[Question]: e/MInference/scbench/run_scbench.py", line 200, in load_model llm = LLM( NameError: name 'LLM' is not defined. Did you mean: 'llm'?

Describe the bug

I followed the install instructions as https://github.com/microsoft/MInference/tree/main/scbench#environment-for-efficient-long-context-methods but came across this issue. Is there any way that I can use this benchmark without installing MInference? I think this is more benift for extending this benchmark to other models.

Steps to reproduce

conda create -n scbench python=3.10 -y && conda activate scbench
pip install torch
pip install minference
pip install flash-attn --no-build-isolation
git clone https://github.com/microsoft/MInference.git && cd MInference/scbench
pip install -r requirements.txt
VLLM_ALLOW_LONG_MAX_MODEL_LEN=1 CUDA_VISIBLE_DEVICES=0 VLLM_WORKER_MULTIPROC_METHOD=spawn bash scripts/run_single_method.sh path/to/model 1 multi-turn vllm dense

Expected Behavior

No response

Logs

No response

Additional Information

No response