nielsr HF staff commited on
Commit
470ab47
1 Parent(s): 6896b7a

Upload fast tokenizer files

Browse files
special_tokens_map.json CHANGED
@@ -1204,5 +1204,6 @@
1204
  ],
1205
  "eos_token": "</s>",
1206
  "pad_token": "<pad>",
 
1207
  "unk_token": "<unk>"
1208
  }
 
1204
  ],
1205
  "eos_token": "</s>",
1206
  "pad_token": "<pad>",
1207
+ "sep_token": "</s>",
1208
  "unk_token": "<unk>"
1209
  }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json CHANGED
@@ -1204,8 +1204,9 @@
1204
  ],
1205
  "eos_token": "</s>",
1206
  "extra_ids": 100,
 
1207
  "loc_extra_ids": 501,
1208
- "model_max_length": 1000000000000000019884624838656,
1209
  "only_label_first_subword": true,
1210
  "other_extra_ids": 200,
1211
  "pad_token": "<pad>",
@@ -1217,6 +1218,7 @@
1217
  ],
1218
  "pad_token_label": -100,
1219
  "processor_class": "UdopProcessor",
 
1220
  "sep_token_box": [
1221
  1000,
1222
  1000,
 
1204
  ],
1205
  "eos_token": "</s>",
1206
  "extra_ids": 100,
1207
+ "from_slow": true,
1208
  "loc_extra_ids": 501,
1209
+ "model_max_length": 512,
1210
  "only_label_first_subword": true,
1211
  "other_extra_ids": 200,
1212
  "pad_token": "<pad>",
 
1218
  ],
1219
  "pad_token_label": -100,
1220
  "processor_class": "UdopProcessor",
1221
+ "sep_token": "</s>",
1222
  "sep_token_box": [
1223
  1000,
1224
  1000,