DerivedFunction commited on
Commit
ed4764f
·
verified ·
1 Parent(s): 8dc7b23

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +151 -149
README.md CHANGED
@@ -116,6 +116,9 @@ language:
116
  - yi
117
  - zh
118
  - zu
 
 
 
119
  datasets:
120
  - wikimedia/wikipedia
121
  - HuggingFaceFW/finetranslations
@@ -125,9 +128,6 @@ datasets:
125
  - DerivedFunction/finetranslations-filtered
126
  - DerivedFunction/tatoeba-filtered
127
  pipeline_tag: token-classification
128
- model-index:
129
- - name: polyglot-tagger
130
- results: []
131
  ---
132
 
133
 
@@ -138,7 +138,7 @@ Fine-tuned `xlm-roberta-base` for sentence-level language tagging across 100 lan
138
  The model predicts BIO-style language tags over tokens, which makes it useful for
139
  language identification, code-switch detection, and multilingual document analysis.
140
 
141
- > Compared to version 2.1, this version had training data that cleaned up mixed script training rows not in the target language, such as Arabic present in Cyrllic languages.
142
 
143
  ## Model description
144
 
@@ -152,7 +152,7 @@ Note that as a general language tagging model, it can potentially get confused f
152
  The model is trained on a sentence with a minimum of four tokens, so it may not accurately classify very short and ambigous statements. Note that this model is experimental
153
  and may produce unexpected results compared to generic text classifiers. It is trained on cleaned text, therefore, "messy" text may unexpectedly produce different results.
154
 
155
- > Note that Romanized versions of any language may only have minor representation in the training set, such as Romanized Russian, and Hindi.
156
 
157
  ### Training and Evaluation Data
158
 
@@ -164,133 +164,132 @@ factors were used to simulate messy text, and to reduce single character bias on
164
  - Random chance to change the casing of compatible language scripts, such as Latin and Cyrllic.
165
  - Low chance of simulating OCR and messy text with character mutation.
166
 
167
- To generalize well on both the target language and code switching a circulumn is provided:
168
  - Pure documents 55%: Single language to learn its vocabulary, simulating a short paragraph of a single language.
169
  - Homogenous 25%: Single language + one foreign sentence to learn simple code switching.
170
  - Spliced 10%: A foreign sentence is centered between two same-language sentence, with the first sentence's punctuation stripped, and second sentence's forced to be lowercased.
171
  - Mixed 10%: Generic mix of any languages.
172
 
