Training done. Epoch: 9
Browse files- added_tokens.json +12 -12
- tokenizer.json +12 -12
- tokenizer_config.json +12 -12
added_tokens.json
CHANGED
|
@@ -1,24 +1,24 @@
|
|
| 1 |
{
|
| 2 |
-
"</
|
| 3 |
"</s_d>": 57530,
|
| 4 |
-
"</
|
| 5 |
"</s_m>": 57532,
|
| 6 |
-
"</
|
| 7 |
-
"</
|
| 8 |
-
"</
|
| 9 |
-
"</
|
| 10 |
"</s_y>": 57534,
|
| 11 |
-
"<
|
| 12 |
"<s_cord-v2>": 57543,
|
| 13 |
"<s_d>": 57529,
|
| 14 |
-
"<
|
| 15 |
"<s_iitcdip>": 57523,
|
| 16 |
"<s_m>": 57531,
|
| 17 |
-
"<
|
| 18 |
-
"<
|
| 19 |
"<s_synthdog>": 57524,
|
| 20 |
-
"<
|
| 21 |
-
"<
|
| 22 |
"<s_y>": 57533,
|
| 23 |
"<sep/>": 57522
|
| 24 |
}
|
|
|
|
| 1 |
{
|
| 2 |
+
"</s_an>": 57526,
|
| 3 |
"</s_d>": 57530,
|
| 4 |
+
"</s_ed>": 57528,
|
| 5 |
"</s_m>": 57532,
|
| 6 |
+
"</s_pn>": 57536,
|
| 7 |
+
"</s_sd>": 57538,
|
| 8 |
+
"</s_t>": 57540,
|
| 9 |
+
"</s_v>": 57542,
|
| 10 |
"</s_y>": 57534,
|
| 11 |
+
"<s_an>": 57525,
|
| 12 |
"<s_cord-v2>": 57543,
|
| 13 |
"<s_d>": 57529,
|
| 14 |
+
"<s_ed>": 57527,
|
| 15 |
"<s_iitcdip>": 57523,
|
| 16 |
"<s_m>": 57531,
|
| 17 |
+
"<s_pn>": 57535,
|
| 18 |
+
"<s_sd>": 57537,
|
| 19 |
"<s_synthdog>": 57524,
|
| 20 |
+
"<s_t>": 57539,
|
| 21 |
+
"<s_v>": 57541,
|
| 22 |
"<s_y>": 57533,
|
| 23 |
"<sep/>": 57522
|
| 24 |
}
|
tokenizer.json
CHANGED
|
@@ -91,7 +91,7 @@
|
|
| 91 |
},
|
| 92 |
{
|
| 93 |
"id": 57525,
|
| 94 |
-
"content": "<
|
| 95 |
"single_word": false,
|
| 96 |
"lstrip": false,
|
| 97 |
"rstrip": false,
|
|
@@ -100,7 +100,7 @@
|
|
| 100 |
},
|
| 101 |
{
|
| 102 |
"id": 57526,
|
| 103 |
-
"content": "</
|
| 104 |
"single_word": false,
|
| 105 |
"lstrip": false,
|
| 106 |
"rstrip": false,
|
|
@@ -109,7 +109,7 @@
|
|
| 109 |
},
|
| 110 |
{
|
| 111 |
"id": 57527,
|
| 112 |
-
"content": "<
|
| 113 |
"single_word": false,
|
| 114 |
"lstrip": false,
|
| 115 |
"rstrip": false,
|
|
@@ -118,7 +118,7 @@
|
|
| 118 |
},
|
| 119 |
{
|
| 120 |
"id": 57528,
|
| 121 |
-
"content": "</
|
| 122 |
"single_word": false,
|
| 123 |
"lstrip": false,
|
| 124 |
"rstrip": false,
|
|
@@ -181,7 +181,7 @@
|
|
| 181 |
},
|
| 182 |
{
|
| 183 |
"id": 57535,
|
| 184 |
-
"content": "<
|
| 185 |
"single_word": false,
|
| 186 |
"lstrip": false,
|
| 187 |
"rstrip": false,
|
|
@@ -190,7 +190,7 @@
|
|
| 190 |
},
|
| 191 |
{
|
| 192 |
"id": 57536,
|
| 193 |
-
"content": "</
|
| 194 |
"single_word": false,
|
| 195 |
"lstrip": false,
|
| 196 |
"rstrip": false,
|
|
@@ -199,7 +199,7 @@
