Upload tokenizer
Browse files- added_tokens.json +319 -3
- special_tokens_map.json +51 -51
- tokenizer_config.json +0 -0
added_tokens.json
CHANGED
|
@@ -1,3 +1,319 @@
|
|
| 1 |
-
{
|
| 2 |
-
"
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"1.600_gia_đì": 64247,
|
| 3 |
+
"27-29": 64018,
|
| 4 |
+
"30/4/2025": 64169,
|
| 5 |
+
"<mask>": 64000,
|
| 6 |
+
"A_i": 64122,
|
| 7 |
+
"Akihi_o": 64048,
|
| 8 |
+
"Aoki_Kazuhiko": 64036,
|
| 9 |
+
"Ba_A": 64041,
|
| 10 |
+
"Bê_cạ": 64147,
|
| 11 |
+
"Bùi_Tha_h_Sơ": 64026,
|
| 12 |
+
"Bả": 64085,
|
| 13 |
+
"Bả_Ishiba": 64004,
|
| 14 |
+
"Chuyế": 64019,
|
| 15 |
+
"Chuyế_hăm": 64059,
|
| 16 |
+
"Chá_h": 64034,
|
| 17 |
+
"Chí_h": 64022,
|
| 18 |
+
"Chủ_ịch": 64218,
|
| 19 |
+
"Cố_vấ": 64037,
|
| 20 |
+
"Hai_bê": 64139,
|
| 21 |
+
"Hi_oo": 64051,
|
| 22 |
+
"Hiệ": 64179,
|
| 23 |
+
"Huy_h": 64285,
|
| 24 |
+
"Hòa": 64225,
|
| 25 |
+
"Hội_gười": 64304,
|
| 26 |
+
"I_o": 64097,
|
| 27 |
+
"Iijima_Isao": 64046,
|
| 28 |
+
"Ishiba": 64007,
|
| 29 |
+
"Jea_e": 64284,
|
| 30 |
+
"Kaiha_a": 64054,
|
| 31 |
+
"Ke": 64055,
|
| 32 |
+
"Koji": 64053,
|
| 33 |
+
"Lê_Thị_Bích_T": 64023,
|
| 34 |
+
"Lươ": 64151,
|
| 35 |
+
"Masa_aka": 64044,
|
| 36 |
+
"Mộ_cộ": 64279,
|
| 37 |
+
"Nagashima_Akihisa": 64039,
|
| 38 |
+
"Nakashima": 64050,
|
| 39 |
+
"Nam_Tow": 64264,
|
| 40 |
+
"Nam_chố": 64159,
|
| 41 |
+
"Nam_hực": 64128,
|
| 42 |
+
"Nam_vẫ": 64282,
|
| 43 |
+
"Nam_ă": 64311,
|
| 44 |
+
"Nam_ại": 64031,
|
| 45 |
+
"Nam_ỉ": 64243,
|
| 46 |
+
"Nam_ừ": 64016,
|
| 47 |
+
"Naoki": 64098,
|
| 48 |
+
"Nguyễ_Mi": 64029,
|
| 49 |
+
"Nhậ": 64003,
|
| 50 |
+
"Nhậ_Bả": 64011,
|
| 51 |
+
"Nhữ": 64229,
|
| 52 |
+
"Oka_o": 64043,
|
| 53 |
+
"Phu_hâ": 64006,
|
| 54 |
+
"Phạm_Mi_h": 64021,
|
| 55 |
+
"Phạm_Qua": 64032,
|
| 56 |
+
"Phầ_lớ": 64249,
|
| 57 |
+
"Phố_Việ": 64281,
|
| 58 |
+
"S_isuk": 64271,
|
| 59 |
+
"Shige_u": 64005,
|
| 60 |
+
"TPHCM.": 64183,
|
| 61 |
+
"TPHCM_phá": 64184,
|
| 62 |
+
"TPHCM_sá": 64160,
|
| 63 |
+
"T_o": 64195,
|
| 64 |
+
"T_u": 64216,
|
| 65 |
+
"Tha_i": 64221,
|
| 66 |
+
"Thá": 64272,
|
| 67 |
+
"Thái_La": 64153,
|
| 68 |
+
