HuggingFaceTB

/

SmolLM-1.7B-Instruct

like 96

Its ChatML (this is from the model's tokenizer_config.json file):

"bos_token": "<|im_start|>",
  "chat_template": "{% for message in messages %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
  "clean_up_tokenization_spaces": false,
  "eos_token": "<|im_end|>",

This prompt template doesn’t really work?