RuntimeError: This function was deprecated since version 1.9 and is now removed.

[s0966066980]

Can anyone solve the problem of normal training when pytorch>1.9?

[TimSC]

I fixed this problem in a separate branch but I don't know if STAR is adding torch 2 support officially.

026d90e

I guess we could just put in this single issue?

Update: Ah this fix was merged into STAR. #11 It should work now?

[s0966066980]

我在一個單獨的分支中修復了這個問題，但我不知道 STAR 是否正式添加了 torch 2 支援。

026d90e

我想我們可以只放入這一期嗎？

更新：啊這個修復已合併到 STAR 中。#11現在應該可以了嗎？

I will later use the torch=1.8 version and replace it with cuda=11.4 because there are too many problems in cuda11.8. Thank you for your help.