Upload folder using huggingface_hub
Browse files- .gitattributes +10 -0
- bn_101_model.pth +3 -0
- bn_scripts.json +22 -0
- bn_words_a4b.json +3 -0
- gom_122_model.pth +3 -0
- gom_scripts.json +24 -0
- gom_words_subset.json +0 -0
- gu_101_model.pth +3 -0
- gu_scripts.json +26 -0
- gu_words_a4b.json +3 -0
- hi_111_model.pth +3 -0
- hi_scripts.json +25 -0
- hi_words_a4b.json +3 -0
- kn_101_model.pth +3 -0
- kn_scripts.json +20 -0
- kn_words_a4b.json +3 -0
- mai_122_model.pth +3 -0
- mai_scripts.json +25 -0
- mai_words_subset.json +0 -0
- ml_101_model.pth +3 -0
- ml_scripts.json +25 -0
- ml_words_a4b.json +3 -0
- mr_102_model.pth +3 -0
- mr_scripts.json +25 -0
- mr_words_a4b.json +3 -0
- pa_101_model.pth +3 -0
- pa_scripts.json +19 -0
- pa_words_a4b.json +0 -0
- sd_101_model.pth +3 -0
- sd_scripts.json +48 -0
- sd_words_ccset.json +0 -0
- si_101_model.pth +3 -0
- si_scripts.json +29 -0
- si_words_ccset.json +3 -0
- ta_101_model.pth +3 -0
- ta_scripts.json +19 -0
- ta_words_a4b.json +3 -0
- te_101_model.pth +3 -0
- te_scripts.json +21 -0
- te_words_a4b.json +3 -0
- ur_101_model.pth +3 -0
- ur_scripts.json +48 -0
- ur_words_ccset.json +3 -0
.gitattributes
CHANGED
@@ -33,3 +33,13 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
36 |
+
bn_words_a4b.json filter=lfs diff=lfs merge=lfs -text
|
37 |
+
gu_words_a4b.json filter=lfs diff=lfs merge=lfs -text
|
38 |
+
hi_words_a4b.json filter=lfs diff=lfs merge=lfs -text
|
39 |
+
kn_words_a4b.json filter=lfs diff=lfs merge=lfs -text
|
40 |
+
ml_words_a4b.json filter=lfs diff=lfs merge=lfs -text
|
41 |
+
mr_words_a4b.json filter=lfs diff=lfs merge=lfs -text
|
42 |
+
si_words_ccset.json filter=lfs diff=lfs merge=lfs -text
|
43 |
+
ta_words_a4b.json filter=lfs diff=lfs merge=lfs -text
|
44 |
+
te_words_a4b.json filter=lfs diff=lfs merge=lfs -text
|
45 |
+
ur_words_ccset.json filter=lfs diff=lfs merge=lfs -text
|
bn_101_model.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:01461368436fd35f600a9d4bdd2bff04056bcbc099cd59b2f6cbf0d7212a1311
|
3 |
+
size 40804622
|
bn_scripts.json
ADDED
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"WARNING" : " !!! Do not modify the Order of Glyph List !!!",
|
3 |
+
"UNICODE" : {"name": "bengali", "begin":2432, "end":2559},
|
4 |
+
"LANGUAGE": "bengali",
|
5 |
+
|
6 |
+
"glyphs" : [
|
7 |
+
"ঁ", "ং", "ঃ", "অ", "আ", "ই", "ঈ", "উ", "ঊ", "ঋ", "ঌ", "এ", "ঐ",
|
8 |
+
"ও", "ঔ", "ক", "খ", "গ", "ঘ", "ঙ", "চ", "ছ", "জ", "ঝ", "ঞ",
|
9 |
+
"ট", "ঠ", "ড", "ঢ", "ণ", "ত", "থ", "দ", "ধ",
|
10 |
+
"ন", "\u09a9", "প", "ফ", "ব", "ভ", "ম", "য", "র", "ল",
|
11 |
+
"শ", "ষ", "স", "হ", "়", "ঽ", "া", "ি", "ী", "ু", "ূ", "ৃ", "ৄ",
|
12 |
+
"ে", "ৈ", "ো", "ৌ", "্", "ৎ", "ৗ", "\u09db", "ড়", "ঢ়", "য়", "ৠ", "ৡ",
|
13 |
+
"ৢ", "ৣ", "ৰ", "ৱ"
|
14 |
+
|
15 |
+
],
|
16 |
+
|
17 |
+
"numsym_map" : {
|
18 |
+
"0" : ["০"], "1" : ["১"], "2" : ["২"], "3" : ["৩"], "4" : ["৪"],
|
19 |
+
"5" : ["৫"], "6" : ["৬"], "7" : ["৭"], "8" : ["৮"], "9" : ["৯"]
|
20 |
+
}
|
21 |
+
|
22 |
+
}
|
bn_words_a4b.json
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:97438d7712fbfaca4ef56ad20cc7fb7f76e2261d70ac76eea49675f15f73e834
