mehedihasanbijoy
/

w2v-bert-2.0-mongolian-colab-CV16.0

Automatic Speech Recognition

Generated from Trainer

Inference Endpoints

Model card Files Files and versions Metrics Training metrics Community

mehedihasanbijoy commited on Sep 24

Commit

e220c35

•

1 Parent(s): 82c9e44

Upload processor

Files changed (3) hide show

added_tokens.json +2 -2
tokenizer_config.json +4 -4
vocab.json +61 -36

added_tokens.json CHANGED Viewed

@@ -1,4 +1,4 @@
 {
-  "</s>": 38,
-  "<s>": 37
 }

 {
+  "</s>": 63,
+  "<s>": 62
 }

tokenizer_config.json CHANGED Viewed

@@ -1,6 +1,6 @@
 {
   "added_tokens_decoder": {
-    "35": {
       "content": "[UNK]",
       "lstrip": true,
       "normalized": false,
@@ -8,7 +8,7 @@
       "single_word": false,
       "special": false
     },
-    "36": {
       "content": "[PAD]",
       "lstrip": true,
       "normalized": false,
@@ -16,7 +16,7 @@
       "single_word": false,
       "special": false
     },
-    "37": {
       "content": "<s>",
       "lstrip": false,
       "normalized": false,
@@ -24,7 +24,7 @@
       "single_word": false,
       "special": true
     },
-    "38": {
       "content": "</s>",
       "lstrip": false,
       "normalized": false,

 {
   "added_tokens_decoder": {
+    "60": {
       "content": "[UNK]",
       "lstrip": true,
       "normalized": false,
       "single_word": false,
       "special": false
     },
+    "61": {
       "content": "[PAD]",
       "lstrip": true,
       "normalized": false,
       "single_word": false,
       "special": false
     },
+    "62": {
       "content": "<s>",
       "lstrip": false,
       "normalized": false,
       "single_word": false,
       "special": true
     },
+    "63": {
       "content": "</s>",
       "lstrip": false,
       "normalized": false,

vocab.json CHANGED Viewed

@@ -1,39 +1,64 @@
 {
-  "[PAD]": 36,
-  "[UNK]": 35,
   "|": 0,
-  "а": 1,
-  "б": 2,
-  "в": 3,
-  "г": 4,
-  "д": 5,
-  "е": 6,
-  "ж": 7,
-  "з": 8,
-  "и": 9,
-  "й": 10,
-  "к": 11,
-  "л": 12,
-  "м": 13,
-  "н": 14,
-  "о": 15,
-  "п": 16,
-  "р": 17,
-  "с": 18,
-  "т": 19,
-  "у": 20,
-  "ф": 21,
-  "х": 22,
-  "ц": 23,
-  "ч": 24,
-  "ш": 25,
-  "ъ": 26,
-  "ы": 27,
-  "ь": 28,
-  "э": 29,
-  "ю": 30,
-  "я": 31,
-  "ё": 32,
-  "ү": 33,
-  "ө": 34
 }

 {
+  "[PAD]": 61,
+  "[UNK]": 60,
   "|": 0,
+  "ঁ": 1,
+  "ং": 2,
+  "ঃ": 3,
+  "অ": 4,
+  "আ": 5,
+  "ই": 6,
+  "ঈ": 7,
+  "উ": 8,
+  "এ": 9,
+  "ঐ": 10,
+  "ও": 11,
+  "ঔ": 12,
+  "ক": 13,
+  "খ": 14,
+  "গ": 15,
+  "ঘ": 16,
+  "ঙ": 17,
+  "চ": 18,
+  "ছ": 19,
+  "জ": 20,
+  "ঝ": 21,
+  "ঞ": 22,
+  "ট": 23,
+  "ঠ": 24,
+  "ড": 25,
+  "ঢ": 26,
+  "ণ": 27,
+  "ত": 28,
+  "থ": 29,
+  "দ": 30,
+  "ধ": 31,
+  "ন": 32,
+  "প": 33,
+  "ফ": 34,
+  "ব": 35,
+  "ভ": 36,
+  "ম": 37,
+  "য": 38,
+  "র": 39,
+  "ল": 40,
+  "শ": 41,
+  "ষ": 42,
+  "স": 43,
+  "হ": 44,
+  "়": 45,
+  "া": 46,
+  "ি": 47,
+  "ী": 48,
+  "ু": 49,
+  "ূ": 50,
+  "ৃ": 51,
+  "ে": 52,
+  "ৈ": 53,
+  "ো": 54,
+  "ৌ": 55,
+  "্": 56,
+  "ৎ": 57,
+  "য়": 58,
+  "’": 59
 }