"Tháp_ù": 64033,
|
| 69 |
+
"Thườ": 64028,
|
| 70 |
+
"Thủ_ướ": 64002,
|
| 71 |
+
"Tsuchimichi": 64047,
|
| 72 |
+
"Tì": 64211,
|
| 73 |
+
"Tổ_g": 64040,
|
| 74 |
+
"Udo": 64220,
|
| 75 |
+
"Udo_Tha": 64244,
|
| 76 |
+
"Vie": 64265,
|
| 77 |
+
"Việ": 64015,
|
| 78 |
+
"Việ_Nam": 64057,
|
| 79 |
+
"Việ_kiều": 64260,
|
| 80 |
+
"Việ_ma": 64256,
|
| 81 |
+
"Việ_ại": 64219,
|
| 82 |
+
"Vũ_Thị_Huỳ_h": 64172,
|
| 83 |
+
"Xuâ": 64273,
|
| 84 |
+
"Xuâ_Hòa": 64152,
|
| 85 |
+
"Yoshi_o": 64052,
|
| 86 |
+
"Yoshiko": 64008,
|
| 87 |
+
"a_g": 64296,
|
| 88 |
+
"a_h": 64278,
|
| 89 |
+
"a_hà_h": 64309,
|
| 90 |
+
"a_o": 64056,
|
| 91 |
+
"aka_Mao": 64049,
|
| 92 |
+
"am_Tow": 64266,
|
| 93 |
+
"biế": 64099,
|
| 94 |
+
"biế_ơ": 64240,
|
| 95 |
+
"biể_hà": 64269,
|
| 96 |
+
"biểu_dươ": 64162,
|
| 97 |
+
"biệ": 64176,
|
| 98 |
+
"buô_bá": 64255,
|
| 99 |
+
"bê_hư": 64145,
|
| 100 |
+
"bì_h": 64071,
|
| 101 |
+
"bả_sắc": 64257,
|
| 102 |
+
"bắ": 64012,
|
| 103 |
+
"bắ_đầu": 64274,
|
| 104 |
+
"bằ": 64251,
|
| 105 |
+
"bổ_g": 64316,
|
| 106 |
+
"chiế": 64226,
|
| 107 |
+
"chiế_lược": 64067,
|
| 108 |
+
"chuyê_cơ": 64001,
|
| 109 |
+
"chuyê_gia": 64205,
|
| 110 |
+
"chuyế": 64062,
|
| 111 |
+
"chuyể": 64091,
|
| 112 |
+
"chuyể_đổi": 64089,
|
| 113 |
+
"chí_h": 64084,
|
| 114 |
+
"chấ": 64077,
|
| 115 |
+
"chức_ă": 64283,
|
| 116 |
+
"co_gười": 64254,
|
| 117 |
+
"co_phố": 64267,
|
| 118 |
+
"co_si": 64250,
|
| 119 |
+
"co_số": 64291,
|
| 120 |
+
"co_xa": 64155,
|
| 121 |
+
"cà_g": 64261,
|
| 122 |
+
"cá_hâ": 64117,
|
| 123 |
+
"cù_g": 64009,
|
| 124 |
+
"cơ_sở_hạ": 64131,
|
| 125 |
+
"cườ": 64102,
|
| 126 |
+
"cải_hiệ": 64313,
|
| 127 |
+
"cậy_chí": 64081,
|
| 128 |
+
"diễ_a": 64199,
|
| 129 |
+
"diệ": 64069,
|
| 130 |
+
"doa": 64207,
|
| 131 |
+
"doa_h": 64193,
|
| 132 |
+
"dâ": 64258,
|
| 133 |
+
"dự_g": 64164,
|
| 134 |
+
"dự_á": 64095,
|
| 135 |
+
"eu_o": 64315,
|
| 136 |
+
"ghiệp": 64134,
|
| 137 |
+
"ghiệp_hố": 64236,
|
| 138 |
+
"ghĩ": 64299,
|
| 139 |
+
"ghĩa": 64060,
|
| 140 |
+
"ghị": 64198,
|
| 141 |
+
"giai_đọa": 64196,
|
| 142 |
+
"goại": 64111,
|
| 143 |
+
"guyệ": 64306,
|
| 144 |
