TimesNet由堆叠的TimesBlock组成,输入序列首先经过嵌入层得到深度特征X_{1D}^{0} \in \mathbb{R}^{T \times d_{model}} 对于第l层TimesBlock,其输入为X_{1D}^{l-1} \in \mathbb{R}^{T \times d_{model}},之后通过2D卷积提取二维时序变化:X_{1D}^{l} = TimesBlock(X_{1D}^{l-1}) + X...