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

[Chore] Remove Sampler from Model Code (#17084)

Browse Source 
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
This commit is contained in:
Woosuk Kwon
2025-04-24 02:49:33 -07:00
committed by GitHub
parent 2bc0f72ae5
commit b411418ff0
103 changed files with 48 additions and 1099 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
vllm/model_executor/models/mamba.py
View File

@@ -14,7 +14,6 @@ from vllm.model_executor.layers.logits_processor import LogitsProcessor
from vllm.model_executor.layers.mamba.mamba_mixer import MambaMixer
from vllm.model_executor.layers.quantization.base_config import (
QuantizationConfig)
from vllm.model_executor.layers.sampler import SamplerOutput, get_sampler
from vllm.model_executor.layers.vocab_parallel_embedding import (
DEFAULT_VOCAB_PADDING_SIZE, ParallelLMHead, VocabParallelEmbedding)
from vllm.model_executor.model_loader.weight_utils import default_weight_loader
@@ -213,7 +212,6 @@ class MambaForCausalLM(nn.Module, HasInnerState, IsAttentionFree, SupportsPP,
self.logits_processor = LogitsProcessor(self.unpadded_vocab_size,
config.vocab_size)
self.sampler = get_sampler()
self.make_empty_intermediate_tensors = (
self.backbone.make_empty_intermediate_tensors)
@@ -267,14 +265,6 @@ class MambaForCausalLM(nn.Module, HasInnerState, IsAttentionFree, SupportsPP,
sampling_metadata)
return logits
def sample(
self,
logits: Optional[torch.Tensor],
sampling_metadata: SamplingMetadata,
) -> Optional[SamplerOutput]:
next_tokens = self.sampler(logits, sampling_metadata)
return next_tokens
def load_weights(self, weights: Iterable[Tuple[str,
torch.Tensor]]) -> Set[str]:
loader = AutoWeightsLoader(self)