xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-15 18:16:46 +08:00
Code Issues Packages Projects Releases Wiki Activity

[CI] Skip tests failing on main (#25326)

Browse Source 
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
This commit is contained in:
Woosuk Kwon 2025-09-20 19:57:46 -07:00 committed by GitHub
parent 572ddf83ce
commit 72dd1595b4
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
4 changed files with 10 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
tests/entrypoints/openai/test_completion_with_prompt_embeds.py
View File

@ -60,6 +60,7 @@ def create_dummy_embeds(num_tokens: int = 5) -> str:
return base64.b64encode(buffer.getvalue()).decode('utf-8') return base64.b64encode(buffer.getvalue()).decode('utf-8')
@pytest.mark.skip("This test is skipped because it is flaky.")
@pytest.mark.asyncio @pytest.mark.asyncio
@pytest.mark.parametrize("model_name", [MODEL_NAME]) @pytest.mark.parametrize("model_name", [MODEL_NAME])
async def test_completions_with_prompt_embeds( async def test_completions_with_prompt_embeds(

5
tests/models/quantization/test_fp8.py
View File

@ -32,7 +32,7 @@ from ..utils import check_logprobs_close
# Due to low-precision numerical divergence, we only test logprob of 4 tokens # Due to low-precision numerical divergence, we only test logprob of 4 tokens
@pytest.mark.parametrize("max_tokens", [4]) @pytest.mark.parametrize("max_tokens", [4])
@pytest.mark.parametrize("enforce_eager", [True]) @pytest.mark.parametrize("enforce_eager", [True])
@pytest.mark.parametrize("backend", ["FLASH_ATTN", "XFORMERS"]) @pytest.mark.parametrize("backend", ["FLASH_ATTN"])
# NOTE: Increasing this in this suite will fail CI because we currently cannot # NOTE: Increasing this in this suite will fail CI because we currently cannot
# reset distributed env properly. Use a value > 1 just when you test. # reset distributed env properly. Use a value > 1 just when you test.
@pytest.mark.parametrize("tensor_parallel_size", [1]) @pytest.mark.parametrize("tensor_parallel_size", [1])
@ -57,6 +57,9 @@ def test_models(
pytest.skip( pytest.skip(
f"{kv_cache_dtype} is currently not supported on ROCm/HIP.") f"{kv_cache_dtype} is currently not supported on ROCm/HIP.")
if not current_platform.is_kv_cache_dtype_supported(kv_cache_dtype, None):
pytest.skip(f"{kv_cache_dtype} is not supported on this platform.")
with monkeypatch.context() as m: with monkeypatch.context() as m:
m.setenv("TOKENIZERS_PARALLELISM", 'true') m.setenv("TOKENIZERS_PARALLELISM", 'true')
m.setenv(STR_BACKEND_ENV_VAR, backend) m.setenv(STR_BACKEND_ENV_VAR, backend)

1
tests/models/test_oot_registration.py
View File

@ -63,6 +63,7 @@ def test_oot_registration_embedding(
image = convert_image_mode(ImageAsset("cherry_blossom").pil_image, "RGB") image = convert_image_mode(ImageAsset("cherry_blossom").pil_image, "RGB")
@pytest.mark.skip(reason="This test is skipped because it failed on V1.")
@create_new_process_for_each_test() @create_new_process_for_each_test()
def test_oot_registration_multimodal( def test_oot_registration_multimodal(
monkeypatch: pytest.MonkeyPatch, monkeypatch: pytest.MonkeyPatch,

5
tests/quantization/test_compressed_tensors.py
View File

@ -357,6 +357,9 @@ def test_compressed_tensors_fp8(vllm_runner):
assert output assert output
@pytest.mark.skipif(
not current_platform.is_kv_cache_dtype_supported("fp8", None),
reason="FP8 KV cache is not supported on this device.")
@pytest.mark.skipif(not current_platform.is_cuda(), @pytest.mark.skipif(not current_platform.is_cuda(),
reason="This test is skipped on non-CUDA platform.") reason="This test is skipped on non-CUDA platform.")
def test_compressed_tensors_kv_cache(vllm_runner): def test_compressed_tensors_kv_cache(vllm_runner):
@ -738,4 +741,4 @@ def test_compressed_tensors_transforms_perplexity(vllm_runner, model, prompt,
with vllm_runner(model, enforce_eager=True) as llm: with vllm_runner(model, enforce_eager=True) as llm:
perplexity = llm.generate_prompt_perplexity([prompt])[0] perplexity = llm.generate_prompt_perplexity([prompt])[0]
print(perplexity) print(perplexity)
assert perplexity <= exp_perplexity assert perplexity <= exp_perplexity