TCN时间卷积网络-处理时间序列数据的网络结构

Temporal Convolutional Network

 

Causal Convolutions

灵活的感受视野
比如改变卷积核的参数，容易控制模型的记忆大小；

Receptive field

• Receptive field = nb_stacks_of_residuals_blocks * kernel_size * last_dilation.
• If a TCN has only one stack of residual blocks with a kernel size of 2 and dilations [1, 2, 4, 8], its receptive field is 2 * 1 * 8 = 16. The image below illustrates it:

KERAS-TCN

 

https://github.com/philipperemy/keras-tcn