|
3 |
+
size 18431869
|
gom_122_model.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8639ed634b882fd7e0089f228870239aa1a20b4b829f36802760fd6cc09f721c
|
3 |
+
size 40847894
|
gom_scripts.json
ADDED
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"WARNING" : " !!! Do not modify the Order of Glyph List !!!",
|
3 |
+
"UNICODE" : {"name": "devanagari", "begin":2304, "end":2431},
|
4 |
+
"LANGUAGE": "konkani_goan",
|
5 |
+
"glyphs" : [
|
6 |
+
|
7 |
+
"ऄ", "अ", "आ", "इ", "ई", "उ", "ऊ","ऍ", "ऎ", "ए", "ऐ",
|
8 |
+
"ऑ", "ऒ", "ओ", "औ","ऋ","ॠ","ऌ","ॡ","ॲ", "ॐ",
|
9 |
+
"क", "ख", "ग", "घ", "ङ", "च", "छ", "ज", "झ", "ञ", "ट", "ठ", "ड", "ढ", "ण",
|
10 |
+
"त", "थ", "द", "ध", "न", "ऩ", "प", "फ", "ब", "भ", "म", "य", "र", "ऱ", "ल",
|
11 |
+
"ळ", "ऴ", "व", "श", "ष", "स", "ह", "क़", "ख़", "ग़", "ज़", "ड़", "ढ़", "फ़", "य़",
|
12 |
+
"्", "ा", "ि", "ी", "ु", "ू", "ॅ", "ॆ", "े", "ै", "ॉ", "ॊ", "ो", "ौ",
|
13 |
+
"ृ", "ॄ", "ॢ", "ॣ", "ँ", "ं", "ः", "़", "॑", "ऽ",
|
14 |
+
"\u200c", "\u200d"
|
15 |
+
|
16 |
+
],
|
17 |
+
|
18 |
+
"numsym_map" : {
|
19 |
+
"0" : ["०"], "1" : ["१"], "2" : ["२"], "3" : ["३"], "4" : ["४"],
|
20 |
+
"5" : ["५"], "6" : ["६"], "7" : ["७"], "8" : ["८"], "9" : ["९"],
|
21 |
+
"." : ["।", "॰"]
|
22 |
+
}
|
23 |
+
|
24 |
+
}
|
gom_words_subset.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
gu_101_model.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:60e4ecbd18d75c45c80bc8d25b6e38682c32534972c34cb0dc66369b48456b6c
|
3 |
+
size 40869530
|
gu_scripts.json
ADDED
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"WARNING" : " !!! Do not modify the Order of Glyph List !!!",
|
3 |
+
"UNICODE" : {"name": "gujarati", "begin":2688, "end":2815},
|
4 |
+
"LANGUAGE": "gujarati",
|
5 |
+
|
6 |
+
"glyphs" : [
|
7 |
+
"ઁ", "ં", "ઃ", "\u0a84", "અ", "આ", "ઇ", "ઈ", "ઉ", "ઊ", "ઋ", "ઌ", "ઍ",
|
8 |
+
"\u0a8e", "એ", "ઐ", "ઑ", "\u0a92", "ઓ", "ઔ", "ક", "ખ", "ગ", "ઘ",
|
9 |
+
"ઙ", "ચ", "છ", "જ", "ઝ", "ઞ", "ટ", "ઠ", "ડ", "ઢ", "ણ",
|
10 |
+
"ત", "થ", "દ", "ધ", "ન", "\u0aa9", "પ", "ફ", "બ", "ભ",
|
11 |
+
"મ", "ય", "ર", "\u0ab1", "લ", "ળ", "\u0ab4", "વ",
|
12 |
+
"શ", "ષ", "સ", "હ", "\u0aba", "\u0abb",
|
13 |
+
"઼", "ઽ", "ા", "િ", "ી", "ુ", "ૂ", "ૃ", "ૄ", "ૅ", "\u0ac6",
|
14 |
+
"ે", "ૈ", "ૉ", "\u0aca", "ો", "ૌ", "્",
|
15 |
+
"\u0ace", "\u0acf", "ૐ", "\u0ad1", "\u0ad2", "\u0ad3", "\u0ad4",
|
16 |
+
"\u0ad6", "\u0ad8", "\u0ad9", "\u0ada", "\u0adb", "\u0adc", "\u0add",
|
17 |
+
"\u0ade", "\u0adf", "ૠ", "ૡ", "ૢ", "ૣ", "\u0af3", "\u0af4", "\u0af5", "ૹ"
|
18 |
+
],
|
19 |
+
|
20 |
+
"numsym_map" : {
|
21 |
+
"0" : ["૦"], "1" : ["૧"], "2" : ["૨"], "3" : ["૩"], "4" : ["૪"],
|
22 |
+
"5" : ["૫"], "6" : ["૬"], "7" : ["૭"], "8" : ["૮"], "9" : ["૯"],
|
23 |
+
".":["૰"]
|
24 |
+
}
|
25 |
+
|
26 |
+
}
|
gu_words_a4b.json
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:399b2fc632e19653bb7e0763292bf350ac74d7b43282514a44956ad99212b8d3
|
3 |
+
size 28844852
|
hi_111_model.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:942c046e9d7ce721e835b65aa5db24b48a4249122cd580153f0573155bdf6b55
