Update README.md
Browse files
README.md
CHANGED
|
@@ -1,693 +1,695 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: cc-by-nc-4.0
|
| 3 |
-
language:
|
| 4 |
-
- ro
|
| 5 |
-
base_model:
|
| 6 |
-
- google/gemma-7b
|
| 7 |
-
datasets:
|
| 8 |
-
- OpenLLM-Ro/ro_sft_alpaca
|
| 9 |
-
- OpenLLM-Ro/ro_sft_alpaca_gpt4
|
| 10 |
-
- OpenLLM-Ro/ro_sft_dolly
|
| 11 |
-
- OpenLLM-Ro/ro_sft_selfinstruct_gpt4
|
| 12 |
-
- OpenLLM-Ro/ro_sft_norobots
|
| 13 |
-
- OpenLLM-Ro/ro_sft_orca
|
| 14 |
-
- OpenLLM-Ro/ro_sft_camel
|
| 15 |
-
- OpenLLM-Ro/ro_sft_oasst
|
| 16 |
-
- OpenLLM-Ro/ro_sft_ultrachat
|
| 17 |
-
- OpenLLM-Ro/ro_sft_magpie_mt
|
| 18 |
-
- OpenLLM-Ro/ro_sft_magpie_reasoning
|
| 19 |
-
model-index:
|
| 20 |
-
- name: OpenLLM-Ro/RoGemma-7b-Instruct-2025-04-23
|
| 21 |
-
results:
|
| 22 |
-
- task:
|
| 23 |
-
type: text-generation
|
| 24 |
-
dataset:
|
| 25 |
-
name: RoMT-Bench
|
| 26 |
-
type: RoMT-Bench
|
| 27 |
-
metrics:
|
| 28 |
-
- name: Score
|
| 29 |
-
type: Score
|
| 30 |
-
value: 6.28
|
| 31 |
-
- task:
|
| 32 |
-
type: text-generation
|
| 33 |
-
dataset:
|
| 34 |
-
name: RoCulturaBench
|
| 35 |
-
type: RoCulturaBench
|
| 36 |
-
metrics:
|
| 37 |
-
- name: Score
|
| 38 |
-
type: Score
|
| 39 |
-
value: 3.65
|
| 40 |
-
- task:
|
| 41 |
-
type: text-generation
|
| 42 |
-
dataset:
|
| 43 |
-
name: Romanian_Academic_Benchmarks
|
| 44 |
-
type: Romanian_Academic_Benchmarks
|
| 45 |
-
metrics:
|
| 46 |
-
- name: Average accuracy
|
| 47 |
-
type: accuracy
|
| 48 |
-
value: 50.52
|
| 49 |
-
- task:
|
| 50 |
-
type: text-generation
|
| 51 |
-
dataset:
|
| 52 |
-
name: OpenLLM-Ro/ro_arc_challenge
|
| 53 |
-
type: OpenLLM-Ro/ro_arc_challenge
|
| 54 |
-
metrics:
|
| 55 |
-
- name: Average accuracy
|
| 56 |
-
type: accuracy
|
| 57 |
-
value: 47.70
|
| 58 |
-
- task:
|
| 59 |
-
type: text-generation
|
| 60 |
-
dataset:
|
| 61 |
-
name: OpenLLM-Ro/ro_mmlu
|
| 62 |
-
type: OpenLLM-Ro/ro_mmlu
|
| 63 |
-
metrics:
|
| 64 |
-
- name: Average accuracy
|
| 65 |
-
type: accuracy
|
| 66 |
-
value: 51.66
|
| 67 |
-
- task:
|
| 68 |
-
type: text-generation
|
| 69 |
-
dataset:
|
| 70 |
-
name: OpenLLM-Ro/ro_winogrande
|
| 71 |
-
type: OpenLLM-Ro/ro_winogrande
|
| 72 |
-
metrics:
|
| 73 |
-
- name: Average accuracy
|
| 74 |
-
type: accuracy
|
| 75 |
-
value: 66.32
|
| 76 |
-
- task:
|
| 77 |
-
type: text-generation
|
| 78 |
-
dataset:
|
| 79 |
-
name: OpenLLM-Ro/ro_hellaswag
|
| 80 |
-
type: OpenLLM-Ro/ro_hellaswag
|
| 81 |
-
metrics:
|
| 82 |
-
- name: Average accuracy
|
| 83 |
-
type: accuracy
|
| 84 |
-
value: 53.59
|
| 85 |
-
- task:
|
| 86 |
-
type: text-generation
|
