Incorrect Classification of CustomSiLU as Binary Operator in _AIMET_V1_BINARY_MODULES

[kenanking]

The custom.CustomSiLU operator appears to be incorrectly classified as a binary operator in the _AIMET_V1_BINARY_MODULES list (line 752), which is causing quantization to fail.

SiLU (Sigmoid Linear Unit) is an activation function defined as f(x) = x * sigmoid(x), which operates on a single input tensor. This is a unary operation, not a binary one. I think it's an accidental mistake.

_AIMET_V1_BINARY_MODULES = [
    custom.MatMul,
    custom.Add,
    custom.Multiply,
    # ... other binary operators ...
    custom.CustomSiLU,  # <-- This appears to be incorrectly classified
    custom.Maximum,
    # ... other operators ...
]

aimet/TrainingExtensions/torch/src/python/aimet_torch/v2/nn/fake_quant/_legacy_impl.py

Line 752 in 414e7c4

custom.CustomSiLU,

[My-captain]

same issue when quantize ovis2, anyone could confirm and correct it in sdk?