khaoulaoub
commited on
Commit
•
a189b42
1
Parent(s):
93d6b1b
Upload tokenizer
Browse files- added_tokens.json +4 -4
- tokenizer_config.json +4 -4
- vocab.json +35 -33
added_tokens.json
CHANGED
@@ -1,6 +1,6 @@
|
|
1 |
{
|
2 |
-
"</s>":
|
3 |
-
"<s>":
|
4 |
-
"[PAD]":
|
5 |
-
"[UNK]":
|
6 |
}
|
|
|
1 |
{
|
2 |
+
"</s>": 36,
|
3 |
+
"<s>": 35,
|
4 |
+
"[PAD]": 34,
|
5 |
+
"[UNK]": 33
|
6 |
}
|
tokenizer_config.json
CHANGED
@@ -1,6 +1,6 @@
|
|
1 |
{
|
2 |
"added_tokens_decoder": {
|
3 |
-
"
|
4 |
"content": "[UNK]",
|
5 |
"lstrip": true,
|
6 |
"normalized": false,
|
@@ -8,7 +8,7 @@
|
|
8 |
"single_word": false,
|
9 |
"special": false
|
10 |
},
|
11 |
-
"
|
12 |
"content": "[PAD]",
|
13 |
"lstrip": true,
|
14 |
"normalized": false,
|
@@ -16,7 +16,7 @@
|
|
16 |
"single_word": false,
|
17 |
"special": false
|
18 |
},
|
19 |
-
"
|
20 |
"content": "<s>",
|
21 |
"lstrip": false,
|
22 |
"normalized": false,
|
@@ -24,7 +24,7 @@
|
|
24 |
"single_word": false,
|
25 |
"special": true
|
26 |
},
|
27 |
-
"
|
28 |
"content": "</s>",
|
29 |
"lstrip": false,
|
30 |
"normalized": false,
|
|
|
1 |
{
|
2 |
"added_tokens_decoder": {
|
3 |
+
"33": {
|
4 |
"content": "[UNK]",
|
5 |
"lstrip": true,
|
6 |
"normalized": false,
|
|
|
8 |
"single_word": false,
|
9 |
"special": false
|
10 |
},
|
11 |
+
"34": {
|
12 |
"content": "[PAD]",
|
13 |
"lstrip": true,
|
14 |
"normalized": false,
|
|
|
16 |
"single_word": false,
|
17 |
"special": false
|
18 |
},
|
19 |
+
"35": {
|
20 |
"content": "<s>",
|
21 |
"lstrip": false,
|
22 |
"normalized": false,
|
|
|
24 |
"single_word": false,
|
25 |
"special": true
|
26 |
},
|
27 |
+
"36": {
|
28 |
"content": "</s>",
|
29 |
"lstrip": false,
|
30 |
"normalized": false,
|
vocab.json
CHANGED
@@ -1,35 +1,37 @@
|
|
1 |
{
|
2 |
-
"[PAD]":
|
3 |
-
"[UNK]":
|
4 |
-
"|":
|
5 |
-
"ء":
|
6 |
-
"ا":
|
7 |
-
"ب":
|
8 |
-
"ة":
|
9 |
-
"ت":
|
10 |
-
"ث":
|
11 |
-
"ج":
|
12 |
-
"ح":
|
13 |
-
"خ":
|
14 |
-
"د":
|
15 |
-
"ذ":
|
16 |
-
"ر":
|
17 |
-
"ز":
|
18 |
-
"س":
|
19 |
-
"ش":
|
20 |
-
"ص":
|
21 |
-
"ض":
|
22 |
-
"ط":
|
23 |
-
"ظ":
|
24 |
-
"ع":
|
25 |
-
"غ":
|
26 |
-
"
|
27 |
-
"
|
28 |
-
"
|
29 |
-
"
|
30 |
-
"
|
31 |
-
"
|
32 |
-
"
|
33 |
-
"
|
34 |
-
"
|
|
|
|
|
35 |
}
|
|
|
1 |
{
|
2 |
+
"[PAD]": 34,
|
3 |
+
"[UNK]": 33,
|
4 |
+
"|": 27,
|
5 |
+
"ء": 23,
|
6 |
+
"ا": 25,
|
7 |
+
"ب": 32,
|
8 |
+
"ة": 1,
|
9 |
+
"ت": 31,
|
10 |
+
"ث": 11,
|
11 |
+
"ج": 17,
|
12 |
+
"ح": 13,
|
13 |
+
"خ": 29,
|
14 |
+
"د": 8,
|
15 |
+
"ذ": 28,
|
16 |
+
"ر": 0,
|
17 |
+
"ز": 7,
|
18 |
+
"س": 15,
|
19 |
+
"ش": 9,
|
20 |
+
"ص": 26,
|
21 |
+
"ض": 19,
|
22 |
+
"ط": 4,
|
23 |
+
"ظ": 21,
|
24 |
+
"ع": 5,
|
25 |
+
"غ": 16,
|
26 |
+
"ـ": 22,
|
27 |
+
"ف": 24,
|
28 |
+
"ق": 18,
|
29 |
+
"ك": 14,
|
30 |
+
"ل": 3,
|
31 |
+
"م": 30,
|
32 |
+
"ن": 6,
|
33 |
+
"ه": 2,
|
34 |
+
"و": 10,
|
35 |
+
"ى": 12,
|
36 |
+
"ي": 20
|
37 |
}
|