. D. J.,
-       Glasstetter, M., Eggensperger, K., Tangermann, M., ... & Ball, T. (2017).
+       Glasstetter, M., Eggensperger, K., Tangermann, M., Hutter, F. & Ball, T. (2017).
        Deep learning with convolutional neural networks for EEG decoding and
        visualization.
-       arXiv preprint arXiv:1703.05051.
+       Human Brain Mapping , Aug. 2017. Online: http://dx.doi.org/10.1002/hbm.23730
     """
     pred_fn_new = lambda x: [pred_fn(x)]
     pred_corrs = spectral_perturbation_correlation(perturb_fn, mean_diff_feature_maps,

{Braindecode-0.4.1 → Braindecode-0.4.4}/braindecode/visualization/plot.py

@@ -67,13 +67,11 @@ def ax_scalp(v, channels,
     References
     ----------
 
-
     .. [1] Schirrmeister, R. T., Springenberg, J. T., Fiederer, L. D. J.,
-       Glasstetter, M., Eggensperger, K., Tangermann, M., ... & Ball, T. (2017).
+       Glasstetter, M., Eggensperger, K., Tangermann, M., Hutter, F. & Ball, T. (2017).
        Deep learning with convolutional neural networks for EEG decoding and
        visualization.
-       arXiv preprint arXiv:1703.05051.
-
+       Human Brain Mapping , Aug. 2017. Online: http://dx.doi.org/10.1002/hbm.23730
     """
     if ax is None:
         ax = plt.gca()

{Braindecode-0.4.1 → Braindecode-0.4.4}/docs/notebooks/BBCI_Data.ipynb

@@ -833,7 +833,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.6.2"
+   "version": "3.6.6"
   }
  },
  "nbformat": 4,

{Braindecode-0.4.1 → Braindecode-0.4.4}/docs/notebooks/BBCI_Data_Start_Stop.ipynb

@@ -891,7 +891,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.6.2"
+   "version": "3.6.6"
   }
  },
  "nbformat": 4,

{Braindecode-0.4.1 → Braindecode-0.4.4}/docs/notebooks/BBCI_Data_Trialwise.ipynb

@@ -707,7 +707,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.6.2"
+   "version": "3.6.6"
   }
  },
  "nbformat": 4,

{Braindecode-0.4.1 → Braindecode-0.4.4}/docs/notebooks/Cropped_Manual_Training_Loop.ipynb

