Closed NGA-MMDDsteven closed 3 years ago
I changed my pytorch version from 1.6.0 to 1.0.0. I have another error :
RuntimeError: could not compute gradients for some functions
I changed my pytorch version from 1.6.0 to 1.0.0. I have another error :
RuntimeError: could not compute gradients for some functions
I uninstall the odl and build the odl source file. Now the training is working!
git clone https://github.com/odlgroup/odl
cd odl
pip install --user --editable .
我觉得这个可以适当的在readme里面补充一下,还有就是开头提到的requirements.txt文件。
Can you provide requirements.txt? My pytorch version is 1.6.0, whose level seems too high for the repo.
Besides, I install odl in command : pip install odl
The function of odl seems having changed! Such as odl.tomo.FanBeamGeometry has changed to odl.tomo.FanFlatGeometry and odl_torch.OperatorModule has changed to odl_torch.OperatorAsModule.
I'm not sure the change is correct, please check!
First, thank you for your interest in our work !
Sorry for the missing. We have added the requirements.txt in our repo. Thanks!
For the installation of ODL, we have downloaded the repo from https://github.com/odlgroup/odl and then pip install.
I changed my pytorch version from 1.6.0 to 1.0.0. I have another error :
RuntimeError: could not compute gradients for some functions
Please first check whether you have installed the packages correctly.
I changed my pytorch version from 1.6.0 to 1.0.0. I have another error : RuntimeError: could not compute gradients for some functions
I uninstall the odl and build the odl source file. Now the training is working!
git clone https://github.com/odlgroup/odl
cd odl
pip install --user --editable .
我觉得这个可以适当的在readme里面补充一下,还有就是开头提到的requirements.txt文件。
谢谢您的提议,也感谢您对我们工作的关注,我们已经在repo中做了对应的注释。
Can you provide requirements.txt? My pytorch version is 1.6.0, whose level seems too high for the repo.
Besides, I install odl in command : pip install odl
The function of odl seems having changed! Such as odl.tomo.FanBeamGeometry has changed to odl.tomo.FanFlatGeometry and odl_torch.OperatorModule has changed to odl_torch.OperatorAsModule.
I'm not sure the change is correct, please check!