braindecode 1.3.0.dev177069446__py3-none-any.whl → 1.3.0.dev177628147__py3-none-any.whl

braindecode/models/eeginception_erp.py

@@ -13,7 +13,7 @@ from braindecode.modules import DepthwiseConv2d, Ensure4d, InceptionBlock
 
 
 class EEGInceptionERP(EEGModuleMixin, nn.Sequential):
-    r"""EEG Inception for ERP-based from Santamaria-Vazquez et al (2020) [santamaria2020]_.
+    """EEG Inception for ERP-based from Santamaria-Vazquez et al (2020) [santamaria2020]_.
 
     :bdg-success:`Convolution`
 
@@ -176,7 +176,7 @@ class EEGInceptionERP(EEGModuleMixin, nn.Sequential):
         drop_prob=0.5,
         scales_samples_s=(0.5, 0.25, 0.125),
         n_filters=8,
-        activation: type[nn.Module] = nn.ELU,
+        activation: nn.Module = nn.ELU,
         batch_norm_alpha=0.01,
         depth_multiplier=2,
         pooling_sizes=(4, 2, 2, 2),

braindecode/models/eeginception_mi.py

@@ -11,7 +11,7 @@ from braindecode.modules import Ensure4d
 
 
 class EEGInceptionMI(EEGModuleMixin, nn.Module):
-    r"""EEG Inception for Motor Imagery, as proposed in Zhang et al. (2021) [1]_
+    """EEG Inception for Motor Imagery, as proposed in Zhang et al. (2021) [1]_
 
     :bdg-success:`Convolution`
 
@@ -71,7 +71,7 @@ class EEGInceptionMI(EEGModuleMixin, nn.Module):
         n_convs: int = 5,
         n_filters: int = 48,
         kernel_unit_s: float = 0.1,
-        activation: type[nn.Module] = nn.ReLU,
+        activation: nn.Module = nn.ReLU,
         chs_info=None,
         n_times=None,
     ):
@@ -208,7 +208,7 @@ class EEGInceptionMI(EEGModuleMixin, nn.Module):
 
 
 class _InceptionModuleMI(nn.Module):
-    r"""
+    """
     Inception module.
 
     This module implements a inception-like architecture that processes input
@@ -247,7 +247,7 @@ class _InceptionModuleMI(nn.Module):
         n_convs,
         kernel_unit_s=0.1,
         sfreq=250,
-        activation: type[nn.Module] = nn.ReLU,
+        activation: nn.Module = nn.ReLU,
     ):
         super().__init__()
         self.in_channels = in_channels
@@ -323,7 +323,7 @@ class _InceptionModuleMI(nn.Module):
 
 
 class _ResidualModuleMI(nn.Module):
-    r"""
+    """
     Residual module.
 
     This module performs a 1x1 convolution followed by batch normalization and an activation function.
@@ -343,7 +343,7 @@ class _ResidualModuleMI(nn.Module):
 
     """
 
-    def __init__(self, in_channels, n_filters, activation: type[nn.Module] = nn.ReLU):
+    def __init__(self, in_channels, n_filters, activation: nn.Module = nn.ReLU):
         super().__init__()
         self.in_channels = in_channels
         self.n_filters = n_filters

braindecode/models/eegitnet.py

@@ -9,7 +9,7 @@ from braindecode.modules import DepthwiseConv2d, Ensure4d, InceptionBlock
 
 
 class EEGITNet(EEGModuleMixin, nn.Sequential):
-    r"""EEG-ITNet from Salami, et al (2022) [Salami2022]_
+    """EEG-ITNet from Salami, et al (2022) [Salami2022]_
 
     :bdg-success:`Convolution` :bdg-secondary:`Recurrent`
 
@@ -83,7 +83,7 @@ class EEGITNet(EEGModuleMixin, nn.Sequential):
         tcn_padding: int = 3,
         drop_prob: float = 0.4,
         tcn_dilatation: int = 1,
-        activation: type[nn.Module] = nn.ELU,
+        activation: nn.Module = nn.ELU,
     ):
         super().__init__(
             n_outputs=n_outputs,
@@ -205,7 +205,7 @@ class EEGITNet(EEGModuleMixin, nn.Sequential):
         out_channels,
         kernel_length,
         depth_multiplier=1,
-        activation: type[nn.Module] = nn.ELU,
+        activation: nn.Module = nn.ELU,
     ):
         return nn.Sequential(
             nn.Conv2d(
@@ -229,7 +229,7 @@ class EEGITNet(EEGModuleMixin, nn.Sequential):
 
 
 class _TCBlock(nn.Module):
-    r"""
+    """
     Temporal Convolutional (TC) block.
 
