formats.json

[
  {
    "model_family": "SmolLM2",
    "chat_format": "llama-3",
    "example": "<|im_start|>user\\n{msg}<|im_end|>\\n<|im_start|>assistant\\n"
  },
  {
    "model_family": "Llama-3.2",
    "chat_format": "llama-3",
    "example": "<|begin_of_text|><|start_header_id|>user<|end_header_id|>\\n{msg}<|eot_id|>"
  },
  {
    "model_family": "Qwen2.5",
    "chat_format": "chatml",
    "example": "<|im_start|>user\\n{msg}<|im_end|>\\n<|im_start|>assistant\\n"
  },
  {
    "model_family": "Gemma-2",
    "chat_format": "gemma",
    "example": "<start_of_turn>user\\n{msg}<end_of_turn>\\n<start_of_turn>model\\n"
  },
  {
    "model_family": "Phi-3.5",
    "chat_format": "chatml",
    "example": "<|im_start|>user\\n{msg}<|im_end|>\\n<|im_start|>assistant\\n"
  },
  {
    "model_family": "TinyLlama",
    "chat_format": "chatml",
    "example": "<|im_start|>user\\n{msg}<|im_end|>\\n<|im_start|>assistant\\n"
  },
  {
    "model_family": "MiniCPM",
    "chat_format": "chatml",
    "example": "<|im_start|>user\\n{msg}<|im_end|>\\n<|im_start|>assistant\\n"
  }
]