biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[CI/Build][CPU] Fix CPU CI and remove all CPU V0 files (#20560)

Browse Source 
Signed-off-by: jiang1.li <jiang1.li@intel.com>
This commit is contained in:
Li, Jiang
2025-07-08 13:13:44 +08:00
committed by GitHub
parent 8369b7c2a9
commit 7721ef1786
9 changed files with 785 additions and 839 deletions
Download Patch File Download Diff File Expand all files Collapse all files

58
tests/basic_correctness/test_chunked_prefill.py
View File

@@ -294,61 +294,3 @@ def test_with_prefix_caching(
name_0="w/o prefix caching",
name_1="with prefix caching",
)
@pytest.mark.parametrize("model", ["facebook/opt-125m"])
@pytest.mark.parametrize("dtype", ["bfloat16", "half"])
@pytest.mark.parametrize("max_tokens", [32])
@pytest.mark.parametrize("chunked_prefill_token_size", [1, 4, 16])
@pytest.mark.parametrize("enforce_eager", [False])
@pytest.mark.parametrize("attention_backend", ["TORCH_SDPA"])
@pytest.mark.cpu_model
@pytest.mark.skipif(not current_platform.is_cpu(), reason="CPU only")
def test_models_cpu(
hf_runner: HfRunner,
vllm_runner: VllmRunner,
example_prompts,
model: str,
dtype: str,
max_tokens: int,
chunked_prefill_token_size: int,
enforce_eager: bool,
attention_backend: str,
monkeypatch: pytest.MonkeyPatch,
) -> None:
test_models(
hf_runner,
vllm_runner,
example_prompts,
model,
dtype,
max_tokens,
chunked_prefill_token_size,
enforce_eager,
1,
attention_backend,
monkeypatch,
)
@pytest.mark.parametrize("max_tokens", [16])
@pytest.mark.parametrize("enforce_eager", [False])
@pytest.mark.parametrize("chunk_size", [30, 32])
@pytest.mark.parametrize("dtype", ["bfloat16", "half"])
@pytest.mark.cpu_model
@pytest.mark.skipif(not current_platform.is_cpu(), reason="CPU only")
def test_with_prefix_caching_cpu(
vllm_runner: VllmRunner,
max_tokens: int,
enforce_eager: bool,
chunk_size: int,
dtype: str,
) -> None:
test_with_prefix_caching(
vllm_runner,
max_tokens,
enforce_eager,
chunk_size,
1,
dtype,
)