+
"guồ": 64132,
|
| 145 |
+
"guồ_lực": 64200,
|
| 146 |
+
"gày": 64017,
|
| 147 |
+
"gôi_hà": 64292,
|
| 148 |
+
"gười": 64154,
|
| 149 |
+
"gầ": 64201,
|
| 150 |
+
"gắ": 64268,
|
| 151 |
+
"h_Vũ": 64030,
|
| 152 |
+
"h_chí": 64270,
|
| 153 |
+
"h_cư": 64287,
|
| 154 |
+
"h_cảm": 64212,
|
| 155 |
+
"h_hổ": 64181,
|
| 156 |
+
"h_hức": 64014,
|
| 157 |
+
"h_leo": 64146,
|
| 158 |
+
"h_vực": 64080,
|
| 159 |
+
"h_đạo": 64115,
|
| 160 |
+
"hau": 64276,
|
| 161 |
+
"hiêu": 64301,
|
| 162 |
+
"hiế": 64227,
|
| 163 |
+
"hiế_hực": 64191,
|
| 164 |
+
"hiều": 64094,
|
| 165 |
+
"hiều_hội": 64197,
|
| 166 |
+
"hiều_ý": 64203,
|
| 167 |
+
"hiệ": 64129,
|
| 168 |
+
"hiệm": 64170,
|
| 169 |
+
"hà_cửa": 64294,
|
| 170 |
+
"hà_g": 64110,
|
| 171 |
+
"hà_h": 64189,
|
| 172 |
+
"hà_lã": 64118,
|
| 173 |
+
"hào_hơ": 64262,
|
| 174 |
+
"hào_ằ": 64231,
|
| 175 |
+
"hách_hức": 64124,
|
| 176 |
+
"hâ": 64166,
|
| 177 |
+
"hâ_lực": 64133,
|
| 178 |
+
"hì_h": 64150,
|
| 179 |
+
"hòa": 64070,
|
| 180 |
+
"hòa_òa": 64302,
|
| 181 |
+
"hóa": 64135,
|
| 182 |
+
"hăm_TPHCM": 64222,
|
| 183 |
+
"hăm_chí": 64013,
|
| 184 |
+
"hăm_diễ": 64020,
|
| 185 |
+
"hăm_hâ": 64192,
|
| 186 |
+
"hăm_lầ": 64113,
|
| 187 |
+
"hăm_đầu": 64063,
|
| 188 |
+
"hươ": 64137,
|
| 189 |
+
"hướ": 64210,
|
| 190 |
+
"hấ": 64138,
|
| 191 |
+
"hấy": 64290,
|
| 192 |
+
"hậ": 64202,
|
| 193 |
+
"hằm": 64101,
|
| 194 |
+
"hế": 64061,
|
| 195 |
+
"hế_giới": 64107,
|
| 196 |
+
"hể": 64096,
|
| 197 |
+
"hể_chế": 64130,
|
| 198 |
+
"hị": 64141,
|
| 199 |
+
"họa_độ": 64190,
|
| 200 |
+
"hỏ": 64235,
|
| 201 |
+
"hỗ_ợ": 64127,
|
| 202 |
+
"hợp_ác": 64082,
|
| 203 |
+
"hứ": 64297,
|
| 204 |
+
"hữ": 64108,
|
| 205 |
+
"hữu_ghị": 64259,
|
| 206 |
+
"i_Masafumi": 64045,
|
| 207 |
+
"i_cậy": 64116,
|
| 208 |
+
"i_h": 64042,
|
| 209 |
+
"i_ước": 64245,
|
| 210 |
+
"iê": 64064,
|
| 211 |
+
"iế": 64092,
|
| 212 |
+
"iếp_ục": 64086,
|
| 213 |
+
"iề": 64156,
|
| 214 |
+
"iể": 64093,
|
| 215 |
+
"iể_hà": 64165,
|
| 216 |
+
"iể_hực": 64076,
|
| 217 |
+
"kha_g": 64295,
|
| 218 |
+
"khô_g": 64289,
|
| 219 |
+
"khă": 64314,
|
| 220 |
+
"khả_ă": 64143,