173
 
174
- ### Training Data Breakdown
175
- | lang | train sentences | train tokens | eval sentences | eval tokens | all sentences | all tokens |
176
- | :--- | ---: | ---: | ---: | ---: | ---: | ---: |
177
- | en | 342138 (2.14%) | 8515554 (1.58%) | 2925 (3.89%) | 29279 (1.57%) | 345063 (2.14%) | 8544833 (1.58%) |
178
- | es | 290248 (1.81%) | 8250416 (1.53%) | 1864 (2.48%) | 19826 (1.06%) | 292112 (1.82%) | 8270242 (1.53%) |
179
- | ru | 289806 (1.81%) | 7911565 (1.47%) | 1963 (2.61%) | 19501 (1.04%) | 291769 (1.81%) | 7931066 (1.47%) |
180
- | ja | 288677 (1.80%) | 6935500 (1.29%) | 1836 (2.44%) | 20115 (1.08%) | 290513 (1.81%) | 6955615 (1.29%) |
181
- | fr | 285022 (1.78%) | 8950594 (1.66%) | 1849 (2.46%) | 22785 (1.22%) | 286871 (1.78%) | 8973379 (1.66%) |
182
- | zh | 282136 (1.76%) | 6585294 (1.22%) | 1780 (2.37%) | 17110 (0.92%) | 283916 (1.76%) | 6602404 (1.22%) |
183
- | de | 279766 (1.75%) | 7421366 (1.38%) | 1761 (2.34%) | 20090 (1.08%) | 281527 (1.75%) | 7441456 (1.37%) |
184
- | pt | 277705 (1.73%) | 7518951 (1.39%) | 1789 (2.38%) | 16872 (0.90%) | 279494 (1.74%) | 7535823 (1.39%) |
185
- | it | 274463 (1.71%) | 7647936 (1.42%) | 1641 (2.18%) | 14284 (0.77%) | 276104 (1.72%) | 7662220 (1.42%) |
186
- | uk | 244679 (1.53%) | 6480580 (1.20%) | 1187 (1.58%) | 9893 (0.53%) | 245866 (1.53%) | 6490473 (1.20%) |
187
- | fi | 243388 (1.52%) | 5886670 (1.09%) | 1521 (2.02%) | 14613 (0.78%) | 244909 (1.52%) | 5901283 (1.09%) |
188
- | ar | 243319 (1.52%) | 5055410 (0.94%) | 1237 (1.65%) | 18277 (0.98%) | 244556 (1.52%) | 5073687 (0.94%) |
189
- | pl | 239763 (1.50%) | 6747779 (1.25%) | 1162 (1.55%) | 11588 (0.62%) | 240925 (1.50%) | 6759367 (1.25%) |
190
- | he | 235680 (1.47%) | 6872873 (1.27%) | 842 (1.12%) | 7301 (0.39%) | 236522 (1.47%) | 6880174 (1.27%) |
191
- | hu | 234082 (1.46%) | 6422364 (1.19%) | 1093 (1.45%) | 10598 (0.57%) | 235175 (1.46%) | 6432962 (1.19%) |
192
- | tr | 231785 (1.45%) | 5649577 (1.05%) | 1078 (1.43%) | 9102 (0.49%) | 232863 (1.45%) | 5658679 (1.05%) |
193
- | cs | 229623 (1.43%) | 6157905 (1.14%) | 1010 (1.34%) | 8638 (0.46%) | 230633 (1.43%) | 6166543 (1.14%) |
194
- | nl | 227744 (1.42%) | 5208953 (0.97%) | 1086 (1.45%) | 9966 (0.53%) | 228830 (1.42%) | 5218919 (0.96%) |
195
- | lt | 220746 (1.38%) | 5523460 (1.02%) | 953 (1.27%) | 8390 (0.45%) | 221699 (1.38%) | 5531850 (1.02%) |
196
- | mk | 219568 (1.37%) | 6321823 (1.17%) | 830 (1.10%) | 7363 (0.39%) | 220398 (1.37%) | 6329186 (1.17%) |
197
- | mr | 218564 (1.36%) | 5737670 (1.06%) | 755 (1.00%) | 6419 (0.34%) | 219319 (1.36%) | 5744089 (1.06%) |
198
- | eo | 214879 (1.34%) | 5575371 (1.03%) | 806 (1.07%) | 9199 (0.49%) | 215685 (1.34%) | 5584570 (1.03%) |
199
- | no | 212572 (1.33%) | 6119841 (1.13%) | 1520 (2.02%) | 46805 (2.51%) | 214092 (1.33%) | 6166646 (1.14%) |
200
- | da | 210913 (1.32%) | 5244616 (0.97%) | 1244 (1.66%) | 11307 (0.61%) | 212157 (1.32%) | 5255923 (0.97%) |
201
- | tl | 198531 (1.24%) | 5385091 (1.00%) | 980 (1.30%) | 11451 (0.61%) | 199511 (1.24%) | 5396542 (1.00%) |
202
- | hy | 197579 (1.23%) | 6260512 (1.16%) | 741 (0.99%) | 14164 (0.76%) | 198320 (1.23%) | 6274676 (1.16%) |
203
- | hi | 196290 (1.23%) | 7473616 (1.39%) | 995 (1.32%) | 43438 (2.33%) | 197285 (1.23%) | 7517054 (1.39%) |
204
- | ko | 196247 (1.23%) | 6256562 (1.16%) | 1021 (1.36%) | 28154 (1.51%) | 197268 (1.23%) | 6284716 (1.16%) |
205
- | el | 192981 (1.21%) | 7010969 (1.30%) | 739 (0.98%) | 17209 (0.92%) | 193720 (1.20%) | 7028178 (1.30%) |
206
- | ro | 185669 (1.16%) | 6171646 (1.14%) | 771 (1.03%) | 21626 (1.16%) | 186440 (1.16%) | 6193272 (1.14%) |
207
- | fa | 182012 (1.14%) | 5634518 (1.04%) | 724 (0.96%) | 21266 (1.14%) | 182736 (1.14%) | 5655784 (1.04%) |
208
- | sk | 181257 (1.13%) | 5323589 (0.99%) | 868 (1.16%) | 30324 (1.62%) | 182125 (1.13%) | 5353913 (0.99%) |
209
- | la | 178735 (1.12%) | 4452161 (0.83%) | 752 (1.00%) | 7803 (0.42%) | 179487 (1.12%) | 4459964 (0.82%) |
210
- | bg | 178477 (1.11%) | 5772043 (1.07%) | 681 (0.91%) | 18306 (0.98%) | 179158 (1.11%) | 5790349 (1.07%) |
211
- | is | 174213 (1.09%) | 6024579 (1.12%) | 1005 (1.34%) | 47601 (2.55%) | 175218 (1.09%) | 6072180 (1.12%) |
212
- | be | 174166 (1.09%) | 6368950 (1.18%) | 880 (1.17%) | 30097 (1.61%) | 175046 (1.09%) | 6399047 (1.18%) |
213
- | lv | 170404 (1.06%) | 5706364 (1.06%) | 702 (0.93%) | 35541 (1.90%) | 171106 (1.06%) | 5741905 (1.06%) |
214
- | ckb | 166543 (1.04%) | 7530043 (1.40%) | 591 (0.79%) | 25103 (1.35%) | 167134 (1.04%) | 7555146 (1.40%) |
215
- | ms | 165285 (1.03%) | 4430830 (0.82%) | 778 (1.04%) | 24919 (1.34%) | 166063 (1.03%) | 4455749 (0.82%) |
216
- | kk | 163582 (1.02%) | 4925313 (0.91%) | 629 (0.84%) | 17536 (0.94%) | 164211 (1.02%) | 4942849 (0.91%) |
217
- | ka | 162558 (1.02%) | 5244466 (0.97%) | 527 (0.70%) | 16440 (0.88%) | 163085 (1.01%) | 5260906 (0.97%) |
218
- | bn | 162058 (1.01%) | 6155732 (1.14%) | 416 (0.55%) | 14535 (0.78%) | 162474 (1.01%) | 6170267 (1.14%) |
219
- | eu | 160479 (1.00%) | 5375791 (1.00%) | 675 (0.90%) | 34708 (1.86%) | 161154 (1.00%) | 5410499 (1.00%) |
220
- | as | 160525 (1.00%) | 8228319 (1.53%) | 377 (0.50%) | 20723 (1.11%) | 160902 (1.00%) | 8249042 (1.52%) |
221
- | mn | 160161 (1.00%) | 5433430 (1.01%) | 645 (0.86%) | 19610 (1.05%) | 160806 (1.00%) | 5453040 (1.01%) |
222
- | ur | 158216 (0.99%) | 5101965 (0.95%) | 582 (0.77%) | 19433 (1.04%) | 158798 (0.99%) | 5121398 (0.95%) |
223
- | ky | 157623 (0.98%) | 4996516 (0.93%) | 640 (0.85%) | 18469 (0.99%) | 158263 (0.98%) | 5014985 (0.93%) |
224
- | ba | 157391 (0.98%) | 7998542 (1.48%) | 591 (0.79%) | 30781 (1.65%) | 157982 (0.98%) | 8029323 (1.48%) |
225
- | oc | 157265 (0.98%) | 5676181 (1.05%) | 669 (0.89%) | 25130 (1.35%) | 157934 (0.98%) | 5701311 (1.05%) |
226
- | th | 156439 (0.98%) | 5222662 (0.97%) | 579 (0.77%) | 20033 (1.07%) | 157018 (0.98%) | 5242695 (0.97%) |
227
- | hr | 156094 (0.97%) | 4838805 (0.90%) | 716 (0.95%) | 31608 (1.69%) | 156810 (0.97%) | 4870413 (0.90%) |
228
- | af | 155450 (0.97%) | 4536957 (0.84%) | 995 (1.32%) | 29783 (1.60%) | 156445 (0.97%) | 4566740 (0.84%) |