| 87 |
-
dataset:
|
| 88 |
-
name: OpenLLM-Ro/ro_gsm8k
|
| 89 |
-
type: OpenLLM-Ro/ro_gsm8k
|
| 90 |
-
metrics:
|
| 91 |
-
- name: Average accuracy
|
| 92 |
-
type: accuracy
|
| 93 |
-
value: 36.04
|
| 94 |
-
- task:
|
| 95 |
-
type: text-generation
|
| 96 |
-
dataset:
|
| 97 |
-
name: OpenLLM-Ro/ro_truthfulqa
|
| 98 |
-
type: OpenLLM-Ro/ro_truthfulqa
|
| 99 |
-
metrics:
|
| 100 |
-
- name: Average accuracy
|
| 101 |
-
type: accuracy
|
| 102 |
-
value: 47.81
|
| 103 |
-
- task:
|
| 104 |
-
type: text-generation
|
| 105 |
-
dataset:
|
| 106 |
-
name: LaRoSeDa_binary
|
| 107 |
-
type: LaRoSeDa_binary
|
| 108 |
-
metrics:
|
| 109 |
-
- name: Average macro-f1
|
| 110 |
-
type: macro-f1
|
| 111 |
-
value: 95.44
|
| 112 |
-
- task:
|
| 113 |
-
type: text-generation
|
| 114 |
-
dataset:
|
| 115 |
-
name: LaRoSeDa_multiclass
|
| 116 |
-
type: LaRoSeDa_multiclass
|
| 117 |
-
metrics:
|
| 118 |
-
- name: Average macro-f1
|
| 119 |
-
type: macro-f1
|
| 120 |
-
value: 59.24
|
| 121 |
-
- task:
|
| 122 |
-
type: text-generation
|
| 123 |
-
dataset:
|
| 124 |
-
name: WMT_EN-RO
|
| 125 |
-
type: WMT_EN-RO
|
| 126 |
-
metrics:
|
| 127 |
-
- name: Average bleu
|
| 128 |
-
type: bleu
|
| 129 |
-
value: 25.17
|
| 130 |
-
- task:
|
| 131 |
-
type: text-generation
|
| 132 |
-
dataset:
|
| 133 |
-
name: WMT_RO-EN
|
| 134 |
-
type: WMT_RO-EN
|
| 135 |
-
metrics:
|
| 136 |
-
- name: Average bleu
|
| 137 |
-
type: bleu
|
| 138 |
-
value: 21.17
|
| 139 |
-
- task:
|
| 140 |
-
type: text-generation
|
| 141 |
-
dataset:
|
| 142 |
-
name: XQuAD
|
| 143 |
-
type: XQuAD
|
| 144 |
-
metrics:
|
| 145 |
-
- name: Average exact_match
|
| 146 |
-
type: exact_match
|
| 147 |
-
value: 15.88
|
| 148 |
-
- task:
|
| 149 |
-
type: text-generation
|
| 150 |
-
dataset:
|
| 151 |
-
name: XQuAD
|
| 152 |
-
type: XQuAD
|
| 153 |
-
metrics:
|
| 154 |
-
- name: Average f1
|
| 155 |
-
type: f1
|
| 156 |
-
value: 29.16
|
| 157 |
-
- task:
|
| 158 |
-
type: text-generation
|
| 159 |
-
dataset:
|
| 160 |
-
name: STS
|
| 161 |
-
type: STS
|
| 162 |
-
metrics:
|
| 163 |
-
- name: Average spearman
|
| 164 |
-
type: spearman
|
| 165 |
-
value: 75.90
|
| 166 |
-
- task:
|
| 167 |
-
type: text-generation
|
| 168 |
-
dataset:
|
| 169 |
-
name: STS
|
| 170 |
-
type: STS
|
| 171 |
-
metrics:
|
| 172 |
-
- name: Average pearson
|
| 173 |
-
type: pearson
|
| 174 |
-
value: 75.16
|
| 175 |
-
- task:
|
| 176 |
-
type: text-generation
|
| 177 |
-
dataset:
|
| 178 |
-
name: RoMT-Bench
|
| 179 |
-
type: RoMT-Bench
|
| 180 |
-
metrics:
|
| 181 |
-
- name: First turn
|
| 182 |
-
type: Score
|
| 183 |
-
value: 6.97
|
| 184 |
