AKU-d_ms-0.5B-chat-v0.1 / tokenizer_config.json
YukiTomita-CC's picture
Upload folder using huggingface_hub
7fb28a1 verified
{
"add_bos_token": true,
"add_eos_token": false,
"add_prefix_space": false,
"added_tokens_decoder": {
"3": {
"content": "<UNK>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"32000": {
"content": "<BOS>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"32001": {
"content": "<EOS>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"32002": {
"content": "<PAD>",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
}
},
"bos_token": "<BOS>",
"chat_template": "{% for message in messages %}{% if message['role'] == 'user' %}<ROLE>User</ROLE>\n{{ message['content'] }}\n{% elif message['role'] == 'assistant' %}<ROLE>Aku</ROLE>\n{{ message['content'] }}<BOS>\n{% endif %}<ROLE>Aku</ROLE>\n{% endfor %}",
"clean_up_tokenization_spaces": false,
"eos_token": "<EOS>",
"legacy": false,
"model_max_length": 1000000000000000019884624838656,
"pad_token": "<PAD>",
"sp_model_kwargs": {},
"spaces_between_special_tokens": false,
"tokenizer_class": "LlamaTokenizer",
"unk_token": "<UNK>",
"use_default_system_prompt": false
}