@@ -50,7 +50,9 @@
   {
    "cell_type": "code",
    "execution_count": 2,
-   "metadata": {},
+   "metadata": {
+    "collapsed": true
+   },
    "outputs": [],
    "source": [
     "import mne\n",
@@ -93,7 +95,9 @@
   {
    "cell_type": "code",
    "execution_count": 3,
-   "metadata": {},
+   "metadata": {
+    "collapsed": true
+   },
    "outputs": [],
    "source": [
     "import numpy as np\n",
@@ -195,7 +199,9 @@
   {
    "cell_type": "code",
    "execution_count": 6,
-   "metadata": {},
+   "metadata": {
+    "collapsed": true
+   },
    "outputs": [],
    "source": [
     "from braindecode.datautil.iterators import CropsFromTrialsIterator\n",
@@ -227,7 +233,9 @@
   {
    "cell_type": "code",
    "execution_count": 7,
-   "metadata": {},
+   "metadata": {
+    "collapsed": true
+   },
    "outputs": [],
    "source": [
     "from braindecode.torch_ext.optimizers import AdamW\n",
@@ -269,105 +277,105 @@
      "output_type": "stream",
      "text": [
       "Epoch 0\n",
-      "Train  Loss: 3.88829\n",
+      "Train  Loss: 3.79010\n",
       "Train  Accuracy: 50.0%\n",
-      "Valid  Loss: 3.18595\n",
-      "Valid  Accuracy: 50.0%\n",
+      "Valid  Loss: 3.12765\n",
+      "Valid  Accuracy: 46.7%\n",
       "Epoch 1\n",
-      "Train  Loss: 1.33431\n",
-      "Train  Accuracy: 55.0%\n",
-      "Valid  Loss: 0.96900\n",
-      "Valid  Accuracy: 60.0%\n",
+      "Train  Loss: 1.91778\n",
+      "Train  Accuracy: 50.0%\n",
+      "Valid  Loss: 1.52058\n",
+      "Valid  Accuracy: 46.7%\n",
       "Epoch 2\n",
-      "Train  Loss: 0.60838\n",
-      "Train  Accuracy: 77.5%\n",
-      "Valid  Loss: 0.55685\n",
-      "Valid  Accuracy: 76.7%\n",
+      "Train  Loss: 1.09943\n",
+      "Train  Accuracy: 60.0%\n",
+      "Valid  Loss: 0.99602\n",
+      "Valid  Accuracy: 56.7%\n",
       "Epoch 3\n",
-      "Train  Loss: 0.49553\n",
-      "Train  Accuracy: 77.5%\n",
-      "Valid  Loss: 0.57463\n",
-      "Valid  Accuracy: 76.7%\n",
+      "Train  Loss: 0.73015\n",
+      "Train  Accuracy: 67.5%\n",
+      "Valid  Loss: 0.82321\n",
+      "Valid  Accuracy: 56.7%\n",
       "Epoch 4\n",
-      "Train  Loss: 0.46749\n",
-      "Train  Accuracy: 80.0%\n",
-      "Valid  Loss: 0.62036\n",
-      "Valid  Accuracy: 76.7%\n",
+      "Train  Loss: 0.55965\n",
+      "Train  Accuracy: 75.0%\n",
+      "Valid  Loss: 0.77549\n",
+      "Valid  Accuracy: 63.3%\n",
       "Epoch 5\n",
-      "Train  Loss: 0.36088\n",
-      "Train  Accuracy: 85.0%\n",
-      "Valid  Loss: 0.59054\n",
-      "Valid  Accuracy: 76.7%\n",
+      "Train  Loss: 0.38075\n",
+      "Train  Accuracy: 82.5%\n",
+      "Valid  Loss: 0.65490\n",
+      "Valid  Accuracy: 73.3%\n",
       "Epoch 6\n",
-      "Train  Loss: 0.26471\n",
+      "Train  Loss: 0.28886\n",
       "Train  Accuracy: 90.0%\n",
-      "Valid  Loss: 0.55745\n",
-      "Valid  Accuracy: 76.7%\n",
+      "Valid  Loss: 0.59481\n",
+      "Valid  Accuracy: 73.3%\n",
       "Epoch 7\n",
-      "Train  Loss: 0.19023\n",
-      "Train  Accuracy: 95.0%\n",
-      "Valid  Loss: 0.53162\n",
+      "Train  Loss: 0.21871\n",
+      "Train  Accuracy: 97.5%\n",
+      "Valid  Loss: 0.52357\n",
       "Valid  Accuracy: 83.3%\n",
       "Epoch 8\n",
-      "Train  Loss: 0.13684\n",
+      "Train  Loss: 0.16713\n",
       "Train  Accuracy: 97.5%\n",
-      "Valid  Loss: 0.50510\n",
-      "Valid  Accuracy: 83.3%\n",
+      "Valid  Loss: 0.44890\n",
+      "Valid  Accuracy: 86.7%\n",
       "Epoch 9\n",
-      "Train  Loss: 0.10682\n",
+      "Train  Loss: 0.13149\n",
       "Train  Accuracy: 97.5%\n",
-      "Valid  Loss: 0.47980\n",
-      "Valid  Accuracy: 86.7%\n",
+      "Valid  Loss: 0.40492\n",
+      "Valid  Accuracy: 83.3%\n",
       "Epoch 10\n",
-      "Train  Loss: 0.08165\n",
+      "Train  Loss: 0.09581\n",
       "Train  Accuracy: 100.0%\n",
-      "Valid  Loss: 0.45300\n",
-      "Valid  Accuracy: 86.7%\n",
+      "Valid  Loss: 0.36021\n",
+      "Valid  Accuracy: 90.0%\n",
