[V0 Deprecation] Remove LLMEngine (#25033)

Signed-off-by: Woosuk Kwon <woosuk@thinkingmachines.ai>
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>

tests/basic_correctness/test_basic_correctness.py

@@ -11,7 +11,7 @@ from unittest.mock import Mock
 import pytest
 import torch
 
-from vllm import LLM, envs
+from vllm import LLM
 from vllm.v1.engine.llm_engine import LLMEngine as LLMEngineV1
 
 from ..conftest import HfRunner, VllmRunner
@@ -26,14 +26,6 @@ MODELS = [
 TARGET_TEST_SUITE = os.environ.get("TARGET_TEST_SUITE", "L4")
 
 
-@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
-
-
 def test_vllm_gc_ed():
     """Verify vllm instance is GC'ed when it is deleted"""
     llm = LLM("distilbert/distilgpt2")
@@ -76,12 +68,6 @@ def test_models(
     model_executor: str,
     enable_prompt_embeds: bool,
 ) -> None:
-    if not envs.VLLM_USE_V1:
-        if async_scheduling:
-            pytest.skip("async_scheduling only supported in v1.")
-        if model_executor != "uni":
-            pytest.skip("only test uniproc executor for v0.")
-
     if backend == "XFORMERS" and model == "google/gemma-2-2b-it":
         pytest.skip(
             f"{backend} does not support gemma2 with full context length.")

tests/basic_correctness/test_cumem.py

@@ -122,11 +122,12 @@ def test_cumem_with_cudagraph():
         # sleep mode with safetensors
         ("meta-llama/Llama-3.2-1B", True),
         # sleep mode with pytorch checkpoint
-        ("facebook/opt-125m", False),
+        ("facebook/opt-125m", True),
     ])
 def test_end_to_end(monkeypatch: pytest.MonkeyPatch, model: str, use_v1: bool):
     with monkeypatch.context() as m:
-        m.setenv("VLLM_USE_V1", "1" if use_v1 else "0")
+        assert use_v1
+        m.setenv("VLLM_USE_V1", "1")
         free, total = torch.cuda.mem_get_info()
         used_bytes_baseline = total - free  # in case other process is running
         llm = LLM(model, enable_sleep_mode=True)