Remove unused kwargs from model definitions (#13555)

vllm/model_executor/models/internvl.py

@@ -17,7 +17,6 @@ import torchvision.transforms as T
 from PIL import Image
 from transformers import BatchFeature, PretrainedConfig, TensorType
 
-from vllm.attention import AttentionMetadata
 from vllm.config import VllmConfig
 from vllm.model_executor.layers.quantization import QuantizationConfig
 from vllm.model_executor.layers.quantization.awq import AWQConfig
@@ -929,8 +928,6 @@ class InternVLChatModel(nn.Module, SupportsMultiModal, SupportsPP):
         self,
         input_ids: torch.Tensor,
         positions: torch.Tensor,
-        kv_caches: List[torch.Tensor],
-        attn_metadata: AttentionMetadata,
         intermediate_tensors: Optional[IntermediateTensors] = None,
         inputs_embeds: Optional[torch.Tensor] = None,
         **kwargs: object,
@@ -951,8 +948,6 @@ class InternVLChatModel(nn.Module, SupportsMultiModal, SupportsPP):
         forward_kwargs = {
             "input_ids": input_ids,
             "positions": positions,
-            "kv_caches": kv_caches,
-            "attn_metadata": attn_metadata,
             "intermediate_tensors": intermediate_tensors,
             "inputs_embeds": inputs_embeds,
         }