Bump vllm version to v0.7.3 (Qwen2.5-VL vllm inference fixes)

[m-evdokimov]

Reminder

• I have read the above rules and searched the existing issues.

System Info

• llamafactory version: 0.9.2.dev0
• Python version: 3.11.11
• PyTorch version: 2.5.1+cu124 (GPU)
• Transformers version: 4.49.0
• Datasets version: 3.2.0
• Accelerate version: 1.2.1
• PEFT version: 0.12.0
• TRL version: 0.9.6
• GPU type: NVIDIA H100 80GB HBM3
• GPU number: 1
• GPU memory: 79.11GB
• vLLM version: 0.7.2

Reproduction

python3.11 LLaMA-Factory/scripts/vllm_infer.py --model_name_or_path "Qwen/Qwen2.5-VL-7B-Instruct" \
                                               --dataset my_dataset \
                                               --dataset_dir /data/dataset \
                                               --template qwen2_vl  \
                                               --save_name /predicts/predictions.jsonl \

Error msg: ValueError: Model architectures ['Qwen2_5_VLForConditionalGeneration'] failed to be inspected. Please check the logs for more details.

Others

VLLM v0.7.3 release fixes Qwen2.5-VL inference. Would you consider updating?

Raw vllm=v0.7.3 inference fixes the bug.

[hiyouga]

done