[V0 Deprecation] Remove V0 sampling metadata (#25345)

Signed-off-by: Woosuk Kwon <woosuk@thinkingmachines.ai>

vllm/model_executor/models/hyperclovax_vision.py

@@ -31,7 +31,6 @@ from transformers.modeling_utils import no_init_weights
 from vllm.config import VllmConfig
 from vllm.inputs import InputProcessingContext
 from vllm.model_executor.layers.quantization import QuantizationConfig
-from vllm.model_executor.sampling_metadata import SamplingMetadata
 from vllm.multimodal import MULTIMODAL_REGISTRY
 from vllm.multimodal.cache import BaseMultiModalProcessorCache
 from vllm.multimodal.inputs import (MultiModalDataDict, MultiModalFieldConfig,
@@ -962,10 +961,8 @@ class HCXVisionForCausalLM(nn.Module, SupportsMultiModal, SupportsPP):
     def compute_logits(
         self,
         hidden_states: torch.Tensor,
-        sampling_metadata: SamplingMetadata,
     ) -> Optional[torch.Tensor]:
-        return self.language_model.compute_logits(hidden_states,
-                                                  sampling_metadata)
+        return self.language_model.compute_logits(hidden_states)
 
     def load_weights(
         self,