frank-xwang/towards-universal-object-detection

No module named 'model.nms._ext.nms._nms'

Closed this issue · 1 comments

Hello, I have tried your script.
Thanks for sharing.
It reports the error when I run scripts "sh scripts/train_universal.sh "

Error log is

File "universal_model.py", line 32, in
from model.faster_rcnn.resnet_uni import resnet
File "/root/junyan/object_detection/towards-universal-object-detection/lib/model/faster_rcnn/resnet_uni.py", line 6, in
from model.faster_rcnn.faster_rcnn_uni import _fasterRCNN
File "/root/junyan/object_detection/towards-universal-object-detection/lib/model/faster_rcnn/faster_rcnn_uni.py", line 10, in
from model.rpn.rpn_universal import _RPN
File "/root/junyan/object_detection/towards-universal-object-detection/lib/model/rpn/rpn_universal.py", line 8, in
from .proposal_layer import _ProposalLayer
File "/root/junyan/object_detection/towards-universal-object-detection/lib/model/rpn/proposal_layer.py", line 20, in
from model.nms.nms_wrapper import nms
File "/root/junyan/object_detection/towards-universal-object-detection/lib/model/nms/nms_wrapper.py", line 10, in
from model.nms.nms_gpu import nms_gpu
File "/root/junyan/object_detection/towards-universal-object-detection/lib/model/nms/nms_gpu.py", line 4, in
from ._ext import nms
File "/root/junyan/object_detection/towards-universal-object-detection/lib/model/nms/_ext/nms/init.py", line 3, in
from ._nms import lib as _lib, ffi as _ffi

Looks like path-setting problem, could you give some guide that the path could be searched correctly?

Thanks & Reagrds!

ModuleNotFoundError: No module named 'model.nms._ext.nms._nms'

Hi, this is cause by mismatched cuda version. You may need to make sure that the working environment is in cuda9.0/cuda9.1. Also, Pytorch need to be installed with cuda9.0/9.1 too.