[CI/Build] Remove V0 LoRA test (#19066)

Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>

tests/lora/test_add_lora.py

@@ -6,6 +6,8 @@ import pytest
 
 import vllm.envs as env
 from vllm.engine.arg_utils import AsyncEngineArgs
+from vllm.entrypoints.openai.api_server import (
+    build_async_engine_client_from_engine_args)
 from vllm.inputs import TextPrompt
 from vllm.lora.request import LoRARequest
 from vllm.sampling_params import SamplingParams
@@ -16,14 +18,6 @@ LORA_RANK = 64
 DEFAULT_MAX_LORAS = 4 * 3
 
 
-@pytest.fixture(autouse=True)
-def v1(run_with_both_engines_lora):
-    # 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
-
-
 def get_lora_requests(lora_path) -> list[LoRARequest]:
     lora_requests: list[LoRARequest] = [
         LoRARequest(lora_name=f"{i}", lora_int_id=i, lora_path=lora_path)
@@ -88,17 +82,6 @@ async def test_add_lora(chatglm3_lora_files):
         trust_remote_code=True,
         enforce_eager=True)
 
-    # The run_with_both_engines_lora fixture sets up the `VLLM_USE_V1`
-    # environment variable. reload vllm.enging.async_llm_engine as
-    # vllm.engine.async_llm_engine.AsyncLLMEgnine changes depending on the
-    # env var.
-    import importlib
-
-    import vllm.engine.async_llm_engine
-    importlib.reload(vllm.engine.async_llm_engine)
-    from vllm.entrypoints.openai.api_server import (
-        build_async_engine_client_from_engine_args)
-
     # split lora_requests into 3 parts
     part_size = len(lora_requests) // 3
     dummy_run_requests = lora_requests[:part_size]

tests/lora/test_chatglm3_tp.py

 # SPDX-License-Identifier: Apache-2.0
 
-import pytest
-
 import vllm
 from vllm.lora.request import LoRARequest
 
@@ -18,14 +16,6 @@ EXPECTED_LORA_OUTPUT = [
 ]
 
 
-@pytest.fixture(autouse=True)
-def v1(run_with_both_engines_lora):
-    # 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
-
-
 def do_sample(llm: vllm.LLM, lora_path: str, lora_id: int) -> list[str]:
     prompts = [
         PROMPT_TEMPLATE.format(query="How many singers do we have?"),

tests/lora/test_llama_tp.py

@@ -33,14 +33,6 @@ EXPECTED_LORA_OUTPUT = [
 ]
 
 
-@pytest.fixture(autouse=True)
-def v1(run_with_both_engines_lora):
-    # 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
-
-
 def do_sample(llm: vllm.LLM,
               lora_path: str,
               lora_id: int,

tests/lora/test_lora_functions.py

@@ -2,26 +2,24 @@
 """
 Script to test add_lora, remove_lora, pin_lora, list_loras functions.
 """
-
-import os
-
 import pytest
 
 from vllm.engine.arg_utils import AsyncEngineArgs, EngineArgs
 from vllm.engine.llm_engine import LLMEngine
+from vllm.entrypoints.openai.api_server import (
+    build_async_engine_client_from_engine_args)
 from vllm.lora.request import LoRARequest
 
 MODEL_PATH = "meta-llama/Llama-2-7b-hf"
 LORA_MODULE_PATH = "yard1/llama-2-7b-sql-lora-test"
 LORA_RANK = 8
 
-
-@pytest.fixture(autouse=True)
-def v1(run_with_both_engines_lora):
-    # 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(autouse=True)
+# def v1(run_with_both_engines_lora):
+#     # 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
 
 
 def make_lora_request(lora_id: int):
@@ -79,22 +77,6 @@ def test_lora_functions_sync():
 @pytest.mark.asyncio
 async def test_lora_functions_async():
 
-    if os.getenv("VLLM_USE_V1") == "0":
-        pytest.skip(
-            reason=
-            "V0 AsyncLLMEngine does not expose remove/list/pin LoRA functions")
-
-    # The run_with_both_engines_lora fixture sets up the `VLLM_USE_V1`
-    # environment variable. reload vllm.enging.async_llm_engine as
-    # vllm.engine.async_llm_engine.AsyncLLMEgnine changes depending on the
-    # env var.
-    import importlib
-
-    import vllm.engine.async_llm_engine
-    importlib.reload(vllm.engine.async_llm_engine)
-    from vllm.entrypoints.openai.api_server import (
-        build_async_engine_client_from_engine_args)
-
     max_loras = 4
     engine_args = AsyncEngineArgs(model=MODEL_PATH,
                                   enable_lora=True,

tests/lora/test_mixtral.py

@@ -10,14 +10,6 @@ from vllm.platforms import current_platform
 MODEL_PATH = "mistralai/Mixtral-8x7B-Instruct-v0.1"
 
 
-@pytest.fixture(autouse=True)
-def v1(run_with_both_engines_lora):
-    # 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
-
-
 def do_sample(llm: vllm.LLM, lora_path: str, lora_id: int,
               prompts: list[str]) -> list[str]:
 

tests/lora/test_quant_model.py

@@ -37,14 +37,6 @@ else:
     ]
 
 
-@pytest.fixture(autouse=True)
-def v1(run_with_both_engines_lora):
-    # 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
-
-
 def do_sample(llm: vllm.LLM,
               lora_path: str,
               lora_id: int,

tests/lora/test_qwen2vl.py

@@ -13,14 +13,6 @@ from vllm.platforms import current_platform
 from vllm.sampling_params import BeamSearchParams
 
 
-@pytest.fixture(autouse=not current_platform.is_cpu())
-def v1(run_with_both_engines_lora):
-    # 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
-
-
 @dataclass
 class TestConfig:
     model_path: str

tests/lora/test_worker.py

@@ -6,8 +6,6 @@ import tempfile
 from typing import Union
 from unittest.mock import patch
 
-import pytest
-
 import vllm.envs as envs
 from vllm.config import (CacheConfig, DeviceConfig, LoadConfig, LoRAConfig,
                          ModelConfig, ParallelConfig, SchedulerConfig,
@@ -18,14 +16,6 @@ from vllm.v1.worker.gpu_worker import Worker as V1Worker
 from vllm.worker.worker import Worker
 
 
-@pytest.fixture(autouse=True)
-def v1(run_with_both_engines_lora):
-    # 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
-
-
 @patch.dict(os.environ, {"RANK": "0"})
 def test_worker_apply_lora(sql_lora_files):