|
|
| 199 |
},
|
| 200 |
{
|
| 201 |
"id": 57537,
|
| 202 |
-
"content": "<
|
| 203 |
"single_word": false,
|
| 204 |
"lstrip": false,
|
| 205 |
"rstrip": false,
|
|
@@ -208,7 +208,7 @@
|
|
| 208 |
},
|
| 209 |
{
|
| 210 |
"id": 57538,
|
| 211 |
-
"content": "</
|
| 212 |
"single_word": false,
|
| 213 |
"lstrip": false,
|
| 214 |
"rstrip": false,
|
|
@@ -217,7 +217,7 @@
|
|
| 217 |
},
|
| 218 |
{
|
| 219 |
"id": 57539,
|
| 220 |
-
"content": "<
|
| 221 |
"single_word": false,
|
| 222 |
"lstrip": false,
|
| 223 |
"rstrip": false,
|
|
@@ -226,7 +226,7 @@
|
|
| 226 |
},
|
| 227 |
{
|
| 228 |
"id": 57540,
|
| 229 |
-
"content": "</
|
| 230 |
"single_word": false,
|
| 231 |
"lstrip": false,
|
| 232 |
"rstrip": false,
|
|
@@ -235,7 +235,7 @@
|
|
| 235 |
},
|
| 236 |
{
|
| 237 |
"id": 57541,
|
| 238 |
-
"content": "<
|
| 239 |
"single_word": false,
|
| 240 |
"lstrip": false,
|
| 241 |
"rstrip": false,
|
|
@@ -244,7 +244,7 @@
|
|
| 244 |
},
|
| 245 |
{
|
| 246 |
"id": 57542,
|
| 247 |
-
"content": "</
|
| 248 |
"single_word": false,
|
| 249 |
"lstrip": false,
|
| 250 |
"rstrip": false,
|
|
|
|
| 91 |
},
|
| 92 |
{
|
| 93 |
"id": 57525,
|
| 94 |
+
"content": "<s_an>",
|
| 95 |
"single_word": false,
|
| 96 |
"lstrip": false,
|
| 97 |
"rstrip": false,
|
|
|
|
| 100 |
},
|
| 101 |
{
|
| 102 |
"id": 57526,
|
| 103 |
+
"content": "</s_an>",
|
| 104 |
"single_word": false,
|
| 105 |
"lstrip": false,
|
| 106 |
"rstrip": false,
|
|
|
|
| 109 |
},
|
| 110 |
{
|
| 111 |
"id": 57527,
|
| 112 |
+
"content": "<s_ed>",
|
| 113 |
"single_word": false,
|
| 114 |
"lstrip": false,
|
| 115 |
"rstrip": false,
|
|
|
|
| 118 |
},
|
| 119 |
{
|
| 120 |
"id": 57528,
|
| 121 |
+
"content": "</s_ed>",
|
| 122 |
"single_word": false,
|
| 123 |
"lstrip": false,
|
| 124 |
"rstrip": false,
|
|
|
|
| 181 |
},
|
| 182 |
{
|
| 183 |
"id": 57535,
|
| 184 |
+
"content": "<s_pn>",
|
| 185 |
"single_word": false,
|
| 186 |
"lstrip": false,
|
| 187 |
"rstrip": false,
|
|
|
|
| 190 |
},
|
| 191 |
{
|
| 192 |
"id": 57536,
|
| 193 |
+
"content": "</s_pn>",
|
| 194 |
"single_word": false,
|
| 195 |
"lstrip": false,
|
| 196 |
"rstrip": false,
|
|
|
|
| 199 |
},
|
| 200 |
{
|
| 201 |
"id": 57537,
|
| 202 |
+
"content": "<s_sd>",
|
| 203 |
"single_word": false,
|
| 204 |
"lstrip": false,
|
| 205 |
"rstrip": false,
|
|
|
|
| 208 |
},
|
| 209 |
{
|
| 210 |
"id": 57538,
|
| 211 |
+
"content": "</s_sd>",
|
| 212 |
