[Model] Pooling models default to using chunked prefill & prefix caching if supported. (#20930)

Signed-off-by: wang.yuqi <noooop@126.com>
This commit is contained in:
wang.yuqi
2025-08-11 09:41:37 -07:00
committed by GitHub
parent 16fb668b61
commit 84cf78acee
31 changed files with 452 additions and 261 deletions
+4 -2
View File
@@ -26,7 +26,8 @@ from vllm.model_executor.pooling_metadata import PoolingMetadata
from vllm.sequence import IntermediateTensors
from vllm.tasks import PoolingTask
from .interfaces import SupportsCrossEncoding, SupportsV0Only
from .interfaces import (SupportsCrossEncoding, SupportsV0Only,
default_pooling_type)
from .utils import WeightsMapper, maybe_prefix
@@ -201,6 +202,7 @@ class ModernBertEncoderLayer(nn.Module):
@support_torch_compile
@default_pooling_type("CLS")
class ModernBertModel(nn.Module):
hf_to_vllm_mapper = WeightsMapper(
orig_to_new_prefix={"layers.": "encoder_layer.layers."})
@@ -264,7 +266,6 @@ class ModernBertPooler(Pooler):
self.pooling = PoolingMethod.from_pooling_type(pooling_type)
self.dense = nn.Linear(config.hidden_size, config.hidden_size,
config.classifier_bias)
self.pooling_type = config.classifier_pooling
self.act = nn.GELU()
self.norm = nn.LayerNorm(config.hidden_size,
eps=config.norm_eps,
@@ -294,6 +295,7 @@ class ModernBertPooler(Pooler):
return pooled_output
@default_pooling_type("CLS")
class ModernBertForSequenceClassification(nn.Module, SupportsV0Only,
SupportsCrossEncoding):