← Back to API reference

neuraltrain.models.simplerconv.SimplerConv

pydantic model neuraltrain.models.simplerconv.SimplerConv

Convolutional encoder inspired by BENDR / wav2vec 2.0.

Parameters:

• subject_layers_config (SubjectLayers or None) – If set, prepend a per-subject linear projection.

• merger_config (ChannelMerger or None) – If set, prepend a ChannelMerger for multi-montage support.

• n_hidden_channels (int) – Number of hidden channels in each convolutional layer.

• kernel_sizes (tuple of int) – Kernel size for each convolutional layer.

• strides (tuple of int) – Stride for each convolutional layer. Must have the same length as kernel_sizes.

• output_nonlin (bool) – Apply GELU after the last convolutional layer.

• dropout (float) – Channel-dropout (Dropout1d) probability after each convolution.

Fields:

• dropout (float)

• kernel_sizes (tuple[int, ...])

• merger_config (neuraltrain.models.common.ChannelMerger | None)

• n_hidden_channels (int)

• output_nonlin (bool)

• strides (tuple[int, ...])

• subject_layers_config (neuraltrain.models.common.SubjectLayers | None)

field subject_layers_config: SubjectLayers | None = None

field merger_config: ChannelMerger | None = None

field n_hidden_channels: int = 512

field kernel_sizes: tuple[int, ...] = (3, 2, 2, 2, 2)

field strides: tuple[int, ...] = (3, 2, 2, 2, 2)

field output_nonlin: bool = True

field dropout: float = 0.0

build(n_in_channels: int, n_outputs: int | None = None) → Module

← Back to API reference