Series decomposition block

Series decomposition block

Series decomposition block

Compose with two layers

Special designed layernorm for the seasonal part

AutoCorrelation Mechanism with the following two phases: (1) period-based dependencies discovery (2) time delay aggregation This bloc

AutoCorrelation Mechanism with the following two phases: (1) period-based dependencies discovery (2) time delay aggregation This bloc

Autoformer encoder

Autoformer decoder layer with the progressive decomposition architecture

Compose with two layers

Autoformer encoder

Autoformer encoder layer with the progressive decomposition architecture

Multi-Head Attention module

1D multiwavelet block.

Two-layer position-wise feed-forward neural network.

Moving average block to highlight the trend of time series

Special designed layernorm for the seasonal part

Bottleneck convolution CSCM

A encoder model with self attention mechanism.

Autoformer encoder

Autoformer decoder layer with the progressive decomposition architecture

A encoder model with self attention mechanism.

A encoder model with self attention mechanism.

Autoformer encoder

Autoformer encoder layer with the progressive decomposition architecture

FEDformer performs the attention mechanism on frequency domain and achieved O(N) complexity

1D Multiwavelet Cross Attention layer.

Scaled Dot-Product Attention

Moving average block to highlight the trend of time series

Moving average block to highlight the trend of time series

Series decomposition block

Extremely slow, please sample < 0.1

Average pooling CSCM

Convolution CSCM