Training done
Browse files- added_tokens.json +8 -10
- tokenizer.json +7 -25
- tokenizer_config.json +7 -23
added_tokens.json
CHANGED
|
@@ -1,18 +1,16 @@
|
|
| 1 |
{
|
| 2 |
-
"</s_address>":
|
| 3 |
-
"</s_company>":
|
| 4 |
-
"</s_date>":
|
| 5 |
-
"</s_gt_parse>": 57529,
|
| 6 |
"</s_nom>": 57526,
|
| 7 |
-
"</s_total>":
|
| 8 |
-
"<s_address>":
|
| 9 |
-
"<s_company>":
|
| 10 |
"<s_cord-v2>": 57527,
|
| 11 |
-
"<s_date>":
|
| 12 |
-
"<s_gt_parse>": 57528,
|
| 13 |
"<s_iitcdip>": 57523,
|
| 14 |
"<s_nom>": 57525,
|
| 15 |
"<s_synthdog>": 57524,
|
| 16 |
-
"<s_total>":
|
| 17 |
"<sep/>": 57522
|
| 18 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"</s_address>": 57535,
|
| 3 |
+
"</s_company>": 57533,
|
| 4 |
+
"</s_date>": 57531,
|
|
|
|
| 5 |
"</s_nom>": 57526,
|
| 6 |
+
"</s_total>": 57529,
|
| 7 |
+
"<s_address>": 57534,
|
| 8 |
+
"<s_company>": 57532,
|
| 9 |
"<s_cord-v2>": 57527,
|
| 10 |
+
"<s_date>": 57530,
|
|
|
|
| 11 |
"<s_iitcdip>": 57523,
|
| 12 |
"<s_nom>": 57525,
|
| 13 |
"<s_synthdog>": 57524,
|
| 14 |
+
"<s_total>": 57528,
|
| 15 |
"<sep/>": 57522
|
| 16 |
}
|
tokenizer.json
CHANGED
|
@@ -104,24 +104,6 @@
|
|
| 104 |
},
|
| 105 |
{
|
| 106 |
"id": 57528,
|
| 107 |
-
"content": "<s_gt_parse>",
|
| 108 |
-
"single_word": false,
|
| 109 |
-
"lstrip": false,
|
| 110 |
-
"rstrip": false,
|
| 111 |
-
"normalized": true,
|
| 112 |
-
"special": false
|
| 113 |
-
},
|
| 114 |
-
{
|
| 115 |
-
"id": 57529,
|
| 116 |
-
"content": "</s_gt_parse>",
|
| 117 |
-
"single_word": false,
|
| 118 |
-
"lstrip": false,
|
| 119 |
-
"rstrip": false,
|
| 120 |
-
"normalized": true,
|
| 121 |
-
"special": false
|
| 122 |
-
},
|
| 123 |
-
{
|
| 124 |
-
"id": 57530,
|
| 125 |
"content": "<s_total>",
|
| 126 |
"single_word": false,
|
| 127 |
"lstrip": false,
|
|
@@ -130,7 +112,7 @@
|
|
| 130 |
"special": false
|
| 131 |
},
|
| 132 |
{
|
| 133 |
-
"id":
|
| 134 |
"content": "</s_total>",
|
| 135 |
"single_word": false,
|
| 136 |
"lstrip": false,
|
|
@@ -139,7 +121,7 @@
|
|
| 139 |
"special": false
|
| 140 |
},
|
| 141 |
{
|
| 142 |
-
"id":
|
| 143 |
"content": "<s_date>",
|
| 144 |
"single_word": false,
|
| 145 |
"lstrip": false,
|
|
@@ -148,7 +130,7 @@
|
|
| 148 |
"special": false
|
| 149 |
},
|
| 150 |
{
|
| 151 |
-
"id":
|
| 152 |
"content": "</s_date>",
|
| 153 |
"single_word": false,
|
| 154 |
"lstrip": false,
|
|
@@ -157,7 +139,7 @@
|
|
| 157 |
"special": false
|
| 158 |
},
|
| 159 |
{
|
| 160 |
-
"id":
|
| 161 |
"content": "<s_company>",
|
| 162 |
"single_word": false,
|
| 163 |
"lstrip": false,
|
|
@@ -166,7 +148,7 @@
|
|
| 166 |
"special": false
|
| 167 |
},
|
| 168 |
{
|
| 169 |
-
"id":
|
| 170 |
"content": "</s_company>",
|
| 171 |
"single_word": false,
|
| 172 |
"lstrip": false,