229
- | ps | 155591 (0.97%) | 4514560 (0.84%) | 537 (0.71%) | 15173 (0.81%) | 156128 (0.97%) | 4529733 (0.84%) |
230
- | id | 154998 (0.97%) | 4010110 (0.74%) | 737 (0.98%) | 18766 (1.01%) | 155735 (0.97%) | 4028876 (0.74%) |
231
- | pa | 155053 (0.97%) | 7277866 (1.35%) | 602 (0.80%) | 29831 (1.60%) | 155655 (0.97%) | 7307697 (1.35%) |
232
- | sw | 154323 (0.96%) | 4695422 (0.87%) | 562 (0.75%) | 20760 (1.11%) | 154885 (0.96%) | 4716182 (0.87%) |
233
- | tt | 152959 (0.96%) | 5329166 (0.99%) | 642 (0.85%) | 9018 (0.48%) | 153601 (0.95%) | 5338184 (0.99%) |
234
- | jv | 149846 (0.94%) | 4508978 (0.84%) | 512 (0.68%) | 18744 (1.00%) | 150358 (0.93%) | 4527722 (0.84%) |
235
- | cy | 147138 (0.92%) | 5344455 (0.99%) | 664 (0.88%) | 27600 (1.48%) | 147802 (0.92%) | 5372055 (0.99%) |
236
- | ga | 144559 (0.90%) | 5145701 (0.95%) | 681 (0.91%) | 28351 (1.52%) | 145240 (0.90%) | 5174052 (0.96%) |
237
- | bs | 142871 (0.89%) | 4246487 (0.79%) | 664 (0.88%) | 23293 (1.25%) | 143535 (0.89%) | 4269780 (0.79%) |
238
- | ca | 142813 (0.89%) | 5286099 (0.98%) | 614 (0.82%) | 20417 (1.09%) | 143427 (0.89%) | 5306516 (0.98%) |
239
- | kn | 142453 (0.89%) | 14440106 (2.68%) | 636 (0.85%) | 48433 (2.60%) | 143089 (0.89%) | 14488539 (2.68%) |
240
- | ne | 141617 (0.88%) | 4633072 (0.86%) | 441 (0.59%) | 13357 (0.72%) | 142058 (0.88%) | 4646429 (0.86%) |
241
- | gl | 140293 (0.88%) | 4370302 (0.81%) | 580 (0.77%) | 17719 (0.95%) | 140873 (0.88%) | 4388021 (0.81%) |
242
- | ku | 140150 (0.88%) | 4681613 (0.87%) | 539 (0.72%) | 25315 (1.36%) | 140689 (0.87%) | 4706928 (0.87%) |
243
- | uz | 137971 (0.86%) | 4402565 (0.82%) | 512 (0.68%) | 18239 (0.98%) | 138483 (0.86%) | 4420804 (0.82%) |
244
- | sl | 137474 (0.86%) | 3689209 (0.68%) | 602 (0.80%) | 15975 (0.86%) | 138076 (0.86%) | 3705184 (0.68%) |
245
- | sv | 136227 (0.85%) | 3888997 (0.72%) | 951 (1.27%) | 7503 (0.40%) | 137178 (0.85%) | 3896500 (0.72%) |
246
- | tg | 129047 (0.81%) | 7212798 (1.34%) | 507 (0.67%) | 30575 (1.64%) | 129554 (0.81%) | 7243373 (1.34%) |
247
- | et | 125334 (0.78%) | 3134904 (0.58%) | 516 (0.69%) | 13566 (0.73%) | 125850 (0.78%) | 3148470 (0.58%) |
248
- | br | 124276 (0.78%) | 4304633 (0.80%) | 555 (0.74%) | 16741 (0.90%) | 124831 (0.78%) | 4321374 (0.80%) |
249
- | su | 123580 (0.77%) | 3996363 (0.74%) | 485 (0.65%) | 18936 (1.01%) | 124065 (0.77%) | 4015299 (0.74%) |
250
- | lb | 123335 (0.77%) | 4218363 (0.78%) | 494 (0.66%) | 18236 (0.98%) | 123829 (0.77%) | 4236599 (0.78%) |
251
- | mt | 122262 (0.76%) | 6326484 (1.17%) | 448 (0.60%) | 23893 (1.28%) | 122710 (0.76%) | 6350377 (1.17%) |
252
- | sr | 121328 (0.76%) | 3374324 (0.63%) | 461 (0.61%) | 4054 (0.22%) | 121789 (0.76%) | 3378378 (0.62%) |
253
- | sq | 114178 (0.71%) | 3917105 (0.73%) | 519 (0.69%) | 17283 (0.93%) | 114697 (0.71%) | 3934388 (0.73%) |
254
- | or | 105680 (0.66%) | 3746174 (0.69%) | 415 (0.55%) | 13597 (0.73%) | 106095 (0.66%) | 3759771 (0.69%) |
255
- | ml | 104847 (0.65%) | 10467180 (1.94%) | 432 (0.58%) | 34289 (1.84%) | 105279 (0.65%) | 10501469 (1.94%) |
256
- | yi | 99296 (0.62%) | 4037607 (0.75%) | 356 (0.47%) | 7656 (0.41%) | 99652 (0.62%) | 4045263 (0.75%) |
257
- | te | 96687 (0.60%) | 9339722 (1.73%) | 383 (0.51%) | 32688 (1.75%) | 97070 (0.60%) | 9372410 (1.73%) |
258
- | ta | 89670 (0.56%) | 7564484 (1.40%) | 378 (0.50%) | 25504 (1.37%) | 90048 (0.56%) | 7589988 (1.40%) |
259
- | mg | 89513 (0.56%) | 3128095 (0.58%) | 343 (0.46%) | 11046 (0.59%) | 89856 (0.56%) | 3139141 (0.58%) |
260
- | si | 88038 (0.55%) | 5030862 (0.93%) | 345 (0.46%) | 17895 (0.96%) | 88383 (0.55%) | 5048757 (0.93%) |
261
- | rm | 71273 (0.45%) | 2704067 (0.50%) | 254 (0.34%) | 10060 (0.54%) | 71527 (0.44%) | 2714127 (0.50%) |
262
- | vi | 70788 (0.44%) | 2476955 (0.46%) | 333 (0.44%) | 3428 (0.18%) | 71121 (0.44%) | 2480383 (0.46%) |
263
- | gu | 67542 (0.42%) | 7557129 (1.40%) | 299 (0.40%) | 28080 (1.50%) | 67841 (0.42%) | 7585209 (1.40%) |
264
- | bo | 66467 (0.42%) | 1308971 (0.24%) | 218 (0.29%) | 4357 (0.23%) | 66685 (0.41%) | 1313328 (0.24%) |
265
- | ug | 61389 (0.38%) | 1398443 (0.26%) | 213 (0.28%) | 4674 (0.25%) | 61602 (0.38%) | 1403117 (0.26%) |
266
- | dv | 57776 (0.36%) | 1485603 (0.28%) | 205 (0.27%) | 5705 (0.31%) | 57981 (0.36%) | 1491308 (0.28%) |
267
- | am | 56487 (0.35%) | 2563835 (0.48%) | 227 (0.30%) | 11582 (0.62%) | 56714 (0.35%) | 2575417 (0.48%) |
268
- | yo | 56313 (0.35%) | 3444596 (0.64%) | 230 (0.31%) | 18061 (0.97%) | 56543 (0.35%) | 3462657 (0.64%) |
269
- | my | 55960 (0.35%) | 2062131 (0.38%) | 212 (0.28%) | 8300 (0.44%) | 56172 (0.35%) | 2070431 (0.38%) |
270
- | km | 53890 (0.34%) | 2889004 (0.54%) | 188 (0.25%) | 9842 (0.53%) | 54078 (0.34%) | 2898846 (0.54%) |
271
- | so | 53862 (0.34%) | 1963263 (0.36%) | 202 (0.27%) | 7961 (0.43%) | 54064 (0.34%) | 1971224 (0.36%) |
272
- | sd | 52494 (0.33%) | 3075283 (0.57%) | 200 (0.27%) | 10379 (0.56%) | 52694 (0.33%) | 3085662 (0.57%) |
273
- | zu | 50081 (0.31%) | 2287742 (0.42%) | 187 (0.25%) | 8913 (0.48%) | 50268 (0.31%) | 2296655 (0.42%) |
274
- | lo | 47982 (0.30%) | 1659406 (0.31%) | 188 (0.25%) | 5980 (0.32%) | 48170 (0.30%) | 1665386 (0.31%) |
275
- | ti | 45884 (0.29%) | 2804218 (0.52%) | 193 (0.26%) | 11862 (0.64%) | 46077 (0.29%) | 2816080 (0.52%) |
276
- | ce | 43040 (0.27%) | 2319707 (0.43%) | 181 (0.24%) | 9398 (0.50%) | 43221 (0.27%) | 2329105 (0.43%) |
277
- | ny | 41897 (0.26%) | 1969591 (0.37%) | 159 (0.21%) | 8122 (0.44%) | 42056 (0.26%) | 1977713 (0.37%) |
278
- | gd | 35576 (0.22%) | 1240445 (0.23%) | 142 (0.19%) | 3449 (0.18%) | 35718 (0.22%) | 1243894 (0.23%) |
279
- | xh | 23590 (0.15%) | 877812 (0.16%) | 96 (0.13%) | 3597 (0.19%) | 23686 (0.15%) | 881409 (0.16%) |
280
- | om | 14738 (0.09%) | 523551 (0.10%) | 55 (0.07%) | 1935 (0.10%) | 14793 (0.09%) | 525486 (0.10%) |
281
- | sco | 8374 (0.05%) | 224424 (0.04%) | 30 (0.04%) | 1055 (0.06%) | 8404 (0.05%) | 225479 (0.04%) |
282
- | **total** | 16012306 (100.00%) | 539378202 (100.00%) | 75126 (100.00%) | 1866305 (100.00%) | 16087432 (100.00%) | 541244507 (100.00%) |
283
 