|
| 221 |
+
"khẳ": 64119,
|
| 222 |
+
"khỏa": 64180,
|
| 223 |
+
"kiế": 64204,
|
| 224 |
+
"kiệ": 64248,
|
| 225 |
+
"kỷ_iệm": 64167,
|
| 226 |
+
"lam_hắ": 64280,
|
| 227 |
+
"liê_hệ": 64182,
|
| 228 |
+
"luô": 64238,
|
| 229 |
+
"luô_dà": 64174,
|
| 230 |
+
"lã": 64114,
|
| 231 |
+
"lĩ": 64079,
|
| 232 |
+
"lượ": 64232,
|
| 233 |
+
"lầ": 64100,
|
| 234 |
+
"lớ": 64233,
|
| 235 |
+
"miề": 64158,
|
| 236 |
+
"miề_Nam": 64230,
|
| 237 |
+
"muố": 64112,
|
| 238 |
+
"máu_hị": 64178,
|
| 239 |
+
"mặ": 64161,
|
| 240 |
+
"mặ_hà": 64144,
|
| 241 |
+
"mừ": 64157,
|
| 242 |
+
"o_Naoki": 64058,
|
| 243 |
+
"o_lớ": 64300,
|
| 244 |
+
"phươ": 64121,
|
| 245 |
+
"phấ_đấu": 64140,
|
| 246 |
+
"phầ": 64177,
|
| 247 |
+
"phồ": 64072,
|
| 248 |
+
"quyề": 64252,
|
| 249 |
+
"quốc_phò": 64123,
|
| 250 |
+
"quốc_ế": 64125,
|
| 251 |
+
"si_h": 64173,
|
| 252 |
+
"suấ": 64312,
|
| 253 |
+
"sâ": 64024,
|
| 254 |
+
"sả_xuấ": 64194,
|
| 255 |
+
"u_g": 64105,
|
| 256 |
+
"vi_h": 64073,
|
| 257 |
+
"viê": 64215,
|
| 258 |
+
"vă": 64035,
|
| 259 |
+
"vượ": 64185,
|
| 260 |
+
"vấ": 64208,
|
| 261 |
+
"vậ_lực": 64234,
|
| 262 |
+
"vữ": 64186,
|
| 263 |
+
"xa_h": 64090,
|
| 264 |
+
"xuố": 64288,
|
| 265 |
+
"xúc_độ": 64303,
|
| 266 |
+
"xươ": 64228,
|
| 267 |
+
"Ô_g": 64213,
|
| 268 |
+
"ác_o": 64088,
|
| 269 |
+
"âm_sự": 64286,
|
| 270 |
+
"âm_ă": 64106,
|
| 271 |
+
"âm_đặc": 64175,
|
| 272 |
+
"âm_ới": 64307,
|
| 273 |
+
"ê_hế": 64263,
|
| 274 |
+
"ê_hế_giới": 64074,
|
| 275 |
+
"ê_ấ": 64078,
|
| 276 |
+
"ì_h": 64149,
|
| 277 |
+
"í_h": 64246,
|
| 278 |
+
"í_hức": 64206,
|
| 279 |
+
"òa": 64068,
|
| 280 |
+
"ô_g": 64310,
|
| 281 |
+
"ă_mó": 64277,
|
| 282 |
+
"Đô_g": 64103,
|
| 283 |
+
"Đả_g": 64241,
|
| 284 |
+
"Đối_ác": 64066,
|
| 285 |
+
"đa_g": 64104,
|
| 286 |
+
"đà_phá": 64075,
|
| 287 |
+
"đòa": 64010,
|
| 288 |
+
"đú": 64223,
|
| 289 |
+
"đơ_sơ": 64293,
|
| 290 |
+
"đạ": 64188,
|
| 291 |
+
"đại_hóa": 64136,
|
| 292 |
+
"đấ_ước": 64168,
|
| 293 |
+
"đầu_àu": 64187,
|
| 294 |
+
"đặc_biệ": 64038,
|
| 295 |
+
"đề_liê": 64209,
|
| 296 |
+
"đị": 64120,
|
| 297 |
+
"đối_ác": 64087,
|
| 298 |
+
