PyTorch: [CPP FakeTensor] Don't look up the op in register_fake (#200798)

Executive Summary

Summary: Forward fix for #199785 (D124671672). #199785 made FakeImplHolder.register call lookup_op(qualname)._schema to get the (name, overload) key for _fake_dispatch_register_custom_op_impl . lookup_op goes through torch.ops , which needs the op's schema to be defined. But register_fake has always worked before the schema is defined, as long as the dispatcher already knows the op (e.g.

Artificial Intelligence Architecture & Model Evaluation

From an artificial intelligence architecture, model weights governance, and inference efficiency perspective: - **Weights Accessibility & Sovereignty:** Evaluates whether weights are open for private self-hosting or locked behind centralized cloud APIs. - **Quantization & Edge Performance:** Kernel optimizations (4-bit/8-bit GGUF, AWQ, EXL2) allow high tokens-per-second on consumer GPUs and Apple Silicon. - **Reasoning & Architectural Scaling:** Scrutinizes mixture-of-experts (MoE), attention mechanisms, and fine-tuning datasets against open community benchmarks.

Impact on the Open Ecosystem

Protects developers and enterprises from proprietary black-box entrapment, fostering auditable, sovereign AI infrastructure.