Upload tokenizer
Browse files- special_tokens_map.json +1 -5
- tokenizer.py +1 -7
- tokenizer_config.json +1 -29
special_tokens_map.json
CHANGED
|
@@ -1,5 +1 @@
|
|
| 1 |
-
{
|
| 2 |
-
"cls_token": "<cls>",
|
| 3 |
-
"mask_token": "<mask>",
|
| 4 |
-
"pad_token": "<pad>"
|
| 5 |
-
}
|
|
|
|
| 1 |
+
{}
|
|
|
|
|
|
|
|
|
|
|
|
tokenizer.py
CHANGED
|
@@ -30,7 +30,6 @@ class BinnedOmicTokenizer(PreTrainedTokenizer):
|
|
| 30 |
|
| 31 |
ids_to_tokens = {i: tok for tok, i in vocab.items()}
|
| 32 |
|
| 33 |
-
# Save vocab attributes before superclass init
|
| 34 |
self.vocab = vocab
|
| 35 |
self.ids_to_tokens = ids_to_tokens
|
| 36 |
|
|
@@ -49,12 +48,7 @@ class BinnedOmicTokenizer(PreTrainedTokenizer):
|
|
| 49 |
self.mask_token = "<mask>"
|
| 50 |
self.cls_token = "<cls>"
|
| 51 |
|
| 52 |
-
super().__init__(
|
| 53 |
-
pad_token=self.pad_token,
|
| 54 |
-
mask_token=self.mask_token,
|
| 55 |
-
cls_token=self.cls_token,
|
| 56 |
-
**kwargs,
|
| 57 |
-
)
|
| 58 |
|
| 59 |
def _convert_token_to_id(self, token: str) -> int:
|
| 60 |
return self.vocab.get(token, self.vocab[self.unk_token])
|
|
|
|
| 30 |
|
| 31 |
ids_to_tokens = {i: tok for tok, i in vocab.items()}
|
| 32 |
|
|
|
|
| 33 |
self.vocab = vocab
|
| 34 |
self.ids_to_tokens = ids_to_tokens
|
| 35 |
|
|
|
|
| 48 |
self.mask_token = "<mask>"
|
| 49 |
self.cls_token = "<cls>"
|
| 50 |
|
| 51 |
+
super().__init__(**kwargs)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 52 |
|
| 53 |
def _convert_token_to_id(self, token: str) -> int:
|
| 54 |
return self.vocab.get(token, self.vocab[self.unk_token])
|
tokenizer_config.json
CHANGED
|
@@ -1,30 +1,5 @@
|
|
| 1 |
{
|
| 2 |
-
"added_tokens_decoder": {
|
| 3 |
-
"64": {
|
| 4 |
-
"content": "<pad>",
|
| 5 |
-
"lstrip": false,
|
| 6 |
-
"normalized": false,
|
| 7 |
-
"rstrip": false,
|
| 8 |
-
"single_word": false,
|
| 9 |
-
"special": true
|
| 10 |
-
},
|
| 11 |
-
"65": {
|
| 12 |
-
"content": "<mask>",
|
| 13 |
-
"lstrip": false,
|
| 14 |
-
"normalized": false,
|
| 15 |
-
"rstrip": false,
|
| 16 |
-
"single_word": false,
|
| 17 |
-
"special": true
|
| 18 |
-
},
|
| 19 |
-
"66": {
|
| 20 |
-
"content": "<cls>",
|
| 21 |
-
"lstrip": false,
|
| 22 |
-
"normalized": false,
|
| 23 |
-
"rstrip": false,
|
| 24 |
-
"single_word": false,
|
| 25 |
-
"special": true
|
| 26 |
-
}
|
| 27 |
-
},
|
| 28 |
"auto_map": {
|
| 29 |
"AutoTokenizer": [
|
| 30 |
"tokenizer.BinnedOmicTokenizer",
|
|
@@ -32,9 +7,6 @@
|
|
| 32 |
]
|
| 33 |
},
|
| 34 |
"clean_up_tokenization_spaces": true,
|
| 35 |
-
"cls_token": "<cls>",
|
| 36 |
-
"mask_token": "<mask>",
|
| 37 |
"model_max_length": 1000000000000000019884624838656,
|
| 38 |
-
"pad_token": "<pad>",
|
| 39 |
"tokenizer_class": "BinnedOmicTokenizer"
|
| 40 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"added_tokens_decoder": {},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
"auto_map": {
|
| 4 |
"AutoTokenizer": [
|
| 5 |
"tokenizer.BinnedOmicTokenizer",
|
|
|
|
| 7 |
]
|
| 8 |
},
|
| 9 |
"clean_up_tokenization_spaces": true,
|
|
|
|
|
|
|
| 10 |
"model_max_length": 1000000000000000019884624838656,
|
|
|
|
| 11 |
"tokenizer_class": "BinnedOmicTokenizer"
|
| 12 |
}
|