nielsr HF staff commited on
Commit
4230724
1 Parent(s): deecae5

Update model

Browse files
preprocessor_config.json CHANGED
@@ -1,4 +1,5 @@
1
  {
 
2
  "do_normalize": true,
3
  "do_pad": true,
4
  "do_resize": true,
 
1
  {
2
+ "do_align_long_axis": false,
3
  "do_normalize": true,
4
  "do_pad": true,
5
  "do_resize": true,
tokenizer.json CHANGED
@@ -5,138 +5,138 @@
5
  "added_tokens": [
6
  {
7
  "id": 0,
 
8
  "content": "<s>",
9
  "single_word": false,
10
  "lstrip": false,
11
  "rstrip": false,
12
- "normalized": false,
13
- "special": true
14
  },
15
  {
16
  "id": 1,
 
17
  "content": "<pad>",
18
  "single_word": false,
19
  "lstrip": false,
20
  "rstrip": false,
21
- "normalized": false,
22
- "special": true
23
  },
24
  {
25
  "id": 2,
 
26
  "content": "</s>",
27
  "single_word": false,
28
  "lstrip": false,
29
  "rstrip": false,
30
- "normalized": false,
31
- "special": true
32
  },
33
  {
34
  "id": 3,
 
35
  "content": "<unk>",
36
  "single_word": false,
37
  "lstrip": false,
38
  "rstrip": false,
39
- "normalized": false,
40
- "special": true
41
  },
42
  {
43
  "id": 57521,
 
44
  "content": "<mask>",
45
  "single_word": false,
46
  "lstrip": true,
47
  "rstrip": false,
48
- "normalized": true,
49
- "special": true
50
  },
51
  {
52
  "id": 57522,
 
53
  "content": "<sep/>",
54
  "single_word": false,
55
  "lstrip": false,
56
  "rstrip": false,
57
- "normalized": true,
58
- "special": false
59
  },
60
  {
61
  "id": 57523,
 
62
  "content": "</s_answer>",
63
  "single_word": false,
64
  "lstrip": false,
65
  "rstrip": false,
66
- "normalized": false,
67
- "special": true
68
  },
69
  {
70
  "id": 57524,
 
71
  "content": "</s_question>",
72
  "single_word": false,
73
  "lstrip": false,
74
  "rstrip": false,
75
- "normalized": false,
76
- "special": true
77
  },
78
  {
79
  "id": 57525,
 
80
  "content": "<no/>",
81
  "single_word": false,
82
  "lstrip": false,
83
  "rstrip": false,
84
- "normalized": false,
85
- "special": true
86
  },
87
  {
88
  "id": 57526,
 
89
  "content": "<s_answer>",
90
  "single_word": false,
91
  "lstrip": false,
92
  "rstrip": false,
93
- "normalized": false,
94
- "special": true
95
  },
96
  {
97
  "id": 57527,
 
98
  "content": "<s_docvqa>",
99
  "single_word": false,
100
  "lstrip": false,
101
  "rstrip": false,
102
- "normalized": false,
103
- "special": true
104
  },
105
  {
106
  "id": 57528,
 
107
  "content": "<s_iitcdip>",
108
  "single_word": false,
109
  "lstrip": false,
110
  "rstrip": false,
111
- "normalized": false,
112
- "special": true
113
  },
114
  {
115
  "id": 57529,
 
116
  "content": "<s_question>",
117
  "single_word": false,
118
  "lstrip": false,
119
  "rstrip": false,
120
- "normalized": false,
121
- "special": true
122
  },
123
  {
124
  "id": 57530,
 
125
  "content": "<s_synthdog>",
126
  "single_word": false,
127
  "lstrip": false,
128
  "rstrip": false,
129
- "normalized": false,
130
- "special": true
131
  },
132
  {
133
  "id": 57531,
 
134
  "content": "<yes/>",
135
  "single_word": false,
136
  "lstrip": false,
137
  "rstrip": false,
138
- "normalized": false,
139
- "special": true
140
  }
141
  ],
142
  "normalizer": {
 
