[Model] Standardize pooling heads

vllm/model_executor/layers/pooler/common.py

@@ -2,11 +2,15 @@
 # SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 from collections.abc import Callable
 from dataclasses import dataclass
+from typing import TypeVar
 
 import torch
 
 from vllm.pooling_params import PoolingParams
 
+_T = TypeVar("_T", bound=torch.Tensor | list[torch.Tensor])
+
+ActivationFn = Callable[[_T], _T]
 ClassifierFn = Callable[[torch.Tensor], torch.Tensor]
 
 
@@ -24,4 +28,4 @@ def apply(self, params: PoolingParams) -> None:
         params.requires_token_ids = self.requires_token_ids
 
 
-__all__ = ["ClassifierFn", "PoolingParamsUpdate"]
+__all__ = ["ActivationFn", "ClassifierFn", "PoolingParamsUpdate"]

vllm/model_executor/layers/pooler/seqwise/heads.py

@@ -7,14 +7,7 @@
 import torch
 import torch.nn as nn
 
-from vllm.config import get_current_vllm_config
-from vllm.model_executor.layers.pooler import ClassifierFn
-from vllm.model_executor.layers.pooler.activations import (
-    PoolerActivation,
-    PoolerNormalize,
-    resolve_classifier_act_fn,
-)
-from vllm.model_executor.models.adapters import _load_st_projector
+from vllm.model_executor.layers.pooler import ActivationFn, ClassifierFn
 from vllm.tasks import PoolingTask
 from vllm.v1.pool.metadata import PoolingMetadata
 
@@ -38,17 +31,17 @@ def forward(
 
 
 class EmbeddingPoolerHead(SequencePoolerHead):
-    def __init__(self) -> None:
+    def __init__(
+        self,
+        projector: nn.Module | None = None,
+        head_dtype: str | None = None,
+        activation: ActivationFn | None = None,
+    ) -> None:
         super().__init__()
 
-        # Load ST projector if available
-        vllm_config = get_current_vllm_config()
-        model_config = vllm_config.model_config
-
-        self.projector = _load_st_projector(model_config)
-        self.head_dtype = model_config.head_dtype
-
-        self.activation = PoolerNormalize()
+        self.projector = projector
+        self.head_dtype = head_dtype
+        self.activation = activation
 
     def get_supported_tasks(self) -> Set[PoolingTask]:
         return {"embed"}
@@ -88,15 +81,16 @@ def forward(
                 ]
 
         # for normalize
-        flags = [p.normalize for p in pooling_params]
-        if len(set(flags)) == 1:
-            if flags[0]:
-                pooled_data = self.activation(pooled_data)
-        else:
-            pooled_data = [
-                self.activation(vecs) if f else vecs
-                for vecs, f in zip(pooled_data, flags)
-            ]
+        if self.activation is not None:
+            flags = [p.normalize for p in pooling_params]
+            if len(set(flags)) == 1:
+                if flags[0]:
+                    pooled_data = self.activation(pooled_data)
+            else:
+                pooled_data = [
+                    self.activation(vecs) if f else vecs
+                    for vecs, f in zip(pooled_data, flags)
+                ]
 
         # pooled_data shape: [batchsize, embedding_dimension]
         return pooled_data
@@ -106,20 +100,16 @@ class ClassifierPoolerHead(SequencePoolerHead):
     def __init__(
         self,
         classifier: ClassifierFn | None = None,
-        act_fn: PoolerActivation | str | None = None,
+        logit_bias: float | None = None,
+        head_dtype: str | None = None,
+        activation: ActivationFn | None = None,
     ) -> None:
         super().__init__()
 
-        vllm_config = get_current_vllm_config()
-        model_config = vllm_config.model_config
-
         self.classifier = classifier
-        self.logit_bias: float | None = model_config.pooler_config.logit_bias
-        self.head_dtype = model_config.head_dtype
-
-        self.act_fn = resolve_classifier_act_fn(
-            model_config, static_num_labels=True, act_fn=act_fn
-        )
+        self.logit_bias = logit_bias
+        self.head_dtype = head_dtype
+        self.activation = activation
 
     def get_supported_tasks(self) -> Set[PoolingTask]:
         return {"classify", "score"}
@@ -145,13 +135,15 @@ def forward(
         if self.logit_bias is not None:
             pooled_data -= self.logit_bias
 
-        flags = [p.use_activation for p in pooling_params]
-        if len(set(flags)) == 1:
-            scores = self.act_fn(pooled_data) if flags[0] else pooled_data
-        else:
-            scores = [
-                self.act_fn(vecs) if f else vecs for vecs, f in zip(pooled_data, flags)
-            ]
+        if self.activation is not None:
+            flags = [p.use_activation for p in pooling_params]
+            if len(set(flags)) == 1:
+                scores = self.activation(pooled_data) if flags[0] else pooled_data
+            else:
+                scores = [
+                    self.activation(vecs) if f else vecs
+                    for vecs, f in zip(pooled_data, flags)
+                ]
 