|
3 |
+
size 40847894
|
hi_scripts.json
ADDED
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"WARNING" : " !!! Do not modify the Order of Glyph List !!!",
|
3 |
+
"UNICODE" : {"name": "devanagari", "begin":2304, "end":2431},
|
4 |
+
"LANGUAGE": "hindi",
|
5 |
+
|
6 |
+
"glyphs" : [
|
7 |
+
|
8 |
+
"ऄ", "अ", "आ", "इ", "ई", "उ", "ऊ","ऍ", "ऎ", "ए", "ऐ",
|
9 |
+
"ऑ", "ऒ", "ओ", "औ","ऋ","ॠ","ऌ","ॡ","ॲ", "ॐ",
|
10 |
+
"क", "ख", "ग", "घ", "ङ", "च", "छ", "ज", "झ", "ञ", "ट", "ठ", "ड", "ढ", "ण",
|
11 |
+
"त", "थ", "द", "ध", "न", "ऩ", "प", "फ", "ब", "भ", "म", "य", "र", "ऱ", "ल",
|
12 |
+
"ळ", "ऴ", "व", "श", "ष", "स", "ह", "क़", "ख़", "ग़", "ज़", "ड़", "ढ़", "फ़", "य़",
|
13 |
+
"्", "ा", "ि", "ी", "ु", "ू", "ॅ", "ॆ", "े", "ै", "ॉ", "ॊ", "ो", "ौ",
|
14 |
+
"ृ", "ॄ", "ॢ", "ॣ", "ँ", "ं", "ः", "़", "॑", "ऽ",
|
15 |
+
"\u200c", "\u200d"
|
16 |
+
|
17 |
+
],
|
18 |
+
|
19 |
+
"numsym_map" : {
|
20 |
+
"0" : ["०"], "1" : ["१"], "2" : ["२"], "3" : ["३"], "4" : ["४"],
|
21 |
+
"5" : ["५"], "6" : ["६"], "7" : ["७"], "8" : ["८"], "9" : ["९"],
|
22 |
+
"." : ["।", "॰"]
|
23 |
+
}
|
24 |
+
|
25 |
+
}
|
hi_words_a4b.json
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:1d67b6810b60d689ac54f1e85ea6681dbe966b171de1516d17fc0280b9e3e13a
|
3 |
+
size 53304458
|
kn_101_model.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e9185b1400decd278923ab51e0b71d4b4700e50445eba26e81dd19dfc14f3dd2
|
3 |
+
size 40799814
|
kn_scripts.json
ADDED
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"WARNING" : " !!! Do not modify the Order of Glyph List !!!",
|
3 |
+
"UNICODE" : {"name": "kannada", "begin":3200, "end":3327},
|
4 |
+
"LANGUAGE": "kannada",
|
5 |
+
|
6 |
+
"glyphs" : [
|
7 |
+
"ಂ", "ಃ", "ಅ", "ಆ", "ಇ", "ಈ", "ಉ", "ಊ", "ಋ", "ಌ", "ಎ", "ಏ", "ಐ",
|
8 |
+
"ಒ", "ಓ", "ಔ", "ಕ", "ಖ", "ಗ", "ಘ", "ಙ", "ಚ", "ಛ", "ಜ", "ಝ", "ಞ",
|
9 |
+
"ಟ", "ಠ", "ಡ", "ಢ", "ಣ", "ತ", "ಥ", "ದ", "ಧ", "ನ", "ಪ", "ಫ", "ಬ", "ಭ",
|
10 |
+
"ಮ", "ಯ", "ರ", "ಱ", "ಲ", "ಳ", "ವ", "ಶ", "ಷ", "ಸ", "ಹ", "಼", "ಽ", "ಾ",
|
11 |
+
"ಿ", "ೀ", "ು", "ೂ", "ೃ", "ೄ", "ೆ", "ೇ", "ೈ", "ೊ", "ೋ", "ೌ",
|
12 |
+
"್", "ೕ", "ೖ", "ೞ", "ೠ", "ೡ"
|
13 |
+
],
|
14 |
+
|
15 |
+
"numsym_map" : {
|
16 |
+
"0" : ["೦"], "1" : ["೧"], "2" : ["೨"], "3" : ["೩"], "4" : ["೪"],
|
17 |
+
"5" : ["೫"], "6" : ["೬"], "7" : ["೭"], "8" : ["೮"], "9" : ["೯"]
|
18 |
+
}
|
19 |
+
|
20 |
+
}
|
kn_words_a4b.json
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e745b1bb3aa4ad3956b6151552a46df32379db5206087c2d660a8f5fe5e529b6
|
3 |
+
size 53105297
|
mai_122_model.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d5d25ef8e279928fbb16eed95f8ca59d99fc86fe5da155bd3779a2335323e96e
|
3 |
+
size 40847894
|
mai_scripts.json
ADDED
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"WARNING" : " !!! Do not modify the Order of Glyph List !!!",
|
3 |
+
"UNICODE" : {"name": "devanagari", "begin":2304, "end":2431},
|
4 |
+
"LANGUAGE": "maithili",
|
5 |
+
|
6 |
+
"glyphs" : [
|
7 |
+
|
8 |
+
"ऄ", "अ", "आ", "इ", "ई", "उ", "ऊ","ऍ", "ऎ", "ए", "ऐ",
|
9 |
+
"ऑ", "ऒ", "ओ", "औ","ऋ","ॠ","ऌ","ॡ","ॲ", "ॐ",
|
10 |
+
