How to batch process如何批量处理？

[logwhy]

我希望使用程序调用janus-pro-7b，让其自动根据提示批量处理图片，要如何做？
I want to use a program to call janus-pro-7b and let it automatically process pictures in batches according to prompts. How can I do this?
根据deepseek模型和chatgpt的方法在使用transfomers库的时候总是提示无法识别多态模型，在使用vllm的时候总是显示路径不正确（即使我布置了正确的路径）
According to the deepseek model and chatgpt method, when using the transfomers library, it always prompts that the polymorphic model cannot be recognized. When using vllm, it always shows that the path is incorrect (even if I have arranged the correct path)

[HaFred]

Increase parallel_size