         # scores shape: [batchsize, num_labels]
         return scores

vllm/model_executor/layers/pooler/seqwise/poolers.py

@@ -5,10 +5,15 @@
 
 import torch
 
-from vllm.config import PoolerConfig
+from vllm.config import PoolerConfig, get_current_vllm_config
 from vllm.model_executor.layers.pooler import ClassifierFn, PoolingParamsUpdate
 from vllm.model_executor.layers.pooler.abstract import Pooler
-from vllm.model_executor.layers.pooler.activations import PoolerActivation
+from vllm.model_executor.layers.pooler.activations import (
+    PoolerActivation,
+    PoolerNormalize,
+    resolve_classifier_act_fn,
+)
+from vllm.model_executor.models.adapters import _load_st_projector
 from vllm.tasks import POOLING_TASKS, PoolingTask
 from vllm.v1.pool.metadata import PoolingMetadata
 
@@ -86,7 +91,14 @@ def forward(
 
 def pooler_for_embed(pooler_config: PoolerConfig):
     pooling = get_seq_pooling_method(pooler_config.get_seq_pooling_type())
-    head = EmbeddingPoolerHead()
+
+    vllm_config = get_current_vllm_config()
+    model_config = vllm_config.model_config
+    head = EmbeddingPoolerHead(
+        projector=_load_st_projector(model_config),
+        head_dtype=model_config.head_dtype,
+        activation=PoolerNormalize(),
+    )
 
     return SequencePooler(pooling=pooling, head=head)
 
@@ -101,6 +113,15 @@ def pooler_for_classify(
     if pooling is None:
         pooling = get_seq_pooling_method(pooler_config.get_seq_pooling_type())
 
-    head = ClassifierPoolerHead(classifier=classifier, act_fn=act_fn)
+    vllm_config = get_current_vllm_config()
+    model_config = vllm_config.model_config
+    head = ClassifierPoolerHead(
+        classifier=classifier,
+        logit_bias=model_config.pooler_config.logit_bias,
+        head_dtype=model_config.head_dtype,
+        activation=resolve_classifier_act_fn(
+            model_config, static_num_labels=True, act_fn=act_fn
+        ),
+    )
 
     return SequencePooler(pooling=pooling, head=head)

vllm/model_executor/layers/pooler/tokwise/heads.py

@@ -7,14 +7,7 @@
 import torch
 import torch.nn as nn
 
-from vllm.config import get_current_vllm_config
-from vllm.model_executor.layers.pooler import ClassifierFn
-from vllm.model_executor.layers.pooler.activations import (
-    PoolerActivation,
-    PoolerNormalize,
-    resolve_classifier_act_fn,
-)
-from vllm.model_executor.models.adapters import _load_st_projector
+from vllm.model_executor.layers.pooler import ActivationFn, ClassifierFn
 from vllm.pooling_params import PoolingParams
 from vllm.tasks import PoolingTask
 from vllm.v1.pool.metadata import PoolingMetadata
@@ -49,17 +42,17 @@ def forward(
 
 
 class TokenEmbeddingPoolerHead(TokenPoolerHead):
-    def __init__(self) -> None:
+    def __init__(
+        self,
+        projector: nn.Module | None = None,
+        head_dtype: str | None = None,
+        activation: ActivationFn | None = None,
+    ) -> None:
         super().__init__()
 
-        # Load ST projector if available
-        vllm_config = get_current_vllm_config()
-        model_config = vllm_config.model_config
-
-        self.projector = _load_st_projector(model_config)
-        self.head_dtype = model_config.head_dtype
-
-        self.activation = PoolerNormalize()
+        self.projector = projector
+        self.head_dtype = head_dtype
+        self.activation = activation
 
     def get_supported_tasks(self) -> Set[PoolingTask]:
         return {"token_embed"}
@@ -85,7 +78,7 @@ def forward_chunk(
         pooled_data = pooled_data[..., : pooling_param.dimensions]
 
         # for normalize
-        if pooling_param.normalize:
+        if self.activation is not None and pooling_param.normalize:
             pooled_data = self.activation(pooled_data)
 
         # pooled_data shape: [n_tokens, embedding_dimension]
@@ -96,20 +89,16 @@ class TokenClassifierPoolerHead(TokenPoolerHead):
     def __init__(
         self,
         classifier: ClassifierFn | None = None,
-        act_fn: PoolerActivation | str | None = None,
+        logit_bias: float | None = None,
+        head_dtype: str | None = None,
+        activation: ActivationFn | None = None,
     ) -> None:
         super().__init__()
 
-        vllm_config = get_current_vllm_config()
-        model_config = vllm_config.model_config
-
         self.classifier = classifier
-        self.logit_bias: float | None = model_config.pooler_config.logit_bias
-        self.head_dtype = model_config.head_dtype
-
-        self.act_fn = resolve_classifier_act_fn(
-            model_config, static_num_labels=False, act_fn=act_fn
-        )
+        self.logit_bias = logit_bias
+        self.head_dtype = head_dtype
+        self.activation = activation
 
     def get_supported_tasks(self) -> Set[PoolingTask]:
         return {"token_classify"}
@@ -135,8 +124,8 @@ def forward_chunk(
         if self.logit_bias is not None:
             scores -= self.logit_bias
 