"क", "ख", "ग", "घ", "ङ", "च", "छ", "ज", "झ", "ञ", "ट", "ठ", "ड", "ढ", "ण",
|
11 |
+
"त", "थ", "द", "ध", "न", "ऩ", "प", "फ", "ब", "भ", "म", "य", "र", "ऱ", "ल",
|
12 |
+
"ळ", "ऴ", "व", "श", "ष", "स", "ह", "क़", "ख़", "ग़", "ज़", "ड़", "ढ़", "फ़", "य़",
|
13 |
+
"्", "ा", "ि", "ी", "ु", "ू", "ॅ", "ॆ", "े", "ै", "ॉ", "ॊ", "ो", "ौ",
|
14 |
+
"ृ", "ॄ", "ॢ", "ॣ", "ँ", "ं", "ः", "़", "॑", "ऽ",
|
15 |
+
"\u200c", "\u200d"
|
16 |
+
|
17 |
+
],
|
18 |
+
|
19 |
+
"numsym_map" : {
|
20 |
+
"0" : ["०"], "1" : ["१"], "2" : ["२"], "3" : ["३"], "4" : ["४"],
|
21 |
+
"5" : ["५"], "6" : ["६"], "7" : ["७"], "8" : ["८"], "9" : ["९"],
|
22 |
+
"." : ["।", "॰"]
|
23 |
+
}
|
24 |
+
|
25 |
+
}
|
mai_words_subset.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
ml_101_model.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:694f981ffd1721d330be7fcd456616df92b6229c6c230285849c98bcfe06e2a4
|
3 |
+
size 40850298
|
ml_scripts.json
ADDED
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"WARNING" : " !!! Do not modify the Order of Glyph List !!!",
|
3 |
+
"UNICODE" : {"name": "malayalam", "begin":3328, "end":3455},
|
4 |
+
"LANGUAGE": "malayalam",
|
5 |
+
|
6 |
+
"glyphs" : [
|
7 |
+
"ം", "ഃ",
|
8 |
+
"അ", "ആ", "ഇ", "ഈ", "ഉ", "ഊ", "ഋ", "ഌ", "എ", "ഏ", "ഐ", "ഒ", "ഓ", "ഔ",
|
9 |
+
"ക", "ഖ", "ഗ", "ഘ", "ങ", "ച", "ഛ", "ജ", "ഝ", "ഞ",
|
10 |
+
"ട", "ഠ", "ഡ", "ഢ", "ണ", "ത", "ഥ", "ദ", "ധ", "ന", "ഩ",
|
11 |
+
"പ", "ഫ", "ബ", "ഭ", "മ", "യ", "ര", "റ", "ല", "ള", "ഴ",
|
12 |
+
"വ", "ശ", "ഷ", "സ", "ഹ", "ഺ", "\u0d3b", "\u0d3c", "ഽ",
|
13 |
+
"ാ", "ി", "ീ", "ു", "ൂ", "ൃ", "ൄ", "െ", "േ", "ൈ", "ൊ", "ോ", "ൌ", "്", "ൎ",
|
14 |
+
"ൖ", "ൗ", "൛", "൜", "൞", "ൟ", "ൠ", "ൡ", "ൢ", "ൣ", "൰", "൱",
|
15 |
+
"൲", "൹", "ൺ", "ൻ", "ർ", "ൽ", "ൾ",
|
16 |
+
"\u200c", "\u200d"
|
17 |
+
|
18 |
+
],
|
19 |
+
|
20 |
+
"numsym_map" : {
|
21 |
+
"0" : ["൦"], "1" : ["൧"], "2" : ["൨"], "3" : ["൩"], "4" : ["൪"],
|
22 |
+
"5" : ["൫"], "6" : ["൬"], "7" : ["൭"], "8" : ["൮"], "9" : ["൯"]
|
23 |
+
}
|
24 |
+
|
25 |
+
}
|
ml_words_a4b.json
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:993252dc17f4dc1af6714a9c65ba507962d175724f4984073560feda7f96207d
|
3 |
+
size 142304885
|
mr_102_model.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4c536333efe74377e4b47935de94bcd6ed1f02c5843b25e557dc1b0b0b5bda21
|
3 |
+
size 40847894
|
mr_scripts.json
ADDED
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"WARNING" : " !!! Do not modify the Order of Glyph List !!!",
|
3 |
+
"UNICODE" : {"name": "devanagari", "begin":2304, "end":2431},
|
4 |
+
"LANGUAGE": "marathi",
|
5 |
+
|
6 |
+
"glyphs" : [
|
7 |
+
|
8 |
+
"ऄ", "अ", "आ", "इ", "ई", "उ", "ऊ","ऍ", "ऎ", "ए", "ऐ",
|
9 |
+
"ऑ", "ऒ", "ओ", "औ","ऋ","ॠ","ऌ","ॡ","ॲ", "ॐ",
|
10 |
+
"क", "ख", "ग", "घ", "ङ", "च", "छ", "ज", "झ", "ञ", "ट", "ठ", "ड", "ढ", "ण",
|
11 |
+
"त", "थ", "द", "ध", "न", "ऩ", "प", "फ", "ब", "भ", "म", "य", "र", "ऱ", "ल",
|
12 |
+
"ळ", "ऴ", "व", "श", "ष", "स", "ह", "क़", "ख़", "ग़", "ज़", "ड़", "ढ़", "फ़", "य़",
|
13 |
+
"्", "ा", "ि", "ी", "ु", "ू", "ॅ", "ॆ", "े", "ै", "ॉ", "ॊ", "ो", "ौ",
|
14 |
+
"ृ", "ॄ", "ॢ", "ॣ", "ँ", "ं", "ः", "़", "॑", "ऽ",
