Parallelize and add more wheel builds

[ddelange]

linux-aarch64 makes up for almost 10% of all platforms ref giampaolo/psutil#2103

aarch64 has already surpassed windows in terms of downloads for this package. Oracle, Amazon, Google, and Microsoft are all offering aarch64 cloud instances at an undeniable price point compared to amd/intel, so the demand will undoubtedly only grow

Same goes for mac and their new arm chips. This PR ads arm64/aarch64 and alpine(musl) wheels.

Example run: https://github.com/ddelange/dlib-wheels/actions/runs/8555157171

• The first step generates a matrix of all wheel combinations (based on python_requires in setup.py, missing here so I added CIBW_PROJECT_REQUIRES_PYTHON)
• For each output wheel, one job is created and so the wheels build in maximum parallelization.
• Use native arm64 mac runners for the arm64 macos wheels

Analogous to vaexio/vaex#2331, ahupp/python-magic#294, MagicStack/asyncpg#954

[ddelange]

cc @alesanfra

[alesanfra]

Hi @ddelange, the PR looks good to me, thank you for contributing to this project!