Yasinjan99
/

wav2vec2-large-xls-r-300m-tr-colab-test-1106

Automatic Speech Recognition

Generated from Trainer

Inference Endpoints

Model card Files Files and versions Community

Yasinjan99 commited on Nov 9, 2023

Commit

69e9711

•

1 Parent(s): 12bc76c

Upload tokenizer

Files changed (3) hide show

added_tokens.json +2 -2
tokenizer_config.json +4 -4
vocab.json +48 -41

added_tokens.json CHANGED Viewed

@@ -1,4 +1,4 @@
 {
-  "</s>": 43,
-  "<s>": 42
 }

 {
+  "</s>": 50,
+  "<s>": 49
 }

tokenizer_config.json CHANGED Viewed

@@ -1,6 +1,6 @@
 {
   "added_tokens_decoder": {
-    "40": {
       "content": "[UNK]",
       "lstrip": true,
       "normalized": false,
@@ -8,7 +8,7 @@
       "single_word": false,
       "special": false
     },
-    "41": {
       "content": "[PAD]",
       "lstrip": true,
       "normalized": false,
@@ -16,7 +16,7 @@
       "single_word": false,
       "special": false
     },
-    "42": {
       "content": "<s>",
       "lstrip": false,
       "normalized": true,
@@ -24,7 +24,7 @@
       "single_word": false,
       "special": true
     },
-    "43": {
       "content": "</s>",
       "lstrip": false,
       "normalized": true,

 {
   "added_tokens_decoder": {
+    "47": {
       "content": "[UNK]",
       "lstrip": true,
       "normalized": false,
       "single_word": false,
       "special": false
     },
+    "48": {
       "content": "[PAD]",
       "lstrip": true,
       "normalized": false,
       "single_word": false,
       "special": false
     },
+    "49": {
       "content": "<s>",
       "lstrip": false,
       "normalized": true,
       "single_word": false,
       "special": true
     },
+    "50": {
       "content": "</s>",
       "lstrip": false,
       "normalized": true,

vocab.json CHANGED Viewed

@@ -1,44 +1,51 @@
 {
-  "(": 1,
-  ")": 2,
-  "[PAD]": 41,
-  "[UNK]": 40,
-  "a": 3,
-  "b": 4,
-  "c": 5,
-  "d": 6,
-  "e": 7,
-  "f": 8,
-  "g": 9,
-  "h": 10,
-  "i": 11,
-  "j": 12,
-  "k": 13,
-  "l": 14,
-  "m": 15,
-  "n": 16,
-  "o": 17,
-  "p": 18,
-  "q": 19,
-  "r": 20,
-  "s": 21,
-  "t": 22,
-  "u": 23,
-  "v": 24,
-  "w": 25,
-  "x": 26,
-  "y": 27,
-  "z": 28,
   "|": 0,
-  "ç": 29,
-  "é": 30,
-  "ë": 31,
-  "ö": 32,
-  "ü": 33,
-  "ğ": 34,
-  "ı": 35,
-  "ş": 36,
-  "̇": 37,
-  "’": 38,
-  "…": 39
 }

 {
+  "[PAD]": 48,
+  "[UNK]": 47,
+  "`": 1,
   "|": 0,
+  "«": 2,
+  "»": 3,
+  "،": 4,
+  "؛": 5,
+  "؟": 6,
+  "ئ": 7,
+  "ا": 8,
+  "ب": 9,
+  "ت": 10,
+  "ج": 11,
+  "خ": 12,
+  "د": 13,
+  "ر": 14,
+  "ز": 15,
+  "س": 16,
+  "ش": 17,
+  "غ": 18,
+  "ف": 19,
+  "ق": 20,
+  "ك": 21,
+  "ل": 22,
+  "م": 23,
+  "ن": 24,
+  "و": 25,
+  "ى": 26,
+  "ي": 27,
+  "پ": 28,
+  "چ": 29,
+  "ژ": 30,
+  "ک": 31,
+  "ڭ": 32,
+  "گ": 33,
+  "ھ": 34,
+  "ۆ": 35,
+  "ۇ": 36,
+  "ۈ": 37,
+  "ۋ": 38,
+  "ی": 39,
+  "ې": 40,
+  "ە": 41,
+  "–": 42,
+  "—": 43,
+  "‹": 44,
+  "›": 45,
+  "−": 46
 }