WesleySantos
commited on
Commit
β’
a22b18f
1
Parent(s):
a8fe092
Upload tokenizer (#5)
Browse files- Upload tokenizer (f0de99376a538b154f3196901e92b4ba0843422e)
- special_tokens_map.json +3 -4
- tokenizer.json +99 -108
- tokenizer_config.json +4 -7
special_tokens_map.json
CHANGED
@@ -2,22 +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 |
-
"pad_token": "β***",
|
17 |
"unk_token": {
|
18 |
"content": "<unk>",
|
19 |
"lstrip": false,
|
20 |
-
"normalized":
|
21 |
"rstrip": false,
|
22 |
"single_word": false
|
23 |
}
|
|
|
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 |
{
|
@@ -27,16 +27,7 @@
|
|
27 |
"single_word": false,
|
28 |
"lstrip": false,
|
29 |
"rstrip": false,
|
30 |
-
"normalized":
|
31 |
-
"special": true
|
32 |
-
},
|
33 |
-
{
|
34 |
-
"id": 18610,
|
35 |
-
"content": "β***",
|
36 |
-
"single_word": false,
|
37 |
-
"lstrip": false,
|
38 |
-
"rstrip": false,
|
39 |
-
"normalized": false,
|
40 |
"special": true
|
41 |
}
|
42 |
],
|
@@ -93275,126 +93266,126 @@
|
|
93275 |
"βlivre s",
|
93276 |
"lu b",
|
93277 |
"l ub",
|
93278 |
-
"β β",
|
93279 |
"ββ ββ",
|
93280 |
-
"
|
93281 |
-
"
|
|
|
93282 |
"ββ ββββββ",
|
93283 |
-
"
|
93284 |
-
"
|
93285 |
-
"
|
93286 |
-
"βββ βββββ",
|
93287 |
-
"βββββββ β",
|
93288 |
-
"β βββββββ",
|
93289 |
-
"ββ βββ",
|
93290 |
-
"ββββ β",
|
93291 |
-
"βββ ββ",
|
93292 |
-
"β ββββ",
|
93293 |
"ββ ββββββββββββββ",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
93294 |
"ββββ ββββββββββββ",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
93295 |
"ββββββββ ββββββββ",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
93296 |
"βββββ βββββββββββ",
|
|
|
|
|
|
|
|
|
|
|
|
|
93297 |
"ββββββ ββββββββββ",
|
|
|
|
|
|
|
|
|
|
|
93298 |
"ββββββββββββ ββββ",
|
|
|
|
|
|
|
93299 |
"βββββββββββββ βββ",
|
|
|
|
|
|
|
|
|
93300 |
"ββββββββββ ββββββ",
|
|
|
|
|
93301 |
"ββββββββββββββ ββ",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
93302 |
"βββ βββββββββββββ",
|
93303 |
-
"
|
93304 |
-
"βββββββ βββββββββ",
|
93305 |
-
"βββββββββββ βββββ",
|
93306 |
-
"βββββββββββββββ β",
|
93307 |
-
"β βββββββββββββββ",
|
93308 |
-
"ββ ββββ",
|
93309 |
-
"ββββ ββ",
|
93310 |
-
"βββββ β",
|
93311 |
"βββ βββ",
|
93312 |
-
"β βββββ",
|
93313 |
-
"ββ ββββββββββ",
|
93314 |
-
"ββββ ββββββββ",
|
93315 |
-
"ββββββββ ββββ",
|
93316 |
-
"βββββ βββββββ",
|
93317 |
-
"ββββββ ββββββ",
|
93318 |
-
"ββββββββββ ββ",
|
93319 |
"βββ βββββββββ",
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
