[Bug] Typo in parameter name 'deprecated_tenosr_subclass' in test and benchmark files

[GiGiKoneti]

Description

While inspecting the codebase, I noticed a minor spelling typo in the parameter name deprecated_tenosr_subclass (spelled tenosr instead of tensor). It is defined in the helper functions of the testing and benchmarking suites. It should be renamed to deprecated_tensor_subclass for code correctness and readability.

Steps to reproduce

# Location 1: test/quantization/test_quant_api.py (line 135)
def _get_ref_change_linear_weights_to_woqtensors(deprecated_tenosr_subclass):
    ...

# Location 2: benchmarks/benchmark_aq.py (line 47)
def _get_ref_change_linear_weights_to_woqtensors(deprecated_tenosr_subclass):
    ...

Environment

• torch version: 2.11.0 (or newer)
• torchao version: 0.15.1
• Hardware: CPU
• Relevant library versions: N/A
• OS: macOS