Don't run duplicated BUCK and CMake runner workflows (#1074)

Summary:
Currently we are running BUCK and CMake runner on CI workflow. This is redundant because we don't expect much difference in the runners from two build systems.

Instead we only run CMake runner on enabled models, and run MV3 (portable + xnnpack with different quantization/delegation options) on BUCK for validation only.

That's because BUCK runner can be validated through internal CI.  Also CMake runner depends on BUCK srcs, so we don't expect failures from BUCK if CMake runner is working.

Pull Request resolved: https://github.com/pytorch/executorch/pull/1074

Reviewed By: huydhn

Differential Revision: D50613805

Pulled By: kirklandsign

fbshipit-source-id: 89e98c90a0eab3b1a50060dccfad3eb35cce6141