284
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
285
 
286
 
287
- This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base).
288
  It achieves the following results on the evaluation set:
289
- - Loss: 0.0345
290
- - Precision: 0.9508
291
- - Recall: 0.9647
292
- - F1: 0.9577
293
- - Accuracy: 0.9908
294
 
295
  ## Training procedure
296
 
@@ -298,10 +297,10 @@ It achieves the following results on the evaluation set:
298
 
299
  The following hyperparameters were used during training:
300
  - learning_rate: 5e-05
301
- - train_batch_size: 18
302
- - eval_batch_size: 18
303
  - seed: 42
304
- - gradient_accumulation_steps: 8
305
  - total_train_batch_size: 144
306
  - optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
307
  - lr_scheduler_type: linear
@@ -312,30 +311,33 @@ The following hyperparameters were used during training:
312
 
313
  | Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
314
  |:-------------:|:------:|:-----:|:---------------:|:---------:|:------:|:------:|:--------:|
315
- | 0.3393 | 0.0804 | 2500 | 0.1078 | 0.7999 | 0.8817 | 0.8388 | 0.9723 |
316
- | 0.3042 | 0.1609 | 5000 | 0.0910 | 0.8422 | 0.9054 | 0.8726 | 0.9766 |
317
- | 0.2323 | 0.2413 | 7500 | 0.0859 | 0.8661 | 0.9174 | 0.8910 | 0.9766 |
318
- | 0.2253 | 0.3218 | 10000 | 0.0707 | 0.8805 | 0.9266 | 0.9029 | 0.9818 |
319
- | 0.2117 | 0.4022 | 12500 | 0.0715 | 0.8943 | 0.9284 | 0.9110 | 0.9805 |
320
- | 0.1895 | 0.4827 | 15000 | 0.0582 | 0.8992 | 0.9372 | 0.9178 | 0.9843 |
321
- | 0.1865 | 0.5631 | 17500 | 0.0557 | 0.9053 | 0.9381 | 0.9214 | 0.9851 |
322
- | 0.1666 | 0.6436 | 20000 | 0.0560 | 0.9047 | 0.9424 | 0.9232 | 0.9852 |
323
- | 0.1623 | 0.7240 | 22500 | 0.0588 | 0.9099 | 0.9405 | 0.9250 | 0.9838 |
324
- | 0.1564 | 0.8045 | 25000 | 0.0476 | 0.9232 | 0.9489 | 0.9359 | 0.9872 |
325
- | 0.1474 | 0.8849 | 27500 | 0.0525 | 0.9200 | 0.9483 | 0.9339 | 0.9855 |
326
- | 0.1580 | 0.9654 | 30000 | 0.0455 | 0.9245 | 0.9502 | 0.9372 | 0.9878 |
327
- | 0.1245 | 1.0458 | 32500 | 0.0447 | 0.9310 | 0.9536 | 0.9422 | 0.9881 |
328
- | 0.1199 | 1.1262 | 35000 | 0.0445 | 0.9316 | 0.9547 | 0.9430 | 0.9881 |
329
- | 0.1093 | 1.2067 | 37500 | 0.0432 | 0.9363 | 0.9559 | 0.9460 | 0.9886 |
330
- | 0.1164 | 1.2871 | 40000 | 0.0410 | 0.9392 | 0.9583 | 0.9487 | 0.9895 |
331
- | 0.1146 | 1.3676 | 42500 | 0.0433 | 0.9314 | 0.9562 | 0.9436 | 0.9886 |
332
- | 0.0913 | 1.4480 | 45000 | 0.0434 | 0.9403 | 0.9585 | 0.9493 | 0.9894 |
333
- | 0.1068 | 1.5285 | 47500 | 0.0397 | 0.9424 | 0.9608 | 0.9515 | 0.9897 |
334
- | 0.0930 | 1.6089 | 50000 | 0.0388 | 0.9419 | 0.9605 | 0.9511 | 0.9898 |
335
- | 0.0896 | 1.6894 | 52500 | 0.0371 | 0.9453 | 0.9623 | 0.9537 | 0.9901 |
336
- | 0.0862 | 1.7698 | 55000 | 0.0362 | 0.9446 | 0.9624 | 0.9534 | 0.9903 |
337
- | 0.0971 | 1.8503 | 57500 | 0.0355 | 0.9478 | 0.9635 | 0.9556 | 0.9908 |
338
- | 0.0947 | 1.9307 | 60000 | 0.0345 | 0.9508 | 0.9647 | 0.9577 | 0.9908 |
 
 
 
339
 
340
 
341
  ### Framework versions
@@ -343,4 +345,4 @@ The following hyperparameters were used during training:
343
  - Transformers 5.0.0
344
  - Pytorch 2.10.0+cu128
345
  - Datasets 4.0.0
346
- - Tokenizers 0.22.2
 
