EEGInception

class selfeeg.models.zoo.EEGInception(nb_classes: int, Chans: int, Samples: int, F1: int = 8, D: int = 2, kernel_size: int = 64, pool: int = 4, dropRate: float = 0.5, ELUalpha: float = 1.0, bias: bool = True, batch_momentum: float = 0.1, max_depth_norm: float = 1.0, return_logits: bool = True, seed: int = None)[source]

Pytorch Implementation of the EEGInception model. Original paper can be found here [eeginc] . A Keras implementation can be found here [eegincgit] .

The expected input is a 3D tensor with size (Batch x Channels x Samples).

Parameters:

  • nb_classes (int) – The number of classes. If less than 2, a binary classification problem is considered (output dimensions will be [batch, 1] in this case).

  • Chans (int) – The number of EEG channels.

  • Samples (int) – The sample length. It will be used to calculate the embedding size (for head initialization).

  • F1 (int, optional) –

    The number of filters in the first temporal convolutional layer. Other output filters will be calculated according to the paper specification.

    Default = 8

  • D (int, optional) –

    The depth of the depthwise convolutional layer.

    Default = 2

  • kernel_size (int, optional) –

    The length of the temporal convolutional layer.

    Default = 64

  • pool (int, optional) –

    The temporal pooling kernel size.

    Default = 4

  • dropRate (float, optional) –

    The dropout percentage in range [0,1].

    Default = 0.5

  • ELUalpha (float, optional) –

    The alpha value of the ELU activation function.

    Default = 1

  • bias (bool, optional) –

    If True, add a learnable bias to the output.

    Default = True

  • batch_momentum (float, optional) –

    The batch normalization momentum.

    Default = 0.9

  • max_depth_norm (float, optional) –

    The maximum norm each filter can have in the depthwise block. If None no constraint will be included.

    Default = 1.

  • return_logits (bool, optional) –

    Whether to return the output as logit or probability. It is suggested to not use False as the pytorch crossentropy applies the softmax internally.

    Default = True

  • seed (int, optional) –

    A custom seed for model initialization. It must be a nonnegative number. If None is passed, no custom seed will be set

    Default = None

References

[eeginc]

Zhang, Ce, Young-Keun Kim, and Azim Eskandarian. “EEG-inception: an accurate and robust end-to-end neural network for EEG-based motor imagery classification.” Journal of Neural Engineering 18.4 (2021): 046014. https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9311146

Example

>>> import selfeeg.models
>>> import torch
>>> x = torch.randn(4,8,64)
>>> mdl = models.EEGInception(4,8,64)
>>> out = mdl(x)
>>> print(out.shape) # shoud return torch.Size([4, 4])
>>> print(torch.isnan(out).sum()) # shoud return 0
forward(x)[source]