mgtoxd commited on
Commit
21b55b1
·
1 Parent(s): bc73c3f

add tokenizer

Browse files
runs/Jun18_16-32-36_container-3d5811a0ae-620c5fad/1655543263.9791784/events.out.tfevents.1655543263.container-3d5811a0ae-620c5fad.86026.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e2e75966e94df8254817faa2879ad07dadcc8295363d40d7a214a53b7aa5c1fe
3
+ size 5380
runs/Jun18_16-32-36_container-3d5811a0ae-620c5fad/events.out.tfevents.1655543263.container-3d5811a0ae-620c5fad.86026.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2d05bb919025a1c4f8a17626efa0a9ebb335a0a39c5a33c99232470d8dafff47
3
+ size 5335
special_tokens_map.json CHANGED
@@ -1 +1,6 @@
1
- {"bos_token": "<s>", "eos_token": "</s>", "unk_token": "[UNK]", "pad_token": "[PAD]"}
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "eos_token": "</s>",
4
+ "pad_token": "[PAD]",
5
+ "unk_token": "[UNK]"
6
+ }
tokenizer_config.json CHANGED
@@ -1 +1,10 @@
1
- {"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"}
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "do_lower_case": false,
4
+ "eos_token": "</s>",
5
+ "pad_token": "[PAD]",
6
+ "replace_word_delimiter_char": " ",
7
+ "tokenizer_class": "Wav2Vec2CTCTokenizer",
8
+ "unk_token": "[UNK]",
9
+ "word_delimiter_token": "|"
10
+ }
vocab.json CHANGED
@@ -1 +1,37 @@
1
- {"q": 0, "i": 1, "a": 2, "2": 3, "y": 4, "f": 5, "x": 6, "c": 7, "h": 8, "l": 9, "r": 10, "j": 11, "p": 12, "s": 13, "4": 14, "e": 15, "z": 16, "m": 18, "3": 19, "d": 20, "n": 21, "u": 22, "v": 23, "6": 24, "t": 25, "o": 26, "w": 27, "b": 28, "k": 29, "g": 30, "5": 31, "'": 32, "|": 17, "[UNK]": 33, "[PAD]": 34}
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "'": 8,
3
+ "2": 14,
4
+ "3": 15,
5
+ "4": 12,
6
+ "5": 25,
7
+ "6": 23,
8
+ "[PAD]": 34,
9
+ "[UNK]": 33,
10
+ "a": 29,
11
+ "b": 9,
12
+ "c": 21,
13
+ "d": 22,
14
+ "e": 17,
15
+ "f": 13,
16
+ "g": 11,
17
+ "h": 18,
18
+ "i": 16,
19
+ "j": 26,
20
+ "k": 10,
21
+ "l": 5,
22
+ "m": 31,
23
+ "n": 3,
24
+ "o": 6,
25
+ "p": 28,
26
+ "q": 7,
27
+ "r": 4,
28
+ "s": 1,
29
+ "t": 0,
30
+ "u": 32,
31
+ "v": 2,
32
+ "w": 30,
33
+ "x": 19,
34
+ "y": 27,
35
+ "z": 24,
36
+ "|": 20
37
+ }