Lock error:matplotlib failed to acquire the following lock file

[willnotapostle]

I used mmclassification to calculate the data on the Beijing Supercomputer, and the training model is resnet18. I checked the log after submitting the calculation, and the following error occurred:"Lock error:matplotlib failed to acquire the following lock file: this may due to another process holding this lock file".
Here is the full log error：

My cuda version is 11.3, torch is 1.12.1, mmcv version is 1.7, the installation path corresponds to cuda and torch, and the environment configuration check is correct.
I saw some suggestions to delete the matplotlib folder and try again. However, we did not have permission to operate this file. How to solve this problem?

[Ezra-Yu]

we did not have permission to operate this file. How to solve this problem?

Can you install a new environment in your own directory?

[willnotapostle]

Yes, I can

[willnotapostle]

In the home directory, only the run directory is usable. So I don't know how to remove this base library file.

[mzr1996]

It seems to be a matplotlib issue, please try to delete the cache file according to https://stackoverflow.com/questions/71690325/timeouterror-lock-error-matplotlib-failed-to-acquire-the-following-lock-file
While if you cannot delete it, please get in touch with your administrator

[willnotapostle]

Thankyou! Delete it did work!