support batch infer in vllm

Former-commit-id: 3ef5ed3b9a44eed2f7e3ff221dfc343d0a97c0b5

src/llamafactory/hparams/parser.py

@@ -122,7 +122,7 @@ def _check_extra_dependencies(
         require_version("mixture-of-depth>=1.1.6", "To fix: pip install mixture-of-depth>=1.1.6")
 
     if model_args.infer_backend == "vllm":
-        require_version("vllm>=0.4.3,<0.6.4", "To fix: pip install vllm>=0.4.3,<0.6.4")
+        require_version("vllm>=0.4.3,<0.6.5", "To fix: pip install vllm>=0.4.3,<0.6.5")
 
     if finetuning_args.use_galore:
         require_version("galore_torch", "To fix: pip install galore_torch")