Upload tokenizer
Browse files- special_tokens_map.json +3 -3
- tokenizer.json +99 -99
- tokenizer_config.json +3 -5
special_tokens_map.json
CHANGED
|
@@ -2,21 +2,21 @@
|
|
| 2 |
"bos_token": {
|
| 3 |
"content": "<s>",
|
| 4 |
"lstrip": false,
|
| 5 |
-
"normalized":
|
| 6 |
"rstrip": false,
|
| 7 |
"single_word": false
|
| 8 |
},
|
| 9 |
"eos_token": {
|
| 10 |
"content": "</s>",
|
| 11 |
"lstrip": false,
|
| 12 |
-
"normalized":
|
| 13 |
"rstrip": false,
|
| 14 |
"single_word": false
|
| 15 |
},
|
| 16 |
"unk_token": {
|
| 17 |
"content": "<unk>",
|
| 18 |
"lstrip": false,
|
| 19 |
-
"normalized":
|
| 20 |
"rstrip": false,
|
| 21 |
"single_word": false
|
| 22 |
}
|
|
|
|
| 2 |
"bos_token": {
|
| 3 |
"content": "<s>",
|
| 4 |
"lstrip": false,
|
| 5 |
+
"normalized": true,
|
| 6 |
"rstrip": false,
|
| 7 |
"single_word": false
|
| 8 |
},
|
| 9 |
"eos_token": {
|
| 10 |
"content": "</s>",
|
| 11 |
"lstrip": false,
|
| 12 |
+
"normalized": true,
|
| 13 |
"rstrip": false,
|
| 14 |
"single_word": false
|
| 15 |
},
|
| 16 |
"unk_token": {
|
| 17 |
"content": "<unk>",
|
| 18 |
"lstrip": false,
|
| 19 |
+
"normalized": true,
|
| 20 |
"rstrip": false,
|
| 21 |
"single_word": false
|
| 22 |
}
|
tokenizer.json
CHANGED
|
@@ -9,7 +9,7 @@
|
|
| 9 |
"single_word": false,
|
| 10 |
"lstrip": false,
|
| 11 |
"rstrip": false,
|
| 12 |
-
"normalized":
|
| 13 |
"special": true
|
| 14 |
},
|
| 15 |
{
|
|
@@ -18,7 +18,7 @@
|
|
| 18 |
"single_word": false,
|
| 19 |
"lstrip": false,
|
| 20 |
"rstrip": false,
|
| 21 |
-
"normalized":
|
| 22 |
"special": true
|
| 23 |
},
|
| 24 |
{
|
|
@@ -27,7 +27,7 @@
|
|
| 27 |
"single_word": false,
|
| 28 |
"lstrip": false,
|
| 29 |
"rstrip": false,
|
| 30 |
-
"normalized":
|
| 31 |
"special": true
|
| 32 |
}
|
| 33 |
],
|
|
@@ -93266,126 +93266,126 @@
|
|
| 93266 |
"βlivre s",
|
| 93267 |
"lu b",
|
| 93268 |
"l ub",
|
| 93269 |
-
"β β",
|
| 93270 |
"ββ ββ",
|
| 93271 |
-
"
|
| 93272 |
-
"
|
|
|
|
| 93273 |
"ββ ββββββ",
|
| 93274 |
-
"
|
| 93275 |
-
"
|
| 93276 |
-
"
|
| 93277 |
-
"βββ βββββ",
|
| 93278 |
-
"βββββββ β",
|
| 93279 |
-
"β βββββββ",
|
| 93280 |
-
"ββ βββ",
|
| 93281 |
-
"ββββ β",
|
| 93282 |
-
"βββ ββ",
|
| 93283 |
-
"β ββββ",
|
| 93284 |
"ββ ββββββββββββββ",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93285 |
"ββββ ββββββββββββ",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93286 |
"ββββββββ ββββββββ",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93287 |
