gluonts.mx.block.enc2dec 模块

class gluonts.mx.block.enc2dec.FutureFeatIntegratorEnc2Dec(**kwargs)

基类: gluonts.mx.block.enc2dec.Seq2SeqEnc2Dec

将 encoder_output_dynamic 和 future_features_dynamic 集成在一起，并将它们作为动态输入传递给解码器。

hybrid_forward(F, encoder_output_static: Union[mxnet.ndarray.ndarray.NDArray, mxnet.symbol.symbol.Symbol], encoder_output_dynamic: Union[mxnet.ndarray.ndarray.NDArray, mxnet.symbol.symbol.Symbol], future_features_dynamic: Union[mxnet.ndarray.ndarray.NDArray, mxnet.symbol.symbol.Symbol]) → Tuple[Union[mxnet.ndarray.ndarray.NDArray, mxnet.symbol.symbol.Symbol], Union[mxnet.ndarray.ndarray.NDArray, mxnet.symbol.symbol.Symbol]]

参数

• encoder_output_static – 形状为 (batch_size, channels_seq[-1] + 1) 或 (N, C)

• encoder_output_dynamic – 形状为 (batch_size, sequence_length, channels_seq[-1] + 1) 或 (N, T, C)

• future_features_dynamic – 形状为 (batch_size, sequence_length, prediction_length=decoder_length, num_feat_dynamic) 或 (N, T, P, C`)

返回

• 张量 – 形状为 (batch_size, channels_seq[-1] + 1) 或 (N, C)

• 张量 – 形状为 (batch_size, prediction_length=decoder_length, channels_seq[-1] + 1 + decoder_length * num_feat_dynamic) 或 (N, T, C)

class gluonts.mx.block.enc2dec.PassThroughEnc2Dec(**kwargs)

基类: gluonts.mx.block.enc2dec.Seq2SeqEnc2Dec

用于将编码器张量传递给解码器的最简单类。

传递张量，但 future_features_dynamic 被丢弃。

hybrid_forward(F, encoder_output_static: Union[mxnet.ndarray.ndarray.NDArray, mxnet.symbol.symbol.Symbol], encoder_output_dynamic: Union[mxnet.ndarray.ndarray.NDArray, mxnet.symbol.symbol.Symbol], future_features_dynamic: Union[mxnet.ndarray.ndarray.NDArray, mxnet.symbol.symbol.Symbol]) → Tuple[Union[mxnet.ndarray.ndarray.NDArray, mxnet.symbol.symbol.Symbol], Union[mxnet.ndarray.ndarray.NDArray, mxnet.symbol.symbol.Symbol]]

参数

• encoder_output_static – 形状为 (batch_size, channels_seq[-1] + 1) 或 (N, C)

• encoder_output_dynamic – 形状为 (batch_size, sequence_length, channels_seq[-1] + 1) 或 (N, T, C)

• future_features_dynamic – 形状为 (batch_size, sequence_length, prediction_length=decoder_length, num_feat_dynamic) 或 (N, T, P, C`)

返回

• 张量 – 形状为 (batch_size, channels_seq[-1] + 1) 或 (N, C)

• 张量 – 形状为 (batch_size, sequence_length, channels_seq[-1] + 1) 或 (N, T, C)

class gluonts.mx.block.enc2dec.Seq2SeqEnc2Dec(**kwargs)

基类: mxnet.gluon.block.HybridBlock

用于将编码器传递给解码器（例如注意力网络）的任何模块的抽象类。

hybrid_forward(F, encoder_output_static: Union[mxnet.ndarray.ndarray.NDArray, mxnet.symbol.symbol.Symbol], encoder_output_dynamic: Union[mxnet.ndarray.ndarray.NDArray, mxnet.symbol.symbol.Symbol], future_features_dynamic: Union[mxnet.ndarray.ndarray.NDArray, mxnet.symbol.symbol.Symbol]) → Tuple[Union[mxnet.ndarray.ndarray.NDArray, mxnet.symbol.symbol.Symbol], Union[mxnet.ndarray.ndarray.NDArray, mxnet.symbol.symbol.Symbol]]

参数

• encoder_output_static – 形状为 (batch_size, channels_seq[-1] + 1) 或 (N, C)

• encoder_output_dynamic – 形状为 (batch_size, sequence_length, channels_seq[-1] + 1) 或 (N, T, C)

• future_features_dynamic – 形状为 (batch_size, sequence_length, prediction_length=decoder_length, num_feat_dynamic) 或 (N, T, P, C`)