libcity.model.trajectory_loc_prediction.ATSTLSTM¶

class libcity.model.trajectory_loc_prediction.ATSTLSTM.ATSTLSTM(config, data_feature)[source]¶

Bases: libcity.model.abstract_model.AbstractModel

ATST_LSTM 轨迹下一跳预测模型.

calculate_loss(batch)[source]¶

计算模型损失（不包含正则项） :param batch: 输入 :type batch: libcity.data.batch

Returns: shape () 损失
Return type: (torch.tensor.Tensor)

forward(batch)[source]¶

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.

predict(batch)[source]¶

Parameters: batch (Batch) – a batch of input
Returns: predict result of this batch
Return type: torch.tensor

training: bool¶

class libcity.model.trajectory_loc_prediction.ATSTLSTM.Attn(hidden_size)[source]¶

Bases: torch.nn.modules.module.Module

Attention 注意力机制模块, 对 LSTM 中间层输出做加权平均.

forward(x)[source]¶

前向传播. :param x: shape (batch, seq_len, hidden_size) 中间层输出序列 :type x: torch.tensor.Tensor

Returns: shape (batch, size)
Return type: (torch.tensor.Tensor)

training: bool¶