birdx0810 / timegan-pytorch

This repository is a non-official implementation of TimeGAN (Yoon et al., NIPS2019) using PyTorch.
MIT License
80 stars 35 forks source link

关于参数 T #10

Closed luoxia-wen closed 1 year ago

luoxia-wen commented 1 year ago

请问, 如果以等长的时间序列输入到RNN,即max_seq_len=every_seq_len 还需要参数T吗?

birdx0810 commented 1 year ago

不需要。

luoxia-wen commented 1 year ago

谢谢

luoxia-wen commented 1 year ago

关于recovery,官方实现源代码中与supervisor一样最后会经过sigmoid,作者给出的理由是:转换为与输入一样的【0,1】的区间