       "Epoch 11\n",
-      "Train  Loss: 0.05927\n",
+      "Train  Loss: 0.07818\n",
       "Train  Accuracy: 100.0%\n",
-      "Valid  Loss: 0.43593\n",
-      "Valid  Accuracy: 86.7%\n",
+      "Valid  Loss: 0.34625\n",
+      "Valid  Accuracy: 90.0%\n",
       "Epoch 12\n",
-      "Train  Loss: 0.04447\n",
+      "Train  Loss: 0.07454\n",
       "Train  Accuracy: 100.0%\n",
-      "Valid  Loss: 0.43543\n",
-      "Valid  Accuracy: 83.3%\n",
+      "Valid  Loss: 0.34489\n",
+      "Valid  Accuracy: 90.0%\n",
       "Epoch 13\n",
-      "Train  Loss: 0.03776\n",
+      "Train  Loss: 0.06694\n",
       "Train  Accuracy: 100.0%\n",
-      "Valid  Loss: 0.43941\n",
-      "Valid  Accuracy: 83.3%\n",
+      "Valid  Loss: 0.33878\n",
+      "Valid  Accuracy: 90.0%\n",
       "Epoch 14\n",
-      "Train  Loss: 0.03341\n",
+      "Train  Loss: 0.05971\n",
       "Train  Accuracy: 100.0%\n",
-      "Valid  Loss: 0.43653\n",
-      "Valid  Accuracy: 83.3%\n",
+      "Valid  Loss: 0.33128\n",
+      "Valid  Accuracy: 90.0%\n",
       "Epoch 15\n",
-      "Train  Loss: 0.02990\n",
+      "Train  Loss: 0.05269\n",
       "Train  Accuracy: 100.0%\n",
-      "Valid  Loss: 0.43027\n",
-      "Valid  Accuracy: 83.3%\n",
+      "Valid  Loss: 0.32202\n",
+      "Valid  Accuracy: 90.0%\n",
       "Epoch 16\n",
-      "Train  Loss: 0.02713\n",
+      "Train  Loss: 0.04354\n",
       "Train  Accuracy: 100.0%\n",
-      "Valid  Loss: 0.42535\n",
-      "Valid  Accuracy: 83.3%\n",
+      "Valid  Loss: 0.31063\n",
+      "Valid  Accuracy: 90.0%\n",
       "Epoch 17\n",
-      "Train  Loss: 0.02483\n",
+      "Train  Loss: 0.03759\n",
       "Train  Accuracy: 100.0%\n",
-      "Valid  Loss: 0.42535\n",
-      "Valid  Accuracy: 83.3%\n",
+      "Valid  Loss: 0.30314\n",
+      "Valid  Accuracy: 90.0%\n",
       "Epoch 18\n",
-      "Train  Loss: 0.02301\n",
+      "Train  Loss: 0.03401\n",
       "Train  Accuracy: 100.0%\n",
-      "Valid  Loss: 0.42749\n",
-      "Valid  Accuracy: 83.3%\n",
+      "Valid  Loss: 0.29997\n",
+      "Valid  Accuracy: 90.0%\n",
       "Epoch 19\n",
-      "Train  Loss: 0.02164\n",
+      "Train  Loss: 0.03145\n",
       "Train  Accuracy: 100.0%\n",
-      "Valid  Loss: 0.42855\n",
-      "Valid  Accuracy: 86.7%\n"
+      "Valid  Loss: 0.29764\n",
+      "Valid  Accuracy: 90.0%\n"
      ]
     }
    ],
@@ -381,7 +389,7 @@
     "for i_epoch in range(20):\n",
     "    # Set model to training mode\n",
     "    model.train()\n",
-    "    for batch_X, batch_y in iterator.get_batches(train_set, shuffle=False):\n",
+    "    for batch_X, batch_y in iterator.get_batches(train_set, shuffle=True):\n",
     "        net_in = np_to_var(batch_X)\n",
     "        if cuda:\n",
     "            net_in = net_in.cuda()\n",
@@ -443,7 +451,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "Eventually, we arrive at 86.7% accuracy, so 26 from 30 trials are correctly predicted, 4 more than for the trialwise decoding method."
+    "Eventually, we arrive at 90.0% accuracy, so 27 from 30 trials are correctly predicted, 5 more than for the trialwise decoding method."
    ]
   },
   {
@@ -462,14 +470,14 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 10,
+   "execution_count": 9,
    "metadata": {},
    "outputs": [
     {
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "Test Loss: 0.43439\n",
+      "Test Loss: 0.42105\n",
       "Test Accuracy: 85.0%\n"
      ]
     }

{Braindecode-0.4.1 → Braindecode-0.4.4}/docs/notebooks/Trialwise_Manual_Training_Loop.ipynb

@@ -92,7 +92,9 @@
   {
    "cell_type": "code",
    "execution_count": 4,
-   "metadata": {},
+   "metadata": {
+    "collapsed": true
+   },
    "outputs": [],
    "source": [
     "import mne\n",
@@ -226,7 +228,9 @@
   {
    "cell_type": "code",
    "execution_count": 21,
-   "metadata": {},
+   "metadata": {
+    "collapsed": true
+   },
    "outputs": [],
    "source": [
     "from braindecode.torch_ext.optimizers import AdamW\n",