tokenizer_config.json

{
  "add_bos_token": true,
  "add_eos_token": false,
  "add_prefix_space": true,
  "added_tokens_decoder": {
    "0": {
      "content": "<unk>",
      "lstrip": false,
      "normalized": false,
      "rstrip": false,
      "single_word": false,
      "special": true
    },
    "1": {
      "content": "<s>",
      "lstrip": false,
      "normalized": false,
      "rstrip": false,
      "single_word": false,
      "special": true
    },
    "2": {
      "content": "</s>",
      "lstrip": false,
      "normalized": false,
      "rstrip": false,
      "single_word": false,
      "special": true
    },
    "32000": {
      "content": "<image>",
      "lstrip": false,
      "normalized": false,
      "rstrip": false,
      "single_word": false,
      "special": true
    },
    "32001": {
      "content": "<pad>",
      "lstrip": false,
      "normalized": false,
      "rstrip": false,
      "single_word": false,
      "special": true
    }
  },
  "additional_special_tokens": [
    "<image>"
  ],
  "bos_token": "<s>",
  "chat_template": "\n{%- set preamble = \"A chat between a curious human and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the human's questions. \" -%}\n{{ preamble }}\n{%- for m in messages %}\n    {%- if m[\"role\"] == \"user\" %}\n        {%- if m[\"content\"] is string -%}\n            USER: {{ m[\"content\"] }} ASSISTANT:\n        {%- elif m[\"content\"] is iterable -%}\n            {%- set user_prompt = m[\"content\"] | selectattr(\"type\", \"equalto\", \"text\") | map(attribute=\"text\") | first -%}\n            {%- set image = m[\"content\"] | selectattr(\"type\", \"equalto\", \"image_url\") | map(attribute=\"image_url\") | first -%}\n            {%- if \"<image>\" not in user_prompt -%}\n                USER: <image>\n{{ user_prompt }} ASSISTANT:\n            {%- else -%}\n                USER: {{ user_prompt }} ASSISTANT:\n            {%- endif -%}\n        {%- endif -%}\n    {%- elif m[\"role\"] == \"assistant\" -%}\n        {{ \" \" + m[\"content\"] }}</s>\n    {%- endif -%}\n{% endfor %}\n",
  "clean_up_tokenization_spaces": false,
  "eos_token": "</s>",
  "legacy": false,
  "model_max_length": 4096,
  "pad_token": "<unk>",
  "padding_side": "left",
  "processor_class": "LlavaProcessor",
  "sp_model_kwargs": {},
  "spaces_between_special_tokens": false,
  "tokenizer_class": "LlamaTokenizer",
  "unk_token": "<unk>",
  "use_default_system_prompt": false
}