Describe the feature you want to add to this project

Many ML libraries (like xgboost, tensorflow, pytorch) publish CPU-only versions not to force users to install heavy GPU support if not needed. I'd like pycaret to support CPU only install too

Describe your proposed solution

Publish pycaret-cpu package that will depend on cpu version of ML libraries (like xgboost-cpu).

Describe alternatives you've considered, if relevant

No response

Additional context

Even separate Nvidia GPU support, xgboost depend on, is heavy:

28K	/tmp/xgboost/lib/python3.10/site-packages/nvidia_nccl_cu12-2.22.3.dist-info
244M	/tmp/xgboost/lib/python3.10/site-packages/nvidia