Upload tokenizer
Browse files- special_tokens_map.json +4 -1
- tokenizer_config.json +4 -1
special_tokens_map.json
CHANGED
@@ -1,3 +1,6 @@
|
|
1 |
{
|
2 |
-
"
|
|
|
|
|
|
|
3 |
}
|
|
|
1 |
{
|
2 |
+
"bos_token": "<|bos|>",
|
3 |
+
"eos_token": "<|eos|>",
|
4 |
+
"pad_token": "<|pad|>",
|
5 |
+
"unk_token": "<|unknown|>"
|
6 |
}
|
tokenizer_config.json
CHANGED
@@ -353,12 +353,15 @@
|
|
353 |
"special": true
|
354 |
}
|
355 |
},
|
|
|
356 |
"clean_up_tokenization_spaces": true,
|
|
|
357 |
"max_length": null,
|
358 |
"model_max_length": 1000000000000000019884624838656,
|
359 |
"pad_to_multiple_of": null,
|
360 |
"pad_token": "<|pad|>",
|
361 |
"pad_token_type_id": 0,
|
362 |
"padding_side": "right",
|
363 |
-
"tokenizer_class": "PreTrainedTokenizerFast"
|
|
|
364 |
}
|
|
|
353 |
"special": true
|
354 |
}
|
355 |
},
|
356 |
+
"bos_token": "<|bos|>",
|
357 |
"clean_up_tokenization_spaces": true,
|
358 |
+
"eos_token": "<|eos|>",
|
359 |
"max_length": null,
|
360 |
"model_max_length": 1000000000000000019884624838656,
|
361 |
"pad_to_multiple_of": null,
|
362 |
"pad_token": "<|pad|>",
|
363 |
"pad_token_type_id": 0,
|
364 |
"padding_side": "right",
|
365 |
+
"tokenizer_class": "PreTrainedTokenizerFast",
|
366 |
+
"unk_token": "<|unknown|>"
|
367 |
}
|