|
|
@@ -175,7 +157,7 @@
|
|
| 175 |
"special": false
|
| 176 |
},
|
| 177 |
{
|
| 178 |
-
"id":
|
| 179 |
"content": "<s_address>",
|
| 180 |
"single_word": false,
|
| 181 |
"lstrip": false,
|
|
@@ -184,7 +166,7 @@
|
|
| 184 |
"special": false
|
| 185 |
},
|
| 186 |
{
|
| 187 |
-
"id":
|
| 188 |
"content": "</s_address>",
|
| 189 |
"single_word": false,
|
| 190 |
"lstrip": false,
|
|
|
|
| 104 |
},
|
| 105 |
{
|
| 106 |
"id": 57528,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 107 |
"content": "<s_total>",
|
| 108 |
"single_word": false,
|
| 109 |
"lstrip": false,
|
|
|
|
| 112 |
"special": false
|
| 113 |
},
|
| 114 |
{
|
| 115 |
+
"id": 57529,
|
| 116 |
"content": "</s_total>",
|
| 117 |
"single_word": false,
|
| 118 |
"lstrip": false,
|
|
|
|
| 121 |
"special": false
|
| 122 |
},
|
| 123 |
{
|
| 124 |
+
"id": 57530,
|
| 125 |
"content": "<s_date>",
|
| 126 |
"single_word": false,
|
| 127 |
"lstrip": false,
|
|
|
|
| 130 |
"special": false
|
| 131 |
},
|
| 132 |
{
|
| 133 |
+
"id": 57531,
|
| 134 |
"content": "</s_date>",
|
| 135 |
"single_word": false,
|
| 136 |
"lstrip": false,
|
|
|
|
| 139 |
"special": false
|
| 140 |
},
|
| 141 |
{
|
| 142 |
+
"id": 57532,
|
| 143 |
"content": "<s_company>",
|
| 144 |
"single_word": false,
|
| 145 |
"lstrip": false,
|
|
|
|
| 148 |
"special": false
|
| 149 |
},
|
| 150 |
{
|
| 151 |
+
"id": 57533,
|
| 152 |
"content": "</s_company>",
|
| 153 |
"single_word": false,
|
| 154 |
"lstrip": false,
|
|
|
|
| 157 |
"special": false
|
| 158 |
},
|
| 159 |
{
|
| 160 |
+
"id": 57534,
|
| 161 |
"content": "<s_address>",
|
| 162 |
"single_word": false,
|
| 163 |
"lstrip": false,
|
|
|
|
| 166 |
"special": false
|
| 167 |
},
|
| 168 |
{
|
| 169 |
+
"id": 57535,
|
| 170 |
"content": "</s_address>",
|
| 171 |
"single_word": false,
|
| 172 |
"lstrip": false,
|
tokenizer_config.json
CHANGED
|
@@ -89,22 +89,6 @@
|
|
| 89 |
"special": false
|
| 90 |
},
|
| 91 |
"57528": {
|
| 92 |
-
"content": "<s_gt_parse>",
|
| 93 |
-
"lstrip": false,
|
| 94 |
-
"normalized": true,
|
| 95 |
-
"rstrip": false,
|
| 96 |
-
"single_word": false,
|
| 97 |
-
"special": false
|
| 98 |
-
},
|
| 99 |
-
"57529": {
|
| 100 |
-
"content": "</s_gt_parse>",
|
| 101 |
-
"lstrip": false,
|
| 102 |
-
"normalized": true,
|
| 103 |
-
"rstrip": false,
|
| 104 |
-
"single_word": false,
|
| 105 |
-
"special": false
|
| 106 |
-
},
|
| 107 |
-
"57530": {
|
| 108 |
"content": "<s_total>",
|
| 109 |
"lstrip": false,
|
| 110 |
"normalized": true,
|
|
@@ -112,7 +96,7 @@
|
|
| 112 |
"single_word": false,
|
| 113 |
"special": false
|
| 114 |
},
|
| 115 |
-
"
|
| 116 |
"content": "</s_total>",
|
| 117 |
"lstrip": false,
|
| 118 |
"normalized": true,
|
|
@@ -120,7 +104,7 @@
|
|
| 120 |
"single_word": false,
