[V0 Deprecation] Remove pooling model support in V0 (#23434)

Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
Signed-off-by: Max de Bayser <mbayser@br.ibm.com>
Co-authored-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>

tests/entrypoints/llm/test_classify.py

@@ -16,14 +16,6 @@ MODEL_NAME = "jason9693/Qwen2.5-1.5B-apeach"
 prompts = ["The chef prepared a delicious meal."]
 
 
-@pytest.fixture(autouse=True)
-def v1(run_with_both_engines):
-    # Simple autouse wrapper to run both engines for each test
-    # This can be promoted up to conftest.py to run for every
-    # test in a package
-    pass
-
-
 @pytest.fixture(scope="module")
 def llm():
     # pytest caches the fixture so we use weakref.proxy to

tests/entrypoints/llm/test_encode.py

@@ -27,14 +27,6 @@ TOKEN_IDS = [
 ]
 
 
-@pytest.fixture(autouse=True)
-def v1(run_with_both_engines):
-    # Simple autouse wrapper to run both engines for each test
-    # This can be promoted up to conftest.py to run for every
-    # test in a package
-    pass
-
-
 @pytest.fixture(scope="module")
 def llm():
     # pytest caches the fixture so we use weakref.proxy to

tests/entrypoints/llm/test_reward.py

@@ -16,14 +16,6 @@ MODEL_NAME = "internlm/internlm2-1_8b-reward"
 prompts = ["The chef prepared a delicious meal."]
 
 
-@pytest.fixture(autouse=True)
-def v1(run_with_both_engines):
-    # Simple autouse wrapper to run both engines for each test
-    # This can be promoted up to conftest.py to run for every
-    # test in a package
-    pass
-
-
 @pytest.fixture(scope="module")
 def llm():
     # pytest caches the fixture so we use weakref.proxy to

tests/entrypoints/llm/test_score.py

@@ -14,14 +14,6 @@ from ...models.utils import softmax
 MODEL_NAME = "tomaarsen/Qwen3-Reranker-0.6B-seq-cls"
 
 
-@pytest.fixture(autouse=True)
-def v1(run_with_both_engines):
-    # Simple autouse wrapper to run both engines for each test
-    # This can be promoted up to conftest.py to run for every
-    # test in a package
-    pass
-
-
 @pytest.fixture(scope="module")
 def llm():
     # pytest caches the fixture so we use weakref.proxy to

tests/entrypoints/offline_mode/test_offline_mode.py

@@ -32,15 +32,16 @@ MODEL_CONFIGS = [
         "tensor_parallel_size": 1,
         "tokenizer_mode": "mistral",
     },
-    {
-        "model": "sentence-transformers/all-MiniLM-L12-v2",
-        "enforce_eager": True,
-        "gpu_memory_utilization": 0.20,
-        "max_model_len": 64,
-        "max_num_batched_tokens": 64,
-        "max_num_seqs": 64,
-        "tensor_parallel_size": 1,
-    },
+    # TODO: re-enable once these tests are run with V1
+    # {
+    #     "model": "sentence-transformers/all-MiniLM-L12-v2",
+    #     "enforce_eager": True,
+    #     "gpu_memory_utilization": 0.20,
+    #     "max_model_len": 64,
+    #     "max_num_batched_tokens": 64,
+    #     "max_num_seqs": 64,
+    #     "tensor_parallel_size": 1,
+    # },
 ]
 
 

tests/entrypoints/openai/test_embedding.py

@@ -24,14 +24,6 @@ DUMMY_CHAT_TEMPLATE = """{% for message in messages %}{{message['role'] + ': ' +
 DTYPE = "bfloat16"
 
 
-@pytest.fixture(autouse=True)
-def v1(run_with_both_engines):
-    # Simple autouse wrapper to run both engines for each test
-    # This can be promoted up to conftest.py to run for every
-    # test in a package
-    pass
-
-
 @pytest.fixture(scope="module")
 def server():
     args = [

tests/entrypoints/openai/test_rerank.py

@@ -14,14 +14,6 @@ MODEL_NAME = "BAAI/bge-reranker-base"
 DTYPE = "bfloat16"
 
 
-@pytest.fixture(autouse=True)
-def v1(run_with_both_engines):
-    # Simple autouse wrapper to run both engines for each test
-    # This can be promoted up to conftest.py to run for every
-    # test in a package
-    pass
-
-
 @pytest.fixture(scope="module")
 def server():
     args = ["--enforce-eager", "--max-model-len", "100", "--dtype", DTYPE]

tests/entrypoints/openai/test_score.py

@@ -12,15 +12,6 @@ from vllm.entrypoints.openai.protocol import ScoreResponse
 
 from ...utils import RemoteOpenAIServer
 
-
-@pytest.fixture(autouse=True)
-def v1(run_with_both_engines):
-    # Simple autouse wrapper to run both engines for each test
-    # This can be promoted up to conftest.py to run for every
-    # test in a package
-    pass
-
-
 MODELS = [
     {
         "name": "BAAI/bge-reranker-v2-m3",