Upload tokenizer

special_tokens_map.json

@@ -1 +1,6 @@
-{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "[UNK]", "pad_token": "[PAD]"}
+{
+  "bos_token": "<s>",
+  "eos_token": "</s>",
+  "pad_token": "[PAD]",
+  "unk_token": "[UNK]"
+}

tokenizer_config.json

@@ -1 +1,10 @@
-{"unk_token": "[UNK]", "bos_token": "<s>", "eos_token": "</s>", "pad_token": "[PAD]", "do_lower_case": false, "word_delimiter_token": "|", "replace_word_delimiter_char": " ", "tokenizer_class": "Wav2Vec2CTCTokenizer"}
+{
+  "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": "|"
+}

vocab.json

@@ -1 +1,53 @@
-{"ㄱ": 1, "ㄲ": 2, "ㄳ": 3, "ㄴ": 4, "ㄵ": 5, "ㄶ": 6, "ㄷ": 7, "ㄸ": 8, "ㄹ": 9, "ㄺ": 10, "ㄻ": 11, "ㄼ": 12, "ㅀ": 13, "ㅁ": 14, "ㅂ": 15, "ㅃ": 16, "ㅄ": 17, "ㅅ": 18, "ㅆ": 19, "ㅇ": 20, "ㅈ": 21, "ㅉ": 22, "ㅊ": 23, "ㅋ": 24, "ㅌ": 25, "ㅍ": 26, "ㅎ": 27, "ㅏ": 28, "ㅐ": 29, "ㅑ": 30, "ㅒ": 31, "ㅓ": 32, "ㅔ": 33, "ㅕ": 34, "ㅖ": 35, "ㅗ": 36, "ㅘ": 37, "ㅙ": 38, "ㅚ": 39, "ㅛ": 40, "ㅜ": 41, "ㅝ": 42, "ㅞ": 43, "ㅟ": 44, "ㅠ": 45, "ㅡ": 46, "ㅢ": 47, "ㅣ": 48, "|": 0, "[UNK]": 49, "[PAD]": 50}
+{
+  "[PAD]": 50,
+  "[UNK]": 49,
+  "|": 0,
+  "ㄱ": 1,
+  "ㄲ": 2,
+  "ㄳ": 3,
+  "ㄴ": 4,
+  "ㄵ": 5,
+  "ㄶ": 6,
+  "ㄷ": 7,
+  "ㄸ": 8,
+  "ㄹ": 9,
+  "ㄺ": 10,
+  "ㄻ": 11,
+  "ㄼ": 12,
+  "ㅀ": 13,
+  "ㅁ": 14,
+  "ㅂ": 15,
+  "ㅃ": 16,
+  "ㅄ": 17,
+  "ㅅ": 18,
+  "ㅆ": 19,
+  "ㅇ": 20,
+  "ㅈ": 21,
+  "ㅉ": 22,
+  "ㅊ": 23,
+  "ㅋ": 24,
+  "ㅌ": 25,
+  "ㅍ": 26,
+  "ㅎ": 27,
+  "ㅏ": 28,
+  "ㅐ": 29,
+  "ㅑ": 30,
+  "ㅒ": 31,
+  "ㅓ": 32,
+  "ㅔ": 33,
+  "ㅕ": 34,
+  "ㅖ": 35,
+  "ㅗ": 36,
+  "ㅘ": 37,
+  "ㅙ": 38,
+  "ㅚ": 39,
+  "ㅛ": 40,
+  "ㅜ": 41,
+  "ㅝ": 42,
+  "ㅞ": 43,
+  "ㅟ": 44,
+  "ㅠ": 45,
+  "ㅡ": 46,
+  "ㅢ": 47,
+  "ㅣ": 48
+}