     This module applies two depthwise separable convolutions with dilation and residual
@@ -260,7 +260,7 @@ class _TCBlock(nn.Module):
         dilatation,
         padding,
         drop_prob=0.4,
-        activation: type[nn.Module] = nn.ELU,
+        activation: nn.Module = nn.ELU,
     ):
         super().__init__()
         self.pad = padding

braindecode/models/eegminer.py

@@ -19,7 +19,7 @@ _eeg_miner_methods = ["mag", "corr", "plv"]
 
 
 class EEGMiner(EEGModuleMixin, nn.Module):
-    r"""EEGMiner from Ludwig et al (2024) [eegminer]_.
+    """EEGMiner from Ludwig et al (2024) [eegminer]_.
 
     :bdg-success:`Convolution` :bdg-warning:`Interpretability`
 

braindecode/models/eegnet.py

@@ -20,7 +20,7 @@ from braindecode.modules import (
 
 
 class EEGNet(EEGModuleMixin, nn.Sequential):
-    r"""EEGNet model from Lawhern et al (2018) [Lawhern2018]_.
+    """EEGNet model from Lawhern et al. (2018) [Lawhern2018]_.
 
     :bdg-success:`Convolution`
 
@@ -143,7 +143,7 @@ class EEGNet(EEGModuleMixin, nn.Sequential):
         pool1_kernel_size: int = 4,
         pool2_kernel_size: int = 8,
         conv_spatial_max_norm: int = 1,
-        activation: type[nn.Module] = nn.ELU,
+        activation: nn.Module = nn.ELU,
         batch_norm_momentum: float = 0.01,
         batch_norm_affine: bool = True,
         batch_norm_eps: float = 1e-3,
@@ -354,6 +354,6 @@ class EEGNet(EEGModuleMixin, nn.Sequential):
     "`EEGNetv4` was renamed to `EEGNet` in v1.12; this alias will be removed in v1.14."
 )
 class EEGNetv4(EEGNet):
-    r"""Deprecated alias for EEGNet."""
+    """Deprecated alias for EEGNet."""
 
     pass

braindecode/models/eegnex.py

@@ -14,7 +14,7 @@ from braindecode.modules import Conv2dWithConstraint, LinearWithConstraint
 
 
 class EEGNeX(EEGModuleMixin, nn.Module):
-    r"""EEGNeX model from Chen et al (2024) [eegnex]_.
+    """EEGNeX model from Chen et al. (2024) [eegnex]_.
 
     :bdg-success:`Convolution`
 
@@ -178,7 +178,7 @@ class EEGNeX(EEGModuleMixin, nn.Module):
         input_window_seconds=None,
         sfreq=None,
         # Model parameters
-        activation: type[nn.Module] = nn.ELU,
+        activation: nn.Module = nn.ELU,
         depth_multiplier: int = 2,
         filter_1: int = 8,
         filter_2: int = 32,

braindecode/models/eegsimpleconv.py

@@ -19,7 +19,7 @@ from braindecode.models.base import EEGModuleMixin
 
 
 class EEGSimpleConv(EEGModuleMixin, torch.nn.Module):
-    r"""EEGSimpleConv from Ouahidi, YE et al (2023) [Yassine2023]_.
+    """EEGSimpleConv from Ouahidi, YE et al. (2023) [Yassine2023]_.
 
     :bdg-success:`Convolution`
 
@@ -107,7 +107,7 @@ class EEGSimpleConv(EEGModuleMixin, torch.nn.Module):
         resampling_freq=80,
         kernel_size=8,
         return_feature=False,
-        activation: type[nn.Module] = nn.ReLU,
+        activation: nn.Module = nn.ReLU,
         # Other ways to initialize the model
         chs_info=None,
         n_times=None,

braindecode/models/eegsym.py

@@ -14,7 +14,7 @@ from braindecode.models.base import EEGModuleMixin
 
 
 class EEGSym(EEGModuleMixin, nn.Module):
-    r"""EEGSym from Pérez-Velasco et al (2022) [eegsym2022]_.
+    """EEGSym from Pérez-Velasco et al (2022) [eegsym2022]_.
 
     :bdg-success:`Convolution` :bdg-dark-line:`Channel`
 
@@ -430,7 +430,7 @@ class EEGSym(EEGModuleMixin, nn.Module):
 
 
 class _InceptionBlock(nn.Module):
-    r"""Inception module used in EEGSym architecture.
+    """Inception module used in EEGSym architecture.
 
     Parameters
     ----------
@@ -543,7 +543,7 @@ class _InceptionBlock(nn.Module):
 
 
 class _ResidualBlock(nn.Module):
-    r"""Residual block used in EEGSym architecture.
+    """Residual block used in EEGSym architecture.
 
