nrshoudi
/

wav2vec2-large-xls-r-300m-Arabic-phoneme-based

Automatic Speech Recognition

Generated from Trainer

Inference Endpoints

Model card Files Files and versions Community

nrshoudi commited on Jun 13, 2023

Commit

38b53b1

•

1 Parent(s): eed16a8

Upload tokenizer

Files changed (2) hide show

added_tokens.json +2 -2
vocab.json +45 -48

added_tokens.json CHANGED Viewed

@@ -1,4 +1,4 @@
 {
-  "</s>": 50,
-  "<s>": 49
 }

 {
+  "</s>": 47,
+  "<s>": 46
 }

vocab.json CHANGED Viewed

@@ -1,51 +1,48 @@
 {
-  "[PAD]": 48,
-  "[UNK]": 47,
-  "e": 1,
-  "g": 2,
-  "t": 3,
   "|": 0,
-  "ء": 4,
-  "آ": 5,
-  "أ": 6,
-  "ؤ": 7,
-  "إ": 8,
-  "ئ": 9,
-  "ا": 10,
-  "ب": 11,
-  "ة": 12,
-  "ت": 13,
-  "ث": 14,
-  "ج": 15,
-  "ح": 16,
-  "خ": 17,
-  "د": 18,
-  "ذ": 19,
-  "ر": 20,
-  "ز": 21,
-  "س": 22,
-  "ش": 23,
-  "ص": 24,
-  "ض": 25,
-  "ط": 26,
-  "ظ": 27,
-  "ع": 28,
-  "غ": 29,
-  "ف": 30,
-  "ق": 31,
-  "ك": 32,
-  "ل": 33,
-  "م": 34,
-  "ن": 35,
-  "ه": 36,
-  "و": 37,
-  "ى": 38,
-  "ي": 39,
-  "چ": 40,
-  "ڨ": 41,
-  "ک": 42,
-  "ھ": 43,
-  "ی": 44,
-  "ﺃ": 45,
-  "ﻻ": 46
 }

 {
+  "[PAD]": 45,
+  "[UNK]": 44,
   "|": 0,
+  "ء": 1,
+  "آ": 2,
+  "أ": 3,
+  "ؤ": 4,
+  "إ": 5,
+  "ئ": 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,
+  "ف": 27,
+  "ق": 28,
+  "ك": 29,
+  "ل": 30,
+  "م": 31,
+  "ن": 32,
+  "ه": 33,
+  "و": 34,
+  "ى": 35,
+  "ي": 36,
+  "چ": 37,
+  "ڨ": 38,
+  "ک": 39,
+  "ھ": 40,
+  "ی": 41,
+  "ﺃ": 42,
+  "ﻻ": 43
 }