dymad.modules.helpers

Module Attributes

ACT_MAP	Mapping of activation names to activation classes.
GCL_MAP	Mapping of graph convolutional layer names to classes.
INIT_MAP_W	Mapping of weight initialization names to functions.
INIT_MAP_B	Mapping of bias initialization names to functions.

dymad.modules.helpers.ACT_MAP: dict[str, type[Module]] = {'elu': <class 'torch.nn.modules.activation.ELU'>, 'gelu': <class 'torch.nn.modules.activation.GELU'>, 'leakyrelu': <class 'torch.nn.modules.activation.LeakyReLU'>, 'mish': <class 'torch.nn.modules.activation.Mish'>, 'none': <class 'torch.nn.modules.linear.Identity'>, 'prelu': <class 'torch.nn.modules.activation.PReLU'>, 'relu': <class 'torch.nn.modules.activation.ReLU'>, 'selu': <class 'torch.nn.modules.activation.SELU'>, 'sigmoid': <class 'torch.nn.modules.activation.Sigmoid'>, 'silu': <class 'torch.nn.modules.activation.SiLU'>, 'softplus': <class 'torch.nn.modules.activation.Softplus'>, 'swish': <class 'torch.nn.modules.activation.SiLU'>, 'tanh': <class 'torch.nn.modules.activation.Tanh'>}

Mapping of activation names to activation classes.

dymad.modules.helpers.GCL_MAP: dict[str, type[Module] | None] = {'cheb': <class 'torch_geometric.nn.conv.cheb_conv.ChebConv'>, 'gat': <class 'torch_geometric.nn.conv.gat_conv.GATConv'>, 'gcn': <class 'torch_geometric.nn.conv.gcn_conv.GCNConv'>, 'gcnv': <class 'torch_geometric.nn.conv.graph_conv.GraphConv'>, 'sage': <class 'torch_geometric.nn.conv.sage_conv.SAGEConv'>}

Mapping of graph convolutional layer names to classes.

dymad.modules.helpers.INIT_MAP_B: dict[str, Callable[[...], Any]] = {'ones': <function ones_>, 'zeros': <function zeros_>}

Mapping of bias initialization names to functions.

dymad.modules.helpers.INIT_MAP_W: dict[str, Callable[[...], Any]] = {'kaiming_normal': <function kaiming_normal_>, 'kaiming_uniform': <function kaiming_uniform_>, 'normal': <function normal_>, 'orthogonal': <function orthogonal_>, 'trunc_normal': <function trunc_normal_>, 'uniform': <function uniform_>, 'xavier_normal': <function xavier_normal_>, 'xavier_uniform': <function xavier_uniform_>}

Mapping of weight initialization names to functions.