evaluation error: TypeError: EngineArgs.__init__() got an unexpected keyword argument 'worker_use_ray'

[ChenDRAG]

The evaluation script using only 1 GPU works out fine. However, when employing data data_parallel_size >1. an error shows up.

TypeError: EngineArgs.init() got an unexpected keyword argument 'worker_use_ray'

RayTaskError(TypeError): [36mray::run_inference_one_model()[39m (pid=718457, ip=10.137.8.13)
  File "/home/huayuc/miniconda3/envs/openr1/lib/python3.11/site-packages/lighteval/models/vllm/vllm_model.py", line 336, in run_inference_one_model
    llm = LLM(**model_args)
          ^^^^^^^^^^^^^^^^^
  File "/home/huayuc/miniconda3/envs/openr1/lib/python3.11/site-packages/vllm/utils.py", line 1039, in inner
    return fn(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^
  File "/home/huayuc/miniconda3/envs/openr1/lib/python3.11/site-packages/vllm/entrypoints/llm.py", line 210, in __init__
    engine_args = EngineArgs(
                  ^^^^^^^^^^^
**TypeError: EngineArgs.__init__() got an unexpected keyword argument 'worker_use_ray'**
(pid=718453) INFO 02-03 16:25:25 __init__.py:183] Automatically detected platform cuda. [repeated 3x across cluster] (Ray deduplicates logs by default. Set RAY_DEDUP_LOGS=0 to disable log deduplication, or see https://docs.ray.io/en/master/ray-observability/user-guides/configure-logging.html#log-deduplication for more options.)
(pid=718453) WARNING 02-03 16:25:25 cuda.py:333] Detected different devices in the system:  [repeated 3x across cluster]
(pid=718453) WARNING 02-03 16:25:25 cuda.py:333] NVIDIA A100-SXM4-80GB [repeated 12x across cluster]
(pid=718453) WARNING 02-03 16:25:25 cuda.py:333] NVIDIA DGX Display [repeated 3x across cluster]
(pid=718453) WARNING 02-03 16:25:25 cuda.py:333] Please make sure to set `CUDA_DEVICE_ORDER=PCI_BUS_ID` to avoid unexpected behavior. [repeated 3x across cluster]

My system setting:
vllm==0.7.1. torch==2.5.1. lighteval == 0.6.0.dev0

Command

MODEL=deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B
NUM_GPUS=1
MODEL_ARGS="pretrained=$MODEL,dtype=float16,data_parallel_size=$NUM_GPUS,max_model_length=32768,gpu_memory_utilisation=0.8"
TASK=math_500
OUTPUT_DIR=data/evals/$MODEL

CUDA_VISIBLE_DEVICES=0,1,2,3 lighteval vllm $MODEL_ARGS "custom|$TASK|0|0" \
    --custom-tasks src/open_r1/evaluate.py \
    --use-chat-template \
    --system-prompt="Please reason step by step, and put your final answer within \boxed{}." \
    --output-dir $OUTPUT_DIR 

Any idea why this happens? Thank you!

[lewtun]

It seems vllm deprecated worker_use_ray in favour of distributed_executor_backend in this PR. I'll fix this if @hynky1999 doesn't beat me to it :)