How can I load this model
#3
by godcat950081 - opened
My previous environment could successfully load the Qwen3-Omni-30B-A3B-Instruct model using transform, but after downloading the AWQ-4bit version and updating the local model path, it now consistently throws the following error when starting up:
raise ValueError(f"{tensor_name} is on the meta device, we need a value to put in on {device}.")
ValueError: weight_scale is on the meta device, we need a value to put in on cpu.
I encountered this issue before, I used pytorch failed but succeeded with vllm
me too