"βββββ βββββββββββ",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93288 |
"ββββββ ββββββββββ",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93289 |
"ββββββββββββ ββββ",
|
|
|
|
|
|
|
|
|
|
| 93290 |
"βββββββββββββ βββ",
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93291 |
"ββββββββββ ββββββ",
|
|
|
|
|
|
|
| 93292 |
"ββββββββββββββ ββ",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93293 |
"βββ βββββββββββββ",
|
| 93294 |
-
"
|
| 93295 |
-
"βββββββ βββββββββ",
|
| 93296 |
-
"βββββββββββ βββββ",
|
| 93297 |
-
"βββββββββββββββ β",
|
| 93298 |
-
"β βββββββββββββββ",
|
| 93299 |
-
"ββ ββββ",
|
| 93300 |
-
"ββββ ββ",
|
| 93301 |
-
"βββββ β",
|
| 93302 |
"βββ βββ",
|
| 93303 |
-
"β βββββ",
|
| 93304 |
-
"ββ ββοΏ½οΏ½βββββββ",
|
| 93305 |
-
"ββββ ββββββββ",
|
| 93306 |
-
"ββββββββ ββββ",
|
| 93307 |
-
"βββββ βββββββ",
|
| 93308 |
-
"ββββββ ββββββ",
|
| 93309 |
-
"ββββββββββ ββ",
|
| 93310 |
"βββ βββββββββ",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93311 |
"βββββββββ βββ",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93312 |
"βββββββ βββββ",
|
| 93313 |
-
"βββββββββββ β",
|
| 93314 |
-
"β βββββββββββ",
|
| 93315 |
-
"ββ βββββββββββ",
|
| 93316 |
-
"ββββ βββββββββ",
|
| 93317 |
-
"ββββββββ βββββ",
|
| 93318 |
-
"βββββ ββββββββ",
|
| 93319 |
-
"ββββββ βββββββ",
|
| 93320 |
-
"ββββββββββββ β",
|
| 93321 |
-
"ββββββββββ βββ",
|
| 93322 |
-
"βββ ββββββββββ",
|
| 93323 |
-
"βββββββββ ββββ",
|
| 93324 |
"βββββββ ββββββ",
|
| 93325 |
-
"βββββββββββ ββ",
|
| 93326 |
-
"β ββββββββββββ",
|
| 93327 |
-
"ββ ββββββββ",
|
| 93328 |
-
"ββββ ββββββ",
|
| 93329 |
-
"ββββββββ ββ",
|
| 93330 |
-
"βββββ βββββ",
|
| 93331 |
-
"ββββββ ββββ",
|
| 93332 |
-
"βββ βββββββ",
|
| 93333 |
-
"βββββββββ β",
|
| 93334 |
"βββββββ βββ",
|
| 93335 |
-
"
|
| 93336 |
-
"ββ ββββββββββββ",
|
| 93337 |
-
"ββββ ββββββββββ",
|
| 93338 |
-
"ββββββββ ββββββ",
|
| 93339 |
-
"βββββ βββββββββ",
|
| 93340 |
-
"ββββββ ββββββββ",
|
| 93341 |
-
"ββββββββββββ ββ",
|
| 93342 |
-
"βββββββββββββ β",
|
| 93343 |
-
"ββββββββββ ββββ",
|
| 93344 |
-
"βββ βββββββββββ",
|
| 93345 |
-
"βββββββββ βββββ",
|
| 93346 |
"βββββββ βββββββ",
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93347 |
"βββββββββββ βββ",
|
| 93348 |
-
"β
|
| 93349 |
-
"
|
| 93350 |
"β ββ",
|
| 93351 |
-
"
|
| 93352 |
-
"ββββ βββββ",
|
| 93353 |
-
"ββββββββ β",
|
| 93354 |
-
"βββββ ββββ",
|
