Upload processor
Browse files- added_tokens.json +5 -5
- tokenizer.json +0 -0
- tokenizer_config.json +12 -5
added_tokens.json
CHANGED
|
@@ -1,11 +1,11 @@
|
|
| 1 |
{
|
| 2 |
-
"</line>":
|
| 3 |
-
"</s_HW>":
|
| 4 |
"</s_P>": 57529,
|
| 5 |
"<MathOCR>": 57526,
|
| 6 |
-
"<line>":
|
| 7 |
-
"<s_HW>":
|
| 8 |
-
"<s_P>":
|
| 9 |
"<s_iitcdip>": 57523,
|
| 10 |
"<s_synthdog>": 57524,
|
| 11 |
"<sep/>": 57522
|
|
|
|
| 1 |
{
|
| 2 |
+
"</line>": 57528,
|
| 3 |
+
"</s_HW>": 57530,
|
| 4 |
"</s_P>": 57529,
|
| 5 |
"<MathOCR>": 57526,
|
| 6 |
+
"<line>": 57525,
|
| 7 |
+
"<s_HW>": 57531,
|
| 8 |
+
"<s_P>": 57527,
|
| 9 |
"<s_iitcdip>": 57523,
|
| 10 |
"<s_synthdog>": 57524,
|
| 11 |
"<sep/>": 57522
|
tokenizer.json
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
CHANGED
|
@@ -65,7 +65,7 @@
|
|
| 65 |
"special": true
|
| 66 |
},
|
| 67 |
"57525": {
|
| 68 |
-
"content": "<
|
| 69 |
"lstrip": false,
|
| 70 |
"normalized": true,
|
| 71 |
"rstrip": false,
|
|
@@ -81,7 +81,7 @@
|
|
| 81 |
"special": false
|
| 82 |
},
|
| 83 |
"57527": {
|
| 84 |
-
"content": "<
|
| 85 |
"lstrip": false,
|
| 86 |
"normalized": true,
|
| 87 |
"rstrip": false,
|
|
@@ -89,7 +89,7 @@
|
|
| 89 |
"special": false
|
| 90 |
},
|
| 91 |
"57528": {
|
| 92 |
-
"content": "<
|
| 93 |
"lstrip": false,
|
| 94 |
"normalized": true,
|
| 95 |
"rstrip": false,
|
|
@@ -105,7 +105,7 @@
|
|
| 105 |
"special": false
|
| 106 |
},
|
| 107 |
"57530": {
|
| 108 |
-
"content": "<
|
| 109 |
"lstrip": false,
|
| 110 |
"normalized": true,
|
| 111 |
"rstrip": false,
|
|
@@ -113,7 +113,7 @@
|
|
| 113 |
"special": false
|
| 114 |
},
|
| 115 |
"57531": {
|
| 116 |
-
"content": "<
|
| 117 |
"lstrip": false,
|
| 118 |
"normalized": true,
|
| 119 |
"rstrip": false,
|
|
@@ -130,11 +130,18 @@
|
|
| 130 |
"cls_token": "<s>",
|
| 131 |
"eos_token": "</s>",
|
| 132 |
"mask_token": "<mask>",
|
|
|
|
| 133 |
"model_max_length": 1000000000000000019884624838656,
|
|
|
|
| 134 |
"pad_token": "<pad>",
|
|
|
|
|
|
|
| 135 |
"processor_class": "DonutProcessor",
|
| 136 |
"sep_token": "</s>",
|
| 137 |
"sp_model_kwargs": {},
|
|
|
|
| 138 |
"tokenizer_class": "XLMRobertaTokenizer",
|
|
|
|
|
|
|
| 139 |
"unk_token": "<unk>"
|
| 140 |
}
|
|
|
|
| 65 |
"special": true
|
| 66 |
},
|
| 67 |
"57525": {
|
| 68 |
+
"content": "<line>",
|
| 69 |
"lstrip": false,
|
| 70 |
"normalized": true,
|
| 71 |
"rstrip": false,
|
|
|
|
| 81 |
"special": false
|
| 82 |
},
|
| 83 |
"57527": {
|
| 84 |
+
"content": "<s_P>",
|
| 85 |
"lstrip": false,
|
| 86 |
"normalized": true,
|
| 87 |
"rstrip": false,
|
|
|
|
| 89 |
"special": false
|
| 90 |
},
|
| 91 |
"57528": {
|
| 92 |
+
"content": "</line>",
|
| 93 |
"lstrip": false,
|
| 94 |
"normalized": true,
|
| 95 |
"rstrip": false,
|
|
|
|
| 105 |
"special": false
|
| 106 |
},
|
| 107 |
"57530": {
|
| 108 |
+
"content": "</s_HW>",
|
| 109 |
"lstrip": false,
|
| 110 |
"normalized": true,
|
| 111 |
"rstrip": false,
|
|
|
|
| 113 |
"special": false
|
| 114 |
},
|
| 115 |
"57531": {
|
| 116 |
+
"content": "<s_HW>",
|
| 117 |
"lstrip": false,
|
| 118 |
"normalized": true,
|
| 119 |
"rstrip": false,
|
|
|
|
| 130 |
"cls_token": "<s>",
|
| 131 |
"eos_token": "</s>",
|
| 132 |
"mask_token": "<mask>",
|
| 133 |
+
"max_length": 512,
|
| 134 |
"model_max_length": 1000000000000000019884624838656,
|
| 135 |
+
"pad_to_multiple_of": null,
|
| 136 |
"pad_token": "<pad>",
|
| 137 |
+
"pad_token_type_id": 0,
|
| 138 |
+
"padding_side": "right",
|
| 139 |
"processor_class": "DonutProcessor",
|
| 140 |
"sep_token": "</s>",
|
| 141 |
"sp_model_kwargs": {},
|
| 142 |
+
"stride": 0,
|
| 143 |
"tokenizer_class": "XLMRobertaTokenizer",
|
| 144 |
+
"truncation_side": "right",
|
| 145 |
+
"truncation_strategy": "longest_first",
|
| 146 |
"unk_token": "<unk>"
|
| 147 |
}
|