"single_word": false,
|
| 213 |
"lstrip": false,
|
| 214 |
"rstrip": false,
|
|
|
|
| 217 |
},
|
| 218 |
{
|
| 219 |
"id": 57539,
|
| 220 |
+
"content": "<s_t>",
|
| 221 |
"single_word": false,
|
| 222 |
"lstrip": false,
|
| 223 |
"rstrip": false,
|
|
|
|
| 226 |
},
|
| 227 |
{
|
| 228 |
"id": 57540,
|
| 229 |
+
"content": "</s_t>",
|
| 230 |
"single_word": false,
|
| 231 |
"lstrip": false,
|
| 232 |
"rstrip": false,
|
|
|
|
| 235 |
},
|
| 236 |
{
|
| 237 |
"id": 57541,
|
| 238 |
+
"content": "<s_v>",
|
| 239 |
"single_word": false,
|
| 240 |
"lstrip": false,
|
| 241 |
"rstrip": false,
|
|
|
|
| 244 |
},
|
| 245 |
{
|
| 246 |
"id": 57542,
|
| 247 |
+
"content": "</s_v>",
|
| 248 |
"single_word": false,
|
| 249 |
"lstrip": false,
|
| 250 |
"rstrip": false,
|
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,
|
|
@@ -73,7 +73,7 @@
|
|
| 73 |
"special": false
|
| 74 |
},
|
| 75 |
"57526": {
|
| 76 |
-
"content": "</
|
| 77 |
"lstrip": false,
|
| 78 |
"normalized": true,
|
| 79 |
"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,
|
|
@@ -145,7 +145,7 @@
|
|
| 145 |
"special": false
|
| 146 |
},
|
| 147 |
"57535": {
|
| 148 |
-
"content": "<
|
| 149 |
"lstrip": false,
|
| 150 |
"normalized": true,
|
| 151 |
"rstrip": false,
|
|
@@ -153,7 +153,7 @@
|
|
| 153 |
"special": false
|
| 154 |
},
|
| 155 |
"57536": {
|
| 156 |
-
"content": "</
|
| 157 |
"lstrip": false,
|
| 158 |
"normalized": true,
|
| 159 |
"rstrip": false,
|
|
@@ -161,7 +161,7 @@
|
|
| 161 |
"special": false
|
| 162 |
},
|
| 163 |
"57537": {
|
| 164 |
-
"content": "<
|
| 165 |
"lstrip": false,
|
| 166 |
"normalized": true,
|
| 167 |
"rstrip": false,
|
|
@@ -169,7 +169,7 @@
|
|
| 169 |
"special": false
|
| 170 |
},
|
| 171 |
"57538": {
|
| 172 |
-
"content": "</
|
| 173 |
"lstrip": false,
|
| 174 |
"normalized": true,
|
| 175 |
"rstrip": false,
|
|
@@ -177,7 +177,7 @@
|
|
| 177 |
"special": false
|
| 178 |
},
|
| 179 |
"57539": {
|
| 180 |
-
"content": "<
|
| 181 |
"lstrip": false,
|
| 182 |
"normalized": true,
|
| 183 |
"rstrip": false,
|
|
@@ -185,7 +185,7 @@
|
|
| 185 |
"special": false
|
| 186 |
},
|
| 187 |
"57540": {
|
| 188 |
-
"content": "</
|
| 189 |
"lstrip": false,
|
| 190 |
"normalized": true,
|
| 191 |
"rstrip": false,
|
|
@@ -193,7 +193,7 @@
|
|
| 193 |
"special": false
|
| 194 |
},
|
| 195 |
"57541": {
|
| 196 |
-
"content": "<
|
| 197 |
"lstrip": false,
|
| 198 |
"normalized": true,
