Flode-Labs / vid2densepose

Convert your videos to densepose and use it on MagicAnimate
MIT License
957 stars 122 forks source link

ModuleNotFoundError: No module named 'detectron2.config' #4

Open alexfredo opened 7 months ago

alexfredo commented 7 months ago

Hi I have a error when I use the "py main.py" command, the error :

File "C:\vid2densepose\main.py", line 1, in from detectron2.config import get_cfg ModuleNotFoundError: No module named 'detectron2.config'

Please helps ! thanks

igorriti commented 7 months ago

Did you do pip install -r requirements.txt ?

I would reccomend to use a WSL if you are on Windows.

alexfredo commented 7 months ago

Yes I tried but I have a error when I pip install the requirement too : ffferror

Questscape commented 7 months ago

I have same issue on windows 10 as alexfredo

alexfredo commented 7 months ago

I have windows 10 too

tadeodonegana commented 7 months ago

Have you cloned the Detectron repository?

Please use git clone https://github.com/facebookresearch/detectron2.git as specified in the readme.

Questscape commented 7 months ago

yes i cloned and installed the detectron repo, installed it, now im getting: Traceback (most recent call last): File "G:\magic_animate\magic-animate\video2dense\vid2densepose\app.py", line 7, in <module> from densepose import add_densepose_config ModuleNotFoundError: No module named 'densepose'

when installing densepose, i get this same error as above: image

igorriti commented 7 months ago

The issue lies with Windows, and while we are actively working on a solution, I would recommend utilizing Windows Subsystem for Linux (WSL) as a temporary workaround. You can easily download WSL from the Microsoft Store, and it functions seamlessly. I will also include this advice in the readme for your convenience.

chen-rn commented 7 months ago

I'm on OSX Ventura and am running into the same issue as well.

Edit: creating a venv, then first installing pytorch+ torchvision before installing the requirement.txt resolved this.

zono50 commented 7 months ago

I'm on linux, installed torch and torch vision, did pip install -r requirements.txt and then git clone of the detectron2 and getting no module named detectron2.config found

igorriti commented 7 months ago

I cannot reproduce your problem @zono50 , could you provide us with more details? Which version of torch and torch vision did you have?

I suggest you to create a virtual env, git clone the detectron repo, then install the torch and torch vision(check the requirements.txt for the versions) and finally the pip install - r requirements.txt

ljzm commented 6 months ago

您是否克隆了 Detectron 存储库?

请按照自述文件中的规定使用。git clone https://github.com/facebookresearch/detectron2.git

请问这个仓库git clone下来应该放在哪个目录下,放在vid2densepose目录下吗,还是同级?期待答复。谢谢

harpy2 commented 6 months ago

!apt install -y nvidia-cuda-dev

ml1969 commented 5 months ago

Hi, I trioed to install it under windows like @alexfredo and I got the same errormessage. After that, I tried to use WSL with Ubuntu under Win11. I´m a noob. I do not get it to work. At first I realised that there are some packeges missing :( At least I installed the wrong CUD Version, and so on.

I will try it on a fresh WSL installation,

For me the best solution will be, no erros under windows 👯

Regards michael