5
  "added_tokens": [
6
  {
7
  "id": 0,
8
+ "special": true,
9
  "content": "<s>",
10
  "single_word": false,
11
  "lstrip": false,
12
  "rstrip": false,
13
+ "normalized": false
 
14
  },
15
  {
16
  "id": 1,
17
+ "special": true,
18
  "content": "<pad>",
19
  "single_word": false,
20
  "lstrip": false,
21
  "rstrip": false,
22
+ "normalized": false
 
23
  },
24
  {
25
  "id": 2,
26
+ "special": true,
27
  "content": "</s>",
28
  "single_word": false,
29
  "lstrip": false,
30
  "rstrip": false,
31
+ "normalized": false
 
32
  },
33
  {
34
  "id": 3,
35
+ "special": true,
36
  "content": "<unk>",
37
  "single_word": false,
38
  "lstrip": false,
39
  "rstrip": false,
40
+ "normalized": false
 
41
  },
42
  {
43
  "id": 57521,
44
+ "special": true,
45
  "content": "<mask>",
46
  "single_word": false,
47
  "lstrip": true,
48
  "rstrip": false,
49
+ "normalized": true
 
50
  },
51
  {
52
  "id": 57522,
53
+ "special": false,
54
  "content": "<sep/>",
55
  "single_word": false,
56
  "lstrip": false,
57
  "rstrip": false,
58
+ "normalized": true
 
59
  },
60
  {
61
  "id": 57523,
62
+ "special": true,
63
  "content": "</s_answer>",
64
  "single_word": false,
65
  "lstrip": false,
66
  "rstrip": false,
67
+ "normalized": false
 
68
  },
69
  {
70
  "id": 57524,
71
+ "special": true,
72
  "content": "</s_question>",
73
  "single_word": false,
74
  "lstrip": false,
75
  "rstrip": false,
76
+ "normalized": false
 
77
  },
78
  {
79
  "id": 57525,
80
+ "special": true,
81
  "content": "<no/>",
82
  "single_word": false,
83
  "lstrip": false,
84
  "rstrip": false,
85
+ "normalized": false
 
86
  },
87
  {
88
  "id": 57526,
89
+ "special": true,
90
  "content": "<s_answer>",
91
  "single_word": false,
92
  "lstrip": false,
93
  "rstrip": false,
94
+ "normalized": false
 
95
  },
96
  {
97
  "id": 57527,
98
+ "special": true,
99
  "content": "<s_docvqa>",
100
  "single_word": false,
101
  "lstrip": false,
102
  "rstrip": false,
103
+ "normalized": false
 
104
  },
105
  {
106
  "id": 57528,
107
+ "special": true,
108
  "content": "<s_iitcdip>",
109
  "single_word": false,
110
  "lstrip": false,
111
  "rstrip": false,
112
+ "normalized": false
 
113
  },
114
  {
115
  "id": 57529,
116
+ "special": true,
117
  "content": "<s_question>",
118
  "single_word": false,
119
  "lstrip": false,
120
  "rstrip": false,
121
+ "normalized": false
 
122
  },
123
  {
124
  "id": 57530,
125
+ "special": true,
126
  "content": "<s_synthdog>",
127
  "single_word": false,
128
  "lstrip": false,
129
  "rstrip": false,
130
+ "normalized": false
 
131
  },
132
  {
133
  "id": 57531,
134
+ "special": true,
135
  "content": "<yes/>",
136
  "single_word": false,
137
  "lstrip": false,
138
  "rstrip": false,
139
+ "normalized": false
 
140
  }
141
  ],
142
  "normalizer": {
tokenizer_config.json CHANGED
@@ -13,6 +13,7 @@
13
  },
14
  "name_or_path": "naver-clova-ix/donut-base-finetuned-docvqa",
15
  "pad_token": "<pad>",
 
16
  "sep_token": "</s>",
17
  "sp_model_kwargs": {},
18
  "special_tokens_map_file": null,
 
13
  },
14
  "name_or_path": "naver-clova-ix/donut-base-finetuned-docvqa",
15
  "pad_token": "<pad>",
16
+ "processor_class": "DonutProcessor",
17
  "sep_token": "</s>",
18
  "sp_model_kwargs": {},
19
  "special_tokens_map_file": null,