Commit
•
9a5d98e
1
Parent(s):
e6178fe
End of training
Browse files- config.json +1 -1
- generation_config.json +6 -0
- pytorch_model.bin +1 -1
- runs/Sep02_01-12-45_frankenstein/1693617247.5637603/events.out.tfevents.1693617247.frankenstein.406473.1 +3 -0
- runs/Sep02_01-12-45_frankenstein/events.out.tfevents.1693617247.frankenstein.406473.0 +3 -0
- runs/Sep02_01-15-36_frankenstein/1693617337.6766403/events.out.tfevents.1693617337.frankenstein.406770.1 +3 -0
- runs/Sep02_01-15-36_frankenstein/events.out.tfevents.1693617337.frankenstein.406770.0 +3 -0
- runs/Sep02_01-28-40_frankenstein/1693618121.2594054/events.out.tfevents.1693618121.frankenstein.407062.1 +3 -0
- runs/Sep02_01-28-40_frankenstein/events.out.tfevents.1693618121.frankenstein.407062.0 +3 -0
- runs/Sep02_07-17-35_frankenstein/1693639056.7861805/events.out.tfevents.1693639056.frankenstein.409682.1 +3 -0
- runs/Sep02_07-17-35_frankenstein/events.out.tfevents.1693639056.frankenstein.409682.0 +3 -0
- runs/Sep02_08-19-09_frankenstein/1693642777.1662269/events.out.tfevents.1693642777.frankenstein.410808.1 +3 -0
- runs/Sep02_08-19-09_frankenstein/events.out.tfevents.1693642777.frankenstein.410808.0 +3 -0
- runs/Sep02_08-33-50_frankenstein/1693643631.6678832/events.out.tfevents.1693643631.frankenstein.410808.3 +3 -0
- runs/Sep02_08-33-50_frankenstein/events.out.tfevents.1693643631.frankenstein.410808.2 +3 -0
- tokenizer.json +206 -206
- tokenizer_config.json +1 -0
- training_args.bin +2 -2
config.json
CHANGED
|
@@ -54,7 +54,7 @@
|
|
| 54 |
}
|
| 55 |
},
|
| 56 |
"torch_dtype": "float32",
|
| 57 |
-
"transformers_version": "4.
|
| 58 |
"use_cache": true,
|
| 59 |
"vocab_size": 32128
|
| 60 |
}
|
|
|
|
| 54 |
}
|
| 55 |
},
|
| 56 |
"torch_dtype": "float32",
|
| 57 |
+
"transformers_version": "4.27.0.dev0",
|
| 58 |
"use_cache": true,
|
| 59 |
"vocab_size": 32128
|
| 60 |
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"decoder_start_token_id": 0,
|
| 3 |
+
"eos_token_id": 1,
|
| 4 |
+
"pad_token_id": 0,
|
| 5 |
+
"transformers_version": "4.27.0.dev0"
|
| 6 |
+
}
|
pytorch_model.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 242071641
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:42d400b010060b236b7aeda8923ca4b2de1e7da64eacd0cf919e38b2a6a135eb
|
| 3 |
size 242071641
|
runs/Sep02_01-12-45_frankenstein/1693617247.5637603/events.out.tfevents.1693617247.frankenstein.406473.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ad8a788512a38c276c22faf5e917d7c636ffbce7ac3bf24bfdbb2573c429972b
|
| 3 |
+
size 6083
|
runs/Sep02_01-12-45_frankenstein/events.out.tfevents.1693617247.frankenstein.406473.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4a8e26fc2e9108ed19794fb67d8bea6d82bbd35f53d03ee7c37120296646ec35
|
| 3 |
+
size 4928
|
runs/Sep02_01-15-36_frankenstein/1693617337.6766403/events.out.tfevents.1693617337.frankenstein.406770.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2345894e5faff00f19498a3a67ad01573b9d992165e19fc8d215092384cd7acb
|
| 3 |
+
size 6083
|
runs/Sep02_01-15-36_frankenstein/events.out.tfevents.1693617337.frankenstein.406770.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8b2cffa35ae5ccfaa1fe43de0a95a73220be62639fcda3b20ccf9f7a8f045f26
|
| 3 |
+
size 4929
|
runs/Sep02_01-28-40_frankenstein/1693618121.2594054/events.out.tfevents.1693618121.frankenstein.407062.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c98ca9918084fd3c067299a6101e34b041d99944ac987a9024fc3193de2da3fe
|
| 3 |
+
size 6083
|
runs/Sep02_01-28-40_frankenstein/events.out.tfevents.1693618121.frankenstein.407062.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2c41897ea422ba0682e3278d0d4c6518360b377798efefda20f753a7f5e35146
|
| 3 |
+
size 11168
|
runs/Sep02_07-17-35_frankenstein/1693639056.7861805/events.out.tfevents.1693639056.frankenstein.409682.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:976dcf2fac9056af287b1e63f8b63b8bbeabf095ab4df0e0f46372c787468a9f
|
| 3 |
+
size 6083
|
runs/Sep02_07-17-35_frankenstein/events.out.tfevents.1693639056.frankenstein.409682.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f05c77eae028dfe6e5a04c34d509733a103e8cba6186fd986fe71acbf6ae69ae
|
| 3 |
+
size 4184
|
runs/Sep02_08-19-09_frankenstein/1693642777.1662269/events.out.tfevents.1693642777.frankenstein.410808.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f61ecdbfffebb6b6d5c28c09240c51c802ecc61f3cf5f855e520faba0ad170cc
|
| 3 |
+
size 6083
|
runs/Sep02_08-19-09_frankenstein/events.out.tfevents.1693642777.frankenstein.410808.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:826ee6dcaa490fa59af4941da92806a625aee02ed41ba0b1a51f70d710fc9efe
|
| 3 |
+
size 5792
|
runs/Sep02_08-33-50_frankenstein/1693643631.6678832/events.out.tfevents.1693643631.frankenstein.410808.3
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a3328efc1813b1b72451cd80fc8d4acfd8ddf5f7d11816eb635d803750587291
|
| 3 |
+
size 6083
|
runs/Sep02_08-33-50_frankenstein/events.out.tfevents.1693643631.frankenstein.410808.2
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:40437befe32b3220e4d040d742809bc5a18ecaf486445b2c4082ed2f9eb7158a
|
| 3 |
+
size 10838
|
tokenizer.json
CHANGED
|
@@ -10,930 +10,930 @@
|
|
| 10 |
"added_tokens": [
|
| 11 |
{
|
| 12 |
"id": 0,
|
| 13 |
-
"special": true,
|
| 14 |
"content": "<pad>",
|
| 15 |
"single_word": false,
|
| 16 |
"lstrip": false,
|
| 17 |
"rstrip": false,
|
| 18 |
-
"normalized": false
|
|
|
|
| 19 |
},
|
| 20 |
{
|
| 21 |
"id": 1,
|
| 22 |
-
"special": true,
|
| 23 |
"content": "</s>",
|
| 24 |
"single_word": false,
|
| 25 |
"lstrip": false,
|
| 26 |
"rstrip": false,
|
| 27 |
-
"normalized": false
|
|
|
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"id": 2,
|
| 31 |
-
"special": true,
|
| 32 |
"content": "<unk>",
|
| 33 |
"single_word": false,
|
| 34 |
"lstrip": false,
|
| 35 |
"rstrip": false,
|
| 36 |
-
"normalized": false
|
|
|
|
| 37 |
},
|
| 38 |
{
|
| 39 |
"id": 32000,
|
| 40 |
-
"special": true,
|
| 41 |
"content": "<extra_id_99>",
|
| 42 |
"single_word": false,
|
| 43 |
"lstrip": false,
|
| 44 |
"rstrip": false,
|
| 45 |
-
"normalized": false
|
|
|
|
| 46 |
},
|
| 47 |
{
|
| 48 |
"id": 32001,
|
| 49 |
-
"special": true,
|
| 50 |
"content": "<extra_id_98>",
|
| 51 |
"single_word": false,
|
| 52 |
"lstrip": false,
|
| 53 |
"rstrip": false,
|
| 54 |
-
"normalized": false
|
|
|
|
| 55 |
},
|
| 56 |
{
|
| 57 |
"id": 32002,
|
| 58 |
-
"special": true,
|
| 59 |
"content": "<extra_id_97>",
|
| 60 |
"single_word": false,
|
| 61 |
"lstrip": false,
|
| 62 |
"rstrip": false,
|
| 63 |
-
"normalized": false
|
|
|
|
| 64 |
},
|
| 65 |
{
|
| 66 |
"id": 32003,
|
| 67 |
-
"special": true,
|
| 68 |
"content": "<extra_id_96>",
|
| 69 |
"single_word": false,
|
| 70 |
"lstrip": false,
|
| 71 |
"rstrip": false,
|
| 72 |
-
"normalized": false
|
|
|
|
| 73 |
},
|
| 74 |
{
|
| 75 |
"id": 32004,
|
| 76 |
-
"special": true,
|
| 77 |
"content": "<extra_id_95>",
