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, 23:45 UTC
[Bugfix] Investigate Qwen2-VL failing test (#21527)
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Isotr0py committed
Jul 26, 2025 at 21:09 UTC
f27fdfc3ed79689ec98ef1c609a31d8c3b11e6a9
1 file changed
+1
tests/models/multimodal/generation/test_common.py
+1
@@ -677,6 +677,7 @@ VLM_TEST_SETTINGS = {
677
prompt_formatter=lambda img_prompt: f"<|im_start|>User\n{img_prompt}<|im_end|>\n<|im_start|>assistant\n", # noqa: E501
678
img_idx_to_prompt=lambda idx: "<|vision_start|><|image_pad|><|vision_end|>", # noqa: E501
679
video_idx_to_prompt=lambda idx: "<|vision_start|><|video_pad|><|vision_end|>", # noqa: E501
680
+ multi_image_prompt="Picture 1: <vlm_image>\nPicture 2: <vlm_image>\nDescribe these two images with one paragraph respectively.", # noqa: E501
681
max_model_len=4096,
682
max_num_seqs=2,
683
auto_cls=AutoModelForVision2Seq,