Masking sucks. Rotobrushing sucks. Even now. You can hide all the weird artifacts that come with codeformer with face parsing. Use any yolo face bbox detector model, and hunt around for the faceparser model. Wade through the dead links, if I found it, so can you. Anyway, I'm uploading this as a template, with an image input and your choice of either an full size or bbox mask output. Just select the features you want to keep in one group and those you want to mask in the other. They'll be composited in the output. You want to run a few facedetailer passes without losing your likeness? Use this.
This is mainly for those already familiar with masking and are comfortable with more advanced techniques. If you have no idea what to do with it, it's probably not for you. If you don't know why I'm not posting something more complete, it's definitely not for you.