Migrate backend quantizers/partitioners to use new torchao quant primitives

[metascroy]

🚀 The feature, motivation and pitch

PT2E will be copied over to torchao, and migrate to using the new torchao quant primitives. The existing PT2E code will remain in pytorch/ao with the old quant primitives to avoid breaking existing code.

Backends should migrate to using PT2E in torchao and recognizing the new torchao quant primitives during lowering.

By standardizing on torchao quant primitives, model lowering will "just work" in common scenarios, even if a user quantizes the model using torchao's quantize_.

Alternatives

No response

Additional context

No response

RFC (Optional)

No response

[metascroy]

Goal is to migrate by end of first week of April per @jerryzh168 . But this may change.