tomfunk commited on
Commit
78683ba
1 Parent(s): 7db5de6

Training in progress, epoch 0

Browse files
added_tokens.json ADDED
@@ -0,0 +1,102 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "</s_income-w2-allocated-tips>": 57525,
3
+ "</s_income-w2-dependent-care-benefits>": 57527,
4
+ "</s_income-w2-employee-ssn>": 57529,
5
+ "</s_income-w2-employer-address-multiline>": 57533,
6
+ "</s_income-w2-employer>": 57531,
7
+ "</s_income-w2-federal-id>": 57535,
8
+ "</s_income-w2-federal-income-tax-withheld>": 57537,
9
+ "</s_income-w2-form-year>": 57539,
10
+ "</s_income-w2-fourteen-a-amount>": 57541,
11
+ "</s_income-w2-fourteen-a-code>": 57543,
12
+ "</s_income-w2-fourteen-b-amount>": 57545,
13
+ "</s_income-w2-fourteen-b-code>": 57547,
14
+ "</s_income-w2-fourteen-c-amount>": 57549,
15
+ "</s_income-w2-fourteen-c-code>": 57551,
16
+ "</s_income-w2-fourteen-d-amount>": 57553,
17
+ "</s_income-w2-fourteen-d-code>": 57555,
18
+ "</s_income-w2-local-name2>": 57559,
19
+ "</s_income-w2-local-name>": 57557,
20
+ "</s_income-w2-local-taxes2>": 57563,
21
+ "</s_income-w2-local-taxes>": 57561,
22
+ "</s_income-w2-local-wages2>": 57567,
23
+ "</s_income-w2-local-wages>": 57565,
24
+ "</s_income-w2-medicare-tax-withheld>": 57569,
25
+ "</s_income-w2-medicare-wages>": 57571,
26
+ "</s_income-w2-nonqualified-plans>": 57573,
27
+ "</s_income-w2-retirement-plan>": 57575,
28
+ "</s_income-w2-sick-pay>": 57577,
29
+ "</s_income-w2-social-security-tax-withheld>": 57579,
30
+ "</s_income-w2-social-security-tips>": 57581,
31
+ "</s_income-w2-social-security-wages>": 57583,
32
+ "</s_income-w2-state-employer-id2>": 57589,
33
+ "</s_income-w2-state-employer-id>": 57587,
34
+ "</s_income-w2-state-income-tax2>": 57593,
35
+ "</s_income-w2-state-income-tax>": 57591,
36
+ "</s_income-w2-state-wages2>": 57597,
37
+ "</s_income-w2-state-wages>": 57595,
38
+ "</s_income-w2-state2>": 57599,
39
+ "</s_income-w2-state>": 57585,
40
+ "</s_income-w2-statutory-employee>": 57601,
41
+ "</s_income-w2-twelve-a-partOne>": 57603,
42
+ "</s_income-w2-twelve-a-partTwo>": 57605,
43
+ "</s_income-w2-twelve-b-partOne>": 57607,
44
+ "</s_income-w2-twelve-b-partTwo>": 57609,
45
+ "</s_income-w2-twelve-c-partOne>": 57611,
46
+ "</s_income-w2-twelve-c-partTwo>": 57613,
47
+ "</s_income-w2-twelve-d-partOne>": 57615,
48
+ "</s_income-w2-twelve-d-partTwo>": 57617,
49
+ "</s_income-w2-wages>": 57619,
50
+ "<s_iitcdip>": 57523,
51
+ "<s_income-w2-allocated-tips>": 57526,
52
+ "<s_income-w2-dependent-care-benefits>": 57528,
53
+ "<s_income-w2-employee-ssn>": 57530,
54
+ "<s_income-w2-employer-address-multiline>": 57534,
55
+ "<s_income-w2-employer>": 57532,
56
+ "<s_income-w2-federal-id>": 57536,
57
+ "<s_income-w2-federal-income-tax-withheld>": 57538,
58
+ "<s_income-w2-form-year>": 57540,
59
+ "<s_income-w2-fourteen-a-amount>": 57542,
60
+ "<s_income-w2-fourteen-a-code>": 57544,
61
+ "<s_income-w2-fourteen-b-amount>": 57546,
62
+ "<s_income-w2-fourteen-b-code>": 57548,
63