|
15 |
+
"\u200c", "\u200d"
|
16 |
+
|
17 |
+
],
|
18 |
+
|
19 |
+
"numsym_map" : {
|
20 |
+
"0" : ["०"], "1" : ["१"], "2" : ["२"], "3" : ["३"], "4" : ["४"],
|
21 |
+
"5" : ["५"], "6" : ["६"], "7" : ["७"], "8" : ["८"], "9" : ["९"],
|
22 |
+
"." : ["।", "॰"]
|
23 |
+
}
|
24 |
+
|
25 |
+
}
|
mr_words_a4b.json
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8f4bbb584ea6935063d06ab44989eadb8d2aedf7a672197427715d508b747ec5
|
3 |
+
size 37157053
|
pa_101_model.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cc92f0bdd2577674d283fbd53a5fdba7e736e951931aa8df48327d9aa87fd3b4
|
3 |
+
size 40792602
|
pa_scripts.json
ADDED
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"WARNING" : " !!! Do not modify the Order of Glyph List !!!",
|
3 |
+
"UNICODE" : {"name": "gurmukhi", "begin":2560, "end":2687},
|
4 |
+
"LANGUAGE": "punjabi_eastern",
|
5 |
+
|
6 |
+
"glyphs" : [
|
7 |
+
"ਁ", "ਂ", "ਃ", "ਅ", "ਆ", "ਇ", "ਈ", "ਉ", "ਊ", "ਏ", "ਐ", "ਓ", "ਔ", "ਕ",
|
8 |
+
"ਖ", "ਗ", "ਘ", "ਙ", "ਚ", "ਛ", "ਜ", "ਝ", "ਞ", "ਟ", "ਠ", "ਡ", "ਢ", "ਣ", "ਤ",
|
9 |
+
"ਥ", "ਦ", "ਧ", "ਨ", "ਪ", "ਫ", "ਬ", "ਭ", "ਮ", "ਯ", "ਰ", "ਲ", "ਲ਼", "ਵ", "ਸ਼",
|
10 |
+
"ਸ", "ਹ", "਼", "ਾ", "ਿ", "ੀ", "ੁ", "ੂ", "ੇ", "ੈ", "ੋ", "ੌ", "੍", "ੑ",
|
11 |
+
"ਖ਼", "ਗ਼", "ਜ਼", "ੜ", "ਫ਼", "ੰ", "ੱ", "ੲ", "ੳ", "ੴ", "ੵ"
|
12 |
+
],
|
13 |
+
|
14 |
+
"numsym_map" : {
|
15 |
+
"0" : ["੦"], "1" : ["੧"], "2" : ["੨"], "3" : ["੩"], "4" : ["੪"],
|
16 |
+
"5" : ["੫"], "6" : ["੬"], "7" : ["੭"], "8" : ["੮"], "9" : ["੯"]
|
17 |
+
}
|
18 |
+
|
19 |
+
}
|
pa_words_a4b.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
sd_101_model.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:01b658262ec8e6970625e9fabb77da65ee58f052536aa5d30ce1f7ca21ac5ce8
|
3 |
+
size 41629202
|
sd_scripts.json
ADDED
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"WARNING" : " !!! Do not modify the Order of Glyph List !!!",
|
3 |
+
"UNICODE" : {"name": "arabic", "begin":[1536,64336,65136], "end":[1791,65023,65278]},
|
4 |
+
"LANGUAGE": "sindhi_arabic",
|
5 |
+
|
6 |
+
"glyphs" : [
|
7 |
+
"\u0600", "\u0601", "\u0602", "\u0603", "\u0604",
|
8 |
+
"؈", "؍", "؎", "؏", "ؐ", "ؑ", "ؒ", "ؓ", "ؔ", "ؕ", "ؠ", "ء", "آ", "أ", "ؤ", "إ", "ئ",
|
9 |
+
"ا", "ب", "ة", "ت", "ث", "ج", "ح", "خ", "د", "ذ", "ر", "ز", "س", "ش", "ص", "ض",
|
10 |
+
"ط", "ظ", "ع", "غ", "ػ", "ؽ", "ؿ", "ـ", "ف", "ق", "ك", "ل", "م", "ن", "ه", "و",
|
11 |
+
"ى", "ي", "ً", "ٌ", "ٍ", "َ", "ُ", "ِ", "ّ", "ْ", "ٓ", "ٔ", "ٕ", "ٖ", "ٗ", "٘", "ٙ", "ٚ", "ٛ",
|
12 |
+
"ٜ", "ٝ", "ٞ", "ٟ", "٫", "٬", "٭", "ٮ", "ٰ", "ٱ", "ٲ", "ٳ", "ٴ", "ٶ", "ٷ", "ٸ", "ٹ",
|
13 |
+
"ٺ", "ٻ", "ټ", "ٽ", "پ", "ٿ", "ڀ", "ځ", "ڃ", "ڄ", "څ", "چ", "ڇ", "ڈ", "ډ", "ڊ",
|
14 |
+
"ڋ", "ڌ", "ڍ", "ڎ", "ڏ", "ڑ", "ړ", "ږ", "ژ", "ڙ", "ښ", "ڠ", "ڤ", "ڦ", "ڨ", "ک",
|
15 |
+
"ڪ", "ګ", "ڬ", "ڭ", "ڮ", "گ", "ڱ", "ڳ", "ڵ", "ڷ", "ڹ", "ں", "ڻ", "ڼ", "ھ",
|
16 |
+
"ۀ", "ہ", "ۂ", "ۃ", "ۄ", "ۆ", "ۇ", "ۈ", "ۉ", "ۋ", "ی", "ۍ", "ێ", "ې", "ۑ", "ے",
|
17 |
+
"ۓ", "ە", "ۖ", "ۗ", "ۘ", "ۙ", "ۚ", "ۛ", "ۜ", "\u06dd", "۟", "۠", "ۡ", "ۢ", "ۣ", "ۤ", "ۥ", "ۦ", "ۧ",
|
18 |
+
"ۨ", "۬", "ۭ", "ۯ", "ۺ", "۽", "۾", "ۿ", "ﭐ", "ﭑ", "ﭒ", "ﭔ", "ﭕ", "ﭖ", "ﭗ", "ﭘ", "ﭙ",
