Closed lucasjinreal closed 4 years ago
Would you provide more details of your run? E.g. config file and system environment? Thank you!
@jieli352 this error happens with torch 1.6 (I believe also 1.5) and root reason is accordingly torchvision which FPN used inside it.
The in_channel can not be 0 in some updates in torchvision, I don't know how to edit it. I am using default code without any modification.
Thanks for the information. The update in PyTorch here PR-1407 now requires filtered ResNet layers as FPN inputs. I have incorporated this change into our recent PR: #5 . That should unblock you from using PyTorch 1.6.
hi @jinfagang I was facing the same error and I resolved it using downgrading the torch and torchvision versions.
pip install future typing numpy awscli , pip install https://download.pytorch.org/whl/cu100/torch-1.1.0-cp36-cp36m-linux_x86_64.whl , pip install https://download.pytorch.org/whl/cu100/torchvision-0.3.0-cp36-cp36m-linux_x86_64.whl
run these commands then code starts working.
Hi! We just merged #5 so please let us know if you still have the issue.