the bug

[markwave]

when i train the Virconv-L, in the beginning, the train runs well ,when it runs serval epochs,it will ocuur the error in the below:
"epochs: 76%|???????? | 38/50 [19:53:52<6:18:46, 1893.87s/it, loss=2.13, lr=0.00328]
Traceback (most recent call last):
File "messagestream.pyx", line 91, in scipy._lib.messagestream.MessageStream.close
TypeError: remove: path should be string, bytes or os.PathLike, not NoneType
Exception ignored in: 'scipy._lib.messagestream.MessageStream.dealloc'
Traceback (most recent call last):
File "messagestream.pyx", line 91, in scipy._lib.messagestream.MessageStream.close
TypeError: remove: path should be string, bytes or os.PathLike, not NoneType
traintest.sh: line 3: 91725 Segmentation fault python3 train.py --cfg_file cfgs/models/kitti/VirConv-L.yaml --extra_tag testpro
",
please tell me how to fix this problem ?

[ckdghks]

Did you resolve this issue?
If so, could you share the solution?