search

MAZiqing / FEDformer

MIT License
641 stars 122 forks source link

请问对于序列分解所使用的多个移动平均核体现在哪里 #76

Closed ooooooover closed 9 months ago

ooooooover commented 9 months ago

我唯一看到的默认设置moving_avg的地方是在run.py的第64行，但这里只使用了一个移动平均核 QQ截图20240116163531 如果在其他地方进行了设置感谢指明

tianzhou2011 commented 9 months ago

嗯，这个可以参数自己设置的，上传的代码default参数就写了一个核，加多个会变慢一点

On Tue, Jan 16, 2024 at 4:36 PM ooooooover @.***> wrote:

我唯一看到的默认设置moving_avg的地方是在run.py的第64行，但这里只使用了一个移动平均核 QQ.20240116163531.png (view on web) https://github.com/MAZiqing/FEDformer/assets/106066982/cbaba784-fa98-4ac0-a626-a15b9dffc789 如果在其他地方进行了设置感谢指明

— Reply to this email directly, view it on GitHub https://github.com/MAZiqing/FEDformer/issues/76, or unsubscribe https://github.com/notifications/unsubscribe-auth/AB3JGO4X2K3Z73MCRIR625LYOY3YZAVCNFSM6AAAAABB4MZMHKVHI2DSMVQWIX3LMV43ASLTON2WKOZSGA4DGMZRGIZTSNQ . You are receiving this because you are subscribed to this thread.Message ID: @.***>

ooooooover commented 9 months ago

嗯，这个可以参数自己设置的，上传的代码default参数就写了一个核，加多个会变慢一点 … On Tue, Jan 16, 2024 at 4:36 PM ooooooover @.> wrote: 我唯一看到的默认设置moving_avg的地方是在run.py的第64行，但这里只使用了一个移动平均核 QQ.20240116163531.png (view on web) https://github.com/MAZiqing/FEDformer/assets/106066982/cbaba784-fa98-4ac0-a626-a15b9dffc789 如果在其他地方进行了设置感谢指明 — Reply to this email directly, view it on GitHub <#76>, or unsubscribe https://github.com/notifications/unsubscribe-auth/AB3JGO4X2K3Z73MCRIR625LYOY3YZAVCNFSM6AAAAABB4MZMHKVHI2DSMVQWIX3LMV43ASLTON2WKOZSGA4DGMZRGIZTSNQ . You are receiving this because you are subscribed to this thread.Message ID: @.>

ok，在附录里看到了效果对比，看起来是一个trade-off的过程，效果提升并不明显。感谢回复