Enable NUMA feature for llama_cpp_python (#4040)

modules/llamacpp_hf.py

@@ -172,6 +172,7 @@ class LlamacppHF(PreTrainedModel):
     @classmethod
     def from_pretrained(cls, pretrained_model_name_or_path: Optional[Union[str, os.PathLike]], *model_args, **kwargs):
         assert len(model_args) == 0 and len(kwargs) == 0, "extra args is currently not supported"
+
         if isinstance(pretrained_model_name_or_path, str):
             pretrained_model_name_or_path = Path(pretrained_model_name_or_path)
 
@@ -198,6 +199,7 @@ class LlamacppHF(PreTrainedModel):
             'use_mlock': shared.args.mlock,
             'mul_mat_q': shared.args.mul_mat_q,
             'low_vram': shared.args.low_vram,
+            'numa': shared.args.numa,
             'n_gpu_layers': shared.args.n_gpu_layers,
             'rope_freq_base': RoPE.get_rope_freq_base(shared.args.alpha_value, shared.args.rope_freq_base),
             'tensor_split': tensor_split_list,