Commit
•
e9bf05f
1
Parent(s):
a076edc
add tokenizer
Browse files- tokenizer_config.json +1 -1
- vocab.json +29 -29
tokenizer_config.json
CHANGED
@@ -2,7 +2,7 @@
|
|
2 |
"bos_token": "<s>",
|
3 |
"do_lower_case": false,
|
4 |
"eos_token": "</s>",
|
5 |
-
"name_or_path": "/tmp/
|
6 |
"pad_token": "<pad>",
|
7 |
"replace_word_delimiter_char": " ",
|
8 |
"special_tokens_map_file": null,
|
|
|
2 |
"bos_token": "<s>",
|
3 |
"do_lower_case": false,
|
4 |
"eos_token": "</s>",
|
5 |
+
"name_or_path": "/tmp/tmpi8f1bi5v",
|
6 |
"pad_token": "<pad>",
|
7 |
"replace_word_delimiter_char": " ",
|
8 |
"special_tokens_map_file": null,
|
vocab.json
CHANGED
@@ -1,36 +1,36 @@
|
|
1 |
{
|
2 |
-
"'":
|
3 |
-
"-":
|
4 |
-
"1":
|
5 |
"</s>": 2,
|
6 |
"<pad>": 0,
|
7 |
"<s>": 1,
|
8 |
"<unk>": 3,
|
9 |
-
"a":
|
10 |
-
"b":
|
11 |
-
"c":
|
12 |
-
"d":
|
13 |
-
"e":
|
14 |
-
"f":
|
15 |
-
"g":
|
16 |
-
"h":
|
17 |
"i": 26,
|
18 |
-
"j":
|
19 |
-
"k":
|
20 |
-
"l":
|
21 |
-
"m":
|
22 |
-
"n":
|
23 |
-
"o":
|
24 |
-
"p":
|
25 |
-
"q":
|
26 |
-
"r":
|
27 |
-
"s":
|
28 |
-
"t":
|
29 |
-
"u":
|
30 |
-
"v":
|
31 |
-
"w":
|
32 |
-
"x":
|
33 |
-
"y":
|
34 |
-
"z":
|
35 |
-
"|":
|
36 |
}
|
|
|
1 |
{
|
2 |
+
"'": 29,
|
3 |
+
"-": 25,
|
4 |
+
"1": 18,
|
5 |
"</s>": 2,
|
6 |
"<pad>": 0,
|
7 |
"<s>": 1,
|
8 |
"<unk>": 3,
|
9 |
+
"a": 23,
|
10 |
+
"b": 33,
|
11 |
+
"c": 22,
|
12 |
+
"d": 21,
|
13 |
+
"e": 14,
|
14 |
+
"f": 11,
|
15 |
+
"g": 6,
|
16 |
+
"h": 9,
|
17 |
"i": 26,
|
18 |
+
"j": 4,
|
19 |
+
"k": 15,
|
20 |
+
"l": 17,
|
21 |
+
"m": 8,
|
22 |
+
"n": 30,
|
23 |
+
"o": 20,
|
24 |
+
"p": 19,
|
25 |
+
"q": 12,
|
26 |
+
"r": 13,
|
27 |
+
"s": 10,
|
28 |
+
"t": 24,
|
29 |
+
"u": 7,
|
30 |
+
"v": 31,
|
31 |
+
"w": 32,
|
32 |
+
"x": 28,
|
33 |
+
"y": 16,
|
34 |
+
"z": 27,
|
35 |
+
"|": 5
|
36 |
}
|