|
19 |
+
"ﭚ", "ﭛ", "ﭜ", "ﭝ", "ﭞ", "ﭟ", "ﭠ", "ﭡ", "ﭣ", "ﭤ", "ﭥ", "ﭦ", "ﭧ", "ﭨ", "ﭩ", "ﭰ",
|
20 |
+
"ﭲ", "ﭳ", "ﭴ", "ﭵ", "ﭷ", "ﭸ", "ﭹ", "ﭺ", "ﭻ", "ﭼ", "ﭽ", "ﮀ", "ﮁ", "ﮂ", "ﮃ", "ﮄ",
|
21 |
+
"ﮅ", "ﮈ", "ﮉ", "ﮋ", "ﮌ", "ﮍ", "ﮎ", "ﮏ", "ﮐ", "ﮑ", "ﮒ", "ﮓ", "ﮔ", "ﮕ", "ﮖ", "ﮘ",
|
22 |
+
"ﮙ", "ﮜ", "ﮝ", "ﮞ", "ﮟ", "ﮠ", "ﮡ", "ﮢ", "ﮣ", "ﮦ", "ﮧ", "ﮨ", "ﮩ", "ﮪ", "ﮫ", "ﮬ",
|
23 |
+
"ﮭ", "ﮮ", "ﮯ", "ﮰ", "ﮱ", "ﯼ", "ﯽ", "ﯾ", "ﯿ", "ﰦ", "ﰷ", "ﱕ", "ﱞ", "ﲉ", "ﳀ", "ﴒ",
|
24 |
+
"ﴝ", "ﴼ", "ﴽ", "﴾", "﴿", "ﶌ", "ﶾ", "\ufdde", "\ufddf", "\ufde1", "\ufde2", "\ufde3",
|
25 |
+
"\ufde5", "\ufde6", "\ufdeb", "ﷲ", "ﷴ", "ﷵ", "ﷹ", "ﷺ", "ﷻ", "﷼", "﷽", "\ufdfe",
|
26 |
+
"ﹰ", "ﹶ", "ﹸ", "ﹺ", "ﺀ", "ﺁ", "ﺂ", "ﺃ", "ﺄ", "ﺅ", "ﺆ", "ﺇ", "ﺉ", "ﺊ", "ﺋ", "ﺌ",
|
27 |
+
"ﺍ", "ﺎ", "ﺏ", "ﺐ", "ﺑ", "ﺒ", "ﺓ", "ﺔ", "ﺕ", "ﺖ", "ﺗ", "ﺘ", "ﺙ", "ﺚ", "ﺛ", "ﺜ",
|
28 |
+
"ﺝ", "ﺞ", "ﺟ", "ﺠ", "ﺡ", "ﺢ", "ﺣ", "ﺤ", "ﺥ", "ﺦ", "ﺧ", "ﺨ", "ﺩ", "ﺪ", "ﺫ", "ﺬ",
|
29 |
+
"ﺭ", "ﺮ", "ﺯ", "ﺰ", "ﺱ", "ﺲ", "ﺳ", "ﺴ", "ﺵ", "ﺶ", "ﺷ", "ﺸ", "ﺹ", "ﺺ", "ﺻ", "ﺼ",
|
30 |
+
"ﺽ", "ﺾ", "ﺿ", "ﻀ", "ﻁ", "ﻂ", "ﻃ", "ﻄ", "ﻅ", "ﻆ", "ﻇ", "ﻈ", "ﻉ", "ﻊ", "ﻋ", "ﻌ",
|
31 |
+
"ﻍ", "ﻎ", "ﻏ", "ﻐ", "ﻑ", "ﻒ", "ﻓ", "ﻔ", "ﻕ", "ﻖ", "ﻗ", "ﻘ", "ﻚ", "ﻛ", "ﻜ", "ﻝ",
|
32 |
+
"ﻞ", "ﻟ", "ﻠ", "ﻡ", "ﻢ", "ﻣ", "ﻤ", "ﻥ", "ﻦ", "ﻧ", "ﻨ", "ﻩ", "ﻪ", "ﻫ", "ﻬ", "ﻭ",
|
33 |
+
"ﻮ", "ﻯ", "ﻰ", "ﻱ", "ﻲ", "ﻳ", "ﻴ", "ﻵ", "ﻶ", "ﻹ", "ﻻ", "ﻼ"
|
34 |
+
],
|
35 |
+
|
36 |
+
"numsym_map" : {
|
37 |
+
"0" : ["٠"], "1" : ["١"], "2" : ["٢"], "3" : ["٣"], "4" : ["٤"],
|
38 |
+
"5" : ["٥"], "6" : ["٦"], "7" : ["٧"], "8" : ["٨"], "9" : ["٩"],
|
39 |
+
"." : ["۔", "۬"," ۫", "۪"],
|
40 |
+
"?":["؟"], ";":["؛"], ",":["،"]
|
41 |
+
|
42 |
+
}
|
43 |
+
|
44 |
+
|
45 |
+
|
46 |
+
}
|
47 |
+
|
48 |
+
|
sd_words_ccset.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
si_101_model.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:8a1950562c5a0bc7d06ae773258536b7e01d039c10ede8c6c761d7e753835f0f
|
3 |
+
size 40915206
|
si_scripts.json
ADDED
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"WARNING" : " !!! Do not modify the Order of Glyph List !!!",
|
3 |
+
"UNICODE" : {"name": "sinhala", "begin":2560, "end":2687},
|
4 |
+
"LANGUAGE": "sinhala",
|
5 |
+
|
6 |
+
"glyphs" : [
|
7 |
+
|
8 |
+
"\u0d80", "\u0d81", "ං", "ඃ", "\u0d84", "අ", "ආ", "ඇ", "ඈ", "ඉ", "ඊ", "උ",
|
9 |
+
"ඌ", "ඍ", "ඎ", "ඏ", "ඐ", "එ", "ඒ", "ඓ", "ඔ", "ඕ", "ඖ", "\u0d97",
|
10 |
+
"\u0d98", "\u0d99", "ක", "ඛ", "ග", "ඝ", "ඞ", "ඟ", "ච", "ඡ", "ජ", "ඣ",
|
11 |
+
"ඤ", "ඥ", "ඦ", "ට", "ඨ", "ඩ", "ඪ", "ණ", "ඬ", "ත", "ථ", "ද", "ධ", "න",
|
12 |
+
"\u0db2", "ඳ", "ප", "ඵ", "බ", "භ", "ම", "ඹ", "ය", "ර", "\u0dbc", "ල",
|
13 |
+
"\u0dbe", "\u0dbf", "ව", "ශ", "ෂ", "ස", "හ", "ළ", "ෆ", "\u0dc7", "\u0dc8",
|
14 |
+
"\u0dc9", "්", "\u0dcb", "\u0dcc", "\u0dcd", "\u0dce", "ා", "ැ", "ෑ", "ි",
|
15 |
+
"ී", "ු", "\u0dd5", "ූ", "\u0dd7", "ෘ", "ෙ", "ේ", "ෛ", "ො", "ෝ", "ෞ", "ෟ",
|
16 |
+
"\u0de0", "\u0de1", "\u0de2", "\u0de3", "\u0de4", "\u0de5", "\u0df0", "\u0df1", "ෲ", "ෳ", "෴",
|
17 |
+
"\u0df5", "\u0df6", "\u0df7", "\u0df8", "\u0df9", "\u0dfa", "\u0dfb",
|
18 |
+
"\u0dfc", "\u0dfd", "\u0dfe", "\u0dff",
