Segmentation fault when process image with https://huggingface.co/taobao-mnn/MiniCPM-V-4-MNN

[gryffindor-rr]

cat image_prompt.txt <img>../tests/data/cat.png</img>描述图片内容

./build/llm_demo ~/Downloads/models/MiniCPM-V-4-MNN/config.json ./image_prompt.txt The device supports: i8sdot:1, fp16:1, i8mm: 1, sve2: 0, sme2: 1 config path is /Users/lzhang/Downloads/models/MiniCPM-V-4-MNN/config.json main, 258, cost time: 2333.435059 ms Prepare for tuning opt Begin Prepare for tuning opt End main, 262, cost time: 570.934998 ms prompt file is ./image_prompt.txt [1] 23282 segmentation fault ./build/llm_demo ~/Downloads/models/MiniCPM-V-4-MNN/config.json

If only set text prompt, it's OK: cat prompt.txt 你好

./build/llm_demo ~/Downloads/models/MiniCPM-V-4-MNN/config.json ./prompt.txt The device supports: i8sdot:1, fp16:1, i8mm: 1, sve2: 0, sme2: 1 config path is /Users/lzhang/Downloads/models/MiniCPM-V-4-MNN/config.json

你好！今天我能为您做些什么？

################################# prompt tokens num = 15 decode tokens num = 9 vision time = 0.00 s audio time = 0.00 s prefill time = 0.24 s decode time = 0.35 s sample time = 0.00 s prefill speed = 61.41 tok/s decode speed = 25.91 tok/s ##################################

It looks a image process issue. Any one can take a look?