peltax

Is there plans to go for a bugfix triage?

I opted to use react-native-device-info to get hms availability instead of this lib as it was already in use. One less native dependency.

Applying [this patch](https://aur.archlinux.org/cgit/aur.git/tree/python-pytorch-fix-cuda-12_4.patch?h=python-pytorch-cxx11abi) manually to miniconda3/lib/python3.12/site-packages/torch/include/ATen/core/boxing/impl/boxing.h (where miniconda3 is the conda prefix folder) fixed it for me. I also have nighly version of faiss-gpu but that shouldn't matter, installed using...