[5/N] pass the whole config to model (#9983)

Signed-off-by: youkaichao <youkaichao@gmail.com>

vllm/model_executor/models/qwen2.py

@@ -29,7 +29,7 @@ from transformers import Qwen2Config
 
 from vllm.attention import Attention, AttentionMetadata
 from vllm.compilation.decorators import support_torch_compile
-from vllm.config import CacheConfig, LoRAConfig
+from vllm.config import CacheConfig, VllmConfig
 from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
 from vllm.model_executor.layers.activation import SiluAndMul
 from vllm.model_executor.layers.layernorm import RMSNorm
@@ -405,12 +405,14 @@ class Qwen2ForCausalLM(nn.Module, SupportsLoRA, SupportsPP):
 
     def __init__(
         self,
-        config: Qwen2Config,
-        cache_config: Optional[CacheConfig] = None,
-        quant_config: Optional[QuantizationConfig] = None,
-        lora_config: Optional[LoRAConfig] = None,
+        vllm_config: VllmConfig,
         prefix: str = "",
     ) -> None:
+        super().__init__()
+        config = vllm_config.model_config.hf_config
+        cache_config = vllm_config.cache_config
+        quant_config = vllm_config.quant_config
+        lora_config = vllm_config.lora_config
         # TODO (@robertgshaw2): see if this can be moved out
         if (cache_config.sliding_window is not None
                 and hasattr(config, "max_window_layers")):
@@ -423,8 +425,6 @@ class Qwen2ForCausalLM(nn.Module, SupportsLoRA, SupportsPP):
                                  config.num_hidden_layers,
                              ))
 
-        super().__init__()
-
         self.config = config
         self.lora_config = lora_config