|
| 199 |
"rstrip": false,
|
|
@@ -201,7 +201,7 @@
|
|
| 201 |
"special": false
|
| 202 |
},
|
| 203 |
"57542": {
|
| 204 |
-
"content": "</
|
| 205 |
"lstrip": false,
|
| 206 |
"normalized": true,
|
| 207 |
"rstrip": false,
|
|
|
|
| 65 |
"special": true
|
| 66 |
},
|
| 67 |
"57525": {
|
| 68 |
+
"content": "<s_an>",
|
| 69 |
"lstrip": false,
|
| 70 |
"normalized": true,
|
| 71 |
"rstrip": false,
|
|
|
|
| 73 |
"special": false
|
| 74 |
},
|
| 75 |
"57526": {
|
| 76 |
+
"content": "</s_an>",
|
| 77 |
"lstrip": false,
|
| 78 |
"normalized": true,
|
| 79 |
"rstrip": false,
|
|
|
|
| 81 |
"special": false
|
| 82 |
},
|
| 83 |
"57527": {
|
| 84 |
+
"content": "<s_ed>",
|
| 85 |
"lstrip": false,
|
| 86 |
"normalized": true,
|
| 87 |
"rstrip": false,
|
|
|
|
| 89 |
"special": false
|
| 90 |
},
|
| 91 |
"57528": {
|
| 92 |
+
"content": "</s_ed>",
|
| 93 |
"lstrip": false,
|
| 94 |
"normalized": true,
|
| 95 |
"rstrip": false,
|
|
|
|
| 145 |
"special": false
|
| 146 |
},
|
| 147 |
"57535": {
|
| 148 |
+
"content": "<s_pn>",
|
| 149 |
"lstrip": false,
|
| 150 |
"normalized": true,
|
| 151 |
"rstrip": false,
|
|
|
|
| 153 |
"special": false
|
| 154 |
},
|
| 155 |
"57536": {
|
| 156 |
+
"content": "</s_pn>",
|
| 157 |
"lstrip": false,
|
| 158 |
"normalized": true,
|
| 159 |
"rstrip": false,
|
|
|
|
| 161 |
"special": false
|
| 162 |
},
|
| 163 |
"57537": {
|
| 164 |
+
"content": "<s_sd>",
|
| 165 |
"lstrip": false,
|
| 166 |
"normalized": true,
|
| 167 |
"rstrip": false,
|
|
|
|
| 169 |
"special": false
|
| 170 |
},
|
| 171 |
"57538": {
|
| 172 |
+
"content": "</s_sd>",
|
| 173 |
"lstrip": false,
|
| 174 |
"normalized": true,
|
| 175 |
"rstrip": false,
|
|
|
|
| 177 |
"special": false
|
| 178 |
},
|
| 179 |
"57539": {
|
| 180 |
+
"content": "<s_t>",
|
| 181 |
"lstrip": false,
|
| 182 |
"normalized": true,
|
| 183 |
"rstrip": false,
|
|
|
|
| 185 |
"special": false
|
| 186 |
},
|
| 187 |
"57540": {
|
| 188 |
+
"content": "</s_t>",
|
| 189 |
"lstrip": false,
|
| 190 |
"normalized": true,
|
| 191 |
"rstrip": false,
|
|
|
|
| 193 |
"special": false
|
| 194 |
},
|
| 195 |
"57541": {
|
| 196 |
+
"content": "<s_v>",
|
| 197 |
"lstrip": false,
|
| 198 |
"normalized": true,
|
| 199 |
"rstrip": false,
|
|
|
|
| 201 |
"special": false
|
| 202 |
},
|
| 203 |
"57542": {
|
| 204 |
+
"content": "</s_v>",
|
| 205 |
"lstrip": false,
|
| 206 |
"normalized": true,
|
| 207 |
"rstrip": false,
|