Meaningless results when running demo.py

[Qihang Cao]

Hey guys! I encountered meaningless generated results when running the demo.py of imageBind_LLM.
Has anyone encountered a similar situation or can anyone provide me with some advice?

[yyyddff]

Me too, but I used my own pre trained model that resulted in meaningless sentences

[Jiaming Han]

Hi @Berta911 , have you correctly loaded the OpenChinese LLaMA weights? Our trained checkpoint does not include LLaMA's weights.

Besides, can you show me you full demo code?

[Qihang Cao]

Thanks for your advice! Problem was resolved when OpenChinese LLaMA weights were correctly loaded.

[somin0214]

where can I get OpenChinese LLaMA weights? I can't find it