116
  - yi
117
  - zh
118
  - zu
119
+ model-index:
120
+ - name: polyglot-tagger
121
+ results: []
122
  datasets:
123
  - wikimedia/wikipedia
124
  - HuggingFaceFW/finetranslations
 
128
  - DerivedFunction/finetranslations-filtered
129
  - DerivedFunction/tatoeba-filtered
130
  pipeline_tag: token-classification
 
 
 
131
  ---
132
 
133
 
 
138
  The model predicts BIO-style language tags over tokens, which makes it useful for
139
  language identification, code-switch detection, and multilingual document analysis.
140
 
141
+ > Compared to version 2.2, this version had training data that attempted to fix the model scoring common grade-school words from major langauges as low confidence or in a minor language bucket.
142
 
143
  ## Model description
144
 
 
152
  The model is trained on a sentence with a minimum of four tokens, so it may not accurately classify very short and ambigous statements. Note that this model is experimental
153
  and may produce unexpected results compared to generic text classifiers. It is trained on cleaned text, therefore, "messy" text may unexpectedly produce different results.
154
 
155
+ > Note that Romanized versions of any language may have no representation in the training set, such as Romanized Russian, and Hindi.
156
 
157
  ### Training and Evaluation Data
158
 
 
164
  - Random chance to change the casing of compatible language scripts, such as Latin and Cyrllic.
165
  - Low chance of simulating OCR and messy text with character mutation.
166
 
167
+ To generalize well on both the target language and code switching a curriculum is provided:
168
  - Pure documents 55%: Single language to learn its vocabulary, simulating a short paragraph of a single language.
169
  - Homogenous 25%: Single language + one foreign sentence to learn simple code switching.
170
  - Spliced 10%: A foreign sentence is centered between two same-language sentence, with the first sentence's punctuation stripped, and second sentence's forced to be lowercased.
171
  - Mixed 10%: Generic mix of any languages.
172
 
