How to update code to allow for multiple GPUs?

[gulzatmamytbek1]

I run this on NVIDIA Tesla T4, but it always shows CUDA out of memory.

I want to try adding GPUs . But dont know how exactly to add additional CUDAs

model = AutoGPTQForCausalLM.from_quantized(
                model_id,
                model_basename=model_basename,
                use_safetensors=True,
                trust_remote_code=True,
                device="cuda:0", **// How to update this line to allow for 2 or more GPUs?**
                use_triton=False,
                quantize_config=None,
            )