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

5
vllm/model_executor/models/nemotron_h.py
View File

@@ -54,7 +54,6 @@ from vllm.model_executor.models.mamba_cache import (MambaCacheManager,
from vllm.model_executor.models.utils import (
AutoWeightsLoader, WeightsMapper, make_empty_intermediate_tensors_factory,
make_layers, maybe_prefix)
from vllm.model_executor.sampling_metadata import SamplingMetadata
from vllm.sequence import IntermediateTensors
from vllm.transformers_utils.configs import NemotronHConfig
from vllm.utils import LayerBlockType
@@ -622,10 +621,8 @@ class NemotronHForCausalLM(nn.Module, HasInnerState, SupportsLoRA, SupportsPP,
def compute_logits(
self,
hidden_states: torch.Tensor,
sampling_metadata: SamplingMetadata,
) -> Optional[torch.Tensor]:
logits = self.logits_processor(self.lm_head, hidden_states,
sampling_metadata)
logits = self.logits_processor(self.lm_head, hidden_states)
return logits
def load_weights(self, weights: Iterable[tuple[str,