[fix] quantization properties for lmi dist and hf acc

[sindhuvahinis]

Description

Found a bug while testing quantization with new properties for HF Accelerate and lmi-dist

LMI Dist rolling batch takes in the parsed properties instead of pydantic, so new bitsandbytes8 needs to be handled in rolling batch handler.

Testing:
Tested HF and LMI Dist in my machine with bitsandbytes option
Tested TnX quantization in my machine , works

vllm, deepseed, trtllm has no changes for quantization