|
| 78 |
"single_word": false,
|
| 79 |
"lstrip": false,
|
| 80 |
"rstrip": false,
|
| 81 |
-
"normalized": false
|
|
|
|
| 82 |
},
|
| 83 |
{
|
| 84 |
"id": 32005,
|
| 85 |
-
"special": true,
|
| 86 |
"content": "<extra_id_94>",
|
| 87 |
"single_word": false,
|
| 88 |
"lstrip": false,
|
| 89 |
"rstrip": false,
|
| 90 |
-
"normalized": false
|
|
|
|
| 91 |
},
|
| 92 |
{
|
| 93 |
"id": 32006,
|
| 94 |
-
"special": true,
|
| 95 |
"content": "<extra_id_93>",
|
| 96 |
"single_word": false,
|
| 97 |
"lstrip": false,
|
| 98 |
"rstrip": false,
|
| 99 |
-
"normalized": false
|
|
|
|
| 100 |
},
|
| 101 |
{
|
| 102 |
"id": 32007,
|
| 103 |
-
"special": true,
|
| 104 |
"content": "<extra_id_92>",
|
| 105 |
"single_word": false,
|
| 106 |
"lstrip": false,
|
| 107 |
"rstrip": false,
|
| 108 |
-
"normalized": false
|
|
|
|
| 109 |
},
|
| 110 |
{
|
| 111 |
"id": 32008,
|
| 112 |
-
"special": true,
|
| 113 |
"content": "<extra_id_91>",
|
| 114 |
"single_word": false,
|
| 115 |
"lstrip": false,
|
| 116 |
"rstrip": false,
|
| 117 |
-
"normalized": false
|
|
|
|
| 118 |
},
|
| 119 |
{
|
| 120 |
"id": 32009,
|
| 121 |
-
"special": true,
|
| 122 |
"content": "<extra_id_90>",
|
| 123 |
"single_word": false,
|
| 124 |
"lstrip": false,
|
| 125 |
"rstrip": false,
|
| 126 |
-
"normalized": false
|
|
|
|
| 127 |
},
|
| 128 |
{
|
| 129 |
"id": 32010,
|
| 130 |
-
"special": true,
|
| 131 |
"content": "<extra_id_89>",
|
| 132 |
"single_word": false,
|
| 133 |
"lstrip": false,
|
| 134 |
"rstrip": false,
|
| 135 |
-
"normalized": false
|
|
|
|
| 136 |
},
|
| 137 |
{
|
| 138 |
"id": 32011,
|
| 139 |
-
"special": true,
|
| 140 |
"content": "<extra_id_88>",
|
| 141 |
"single_word": false,
|
| 142 |
"lstrip": false,
|
| 143 |
"rstrip": false,
|
| 144 |
-
"normalized": false
|
|
|
|
| 145 |
},
|
| 146 |
{
|
| 147 |
"id": 32012,
|
| 148 |
-
"special": true,
|
| 149 |
"content": "<extra_id_87>",
|
| 150 |
"single_word": false,
|
| 151 |
"lstrip": false,
|
| 152 |
"rstrip": false,
|
| 153 |
-
"normalized": false
|
|
|
|
| 154 |
},
|
| 155 |
{
|
| 156 |
"id": 32013,
|
| 157 |
-
"special": true,
|
| 158 |
"content": "<extra_id_86>",
|
| 159 |
"single_word": false,
|
| 160 |
"lstrip": false,
|
| 161 |
"rstrip": false,
|
| 162 |
-
"normalized": false
|
|
|
|
| 163 |
},
|
| 164 |
{
|
| 165 |
"id": 32014,
|
| 166 |
-
"special": true,
|
| 167 |
"content": "<extra_id_85>",
|
| 168 |
"single_word": false,
|
| 169 |
"lstrip": false,
|
| 170 |
"rstrip": false,
|
| 171 |
-
"normalized": false
|
|
|
|
| 172 |
},
|
| 173 |
{
|
| 174 |
"id": 32015,
|
| 175 |
-
"special": true,
|
| 176 |
"content": "<extra_id_84>",
|
| 177 |
"single_word": false,
|
| 178 |
"lstrip": false,
|
| 179 |
"rstrip": false,
|
| 180 |
-
"normalized": false
|
|
|
|
| 181 |
},
|
| 182 |
{
|
| 183 |
"id": 32016,
|
| 184 |
-
"special": true,
|
| 185 |
"content": "<extra_id_83>",
|
| 186 |
"single_word": false,
|
| 187 |
"lstrip": false,
|
| 188 |
"rstrip": false,
|
| 189 |
-
"normalized": false
|
|
|
|
| 190 |
},
|
| 191 |
{
|
| 192 |
"id": 32017,
|
| 193 |
-
"special": true,
|
| 194 |
"content": "<extra_id_82>",
|
| 195 |
"single_word": false,
|
| 196 |
"lstrip": false,
|
| 197 |
"rstrip": false,
|
| 198 |
-
"normalized": false
|
|
|
|
| 199 |
},
|
| 200 |
{
|
| 201 |
"id": 32018,
|
| 202 |
-
"special": true,
|
| 203 |
"content": "<extra_id_81>",
|
| 204 |
"single_word": false,
|
| 205 |
"lstrip": false,
|
| 206 |
"rstrip": false,
|
| 207 |
-
"normalized": false
|
|
|
|
| 208 |
},
|
| 209 |
{
|
| 210 |
"id": 32019,
|
| 211 |
-
"special": true,
|
| 212 |
"content": "<extra_id_80>",
|
| 213 |
"single_word": false,
|
| 214 |
"lstrip": false,
|
| 215 |
"rstrip": false,
|
| 216 |
-
"normalized": false
|
|
|
|
| 217 |
},
|
| 218 |
{
|
| 219 |
"id": 32020,
|
| 220 |
-
"special": true,
|
| 221 |
"content": "<extra_id_79>",
|
| 222 |
"single_word": false,
|
| 223 |
"lstrip": false,
|
| 224 |
"rstrip": false,
|
| 225 |
-
"normalized": false
|
|
|
|
| 226 |
},
|
| 227 |
{
|
| 228 |
"id": 32021,
|
| 229 |
-
"special": true,
|
| 230 |
"content": "<extra_id_78>",
|
| 231 |
"single_word": false,
|
| 232 |
"lstrip": false,
|
| 233 |
"rstrip": false,
|
| 234 |
-
"normalized": false
|
|
|
|
| 235 |
},
|
| 236 |
{
|
| 237 |
"id": 32022,
|
| 238 |
-
"special": true,
|
| 239 |
"content": "<extra_id_77>",
|
| 240 |
"single_word": false,
|
| 241 |
"lstrip": false,
|
| 242 |
"rstrip": false,
|
| 243 |
-
"normalized": false
|
|
|
|
| 244 |
},
|
| 245 |
{
|
| 246 |
"id": 32023,
|
| 247 |
-
"special": true,
|
| 248 |
"content": "<extra_id_76>",
|
| 249 |
"single_word": false,
|
| 250 |
"lstrip": false,
|
| 251 |
"rstrip": false,
|
| 252 |
-
"normalized": false
|
|
|
|
| 253 |
},
|
| 254 |
{
|
| 255 |
"id": 32024,
|
| 256 |
-
"special": true,
|
| 257 |
"content": "<extra_id_75>",
|
| 258 |
"single_word": false,
|
| 259 |
"lstrip": false,
|
| 260 |
"rstrip": false,
|
| 261 |
-
"normalized": false
|
|
|
|
| 262 |
},
|
| 263 |
{
|
| 264 |
"id": 32025,
|
| 265 |
-
"special": true,
|
| 266 |
"content": "<extra_id_74>",
|
| 267 |
"single_word": false,
|
| 268 |
"lstrip": false,
|
| 269 |
"rstrip": false,
|
| 270 |
-
"normalized": false
|
|
|
|
| 271 |
},
|
| 272 |
{
|
| 273 |
"id": 32026,
|
| 274 |
-
"special": true,
|
| 275 |
"content": "<extra_id_73>",
|
| 276 |
"single_word": false,
|
| 277 |
"lstrip": false,
|
| 278 |
"rstrip": false,
|
| 279 |
-
"normalized": false
|
|
|
|
| 280 |
},
|
| 281 |
{
|
| 282 |
"id": 32027,
|
| 283 |
-
"special": true,
|
| 284 |
"content": "<extra_id_72>",
|
| 285 |
"single_word": false,
|
| 286 |
"lstrip": false,
|
| 287 |
"rstrip": false,
|
| 288 |
-
"normalized": false
|
|
|
|
| 289 |
},
|
| 290 |
{
|
| 291 |
"id": 32028,
|
| 292 |
-
"special": true,
|
| 293 |
"content": "<extra_id_71>",
|
| 294 |
"single_word": false,
|
| 295 |
"lstrip": false,
|
| 296 |
"rstrip": false,
|
| 297 |
-
"normalized": false
|
|
|
|
| 298 |
},
|
| 299 |
{
|
| 300 |
"id": 32029,
|
| 301 |
-
"special": true,
|
| 302 |
"content": "<extra_id_70>",
|
| 303 |
"single_word": false,
|
| 304 |
"lstrip": false,
|
| 305 |
"rstrip": false,
|
| 306 |
-
"normalized": false
|
|
|
|
| 307 |
},
|
| 308 |
{
|
| 309 |
"id": 32030,
|
| 310 |
-
"special": true,
|
| 311 |
"content": "<extra_id_69>",
|
| 312 |
"single_word": false,
|
| 313 |
"lstrip": false,
|
| 314 |
"rstrip": false,
|
| 315 |
-
"normalized": false
|
|
|
|
| 316 |
},
|
| 317 |
{
|
| 318 |
"id": 32031,
|
| 319 |
-
"special": true,
|
| 320 |
"content": "<extra_id_68>",
|
| 321 |
"single_word": false,
|
| 322 |
"lstrip": false,
|
| 323 |
"rstrip": false,
|
| 324 |
-
"normalized": false
|
|
|
|
| 325 |
},
|
| 326 |
{
|
| 327 |
"id": 32032,
|
| 328 |
-
"special": true,
|
| 329 |
"content": "<extra_id_67>",
|
| 330 |
"single_word": false,
|
| 331 |
"lstrip": false,
|
| 332 |
"rstrip": false,
|
| 333 |
-
"normalized": false
|
|
|
|
| 334 |
},
|
| 335 |
{
|
| 336 |
"id": 32033,
|
| 337 |
-
