Faaz commited on
Commit ·
02eef51
1
Parent(s): 24b5fb1
Fix extra_special_tokens: list to dict for transformers 4.55
Browse files
data/tokenizer/mindi_tokenizer/tokenizer_config.json
CHANGED
|
@@ -5,34 +5,34 @@
|
|
| 5 |
"clean_up_tokenization_spaces": false,
|
| 6 |
"eos_token": "<|im_end|>",
|
| 7 |
"errors": "replace",
|
| 8 |
-
"extra_special_tokens":
|
| 9 |
-
"<|mindi_start|>",
|
| 10 |
-
"<|mindi_end|>",
|
| 11 |
-
"<|code_start|>",
|
| 12 |
-
"<|code_end|>",
|
| 13 |
-
"<|vision_start|>",
|
| 14 |
-
"<|vision_end|>",
|
| 15 |
-
"<|critique_start|>",
|
| 16 |
-
"<|critique_end|>",
|
| 17 |
-
"<|suggest_start|>",
|
| 18 |
-
"<|suggest_end|>",
|
| 19 |
-
"<|think_start|>",
|
| 20 |
-
"<|think_end|>",
|
| 21 |
-
"<|file_start|>",
|
| 22 |
-
"<|file_end|>",
|
| 23 |
-
"<|search_start|>",
|
| 24 |
-
"<|search_end|>",
|
| 25 |
-
"<|sandbox_start|>",
|
| 26 |
-
"<|sandbox_end|>",
|
| 27 |
-
"<|error_start|>",
|
| 28 |
-
"<|error_end|>",
|
| 29 |
-
"<|fix_start|>",
|
| 30 |
-
"<|fix_end|>"
|
| 31 |
-
|
| 32 |
"is_local": true,
|
| 33 |
"model_max_length": 32768,
|
| 34 |
"pad_token": "<|endoftext|>",
|
| 35 |
"split_special_tokens": false,
|
| 36 |
"tokenizer_class": "Qwen2Tokenizer",
|
| 37 |
"unk_token": null
|
| 38 |
-
}
|
|
|
|
| 5 |
"clean_up_tokenization_spaces": false,
|
| 6 |
"eos_token": "<|im_end|>",
|
| 7 |
"errors": "replace",
|
| 8 |
+
"extra_special_tokens": {
|
| 9 |
+
"<|mindi_start|>": "<|mindi_start|>",
|
| 10 |
+
"<|mindi_end|>": "<|mindi_end|>",
|
| 11 |
+
"<|code_start|>": "<|code_start|>",
|
| 12 |
+
"<|code_end|>": "<|code_end|>",
|
| 13 |
+
"<|vision_start|>": "<|vision_start|>",
|
| 14 |
+
"<|vision_end|>": "<|vision_end|>",
|
| 15 |
+
"<|critique_start|>": "<|critique_start|>",
|
| 16 |
+
"<|critique_end|>": "<|critique_end|>",
|
| 17 |
+
"<|suggest_start|>": "<|suggest_start|>",
|
| 18 |
+
"<|suggest_end|>": "<|suggest_end|>",
|
| 19 |
+
"<|think_start|>": "<|think_start|>",
|
| 20 |
+
"<|think_end|>": "<|think_end|>",
|
| 21 |
+
"<|file_start|>": "<|file_start|>",
|
| 22 |
+
"<|file_end|>": "<|file_end|>",
|
| 23 |
+
"<|search_start|>": "<|search_start|>",
|
| 24 |
+
"<|search_end|>": "<|search_end|>",
|
| 25 |
+
"<|sandbox_start|>": "<|sandbox_start|>",
|
| 26 |
+
"<|sandbox_end|>": "<|sandbox_end|>",
|
| 27 |
+
"<|error_start|>": "<|error_start|>",
|
| 28 |
+
"<|error_end|>": "<|error_end|>",
|
| 29 |
+
"<|fix_start|>": "<|fix_start|>",
|
| 30 |
+
"<|fix_end|>": "<|fix_end|>"
|
| 31 |
+
},
|
| 32 |
"is_local": true,
|
| 33 |
"model_max_length": 32768,
|
| 34 |
"pad_token": "<|endoftext|>",
|
| 35 |
"split_special_tokens": false,
|
| 36 |
"tokenizer_class": "Qwen2Tokenizer",
|
| 37 |
"unk_token": null
|
| 38 |
+
}
|