|
19 |
+
"\u200c", "\u200d"
|
20 |
+
|
21 |
+
|
22 |
+
],
|
23 |
+
|
24 |
+
"numsym_map" : {
|
25 |
+
"0" : ["෦"], "1" : ["෧"], "2" : ["෨"], "3" : ["෩"], "4" : ["෪"],
|
26 |
+
"5" : ["෫"], "6" : ["෬"], "7" : ["෭"], "8" : ["෮"], "9" : ["෯"]
|
27 |
+
}
|
28 |
+
|
29 |
+
}
|
si_words_ccset.json
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:992472b1eed8c78a276b0ce607b22d5d4fbfd74d3e5dc05b8741f4f533578c4d
|
3 |
+
size 18073516
|
ta_101_model.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:26e8cb09d0268df1f78e797f89f8cdda86d3a8c054d223cd3abb92d9e429eb77
|
3 |
+
size 40744522
|
ta_scripts.json
ADDED
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"WARNING" : " !!! Do not modify the Order of Glyph List !!!",
|
3 |
+
"UNICODE" : {"name": "tamil", "begin":2944, "end":3071},
|
4 |
+
"LANGUAGE": "tamil",
|
5 |
+
|
6 |
+
"glyphs" : [
|
7 |
+
"ஂ","ஃ","அ","ஆ","இ","ஈ","உ","ஊ","எ","ஏ","ஐ","ஒ","ஓ","ஔ",
|
8 |
+
"க","ங","ச","ஜ","ஞ","ட","ண","த","ந","ன","ப","ம","ய","ர","ற",
|
9 |
+
"ல","ள","ழ","வ","ஷ","ஸ","ஹ",
|
10 |
+
"ா","ி","ீ","ு","ூ","ெ","ே","ை","ொ","ோ","ௌ","்","ௗ"
|
11 |
+
],
|
12 |
+
|
13 |
+
"numsym_map" : {
|
14 |
+
"0" : ["௦"], "1" : ["௧"], "2" : ["௨"], "3" : ["௩"], "4" : ["௪"],
|
15 |
+
"5" : ["௫"], "6" : ["௬"], "7" : ["௭"], "8" : ["௮"], "9" : ["௯"],
|
16 |
+
"10" : ["௰"], "100": ["௱"], "1000": ["௲"]
|
17 |
+
}
|
18 |
+
|
19 |
+
}
|
ta_words_a4b.json
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c76715b5b11630214496a2a3f2dbf4e37deb3bd25194c48e6bebcffe437c7368
|
3 |
+
size 160705281
|
te_101_model.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:9c51ee95ae014a409d21b69040181096f8d13aa5d5ca6790e26cd6bda1315ef4
|
3 |
+
size 40799814
|
te_scripts.json
ADDED
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"WARNING" : " !!! Do not modify the Order of Glyph List !!!",
|
3 |
+
"UNICODE" : {"name": "telugu", "begin": 3072, "end":3199},
|
4 |
+
"LANGUAGE": "telugu",
|
5 |
+
|
6 |
+
"glyphs" : [
|
7 |
+
"ఁ", "ం", "ః", "అ", "ఆ", "ఇ", "ఈ", "ఉ", "ఊ", "ఋ", "ఌ", "ఎ",
|
8 |
+
"ఏ", "ఐ", "ఒ", "ఓ", "ఔ", "క", "ఖ", "గ", "ఘ", "ఙ", "చ", "ఛ", "జ", "ఝ", "ఞ", "ట",
|
9 |
+
"ఠ", "డ", "ఢ", "ణ", "త", "థ", "ద", "ధ", "న", "ప", "ఫ", "బ", "భ", "మ", "య", "ర",
|
10 |
+
"ఱ", "ల", "ళ", "వ", "శ", "ష", "స", "హ", "ా", "ి", "ీ", "ు", "ూ",
|
11 |
+
"ృ", "ౄ", "ె", "ే", "ై", "ొ", "ో", "ౌ", "్", "ౕ", "ౖ","ౠ", "ౡ",
|
12 |
+
|
13 |
+
"\u200c", "\u200d"
|
14 |
+
],
|
15 |
+
|
16 |
+
"numsym_map" : {
|
17 |
+
"0" : ["௦"], "1" : ["౧"], "2" : ["౨"], "3" : ["౩"], "4" : ["౪"],
|
18 |
+
"5" : ["౫"], "6" : ["౬"], "7" : ["౭"], "8" : ["౮"], "9" : ["౯"]
|
19 |
+
}
|
20 |
+
|
21 |
+
}
|
te_words_a4b.json
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4da677fb940729e7d33fb4d73fac77faf883cec8c215f4dfa25e0c4e9f57eee6
|
3 |
+
size 63930508
|
ur_101_model.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:e8b6845560e1cfbc3003024e994faaa5eb5197c5c5607558ac718e0e723cb0cf
|
3 |
+
size 41629202
|
ur_scripts.json
ADDED
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"WARNING" : " !!! Do not modify the Order of Glyph List !!!",
