braindecode 1.6.1.dev1049__tar.gz → 1.6.1.dev1053__tar.gz

{braindecode-1.6.1.dev1049/braindecode.egg-info → braindecode-1.6.1.dev1053}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: braindecode
-Version: 1.6.1.dev1049
+Version: 1.6.1.dev1053
 Summary: Deep learning software to decode EEG, ECG or MEG signals
 Author-email: Robin Tibor Schirrmeister <robintibor@gmail.com>, Bruno Aristimunha Pinto <b.aristimunha@gmail.com>, Alexandre Gramfort <agramfort@meta.com>
 Maintainer-email: Alexandre Gramfort <agramfort@meta.com>, Bruno Aristimunha Pinto <b.aristimunha@gmail.com>, Robin Tibor Schirrmeister <robintibor@gmail.com>

{braindecode-1.6.1.dev1049 → braindecode-1.6.1.dev1053}/braindecode/eegneuralnet.py

@@ -7,6 +7,7 @@
 import abc
 import inspect
 import logging
+import warnings
 from typing import Literal
 
 import mne
@@ -41,6 +42,7 @@ def _get_model(model: str):
 
 class _EEGNeuralNet(NeuralNet, abc.ABC):
     signal_args_set_ = False
+    _warned_drop_last_ = False
 
     @property
     def log(self):
@@ -267,6 +269,44 @@ class _EEGNeuralNet(NeuralNet, abc.ABC):
             X = X.get_data(units="uV")
         return super().get_dataset(X, y)
 
+    def get_iterator(self, dataset, training=False):
+        iterator = super().get_iterator(dataset, training=training)
+        # Warn (once per fit) if drop_last would discard every training batch,
+        # e.g. a training set smaller than batch_size. This silently skips
+        # training without raising any error, which is hard to diagnose.
+        if training and not self._warned_drop_last_:
+            self._warn_if_training_batches_dropped(iterator)
+        return iterator
+
+    def _warn_if_training_batches_dropped(self, loader):
+        """Warn if ``drop_last`` discards all training batches.
+
+        With ``iterator_train__drop_last=True`` (the braindecode default), a
+        training set smaller than ``batch_size`` yields zero batches, so no
+        forward/backward pass runs and the model is left untrained without any
+        error. We inspect the resolved training ``DataLoader`` (after
+        ``train_split`` has been applied), so the count reflects the actual
+        number of training examples.
+        """
+        self._warned_drop_last_ = True
+        if not getattr(loader, "drop_last", False):
+            return
+        batch_size = getattr(loader, "batch_size", None)
+        try:
+            n_train = len(loader.dataset)
+        except TypeError:
+            return  # e.g. an IterableDataset whose length is unknown
+        if not batch_size or n_train == 0 or n_train >= batch_size:
+            return
+        warnings.warn(
+            f"The training set has {n_train} example(s), which is smaller than "
+            f"batch_size={batch_size}. With iterator_train__drop_last=True "
+            f"(the braindecode default), all training batches are dropped and "
+            f"the model is not trained. Reduce batch_size to at most {n_train}, "
+            f"or pass iterator_train__drop_last=False.",
+            UserWarning,
+        )
+
     def partial_fit(self, X, y=None, classes=None, **fit_params):
         """Fit the module.
 
@@ -366,6 +406,7 @@ class _EEGNeuralNet(NeuralNet, abc.ABC):
           the module and to the ``self.train_split`` call.
         """
         # this needs to be executed before the net is initialized:
+        self._warned_drop_last_ = False  # re-check on every fresh fit
         if not self.signal_args_set_:
             self._set_signal_args(X, y, classes=None)
             self.signal_args_set_ = True

{braindecode-1.6.1.dev1049 → braindecode-1.6.1.dev1053}/braindecode/preprocessing/mne_preprocess.py

@@ -13,6 +13,28 @@ import mne.preprocessing
 from braindecode.preprocessing.preprocess import Preprocessor
 from braindecode.util import _clean_docstring_sections
 
+# Standalone MNE functions that are safe to wrap as a *callable* preprocessor.
+#
+# When ``fn`` is a callable and ``apply_on_array=False``, ``Preprocessor`` calls
+# ``fn(raw, **kwargs)`` and reassigns the dataset to whatever ``fn`` returns. That
+# is only correct for functions that return the modified ``Raw``/``Epochs``
+# instance. Many standalone MNE functions instead return auxiliary data
+# (annotations, a list of bad channels, a ``(inst, ...)`` tuple, ...), so passing
+# them as callables would silently replace the recording with that auxiliary
+# object. We therefore restrict the callable path to functions verified to return
+# the modified instance; every other standalone function keeps the existing
+# (string-name) behaviour. See :gh:`885`.
+_SAFE_STANDALONE_FUNCTIONS = frozenset(
+    {
+        # returns the Raw with the CSD-transformed data
+        mne.preprocessing.compute_current_source_density,
+        # returns the Raw with the new bipolar reference channel
+        mne.set_bipolar_reference,
+        # returns the Raw with the OTP-cleaned data
+        mne.preprocessing.oversampled_temporal_projection,
+    }
+)
+
 
