Upload tokenizer
Browse files- tokenizer_config.json +30 -1
tokenizer_config.json
CHANGED
|
@@ -243,10 +243,39 @@
|
|
| 243 |
},
|
| 244 |
"thinking": {
|
| 245 |
"type": "string"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 246 |
}
|
| 247 |
},
|
| 248 |
"type": "object",
|
| 249 |
-
"x-regex": "^(?:<think>)?\\s*(?P<thinking>.+?)\\s*</think>\\s*(?P<content>.+?)\\s*$"
|
| 250 |
},
|
| 251 |
"split_special_tokens": false,
|
| 252 |
"tokenizer_class": "Qwen2Tokenizer",
|
|
|
|
| 243 |
},
|
| 244 |
"thinking": {
|
| 245 |
"type": "string"
|
| 246 |
+
},
|
| 247 |
+
"tool_calls": {
|
| 248 |
+
"items": {
|
| 249 |
+
"properties": {
|
| 250 |
+
"function": {
|
| 251 |
+
"properties": {
|
| 252 |
+
"arguments": {
|
| 253 |
+
"additionalProperties": {
|
| 254 |
+
"type": "any"
|
| 255 |
+
},
|
| 256 |
+
"type": "object"
|
| 257 |
+
},
|
| 258 |
+
"name": {
|
| 259 |
+
"type": "string"
|
| 260 |
+
}
|
| 261 |
+
},
|
| 262 |
+
"type": "object"
|
| 263 |
+
},
|
| 264 |
+
"type": {
|
| 265 |
+
"const": "function"
|
| 266 |
+
}
|
| 267 |
+
},
|
| 268 |
+
"type": "object"
|
| 269 |
+
},
|
| 270 |
+
"type": "array",
|
| 271 |
+
"x-parser": "json",
|
| 272 |
+
"x-parser-args": {
|
| 273 |
+
"transform": "[{type: 'function', function: @}]"
|
| 274 |
+
}
|
| 275 |
}
|
| 276 |
},
|
| 277 |
"type": "object",
|
| 278 |
+
"x-regex": "^(?:<think>)?\\s*(?P<thinking>.+?)\\s*</think>\\s*(?:<tool_call>(?P<tool_calls>.+?)</tool_call>|(?P<content>.+?))\\s*$"
|
| 279 |
},
|
| 280 |
"split_special_tokens": false,
|
| 281 |
"tokenizer_class": "Qwen2Tokenizer",
|