|
3 |
+
"UNICODE" : {"name": "arabic", "begin":[1536,64336,65136], "end":[1791,65023,65278]},
|
4 |
+
"LANGUAGE": "urdu",
|
5 |
+
|
6 |
+
"glyphs" : [
|
7 |
+
"\u0600", "\u0601", "\u0602", "\u0603", "\u0604",
|
8 |
+
"؈", "؍", "؎", "؏", "ؐ", "ؑ", "ؒ", "ؓ", "ؔ", "ؕ", "ؠ", "ء", "آ", "أ", "ؤ", "إ", "ئ",
|
9 |
+
"ا", "ب", "ة", "ت", "ث", "ج", "ح", "خ", "د", "ذ", "ر", "ز", "س", "ش", "ص", "ض",
|
10 |
+
"ط", "ظ", "ع", "غ", "ػ", "ؽ", "ؿ", "ـ", "ف", "ق", "ك", "ل", "م", "ن", "ه", "و",
|
11 |
+
"ى", "ي", "ً", "ٌ", "ٍ", "َ", "ُ", "ِ", "ّ", "ْ", "ٓ", "ٔ", "ٕ", "ٖ", "ٗ", "٘", "ٙ", "ٚ", "ٛ",
|
12 |
+
"ٜ", "ٝ", "ٞ", "ٟ", "٫", "٬", "٭", "ٮ", "ٰ", "ٱ", "ٲ", "ٳ", "ٴ", "ٶ", "ٷ", "ٸ", "ٹ",
|
13 |
+
"ٺ", "ٻ", "ټ", "ٽ", "پ", "ٿ", "ڀ", "ځ", "ڃ", "ڄ", "څ", "چ", "ڇ", "ڈ", "ډ", "ڊ",
|
14 |
+
"ڋ", "ڌ", "ڍ", "ڎ", "ڏ", "ڑ", "ړ", "ږ", "ژ", "ڙ", "ښ", "ڠ", "ڤ", "ڦ", "ڨ", "ک",
|
15 |
+
"ڪ", "ګ", "ڬ", "ڭ", "ڮ", "گ", "ڱ", "ڳ", "ڵ", "ڷ", "ڹ", "ں", "ڻ", "ڼ", "ھ",
|
16 |
+
"ۀ", "ہ", "ۂ", "ۃ", "ۄ", "ۆ", "ۇ", "ۈ", "ۉ", "ۋ", "ی", "ۍ", "ێ", "ې", "ۑ", "ے",
|
17 |
+
"ۓ", "ە", "ۖ", "ۗ", "ۘ", "ۙ", "ۚ", "ۛ", "ۜ", "\u06dd", "۟", "۠", "ۡ", "ۢ", "ۣ", "ۤ", "ۥ", "ۦ", "ۧ",
|
18 |
+
"ۨ", "۬", "ۭ", "ۯ", "ۺ", "۽", "۾", "ۿ", "ﭐ", "ﭑ", "ﭒ", "ﭔ", "ﭕ", "ﭖ", "ﭗ", "ﭘ", "ﭙ",
|
19 |
+
"ﭚ", "ﭛ", "ﭜ", "ﭝ", "ﭞ", "ﭟ", "ﭠ", "ﭡ", "ﭣ", "ﭤ", "ﭥ", "ﭦ", "ﭧ", "ﭨ", "ﭩ", "ﭰ",
|
20 |
+
"ﭲ", "ﭳ", "ﭴ", "ﭵ", "ﭷ", "ﭸ", "ﭹ", "ﭺ", "ﭻ", "ﭼ", "ﭽ", "ﮀ", "ﮁ", "ﮂ", "ﮃ", "ﮄ",
|
21 |
+
"ﮅ", "ﮈ", "ﮉ", "ﮋ", "ﮌ", "ﮍ", "ﮎ", "ﮏ", "ﮐ", "ﮑ", "ﮒ", "ﮓ", "ﮔ", "ﮕ", "ﮖ", "ﮘ",
|
22 |
+
"ﮙ", "ﮜ", "ﮝ", "ﮞ", "ﮟ", "ﮠ", "ﮡ", "ﮢ", "ﮣ", "ﮦ", "ﮧ", "ﮨ", "ﮩ", "ﮪ", "ﮫ", "ﮬ",
|
23 |
+
"ﮭ", "ﮮ", "ﮯ", "ﮰ", "ﮱ", "ﯼ", "ﯽ", "ﯾ", "ﯿ", "ﰦ", "ﰷ", "ﱕ", "ﱞ", "ﲉ", "ﳀ", "ﴒ",
|
24 |
+
"ﴝ", "ﴼ", "ﴽ", "﴾", "﴿", "ﶌ", "ﶾ", "\ufdde", "\ufddf", "\ufde1", "\ufde2", "\ufde3",
|
25 |
+
"\ufde5", "\ufde6", "\ufdeb", "ﷲ", "ﷴ", "ﷵ", "ﷹ", "ﷺ", "ﷻ", "﷼", "﷽", "\ufdfe",
|
26 |
+
"ﹰ", "ﹶ", "ﹸ", "ﹺ", "ﺀ", "ﺁ", "ﺂ", "ﺃ", "ﺄ", "ﺅ", "ﺆ", "ﺇ", "ﺉ", "ﺊ", "ﺋ", "ﺌ",
|
27 |
+
"ﺍ", "ﺎ", "ﺏ", "ﺐ", "ﺑ", "ﺒ", "ﺓ", "ﺔ", "ﺕ", "ﺖ", "ﺗ", "ﺘ", "ﺙ", "ﺚ", "ﺛ", "ﺜ",
|
28 |
+
"ﺝ", "ﺞ", "ﺟ", "ﺠ", "ﺡ", "ﺢ", "ﺣ", "ﺤ", "ﺥ", "ﺦ", "ﺧ", "ﺨ", "ﺩ", "ﺪ", "ﺫ", "ﺬ",
|
29 |
+
"ﺭ", "ﺮ", "ﺯ", "ﺰ", "ﺱ", "ﺲ", "ﺳ", "ﺴ", "ﺵ", "ﺶ", "ﺷ", "ﺸ", "ﺹ", "ﺺ", "ﺻ", "ﺼ",
|
30 |
+
"ﺽ", "ﺾ", "ﺿ", "ﻀ", "ﻁ", "ﻂ", "ﻃ", "ﻄ", "ﻅ", "ﻆ", "ﻇ", "ﻈ", "ﻉ", "ﻊ", "ﻋ", "ﻌ",
|
31 |
+
"ﻍ", "ﻎ", "ﻏ", "ﻐ", "ﻑ", "ﻒ", "ﻓ", "ﻔ", "ﻕ", "ﻖ", "ﻗ", "ﻘ", "ﻚ", "ﻛ", "ﻜ", "ﻝ",
|
32 |
+
"ﻞ", "ﻟ", "ﻠ", "ﻡ", "ﻢ", "ﻣ", "ﻤ", "ﻥ", "ﻦ", "ﻧ", "ﻨ", "ﻩ", "ﻪ", "ﻫ", "ﻬ", "ﻭ",
|
33 |
+
"ﻮ", "ﻯ", "ﻰ", "ﻱ", "ﻲ", "ﻳ", "ﻴ", "ﻵ", "ﻶ", "ﻹ", "ﻻ", "ﻼ"
|
34 |
+
],
|
35 |
+
|
36 |
+
"numsym_map" : {
|
37 |
+
"0" : ["٠"], "1" : ["١"], "2" : ["٢"], "3" : ["٣"], "4" : ["٤"],
|
38 |
+
"5" : ["٥"], "6" : ["٦"], "7" : ["٧"], "8" : ["٨"], "9" : ["٩"],
|
39 |
+
"." : ["۔", "۬"," ۫", "۪"],
|
40 |
+
"?":["؟"], ";":["؛"], ",":["،"]
|
41 |
+
|
42 |
+
}
|
43 |
+
|
44 |
+
|
45 |
+
|
46 |
+
}
|
47 |
+
|
48 |
+
|
ur_words_ccset.json
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:75721c75452ec229230a063ca2a847c8a568ef8ba2a26aebcfc2c3ea31b8b37b
|
3 |
+
size 14591474
|