93320 |
"βββββββββ βββ",
|
|
|
|
|
|
|
|
|
|
|
93321 |
"βββββββ βββββ",
|
93322 |
-
"βββββββββββ β",
|
93323 |
-
"β βββββββββββ",
|
93324 |
-
"ββ βββββββββββ",
|
93325 |
-
"ββββ βββββββββ",
|
93326 |
-
"ββββββββ βββββ",
|
93327 |
-
"βββββ ββββββββ",
|
93328 |
-
"ββββββ βββββββ",
|
93329 |
-
"ββββββββββββ β",
|
93330 |
-
"ββββββββββ βββ",
|
93331 |
-
"βββ ββββββββββ",
|
93332 |
-
"βββββββββ ββββ",
|
93333 |
"βββββββ ββββββ",
|
93334 |
-
"βββββββββββ ββ",
|
93335 |
-
"β ββββββββββββ",
|
93336 |
-
"ββ ββββββββ",
|
93337 |
-
"ββββ ββββββ",
|
93338 |
-
"ββββββββ ββ",
|
93339 |
-
"βββββ βββββ",
|
93340 |
-
"ββββββ ββββ",
|
93341 |
-
"βββ βββββββ",
|
93342 |
-
"βββββββββ β",
|
93343 |
"βββββββ βββ",
|
93344 |
-
"
|
93345 |
-
"ββ ββββββββββββ",
|
93346 |
-
"ββββ ββββββββββ",
|
93347 |
-
"ββββββββ ββββββ",
|
93348 |
-
"βββββ βββββββββ",
|
93349 |
-
"ββββββ ββββββββ",
|
93350 |
-
"ββββββββββββ ββ",
|
93351 |
-
"βββββββββββββ β",
|
93352 |
-
"ββββββββββ ββββ",
|
93353 |
-
"βββ βββββββββββ",
|
93354 |
-
"βββββββββ βββββ",
|
93355 |
"βββββββ βββββββ",
|
|
|
|
|
|
|
|
|
93356 |
"βββββββββββ βββ",
|
93357 |
-
"β
|
93358 |
-
"
|
93359 |
"β ββ",
|
93360 |
-
"
|
93361 |
-
"ββββ βββββ",
|
93362 |
-
"ββββββββ β",
|
93363 |
-
"βββββ ββββ",
|
93364 |
-
"ββββββ βββ",
|
93365 |
-
"βββ ββββββ",
|
93366 |
-
"βββββββ ββ",
|
93367 |
"β ββββββββ",
|
93368 |
-
"
|
93369 |
-
"ββββ βββ",
|
93370 |
-
"βββββ ββ",
|
93371 |
-
"ββββββ β",
|
93372 |
-
"βββ ββββ",
|
93373 |
"β ββββββ",
|
93374 |
-
"
|
93375 |
-
"
|
93376 |
-
"ββββββββ βββ",
|
93377 |
-
"βββββ ββββββ",
|
93378 |
-
"ββββββ βββββ",
|
93379 |
-
"ββββββββββ β",
|
93380 |
-
"βββ ββββββββ",
|
93381 |
-
"βββββββββ ββ",
|
93382 |
-
"βββββββ ββββ",
|
93383 |
"β ββββββββββ",
|
93384 |
-
"
|
93385 |
-
"
|
93386 |
-
"
|
93387 |
-
"
|
93388 |
-
"
|
93389 |
-
"
|
93390 |
-
"
|
93391 |
-
"ββββββββββ βββββ",
|
93392 |
-
"ββββββββββββββ β",
|
93393 |
-
"βββ ββββββββββββ",
|
93394 |
-
"βββββββββ ββββββ",
|
93395 |
-
"βββββββ ββββββββ",
|
93396 |
-
"βββββββββββ ββββ",
|
93397 |
-
"β ββββββββββββββ"
|
93398 |
]
|
93399 |
}
|
93400 |
}
|
|
|
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 ββββββββββββββ",
|
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 |
},
|
@@ -12,23 +12,20 @@
|
|
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 |
-
}
|
33 |
-
"use_default_system_prompt": true
|
34 |
}
|
|
|
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 |
+
}
|
|
|
31 |
}
|