khaoulaoub
/

wav2vec2-darija

Automatic Speech Recognition

Generated from Trainer

Inference Endpoints

Model card Files Files and versions Community

khaoulaoub commited on Aug 15, 2023

Commit

c1aea73

•

1 Parent(s): 73ab1ca

Upload tokenizer

Files changed (2) hide show

tokenizer_config.json +3 -0
vocab.json +41 -40

tokenizer_config.json CHANGED Viewed

@@ -1,9 +1,12 @@
 {
   "bos_token": "<s>",
   "do_lower_case": false,
   "eos_token": "</s>",
   "pad_token": "[PAD]",
   "replace_word_delimiter_char": " ",
   "tokenizer_class": "Wav2Vec2CTCTokenizer",
   "unk_token": "[UNK]",
   "word_delimiter_token": "|"

 {
   "bos_token": "<s>",
+  "clean_up_tokenization_spaces": true,
   "do_lower_case": false,
   "eos_token": "</s>",
+  "model_max_length": 1000000000000000019884624838656,
   "pad_token": "[PAD]",
   "replace_word_delimiter_char": " ",
+  "target_lang": null,
   "tokenizer_class": "Wav2Vec2CTCTokenizer",
   "unk_token": "[UNK]",
   "word_delimiter_token": "|"

vocab.json CHANGED Viewed

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

 {
+  "/": 11,
+  "[PAD]": 40,
+  "[UNK]": 39,
+  "|": 21,
+  "ء": 10,
+  "آ": 36,
+  "أ": 22,
+  "ؤ": 2,
+  "ئ": 16,
+  "ا": 28,
+  "ب": 29,
+  "ة": 24,
+  "ت": 0,
+  "ث": 25,
+  "ج": 38,
+  "ح": 20,
+  "خ": 19,
+  "د": 30,
+  "ذ": 4,
+  "ر": 27,
+  "ز": 37,
+  "س": 32,
+  "ش": 7,
+  "ص": 14,
+  "ض": 12,
+  "ط": 18,
+  "ظ": 1,
+  "ع": 8,
+  "غ": 31,
+  "ف": 5,
+  "ق": 34,
+  "ك": 17,
+  "ل": 9,
+  "م": 35,
+  "ن": 26,
+  "ه": 13,
+  "و": 23,
+  "ى": 6,
+  "ي": 15,
+  "َ": 3,
+  "ْ": 33
 }