ai-agents-chat-template-mismat-when-using-langchain-conversationchain-with-a-hugg-1c816ee5

THEN Ensure that the input prompt is formatted using the model's own chat template. For HuggingFace models, use tokenizer.apply_chat_template() or leverage LangChain's ChatHuggingFace wrapper which handles formatting automatically. Avoid manually writing a template that doesn't match the model's expected tokens (e.g., system/user/assistant).

Instead of a raw string template like 'Human: {input}\nAssistant:', use the model's chat template: messages = [HumanMessage(content=input)]; formatted = tokenizer.apply_chat_template(messages, tokenize=False). In LangChain, use ChatHuggingFace(model=model, tokenizer=tokenizer) and pass messages as a list of Message objects.