TinyLlama1B-spider-v4-18500steps / tokenizer_config.json
lleticiasilvaa's picture
Upload tokenizer
bcc7b53 verified
raw
history blame contribute delete
No virus
2.89 kB
{
"add_bos_token": true,
"add_eos_token": false,
"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": "[SQL]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"32001": {
"content": "[/SQL]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"32002": {
"content": "[QUESTION]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
},
"32003": {
"content": "[/QUESTION]",
"lstrip": false,
"normalized": false,
"rstrip": false,
"single_word": false,
"special": true
}
},
"additional_special_tokens": [
"[SQL]",
"[/SQL]",
"[QUESTION]",
"[/QUESTION]"
],
"bos_token": "<s>",
"chat_template": "{% if messages[0]['role'] == 'system' %}{% set loop_messages = messages[1:] %}{% set system_message = messages[0]['content'] %}{% else %}{% set loop_messages = messages %}{% set system_message = 'You are an SQL expert tasked with answering questions about an SQLite database.\nYour responses must always consist of valid SQL code and only that.' %}{% endif %}{{ '# <|system|>\n' + system_message + '\n\n' }}{% if messages|selectattr(\"role\", \"equalto\", \"example\")|list %}Below are some examples of question and their corresponding SQL queries:\n\n{% else %}\n{% endif %}{% for message in loop_messages %}{% if message['role'] == 'example' %}{{ message['content'] }}\n\n{% elif message['role'] == 'schema' %}# <|schema|>\nThe query will run on a database with the following schema:\n{{ message['content'] }}\n\n{% elif message['role'] == 'user' %}# <|user|>\n[QUESTION]{{ message['content'] }}[/QUESTION]\n\n{% elif message['role'] == 'assistant' %}# <|assistant|>\n[SQL]{{ message['content'] }}[/SQL]{% endif %}{% endfor %}{% if add_generation_prompt %}# <|assistant|>\n[SQL]{% endif %}",
"clean_up_tokenization_spaces": false,
"eos_token": "</s>",
"max_length": 2048,
"model_max_length": 2048,
"pad_token": "</s>",
"padding_side": "right",
"sp_model_kwargs": {},
"stride": 0,
"tokenizer_class": "LlamaTokenizer",
"truncation_side": "right",
"truncation_strategy": "longest_first",
"unk_token": "<unk>",
"use_default_system_prompt": false
}