Enable GPTQv2 from gptqmodel

[chichun-charlie-liu]

Is your feature request related to a problem? Please describe.

GPTQv2 (aka GPTAQ) is supported by gptqmodel, we just need to add a flag to trigger it.

Describe the solution you'd like

add a flag and some description in GPTQ example README.MD

Additional context

GPTAQ may help the accuracy for some models, could be handy when needed.