Cortex-M backend: Refactor quantized_op_fusion_pass (#20179)

Move to use the AtenToCortexMPass instead.


Signed-off-by: Adrian Lundell <adrian.lundell@arm.com>
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>

Author: AdrianLundell

backends/cortex_m/passes/BUCK

@@ -13,7 +13,6 @@ fbcode_target(_kind = runtime.python_library,
     name="replace_quant_nodes_pass",
     srcs=[
         "replace_quant_nodes_pass.py",
-        "quantized_op_fusion_pass.py",
     ],
     deps=[
         "//caffe2:torch",

backends/cortex_m/passes/__init__.py

@@ -41,6 +41,5 @@ def _ensure_cortex_m_dependencies() -> None:
 from .decompose_hardswish_pass import DecomposeHardswishPass  # noqa
 from .decompose_mean_pass import DecomposeMeanPass  # noqa
 from .quantized_clamp_activation_pass import QuantizedClampActivationPass  # noqa
-from .quantized_op_fusion_pass import QuantizedOpFusionPass  # noqa
 from .replace_quant_nodes_pass import ReplaceQuantNodesPass  # noqa
 from .cortex_m_pass_manager import CortexMPassManager  # noqa  # usort: skip