Upload tokenizer
Browse files- special_tokens_map.json +1 -7
- tokenizer.json +2 -2
- tokenizer_config.json +3 -3
special_tokens_map.json
CHANGED
|
@@ -13,13 +13,7 @@
|
|
| 13 |
"rstrip": false,
|
| 14 |
"single_word": false
|
| 15 |
},
|
| 16 |
-
"pad_token":
|
| 17 |
-
"content": "[control_14]",
|
| 18 |
-
"lstrip": false,
|
| 19 |
-
"normalized": false,
|
| 20 |
-
"rstrip": false,
|
| 21 |
-
"single_word": false
|
| 22 |
-
},
|
| 23 |
"unk_token": {
|
| 24 |
"content": "<unk>",
|
| 25 |
"lstrip": false,
|
|
|
|
| 13 |
"rstrip": false,
|
| 14 |
"single_word": false
|
| 15 |
},
|
| 16 |
+
"pad_token": "[control_748]",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 17 |
"unk_token": {
|
| 18 |
"content": "<unk>",
|
| 19 |
"lstrip": false,
|
tokenizer.json
CHANGED
|
@@ -154,7 +154,7 @@
|
|
| 154 |
"lstrip": false,
|
| 155 |
"rstrip": false,
|
| 156 |
"normalized": false,
|
| 157 |
-
"special":
|
| 158 |
},
|
| 159 |
{
|
| 160 |
"id": 17,
|
|
@@ -6760,7 +6760,7 @@
|
|
| 6760 |
"lstrip": false,
|
| 6761 |
"rstrip": false,
|
| 6762 |
"normalized": false,
|
| 6763 |
-
"special":
|
| 6764 |
},
|
| 6765 |
{
|
| 6766 |
"id": 751,
|
|
|
|
| 154 |
"lstrip": false,
|
| 155 |
"rstrip": false,
|
| 156 |
"normalized": false,
|
| 157 |
+
"special": false
|
| 158 |
},
|
| 159 |
{
|
| 160 |
"id": 17,
|
|
|
|
| 6760 |
"lstrip": false,
|
| 6761 |
"rstrip": false,
|
| 6762 |
"normalized": false,
|
| 6763 |
+
"special": true
|
| 6764 |
},
|
| 6765 |
{
|
| 6766 |
"id": 751,
|
tokenizer_config.json
CHANGED
|
@@ -137,7 +137,7 @@
|
|
| 137 |
"normalized": false,
|
| 138 |
"rstrip": false,
|
| 139 |
"single_word": false,
|
| 140 |
-
"special":
|
| 141 |
},
|
| 142 |
"17": {
|
| 143 |
"content": "[control_15]",
|
|
@@ -6009,7 +6009,7 @@
|
|
| 6009 |
"normalized": false,
|
| 6010 |
"rstrip": false,
|
| 6011 |
"single_word": false,
|
| 6012 |
-
"special":
|
| 6013 |
},
|
| 6014 |
"751": {
|
| 6015 |
"content": "[REFERENCE_DOC_19]",
|
|
@@ -6178,7 +6178,7 @@
|
|
| 6178 |
"eos_token": "</s>",
|
| 6179 |
"legacy": true,
|
| 6180 |
"model_max_length": 1000000000000000019884624838656,
|
| 6181 |
-
"pad_token": "[
|
| 6182 |
"padding_side": "left",
|
| 6183 |
"sp_model_kwargs": {},
|
| 6184 |
"spaces_between_special_tokens": false,
|
|
|
|
| 137 |
"normalized": false,
|
| 138 |
"rstrip": false,
|
| 139 |
"single_word": false,
|
| 140 |
+
"special": false
|
| 141 |
},
|
| 142 |
"17": {
|
| 143 |
"content": "[control_15]",
|
|
|
|
| 6009 |
"normalized": false,
|
| 6010 |
"rstrip": false,
|
| 6011 |
"single_word": false,
|
| 6012 |
+
"special": true
|
| 6013 |
},
|
| 6014 |
"751": {
|
| 6015 |
"content": "[REFERENCE_DOC_19]",
|
|
|
|
| 6178 |
"eos_token": "</s>",
|
| 6179 |
"legacy": true,
|
| 6180 |
"model_max_length": 1000000000000000019884624838656,
|
| 6181 |
+
"pad_token": "[control_748]",
|
| 6182 |
"padding_side": "left",
|
| 6183 |
"sp_model_kwargs": {},
|
| 6184 |
"spaces_between_special_tokens": false,
|