A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 14, 05:30 UTC
[CI/Build] Skip prompt embeddings tests on V1-only CPU backend (#24721)
Signed-off-by: jiang1.li <jiang1.li@intel.com>
Li, Jiang committed
Sep 12, 2025 at 18:51 UTC
59d5d2c736028d8a67f230149ad6470adb1e71e5
1 file changed
+6
tests/models/language/generation/test_common.py
+6
@@ -119,6 +119,12 @@ def test_models(hf_runner, vllm_runner, example_prompts, model: str,
119
# in parts of the operators
120
pytest.skip(f"Skipping '{model}' model test with AITER kernel.")
121
122
+ # Note: can be removed when
123
+ # https://github.com/vllm-project/vllm/pull/24278 finished
124
+ if current_platform.is_cpu() and use_prompt_embeds:
125
+ pytest.skip("Skipping use_prompt_embeds=True with "
126
+ "V1-only CPU backend.")
127
+
128
with hf_runner(model) as hf_model:
129
hf_outputs = hf_model.generate_greedy_logprobs_limit(
130
example_prompts, max_tokens, num_logprobs)