Add torchao kernels to xcframework

[metascroy]

🚀 The feature, motivation and pitch

We can build low-bit kernels on Macbook (https://github.com/pytorch/executorch/tree/main/examples/models/llama#running-with-low-bit-kernels), but we don't currently build these as part of the main ExecuTorch CMakeLists.txt.

Instead they are built inside the runner code here: https://github.com/pytorch/executorch/blob/main/examples/models/llama/CMakeLists.txt#L118-L137

In this task, we want to add an option to build this kernel library to the ET CMakeLists.txt (https://github.com/pytorch/executorch/blob/main/CMakeLists.txt) guarded under EXECUTORCH_BUILD_KERNELS_TORCHAO, and add them to the xcframeworks (https://github.com/pytorch/executorch/blob/main/scripts/build_apple_frameworks.sh).

This will let us run low-bit kernels on the iOS app.

cc @larryliu0820 @jathu @kimishpatel @jerryzh168 @shoumikhin @cbilgin

[hinriksnaer]

I'd be happy to take this on!

[shoumikhin]

@hinriksnaer please go ahead