"special": true,
|
| 338 |
"content": "<extra_id_66>",
|
| 339 |
"single_word": false,
|
| 340 |
"lstrip": false,
|
| 341 |
"rstrip": false,
|
| 342 |
-
"normalized": false
|
|
|
|
| 343 |
},
|
| 344 |
{
|
| 345 |
"id": 32034,
|
| 346 |
-
"special": true,
|
| 347 |
"content": "<extra_id_65>",
|
| 348 |
"single_word": false,
|
| 349 |
"lstrip": false,
|
| 350 |
"rstrip": false,
|
| 351 |
-
"normalized": false
|
|
|
|
| 352 |
},
|
| 353 |
{
|
| 354 |
"id": 32035,
|
| 355 |
-
"special": true,
|
| 356 |
"content": "<extra_id_64>",
|
| 357 |
"single_word": false,
|
| 358 |
"lstrip": false,
|
| 359 |
"rstrip": false,
|
| 360 |
-
"normalized": false
|
|
|
|
| 361 |
},
|
| 362 |
{
|
| 363 |
"id": 32036,
|
| 364 |
-
"special": true,
|
| 365 |
"content": "<extra_id_63>",
|
| 366 |
"single_word": false,
|
| 367 |
"lstrip": false,
|
| 368 |
"rstrip": false,
|
| 369 |
-
"normalized": false
|
|
|
|
| 370 |
},
|
| 371 |
{
|
| 372 |
"id": 32037,
|
| 373 |
-
"special": true,
|
| 374 |
"content": "<extra_id_62>",
|
| 375 |
"single_word": false,
|
| 376 |
"lstrip": false,
|
| 377 |
"rstrip": false,
|
| 378 |
-
"normalized": false
|
|
|
|
| 379 |
},
|
| 380 |
{
|
| 381 |
"id": 32038,
|
| 382 |
-
"special": true,
|
| 383 |
"content": "<extra_id_61>",
|
| 384 |
"single_word": false,
|
| 385 |
"lstrip": false,
|
| 386 |
"rstrip": false,
|
| 387 |
-
"normalized": false
|
|
|
|
| 388 |
},
|
| 389 |
{
|
| 390 |
"id": 32039,
|
| 391 |
-
"special": true,
|
| 392 |
"content": "<extra_id_60>",
|
| 393 |
"single_word": false,
|
| 394 |
"lstrip": false,
|
| 395 |
"rstrip": false,
|
| 396 |
-
"normalized": false
|
|
|
|
| 397 |
},
|
| 398 |
{
|
| 399 |
"id": 32040,
|
| 400 |
-
"special": true,
|
| 401 |
"content": "<extra_id_59>",
|
| 402 |
"single_word": false,
|
| 403 |
"lstrip": false,
|
| 404 |
"rstrip": false,
|
| 405 |
-
"normalized": false
|
|
|
|
| 406 |
},
|
| 407 |
{
|
| 408 |
"id": 32041,
|
| 409 |
-
"special": true,
|
| 410 |
"content": "<extra_id_58>",
|
| 411 |
"single_word": false,
|
| 412 |
"lstrip": false,
|
| 413 |
"rstrip": false,
|
| 414 |
-
"normalized": false
|
|
|
|
| 415 |
},
|
| 416 |
{
|
| 417 |
"id": 32042,
|
| 418 |
-
"special": true,
|
| 419 |
"content": "<extra_id_57>",
|
| 420 |
"single_word": false,
|
| 421 |
"lstrip": false,
|
| 422 |
"rstrip": false,
|
| 423 |
-
"normalized": false
|
|
|
|
| 424 |
},
|
| 425 |
{
|
| 426 |
"id": 32043,
|
| 427 |
-
"special": true,
|
| 428 |
"content": "<extra_id_56>",
|
| 429 |
"single_word": false,
|
| 430 |
"lstrip": false,
|
| 431 |
"rstrip": false,
|
| 432 |
-
"normalized": false
|
|
|
|
| 433 |
},
|
| 434 |
{
|
| 435 |
"id": 32044,
|
| 436 |
-
"special": true,
|
| 437 |
"content": "<extra_id_55>",
|
| 438 |
"single_word": false,
|
| 439 |
"lstrip": false,
|
| 440 |
"rstrip": false,
|
| 441 |
-
"normalized": false
|
|
|
|
| 442 |
},
|
| 443 |
{
|
| 444 |
"id": 32045,
|
| 445 |
-
"special": true,
|
| 446 |
"content": "<extra_id_54>",
|
| 447 |
"single_word": false,
|
| 448 |
"lstrip": false,
|
| 449 |
"rstrip": false,
|
| 450 |
-
"normalized": false
|
|
|
|
| 451 |
},
|
| 452 |
{
|
| 453 |
"id": 32046,
|
| 454 |
-
"special": true,
|
| 455 |
"content": "<extra_id_53>",
|
| 456 |
"single_word": false,
|
| 457 |
"lstrip": false,
|
| 458 |
"rstrip": false,
|
| 459 |
-
"normalized": false
|
|
|
|
| 460 |
},
|
| 461 |
{
|
| 462 |
"id": 32047,
|
| 463 |
-
"special": true,
|
| 464 |
"content": "<extra_id_52>",
|
| 465 |
"single_word": false,
|
| 466 |
"lstrip": false,
|
| 467 |
"rstrip": false,
|
| 468 |
-
"normalized": false
|
|
|
|
| 469 |
},
|
| 470 |
{
|
| 471 |
"id": 32048,
|
| 472 |
-
"special": true,
|
| 473 |
"content": "<extra_id_51>",
|
| 474 |
"single_word": false,
|
| 475 |
"lstrip": false,
|
| 476 |
"rstrip": false,
|
| 477 |
-
"normalized": false
|
|
|
|
| 478 |
},
|
| 479 |
{
|
| 480 |
"id": 32049,
|
| 481 |
-
"special": true,
|
| 482 |
"content": "<extra_id_50>",
|
| 483 |
"single_word": false,
|
| 484 |
"lstrip": false,
|
| 485 |
"rstrip": false,
|
| 486 |
-
"normalized": false
|
|
|
|
| 487 |
},
|
| 488 |
{
|
| 489 |
"id": 32050,
|
| 490 |
-
"special": true,
|
| 491 |
"content": "<extra_id_49>",
|
| 492 |
"single_word": false,
|
| 493 |
"lstrip": false,
|
| 494 |
"rstrip": false,
|
| 495 |
-
"normalized": false
|
|
|
|
| 496 |
},
|
| 497 |
{
|
| 498 |
"id": 32051,
|
| 499 |
-
"special": true,
|
| 500 |
"content": "<extra_id_48>",
|
| 501 |
"single_word": false,
|
| 502 |
"lstrip": false,
|
| 503 |
"rstrip": false,
|
| 504 |
-
"normalized": false
|
|
|
|
| 505 |
},
|
| 506 |
{
|
| 507 |
"id": 32052,
|
| 508 |
-
"special": true,
|
| 509 |
"content": "<extra_id_47>",
|
| 510 |
"single_word": false,
|
| 511 |
"lstrip": false,
|
| 512 |
"rstrip": false,
|
| 513 |
-
"normalized": false
|
|
|
|
| 514 |
},
|
| 515 |
{
|
| 516 |
"id": 32053,
|
| 517 |
-
"special": true,
|
| 518 |
"content": "<extra_id_46>",
|
| 519 |
"single_word": false,
|
| 520 |
"lstrip": false,
|
| 521 |
"rstrip": false,
|
| 522 |
-
"normalized": false
|
|
|
|
| 523 |
},
|
| 524 |
{
|
| 525 |
"id": 32054,
|
| 526 |
-
"special": true,
|
| 527 |
"content": "<extra_id_45>",
|
| 528 |
"single_word": false,
|
| 529 |
"lstrip": false,
|
| 530 |
"rstrip": false,
|
| 531 |
-
"normalized": false
|
|
|
|
| 532 |
},
|
| 533 |
{
|
| 534 |
"id": 32055,
|
| 535 |
-
"special": true,
|
| 536 |
"content": "<extra_id_44>",
|
| 537 |
"single_word": false,
|
| 538 |
"lstrip": false,
|
| 539 |
"rstrip": false,
|
| 540 |
-
"normalized": false
|
|
|
|
| 541 |
},
|
| 542 |
{
|
| 543 |
"id": 32056,
|
| 544 |
-
"special": true,
|
| 545 |
"content": "<extra_id_43>",
|
| 546 |
"single_word": false,
|
| 547 |
"lstrip": false,
|
| 548 |
"rstrip": false,
|
| 549 |
-
"normalized": false
|
|
|
|
| 550 |
},
|
| 551 |
{
|
| 552 |
"id": 32057,
|
| 553 |
-
"special": true,
|
| 554 |
"content": "<extra_id_42>",
|
| 555 |
"single_word": false,
|
| 556 |
"lstrip": false,
|
| 557 |
"rstrip": false,
|
| 558 |
-
"normalized": false
|
|
|
|
| 559 |
},
|
| 560 |
{
|
| 561 |
"id": 32058,
|
| 562 |
-
"special": true,
|
| 563 |
"content": "<extra_id_41>",
|
| 564 |
"single_word": false,
|
| 565 |
"lstrip": false,
|
| 566 |
"rstrip": false,
|
| 567 |
-
"normalized": false
|
|
|
|
| 568 |
},
|
| 569 |
{
|
| 570 |
"id": 32059,
|
| 571 |
-
"special": true,
|
| 572 |
"content": "<extra_id_40>",
|
| 573 |
"single_word": false,
|
| 574 |
"lstrip": false,
|
| 575 |
"rstrip": false,
|
| 576 |
-
"normalized": false
|
|
|
|
| 577 |
},
|
| 578 |
{
|
| 579 |
"id": 32060,
|
| 580 |
-
"special": true,
|
| 581 |
"content": "<extra_id_39>",
|
| 582 |
"single_word": false,
|
| 583 |
"lstrip": false,
|
| 584 |
"rstrip": false,
|
| 585 |
-
"normalized": false
|
|
|
|
| 586 |
},
|
| 587 |
{
|
| 588 |
"id": 32061,
|
| 589 |
-
"special": true,
|
| 590 |
"content": "<extra_id_38>",
|
| 591 |
"single_word": false,
|
| 592 |
"lstrip": false,
|
| 593 |
"rstrip": false,
|
| 594 |
-
"normalized": false
