kingabzpro commited on
Commit
035fe76
1 Parent(s): a04c2fe

add tokenizer

Browse files
runs/Jan21_12-53-12_f896580bb99d/1642769611.1221886/events.out.tfevents.1642769611.f896580bb99d.1392.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b84e3a3f9080892f9f6956135d63ebc1ae2ec7c966296970e8ecd82153b4ac68
3
+ size 4763
runs/Jan21_12-53-12_f896580bb99d/events.out.tfevents.1642769611.f896580bb99d.1392.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:37f1cb4f7c427b2486032a24a9a00c9e61695efbc24bb6b40fd0369b5236529a
3
+ size 4947
tokenizer_config.json CHANGED
@@ -1 +1 @@
1
- {"unk_token": "<unk>", "bos_token": "<s>", "eos_token": "</s>", "pad_token": "<pad>", "do_lower_case": false, "word_delimiter_token": "|", "tokenizer_class": "Wav2Vec2CTCTokenizer"}
 
1
+ {"unk_token": "<unk>", "bos_token": "<s>", "eos_token": "</s>", "pad_token": "<pad>", "do_lower_case": false, "word_delimiter_token": "|", "special_tokens_map_file": "/root/.cache/huggingface/transformers/898afae2faf3d20dd22149b80863bb5d448c67d9c9b821dd66753253a4d77c64.9d6cd81ef646692fb1c169a880161ea1cb95f49694f220aced9b704b457e51dd", "tokenizer_file": null, "name_or_path": "Harveenchadha/vakyansh-wav2vec2-urdu-urm-60", "tokenizer_class": "Wav2Vec2CTCTokenizer"}
vocab.json CHANGED
@@ -1 +1 @@
1
- {"ا": 1, "ب": 2, "ت": 3, "ث": 4, "ج": 5, "ح": 6, "خ": 7, "د": 8, "ذ": 9, "ر": 10, "ز": 11, "س": 12, "ش": 13, "ص": 14, "ض": 15, "ط": 16, "ظ": 17, "ع": 18, "غ": 19, "ف": 20, "ق": 21, "ل": 22, "م": 23, "ن": 24, "و": 25, "ٹ": 26, "پ": 27, "چ": 28, "ڈ": 29, "ڑ": 30, "ژ": 31, "ک": 32, "گ": 33, "ھ": 34, "ہ": 35, "ی": 36, "ے": 37, "|": 0, "<unk>": 38, "<pad>": 39}
 
1
+ {"<s>": 0, "<pad>": 1, "</s>": 2, "<unk>": 3, "|": 4, "ء": 5, "آ": 6, "ؤ": 7, "ئ": 8, "ا": 9, "ب": 10, "ت": 11, "ث": 12, "ج": 13, "ح": 14, "خ": 15, "د": 16, "ذ": 17, "ر": 18, "ز": 19, "س": 20, "ش": 21, "ص": 22, "ض": 23, "ط": 24, "ظ": 25, "ع": 26, "غ": 27, "ف": 28, "ق": 29, "ل": 30, "م": 31, "ن": 32, "و": 33, "ٹ": 34, "پ": 35, "چ": 36, "ڈ": 37, "ڑ": 38, "ژ": 39, "ک": 40, "گ": 41, "ں": 42, "ھ": 43, "ہ": 44, "ی": 45, "ے": 46, "ۓ": 47}