Can we use small model like LLAMA to get layout?

[hujunchao]

Sometimes, there is no way to use Chatgpt or GPT4. Can we use smaller model than GPT like llama to get layout? If we can use small model to get layout, It will be more and more popular for this job. Looking forward to your reply! Thanks!

[TonyLianLong]

Yes. I tried Vicuna (https://github.com/lm-sys/FastChat) and it also works. However, proprietary models (e.g. gpt-3.5-turbo) are still better.

[TonyLianLong]

I tried LLaMA-2-7b and it seems to work:

I used the non-chat version to perform the text completion.

[hujunchao]

Oh, It is amazing! I will try it.