This combination is needed for Google Colab and local machines without CUDA, which want to be able to evaluate the models trained with torch 1.13.1
I admit the installation section is kind of a mess and this does not improve it further, but so is python package management. So I would add it anyway and keep fingers crossed we do not need to add more.
This combination is needed for Google Colab and local machines without CUDA, which want to be able to evaluate the models trained with torch 1.13.1
I admit the installation section is kind of a mess and this does not improve it further, but so is python package management. So I would add it anyway and keep fingers crossed we do not need to add more.