Update tokenizer_config.json
#2
by
saattrupdan
- opened
- tokenizer_config.json +12 -5
tokenizer_config.json
CHANGED
@@ -27,16 +27,23 @@
|
|
27 |
"special": true
|
28 |
}
|
29 |
},
|
30 |
-
"additional_special_tokens": [
|
|
|
|
|
|
|
|
|
31 |
"bos_token": "<s>",
|
32 |
"clean_up_tokenization_spaces": false,
|
33 |
"eos_token": "</s>",
|
34 |
"legacy": true,
|
35 |
-
"model_max_length":
|
36 |
-
"pad_token":
|
|
|
37 |
"sp_model_kwargs": {},
|
38 |
"spaces_between_special_tokens": false,
|
|
|
39 |
"tokenizer_class": "LlamaTokenizer",
|
40 |
"unk_token": "<unk>",
|
41 |
-
"
|
42 |
-
|
|
|
|
27 |
"special": true
|
28 |
}
|
29 |
},
|
30 |
+
"additional_special_tokens": [
|
31 |
+
"<unk>",
|
32 |
+
"<s>",
|
33 |
+
"</s>"
|
34 |
+
],
|
35 |
"bos_token": "<s>",
|
36 |
"clean_up_tokenization_spaces": false,
|
37 |
"eos_token": "</s>",
|
38 |
"legacy": true,
|
39 |
+
"model_max_length": 8192,
|
40 |
+
"pad_token": "</s>",
|
41 |
+
"padding_side": "left",
|
42 |
"sp_model_kwargs": {},
|
43 |
"spaces_between_special_tokens": false,
|
44 |
+
"split_special_tokens": false,
|
45 |
"tokenizer_class": "LlamaTokenizer",
|
46 |
"unk_token": "<unk>",
|
47 |
+
"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 %}",
|
48 |
+
"use_default_system_prompt": true
|
49 |
+
}
|