danielhanchen
commited on
Commit
β’
f36ad61
1
Parent(s):
1f7efee
Upload tokenizer
Browse files- special_tokens_map.json +7 -1
- tokenizer.json +1 -0
- tokenizer_config.json +2 -2
special_tokens_map.json
CHANGED
@@ -19,7 +19,13 @@
|
|
19 |
"rstrip": false,
|
20 |
"single_word": false
|
21 |
},
|
22 |
-
"pad_token":
|
|
|
|
|
|
|
|
|
|
|
|
|
23 |
"unk_token": {
|
24 |
"content": "<unk>",
|
25 |
"lstrip": false,
|
|
|
19 |
"rstrip": false,
|
20 |
"single_word": false
|
21 |
},
|
22 |
+
"pad_token": {
|
23 |
+
"content": "<unk>",
|
24 |
+
"lstrip": false,
|
25 |
+
"normalized": false,
|
26 |
+
"rstrip": false,
|
27 |
+
"single_word": false
|
28 |
+
},
|
29 |
"unk_token": {
|
30 |
"content": "<unk>",
|
31 |
"lstrip": false,
|
tokenizer.json
CHANGED
@@ -170,6 +170,7 @@
|
|
170 |
"end_of_word_suffix": null,
|
171 |
"fuse_unk": true,
|
172 |
"byte_fallback": true,
|
|
|
173 |
"vocab": {
|
174 |
"<unk>": 0,
|
175 |
"<s>": 1,
|
|
|
170 |
"end_of_word_suffix": null,
|
171 |
"fuse_unk": true,
|
172 |
"byte_fallback": true,
|
173 |
+
"ignore_merges": false,
|
174 |
"vocab": {
|
175 |
"<unk>": 0,
|
176 |
"<s>": 1,
|
tokenizer_config.json
CHANGED
@@ -72,9 +72,9 @@
|
|
72 |
"fill_token": "<FILL_ME>",
|
73 |
"legacy": null,
|
74 |
"middle_token": "β<MID>",
|
75 |
-
"model_max_length":
|
76 |
"pad_token": "<unk>",
|
77 |
-
"padding_side": "
|
78 |
"prefix_token": "β<PRE>",
|
79 |
"sp_model_kwargs": {},
|
80 |
"suffix_token": "β<SUF>",
|
|
|
72 |
"fill_token": "<FILL_ME>",
|
73 |
"legacy": null,
|
74 |
"middle_token": "β<MID>",
|
75 |
+
"model_max_length": 1000000000000000019884624838656,
|
76 |
"pad_token": "<unk>",
|
77 |
+
"padding_side": "left",
|
78 |
"prefix_token": "β<PRE>",
|
79 |
"sp_model_kwargs": {},
|
80 |
"suffix_token": "β<SUF>",
|