|
|
|
|
| 595 |
},
|
| 596 |
{
|
| 597 |
"id": 32062,
|
| 598 |
-
"special": true,
|
| 599 |
"content": "<extra_id_37>",
|
| 600 |
"single_word": false,
|
| 601 |
"lstrip": false,
|
| 602 |
"rstrip": false,
|
| 603 |
-
"normalized": false
|
|
|
|
| 604 |
},
|
| 605 |
{
|
| 606 |
"id": 32063,
|
| 607 |
-
"special": true,
|
| 608 |
"content": "<extra_id_36>",
|
| 609 |
"single_word": false,
|
| 610 |
"lstrip": false,
|
| 611 |
"rstrip": false,
|
| 612 |
-
"normalized": false
|
|
|
|
| 613 |
},
|
| 614 |
{
|
| 615 |
"id": 32064,
|
| 616 |
-
"special": true,
|
| 617 |
"content": "<extra_id_35>",
|
| 618 |
"single_word": false,
|
| 619 |
"lstrip": false,
|
| 620 |
"rstrip": false,
|
| 621 |
-
"normalized": false
|
|
|
|
| 622 |
},
|
| 623 |
{
|
| 624 |
"id": 32065,
|
| 625 |
-
"special": true,
|
| 626 |
"content": "<extra_id_34>",
|
| 627 |
"single_word": false,
|
| 628 |
"lstrip": false,
|
| 629 |
"rstrip": false,
|
| 630 |
-
"normalized": false
|
|
|
|
| 631 |
},
|
| 632 |
{
|
| 633 |
"id": 32066,
|
| 634 |
-
"special": true,
|
| 635 |
"content": "<extra_id_33>",
|
| 636 |
"single_word": false,
|
| 637 |
"lstrip": false,
|
| 638 |
"rstrip": false,
|
| 639 |
-
"normalized": false
|
|
|
|
| 640 |
},
|
| 641 |
{
|
| 642 |
"id": 32067,
|
| 643 |
-
"special": true,
|
| 644 |
"content": "<extra_id_32>",
|
| 645 |
"single_word": false,
|
| 646 |
"lstrip": false,
|
| 647 |
"rstrip": false,
|
| 648 |
-
"normalized": false
|
|
|
|
| 649 |
},
|
| 650 |
{
|
| 651 |
"id": 32068,
|
| 652 |
-
"special": true,
|
| 653 |
"content": "<extra_id_31>",
|
| 654 |
"single_word": false,
|
| 655 |
"lstrip": false,
|
| 656 |
"rstrip": false,
|
| 657 |
-
"normalized": false
|
|
|
|
| 658 |
},
|
| 659 |
{
|
| 660 |
"id": 32069,
|
| 661 |
-
"special": true,
|
| 662 |
"content": "<extra_id_30>",
|
| 663 |
"single_word": false,
|
| 664 |
"lstrip": false,
|
| 665 |
"rstrip": false,
|
| 666 |
-
"normalized": false
|
|
|
|
| 667 |
},
|
| 668 |
{
|
| 669 |
"id": 32070,
|
| 670 |
-
"special": true,
|
| 671 |
"content": "<extra_id_29>",
|
| 672 |
"single_word": false,
|
| 673 |
"lstrip": false,
|
| 674 |
"rstrip": false,
|
| 675 |
-
"normalized": false
|
|
|
|
| 676 |
},
|
| 677 |
{
|
| 678 |
"id": 32071,
|
| 679 |
-
"special": true,
|
| 680 |
"content": "<extra_id_28>",
|
| 681 |
"single_word": false,
|
| 682 |
"lstrip": false,
|
| 683 |
"rstrip": false,
|
| 684 |
-
"normalized": false
|
|
|
|
| 685 |
},
|
| 686 |
{
|
| 687 |
"id": 32072,
|
| 688 |
-
"special": true,
|
| 689 |
"content": "<extra_id_27>",
|
| 690 |
"single_word": false,
|
| 691 |
"lstrip": false,
|
| 692 |
"rstrip": false,
|
| 693 |
-
"normalized": false
|
|
|
|
| 694 |
},
|
| 695 |
{
|
| 696 |
"id": 32073,
|
| 697 |
-
"special": true,
|
| 698 |
"content": "<extra_id_26>",
|
| 699 |
"single_word": false,
|
| 700 |
"lstrip": false,
|
| 701 |
"rstrip": false,
|
| 702 |
-
"normalized": false
|
|
|
|
| 703 |
},
|
| 704 |
{
|
| 705 |
"id": 32074,
|
| 706 |
-
"special": true,
|
| 707 |
"content": "<extra_id_25>",
|
| 708 |
"single_word": false,
|
| 709 |
"lstrip": false,
|
| 710 |
"rstrip": false,
|
| 711 |
-
"normalized": false
|
|
|
|
| 712 |
},
|
| 713 |
{
|
| 714 |
"id": 32075,
|
| 715 |
-
"special": true,
|
| 716 |
"content": "<extra_id_24>",
|
| 717 |
"single_word": false,
|
| 718 |
"lstrip": false,
|
| 719 |
"rstrip": false,
|
| 720 |
-
"normalized": false
|
|
|
|
| 721 |
},
|
| 722 |
{
|
| 723 |
"id": 32076,
|
| 724 |
-
"special": true,
|
| 725 |
"content": "<extra_id_23>",
|
| 726 |
"single_word": false,
|
| 727 |
"lstrip": false,
|
| 728 |
"rstrip": false,
|
| 729 |
-
"normalized": false
|
|
|
|
| 730 |
},
|
| 731 |
{
|
| 732 |
"id": 32077,
|
| 733 |
-
"special": true,
|
| 734 |
"content": "<extra_id_22>",
|
| 735 |
"single_word": false,
|
| 736 |
"lstrip": false,
|
| 737 |
"rstrip": false,
|
| 738 |
-
"normalized": false
|
|
|
|
| 739 |
},
|
| 740 |
{
|
| 741 |
"id": 32078,
|
| 742 |
-
"special": true,
|
| 743 |
"content": "<extra_id_21>",
|
| 744 |
"single_word": false,
|
| 745 |
"lstrip": false,
|
| 746 |
"rstrip": false,
|
| 747 |
-
"normalized": false
|
|
|
|
| 748 |
},
|
| 749 |
{
|
| 750 |
"id": 32079,
|
| 751 |
-
"special": true,
|
| 752 |
"content": "<extra_id_20>",
|
| 753 |
"single_word": false,
|
| 754 |
"lstrip": false,
|
| 755 |
"rstrip": false,
|
| 756 |
-
"normalized": false
|
|
|
|
| 757 |
},
|
| 758 |
{
|
| 759 |
"id": 32080,
|
| 760 |
-
"special": true,
|
| 761 |
"content": "<extra_id_19>",
|
| 762 |
"single_word": false,
|
| 763 |
"lstrip": false,
|
| 764 |
"rstrip": false,
|
| 765 |
-
"normalized": false
|
|
|
|
| 766 |
},
|
| 767 |
{
|
| 768 |
"id": 32081,
|
| 769 |
-
"special": true,
|
| 770 |
"content": "<extra_id_18>",
|
| 771 |
"single_word": false,
|
| 772 |
"lstrip": false,
|
| 773 |
"rstrip": false,
|
| 774 |
-
"normalized": false
|
|
|
|
| 775 |
},
|
| 776 |
{
|
| 777 |
"id": 32082,
|
| 778 |
-
"special": true,
|
| 779 |
"content": "<extra_id_17>",
|
| 780 |
"single_word": false,
|
| 781 |
"lstrip": false,
|
| 782 |
"rstrip": false,
|
| 783 |
-
"normalized": false
|
|
|
|
| 784 |
},
|
| 785 |
{
|
| 786 |
"id": 32083,
|
| 787 |
-
"special": true,
|
| 788 |
"content": "<extra_id_16>",
|
| 789 |
"single_word": false,
|
| 790 |
"lstrip": false,
|
| 791 |
"rstrip": false,
|
| 792 |
-
"normalized": false
|
|
|
|
| 793 |
},
|
| 794 |
{
|
| 795 |
"id": 32084,
|
| 796 |
-
"special": true,
|
| 797 |
"content": "<extra_id_15>",
|
| 798 |
"single_word": false,
|
| 799 |
"lstrip": false,
|
| 800 |
"rstrip": false,
|
| 801 |
-
"normalized": false
|
|
|
|
| 802 |
},
|
| 803 |
{
|
| 804 |
"id": 32085,
|
| 805 |
-
"special": true,
|
| 806 |
"content": "<extra_id_14>",
|
| 807 |
"single_word": false,
|
| 808 |
"lstrip": false,
|
| 809 |
"rstrip": false,
|
| 810 |
-
"normalized": false
|
|
|
|
| 811 |
},
|
| 812 |
{
|
| 813 |
"id": 32086,
|
| 814 |
-
"special": true,
|
| 815 |
"content": "<extra_id_13>",
|
| 816 |
"single_word": false,
|
| 817 |
"lstrip": false,
|
| 818 |
"rstrip": false,
|
| 819 |
-
"normalized": false
|
|
|
|
| 820 |
},
|
| 821 |
{
|
| 822 |
"id": 32087,
|
| 823 |
-
"special": true,
|
| 824 |
"content": "<extra_id_12>",
|
| 825 |
"single_word": false,
|
| 826 |
"lstrip": false,
|
| 827 |
"rstrip": false,
|
| 828 |
-
"normalized": false
|
|
|
|
| 829 |
},
|
| 830 |
{
|
| 831 |
"id": 32088,
|
| 832 |
-
"special": true,
|
| 833 |
"content": "<extra_id_11>",
|
| 834 |
"single_word": false,
|
| 835 |
"lstrip": false,
|
| 836 |
"rstrip": false,
|
| 837 |
-
"normalized": false
|
|
|
|
| 838 |
},
|
| 839 |
{
|
| 840 |
"id": 32089,
|
| 841 |
-
"special": true,
|
| 842 |
"content": "<extra_id_10>",
|
| 843 |
"single_word": false,
|
| 844 |
"lstrip": false,
|
| 845 |
"rstrip": false,
|
| 846 |
-
"normalized": false
|
|
|
|
| 847 |
},
|
| 848 |
{
|
| 849 |
"id": 32090,
|
| 850 |
-
"special": true,
|
| 851 |
"content": "<extra_id_9>",
|
| 852 |
"single_word": false,
|
| 853 |
