Correct prompt?
Was looking at the source model discussion and someone mentioned that the correct prompt is:
A chat between a curious human and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions.
Human: {input}
Assistant: {output}
(https://huggingface.co/timdettmers/guanaco-33b-merged/discussions/4)
Not sure if they are correct or not, but perhaps something to consider for updating your model cards.
Thanks again for doing all this !
That looks mostly like that LLaVA.yaml template. Do I need to modify it, or can I use it as is?
user: "### Human:"
bot: "### Assistant:"
turn_template: "<|user|> <|user-message|><|bot|> <|bot-message|>\n"
context: "You are LLaVA, a large language and vision assistant trained by UW Madison WAIV Lab. You are able to understand the visual content that the user provides, and assist the user with a variety of tasks using natural language. Follow the instructions carefully and explain your answers in detail.### Human: Hi!### Assistant: Hi there! How can I help you today?\n"
OK interesting, thanks @mancub
I tested it with Alpaca prompt and it worked OK. But maybe it works even better with this.
I will test and then update the README as appropriate. And you are right this should be in the README. I have mentioned prompt templates when it's unusual but not specifically when I think it's Alpaca format. It should really be there for all of them.
@DragonForged I'd definitely modify the context to match what mancub showed. I can't imagine all that other stuff will help, and it might confuse it.