"đứ": 64308,
|
| 299 |
+
"ơ_Chí_h": 64242,
|
| 300 |
+
"ưu_iê": 64109,
|
| 301 |
+
"ươ": 64217,
|
| 302 |
+
"ước_goài": 64163,
|
| 303 |
+
"ước_hà": 64237,
|
| 304 |
+
"ước_hằm": 64148,
|
| 305 |
+
"ước_â": 64065,
|
| 306 |
+
"ước_đa": 64224,
|
| 307 |
+
"ườ": 64142,
|
| 308 |
+
"ưở": 64027,
|
| 309 |
+
"ưở_g": 64025,
|
| 310 |
+
"ấ_hiều": 64298,
|
| 311 |
+
"ấ_đô": 64253,
|
| 312 |
+
"ẩm_hực": 64275,
|
| 313 |
+
"ập_hợp": 64305,
|
| 314 |
+
"ập_u": 64126,
|
| 315 |
+
"ụ_cộ": 64083,
|
| 316 |
+
"Ủy": 64171,
|
| 317 |
+
"ủy": 64214,
|
| 318 |
+
"ự_hào": 64239
|
| 319 |
+
}
|
special_tokens_map.json
CHANGED
|
@@ -1,51 +1,51 @@
|
|
| 1 |
-
{
|
| 2 |
-
"bos_token": {
|
| 3 |
-
"content": "<s>",
|
| 4 |
-
"lstrip": false,
|
| 5 |
-
"normalized": false,
|
| 6 |
-
"rstrip": false,
|
| 7 |
-
"single_word": false
|
| 8 |
-
},
|
| 9 |
-
"cls_token": {
|
| 10 |
-
"content": "<s>",
|
| 11 |
-
"lstrip": false,
|
| 12 |
-
"normalized": false,
|
| 13 |
-
"rstrip": false,
|
| 14 |
-
"single_word": false
|
| 15 |
-
},
|
| 16 |
-
"eos_token": {
|
| 17 |
-
"content": "</s>",
|
| 18 |
-
"lstrip": false,
|
| 19 |
-
"normalized": false,
|
| 20 |
-
"rstrip": false,
|
| 21 |
-
"single_word": false
|
| 22 |
-
},
|
| 23 |
-
"mask_token": {
|
| 24 |
-
"content": "<mask>",
|
| 25 |
-
"lstrip": false,
|
| 26 |
-
"normalized": false,
|
| 27 |
-
"rstrip": false,
|
| 28 |
-
"single_word": false
|
| 29 |
-
},
|
| 30 |
-
"pad_token": {
|
| 31 |
-
"content": "<pad>",
|
| 32 |
-
"lstrip": false,
|
| 33 |
-
"normalized": false,
|
| 34 |
-
"rstrip": false,
|
| 35 |
-
"single_word": false
|
| 36 |
-
},
|
| 37 |
-
"sep_token": {
|
| 38 |
-
"content": "</s>",
|
| 39 |
-
"lstrip": false,
|
| 40 |
-
"normalized": false,
|
| 41 |
-
"rstrip": false,
|
| 42 |
-
"single_word": false
|
| 43 |
-
},
|
| 44 |
-
"unk_token": {
|
| 45 |
-
"content": "<unk>",
|
| 46 |
-
"lstrip": false,
|
| 47 |
-
"normalized": false,
|
| 48 |
-
"rstrip": false,
|
| 49 |
-
"single_word": false
|
| 50 |
-
}
|
| 51 |
-
}
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
tokenizer_config.json
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|