>>11044541Kreyzig is pretty good (the formulas can get pretty ginourmous due to the use of tensors), Do Carmo is a classic. Stoker is probably too hard (and too weird a selection of topics). Wildberger's lectures are pretty legit.
I think IMPA's page has a set of diffgeo lectures on video.
If you care about the discrete case, Keenan Crane's lecture notes address some interesting applications and also how some combinatorial properties of the discrete version relate to analytic properties of the continuous case. Still, he doesn't really develop the classic theory (curvatures, gauss-bonnet, theorem egregium, geodesic coordinates, etc) so it's more of a companion text.