     Parameters
     ----------
@@ -652,7 +652,7 @@ class _ResidualBlock(nn.Module):
 
 
 class _TemporalBlock(nn.Module):
-    r"""Temporal reduction block used in EEGSym architecture.
+    """Temporal reduction block used in EEGSym architecture.
 
     Parameters
     ----------
@@ -706,7 +706,7 @@ class _TemporalBlock(nn.Module):
 
 
 class _ChannelMergingBlock(nn.Module):
-    r"""Channel merging block used in EEGSym architecture.
+    """Channel merging block used in EEGSym architecture.
 
     This block performs hemisphere merging through:
     1. Two residual convolution iterations (with full spatial kernel)
@@ -791,7 +791,7 @@ class _ChannelMergingBlock(nn.Module):
 
 
 class _TemporalMergingBlock(nn.Module):
-    r"""Temporal merging block used in EEGSym architecture.
+    """Temporal merging block used in EEGSym architecture.
 
     This block performs temporal dimension collapse through:
     1. One residual convolution (temporal collapse with residual connection)
@@ -869,7 +869,7 @@ class _TemporalMergingBlock(nn.Module):
 
 
 class _OutputBlock(nn.Module):
-    r"""Output block used in EEGSym architecture.
+    """Output block used in EEGSym architecture.
 
     Parameters
     ----------

braindecode/models/eegtcnet.py

@@ -13,7 +13,7 @@ from braindecode.modules import Chomp1d, MaxNormLinear
 
 
 class EEGTCNet(EEGModuleMixin, nn.Module):
-    r"""EEGTCNet model from Ingolfsson et al (2020) [ingolfsson2020]_.
+    """EEGTCNet model from Ingolfsson et al. (2020) [ingolfsson2020]_.
 
     :bdg-success:`Convolution` :bdg-secondary:`Recurrent`
 
@@ -63,7 +63,7 @@ class EEGTCNet(EEGModuleMixin, nn.Module):
         input_window_seconds=None,
         sfreq=None,
         # Model parameters
-        activation: type[nn.Module] = nn.ELU,
+        activation: nn.Module = nn.ELU,
         depth_multiplier: int = 2,
         filter_1: int = 8,
         kern_length: int = 64,
@@ -157,7 +157,7 @@ class EEGTCNet(EEGModuleMixin, nn.Module):
 
 
 class _EEGNetTC(nn.Module):
-    r"""EEGNet Temporal Convolutional Network (TCN) block.
+    """EEGNet Temporal Convolutional Network (TCN) block.
 
     The main difference from our :class:`EEGNet` (braindecode) implementation is the
     kernel and dimensional order. Because of this, we decided to keep this
@@ -187,7 +187,7 @@ class _EEGNetTC(nn.Module):
         kern_length: int = 64,
         depth_multiplier: int = 2,
         drop_prob: float = 0.5,
-        activation: type[nn.Module] = nn.ELU,
+        activation: nn.Module = nn.ELU,
     ):
         super().__init__()
         self.activation = activation()
@@ -260,7 +260,7 @@ class _EEGNetTC(nn.Module):
 
 
 class _TCNBlock(nn.Module):
-    r"""
+    """
     Many differences from our Temporal Block (braindecode) implementation.
     Because of this, we decided to keep this implementation in a future issue;
     we will re-evaluate if it is necessary to maintain this separate
@@ -276,7 +276,7 @@ class _TCNBlock(nn.Module):
         kernel_size: int,
         filters: int,
         drop_prob: float,
-        activation: type[nn.Module] = nn.ELU,
+        activation: nn.Module = nn.ELU,
     ):
         super().__init__()
         self.activation = activation()

braindecode/models/fbcnet.py

@@ -29,7 +29,7 @@ _valid_layers = {
 
 
 class FBCNet(EEGModuleMixin, nn.Module):
-    r"""FBCNet from Mane, R et al (2021) [fbcnet2021]_.
+    """FBCNet from Mane, R et al (2021) [fbcnet2021]_.
 
     :bdg-success:`Convolution` :bdg-primary:`Filterbank`
 
@@ -97,7 +97,7 @@ class FBCNet(EEGModuleMixin, nn.Module):
         temporal_layer: str = "LogVarLayer",
         n_dim: int = 3,
         stride_factor: int = 4,
-        activation: type[nn.Module] = nn.SiLU,
+        activation: nn.Module = nn.SiLU,
         linear_max_norm: float = 0.5,
         cnn_max_norm: float = 2.0,
         filter_parameters: dict[Any, Any] | None = None,

braindecode/models/fblightconvnet.py

@@ -16,7 +16,7 @@ from braindecode.modules import (
 
 
 class FBLightConvNet(EEGModuleMixin, nn.Module):
-    r"""LightConvNet from Ma, X et al (2023) [lightconvnet]_.
+    """LightConvNet from Ma, X et al (2023) [lightconvnet]_.
 
