Add Vulkan Quantizer to Llama export lib (#6169)

Summary:
Pull Request resolved: https://github.com/pytorch/executorch/pull/6169

TSIA.

Note that only 8 bit weight only quantization is supported for now since `VulkanQuantizer` does not support 4 bit weight only quantization at the moment.
ghstack-source-id: 247613963
exported-using-ghexport

Reviewed By: jorgep31415

Differential Revision: D64249615

fbshipit-source-id: 33ef8d06e56838da5f7866832e50fe74d8878811