CUDA 9.0 support available?

[whikwon]

I'm using cuda 9.0, cudnn7 and conda_packagelist.txt says i have to install cudatoolkit, pytorch according to cuda 8.0, cudnn7.

As I don't want to reinstall the cuda things, I've tried docker but got errors.

AttributeError: 'NoneType' object has no attribute 'shape'

and

What should I do to use cuda 9.0?

[whikwon]

If I use cuda 9.0 and continue installing as conda_packagelist.txt, I get error during execute

cd <CornetNet dir>/models/py_utils/_cpools/
python setup.py install --user

[heilaw]

The first error 'Attribute Error' should have nothing to do with CUDA 9.0. You can see this post.

If you are using CUDA 9.0, you need to update PyTorch. I tried CornerNet with CUDA 9.0. The code should work fine.

[whikwon]

Thanks for reply!

Could you let me know which packages did you use for CUDA 9.0 instead of below?

https://conda.anaconda.org/caffe2/linux-64/caffe2-cuda8.0-cudnn7-0.8.dev-py36_2018.05.14.tar.bz2
https://repo.continuum.io/pkgs/free/linux-64/cudatoolkit-8.0-3.tar.bz2
https://conda.anaconda.org/pytorch/linux-64/pytorch-0.4.0-py36_cuda8.0.61_cudnn7.1.2_1.tar.bz2

[heilaw]

You can follow the instructions here to install PyTorch 0.4.1 w/ CUDA 9.0.

[guyubit91]

Thanks for reply!

Could you let me know which packages did you use for CUDA 9.0 instead of below?

https://conda.anaconda.org/caffe2/linux-64/caffe2-cuda8.0-cudnn7-0.8.dev-py36_2018.05.14.tar.bz2
https://repo.continuum.io/pkgs/free/linux-64/cudatoolkit-8.0-3.tar.bz2
https://conda.anaconda.org/pytorch/linux-64/pytorch-0.4.0-py36_cuda8.0.61_cudnn7.1.2_1.tar.bz2

hi，do u know how to do with cuda9.0?thanks

[Kangzf1996]

You can follow the instructions here to install PyTorch 0.4.1 w/ CUDA 9.0.

hello, when I install it, there will be a error. Can you help me? Thank you! loading model from cache/nnet/CenterNet-104/CenterNet-104_480000.pkl locating kps: 0%| | 0/5000 [00:00<?, ?it/s]/home/travail/miniconda3/envs/CornerNet/lib/python3.6/site-packages/torch/nn/modules/upsampling.py:122: UserWarning: nn.Upsampling is deprecated. Use nn.functional.interpolate instead. warnings.warn("nn.Upsampling is deprecated. Use nn.functional.interpolate instead.") Segmentation fault (core dumped)