lgatr.layers.mlp.config.MLPConfig

class lgatr.layers.mlp.config.MLPConfig(mv_channels=None, s_channels=0, dropout_prob=None, nonlinearity='gelu', mlp_ratio=4, num_layers_mlp=2)

Bases: object

Geometric-MLP configuration.

Parameters:

• mv_channels (int | None) – Number of input multivector channels. Set automatically by the parent network.

• s_channels (int) – Number of input scalar channels. Use 0 for no scalar stream. Set automatically by the parent network.

• dropout_prob (float | None) – Dropout probability. Set automatically by the parent network.

• nonlinearity (str) – Which (gated) activation function to use. One of "relu", "sigmoid", "tanh", "gelu", "silu".

• mlp_ratio (int) – Factor by which to increase the number of hidden channels (both multivectors and scalars).

• num_layers_mlp (int) – Total number of layers, including input and output layers (must be >= 1).

classmethod cast(config)

Cast an MLPConfig or mapping to an MLPConfig.

Return type:

MLPConfig

dropout_prob: float | None = None

mlp_ratio: int = 4

mv_channels: int | None = None

nonlinearity: str = 'gelu'

num_layers_mlp: int = 2

s_channels: int = 0