pansophic commited on
Commit
b09a607
1 Parent(s): 6b65196

Fix tokenizer_config.json EOS token (#3)

Browse files

- Use `<|im_end|>` as EOS token in `tokenizer_config.json` (9e2f067c57c073064f34f85df854b712ffd0ba5a)

Files changed (1) hide show
  1. tokenizer_config.json +1 -1
tokenizer_config.json CHANGED
@@ -2,7 +2,7 @@
2
  "add_prefix_space": false,
3
  "bos_token": "<|endoftext|>",
4
  "clean_up_tokenization_spaces": true,
5
- "eos_token": "<|endoftext|>",
6
  "model_max_length": 1000000000000000019884624838656,
7
  "tokenizer_class": "GPTNeoXTokenizer",
8
  "unk_token": "<|endoftext|>"
 
2
  "add_prefix_space": false,
3
  "bos_token": "<|endoftext|>",
4
  "clean_up_tokenization_spaces": true,
5
+ "eos_token": "<|im_end|>",
6
  "model_max_length": 1000000000000000019884624838656,
7
  "tokenizer_class": "GPTNeoXTokenizer",
8
  "unk_token": "<|endoftext|>"