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 15, 02:15 UTC
[CI/Build] Enlarge tolerance for a CPU multi-modal test (#20684)
Signed-off-by: jiang1.li <jiang1.li@intel.com>
Li, Jiang committed
Jul 10, 2025 at 01:48 UTC
e59ba9e14235dc25ae599bf14d49124d76f3b062
1 file changed
+1
tests/models/multimodal/generation/test_common.py
+1
@@ -152,6 +152,7 @@ VLM_TEST_SETTINGS = {
152
video_idx_to_prompt=lambda idx: "<|vision_bos|><|VIDEO|><|vision_eos|>", # noqa: E501
153
max_model_len=4096,
154
max_num_seqs=2,
155
+ num_logprobs= 6 if current_platform.is_cpu() else 5,
156
auto_cls=AutoModelForTextToWaveform,
157
vllm_output_post_proc=model_utils.qwen2_vllm_to_hf_output,
158
patch_hf_runner=model_utils.qwen2_5_omni_patch_hf_runner,