Update Tokenizer Chat Template (#1)
Browse files- Update Tokenizer Chat Template (fe1c29cada0b72760adb64f24851a0c360833adb)
Co-authored-by: Equinox Psychosis <[email protected]>
- tokenizer_config.json +1 -1
tokenizer_config.json
CHANGED
@@ -2050,7 +2050,7 @@
|
|
2050 |
}
|
2051 |
},
|
2052 |
"bos_token": "<|begin_of_text|>",
|
2053 |
-
"chat_template": "{
|
2054 |
"clean_up_tokenization_spaces": true,
|
2055 |
"eos_token": "<|eot_id|>",
|
2056 |
"extra_special_tokens": {},
|
|
|
2050 |
}
|
2051 |
},
|
2052 |
"bos_token": "<|begin_of_text|>",
|
2053 |
+
"chat_template": "{%- for message in messages %}\n {%- if loop.index0 == 0 %}\n {{- bos_token }}\n {%- endif %}\n {%- if message['role'] == 'system' %}\n {{- '<|start_header_id|>Metadata<|end_header_id|>\n\n' + message['content'] + eos_token }}\n {%- endif %}\n {%- if message['role'] == 'user' %}\n {{- '<|start_header_id|>Japanese<|end_header_id|>\n\n' + message['content'] + eos_token }}\n {%- endif %}\n {%- if message['role'] == 'assistant' %}\n {{- '<|start_header_id|>English<|end_header_id|>\n\n' + message['content'] + eos_token }}\n {%- endif %}\n{%- endfor %}\n{%- if add_generation_prompt %}\n {{- '<|start_header_id|>English<|end_header_id|>\n\n' }}\n{%- endif %}",
|
2054 |
"clean_up_tokenization_spaces": true,
|
2055 |
"eos_token": "<|eot_id|>",
|
2056 |
"extra_special_tokens": {},
|