santoshpandey
commited on
Upload tokenizer
Browse files- added_tokens.json +4 -4
- tokenizer_config.json +4 -4
- vocab.json +2 -4
added_tokens.json
CHANGED
@@ -1,6 +1,6 @@
|
|
1 |
{
|
2 |
-
"</s>":
|
3 |
-
"<s>":
|
4 |
-
"[PAD]":
|
5 |
-
"[UNK]":
|
6 |
}
|
|
|
1 |
{
|
2 |
+
"</s>": 67,
|
3 |
+
"<s>": 66,
|
4 |
+
"[PAD]": 65,
|
5 |
+
"[UNK]": 64
|
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 |
+
"64": {
|
4 |
"content": "[UNK]",
|
5 |
"lstrip": true,
|
6 |
"normalized": false,
|
|
|
8 |
"single_word": false,
|
9 |
"special": false
|
10 |
},
|
11 |
+
"65": {
|
12 |
"content": "[PAD]",
|
13 |
"lstrip": true,
|
14 |
"normalized": false,
|
|
|
16 |
"single_word": false,
|
17 |
"special": false
|
18 |
},
|
19 |
+
"66": {
|
20 |
"content": "<s>",
|
21 |
"lstrip": false,
|
22 |
"normalized": false,
|
|
|
24 |
"single_word": false,
|
25 |
"special": true
|
26 |
},
|
27 |
+
"67": {
|
28 |
"content": "</s>",
|
29 |
"lstrip": false,
|
30 |
"normalized": false,
|
vocab.json
CHANGED
@@ -1,6 +1,6 @@
|
|
1 |
{
|
2 |
-
"[PAD]":
|
3 |
-
"[UNK]":
|
4 |
"ँ": 1,
|
5 |
"ं": 2,
|
6 |
"अ": 3,
|
@@ -52,11 +52,9 @@
|
|
52 |
"ु": 49,
|
53 |
"ू": 50,
|
54 |
"ृ": 51,
|
55 |
-
"े": 52,
|
56 |
"ै": 53,
|
57 |
"ो": 54,
|
58 |
"ौ": 55,
|
59 |
-
"्": 56,
|
60 |
"०": 58,
|
61 |
"१": 59,
|
62 |
"२": 60,
|
|
|
1 |
{
|
2 |
+
"[PAD]": 65,
|
3 |
+
"[UNK]": 64,
|
4 |
"ँ": 1,
|
5 |
"ं": 2,
|
6 |
"अ": 3,
|
|
|
52 |
"ु": 49,
|
53 |
"ू": 50,
|
54 |
"ृ": 51,
|
|
|
55 |
"ै": 53,
|
56 |
"ो": 54,
|
57 |
"ौ": 55,
|
|
|
58 |
"०": 58,
|
59 |
"१": 59,
|
60 |
"२": 60,
|