+ "<s_income-w2-fourteen-c-amount>": 57550,
64
+ "<s_income-w2-fourteen-c-code>": 57552,
65
+ "<s_income-w2-fourteen-d-amount>": 57554,
66
+ "<s_income-w2-fourteen-d-code>": 57556,
67
+ "<s_income-w2-local-name2>": 57560,
68
+ "<s_income-w2-local-name>": 57558,
69
+ "<s_income-w2-local-taxes2>": 57564,
70
+ "<s_income-w2-local-taxes>": 57562,
71
+ "<s_income-w2-local-wages2>": 57568,
72
+ "<s_income-w2-local-wages>": 57566,
73
+ "<s_income-w2-medicare-tax-withheld>": 57570,
74
+ "<s_income-w2-medicare-wages>": 57572,
75
+ "<s_income-w2-nonqualified-plans>": 57574,
76
+ "<s_income-w2-retirement-plan>": 57576,
77
+ "<s_income-w2-sick-pay>": 57578,
78
+ "<s_income-w2-social-security-tax-withheld>": 57580,
79
+ "<s_income-w2-social-security-tips>": 57582,
80
+ "<s_income-w2-social-security-wages>": 57584,
81
+ "<s_income-w2-state-employer-id2>": 57590,
82
+ "<s_income-w2-state-employer-id>": 57588,
83
+ "<s_income-w2-state-income-tax2>": 57594,
84
+ "<s_income-w2-state-income-tax>": 57592,
85
+ "<s_income-w2-state-wages2>": 57598,
86
+ "<s_income-w2-state-wages>": 57596,
87
+ "<s_income-w2-state2>": 57600,
88
+ "<s_income-w2-state>": 57586,
89
+ "<s_income-w2-statutory-employee>": 57602,
90
+ "<s_income-w2-twelve-a-partOne>": 57604,
91
+ "<s_income-w2-twelve-a-partTwo>": 57606,
92
+ "<s_income-w2-twelve-b-partOne>": 57608,
93
+ "<s_income-w2-twelve-b-partTwo>": 57610,
94
+ "<s_income-w2-twelve-c-partOne>": 57612,
95
+ "<s_income-w2-twelve-c-partTwo>": 57614,
96
+ "<s_income-w2-twelve-d-partOne>": 57616,
97
+ "<s_income-w2-twelve-d-partTwo>": 57618,
98
+ "<s_income-w2-wages>": 57620,
99
+ "<s_income-w2>": 57621,
100
+ "<s_synthdog>": 57524,
101
+ "<sep/>": 57522
102
+ }
sentencepiece.bpe.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cb9e3dce4c326195d08fc3dd0f7e2eee1da8595c847bf4c1a9c78b7a82d47e2d
3
+ size 1296245
special_tokens_map.json ADDED
@@ -0,0 +1,18 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "additional_special_tokens": [
3
+ "<s_income-w2>"
4
+ ],
5
+ "bos_token": "<s>",
6
+ "cls_token": "<s>",
7
+ "eos_token": "</s>",
8
+ "mask_token": {
9
+ "content": "<mask>",
10
+ "lstrip": true,
11
+ "normalized": true,
12
+ "rstrip": false,
13
+ "single_word": false
14
+ },
15
+ "pad_token": "<pad>",
16
+ "sep_token": "</s>",
17
+ "unk_token": "<unk>"
18
+ }
tokenizer_config.json ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "cls_token": "<s>",
4
+ "eos_token": "</s>",
5
+ "from_slow": true,
6
+ "mask_token": {
7
+ "__type": "AddedToken",
8
+ "content": "<mask>",
9
+ "lstrip": true,
10
+ "normalized": true,
11
+ "rstrip": false,
12
+ "single_word": false
13
+ },
14
+ "model_max_length": 1000000000000000019884624838656,
15
+ "name_or_path": "naver-clova-ix/donut-base",
16
+ "pad_token": "<pad>",
17
+ "processor_class": "DonutProcessor",
18
+ "sep_token": "</s>",
19
+ "sp_model_kwargs": {},
20
+ "special_tokens_map_file": null,
21
+ "tokenizer_class": "XLMRobertaTokenizer",
22
+ "unk_token": "<unk>"
23
+ }