Upload tokenizer
Browse files- tokenizer_config.json +2 -1
tokenizer_config.json
CHANGED
|
@@ -71,9 +71,10 @@
|
|
| 71 |
"<unk>NOTUSED"
|
| 72 |
],
|
| 73 |
"bos_token": "<s>",
|
| 74 |
-
"clean_up_tokenization_spaces":
|
| 75 |
"cls_token": "<s>",
|
| 76 |
"eos_token": "</s>",
|
|
|
|
| 77 |
"mask_token": "<mask>",
|
| 78 |
"model_max_length": 512,
|
| 79 |
"pad_token": "<pad>",
|
|
|
|
| 71 |
"<unk>NOTUSED"
|
| 72 |
],
|
| 73 |
"bos_token": "<s>",
|
| 74 |
+
"clean_up_tokenization_spaces": false,
|
| 75 |
"cls_token": "<s>",
|
| 76 |
"eos_token": "</s>",
|
| 77 |
+
"extra_special_tokens": {},
|
| 78 |
"mask_token": "<mask>",
|
| 79 |
"model_max_length": 512,
|
| 80 |
"pad_token": "<pad>",
|