"lstrip": false,
|
| 854 |
"rstrip": false,
|
| 855 |
-
"normalized": false
|
|
|
|
| 856 |
},
|
| 857 |
{
|
| 858 |
"id": 32091,
|
| 859 |
-
"special": true,
|
| 860 |
"content": "<extra_id_8>",
|
| 861 |
"single_word": false,
|
| 862 |
"lstrip": false,
|
| 863 |
"rstrip": false,
|
| 864 |
-
"normalized": false
|
|
|
|
| 865 |
},
|
| 866 |
{
|
| 867 |
"id": 32092,
|
| 868 |
-
"special": true,
|
| 869 |
"content": "<extra_id_7>",
|
| 870 |
"single_word": false,
|
| 871 |
"lstrip": false,
|
| 872 |
"rstrip": false,
|
| 873 |
-
"normalized": false
|
|
|
|
| 874 |
},
|
| 875 |
{
|
| 876 |
"id": 32093,
|
| 877 |
-
"special": true,
|
| 878 |
"content": "<extra_id_6>",
|
| 879 |
"single_word": false,
|
| 880 |
"lstrip": false,
|
| 881 |
"rstrip": false,
|
| 882 |
-
"normalized": false
|
|
|
|
| 883 |
},
|
| 884 |
{
|
| 885 |
"id": 32094,
|
| 886 |
-
"special": true,
|
| 887 |
"content": "<extra_id_5>",
|
| 888 |
"single_word": false,
|
| 889 |
"lstrip": false,
|
| 890 |
"rstrip": false,
|
| 891 |
-
"normalized": false
|
|
|
|
| 892 |
},
|
| 893 |
{
|
| 894 |
"id": 32095,
|
| 895 |
-
"special": true,
|
| 896 |
"content": "<extra_id_4>",
|
| 897 |
"single_word": false,
|
| 898 |
"lstrip": false,
|
| 899 |
"rstrip": false,
|
| 900 |
-
"normalized": false
|
|
|
|
| 901 |
},
|
| 902 |
{
|
| 903 |
"id": 32096,
|
| 904 |
-
"special": true,
|
| 905 |
"content": "<extra_id_3>",
|
| 906 |
"single_word": false,
|
| 907 |
"lstrip": false,
|
| 908 |
"rstrip": false,
|
| 909 |
-
"normalized": false
|
|
|
|
| 910 |
},
|
| 911 |
{
|
| 912 |
"id": 32097,
|
| 913 |
-
"special": true,
|
| 914 |
"content": "<extra_id_2>",
|
| 915 |
"single_word": false,
|
| 916 |
"lstrip": false,
|
| 917 |
"rstrip": false,
|
| 918 |
-
"normalized": false
|
|
|
|
| 919 |
},
|
| 920 |
{
|
| 921 |
"id": 32098,
|
| 922 |
-
"special": true,
|
| 923 |
"content": "<extra_id_1>",
|
| 924 |
"single_word": false,
|
| 925 |
"lstrip": false,
|
| 926 |
"rstrip": false,
|
| 927 |
-
"normalized": false
|
|
|
|
| 928 |
},
|
| 929 |
{
|
| 930 |
"id": 32099,
|
| 931 |
-
"special": true,
|
| 932 |
"content": "<extra_id_0>",
|
| 933 |
"single_word": false,
|
| 934 |
"lstrip": false,
|
| 935 |
"rstrip": false,
|
| 936 |
-
"normalized": false
|
|
|
|
| 937 |
}
|
| 938 |
],
|
| 939 |
"normalizer": {
|
|
|
|
| 10 |
"added_tokens": [
|
| 11 |
{
|
| 12 |
"id": 0,
|
|
|
|
| 13 |
"content": "<pad>",
|
| 14 |
"single_word": false,
|
| 15 |
"lstrip": false,
|
| 16 |
"rstrip": false,
|
| 17 |
+
"normalized": false,
|
| 18 |
+
"special": true
|
| 19 |
},
|
| 20 |
{
|
| 21 |
"id": 1,
|
|
|
|
| 22 |
"content": "</s>",
|
| 23 |
"single_word": false,
|
| 24 |
"lstrip": false,
|
| 25 |
"rstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"special": true
|
| 28 |
},
|
| 29 |
{
|
| 30 |
"id": 2,
|
|
|
|
| 31 |
"content": "<unk>",
|
| 32 |
"single_word": false,
|
| 33 |
"lstrip": false,
|
| 34 |
"rstrip": false,
|
| 35 |
+
"normalized": false,
|
| 36 |
+
"special": true
|
| 37 |
},
|
| 38 |
{
|
| 39 |
"id": 32000,
|
|
|
|
| 40 |
"content": "<extra_id_99>",
|
| 41 |
"single_word": false,
|
| 42 |
"lstrip": false,
|
| 43 |
"rstrip": false,
|
| 44 |
+
"normalized": false,
|
| 45 |
+
"special": true
|
| 46 |
},
|
| 47 |
{
|
| 48 |
"id": 32001,
|
|
|
|
| 49 |
"content": "<extra_id_98>",
|
| 50 |
"single_word": false,
|
| 51 |
"lstrip": false,
|
| 52 |
"rstrip": false,
|
| 53 |
+
"normalized": false,
|
| 54 |
+
"special": true
|
| 55 |
},
|
| 56 |
{
|
| 57 |
"id": 32002,
|
|
|
|
| 58 |
"content": "<extra_id_97>",
|
| 59 |
"single_word": false,
|
| 60 |
"lstrip": false,
|
| 61 |
"rstrip": false,
|
| 62 |
+
"normalized": false,
|
| 63 |
+
"special": true
|
| 64 |
},
|
| 65 |
{
|
| 66 |
"id": 32003,
|
|
|
|
| 67 |
"content": "<extra_id_96>",
|
| 68 |
"single_word": false,
|
| 69 |
"lstrip": false,
|
| 70 |
"rstrip": false,
|
| 71 |
+
"normalized": false,
|
| 72 |
+
"special": true
|
| 73 |
},
|
| 74 |
{
|
| 75 |
"id": 32004,
|
|
|
|
| 76 |
"content": "<extra_id_95>",
|
| 77 |
"single_word": false,
|
| 78 |
"lstrip": false,
|
| 79 |
"rstrip": false,
|
| 80 |
+
"normalized": false,
|
| 81 |
+
"special": true
|
| 82 |
},
|
| 83 |
{
|
| 84 |
"id": 32005,
|
|
|
|
| 85 |
"content": "<extra_id_94>",
|
| 86 |
"single_word": false,
|
| 87 |
"lstrip": false,
|
| 88 |
"rstrip": false,
|
| 89 |
+
"normalized": false,
|
| 90 |
+
"special": true
|
| 91 |
},
|
| 92 |
{
|
| 93 |
"id": 32006,
|
|
|
|
| 94 |
"content": "<extra_id_93>",
|
| 95 |
"single_word": false,
|
| 96 |
"lstrip": false,
|
| 97 |
"rstrip": false,
|
| 98 |
+
"normalized": false,
|
| 99 |
+
"special": true
|
| 100 |
},
|
| 101 |
{
|
| 102 |
"id": 32007,
|
|
|
|
| 103 |
"content": "<extra_id_92>",
|
| 104 |
"single_word": false,
|
| 105 |
"lstrip": false,
|
| 106 |
"rstrip": false,
|
| 107 |
+
"normalized": false,
|
| 108 |
+
"special": true
|
| 109 |
},
|
| 110 |
{
|
| 111 |
"id": 32008,
|
|
|
|
| 112 |
"content": "<extra_id_91>",
|
| 113 |
"single_word": false,
|
| 114 |
"lstrip": false,
|
| 115 |
"rstrip": false,
|
| 116 |
+
"normalized": false,
|
| 117 |
+
"special": true
|
| 118 |
},
|
| 119 |
{
|
| 120 |
"id": 32009,
|
|
|
|
| 121 |
"content": "<extra_id_90>",
|
| 122 |
"single_word": false,
|
| 123 |
"lstrip": false,
|
| 124 |
"rstrip": false,
|
| 125 |
+
"normalized": false,
|
| 126 |
+
"special": true
|
| 127 |
},
|
| 128 |
{
|
| 129 |
"id": 32010,
|
|
|
|
| 130 |
"content": "<extra_id_89>",
|
| 131 |
"single_word": false,
|
| 132 |
"lstrip": false,
|
| 133 |
"rstrip": false,
|
| 134 |
+
"normalized": false,
|
| 135 |
+
"special": true
|
| 136 |
},
|
| 137 |
{
|
| 138 |
"id": 32011,
|
|
|
|
| 139 |
"content": "<extra_id_88>",
|
| 140 |
"single_word": false,
|
| 141 |
"lstrip": false,
|
| 142 |
"rstrip": false,
|
| 143 |
+
"normalized": false,
|
| 144 |
+
"special": true
|
| 145 |
},
|
| 146 |
{
|
| 147 |
"id": 32012,
|
|
|
|
| 148 |
"content": "<extra_id_87>",
|
| 149 |
"single_word": false,
|
| 150 |
"lstrip": false,
|
| 151 |
"rstrip": false,
|
| 152 |
+
"normalized": false,
|
| 153 |
+
"special": true
|
| 154 |
},
|
| 155 |
{
|
| 156 |
"id": 32013,
|
|
|
|
| 157 |
"content": "<extra_id_86>",
|
| 158 |
"single_word": false,
|
| 159 |
"lstrip": false,
|
| 160 |
"rstrip": false,
|
| 161 |
+
"normalized": false,
|
| 162 |
+
"special": true
|
| 163 |
},
|
| 164 |
{
|
| 165 |
"id": 32014,
|
|
|
|
| 166 |
"content": "<extra_id_85>",
|
| 167 |
"single_word": false,
|
| 168 |
"lstrip": false,
|
| 169 |
"rstrip": false,
|
| 170 |
+
"normalized": false,
|
| 171 |
+
"special": true
|
| 172 |
},
|
| 173 |
{
|
| 174 |
"id": 32015,
|
|
|
|
| 175 |
"content": "<extra_id_84>",
|
| 176 |
"single_word": false,
|
| 177 |
"lstrip": false,
|
| 178 |
"rstrip": false,
|
| 179 |
+
"normalized": false,
|
| 180 |
+
"special": true
|
| 181 |
},
|
| 182 |
{
|
| 183 |
"id": 32016,
|
|
|
|
| 184 |
"content": "<extra_id_83>",
|
| 185 |
"single_word": false,
|
| 186 |
"lstrip": false,
|
