libcity.model.traffic_speed_prediction.TGCN¶
-
class
libcity.model.traffic_speed_prediction.TGCN.TGCN(config, data_feature)[源代码]¶ 基类：
libcity.model.abstract_traffic_state_model.AbstractTrafficStateModel-
calculate_loss(batch)[源代码]¶ 输入一个batch的数据，返回训练过程的loss，也就是需要定义一个loss函数
- 参数
batch (Batch) – a batch of input
- 返回
return training loss
- 返回类型
torch.tensor
-
forward(batch)[源代码]¶ - 参数
batch –
a batch of input, batch[‘X’]: shape (batch_size, input_window, num_nodes, input_dim)
batch[‘y’]: shape (batch_size, output_window, num_nodes, output_dim)
- 返回
(batch_size, self.output_window, self.num_nodes, self.output_dim)
- 返回类型
torch.tensor
-
predict(batch)[源代码]¶ 输入一个batch的数据，返回对应的预测值，一般应该是**多步预测**的结果，一般会调用nn.Moudle的forward()方法
- 参数
batch (Batch) – a batch of input
- 返回
predict result of this batch
- 返回类型
torch.tensor
-
training: bool¶
-
-
class
libcity.model.traffic_speed_prediction.TGCN.TGCNCell(num_units, adj_mx, num_nodes, device, input_dim=1)[源代码]¶ 基类：
torch.nn.modules.module.Module-
_gc(inputs, state, output_size, bias_start=0.0)[源代码]¶ GCN
- 参数
inputs – (batch, self.num_nodes * self.dim)
state – (batch, self.num_nodes * self.gru_units)
output_size –
bias_start –
- 返回
(B, num_nodes , output_size)
- 返回类型
torch.tensor
-
forward(inputs, state)[源代码]¶ Gated recurrent unit (GRU) with Graph Convolution.
- 参数
inputs – shape (batch, self.num_nodes * self.dim)
state – shape (batch, self.num_nodes * self.gru_units)
- 返回
shape (B, num_nodes * gru_units)
- 返回类型
torch.tensor
-
training: bool¶
-