EasyVolcap icon indicating copy to clipboard operation
EasyVolcap copied to clipboard

Error in 'mamba env update'

Open r530044129 opened this issue 1 year ago • 4 comments

When I prepared the env, I met this error:

(easyvolcap) D:\Git\4K4D>mamba env update
pytorch/win-64                                              Using cache
pytorch/noarch                                              Using cache
nvidia/win-64                                               Using cache
nvidia/noarch                                               Using cache
conda-forge/win-64                                          Using cache
conda-forge/noarch                                          Using cache
pkgs/msys2/win-64                                             No change
pkgs/msys2/noarch                                             No change
pkgs/main/win-64                                              No change
pkgs/main/noarch                                              No change
pkgs/r/noarch                                                 No change
pkgs/r/win-64                                                 No change


Looking for: ["python[version='>=3.9']", "pytorch[version='>=1.13']", "pytorch-cuda[version='>=11.6']", 'torchvision', ]


Could not solve for environment specs
The following packages are incompatible
├─ ncurses does not exist (perhaps a typo or a missing channel);
└─ tmux does not exist (perhaps a typo or a missing channel).

(easyvolcap) D:\Git\4K4D>pip install ncurses
ERROR: Could not find a version that satisfies the requirement ncurses (from versions: none)
ERROR: No matching distribution found for ncurses

r530044129 avatar Jan 04 '24 03:01 r530044129

Here is window, does this cmd only support linux?

r530044129 avatar Jan 04 '24 03:01 r530044129

Hi @r530044129 . You're right, the mamba env update command only supports Linux for now. Our official installation instructions are tailored for Linux but we're working hard on adding more documentations about dependencies. In the future we might introduce new systems for managing dependencies.

On the other hand, our dependency demands are very loose and I do personally use a Windows machine for the viewer. The only core requirement is pytorch, thus you can try installing pytorch following their official instruction in any way. After this you might want to install missing packages if it's related to the algorithm or module you want to run.

For example, if you're using the real-time viewer, opengl-related packages should be installed:

glfw
PyGLM
pyperclip
pyopengl
imgui-bundle
opencv-python
cuda-python
pdbr

Use the command evc-gui to test your GUI installation. You might see warnings about missing imports from some specific modules but they can be ignored as long as they don't interfere with your usage. (for example, pytorch3d)

dendenxu avatar Jan 04 '24 06:01 dendenxu

Hi @lixiangweiabc , I updated the installation guide to include more description about Windows. Could you please check whether it's easier to follow now?

dendenxu avatar Jan 08 '24 05:01 dendenxu

Hi @lixiangweiabc , I updated the installation guide to include more description about Windows. Could you please check whether it's easier to follow now?

Hi,thanks for your updating. For issue https://github.com/zju3dv/4K4D/issues/14#issue-2067001433, I founde that the nvidia-driver and the cuda version is incompatible. For some reason, I can't update the Nvidia driver version of my current machine. I will follow your guide in a new ubuntu 18.04 machine.

lixiangweiabc avatar Jan 08 '24 06:01 lixiangweiabc