Enable GPTQModel to handle GraniteMoeParallelExperts

[fabianlim]

Granite MoE uses a 3D tensor to hold the expert weights, so GPTQModel does not work out of the box.

There are two options

1. module swap GraniteMoeParallelExperts to hold a ModuleList of Linears, then AutoGPTQ will be able to detect them and replace them with QuantLinears
2. write a custom gptq module that handles the GraniteMoeParallelExperts case

Either of the two approaches will solve both quant + inference paths. Option 1 should be easier than Option 2, but in some sense Option 2 should be more proper.

When doing option 2 we should be reusing code from the original gptq.

• also it should be written generally, to not just only handle this particular GraniteMoeParallelExperts instance, but all cases with 3D tensors

[Abhishek-TAMU]

Test script used to test the draft PR to quantize Granite3.0-3b-MoE model (/ibm-dmf-lakehouse/models/base_training/shared/granite-3.0-3b-a800m-base/r240924a).

As per Praveen, this work could be further continued later in Q2 of 2025 when new Granite MoE 120b model will come out.