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

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

vllm/model_executor/models/llava_next.py

@@ -11,11 +11,10 @@ from transformers.models.llava_next.modeling_llava_next import (
 from typing_extensions import NotRequired
 
 from vllm.attention import AttentionMetadata
-from vllm.config import CacheConfig, MultiModalConfig, PoolerConfig
+from vllm.config import VllmConfig
 from vllm.inputs import (INPUT_REGISTRY, DecoderOnlyInputs, DummyData,
                          InputContext)
 from vllm.model_executor.layers.pooler import Pooler, PoolingType
-from vllm.model_executor.layers.quantization import QuantizationConfig
 from vllm.model_executor.layers.sampler import SamplerOutput, get_sampler
 from vllm.model_executor.pooling_metadata import PoolingMetadata
 from vllm.model_executor.sampling_metadata import SamplingMetadata
@@ -282,13 +281,12 @@ def input_processor_for_llava_next(ctx: InputContext,
 class LlavaNextForConditionalGeneration(nn.Module, SupportsMultiModal,
                                         SupportsPP):
 
-    def __init__(self,
-                 config: LlavaNextConfig,
-                 multimodal_config: MultiModalConfig,
-                 cache_config: Optional[CacheConfig] = None,
-                 quant_config: Optional[QuantizationConfig] = None,
-                 pooler_config: Optional[PoolerConfig] = None) -> None:
+    def __init__(self, vllm_config: VllmConfig, prefix: str = "") -> None:
         super().__init__()
+        config = vllm_config.model_config.hf_config
+        quant_config = vllm_config.quant_config
+        pooler_config = vllm_config.model_config.pooler_config
+        multimodal_config = vllm_config.model_config.multimodal_config
 
         self.config = config
         self.multimodal_config = multimodal_config
@@ -308,8 +306,7 @@ class LlavaNextForConditionalGeneration(nn.Module, SupportsMultiModal,
 
         self.language_model = init_vllm_registered_model(
             config.text_config,
-            cache_config,
-            quant_config,
+            vllm_config=vllm_config,
             prefix="language_model")
 
         # The same model class supports both language generation and embedding