hustvl / WeakTr

WeakTr: Exploring Plain Vision Transformer for Weakly-supervised Semantic Segmentation
MIT License
120 stars 2 forks source link

Installation Leads to Dependency Conflict #22

Closed HamidrezaZarrabi closed 7 months ago

HamidrezaZarrabi commented 8 months ago

Search before asking

Question

Issue

I have followed Install.md, but I encountered the error dependency conflicts and I failed to install the packages.

To Reproduce

There are two ways (by using pip or openmim) to install the package mmcv-full.

Openmim

When I run,

pip install -U openmim

I get,

ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts. mlflow 1.30.1 requires pytz<2023, but you have pytz 2023.3.post1 which is incompatible.

I found, the packages pytz==2022.7.1 and mlflow==1.30.1 have been installed by requirements.txt while their versions have not been specified in requirements.txt. Then, when package openmim has been installed, the package pytz has been upgraded automatically. So, I tried to install previous versions of the package openmim. But I could not fix the error dependency conflicts.

Pip

When I run,

pip install mmcv-full==1.4.0 -f https://download.openmmlab.com/mmcv/dist/cu111/torch1.8.0/index.html

I get, > ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts. mlflow 1.30.1 requires importlib-metadata!=4.7.0,<6,>=3.7.0, but you have importlib-metadata 6.7.0 which is incompatible. I found, the packages `importlib-metadata==5.2.0` and `mlflow==1.30.1` have been installed by `requirements.txt` while their versions have not been specified in `requirements.txt`. Then, when package `mmcv-full` has been installed, the package `importlib-metadata` has been downgraded automatically. So, I tried to install other versions of the package `mmcv-full`. But I could not fix the error `dependency conflicts`. # My Environment - OS: Ubuntu 22.04 - Python version: 3.7.16 - torch version: 1.8.0+cu111 - torchvision version: 0.9.0+cu111 ## Request Please let me know, if I should install specific versions of packages to resolve this issue or if there's any additional information needed to assist in troubleshooting.
Yingyue-L commented 8 months ago

Thanks for you question! I have checked package version of my conda environment that you have mentioned as shown below, you can try to install the specific version of these packages.

HamidrezaZarrabi commented 7 months ago

Fixed. I appreciate your swift reply.