Is it compatible with silicon chip

[ankitbioinfo]

Thanks for your niche niche tool.
Currently I am running your tool in macbook with silicon chips. This do not support CUDA as far as I know.
I am running your tutorial_simulated.ipynb with folloiwng modification.

# training
model = sn.tr.Runner(adata=adata, device='cpu',  verbose=False)
adata = model.fit(lr=lr, epochs=epochs)
                
# clustering
#target_num = len(set(adata.obs['Niche']))
#adata = sn.tr.clustering(adata=adata, target_k=target_num)

And getting following error. Any help will be appreciated about rectifying the error.
Thanks.

[ProDong0512]

scNiche works well on cpu.

Remember to install the corresponding cpu version of torch and dgl with that in tutorial. And we found that you're runing on python3.11 while scNiche is recommended on python3.9. Please check the version of python, torch, dgl and numpy in your environment.

Here is a feasible manner to install a cpu version scNiche:

_cd scNiche-main

conda env create -f scniche_dev.yaml -n scniche
conda activate scniche_
pip install torch==1.12.1+cpu torchvision==0.13.1+cpu torchaudio==0.12.1 --extra-index-url https://download.pytorch.org/whl/cpu
or
conda install pytorch==1.12.1 torchvision==0.13.1 torchaudio==0.12.1 cpuonly -c pytorch

pip install dgl==1.1.0
or
conda install -c dglteam dgl==1.1.0

python setup.py build
python setup.py install

Hope this would be helpful.