173
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
174
 
175
 
176
+ | lang | train sentences | train tokens | eval sentences | eval tokens | all sentences | all tokens |
177
+ | :--- | ---: | ---: | ---: | ---: | ---: | ---: |
178
+ | en | 423264 (2.41%) | 9841704 (1.71%) | 3157 (3.84%) | 35025 (1.79%) | 426421 (2.41%) | 9876729 (1.71%) |
179
+ | es | 359106 (2.04%) | 9729675 (1.69%) | 2201 (2.68%) | 22340 (1.14%) | 361307 (2.04%) | 9752015 (1.69%) |
180
+ | ru | 356083 (2.02%) | 8945224 (1.56%) | 2226 (2.71%) | 21578 (1.10%) | 358309 (2.03%) | 8966802 (1.55%) |
181
+ | fr | 354645 (2.02%) | 10591338 (1.84%) | 2213 (2.69%) | 26148 (1.34%) | 356858 (2.02%) | 10617486 (1.84%) |
182
+ | ja | 352243 (2.00%) | 7945312 (1.38%) | 2219 (2.70%) | 25849 (1.32%) | 354462 (2.01%) | 7971161 (1.38%) |
183
+ | pt | 346042 (1.97%) | 8793674 (1.53%) | 2059 (2.50%) | 20881 (1.07%) | 348101 (1.97%) | 8814555 (1.53%) |
184
+ | de | 344644 (1.96%) | 8847457 (1.54%) | 2151 (2.61%) | 24958 (1.27%) | 346795 (1.96%) | 8872415 (1.54%) |
185
+ | it | 343887 (1.95%) | 8790806 (1.53%) | 2000 (2.43%) | 17342 (0.89%) | 345887 (1.96%) | 8808148 (1.53%) |
186
+ | fi | 299568 (1.70%) | 6905536 (1.20%) | 1576 (1.92%) | 14458 (0.74%) | 301144 (1.70%) | 6919994 (1.20%) |
187
+ | uk | 297565 (1.69%) | 7228987 (1.26%) | 1398 (1.70%) | 11391 (0.58%) | 298963 (1.69%) | 7240378 (1.26%) |
188
+ | zh | 294064 (1.67%) | 7329413 (1.28%) | 1717 (2.09%) | 33433 (1.71%) | 295781 (1.67%) | 7362846 (1.28%) |
189
+ | tr | 289328 (1.64%) | 6606625 (1.15%) | 1384 (1.68%) | 11089 (0.57%) | 290712 (1.64%) | 6617714 (1.15%) |
190
+ | he | 289239 (1.64%) | 7792338 (1.36%) | 1100 (1.34%) | 10342 (0.53%) | 290339 (1.64%) | 7802680 (1.35%) |
191
+ | pl | 288423 (1.64%) | 7707293 (1.34%) | 1305 (1.59%) | 11306 (0.58%) | 289728 (1.64%) | 7718599 (1.34%) |
192
+ | hu | 286880 (1.63%) | 7547282 (1.31%) | 1232 (1.50%) | 11115 (0.57%) | 288112 (1.63%) | 7558397 (1.31%) |
193
+ | nl | 280682 (1.60%) | 6057971 (1.05%) | 1296 (1.58%) | 10453 (0.53%) | 281978 (1.60%) | 6068424 (1.05%) |
194
+ | lt | 273386 (1.55%) | 6507776 (1.13%) | 1165 (1.42%) | 10491 (0.54%) | 274551 (1.55%) | 6518267 (1.13%) |
195
+ | eo | 267885 (1.52%) | 6554654 (1.14%) | 1055 (1.28%) | 13271 (0.68%) | 268940 (1.52%) | 6567925 (1.14%) |
196
+ | ar | 257437 (1.46%) | 5125573 (0.89%) | 1327 (1.61%) | 15180 (0.78%) | 258764 (1.46%) | 5140753 (0.89%) |
197
+ | cs | 240507 (1.37%) | 6406086 (1.11%) | 1082 (1.32%) | 9473 (0.48%) | 241589 (1.37%) | 6415559 (1.11%) |
198
+ | mk | 231103 (1.31%) | 6478376 (1.13%) | 953 (1.16%) | 7713 (0.39%) | 232056 (1.31%) | 6486089 (1.12%) |
199
+ | mr | 228596 (1.30%) | 5886608 (1.02%) | 776 (0.94%) | 6332 (0.32%) | 229372 (1.30%) | 5892940 (1.02%) |
200
+ | no | 223605 (1.27%) | 6137131 (1.07%) | 1396 (1.70%) | 40226 (2.05%) | 225001 (1.27%) | 6177357 (1.07%) |
201
+ | da | 222243 (1.26%) | 5375746 (0.94%) | 1201 (1.46%) | 10373 (0.53%) | 223444 (1.26%) | 5386119 (0.93%) |
202
+ | hy | 207937 (1.18%) | 6345675 (1.10%) | 791 (0.96%) | 9276 (0.47%) | 208728 (1.18%) | 6354951 (1.10%) |
203
+ | tl | 207674 (1.18%) | 5561702 (0.97%) | 1017 (1.24%) | 10926 (0.56%) | 208691 (1.18%) | 5572628 (0.97%) |
204
+ | hi | 206552 (1.17%) | 7796062 (1.36%) | 1079 (1.31%) | 47351 (2.42%) | 207631 (1.17%) | 7843413 (1.36%) |
205
+ | ko | 205625 (1.17%) | 6481034 (1.13%) | 1156 (1.41%) | 32355 (1.65%) | 206781 (1.17%) | 6513389 (1.13%) |
206
+ | el | 202334 (1.15%) | 7105554 (1.24%) | 826 (1.00%) | 13412 (0.68%) | 203160 (1.15%) | 7118966 (1.23%) |
207
+ | ro | 194999 (1.11%) | 6206913 (1.08%) | 820 (1.00%) | 14575 (0.74%) | 195819 (1.11%) | 6221488 (1.08%) |
208
+ | fa | 192050 (1.09%) | 5728246 (1.00%) | 696 (0.85%) | 14765 (0.75%) | 192746 (1.09%) | 5743011 (1.00%) |
209
+ | sk | 189330 (1.08%) | 5318617 (0.93%) | 873 (1.06%) | 20779 (1.06%) | 190203 (1.08%) | 5339396 (0.93%) |
210
+ | la | 188201 (1.07%) | 4591159 (0.80%) | 824 (1.00%) | 8557 (0.44%) | 189025 (1.07%) | 4599716 (0.80%) |
211
+ | bg | 187685 (1.07%) | 5860353 (1.02%) | 762 (0.93%) | 16804 (0.86%) | 188447 (1.07%) | 5877157 (1.02%) |
212
+ | be | 181543 (1.03%) | 6528657 (1.14%) | 869 (1.06%) | 25944 (1.32%) | 182412 (1.03%) | 6554601 (1.14%) |
213
+ | is | 180452 (1.03%) | 6146455 (1.07%) | 959 (1.17%) | 39591 (2.02%) | 181411 (1.03%) | 6186046 (1.07%) |
214
+ | lv | 179142 (1.02%) | 5867897 (1.02%) | 762 (0.93%) | 33481 (1.71%) | 179904 (1.02%) | 5901378 (1.02%) |
215
+ | ckb | 174282 (0.99%) | 7825141 (1.36%) | 667 (0.81%) | 28756 (1.47%) | 174949 (0.99%) | 7853897 (1.36%) |
216
+ | ms | 172573 (0.98%) | 4614764 (0.80%) | 815 (0.99%) | 24769 (1.26%) | 173388 (0.98%) | 4639533 (0.80%) |
217
+ | ka | 170876 (0.97%) | 5505127 (0.96%) | 673 (0.82%) | 20651 (1.05%) | 171549 (0.97%) | 5525778 (0.96%) |
218
+ | kk | 170695 (0.97%) | 5132560 (0.89%) | 676 (0.82%) | 18695 (0.95%) | 171371 (0.97%) | 5151255 (0.89%) |
219
+ | bn | 170721 (0.97%) | 6393448 (1.11%) | 441 (0.54%) | 14727 (0.75%) | 171162 (0.97%) | 6408175 (1.11%) |
220
+ | eu | 168462 (0.96%) | 5737310 (1.00%) | 746 (0.91%) | 37196 (1.90%) | 169208 (0.96%) | 5774506 (1.00%) |
221
+ | as | 168746 (0.96%) | 8564682 (1.49%) | 445 (0.54%) | 24444 (1.25%) | 169191 (0.96%) | 8589126 (1.49%) |
222
+ | mn | 167543 (0.95%) | 5678049 (0.99%) | 703 (0.85%) | 20347 (1.04%) | 168246 (0.95%) | 5698396 (0.99%) |
223
+ | ur | 165992 (0.94%) | 5361179 (0.93%) | 684 (0.83%) | 22622 (1.16%) | 166676 (0.94%) | 5383801 (0.93%) |
224
+ | oc | 165863 (0.94%) | 5735536 (1.00%) | 730 (0.89%) | 18599 (0.95%) | 166593 (0.94%) | 5754135 (1.00%) |
225
+ | ba | 164919 (0.94%) | 8387828 (1.46%) | 699 (0.85%) | 35927 (1.83%) | 165618 (0.94%) | 8423755 (1.46%) |
226
+ | th | 164429 (0.93%) | 5495248 (0.96%) | 649 (0.79%) | 22113 (1.13%) | 165078 (0.93%) | 5517361 (0.96%) |
227
+ | ky | 164374 (0.93%) | 5199548 (0.90%) | 683 (0.83%) | 18956 (0.97%) | 165057 (0.93%) | 5218504 (0.90%) |
228
+ | hr | 163828 (0.93%) | 5183677 (0.90%) | 711 (0.86%) | 33845 (1.73%) | 164539 (0.93%) | 5217522 (0.90%) |
229
+ | ps | 163238 (0.93%) | 4735113 (0.82%) | 674 (0.82%) | 18515 (0.95%) | 163912 (0.93%) | 4753628 (0.82%) |
