Merge tag 'v0.7.2' into v0.7.2-dev

tests/quantization/test_ipex_quant.py

+# SPDX-License-Identifier: Apache-2.0
 """Test model set-up and inference for quantized HF models supported
  on the CPU/GPU backend using IPEX (including AWQ/GPTQ).
  

tests/quantization/test_lm_head.py

+# SPDX-License-Identifier: Apache-2.0
 """Tests whether gptq models with quantized lm_head can be loaded.
 
 Run `pytest tests/quantization/test_quant_lm_head_true.py --forked`.

tests/quantization/test_quark.py

+# SPDX-License-Identifier: Apache-2.0
 """Test model set-up and weight loading for quark-quantized models.
 
 Run `pytest tests/quantization/test_quark.py`.

tests/quantization/test_register_quantization_config.py

+# SPDX-License-Identifier: Apache-2.0
 """Tests register custom quantization config.
 
 See https://github.com/vllm-project/vllm/issues/11926 for more details.

tests/quantization/utils.py

+# SPDX-License-Identifier: Apache-2.0
+
 from vllm.model_executor.layers.quantization import get_quantization_config
 from vllm.platforms import current_platform
 

tests/runai_model_streamer/test_runai_model_streamer_loader.py

+# SPDX-License-Identifier: Apache-2.0
+
 from vllm import SamplingParams
 from vllm.config import LoadConfig, LoadFormat
 from vllm.model_executor.model_loader.loader import (RunaiModelStreamerLoader,

tests/runai_model_streamer/test_weight_utils.py

+# SPDX-License-Identifier: Apache-2.0
+
 import glob
 import tempfile
 

tests/samplers/test_beam_search.py

+# SPDX-License-Identifier: Apache-2.0
 """Compare the outputs of HF and vLLM when using beam search.
 
 Run `pytest tests/samplers/test_beam_search.py`.

tests/samplers/test_ignore_eos.py

+# SPDX-License-Identifier: Apache-2.0
 """Make sure ignore_eos works.
 
 Run `pytest tests/samplers/test_ignore_eos.py`.

tests/samplers/test_logits_processor.py

+# SPDX-License-Identifier: Apache-2.0
+
 import pytest
 import torch
 import os

tests/samplers/test_logprobs.py

+# SPDX-License-Identifier: Apache-2.0
+
 from typing import List
 
 import pytest

tests/samplers/test_no_bad_words.py

+# SPDX-License-Identifier: Apache-2.0
 """Make sure bad_words works.
 
 Run `pytest tests/samplers/test_no_bad_words.py`.

tests/samplers/test_ranks.py

+# SPDX-License-Identifier: Apache-2.0
+
 import pytest
 import os
 

tests/samplers/test_rejection_sampler.py

+# SPDX-License-Identifier: Apache-2.0
 """Tests for rejection sampling."""
 from typing import List, Tuple
 

tests/samplers/test_sampler.py

+# SPDX-License-Identifier: Apache-2.0
+
 import itertools
 import random
 from dataclasses import dataclass

tests/samplers/test_seeded_generate.py

+# SPDX-License-Identifier: Apache-2.0
 """Verify that seeded random sampling is deterministic.
 
 Run `pytest tests/samplers/test_seeded_generate.py`.

tests/samplers/test_typical_acceptance_sampler.py

+# SPDX-License-Identifier: Apache-2.0
 """Tests for rejection sampling."""
 
 import pytest

tests/spec_decode/e2e/conftest.py

+# SPDX-License-Identifier: Apache-2.0
+
 from itertools import cycle
 from typing import List, Optional, Sequence, Tuple, Union
 

tests/spec_decode/e2e/test_compatibility.py

+# SPDX-License-Identifier: Apache-2.0
+
 import pytest
 import os
 

tests/spec_decode/e2e/test_eagle_correctness.py

+# SPDX-License-Identifier: Apache-2.0
 """This docstring details important information on the testing methodology.
 
 Most of the tests rely on "greedy equality", where we expect the output of