[V1] Support LLM.apply_model (#18465)

Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
2025-09-20 15:14:35 +08:00
parent be874c0201
commit 3d9a1d2de5
17 changed files with 194 additions and 169 deletions
--- a/tests/quantization/test_lm_head.py
+++ b/tests/quantization/test_lm_head.py
@@ -29,8 +29,8 @@ def test_lm_head(
    lm_head_quantized: bool,
    monkeypatch,
 ) -> None:
-    # vllm_runner.apply_model() relies on V0 internals.
-    monkeypatch.setenv("VLLM_USE_V1", "0")
+    # `LLM.apply_model` requires pickling a function.
+    monkeypatch.setenv("VLLM_ALLOW_INSECURE_SERIALIZATION", "1")
    with vllm_runner(model_id, dtype=torch.float16,
                     max_model_len=2048) as vllm_model: