Upload tokenizer
Browse files- tokenizer_config.json +5 -2
tokenizer_config.json
CHANGED
@@ -37,18 +37,21 @@
|
|
37 |
},
|
38 |
"additional_special_tokens": [],
|
39 |
"bos_token": "<s>",
|
|
|
40 |
"clean_up_tokenization_spaces": false,
|
41 |
"eos_token": "</s>",
|
42 |
"legacy": true,
|
43 |
"max_length": 2048,
|
44 |
-
"model_max_length":
|
45 |
"pad_token": "[PAD]",
|
|
|
46 |
"sp_model_kwargs": {},
|
47 |
"spaces_between_special_tokens": false,
|
|
|
48 |
"stride": 0,
|
49 |
"tokenizer_class": "LlamaTokenizer",
|
50 |
"truncation_side": "right",
|
51 |
"truncation_strategy": "longest_first",
|
52 |
"unk_token": "<unk>",
|
53 |
-
"use_default_system_prompt":
|
54 |
}
|
|
|
37 |
},
|
38 |
"additional_special_tokens": [],
|
39 |
"bos_token": "<s>",
|
40 |
+
"chat_template": "{% for message in messages %}{{bos_token + message['role'] + '\n' + message['content'] + eos_token + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ bos_token + 'assistant\n' }}{% endif %}",
|
41 |
"clean_up_tokenization_spaces": false,
|
42 |
"eos_token": "</s>",
|
43 |
"legacy": true,
|
44 |
"max_length": 2048,
|
45 |
+
"model_max_length": 8192,
|
46 |
"pad_token": "[PAD]",
|
47 |
+
"padding_side": "right",
|
48 |
"sp_model_kwargs": {},
|
49 |
"spaces_between_special_tokens": false,
|
50 |
+
"split_special_tokens": false,
|
51 |
"stride": 0,
|
52 |
"tokenizer_class": "LlamaTokenizer",
|
53 |
"truncation_side": "right",
|
54 |
"truncation_strategy": "longest_first",
|
55 |
"unk_token": "<unk>",
|
56 |
+
"use_default_system_prompt": true
|
57 |
}
|