| 93355 |
-
"ββββββ βββ",
|
| 93356 |
-
"βββ ββββββ",
|
| 93357 |
-
"βββββββ ββ",
|
| 93358 |
"β ββββββββ",
|
| 93359 |
-
"
|
| 93360 |
-
"ββββ βββ",
|
| 93361 |
-
"βββββ ββ",
|
| 93362 |
-
"ββββββ β",
|
| 93363 |
-
"βββ ββββ",
|
| 93364 |
"β ββββββ",
|
| 93365 |
-
"
|
| 93366 |
-
"
|
| 93367 |
-
"ββββββββ βββ",
|
| 93368 |
-
"βββββ ββββββ",
|
| 93369 |
-
"ββββββ βββββ",
|
| 93370 |
-
"ββββββββββ β",
|
| 93371 |
-
"βββ ββββββββ",
|
| 93372 |
-
"βββββββββ ββ",
|
| 93373 |
-
"βββββββ ββββ",
|
| 93374 |
"β ββββββββββ",
|
| 93375 |
-
"
|
| 93376 |
-
"
|
| 93377 |
-
"
|
| 93378 |
-
"
|
| 93379 |
-
"
|
| 93380 |
-
"
|
| 93381 |
-
"
|
| 93382 |
-
"ββββββββββ βββββ",
|
| 93383 |
-
"ββββββββββββββ β",
|
| 93384 |
-
"βββ ββββββββββββ",
|
| 93385 |
-
"βββββββββ ββββββ",
|
| 93386 |
-
"βββββββ ββββββββ",
|
| 93387 |
-
"βββββββββββ ββββ",
|
| 93388 |
-
"β ββββββββββββββ"
|
| 93389 |
]
|
| 93390 |
}
|
| 93391 |
}
|
|
|
|
| 9 |
"single_word": false,
|
| 10 |
"lstrip": false,
|
| 11 |
"rstrip": false,
|
| 12 |
+
"normalized": true,
|
| 13 |
"special": true
|
| 14 |
},
|
| 15 |
{
|
|
|
|
| 18 |
"single_word": false,
|
| 19 |
"lstrip": false,
|
| 20 |
"rstrip": false,
|
| 21 |
+
"normalized": true,
|
| 22 |
"special": true
|
| 23 |
},
|
| 24 |
{
|
|
|
|
| 27 |
"single_word": false,
|
| 28 |
"lstrip": false,
|
| 29 |
"rstrip": false,
|
| 30 |
+
"normalized": true,
|
| 31 |
"special": true
|
| 32 |
}
|
| 33 |
],
|
|
|
|
| 93266 |
"βlivre s",
|
| 93267 |
"lu b",
|
| 93268 |
"l ub",
|
|
|
|
| 93269 |
"ββ ββ",
|
| 93270 |
+
"ββ ββββ",
|
| 93271 |
+
"ββ ββββββββ",
|
| 93272 |
+
"ββ βββββ",
|
| 93273 |
"ββ ββββββ",
|
| 93274 |
+
"ββ ββββββββββββ",
|
| 93275 |
+
"ββ βββββββββββββ",
|
| 93276 |
+
"ββ ββββββββββ",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93277 |
"ββ ββββββββββββββ",
|
| 93278 |
+
"ββ βββ",
|
| 93279 |
+
"ββ βββββββββ",
|
| 93280 |
+
"ββ βββββββ",
|
| 93281 |
+
"ββ βββββββββββ",
|
| 93282 |
+
"ββ β",
|
| 93283 |
+
"ββββ ββ",
|
| 93284 |
+
"ββββ ββββ",
|
| 93285 |
+
"ββββ ββββββββ",
|
| 93286 |
+
"ββββ βββββ",
|
| 93287 |
+
"ββββ ββββββ",
|
| 93288 |
"ββββ ββββββββββββ",
|
| 93289 |
+
"ββββ ββββββββββ",
|
| 93290 |
+
"ββββ βββ",
|
| 93291 |
+
"ββββ βββββββββ",
|
| 93292 |
+
"ββββ βββββββ",
|
| 93293 |
+
"ββββ βββββββββββ",
|
| 93294 |
+
"ββββ β",
|
| 93295 |
+
"ββββββββ ββ",
|
| 93296 |
+
"ββββββββ ββββ",
|
| 93297 |
"ββββββββ ββββββββ",
|
| 93298 |
+
"ββββββββ βββββ",
|
| 93299 |