-        if pooling_param.use_activation:
-            scores = self.act_fn(scores)
+        if self.activation is not None and pooling_param.use_activation:
+            scores = self.activation(scores)
 
         # scores shape: [n_token, num_labels]
         return scores

vllm/model_executor/layers/pooler/tokwise/poolers.py

@@ -5,10 +5,15 @@
 
 import torch
 
-from vllm.config import PoolerConfig
+from vllm.config import PoolerConfig, get_current_vllm_config
 from vllm.model_executor.layers.pooler import ClassifierFn, PoolingParamsUpdate
 from vllm.model_executor.layers.pooler.abstract import Pooler
-from vllm.model_executor.layers.pooler.activations import PoolerActivation
+from vllm.model_executor.layers.pooler.activations import (
+    PoolerActivation,
+    PoolerNormalize,
+    resolve_classifier_act_fn,
+)
+from vllm.model_executor.models.adapters import _load_st_projector
 from vllm.tasks import POOLING_TASKS, PoolingTask
 from vllm.v1.pool.metadata import PoolingMetadata
 
@@ -86,7 +91,14 @@ def forward(
 
 def pooler_for_token_embed(pooler_config: PoolerConfig):
     pooling = get_tok_pooling_method(pooler_config.get_tok_pooling_type())
-    head = TokenEmbeddingPoolerHead()
+
+    vllm_config = get_current_vllm_config()
+    model_config = vllm_config.model_config
+    head = TokenEmbeddingPoolerHead(
+        projector=_load_st_projector(model_config),
+        head_dtype=model_config.head_dtype,
+        activation=PoolerNormalize(),
+    )
 
     return TokenPooler(pooling=pooling, head=head)
 
@@ -101,6 +113,15 @@ def pooler_for_token_classify(
     if pooling is None:
         pooling = get_tok_pooling_method(pooler_config.get_tok_pooling_type())
 
-    head = TokenClassifierPoolerHead(classifier=classifier, act_fn=act_fn)
+    vllm_config = get_current_vllm_config()
+    model_config = vllm_config.model_config
+    head = TokenClassifierPoolerHead(
+        classifier=classifier,
+        logit_bias=model_config.pooler_config.logit_bias,
+        head_dtype=model_config.head_dtype,
+        activation=resolve_classifier_act_fn(
+            model_config, static_num_labels=False, act_fn=act_fn
+        ),
+    )
 
     return TokenPooler(pooling=pooling, head=head)

vllm/model_executor/models/bert.py

@@ -25,10 +25,9 @@
     PoolingParamsUpdate,
 )
 from vllm.model_executor.layers.pooler.seqwise import (
+    EmbeddingPoolerHead,
     SequencePooler,
-    SequencePoolerHeadOutput,
     SequencePoolerOutput,
-    SequencePoolingMethodOutput,
     get_seq_pooling_method,
 )
 from vllm.model_executor.layers.pooler.tokwise import (
@@ -94,26 +93,24 @@ def forward(
 
 
 class BertPooler(SequencePooler):
-    def __init__(self, config: BertConfig, pooler_config: PoolerConfig):
-        super().__init__(
-            pooling=get_seq_pooling_method(pooler_config.seq_pooling_type),
-            head=self.head,
-        )
+    def __init__(self, vllm_config: VllmConfig):
+        model_config = vllm_config.model_config
+        pooler_config = vllm_config.pooler_config
+        assert pooler_config is not None
+
+        config: BertConfig = model_config.hf_config
 
         self.dense = nn.Linear(config.hidden_size, config.hidden_size)
         self.activation = nn.Tanh()
 
-    def head(
-        self,
-        pooled_data: SequencePoolingMethodOutput,
-        pooling_metadata: PoolingMetadata,
-    ) -> SequencePoolerHeadOutput:
-        if isinstance(pooled_data, list):
-            pooled_data = torch.stack(pooled_data)
-
-        pooled_data = self.dense(pooled_data)
-        pooled_data = self.activation(pooled_data)
-        return pooled_data
+        super().__init__(
+            pooling=get_seq_pooling_method(pooler_config.seq_pooling_type),
+            head=EmbeddingPoolerHead(
+                projector=self.dense,
+                head_dtype=model_config.head_dtype,
+                activation=self.activation,
+            ),
+        )
 
 
 class BertEncoder(nn.Module):
@@ -449,12 +446,7 @@ def __init__(
             embedding_class=embedding_class,
         )
 
-        config = vllm_config.model_config.hf_config
-
-        pooler_config = vllm_config.model_config.pooler_config
-        assert pooler_config is not None
-
-        self.pooler = BertPooler(config, pooler_config)
+        self.pooler = BertPooler(vllm_config)
 
     def load_weights(self, weights: Iterable[tuple[str, torch.Tensor]]) -> set[str]:
         other_weights, loaded_stacked_params = self._load_weights(weights)