     :bdg-success:`Convolution` :bdg-primary:`Filterbank`
 
@@ -111,7 +111,7 @@ class FBLightConvNet(EEGModuleMixin, nn.Module):
         heads: int = 8,
         weight_softmax: bool = True,
         bias: bool = False,
-        activation: type[nn.Module] = nn.ELU,
+        activation: nn.Module = nn.ELU,
         verbose: bool = False,
         filter_parameters: Optional[dict] = None,
     ):
@@ -235,7 +235,7 @@ class FBLightConvNet(EEGModuleMixin, nn.Module):
 
 
 class _LightweightConv1d(nn.Module):
-    r"""Lightweight 1D Convolution Module.
+    """Lightweight 1D Convolution Module.
 
     Applies a convolution operation with multiple heads, allowing for
     parallel filter applications. Optionally applies a softmax normalization

braindecode/models/fbmsnet.py

@@ -17,7 +17,7 @@ from braindecode.modules import (
 
 
 class FBMSNet(EEGModuleMixin, nn.Module):
-    r"""FBMSNet from Liu et al (2022) [fbmsnet]_.
+    """FBMSNet from Liu et al (2022) [fbmsnet]_.
 
     :bdg-success:`Convolution` :bdg-primary:`Filterbank`
 
@@ -110,7 +110,7 @@ class FBMSNet(EEGModuleMixin, nn.Module):
         n_dim: int = 3,
         stride_factor: int = 4,
         dilatability: int = 8,
-        activation: type[nn.Module] = nn.SiLU,
+        activation: nn.Module = nn.SiLU,
         kernels_weights: Sequence[int] = (15, 31, 63, 125),
         cnn_max_norm: float = 2,
         linear_max_norm: float = 0.5,
@@ -262,7 +262,7 @@ class FBMSNet(EEGModuleMixin, nn.Module):
 
 
 class _MixedConv2d(nn.Module):
-    r"""Mixed Grouped Convolution for multiscale feature extraction."""
+    """Mixed Grouped Convolution for multiscale feature extraction."""
 
     def __init__(
         self,

braindecode/models/hybrid.py

@@ -11,7 +11,7 @@ from braindecode.models.shallow_fbcsp import ShallowFBCSPNet
 
 
 class HybridNet(nn.Module):
-    r"""Hybrid ConvNet model from Schirrmeister, R T et al (2017)  [Schirrmeister2017]_.
+    """Hybrid ConvNet model from Schirrmeister, R T et al (2017)  [Schirrmeister2017]_.
 
     See [Schirrmeister2017]_ for details.
 
@@ -34,7 +34,7 @@ class HybridNet(nn.Module):
         input_window_seconds=None,
         sfreq=None,
         chs_info=None,
-        activation: type[nn.Module] = nn.ELU,
+        activation: nn.Module = nn.ELU,
         drop_prob: float = 0.5,
     ):
         super().__init__()

braindecode/models/ifnet.py

@@ -29,7 +29,7 @@ from braindecode.modules import (
 
 
 class IFNet(EEGModuleMixin, nn.Module):
-    r"""IFNetV2 from Wang J et al (2023) [ifnet]_.
+    """IFNetV2 from Wang J et al (2023) [ifnet]_.
 
     :bdg-success:`Convolution` :bdg-primary:`Filterbank`
 
@@ -238,9 +238,9 @@ class IFNet(EEGModuleMixin, nn.Module):
 
 
 class _InterFrequencyModule(nn.Module):
-    r"""Module that combines outputs from different frequency bands."""
+    """Module that combines outputs from different frequency bands."""
 
-    def __init__(self, activation: type[nn.Module] = nn.GELU):
+    def __init__(self, activation: nn.Module = nn.GELU):
         """
 
         Parameters
@@ -272,7 +272,7 @@ class _InterFrequencyModule(nn.Module):
 
 
 class _SpatioTemporalFeatureBlock(nn.Module):
-    r"""SpatioTemporal Feature Block consisting of spatial and temporal convolutions."""
+    """SpatioTemporal Feature Block consisting of spatial and temporal convolutions."""
 
     def __init__(
         self,
@@ -283,7 +283,7 @@ class _SpatioTemporalFeatureBlock(nn.Module):
         stride_factor: int = 8,
         n_bands: int = 2,
         drop_prob: float = 0.5,
-        activation: type[nn.Module] = nn.GELU,
+        activation: nn.Module = nn.GELU,
         dim: int = 3,
     ):
         """