Commit
•
f14e2da
1
Parent(s):
2f5a1cd
Upload tokenizer.json
Browse files- tokenizer.json +49 -0
tokenizer.json
ADDED
|
@@ -0,0 +1,49 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"version": "1.0",
|
| 3 |
+
"truncation": null,
|
| 4 |
+
"padding": null,
|
| 5 |
+
"added_tokens": [
|
| 6 |
+
{
|
| 7 |
+
"id": 0,
|
| 8 |
+
"content": "hello",
|
| 9 |
+
"single_word": false,
|
| 10 |
+
"lstrip": false,
|
| 11 |
+
"rstrip": false,
|
| 12 |
+
"normalized": true,
|
| 13 |
+
"special": false
|
| 14 |
+
}
|
| 15 |
+
],
|
| 16 |
+
"normalizer": {
|
| 17 |
+
"type": "NFC"
|
| 18 |
+
},
|
| 19 |
+
"pre_tokenizer": {
|
| 20 |
+
"type": "ByteLevel",
|
| 21 |
+
"add_prefix_space": false,
|
| 22 |
+
"trim_offsets": true,
|
| 23 |
+
"use_regex": true
|
| 24 |
+
},
|
| 25 |
+
"post_processor": {
|
| 26 |
+
"type": "ByteLevel",
|
| 27 |
+
"add_prefix_space": false,
|
| 28 |
+
"trim_offsets": true,
|
| 29 |
+
"use_regex": true
|
| 30 |
+
},
|
| 31 |
+
"decoder": {
|
| 32 |
+
"type": "ByteLevel",
|
| 33 |
+
"add_prefix_space": false,
|
| 34 |
+
"trim_offsets": true,
|
| 35 |
+
"use_regex": true
|
| 36 |
+
},
|
| 37 |
+
"model": {
|
| 38 |
+
"type": "BPE",
|
| 39 |
+
"dropout": null,
|
| 40 |
+
"unk_token": null,
|
| 41 |
+
"continuing_subword_prefix": null,
|
| 42 |
+
"end_of_word_suffix": null,
|
| 43 |
+
"fuse_unk": false,
|
| 44 |
+
"vocab": {
|
| 45 |
+
"hello": 0
|
| 46 |
+
},
|
| 47 |
+
"merges": []
|
| 48 |
+
}
|
| 49 |
+
}
|