train_problem

[jtlxlf]

I try to train with your code, and show that only five g trian sharp and 300MB trian sharp bicubic are used as data sets, but later there is a problem that there are too few data sets. Later, I set GPUs to 1, which still means that there are too few data sets. I want to ask what the reason may be， and whether the 3090 single graphics card or the 3080ti dual graphics card can be traine d

[jtlxlf]

Is the data set missing or misconfigure d

[SongYxing]

I want to ask you how to solve the CUDA version problem when you use 3090 for training. When the CUDA version is below 11.0, I use 3090 to report an error. The error information is as follows: importerror: libcudart. So. 10.1: cannot open shared object file: no such file or directory

[jtlxlf]

It seems that the 30 series graphics card must use the version above cuda11, and the lower version can not be used.

---Original--- From: @.> Date: Sat, Aug 13, 2022 15:49 PM To: @.>; Cc: @.**@.>; Subject: Re: [ckkelvinchan/RealBasicVSR] train_problem (Issue #63)

I want to ask you how to solve the CUDA version problem when you use 3090 for training. When the CUDA version is below 11.0, I use 3090 to report an error. The error information is as follows: importerror: libcudart. So. 10.1: cannot open shared object file: no such file or directory

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>

[fmcfmcfmc]

Hello, I have also encountered the same problem. I set GPU=1. Have you solved this problem? Thank you very much! @jtlxlf