biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

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

Browse Source 
Signed-off-by: Woosuk Kwon <woosuk@thinkingmachines.ai>
This commit is contained in:
Woosuk Kwon
2025-09-21 10:37:11 -07:00
committed by GitHub
parent c438b2951c
commit 1c3ffdbecc
141 changed files with 172 additions and 583 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
vllm/model_executor/models/whisper.py
View File

@@ -31,7 +31,6 @@ from vllm.model_executor.layers.quantization.base_config import (
from vllm.model_executor.layers.vocab_parallel_embedding import ParallelLMHead
from vllm.model_executor.model_loader.utils import set_default_torch_dtype
from vllm.model_executor.model_loader.weight_utils import default_weight_loader
from vllm.model_executor.sampling_metadata import SamplingMetadata
from vllm.multimodal import MULTIMODAL_REGISTRY, NestedTensors
from vllm.multimodal.inputs import (MultiModalDataDict, MultiModalFieldConfig,
MultiModalKwargsItems)
@@ -936,10 +935,8 @@ class WhisperForConditionalGeneration(nn.Module, SupportsTranscription,
return WhisperAudioInputs(input_features=input_features)
def compute_logits(self, hidden_states: torch.Tensor,
sampling_metadata: SamplingMetadata) -> torch.Tensor:
logits = self.logits_processor(self.proj_out, hidden_states,
sampling_metadata)
def compute_logits(self, hidden_states: torch.Tensor) -> torch.Tensor:
logits = self.logits_processor(self.proj_out, hidden_states)
return logits
def load_weights(self, weights: Iterable[tuple[str,