+
"ββββββββ ββββββ",
|
| 93300 |
+
"ββββββββ βββ",
|
| 93301 |
+
"ββββββββ βββββββ",
|
| 93302 |
+
"ββββββββ β",
|
| 93303 |
+
"βββββ ββ",
|
| 93304 |
+
"βββββ ββββ",
|
| 93305 |
+
"βββββ ββββββββ",
|
| 93306 |
+
"βββββ βββββ",
|
| 93307 |
+
"βββββ ββββββ",
|
| 93308 |
+
"βββββ ββββββββββ",
|
| 93309 |
+
"βββββ βββ",
|
| 93310 |
+
"βββββ βββββββββ",
|
| 93311 |
+
"βββββ βββββββ",
|
| 93312 |
"βββββ βββββββββββ",
|
| 93313 |
+
"βββββ β",
|
| 93314 |
+
"ββββββ ββ",
|
| 93315 |
+
"ββββββ ββββ",
|
| 93316 |
+
"ββββββ ββββββββ",
|
| 93317 |
+
"ββββββ βββββ",
|
| 93318 |
+
"ββββββ ββββββ",
|
| 93319 |
"ββββββ ββββββββββ",
|
| 93320 |
+
"ββββββ βββ",
|
| 93321 |
+
"ββββββ βββββββββ",
|
| 93322 |
+
"ββββββ βββββββ",
|
| 93323 |
+
"ββββββ β",
|
| 93324 |
+
"ββββββββββββ ββ",
|
| 93325 |
"ββββββββββββ ββββ",
|
| 93326 |
+
"ββββββββββββ βββ",
|
| 93327 |
+
"ββββββββββββ β",
|
| 93328 |
+
"βββββββββββββ ββ",
|
| 93329 |
"βββββββββββββ βββ",
|
| 93330 |
+
"βββββββββββββ β",
|
| 93331 |
+
"ββββββββββ ββ",
|
| 93332 |
+
"ββββββββββ ββββ",
|
| 93333 |
+
"ββββββββββ βββββ",
|
| 93334 |
"ββββββββββ ββββββ",
|
| 93335 |
+
"ββββββββββ βββ",
|
| 93336 |
+
"ββββββββββ β",
|
| 93337 |
"ββββββββββββββ ββ",
|
| 93338 |
+
"ββββββββββββββ β",
|
| 93339 |
+
"βββ ββ",
|
| 93340 |
+
"βββ ββββ",
|
| 93341 |
+
"βββ ββββββββ",
|
| 93342 |
+
"βββ βββββ",
|
| 93343 |
+
"βββ ββββββ",
|
| 93344 |
+
"βββ ββββββββββββ",
|
| 93345 |
"βββ βββββββββββββ",
|
| 93346 |
+
"βββ ββββββββββ",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93347 |
"βββ βββ",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93348 |
"βββ βββββββββ",
|
| 93349 |
+
"βββ βββββββ",
|
| 93350 |
+
"βββ βββββββββββ",
|
| 93351 |
+
"βββ β",
|
| 93352 |
+
"βββββββββ ββ",
|
| 93353 |
+
"βββββββββ ββββ",
|
| 93354 |
+
"βββββββββ βββββ",
|
| 93355 |
+
"βββββββββ ββββββ",
|
| 93356 |
"βββββββββ βββ",
|
| 93357 |
+
"βββββββββ βββββββ",
|
| 93358 |
+
"βββββββββ β",
|
| 93359 |
+
"βββββββ ββ",
|
| 93360 |
+
"βββββββ ββββ",
|
| 93361 |
+
"βββββββ ββββββββ",
|
| 93362 |
"βββββββ βββββ",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93363 |
"βββββββ ββββββ",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93364 |
"βββββββ βββ",
|
| 93365 |
+
"βββββββ βββββββββ",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93366 |
"βββββββ βββββββ",
|
| 93367 |
+
"βββββββ β",
|
| 93368 |
+
"βββββββββββ ββ",
|
| 93369 |
+
"βββββββββββ ββββ",
|
