Adding a fix for a common macOS failure mode

Fix minor error in documentation detailing the default model in use for face_encodings

docs: Fix simple typo, eucledian -> euclidean

Otherwise smaller faces can't be located via command line

There is a small typo in examples/face_recognition_knn.py, examples/facerec_ipcamera_knn.py.

Should read `euclidean` rather than `eucledian`.

Real time facial recognition on ip cameras using knn.

Real time fast face recognition on ip cameras using knn.

revise comments

The goal of these changes is to provide actual
information about supported Python versions.
I believe that only versions for those you have
tests on CI should be listed as supported.

Show a message if a training image contains none or more than one faces

Added a link for a new Python code example using a SVM