[Model][2/N] Improve all pooling task | Support multi-vector retrieval (#25370)

Signed-off-by: wang.yuqi <noooop@126.com>

tests/entrypoints/pooling/openai/test_rerank.py

@@ -7,7 +7,7 @@ import torch
 import torch.nn.functional as F
 
 from tests.utils import RemoteOpenAIServer
-from vllm.entrypoints.openai.protocol import RerankResponse
+from vllm.entrypoints.openai.protocol import PoolingResponse, RerankResponse
 
 MODEL_NAME = "BAAI/bge-reranker-base"
 DTYPE = "bfloat16"
@@ -159,3 +159,20 @@ async def test_activation(server: RemoteOpenAIServer, model_name: str):
     assert torch.allclose(F.sigmoid(wo_activation), w_activation, atol=1e-2), (
         "w_activation should be close to activation(wo_activation)."
     )
+
+
+@pytest.mark.asyncio
+@pytest.mark.parametrize("model_name", [MODEL_NAME])
+async def test_pooling(server: RemoteOpenAIServer, model_name: str):
+    input_text = ["The chef prepared a delicious meal."]
+
+    response = requests.post(
+        server.url_for("pooling"),
+        json={"model": model_name, "input": input_text, "encoding_format": "float"},
+    )
+
+    poolings = PoolingResponse.model_validate(response.json())
+
+    assert len(poolings.data) == 1
+    assert len(poolings.data[0].data) == 11
+    assert len(poolings.data[0].data[0]) == 1