After vllm is started, a template problem will be prompted

#1
by Gin10086 - opened

Can it be solved?

Yes I will work on it today. I suspect it's an issue with the tokenizer config from the base model not adhering to original Llama template.

I have fixed the chat template setting. Let me know if any other issues arise.

ZeroXClem changed discussion status to closed

Sign up or log in to comment