| 187 |
"rstrip": false,
|
| 188 |
+
"normalized": false,
|
| 189 |
+
"special": true
|
| 190 |
},
|
| 191 |
{
|
| 192 |
"id": 32017,
|
|
|
|
| 193 |
"content": "<extra_id_82>",
|
| 194 |
"single_word": false,
|
| 195 |
"lstrip": false,
|
| 196 |
"rstrip": false,
|
| 197 |
+
"normalized": false,
|
| 198 |
+
"special": true
|
| 199 |
},
|
| 200 |
{
|
| 201 |
"id": 32018,
|
|
|
|
| 202 |
"content": "<extra_id_81>",
|
| 203 |
"single_word": false,
|
| 204 |
"lstrip": false,
|
| 205 |
"rstrip": false,
|
| 206 |
+
"normalized": false,
|
| 207 |
+
"special": true
|
| 208 |
},
|
| 209 |
{
|
| 210 |
"id": 32019,
|
|
|
|
| 211 |
"content": "<extra_id_80>",
|
| 212 |
"single_word": false,
|
| 213 |
"lstrip": false,
|
| 214 |
"rstrip": false,
|
| 215 |
+
"normalized": false,
|
| 216 |
+
"special": true
|
| 217 |
},
|
| 218 |
{
|
| 219 |
"id": 32020,
|
|
|
|
| 220 |
"content": "<extra_id_79>",
|
| 221 |
"single_word": false,
|
| 222 |
"lstrip": false,
|
| 223 |
"rstrip": false,
|
| 224 |
+
"normalized": false,
|
| 225 |
+
"special": true
|
| 226 |
},
|
| 227 |
{
|
| 228 |
"id": 32021,
|
|
|
|
| 229 |
"content": "<extra_id_78>",
|
| 230 |
"single_word": false,
|
| 231 |
"lstrip": false,
|
| 232 |
"rstrip": false,
|
| 233 |
+
"normalized": false,
|
| 234 |
+
"special": true
|
| 235 |
},
|
| 236 |
{
|
| 237 |
"id": 32022,
|
|
|
|
| 238 |
"content": "<extra_id_77>",
|
| 239 |
"single_word": false,
|
| 240 |
"lstrip": false,
|
| 241 |
"rstrip": false,
|
| 242 |
+
"normalized": false,
|
| 243 |
+
"special": true
|
| 244 |
},
|
| 245 |
{
|
| 246 |
"id": 32023,
|
|
|
|
| 247 |
"content": "<extra_id_76>",
|
| 248 |
"single_word": false,
|
| 249 |
"lstrip": false,
|
| 250 |
"rstrip": false,
|
| 251 |
+
"normalized": false,
|
| 252 |
+
"special": true
|
| 253 |
},
|
| 254 |
{
|
| 255 |
"id": 32024,
|
|
|
|
| 256 |
"content": "<extra_id_75>",
|
| 257 |
"single_word": false,
|
| 258 |
"lstrip": false,
|
| 259 |
"rstrip": false,
|
| 260 |
+
"normalized": false,
|
| 261 |
+
"special": true
|
| 262 |
},
|
| 263 |
{
|
| 264 |
"id": 32025,
|
|
|
|
| 265 |
"content": "<extra_id_74>",
|
| 266 |
"single_word": false,
|
| 267 |
"lstrip": false,
|
| 268 |
"rstrip": false,
|
| 269 |
+
"normalized": false,
|
| 270 |
+
"special": true
|
| 271 |
},
|
| 272 |
{
|
| 273 |
"id": 32026,
|
|
|
|
| 274 |
"content": "<extra_id_73>",
|
| 275 |
"single_word": false,
|
| 276 |
"lstrip": false,
|
| 277 |
"rstrip": false,
|
| 278 |
+
"normalized": false,
|
| 279 |
+
"special": true
|
| 280 |
},
|
| 281 |
{
|
| 282 |
"id": 32027,
|
|
|
|
| 283 |
"content": "<extra_id_72>",
|
| 284 |
"single_word": false,
|
| 285 |
"lstrip": false,
|
| 286 |
"rstrip": false,
|
| 287 |
+
"normalized": false,
|
| 288 |
+
"special": true
|
| 289 |
},
|
| 290 |
{
|
| 291 |
"id": 32028,
|
|
|
|
| 292 |
"content": "<extra_id_71>",
|
| 293 |
"single_word": false,
|
| 294 |
"lstrip": false,
|
| 295 |
"rstrip": false,
|
| 296 |
+
"normalized": false,
|
| 297 |
+
"special": true
|
| 298 |
},
|
| 299 |
{
|
| 300 |
"id": 32029,
|
|
|
|
| 301 |
"content": "<extra_id_70>",
|
| 302 |
"single_word": false,
|
| 303 |
"lstrip": false,
|
| 304 |
"rstrip": false,
|
| 305 |
+
"normalized": false,
|
| 306 |
+
"special": true
|
| 307 |
},
|
| 308 |
{
|
| 309 |
"id": 32030,
|
|
|
|
| 310 |
"content": "<extra_id_69>",
|
| 311 |
"single_word": false,
|
| 312 |
"lstrip": false,
|
| 313 |
"rstrip": false,
|
| 314 |
+
"normalized": false,
|
| 315 |
+
"special": true
|
| 316 |
},
|
| 317 |
{
|
| 318 |
"id": 32031,
|
|
|
|
| 319 |
"content": "<extra_id_68>",
|
| 320 |
"single_word": false,
|
| 321 |
"lstrip": false,
|
| 322 |
"rstrip": false,
|
| 323 |
+
"normalized": false,
|
| 324 |
+
"special": true
|
| 325 |
},
|
| 326 |
{
|
| 327 |
"id": 32032,
|
|
|
|
| 328 |
"content": "<extra_id_67>",
|
| 329 |
"single_word": false,
|
| 330 |
"lstrip": false,
|
| 331 |
"rstrip": false,
|
| 332 |
+
"normalized": false,
|
| 333 |
+
"special": true
|
| 334 |
},
|
| 335 |
{
|
| 336 |
"id": 32033,
|
|
|
|
| 337 |
"content": "<extra_id_66>",
|
| 338 |
"single_word": false,
|
| 339 |
"lstrip": false,
|
| 340 |
"rstrip": false,
|
| 341 |
+
"normalized": false,
|
| 342 |
+
"special": true
|
| 343 |
},
|
| 344 |
{
|
| 345 |
"id": 32034,
|
|
|
|
| 346 |
"content": "<extra_id_65>",
|
| 347 |
"single_word": false,
|
| 348 |
"lstrip": false,
|
| 349 |
"rstrip": false,
|
| 350 |
+
"normalized": false,
|
| 351 |
+
"special": true
|
| 352 |
},
|
| 353 |
{
|
| 354 |
"id": 32035,
|
|
|
|
| 355 |
"content": "<extra_id_64>",
|
| 356 |
"single_word": false,
|
| 357 |
"lstrip": false,
|
| 358 |
"rstrip": false,
|
| 359 |
+
"normalized": false,
|
| 360 |
+
"special": true
|
| 361 |
},
|
| 362 |
{
|
| 363 |
"id": 32036,
|
|
|
|
| 364 |
"content": "<extra_id_63>",
|
| 365 |
"single_word": false,
|
| 366 |
"lstrip": false,
|
| 367 |
"rstrip": false,
|
| 368 |
+
"normalized": false,
|
| 369 |
+
"special": true
|
| 370 |
},
|
| 371 |
{
|
| 372 |
"id": 32037,
|
|
|
|
| 373 |
"content": "<extra_id_62>",
|
| 374 |
"single_word": false,
|
| 375 |
"lstrip": false,
|
| 376 |
"rstrip": false,
|
| 377 |
+
"normalized": false,
|
| 378 |
+
"special": true
|
| 379 |
},
|
| 380 |
{
|
| 381 |
"id": 32038,
|
|
|
|
| 382 |
"content": "<extra_id_61>",
|
| 383 |
"single_word": false,
|
| 384 |
"lstrip": false,
|
| 385 |
"rstrip": false,
|
| 386 |
+
"normalized": false,
|
| 387 |
+
"special": true
|
| 388 |
},
|
| 389 |
{
|
| 390 |
"id": 32039,
|
|
|
|
| 391 |
"content": "<extra_id_60>",
|
| 392 |
"single_word": false,
|
| 393 |
"lstrip": false,
|
| 394 |
"rstrip": false,
|
| 395 |
+
"normalized": false,
|
| 396 |
+
"special": true
|
| 397 |
},
|
| 398 |
{
|
| 399 |
"id": 32040,
|
|
|
|
| 400 |
"content": "<extra_id_59>",
|
| 401 |
"single_word": false,
|
| 402 |
"lstrip": false,
|
| 403 |
"rstrip": false,
|
| 404 |
+
"normalized": false,
|
| 405 |
+
"special": true
|
| 406 |
},
|
| 407 |
{
|
| 408 |
"id": 32041,
|
|
|
|
| 409 |
"content": "<extra_id_58>",
|
| 410 |
"single_word": false,
|
| 411 |
"lstrip": false,
|
| 412 |
"rstrip": false,
|
| 413 |
+
"normalized": false,
|
| 414 |
+
"special": true
|
| 415 |
},
|
| 416 |
{
|
| 417 |
"id": 32042,
|
|
|
|
| 418 |
"content": "<extra_id_57>",
|
| 419 |
"single_word": false,
|
| 420 |
"lstrip": false,
|
| 421 |
"rstrip": false,
|
| 422 |
+
"normalized": false,
|
| 423 |
+
"special": true
|
| 424 |
},
|
| 425 |
{
|
| 426 |
"id": 32043,
|
|
|
|
| 427 |
"content": "<extra_id_56>",
|
| 428 |
"single_word": false,
|
| 429 |
"lstrip": false,
|
| 430 |
"rstrip": false,
|
| 431 |
+
"normalized": false,
|
| 432 |
+
"special": true
|
| 433 |
},
|
| 434 |
{
|
| 435 |
"id": 32044,
|
|
|
|
| 436 |
"content": "<extra_id_55>",
|
| 437 |
"single_word": false,
|
| 438 |
"lstrip": false,
|
| 439 |
"rstrip": false,
|
| 440 |
+
"normalized": false,
|
| 441 |
+
"special": true
|
| 442 |
},
|
| 443 |
{
|
| 444 |
"id": 32045,
|
|
|
|
| 445 |
"content": "<extra_id_54>",
