Import error during training

[albusdemens]

Hi, when I train the model as recommended in the README (python train.py --name pose --dataset_mode fewshot_pose --adaptive_spade --warp_ref --spade_combine --remove_face_labels --add_face_D --niter_single 100 --niter 200 --batchSize 2), I get the error import resample2d_cuda due to an undefined symbol (full output below). Does anyone have tips on how to fix that? I followed the recommended steps to compile a snapshot of FlowNet2.

File "train.py", line 34, in train
    model, flowNet = create_model(opt, trainer.start_epoch)
  File "/home/alberto/Documents/few-shot-vid2vid/models/models.py", line 27, in create_model
    flowNet.initialize(opt)
  File "/home/alberto/Documents/few-shot-vid2vid/models/flownet.py", line 23, in initialize
    from .networks.flownet2_pytorch import models as flownet2_models
  File "/home/alberto/Documents/few-shot-vid2vid/models/networks/flownet2_pytorch/models.py", line 8, in <module>
    from .networks.resample2d_package.resample2d import Resample2d
  File "/home/alberto/Documents/few-shot-vid2vid/models/networks/flownet2_pytorch/networks/resample2d_package/resample2d.py", line 3, in <module>
    import resample2d_cuda
ImportError: /home/alberto/.local/lib/python3.6/site-packages/resample2d_cuda-0.0.0-py3.6-linux-x86_64.egg/resample2d_cuda.cpython-36m-x86_64-linux-gnu.so: undefined symbol: _ZN3c105ErrorC1ENS_14SourceLocationERKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEE

[thefattestcat]

I have this issue as well. Have you tried this?
NVIDIA/vid2vid#30 (comment)

[albusdemens]

Hi, I tried following the commands in the link but they don't fix the issue. In detail, python3 setup.py build gives the errors reported here (error: command '/usr/bin/nvcc' failed with exit status 1).