Getting error while running ./run_evaluate.sh

[saket-m]

Traceback (most recent call last):
File "evaluate.py", line 6, in
from networks.CE2P import Res_Deeplab
File "/home/saket/vision/semantic-segmentation/CE2P/networks/CE2P.py", line 13, in
from libs import InPlaceABN, InPlaceABNSync
File "/home/saket/vision/semantic-segmentation/CE2P/libs/init.py", line 1, in
from .bn import ABN, InPlaceABN, InPlaceABNWrapper, InPlaceABNSync, InPlaceABNSyncWrapper
File "/home/saket/vision/semantic-segmentation/CE2P/libs/bn.py", line 15, in
from .functions import inplace_abn, inplace_abn_sync
File "/home/saket/vision/semantic-segmentation/CE2P/libs/functions.py", line 5, in
from . import _ext
File "/home/saket/vision/semantic-segmentation/CE2P/libs/_ext/init.py", line 3, in
from .__ext import lib as _lib, ffi as _ffi
ImportError: /home/saket/vision/semantic-segmentation/CE2P/libs/_ext/__ext.so: undefined symbol: __cudaPopCallConfiguration

I've followed the steps mentioned in the README.md file and it seems I've not compiled the file properly.
I'm using pytorch 0.4.1.
Please help me.

[liutinglt]

@saket-m I just test it on cuda 9.0, you can try to use the same cuda version