fishy.models.deep.mamba

Mamba model for spectral classification.

Classes

class fishy.models.deep.mamba.Mamba(input_dim: int, output_dim: int, hidden_dim: int = 128, num_layers: int = 4, dropout: float = 0.2, d_state: int = 16, d_conv: int = 4, expand: int = 2, **kwargs)

Bases: Module

Mamba-based model for spectral data classification.

__init__(input_dim: int, output_dim: int, hidden_dim: int = 128, num_layers: int = 4, dropout: float = 0.2, d_state: int = 16, d_conv: int = 4, expand: int = 2, **kwargs) → None

Initializes the Mamba model.

Parameters:

• input_dim (int) – Number of input features.

• output_dim (int) – Number of output classes.

• hidden_dim (int, optional) – Hidden layer dimension. Defaults to 128.

• num_layers (int, optional) – Number of Mamba layers. Defaults to 4.

• dropout (float, optional) – Dropout rate. Defaults to 0.2.

• d_state (int, optional) – State dimension. Defaults to 16.

• d_conv (int, optional) – Convolution kernel size. Defaults to 4.

• expand (int, optional) – Expansion factor. Defaults to 2.

forward(x: Tensor) → Tensor

Forward pass of the Mamba model.

class fishy.models.deep.mamba.MambaBlock(d_model: int, d_state: int = 16, d_conv: int = 4, expand: int = 2, dropout: float = 0.2, layer_norm_eps: float = 1e-05)

Bases: Module

Mamba block implementing the inner and outer functions of the Mamba model.

__init__(d_model: int, d_state: int = 16, d_conv: int = 4, expand: int = 2, dropout: float = 0.2, layer_norm_eps: float = 1e-05) → None

Initialize internal Module state, shared by both nn.Module and ScriptModule.

forward(x: Tensor) → Tensor

Forward pass of the Mamba block.

selective_scan(x, delta, B, C)

A placeholder for the selective scan operation.

class fishy.models.deep.mamba.SiameseMamba(input_dim: int, output_dim: int, hidden_dim: int = 128, num_layers: int = 4, dropout: float = 0.2, **kwargs)

Bases: Module

Siamese Mamba architecture for instance recognition.

__init__(input_dim: int, output_dim: int, hidden_dim: int = 128, num_layers: int = 4, dropout: float = 0.2, **kwargs) → None

Initialize internal Module state, shared by both nn.Module and ScriptModule.

forward(x1: Tensor, x2: Tensor) → Tensor | tuple

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

forward_one(x: Tensor) → Tensor

s