Upload tokenizer
Browse files- special_tokens_map.json +6 -0
- tokenizer_config.json +6 -0
- vocab.txt +75 -0
special_tokens_map.json
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"cls_token": "<cls>",
|
3 |
+
"mask_token": "<mask>",
|
4 |
+
"pad_token": "<pad>",
|
5 |
+
"unk_token": "<unk>"
|
6 |
+
}
|
tokenizer_config.json
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"clean_up_tokenization_spaces": true,
|
3 |
+
"eos_token": null,
|
4 |
+
"model_max_length": 2048,
|
5 |
+
"tokenizer_class": "EsmTokenizer"
|
6 |
+
}
|
vocab.txt
ADDED
@@ -0,0 +1,75 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
<unk>
|
2 |
+
<pad>
|
3 |
+
<mask>
|
4 |
+
<cls>
|
5 |
+
<eos>
|
6 |
+
<bos>
|
7 |
+
AAA
|
8 |
+
AAT
|
9 |
+
AAC
|
10 |
+
AAG
|
11 |
+
ATA
|
12 |
+
ATT
|
13 |
+
ATC
|
14 |
+
ATG
|
15 |
+
ACA
|
16 |
+
ACT
|
17 |
+
ACC
|
18 |
+
ACG
|
19 |
+
AGA
|
20 |
+
AGT
|
21 |
+
AGC
|
22 |
+
AGG
|
23 |
+
TAA
|
24 |
+
TAT
|
25 |
+
TAC
|
26 |
+
TAG
|
27 |
+
TTA
|
28 |
+
TTT
|
29 |
+
TTC
|
30 |
+
TTG
|
31 |
+
TCA
|
32 |
+
TCT
|
33 |
+
TCC
|
34 |
+
TCG
|
35 |
+
TGA
|
36 |
+
TGT
|
37 |
+
TGC
|
38 |
+
TGG
|
39 |
+
CAA
|
40 |
+
CAT
|
41 |
+
CAC
|
42 |
+
CAG
|
43 |
+
CTA
|
44 |
+
CTT
|
45 |
+
CTC
|
46 |
+
CTG
|
47 |
+
CCA
|
48 |
+
CCT
|
49 |
+
CCC
|
50 |
+
CCG
|
51 |
+
CGA
|
52 |
+
CGT
|
53 |
+
CGC
|
54 |
+
CGG
|
55 |
+
GAA
|
56 |
+
GAT
|
57 |
+
GAC
|
58 |
+
GAG
|
59 |
+
GTA
|
60 |
+
GTT
|
61 |
+
GTC
|
62 |
+
GTG
|
63 |
+
GCA
|
64 |
+
GCT
|
65 |
+
GCC
|
66 |
+
GCG
|
67 |
+
GGA
|
68 |
+
GGT
|
69 |
+
GGC
|
70 |
+
GGG
|
71 |
+
A
|
72 |
+
T
|
73 |
+
C
|
74 |
+
G
|
75 |
+
N
|