Upload tokenizer
Browse files- added_tokens.json +29 -0
- special_tokens_map.json +42 -6
- tokenizer_config.json +1 -1
added_tokens.json
ADDED
|
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 250053,
|
| 3 |
+
"az_AZ": 250027,
|
| 4 |
+
"bn_IN": 250028,
|
| 5 |
+
"fa_IR": 250029,
|
| 6 |
+
"gl_ES": 250051,
|
| 7 |
+
"he_IL": 250030,
|
| 8 |
+
"hr_HR": 250031,
|
| 9 |
+
"id_ID": 250032,
|
| 10 |
+
"ka_GE": 250033,
|
| 11 |
+
"km_KH": 250034,
|
| 12 |
+
"mk_MK": 250035,
|
| 13 |
+
"ml_IN": 250036,
|
| 14 |
+
"mn_MN": 250037,
|
| 15 |
+
"mr_IN": 250038,
|
| 16 |
+
"pl_PL": 250039,
|
| 17 |
+
"ps_AF": 250040,
|
| 18 |
+
"pt_XX": 250041,
|
| 19 |
+
"sl_SI": 250052,
|
| 20 |
+
"sv_SE": 250042,
|
| 21 |
+
"sw_KE": 250043,
|
| 22 |
+
"ta_IN": 250044,
|
| 23 |
+
"te_IN": 250045,
|
| 24 |
+
"th_TH": 250046,
|
| 25 |
+
"tl_XX": 250047,
|
| 26 |
+
"uk_UA": 250048,
|
| 27 |
+
"ur_PK": 250049,
|
| 28 |
+
"xh_ZA": 250050
|
| 29 |
+
}
|
special_tokens_map.json
CHANGED
|
@@ -53,9 +53,27 @@
|
|
| 53 |
"gl_ES",
|
| 54 |
"sl_SI"
|
| 55 |
],
|
| 56 |
-
"bos_token":
|
| 57 |
-
|
| 58 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 59 |
"mask_token": {
|
| 60 |
"content": "<mask>",
|
| 61 |
"lstrip": true,
|
|
@@ -63,7 +81,25 @@
|
|
| 63 |
"rstrip": false,
|
| 64 |
"single_word": false
|
| 65 |
},
|
| 66 |
-
"pad_token":
|
| 67 |
-
|
| 68 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 69 |
}
|
|
|
|
| 53 |
"gl_ES",
|
| 54 |
"sl_SI"
|
| 55 |
],
|
| 56 |
+
"bos_token": {
|
| 57 |
+
"content": "<s>",
|
| 58 |
+
"lstrip": false,
|
| 59 |
+
"normalized": false,
|
| 60 |
+
"rstrip": false,
|
| 61 |
+
"single_word": false
|
| 62 |
+
},
|
| 63 |
+
"cls_token": {
|
| 64 |
+
"content": "<s>",
|
| 65 |
+
"lstrip": false,
|
| 66 |
+
"normalized": false,
|
| 67 |
+
"rstrip": false,
|
| 68 |
+
"single_word": false
|
| 69 |
+
},
|
| 70 |
+
"eos_token": {
|
| 71 |
+
"content": "</s>",
|
| 72 |
+
"lstrip": false,
|
| 73 |
+
"normalized": false,
|
| 74 |
+
"rstrip": false,
|
| 75 |
+
"single_word": false
|
| 76 |
+
},
|
| 77 |
"mask_token": {
|
| 78 |
"content": "<mask>",
|
| 79 |
"lstrip": true,
|
|
|
|
| 81 |
"rstrip": false,
|
| 82 |
"single_word": false
|
| 83 |
},
|
| 84 |
+
"pad_token": {
|
| 85 |
+
"content": "<pad>",
|
| 86 |
+
"lstrip": false,
|
| 87 |
+
"normalized": false,
|
| 88 |
+
"rstrip": false,
|
| 89 |
+
"single_word": false
|
| 90 |
+
},
|
| 91 |
+
"sep_token": {
|
| 92 |
+
"content": "</s>",
|
| 93 |
+
"lstrip": false,
|
| 94 |
+
"normalized": false,
|
| 95 |
+
"rstrip": false,
|
| 96 |
+
"single_word": false
|
| 97 |
+
},
|
| 98 |
+
"unk_token": {
|
| 99 |
+
"content": "<unk>",
|
| 100 |
+
"lstrip": false,
|
| 101 |
+
"normalized": false,
|
| 102 |
+
"rstrip": false,
|
| 103 |
+
"single_word": false
|
| 104 |
+
}
|
| 105 |
}
|
tokenizer_config.json
CHANGED
|
@@ -523,6 +523,6 @@
|
|
| 523 |
"sp_model_kwargs": {},
|
| 524 |
"src_lang": "en_XX",
|
| 525 |
"tgt_lang": "en_XX",
|
| 526 |
-
"tokenizer_class": "
|
| 527 |
"unk_token": "<unk>"
|
| 528 |
}
|
|
|
|
| 523 |
"sp_model_kwargs": {},
|
| 524 |
"src_lang": "en_XX",
|
| 525 |
"tgt_lang": "en_XX",
|
| 526 |
+
"tokenizer_class": "MBartTokenizer",
|
| 527 |
"unk_token": "<unk>"
|
| 528 |
}
|