libcity.model.road_representation.GeomGCN

class libcity.model.road_representation.GeomGCN.GeomGCN(config, data_feature)

Bases: libcity.model.abstract_traffic_state_model.AbstractTrafficStateModel

calculate_loss(batch)

Parameters

batch – dict, need key ‘node_features’, ‘node_labels’, ‘mask’

Returns:

forward(batch)

自回归任务

Parameters

batch – dict, need key ‘node_features’ contains tensor shape=(N, feature_dim)

Returns

N, output_classes

Return type

torch.tensor

get_input(config, data_feature)

predict(batch)

Parameters

batch – dict, need key ‘node_features’

Returns

torch.tensor

training: bool

class libcity.model.road_representation.GeomGCN.GeomGCNNet(g, in_feats, out_feats, num_divisions, activation, num_heads, dropout_prob, ggcn_merge, channel_merge, device)

Bases: torch.nn.modules.module.Module

forward(feature)

Defines 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.

training: bool

class libcity.model.road_representation.GeomGCN.GeomGCNSingleChannel(g, in_feats, out_feats, num_divisions, activation, dropout_prob, merge, device)

Bases: torch.nn.modules.module.Module

forward(feature)

Defines 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.

get_node_subgraphs(g)

get_subgraphs(g)

training: bool