|
| 446 |
"single_word": false,
|
| 447 |
"lstrip": false,
|
| 448 |
"rstrip": false,
|
| 449 |
+
"normalized": false,
|
| 450 |
+
"special": true
|
| 451 |
},
|
| 452 |
{
|
| 453 |
"id": 32046,
|
|
|
|
| 454 |
"content": "<extra_id_53>",
|
| 455 |
"single_word": false,
|
| 456 |
"lstrip": false,
|
| 457 |
"rstrip": false,
|
| 458 |
+
"normalized": false,
|
| 459 |
+
"special": true
|
| 460 |
},
|
| 461 |
{
|
| 462 |
"id": 32047,
|
|
|
|
| 463 |
"content": "<extra_id_52>",
|
| 464 |
"single_word": false,
|
| 465 |
"lstrip": false,
|
| 466 |
"rstrip": false,
|
| 467 |
+
"normalized": false,
|
| 468 |
+
"special": true
|
| 469 |
},
|
| 470 |
{
|
| 471 |
"id": 32048,
|
|
|
|
| 472 |
"content": "<extra_id_51>",
|
| 473 |
"single_word": false,
|
| 474 |
"lstrip": false,
|
| 475 |
"rstrip": false,
|
| 476 |
+
"normalized": false,
|
| 477 |
+
"special": true
|
| 478 |
},
|
| 479 |
{
|
| 480 |
"id": 32049,
|
|
|
|
| 481 |
"content": "<extra_id_50>",
|
| 482 |
"single_word": false,
|
| 483 |
"lstrip": false,
|
| 484 |
"rstrip": false,
|
| 485 |
+
"normalized": false,
|
| 486 |
+
"special": true
|
| 487 |
},
|
| 488 |
{
|
| 489 |
"id": 32050,
|
|
|
|
| 490 |
"content": "<extra_id_49>",
|
| 491 |
"single_word": false,
|
| 492 |
"lstrip": false,
|
| 493 |
"rstrip": false,
|
| 494 |
+
"normalized": false,
|
| 495 |
+
"special": true
|
| 496 |
},
|
| 497 |
{
|
| 498 |
"id": 32051,
|
|
|
|
| 499 |
"content": "<extra_id_48>",
|
| 500 |
"single_word": false,
|
| 501 |
"lstrip": false,
|
| 502 |
"rstrip": false,
|
| 503 |
+
"normalized": false,
|
| 504 |
+
"special": true
|
| 505 |
},
|
| 506 |
{
|
| 507 |
"id": 32052,
|
|
|
|
| 508 |
"content": "<extra_id_47>",
|
| 509 |
"single_word": false,
|
| 510 |
"lstrip": false,
|
| 511 |
"rstrip": false,
|
| 512 |
+
"normalized": false,
|
| 513 |
+
"special": true
|
| 514 |
},
|
| 515 |
{
|
| 516 |
"id": 32053,
|
|
|
|
| 517 |
"content": "<extra_id_46>",
|
| 518 |
"single_word": false,
|
| 519 |
"lstrip": false,
|
| 520 |
"rstrip": false,
|
| 521 |
+
"normalized": false,
|
| 522 |
+
"special": true
|
| 523 |
},
|
| 524 |
{
|
| 525 |
"id": 32054,
|
|
|
|
| 526 |
"content": "<extra_id_45>",
|
| 527 |
"single_word": false,
|
| 528 |
"lstrip": false,
|
| 529 |
"rstrip": false,
|
| 530 |
+
"normalized": false,
|
| 531 |
+
"special": true
|
| 532 |
},
|
| 533 |
{
|
| 534 |
"id": 32055,
|
|
|
|
| 535 |
"content": "<extra_id_44>",
|
| 536 |
"single_word": false,
|
| 537 |
"lstrip": false,
|
| 538 |
"rstrip": false,
|
| 539 |
+
"normalized": false,
|
| 540 |
+
"special": true
|
| 541 |
},
|
| 542 |
{
|
| 543 |
"id": 32056,
|
|
|
|
| 544 |
"content": "<extra_id_43>",
|
| 545 |
"single_word": false,
|
| 546 |
"lstrip": false,
|
| 547 |
"rstrip": false,
|
| 548 |
+
"normalized": false,
|
| 549 |
+
"special": true
|
| 550 |
},
|
| 551 |
{
|
| 552 |
"id": 32057,
|
|
|
|
| 553 |
"content": "<extra_id_42>",
|
| 554 |
"single_word": false,
|
| 555 |
"lstrip": false,
|
| 556 |
"rstrip": false,
|
| 557 |
+
"normalized": false,
|
| 558 |
+
"special": true
|
| 559 |
},
|
| 560 |
{
|
| 561 |
"id": 32058,
|
|
|
|
| 562 |
"content": "<extra_id_41>",
|
| 563 |
"single_word": false,
|
| 564 |
"lstrip": false,
|
| 565 |
"rstrip": false,
|
| 566 |
+
"normalized": false,
|
| 567 |
+
"special": true
|
| 568 |
},
|
| 569 |
{
|
| 570 |
"id": 32059,
|
|
|
|
| 571 |
"content": "<extra_id_40>",
|
| 572 |
"single_word": false,
|
| 573 |
"lstrip": false,
|
| 574 |
"rstrip": false,
|
| 575 |
+
"normalized": false,
|
| 576 |
+
"special": true
|
| 577 |
},
|
| 578 |
{
|
| 579 |
"id": 32060,
|
|
|
|
| 580 |
"content": "<extra_id_39>",
|
| 581 |
"single_word": false,
|
| 582 |
"lstrip": false,
|
| 583 |
"rstrip": false,
|
| 584 |
+
"normalized": false,
|
| 585 |
+
"special": true
|
| 586 |
},
|
| 587 |
{
|
| 588 |
"id": 32061,
|
|
|
|
| 589 |
"content": "<extra_id_38>",
|
| 590 |
"single_word": false,
|
| 591 |
"lstrip": false,
|
| 592 |
"rstrip": false,
|
| 593 |
+
"normalized": false,
|
| 594 |
+
"special": true
|
| 595 |
},
|
| 596 |
{
|
| 597 |
"id": 32062,
|
|
|
|
| 598 |
"content": "<extra_id_37>",
|
| 599 |
"single_word": false,
|
| 600 |
"lstrip": false,
|
| 601 |
"rstrip": false,
|
| 602 |
+
"normalized": false,
|
| 603 |
+
"special": true
|
| 604 |
},
|
| 605 |
{
|
| 606 |
"id": 32063,
|
|
|
|
| 607 |
"content": "<extra_id_36>",
|
| 608 |
"single_word": false,
|
| 609 |
"lstrip": false,
|
| 610 |
"rstrip": false,
|
| 611 |
+
"normalized": false,
|
| 612 |
+
"special": true
|
| 613 |
},
|
| 614 |
{
|
| 615 |
"id": 32064,
|
|
|
|
| 616 |
"content": "<extra_id_35>",
|
| 617 |
"single_word": false,
|
| 618 |
"lstrip": false,
|
| 619 |
"rstrip": false,
|
| 620 |
+
"normalized": false,
|
| 621 |
+
"special": true
|
| 622 |
},
|
| 623 |
{
|
| 624 |
"id": 32065,
|
|
|
|
| 625 |
"content": "<extra_id_34>",
|
| 626 |
"single_word": false,
|
| 627 |
"lstrip": false,
|
| 628 |
"rstrip": false,
|
| 629 |
+
"normalized": false,
|
| 630 |
+
"special": true
|
| 631 |
},
|
| 632 |
{
|
| 633 |
"id": 32066,
|
|
|
|
| 634 |
"content": "<extra_id_33>",
|
| 635 |
"single_word": false,
|
| 636 |
"lstrip": false,
|
| 637 |
"rstrip": false,
|
| 638 |
+
"normalized": false,
|
| 639 |
+
"special": true
|
| 640 |
},
|
| 641 |
{
|
| 642 |
"id": 32067,
|
|
|
|
| 643 |
"content": "<extra_id_32>",
|
| 644 |
"single_word": false,
|
| 645 |
"lstrip": false,
|
| 646 |
"rstrip": false,
|
| 647 |
+
"normalized": false,
|
| 648 |
+
"special": true
|
| 649 |
},
|
| 650 |
{
|
| 651 |
"id": 32068,
|
|
|
|
| 652 |
"content": "<extra_id_31>",
|
| 653 |
"single_word": false,
|
| 654 |
"lstrip": false,
|
| 655 |
"rstrip": false,
|
| 656 |
+
"normalized": false,
|
| 657 |
+
"special": true
|
| 658 |
},
|
| 659 |
{
|
| 660 |
"id": 32069,
|
|
|
|
| 661 |
"content": "<extra_id_30>",
|
| 662 |
"single_word": false,
|
| 663 |
"lstrip": false,
|
| 664 |
"rstrip": false,
|
| 665 |
+
"normalized": false,
|
| 666 |
+
"special": true
|
| 667 |
},
|
| 668 |
{
|
| 669 |
"id": 32070,
|
|
|
|
| 670 |
"content": "<extra_id_29>",
|
| 671 |
"single_word": false,
|
| 672 |
"lstrip": false,
|
| 673 |
"rstrip": false,
|
| 674 |
+
"normalized": false,
|
| 675 |
+
"special": true
|
| 676 |
},
|
| 677 |
{
|
| 678 |
"id": 32071,
|
|
|
|
| 679 |
"content": "<extra_id_28>",
|
| 680 |
"single_word": false,
|
| 681 |
"lstrip": false,
|
| 682 |
"rstrip": false,
|
| 683 |
+
"normalized": false,
|
| 684 |
+
"special": true
|
| 685 |
},
|
| 686 |
{
|
| 687 |
"id": 32072,
|
|
|
|
| 688 |
"content": "<extra_id_27>",
|
| 689 |
"single_word": false,
|
| 690 |
"lstrip": false,
|
| 691 |
"rstrip": false,
|
| 692 |
+
"normalized": false,
|
| 693 |
+
"special": true
|
| 694 |
},
|
| 695 |
{
|
| 696 |
"id": 32073,
|
|
|
|
| 697 |
"content": "<extra_id_26>",
|
| 698 |
"single_word": false,
|
| 699 |
"lstrip": false,
|
| 700 |
"rstrip": false,
|
| 701 |
+
"normalized": false,
|
| 702 |
+
"special": true
|
| 703 |
},
