Upload tokenizer
Browse files- special_tokens_map.json +1 -0
- tokenizer.json +11 -1
- tokenizer_config.json +9 -0
special_tokens_map.json
CHANGED
|
@@ -2,6 +2,7 @@
|
|
| 2 |
"bos_token": "[bos]",
|
| 3 |
"cls_token": "[cls]",
|
| 4 |
"eos_token": "[eos]",
|
|
|
|
| 5 |
"pad_token": "[pad]",
|
| 6 |
"sep_token": "[sep]",
|
| 7 |
"unk_token": "[unk]"
|
|
|
|
| 2 |
"bos_token": "[bos]",
|
| 3 |
"cls_token": "[cls]",
|
| 4 |
"eos_token": "[eos]",
|
| 5 |
+
"mask_token": "[mask]",
|
| 6 |
"pad_token": "[pad]",
|
| 7 |
"sep_token": "[sep]",
|
| 8 |
"unk_token": "[unk]"
|
tokenizer.json
CHANGED
|
@@ -50,6 +50,15 @@
|
|
| 50 |
},
|
| 51 |
{
|
| 52 |
"id": 258,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 53 |
"content": "[unk]",
|
| 54 |
"single_word": false,
|
| 55 |
"lstrip": false,
|
|
@@ -338,7 +347,8 @@
|
|
| 338 |
"[pad]": 255,
|
| 339 |
"[eos]": 256,
|
| 340 |
"[sep]": 257,
|
| 341 |
-
"[
|
|
|
|
| 342 |
},
|
| 343 |
"unk_token": "[unk]"
|
| 344 |
}
|
|
|
|
| 50 |
},
|
| 51 |
{
|
| 52 |
"id": 258,
|
| 53 |
+
"content": "[mask]",
|
| 54 |
+
"single_word": false,
|
| 55 |
+
"lstrip": false,
|
| 56 |
+
"rstrip": false,
|
| 57 |
+
"normalized": false,
|
| 58 |
+
"special": true
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"id": 259,
|
| 62 |
"content": "[unk]",
|
| 63 |
"single_word": false,
|
| 64 |
"lstrip": false,
|
|
|
|
| 347 |
"[pad]": 255,
|
| 348 |
"[eos]": 256,
|
| 349 |
"[sep]": 257,
|
| 350 |
+
"[mask]": 258,
|
| 351 |
+
"[unk]": 259
|
| 352 |
},
|
| 353 |
"unk_token": "[unk]"
|
| 354 |
}
|
tokenizer_config.json
CHANGED
|
@@ -41,6 +41,14 @@
|
|
| 41 |
"special": true
|
| 42 |
},
|
| 43 |
"258": {
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 44 |
"content": "[unk]",
|
| 45 |
"lstrip": false,
|
| 46 |
"normalized": false,
|
|
@@ -54,6 +62,7 @@
|
|
| 54 |
"cls_token": "[cls]",
|
| 55 |
"eos_token": "[eos]",
|
| 56 |
"extra_special_tokens": {},
|
|
|
|
| 57 |
"model_max_length": 1000000000000000019884624838656,
|
| 58 |
"pad_token": "[pad]",
|
| 59 |
"sep_token": "[sep]",
|
|
|
|
| 41 |
"special": true
|
| 42 |
},
|
| 43 |
"258": {
|
| 44 |
+
"content": "[mask]",
|
| 45 |
+
"lstrip": false,
|
| 46 |
+
"normalized": false,
|
| 47 |
+
"rstrip": false,
|
| 48 |
+
"single_word": false,
|
| 49 |
+
"special": true
|
| 50 |
+
},
|
| 51 |
+
"259": {
|
| 52 |
"content": "[unk]",
|
| 53 |
"lstrip": false,
|
| 54 |
"normalized": false,
|
|
|
|
| 62 |
"cls_token": "[cls]",
|
| 63 |
"eos_token": "[eos]",
|
| 64 |
"extra_special_tokens": {},
|
| 65 |
+
"mask_token": "[mask]",
|
| 66 |
"model_max_length": 1000000000000000019884624838656,
|
| 67 |
"pad_token": "[pad]",
|
| 68 |
"sep_token": "[sep]",
|