tokenizer_config.json

{
  "add_prefix_space": false,
  "added_tokens_decoder": {
    "151643": {
      "content": "<|endoftext|>",
      "lstrip": false,
      "normalized": false,
      "rstrip": false,
      "single_word": false,
      "special": true
    },
    "151644": {
      "content": "<|im_start|>",
      "lstrip": false,
      "normalized": false,
      "rstrip": false,
      "single_word": false,
      "special": true
    },
    "151645": {
      "content": "<|im_end|>",
      "lstrip": false,
      "normalized": false,
      "rstrip": false,
      "single_word": false,
      "special": true
    },
    "151646": {
      "content": "<image>",
      "lstrip": false,
      "normalized": false,
      "rstrip": false,
      "single_word": false,
      "special": true
    }
  },
  "additional_special_tokens": [
    "<image>"
  ],
  "bos_token": "<|im_start|>",
  "chat_template": "{% if messages[0]['role'] == 'system' %}{{ bos_token + 'system\n' + messages[0]['content'] + eos_token}}{% set messages = messages[1:] %}{% endif %}{% for i in range(messages | length) %}{% set message = messages[i] %}{% if message['role'] == 'user' %}{% if add_stream_query_prompt %}{{ eos_token + '\n' + bos_token + 'user\n' + message['content'] + eos_token }}{% else %}{{ '\n' + bos_token + 'user\n' + message['content'] + eos_token }}{% endif %}{% elif message['role'] == 'assistant' %}{{ '\n' + bos_token + 'assistant\n' + message['content'] + eos_token }}{% elif message['role'] == 'stream' and message['num_frames'] > 0 %}{{ '\n' + bos_token + 'stream\n' + ''.join([49 * '<image>'] * message['num_frames']) + eos_token }}{% endif %}{% endfor %}{% if add_generation_prompt %}{{ '\n' + bos_token + 'assistant\n' }}{% elif add_stream_prompt %}{{ '\n' + bos_token + 'stream\n' }}{% elif add_stream_generation_prompt %}{{ eos_token + '\n' + bos_token + 'assistant\n' }}{% endif %}",
  "clean_up_tokenization_spaces": false,
  "eos_token": "<|im_end|>",
  "errors": "replace",
  "model_max_length": 32768,
  "pad_token": "<|endoftext|>",
  "padding_side": "left",
  "processor_class": "LlavaProcessor",
  "split_special_tokens": false,
  "tokenizer_class": "Qwen2Tokenizer",
  "unk_token": null
}