danielhanchen commited on
Commit
f97e297
1 Parent(s): cb204e3

Upload tokenizer

Browse files
Files changed (1) hide show
  1. tokenizer_config.json +2 -2
tokenizer_config.json CHANGED
@@ -8005,11 +8005,11 @@
8005
  }
8006
  },
8007
  "bos_token": "<s>",
8008
- "clean_up_tokenization_spaces": true,
8009
  "eos_token": "</s>",
8010
  "model_max_length": 1000000000000000019884624838656,
8011
  "pad_token": "<pad>",
8012
  "padding_side": "left",
8013
- "tokenizer_class": "GPT2Tokenizer",
8014
  "unk_token": "<unk>"
8015
  }
 
8005
  }
8006
  },
8007
  "bos_token": "<s>",
8008
+ "clean_up_tokenization_spaces": false,
8009
  "eos_token": "</s>",
8010
  "model_max_length": 1000000000000000019884624838656,
8011
  "pad_token": "<pad>",
8012
  "padding_side": "left",
8013
+ "tokenizer_class": "PreTrainedTokenizerFast",
8014
  "unk_token": "<unk>"
8015
  }