chanicpanic
commited on
Fix tokenizer special tokens
Browse files- special_tokens_map.json +5 -7
- tokenizer_config.json +6 -2
special_tokens_map.json
CHANGED
@@ -1,9 +1,7 @@
|
|
1 |
{
|
2 |
-
"
|
3 |
-
|
4 |
-
|
5 |
-
|
6 |
-
|
7 |
-
"single_word": false
|
8 |
-
}
|
9 |
}
|
|
|
1 |
{
|
2 |
+
"bos_token": "<s>",
|
3 |
+
"eos_token": "</s>",
|
4 |
+
"mask_token": "<mask>",
|
5 |
+
"pad_token": "<pad>",
|
6 |
+
"unk_token": "<unk>"
|
|
|
|
|
7 |
}
|
tokenizer_config.json
CHANGED
@@ -41,12 +41,16 @@
|
|
41 |
"special": true
|
42 |
}
|
43 |
},
|
44 |
-
"
|
|
|
|
|
|
|
45 |
"model_input_names": [
|
46 |
"labels",
|
47 |
"pixel_values"
|
48 |
],
|
49 |
"model_max_length": 1000000000000000019884624838656,
|
50 |
"pad_token": "<pad>",
|
51 |
-
"tokenizer_class": "PreTrainedTokenizerFast"
|
|
|
52 |
}
|
|
|
41 |
"special": true
|
42 |
}
|
43 |
},
|
44 |
+
"bos_token": "<s>",
|
45 |
+
"clean_up_tokenization_spaces": false,
|
46 |
+
"eos_token": "</s>",
|
47 |
+
"mask_token": "<mask>",
|
48 |
"model_input_names": [
|
49 |
"labels",
|
50 |
"pixel_values"
|
51 |
],
|
52 |
"model_max_length": 1000000000000000019884624838656,
|
53 |
"pad_token": "<pad>",
|
54 |
+
"tokenizer_class": "PreTrainedTokenizerFast",
|
55 |
+
"unk_token": "<unk>"
|
56 |
}
|