|
| 121 |
"special": false
|
| 122 |
},
|
| 123 |
-
"
|
| 124 |
"content": "<s_date>",
|
| 125 |
"lstrip": false,
|
| 126 |
"normalized": true,
|
|
@@ -128,7 +112,7 @@
|
|
| 128 |
"single_word": false,
|
| 129 |
"special": false
|
| 130 |
},
|
| 131 |
-
"
|
| 132 |
"content": "</s_date>",
|
| 133 |
"lstrip": false,
|
| 134 |
"normalized": true,
|
|
@@ -136,7 +120,7 @@
|
|
| 136 |
"single_word": false,
|
| 137 |
"special": false
|
| 138 |
},
|
| 139 |
-
"
|
| 140 |
"content": "<s_company>",
|
| 141 |
"lstrip": false,
|
| 142 |
"normalized": true,
|
|
@@ -144,7 +128,7 @@
|
|
| 144 |
"single_word": false,
|
| 145 |
"special": false
|
| 146 |
},
|
| 147 |
-
"
|
| 148 |
"content": "</s_company>",
|
| 149 |
"lstrip": false,
|
| 150 |
"normalized": true,
|
|
@@ -152,7 +136,7 @@
|
|
| 152 |
"single_word": false,
|
| 153 |
"special": false
|
| 154 |
},
|
| 155 |
-
"
|
| 156 |
"content": "<s_address>",
|
| 157 |
"lstrip": false,
|
| 158 |
"normalized": true,
|
|
@@ -160,7 +144,7 @@
|
|
| 160 |
"single_word": false,
|
| 161 |
"special": false
|
| 162 |
},
|
| 163 |
-
"
|
| 164 |
"content": "</s_address>",
|
| 165 |
"lstrip": false,
|
| 166 |
"normalized": true,
|
|
|
|
| 89 |
"special": false
|
| 90 |
},
|
| 91 |
"57528": {
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 92 |
"content": "<s_total>",
|
| 93 |
"lstrip": false,
|
| 94 |
"normalized": true,
|
|
|
|
| 96 |
"single_word": false,
|
| 97 |
"special": false
|
| 98 |
},
|
| 99 |
+
"57529": {
|
| 100 |
"content": "</s_total>",
|
| 101 |
"lstrip": false,
|
| 102 |
"normalized": true,
|
|
|
|
| 104 |
"single_word": false,
|
| 105 |
"special": false
|
| 106 |
},
|
| 107 |
+
"57530": {
|
| 108 |
"content": "<s_date>",
|
| 109 |
"lstrip": false,
|
| 110 |
"normalized": true,
|
|
|
|
| 112 |
"single_word": false,
|
| 113 |
"special": false
|
| 114 |
},
|
| 115 |
+
"57531": {
|
| 116 |
"content": "</s_date>",
|
| 117 |
"lstrip": false,
|
| 118 |
"normalized": true,
|
|
|
|
| 120 |
"single_word": false,
|
| 121 |
"special": false
|
| 122 |
},
|
| 123 |
+
"57532": {
|
| 124 |
"content": "<s_company>",
|
| 125 |
"lstrip": false,
|
| 126 |
"normalized": true,
|
|
|
|
| 128 |
"single_word": false,
|
| 129 |
"special": false
|
| 130 |
},
|
| 131 |
+
"57533": {
|
| 132 |
"content": "</s_company>",
|
| 133 |
"lstrip": false,
|
| 134 |
"normalized": true,
|
|
|
|
| 136 |
"single_word": false,
|
| 137 |
"special": false
|
| 138 |
},
|
| 139 |
+
"57534": {
|
| 140 |
"content": "<s_address>",
|
| 141 |
"lstrip": false,
|
| 142 |
"normalized": true,
|
|
|
|
| 144 |
"single_word": false,
|
| 145 |
"special": false
|
| 146 |
},
|
| 147 |
+
"57535": {
|
| 148 |
"content": "</s_address>",
|
| 149 |
"lstrip": false,
|
| 150 |
"normalized": true,
|