230
+ | id | 163187 (0.93%) | 4025079 (0.70%) | 723 (0.88%) | 13371 (0.68%) | 163910 (0.93%) | 4038450 (0.70%) |
231
+ | pa | 162180 (0.92%) | 7621059 (1.33%) | 581 (0.71%) | 29036 (1.48%) | 162761 (0.92%) | 7650095 (1.33%) |
232
+ | sw | 161777 (0.92%) | 5013161 (0.87%) | 653 (0.79%) | 26493 (1.35%) | 162430 (0.92%) | 5039654 (0.87%) |
233
+ | af | 160455 (0.91%) | 4676798 (0.81%) | 932 (1.13%) | 27369 (1.40%) | 161387 (0.91%) | 4704167 (0.82%) |
234
+ | jv | 156292 (0.89%) | 4752381 (0.83%) | 576 (0.70%) | 22573 (1.15%) | 156868 (0.89%) | 4774954 (0.83%) |
235
+ | tt | 154833 (0.88%) | 5165763 (0.90%) | 578 (0.70%) | 7298 (0.37%) | 155411 (0.88%) | 5173061 (0.90%) |
236
+ | cy | 153551 (0.87%) | 5656404 (0.98%) | 653 (0.79%) | 29503 (1.51%) | 154204 (0.87%) | 5685907 (0.99%) |
237
+ | ga | 150458 (0.86%) | 5488243 (0.95%) | 680 (0.83%) | 33471 (1.71%) | 151138 (0.86%) | 5521714 (0.96%) |
238
+ | kn | 150184 (0.85%) | 14992479 (2.61%) | 697 (0.85%) | 49288 (2.52%) | 150881 (0.85%) | 15041767 (2.61%) |
239
+ | bs | 150037 (0.85%) | 4582900 (0.80%) | 649 (0.79%) | 25588 (1.31%) | 150686 (0.85%) | 4608488 (0.80%) |
240
+ | ca | 149401 (0.85%) | 5477662 (0.95%) | 629 (0.76%) | 21391 (1.09%) | 150030 (0.85%) | 5499053 (0.95%) |
241
+ | ne | 148716 (0.85%) | 4855198 (0.84%) | 535 (0.65%) | 16246 (0.83%) | 149251 (0.84%) | 4871444 (0.84%) |
242
+ | ku | 147702 (0.84%) | 4973601 (0.87%) | 574 (0.70%) | 28196 (1.44%) | 148276 (0.84%) | 5001797 (0.87%) |
243
+ | gl | 147011 (0.84%) | 4554907 (0.79%) | 658 (0.80%) | 20127 (1.03%) | 147669 (0.84%) | 4575034 (0.79%) |
244
+ | uz | 145433 (0.83%) | 4704898 (0.82%) | 573 (0.70%) | 21862 (1.12%) | 146006 (0.83%) | 4726760 (0.82%) |
245
+ | sl | 144084 (0.82%) | 3851696 (0.67%) | 651 (0.79%) | 18164 (0.93%) | 144735 (0.82%) | 3869860 (0.67%) |
246
+ | sv | 143041 (0.81%) | 4006332 (0.70%) | 905 (1.10%) | 7012 (0.36%) | 143946 (0.81%) | 4013344 (0.70%) |
247
+ | tg | 136703 (0.78%) | 7664329 (1.33%) | 572 (0.70%) | 34220 (1.75%) | 137275 (0.78%) | 7698549 (1.33%) |
248
+ | et | 131007 (0.74%) | 3280590 (0.57%) | 549 (0.67%) | 14021 (0.72%) | 131556 (0.74%) | 3294611 (0.57%) |
249
+ | br | 130223 (0.74%) | 4495403 (0.78%) | 546 (0.66%) | 17304 (0.88%) | 130769 (0.74%) | 4512707 (0.78%) |
250
+ | lb | 129528 (0.74%) | 4421411 (0.77%) | 495 (0.60%) | 17761 (0.91%) | 130023 (0.74%) | 4439172 (0.77%) |
251
+ | su | 129144 (0.73%) | 4215719 (0.73%) | 535 (0.65%) | 21391 (1.09%) | 129679 (0.73%) | 4237110 (0.73%) |
252
+ | mt | 128626 (0.73%) | 6671441 (1.16%) | 508 (0.62%) | 26729 (1.36%) | 129134 (0.73%) | 6698170 (1.16%) |
253
+ | sq | 119431 (0.68%) | 4107917 (0.71%) | 561 (0.68%) | 18633 (0.95%) | 119992 (0.68%) | 4126550 (0.72%) |
254
+ | sr | 117855 (0.67%) | 3160599 (0.55%) | 427 (0.52%) | 3505 (0.18%) | 118282 (0.67%) | 3164104 (0.55%) |
255
+ | or | 110709 (0.63%) | 3922431 (0.68%) | 410 (0.50%) | 13094 (0.67%) | 111119 (0.63%) | 3935525 (0.68%) |
256
+ | ml | 110085 (0.63%) | 10929013 (1.90%) | 464 (0.56%) | 36922 (1.89%) | 110549 (0.63%) | 10965935 (1.90%) |
257
+ | yi | 104494 (0.59%) | 4085563 (0.71%) | 400 (0.49%) | 6005 (0.31%) | 104894 (0.59%) | 4091568 (0.71%) |
258
+ | te | 101076 (0.57%) | 9757033 (1.70%) | 430 (0.52%) | 37897 (1.94%) | 101506 (0.57%) | 9794930 (1.70%) |
259
+ | ta | 94122 (0.53%) | 7917169 (1.38%) | 386 (0.47%) | 26610 (1.36%) | 94508 (0.53%) | 7943779 (1.38%) |
260
+ | mg | 93939 (0.53%) | 3291017 (0.57%) | 391 (0.48%) | 11698 (0.60%) | 94330 (0.53%) | 3302715 (0.57%) |
261
+ | si | 92723 (0.53%) | 5275463 (0.92%) | 364 (0.44%) | 18426 (0.94%) | 93087 (0.53%) | 5293889 (0.92%) |
262
+ | vi | 74916 (0.43%) | 2535825 (0.44%) | 335 (0.41%) | 3396 (0.17%) | 75251 (0.43%) | 2539221 (0.44%) |
263
+ | rm | 74806 (0.43%) | 2826708 (0.49%) | 318 (0.39%) | 12654 (0.65%) | 75124 (0.43%) | 2839362 (0.49%) |
264
+ | gu | 70961 (0.40%) | 7859622 (1.37%) | 335 (0.41%) | 28389 (1.45%) | 71296 (0.40%) | 7888011 (1.37%) |
265
+ | bo | 69565 (0.40%) | 1378245 (0.24%) | 263 (0.32%) | 5407 (0.28%) | 69828 (0.40%) | 1383652 (0.24%) |
266
+ | ug | 64297 (0.37%) | 1427585 (0.25%) | 260 (0.32%) | 4769 (0.24%) | 64557 (0.37%) | 1432354 (0.25%) |
267
+ | dv | 60328 (0.34%) | 1557497 (0.27%) | 215 (0.26%) | 5844 (0.30%) | 60543 (0.34%) | 1563341 (0.27%) |
268
+ | am | 59339 (0.34%) | 2705311 (0.47%) | 235 (0.29%) | 10768 (0.55%) | 59574 (0.34%) | 2716079 (0.47%) |
269
+ | yo | 59246 (0.34%) | 3649130 (0.63%) | 260 (0.32%) | 21157 (1.08%) | 59506 (0.34%) | 3670287 (0.64%) |
270
+ | my | 58575 (0.33%) | 2165089 (0.38%) | 214 (0.26%) | 8142 (0.42%) | 58789 (0.33%) | 2173231 (0.38%) |
271
+ | km | 57081 (0.32%) | 3056236 (0.53%) | 193 (0.23%) | 10606 (0.54%) | 57274 (0.32%) | 3066842 (0.53%) |
272
+ | so | 56160 (0.32%) | 2044409 (0.36%) | 212 (0.26%) | 8847 (0.45%) | 56372 (0.32%) | 2053256 (0.36%) |
273
+ | sd | 55359 (0.31%) | 3226018 (0.56%) | 217 (0.26%) | 10847 (0.55%) | 55576 (0.31%) | 3236865 (0.56%) |
274
+ | zu | 52465 (0.30%) | 2406841 (0.42%) | 203 (0.25%) | 9751 (0.50%) | 52668 (0.30%) | 2416592 (0.42%) |
275
+ | lo | 50641 (0.29%) | 1747495 (0.30%) | 189 (0.23%) | 6221 (0.32%) | 50830 (0.29%) | 1753716 (0.30%) |
276
+ | ti | 47785 (0.27%) | 2895617 (0.50%) | 195 (0.24%) | 12699 (0.65%) | 47980 (0.27%) | 2908316 (0.50%) |
277
+ | ce | 45014 (0.26%) | 2425219 (0.42%) | 188 (0.23%) | 9950 (0.51%) | 45202 (0.26%) | 2435169 (0.42%) |
278
+ | ny | 43552 (0.25%) | 2051132 (0.36%) | 171 (0.21%) | 8286 (0.42%) | 43723 (0.25%) | 2059418 (0.36%) |
279
+ | gd | 36623 (0.21%) | 1273243 (0.22%) | 156 (0.19%) | 3615 (0.18%) | 36779 (0.21%) | 1276858 (0.22%) |
280
+ | xh | 24432 (0.14%) | 911850 (0.16%) | 93 (0.11%) | 3528 (0.18%) | 24525 (0.14%) | 915378 (0.16%) |
281
+ | om | 15372 (0.09%) | 545603 (0.09%) | 77 (0.09%) | 2564 (0.13%) | 15449 (0.09%) | 548167 (0.10%) |
282
+ | sco | 8772 (0.05%) | 233030 (0.04%) | 37 (0.04%) | 828 (0.04%) | 8809 (0.05%) | 233858 (0.04%) |
283
+ | **total** | 17593786 (100.00%) | 574735483 (100.00%) | 82270 (100.00%) | 1958217 (100.00%) | 17676056 (100.00%) | 576693700 (100.00%) |
284
 
