search

akanazawa / cmr

Project repo for Learning Category-Specific Mesh Reconstruction from Image Collections
MIT License
474 stars 84 forks source link

What version of Cuda is required for installing? #6

Closed BornInWater closed 5 years ago

BornInWater commented 5 years ago

As in the title. I have cuda version 9.0 and this is causing a lot of problems in installing cupy. I installed it seperately using "pip install cupy-cuda90", removed it from requirements.txt and then installed everything else. When I try to install the external repos, i.e, Perceptual Loss and Neural Renderer, I got this Traceback: Traceback (most recent call last): File "setup.py", line 3, in import neural_renderer File "/home/nsajjan/WORK/Capstone/cmr/external/neural_renderer/neural_renderer/init.py", line 1, in from cross import cross File "/home/nsajjan/WORK/Capstone/cmr/external/neural_renderer/neural_renderer/cross.py", line 2, in import cupy as cp File "/home/nsajjan/WORK/Capstone/cmr/venv_cmr/lib/python2.7/site-packages/cupy/init.py", line 7, in from cupy import _version ImportError: cannot import name _version

Any suggestions?

shubhtuls commented 5 years ago

We used CUDA 8.0 for our experiments. Some users have been trying out other recent version and maybe their comments can help - see here.

BornInWater commented 5 years ago

Also, is pytorch 0.3 that only version compatible ? I was not able to solve the problem with CUDA 9 and hence shifted to CUDA 8. But I have tesla v100 and it seems to be horribly slow with cuda 8 and pytorch 0.3 combination. With Pytorch 1.0 it seems to work faster but I am getting this error when I run the first demo code

cupy.cuda.compiler.CompileException: nvrtc: error: failed to load builtins

BornInWater commented 5 years ago

Solved. With CUDA 9.0, Chainer 5.2.0, Cupy 5.2.0, Torch 1.0 and by installing Neural Mesh Renderer from most recent update.

YujieLu10 commented 4 years ago

2 errors detected in the compilation of "/tmp/../cubin.cu" solved by chainer==3.3.0 and cupy==2.3.0 with CUDA9.0 https://github.com/hiroharu-kato/neural_renderer/issues/12

dahyun-kang commented 4 years ago

Python 3.5, CUDA 10.1, chainer 3.3.0, cupy 2.3.0, and PyTorch 0.3.1 works as well

ShougangShen commented 4 years ago

python2.7, CUDA8.0, chainer 3.3.0, cupy 2.3.0, pytorch 0.3.1 works for me. Before running the demo command, I run: export PATH="$PATH:/usr/local/cuda-8.0/bin" ，export LD_LIBRARY_PATH="/usr/local/cuda-8.0/lib64" and reinstall the chainer and cupy.

felixboevr commented 4 years ago

Python 3.5, CUDA 10.1, chainer 3.3.0, cupy 2.3.0, and PyTorch 0.3.1 works as well

I get errors using that. What Version of Neural Mesh Renderer are you using?

dahyun-kang commented 4 years ago

Python 3.5, CUDA 10.1, chainer 3.3.0, cupy 2.3.0, and PyTorch 0.3.1 works as well

I get errors using that. What Version of Neural Mesh Renderer are you using?

It is 1.1.0.

felixboevr commented 4 years ago

Thank you, but here chainer==3.3.0 and cupy==2.3.0 both error out when trying pip install in python3.5 environment :/