Commit ·
5c1213e
1
Parent(s): 16df352
Upload tokenizer
Browse files- added_tokens.json +1 -1
- special_tokens_map.json +3 -3
- tokenizer.json +5 -5
added_tokens.json
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
{
|
| 2 |
-
"
|
| 3 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"\n": 32000
|
| 3 |
}
|
special_tokens_map.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
| 1 |
{
|
| 2 |
-
"bos_token": "<
|
| 3 |
"eos_token": "</s>",
|
| 4 |
-
"pad_token": "
|
| 5 |
-
"unk_token": "<
|
| 6 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
"eos_token": "</s>",
|
| 4 |
+
"pad_token": "<unk>",
|
| 5 |
+
"unk_token": "<unk>"
|
| 6 |
}
|
tokenizer.json
CHANGED
|
@@ -9,7 +9,7 @@
|
|
| 9 |
"single_word": false,
|
| 10 |
"lstrip": false,
|
| 11 |
"rstrip": false,
|
| 12 |
-
"normalized":
|
| 13 |
"special": true
|
| 14 |
},
|
| 15 |
{
|
|
@@ -18,7 +18,7 @@
|
|
| 18 |
"single_word": false,
|
| 19 |
"lstrip": false,
|
| 20 |
"rstrip": false,
|
| 21 |
-
"normalized":
|
| 22 |
"special": true
|
| 23 |
},
|
| 24 |
{
|
|
@@ -32,12 +32,12 @@
|
|
| 32 |
},
|
| 33 |
{
|
| 34 |
"id": 32000,
|
| 35 |
-
"content": "
|
| 36 |
"single_word": false,
|
| 37 |
"lstrip": false,
|
| 38 |
"rstrip": false,
|
| 39 |
-
"normalized":
|
| 40 |
-
"special":
|
| 41 |
}
|
| 42 |
],
|
| 43 |
"normalizer": {
|
|
|
|
| 9 |
"single_word": false,
|
| 10 |
"lstrip": false,
|
| 11 |
"rstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
"special": true
|
| 14 |
},
|
| 15 |
{
|
|
|
|
| 18 |
"single_word": false,
|
| 19 |
"lstrip": false,
|
| 20 |
"rstrip": false,
|
| 21 |
+
"normalized": false,
|
| 22 |
"special": true
|
| 23 |
},
|
| 24 |
{
|
|
|
|
| 32 |
},
|
| 33 |
{
|
| 34 |
"id": 32000,
|
| 35 |
+
"content": "\n",
|
| 36 |
"single_word": false,
|
| 37 |
"lstrip": false,
|
| 38 |
"rstrip": false,
|
| 39 |
+
"normalized": true,
|
| 40 |
+
"special": false
|
| 41 |
}
|
| 42 |
],
|
| 43 |
"normalizer": {
|