Conflicting dependency error on Windows machine

Hello,

I am following the installation steps on anaconda3 and I get the following error:

ERROR: Cannot install cell2location and scvi-tools because these package versions have conflicting dependencies.

The conflict is caused by:
    numpyro 0.9.2 depends on jaxlib>=0.1.65
    numpyro 0.9.1 depends on jaxlib>=0.1.65
    numpyro 0.9.0 depends on jaxlib>=0.1.65
    numpyro 0.8.0 depends on jaxlib>=0.1.65
    numpyro 0.7.2 depends on jaxlib>=0.1.65
    numpyro 0.7.1 depends on jaxlib>=0.1.65
    numpyro 0.7.0 depends on jaxlib>=0.1.65
    numpyro 0.6.0 depends on jaxlib==0.1.62
    scvi-tools 0.15.3 depends on jax>=0.3
    flax 0.1.0 depends on jax>=0.1.59
    numpyro 0.5.0 depends on jax==0.2.8
    scvi-tools 0.15.3 depends on jax>=0.3
    flax 0.1.0 depends on jax>=0.1.59
    numpyro 0.4.1 depends on jax==0.2.3
    numpyro 0.4.0 depends on jaxlib==0.1.55
    scvi-tools 0.15.3 depends on jax>=0.3
    flax 0.1.0 depends on jax>=0.1.59
    numpyro 0.3.0 depends on jax==0.1.73
    numpyro 0.2.4 depends on jaxlib==0.1.37
    numpyro 0.2.3 depends on jaxlib==0.1.36
    numpyro 0.2.2 depends on jaxlib==0.1.36
    scvi-tools 0.15.3 depends on jax>=0.3
    flax 0.1.0 depends on jax>=0.1.59
    numpyro 0.2.1 depends on jax==0.1.49
    numpyro 0.2.0 depends on jaxlib==0.1.27
    scvi-tools 0.15.3 depends on jax>=0.3
    flax 0.1.0 depends on jax>=0.1.59
    numpyro 0.1.0 depends on jax==0.1.35

To fix this you could try to:
1. loosen the range of package versions you've specified
2. remove package versions to allow pip attempt to solve the dependency conflict

ERROR: ResolutionImpossible: for help visit https://pip.pypa.io/en/latest/topics/dependency-resolution/#dealing-with-dependency-conflicts

Is it possible to loosen the requirements?

Thank you!

[rimelof]

My windows machine does not have GPU. I ended up getting cell2location installed on a linux cluster and running it there. I have no success in windows.

[vitkl]

My best guess would be that the issue is with python user site being used instead of the conda environment. Try using

export PYTHONNOUSERSITE="literallyanyletters"

Before creating and every time before activating the environment.

Are you doing this in the HPC system? This seems to be a very common issues with how permissions on compute cluster work with pip.