 def _is_standalone_function(func):
     """
@@ -79,7 +101,12 @@ def _generate_init_method(func, force_copy_false=False):
             raise TypeError(
                 f"{func_name}() missing required arguments: {', '.join(mandatory)}"
             )
-        Preprocessor.__init__(self, fn=func_name, apply_on_array=False, **init_kwargs)
+        # For standalone functions verified to return the modified instance,
+        # pass the callable itself so it is actually applied. Every other
+        # function keeps the legacy string-name behaviour (see #885 and
+        # ``_SAFE_STANDALONE_FUNCTIONS`` above).
+        fn = func if func in _SAFE_STANDALONE_FUNCTIONS else func_name
+        Preprocessor.__init__(self, fn=fn, apply_on_array=False, **init_kwargs)
 
     init_method.__signature__ = inspect.signature(func)
     return init_method

braindecode-1.6.1.dev1053/braindecode/version.py

@@ -0,0 +1 @@
+__version__ = "1.6.1.dev1053"

{braindecode-1.6.1.dev1049 → braindecode-1.6.1.dev1053/braindecode.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: braindecode
-Version: 1.6.1.dev1049
+Version: 1.6.1.dev1053
 Summary: Deep learning software to decode EEG, ECG or MEG signals
 Author-email: Robin Tibor Schirrmeister <robintibor@gmail.com>, Bruno Aristimunha Pinto <b.aristimunha@gmail.com>, Alexandre Gramfort <agramfort@meta.com>
 Maintainer-email: Alexandre Gramfort <agramfort@meta.com>, Bruno Aristimunha Pinto <b.aristimunha@gmail.com>, Robin Tibor Schirrmeister <robintibor@gmail.com>

{braindecode-1.6.1.dev1049 → braindecode-1.6.1.dev1053}/docs/whats_new.rst

@@ -70,6 +70,15 @@ Bug fixes
   use running statistics when ``batch_size == 1`` in train mode.
   By `Bruno Aristimunha`_.
 
+- Fix the auto-generated standalone-function preprocessors
+  (:class:`braindecode.preprocessing.ComputeCurrentSourceDensity`,
+  :class:`braindecode.preprocessing.SetBipolarReference`, and
+  :class:`braindecode.preprocessing.OversampledTemporalProjection`) passing the
+  function name as a string instead of the callable, so they failed to apply.
+  These functions return the modified instance and are now wrapped as callables.
+  Standalone functions that return auxiliary data (e.g. annotations or bad
+  channels) are intentionally left on the existing path for now. (:gh:`885` by
+  `Yiheng Li`_)
 - Fix incorrect import path in CONTRIBUTING.md by `Yiheng Li`_
 
 - Fix the broken EEGNeX quickstart snippet on the documentation landing page,
@@ -79,9 +88,23 @@ Bug fixes
   :class:`~braindecode.EEGClassifier` infer the signal dimensions from the data.
   By `Bhargav Kowshik`_
 
+- Fix :class:`~braindecode.EEGClassifier` and
+  :class:`~braindecode.EEGRegressor` silently skipping all training when the
+  training set is smaller than ``batch_size`` (with the default
+  ``iterator_train__drop_last=True``). Previously every batch was dropped and
+  the model was left untrained without any message; a :class:`UserWarning` is
+  now raised telling the user to lower ``batch_size`` or set
+  ``iterator_train__drop_last=False``.
+  (:gh:`1053` by `Bhargav Kowshik`_)
+
 Code health
 ============
 
+- Install CPU-only PyTorch wheels in the ``tests`` and ``docs`` CI
+  workflows via ``UV_TORCH_BACKEND=cpu``. GitHub runners have no GPU, so
+  the default CUDA build pulled ~1.8 GiB of unused ``nvidia-*`` wheels and
+  contributed to a disk-exhaustion crash. (:gh:`1054` by `Bhargav Kowshik`_)
+
 - Add a monthly scheduled workflow that cuts a stable PyPI release on
   the 1st of every month, complementing the existing per-push ``.devN``
   pipeline. (:gh:`1030` by `Bruno Aristimunha`_)

braindecode-1.6.1.dev1049/braindecode/version.py

@@ -1 +0,0 @@
-__version__ = "1.6.1.dev1049"