The prompt template for Yi-34b

[aisensiy]

I am using yi-34b-chat.Q4_K_M.gguf with llama.cpp , but I have issues like this:

yi-error.mp4

Why the llm continue print out things after <|im_end|>. How can I config it to make it work like chatgpt.

[Jeximo]

Why the llm continue print out things after <|im_end|>

Do you use --chat-ml parameter?