285
 
286
+ This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on an unknown dataset.
287
  It achieves the following results on the evaluation set:
288
+ - Loss: 0.0306
289
+ - Precision: 0.9507
290
+ - Recall: 0.9644
291
+ - F1: 0.9575
292
+ - Accuracy: 0.9917
293
 
294
  ## Training procedure
295
 
 
297
 
298
  The following hyperparameters were used during training:
299
  - learning_rate: 5e-05
300
+ - train_batch_size: 72
301
+ - eval_batch_size: 36
302
  - seed: 42
303
+ - gradient_accumulation_steps: 2
304
  - total_train_batch_size: 144
305
  - optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
306
  - lr_scheduler_type: linear
 
311
 
312
  | Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
313
  |:-------------:|:------:|:-----:|:---------------:|:---------:|:------:|:------:|:--------:|
314
+ | 0.0918 | 0.0731 | 2500 | 0.1050 | 0.7984 | 0.8818 | 0.8381 | 0.9735 |
315
+ | 0.0717 | 0.1463 | 5000 | 0.0797 | 0.8393 | 0.9041 | 0.8705 | 0.9782 |
316
+ | 0.0624 | 0.2194 | 7500 | 0.0762 | 0.8664 | 0.9166 | 0.8908 | 0.9804 |
317
+ | 0.0562 | 0.2925 | 10000 | 0.0620 | 0.8758 | 0.9247 | 0.8995 | 0.9830 |
318
+ | 0.0516 | 0.3657 | 12500 | 0.0576 | 0.8844 | 0.9298 | 0.9065 | 0.9845 |
319
+ | 0.0465 | 0.4388 | 15000 | 0.0543 | 0.8993 | 0.9357 | 0.9172 | 0.9857 |
320
+ | 0.0433 | 0.5119 | 17500 | 0.0558 | 0.9005 | 0.9356 | 0.9177 | 0.9856 |
321
+ | 0.0411 | 0.5851 | 20000 | 0.0499 | 0.9012 | 0.9385 | 0.9195 | 0.9867 |
322
+ | 0.0420 | 0.6582 | 22500 | 0.0460 | 0.9167 | 0.9438 | 0.9300 | 0.9873 |
323
+ | 0.0392 | 0.7313 | 25000 | 0.0441 | 0.9149 | 0.9448 | 0.9296 | 0.9878 |
324
+ | 0.0386 | 0.8045 | 27500 | 0.0434 | 0.9200 | 0.9476 | 0.9336 | 0.9885 |
325
+ | 0.0357 | 0.8776 | 30000 | 0.0422 | 0.9235 | 0.9503 | 0.9367 | 0.9886 |
326
+ | 0.0356 | 0.9507 | 32500 | 0.0404 | 0.9272 | 0.9520 | 0.9395 | 0.9890 |
327
+ | 0.0261 | 1.0238 | 35000 | 0.0381 | 0.9293 | 0.9529 | 0.9409 | 0.9898 |
328
+ | 0.0322 | 1.0970 | 37500 | 0.0371 | 0.9346 | 0.9558 | 0.9451 | 0.9899 |
329
+ | 0.0303 | 1.1701 | 40000 | 0.0374 | 0.9375 | 0.9580 | 0.9476 | 0.9903 |
330
+ | 0.0276 | 1.2432 | 42500 | 0.0378 | 0.9355 | 0.9566 | 0.9460 | 0.9901 |
331
+ | 0.0264 | 1.3164 | 45000 | 0.0353 | 0.9373 | 0.9574 | 0.9472 | 0.9904 |
332
+ | 0.0228 | 1.3895 | 47500 | 0.0366 | 0.9398 | 0.9589 | 0.9493 | 0.9903 |
333
+ | 0.0234 | 1.4626 | 50000 | 0.0343 | 0.9430 | 0.9602 | 0.9516 | 0.9907 |
334
+ | 0.0274 | 1.5358 | 52500 | 0.0339 | 0.9396 | 0.9591 | 0.9492 | 0.9906 |
335
+ | 0.0236 | 1.6089 | 55000 | 0.0324 | 0.9438 | 0.9613 | 0.9525 | 0.9913 |
336
+ | 0.0244 | 1.6820 | 57500 | 0.0322 | 0.9478 | 0.9624 | 0.9551 | 0.9914 |
337
+ | 0.0222 | 1.7552 | 60000 | 0.0323 | 0.9483 | 0.9628 | 0.9555 | 0.9914 |
338
+ | 0.0238 | 1.8283 | 62500 | 0.0320 | 0.9480 | 0.9630 | 0.9554 | 0.9913 |
339
+ | 0.0223 | 1.9014 | 65000 | 0.0320 | 0.9485 | 0.9637 | 0.9560 | 0.9913 |
340
+ | 0.0208 | 1.9746 | 67500 | 0.0306 | 0.9507 | 0.9644 | 0.9575 | 0.9917 |
341
 
342
 
343
  ### Framework versions
 
345
  - Transformers 5.0.0
346
  - Pytorch 2.10.0+cu128
347
  - Datasets 4.0.0
348
+ - Tokenizers 0.22.2