| 93370 |
+
"βββββββββββ βββββ",
|
| 93371 |
"βββββββββββ βββ",
|
| 93372 |
+
"βββββββββββ β",
|
| 93373 |
+
"βββββββββββββββ β",
|
| 93374 |
"β ββ",
|
| 93375 |
+
"β ββββ",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93376 |
"β ββββββββ",
|
| 93377 |
+
"β βββββ",
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93378 |
"β ββββββ",
|
| 93379 |
+
"β ββββββββββββ",
|
| 93380 |
+
"β βββββββββββββ",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93381 |
"β ββββββββββ",
|
| 93382 |
+
"β ββοΏ½οΏ½βββββββββββ",
|
| 93383 |
+
"β βββ",
|
| 93384 |
+
"β βββββββββ",
|
| 93385 |
+
"β βββββββ",
|
| 93386 |
+
"β βββββββββββ",
|
| 93387 |
+
"β βββββββββββββββ",
|
| 93388 |
+
"β β"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93389 |
]
|
| 93390 |
}
|
| 93391 |
}
|
tokenizer_config.json
CHANGED
|
@@ -3,7 +3,7 @@
|
|
| 3 |
"__type": "AddedToken",
|
| 4 |
"content": "<s>",
|
| 5 |
"lstrip": false,
|
| 6 |
-
"normalized":
|
| 7 |
"rstrip": false,
|
| 8 |
"single_word": false
|
| 9 |
},
|
|
@@ -12,21 +12,19 @@
|
|
| 12 |
"__type": "AddedToken",
|
| 13 |
"content": "</s>",
|
| 14 |
"lstrip": false,
|
| 15 |
-
"normalized":
|
| 16 |
"rstrip": false,
|
| 17 |
"single_word": false
|
| 18 |
},
|
| 19 |
-
"legacy": false,
|
| 20 |
"model_max_length": 1000000000000000019884624838656,
|
| 21 |
"pad_token": null,
|
| 22 |
-
"padding_side": "right",
|
| 23 |
"sp_model_kwargs": {},
|
| 24 |
"tokenizer_class": "LlamaTokenizer",
|
| 25 |
"unk_token": {
|
| 26 |
"__type": "AddedToken",
|
| 27 |
"content": "<unk>",
|
| 28 |
"lstrip": false,
|
| 29 |
-
"normalized":
|
| 30 |
"rstrip": false,
|
| 31 |
"single_word": false
|
| 32 |
},
|
|
|
|
| 3 |
"__type": "AddedToken",
|
| 4 |
"content": "<s>",
|
| 5 |
"lstrip": false,
|
| 6 |
+
"normalized": true,
|
| 7 |
"rstrip": false,
|
| 8 |
"single_word": false
|
| 9 |
},
|
|
|
|
| 12 |
"__type": "AddedToken",
|
| 13 |
"content": "</s>",
|
| 14 |
"lstrip": false,
|
| 15 |
+
"normalized": true,
|
| 16 |
"rstrip": false,
|
| 17 |
"single_word": false
|
| 18 |
},
|
|
|
|
| 19 |
"model_max_length": 1000000000000000019884624838656,
|
| 20 |
"pad_token": null,
|
|
|
|
| 21 |
"sp_model_kwargs": {},
|
| 22 |
"tokenizer_class": "LlamaTokenizer",
|
| 23 |
"unk_token": {
|
| 24 |
"__type": "AddedToken",
|
| 25 |
"content": "<unk>",
|
| 26 |
"lstrip": false,
|
| 27 |
+
"normalized": true,
|
| 28 |
"rstrip": false,
|
| 29 |
"single_word": false
|
| 30 |
},
|