Add support for getting executorch quantized model (non-lowered) from ModAI

Summary: Adding support in this diff for getting the executorch quantized model from ModAI manager. This is the non lowered quantized model (with dq->op->q patterns) running on CPU that will be used for CPU eval's in various places.

Reviewed By: navsud, lucylq, larryliu0820

Differential Revision: D56740679

fbshipit-source-id: 27751098127136e3648f7b441c9d36920a395bc8