|
| 704 |
{
|
| 705 |
"id": 32074,
|
|
|
|
| 706 |
"content": "<extra_id_25>",
|
| 707 |
"single_word": false,
|
| 708 |
"lstrip": false,
|
| 709 |
"rstrip": false,
|
| 710 |
+
"normalized": false,
|
| 711 |
+
"special": true
|
| 712 |
},
|
| 713 |
{
|
| 714 |
"id": 32075,
|
|
|
|
| 715 |
"content": "<extra_id_24>",
|
| 716 |
"single_word": false,
|
| 717 |
"lstrip": false,
|
| 718 |
"rstrip": false,
|
| 719 |
+
"normalized": false,
|
| 720 |
+
"special": true
|
| 721 |
},
|
| 722 |
{
|
| 723 |
"id": 32076,
|
|
|
|
| 724 |
"content": "<extra_id_23>",
|
| 725 |
"single_word": false,
|
| 726 |
"lstrip": false,
|
| 727 |
"rstrip": false,
|
| 728 |
+
"normalized": false,
|
| 729 |
+
"special": true
|
| 730 |
},
|
| 731 |
{
|
| 732 |
"id": 32077,
|
|
|
|
| 733 |
"content": "<extra_id_22>",
|
| 734 |
"single_word": false,
|
| 735 |
"lstrip": false,
|
| 736 |
"rstrip": false,
|
| 737 |
+
"normalized": false,
|
| 738 |
+
"special": true
|
| 739 |
},
|
| 740 |
{
|
| 741 |
"id": 32078,
|
|
|
|
| 742 |
"content": "<extra_id_21>",
|
| 743 |
"single_word": false,
|
| 744 |
"lstrip": false,
|
| 745 |
"rstrip": false,
|
| 746 |
+
"normalized": false,
|
| 747 |
+
"special": true
|
| 748 |
},
|
| 749 |
{
|
| 750 |
"id": 32079,
|
|
|
|
| 751 |
"content": "<extra_id_20>",
|
| 752 |
"single_word": false,
|
| 753 |
"lstrip": false,
|
| 754 |
"rstrip": false,
|
| 755 |
+
"normalized": false,
|
| 756 |
+
"special": true
|
| 757 |
},
|
| 758 |
{
|
| 759 |
"id": 32080,
|
|
|
|
| 760 |
"content": "<extra_id_19>",
|
| 761 |
"single_word": false,
|
| 762 |
"lstrip": false,
|
| 763 |
"rstrip": false,
|
| 764 |
+
"normalized": false,
|
| 765 |
+
"special": true
|
| 766 |
},
|
| 767 |
{
|
| 768 |
"id": 32081,
|
|
|
|
| 769 |
"content": "<extra_id_18>",
|
| 770 |
"single_word": false,
|
| 771 |
"lstrip": false,
|
| 772 |
"rstrip": false,
|
| 773 |
+
"normalized": false,
|
| 774 |
+
"special": true
|
| 775 |
},
|
| 776 |
{
|
| 777 |
"id": 32082,
|
|
|
|
| 778 |
"content": "<extra_id_17>",
|
| 779 |
"single_word": false,
|
| 780 |
"lstrip": false,
|
| 781 |
"rstrip": false,
|
| 782 |
+
"normalized": false,
|
| 783 |
+
"special": true
|
| 784 |
},
|
| 785 |
{
|
| 786 |
"id": 32083,
|
|
|
|
| 787 |
"content": "<extra_id_16>",
|
| 788 |
"single_word": false,
|
| 789 |
"lstrip": false,
|
| 790 |
"rstrip": false,
|
| 791 |
+
"normalized": false,
|
| 792 |
+
"special": true
|
| 793 |
},
|
| 794 |
{
|
| 795 |
"id": 32084,
|
|
|
|
| 796 |
"content": "<extra_id_15>",
|
| 797 |
"single_word": false,
|
| 798 |
"lstrip": false,
|
| 799 |
"rstrip": false,
|
| 800 |
+
"normalized": false,
|
| 801 |
+
"special": true
|
| 802 |
},
|
| 803 |
{
|
| 804 |
"id": 32085,
|
|
|
|
| 805 |
"content": "<extra_id_14>",
|
| 806 |
"single_word": false,
|
| 807 |
"lstrip": false,
|
| 808 |
"rstrip": false,
|
| 809 |
+
"normalized": false,
|
| 810 |
+
"special": true
|
| 811 |
},
|
| 812 |
{
|
| 813 |
"id": 32086,
|
|
|
|
| 814 |
"content": "<extra_id_13>",
|
| 815 |
"single_word": false,
|
| 816 |
"lstrip": false,
|
| 817 |
"rstrip": false,
|
| 818 |
+
"normalized": false,
|
| 819 |
+
"special": true
|
| 820 |
},
|
| 821 |
{
|
| 822 |
"id": 32087,
|
|
|
|
| 823 |
"content": "<extra_id_12>",
|
| 824 |
"single_word": false,
|
| 825 |
"lstrip": false,
|
| 826 |
"rstrip": false,
|
| 827 |
+
"normalized": false,
|
| 828 |
+
"special": true
|
| 829 |
},
|
| 830 |
{
|
| 831 |
"id": 32088,
|
|
|
|
| 832 |
"content": "<extra_id_11>",
|
| 833 |
"single_word": false,
|
| 834 |
"lstrip": false,
|
| 835 |
"rstrip": false,
|
| 836 |
+
"normalized": false,
|
| 837 |
+
"special": true
|
| 838 |
},
|
| 839 |
{
|
| 840 |
"id": 32089,
|
|
|
|
| 841 |
"content": "<extra_id_10>",
|
| 842 |
"single_word": false,
|
| 843 |
"lstrip": false,
|
| 844 |
"rstrip": false,
|
| 845 |
+
"normalized": false,
|
| 846 |
+
"special": true
|
| 847 |
},
|
| 848 |
{
|
| 849 |
"id": 32090,
|
|
|
|
| 850 |
"content": "<extra_id_9>",
|
| 851 |
"single_word": false,
|
| 852 |
"lstrip": false,
|
| 853 |
"rstrip": false,
|
| 854 |
+
"normalized": false,
|
| 855 |
+
"special": true
|
| 856 |
},
|
| 857 |
{
|
| 858 |
"id": 32091,
|
|
|
|
| 859 |
"content": "<extra_id_8>",
|
| 860 |
"single_word": false,
|
| 861 |
"lstrip": false,
|
| 862 |
"rstrip": false,
|
| 863 |
+
"normalized": false,
|
| 864 |
+
"special": true
|
| 865 |
},
|
| 866 |
{
|
| 867 |
"id": 32092,
|
|
|
|
| 868 |
"content": "<extra_id_7>",
|
| 869 |
"single_word": false,
|
| 870 |
"lstrip": false,
|
| 871 |
"rstrip": false,
|
| 872 |
+
"normalized": false,
|
| 873 |
+
"special": true
|
| 874 |
},
|
| 875 |
{
|
| 876 |
"id": 32093,
|
|
|
|
| 877 |
"content": "<extra_id_6>",
|
| 878 |
"single_word": false,
|
| 879 |
"lstrip": false,
|
| 880 |
"rstrip": false,
|
| 881 |
+
"normalized": false,
|
| 882 |
+
"special": true
|
| 883 |
},
|
| 884 |
{
|
| 885 |
"id": 32094,
|
|
|
|
| 886 |
"content": "<extra_id_5>",
|
| 887 |
"single_word": false,
|
| 888 |
"lstrip": false,
|
| 889 |
"rstrip": false,
|
| 890 |
+
"normalized": false,
|
| 891 |
+
"special": true
|
| 892 |
},
|
| 893 |
{
|
| 894 |
"id": 32095,
|
|
|
|
| 895 |
"content": "<extra_id_4>",
|
| 896 |
"single_word": false,
|
| 897 |
"lstrip": false,
|
| 898 |
"rstrip": false,
|
| 899 |
+
"normalized": false,
|
| 900 |
+
"special": true
|
| 901 |
},
|
| 902 |
{
|
| 903 |
"id": 32096,
|
|
|
|
| 904 |
"content": "<extra_id_3>",
|
| 905 |
"single_word": false,
|
| 906 |
"lstrip": false,
|
| 907 |
"rstrip": false,
|
| 908 |
+
"normalized": false,
|
| 909 |
+
"special": true
|
| 910 |
},
|
| 911 |
{
|
| 912 |
"id": 32097,
|
|
|
|
| 913 |
"content": "<extra_id_2>",
|
| 914 |
"single_word": false,
|
| 915 |
"lstrip": false,
|
| 916 |
"rstrip": false,
|
| 917 |
+
"normalized": false,
|
| 918 |
+
"special": true
|
| 919 |
},
|
| 920 |
{
|
| 921 |
"id": 32098,
|
|
|
|
| 922 |
"content": "<extra_id_1>",
|
| 923 |
"single_word": false,
|
| 924 |
"lstrip": false,
|
| 925 |
"rstrip": false,
|
| 926 |
+
"normalized": false,
|
| 927 |
+
"special": true
|
| 928 |
},
|
| 929 |
{
|
| 930 |
"id": 32099,
|
|
|
|
| 931 |
"content": "<extra_id_0>",
|
| 932 |
"single_word": false,
|
| 933 |
"lstrip": false,
|
| 934 |
"rstrip": false,
|
| 935 |
+
"normalized": false,
|
| 936 |
+
"special": true
|
| 937 |
}
|
| 938 |
],
|
| 939 |
"normalizer": {
|
tokenizer_config.json
CHANGED
|
@@ -106,6 +106,7 @@
|
|
| 106 |
"extra_ids": 100,
|
| 107 |
"model_max_length": 512,
|
| 108 |
"pad_token": "<pad>",
|
|
|
|
| 109 |
"tokenizer_class": "T5Tokenizer",
|
| 110 |
"unk_token": "<unk>"
|
| 111 |
}
|
|
|
|
| 106 |
"extra_ids": 100,
|
| 107 |
"model_max_length": 512,
|
| 108 |
"pad_token": "<pad>",
|
| 109 |
+
"special_tokens_map_file": null,
|
| 110 |
"tokenizer_class": "T5Tokenizer",
|
| 111 |
"unk_token": "<unk>"
|
| 112 |
}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:33b5537881331cd162c1a93fe5b26232f6116acdf2ffa6eb833d9c3e137e2e6d
|
| 3 |
+
size 3707
|