Merge tag 'v0.9.1' into v0.9.1-ori

tests/entrypoints/conftest.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 import pytest
 

tests/entrypoints/llm/test_accuracy.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 """
 This file test accuracy of the vLLM server via LMEval.
 It uses local-completions, which interacts with vLLM

tests/entrypoints/llm/test_chat.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 import weakref
 
 import pytest

tests/entrypoints/llm/test_collective_rpc.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 import pytest
 

tests/entrypoints/llm/test_encode.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 import weakref
 

tests/entrypoints/llm/test_generate.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 import weakref
 
@@ -24,6 +25,12 @@ TOKEN_IDS = [
 ]
 
 
+@pytest.fixture(autouse=True)
+def v1(run_with_both_engines):
+    """We can run both engines for this test."""
+    pass
+
+
 @pytest.fixture(scope="module")
 def llm():
     # pytest caches the fixture so we use weakref.proxy to
@@ -103,3 +110,19 @@ def test_multiple_sampling_params(llm: LLM):
     # sampling_params is None, default params should be applied
     outputs = llm.generate(PROMPTS, sampling_params=None)
     assert len(PROMPTS) == len(outputs)
+
+
+def test_max_model_len():
+    max_model_len = 20
+    llm = LLM(
+        model=MODEL_NAME,
+        max_model_len=max_model_len,
+        gpu_memory_utilization=0.10,
+        enforce_eager=True,  # reduce test time
+    )
+    sampling_params = SamplingParams(max_tokens=max_model_len + 10)
+    outputs = llm.generate(PROMPTS, sampling_params)
+    for output in outputs:
+        num_total_tokens = len(output.prompt_token_ids) + len(
+            output.outputs[0].token_ids)
+        assert num_total_tokens == max_model_len

tests/entrypoints/llm/test_generate_multiple_loras.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 import weakref
 

tests/entrypoints/llm/test_gpu_utilization.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 from vllm import LLM, SamplingParams
 

tests/entrypoints/llm/test_guided_generate.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 import json
 import weakref

tests/entrypoints/llm/test_init.py

-# SPDX-License-Identifier: Apache-2.0
-
-import pytest
-
-from vllm import LLM
-
-from ...utils import error_on_warning
-
-MODEL_NAME = "facebook/opt-125m"
-
-
-def test_pos_args_deprecated():
-    with error_on_warning(DeprecationWarning):
-        LLM(model=MODEL_NAME, tokenizer=MODEL_NAME)
-
-    with error_on_warning(DeprecationWarning):
-        LLM(MODEL_NAME, tokenizer=MODEL_NAME)
-
-    with pytest.warns(DeprecationWarning, match="'tokenizer'"):
-        LLM(MODEL_NAME, MODEL_NAME)
-
-    with pytest.warns(DeprecationWarning,
-                      match="'tokenizer', 'tokenizer_mode'"):
-        LLM(MODEL_NAME, MODEL_NAME, "auto")

tests/entrypoints/llm/test_lazy_outlines.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 import sys
 from contextlib import nullcontext

tests/entrypoints/llm/test_prompt_validation.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 import pytest
 

tests/entrypoints/offline_mode/test_offline_mode.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 """Tests for HF_HUB_OFFLINE mode"""
 import importlib
 import sys

tests/entrypoints/openai/correctness/test_lmeval.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 """
 This file test accuracy of the vLLM server via LMEval.
 It uses local-completions, which interacts with vLLM

tests/entrypoints/openai/correctness/test_mteb.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 import os
 
 import pytest
 
 from tests.models.language.pooling.mteb_utils import (MTEB_EMBED_TASKS,
+                                                      MTEB_EMBED_TOL,
                                                       OpenAIClientMtebEncoder,
                                                       run_mteb_embed_task,
                                                       run_mteb_embed_task_st)
@@ -38,4 +40,4 @@ def test_mteb(server):
     print("SentenceTransformer main score: ", st_main_score)
     print("Difference: ", st_main_score - vllm_main_score)
 
-    assert st_main_score == pytest.approx(vllm_main_score, rel=1e-4)
+    assert st_main_score == pytest.approx(vllm_main_score, abs=MTEB_EMBED_TOL)

tests/entrypoints/openai/correctness/test_transcription_api_correctness.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 """
 Evaluate Transcription API correctness by computing Word Error Rate (WER)
 on a given ASR dataset. When provided, it will also compare the WER against

tests/entrypoints/openai/test_async_tokenization.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 import asyncio
 import contextlib

tests/entrypoints/openai/test_audio.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 import json
 

tests/entrypoints/openai/test_basic.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 import asyncio
 from http import HTTPStatus

tests/entrypoints/openai/test_chat.py

 # SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
 
 # imports for guided decoding tests
 import json