-
- name: Second turn
|
| 185 |
-
type: Score
|
| 186 |
-
value: 5.58
|
| 187 |
-
- task:
|
| 188 |
-
type: text-generation
|
| 189 |
-
dataset:
|
| 190 |
-
name: OpenLLM-Ro/ro_arc_challenge
|
| 191 |
-
type: OpenLLM-Ro/ro_arc_challenge
|
| 192 |
-
metrics:
|
| 193 |
-
- name: 0-shot
|
| 194 |
-
type: accuracy
|
| 195 |
-
value: 46.19
|
| 196 |
-
- name: 1-shot
|
| 197 |
-
type: accuracy
|
| 198 |
-
value: 46.53
|
| 199 |
-
- name: 3-shot
|
| 200 |
-
type: accuracy
|
| 201 |
-
value: 46.02
|
| 202 |
-
- name: 5-shot
|
| 203 |
-
type: accuracy
|
| 204 |
-
value: 48.33
|
| 205 |
-
- name: 10-shot
|
| 206 |
-
type: accuracy
|
| 207 |
-
value: 49.27
|
| 208 |
-
- name: 25-shot
|
| 209 |
-
type: accuracy
|
| 210 |
-
value: 49.87
|
| 211 |
-
- task:
|
| 212 |
-
type: text-generation
|
| 213 |
-
dataset:
|
| 214 |
-
name: OpenLLM-Ro/ro_mmlu
|
| 215 |
-
type: OpenLLM-Ro/ro_mmlu
|
| 216 |
-
metrics:
|
| 217 |
-
- name: 0-shot
|
| 218 |
-
type: accuracy
|
| 219 |
-
value: 51.13
|
| 220 |
-
- name: 1-shot
|
| 221 |
-
type: accuracy
|
| 222 |
-
value: 50.94
|
| 223 |
-
- name: 3-shot
|
| 224 |
-
type: accuracy
|
| 225 |
-
value: 52.67
|
| 226 |
-
- name: 5-shot
|
| 227 |
-
type: accuracy
|
| 228 |
-
value: 51.90
|
| 229 |
-
- task:
|
| 230 |
-
type: text-generation
|
| 231 |
-
dataset:
|
| 232 |
-
name: OpenLLM-Ro/ro_winogrande
|
| 233 |
-
type: OpenLLM-Ro/ro_winogrande
|
| 234 |
-
metrics:
|
| 235 |
-
- name: 0-shot
|
| 236 |
-
type: accuracy
|
| 237 |
-
value: 67.40
|
| 238 |
-
- name: 1-shot
|
| 239 |
-
type: accuracy
|
| 240 |
-
value: 65.04
|
| 241 |
-
- name: 3-shot
|
| 242 |
-
type: accuracy
|
| 243 |
-
value: 65.67
|
| 244 |
-
- name: 5-shot
|
| 245 |
-
type: accuracy
|
| 246 |
-
value: 67.17
|
| 247 |
-
- task:
|
| 248 |
-
type: text-generation
|
| 249 |
-
dataset:
|
| 250 |
-
name: OpenLLM-Ro/ro_hellaswag
|
| 251 |
-
type: OpenLLM-Ro/ro_hellaswag
|
| 252 |
-
metrics:
|
| 253 |
-
- name: 0-shot
|
| 254 |
-
type: accuracy
|
| 255 |
-
value: 58.03
|
| 256 |
-
- name: 1-shot
|
| 257 |
-
type: accuracy
|
| 258 |
-
value: 56.63
|
| 259 |
-
- name: 3-shot
|
| 260 |
-
type: accuracy
|
| 261 |
-
value: 52.47
|
| 262 |
-
- name: 5-shot
|
| 263 |
-
type: accuracy
|
| 264 |
-
value: 48.63
|
| 265 |
-
- name: 10-shot
|
| 266 |
-
type: accuracy
|
| 267 |
-
value: 52.18
|
| 268 |
-
- task:
|
| 269 |
-
type: text-generation
|
| 270 |
-
dataset:
|
| 271 |
-
name: OpenLLM-Ro/ro_gsm8k
|
| 272 |
-
type: OpenLLM-Ro/ro_gsm8k
|
| 273 |
-
metrics:
|
| 274 |
-
- name: 1-shot
|
| 275 |
-
type: accuracy
|
| 276 |
-
value: 24.11
|
| 277 |
-
- name: 3-shot
|
| 278 |
-
type: accuracy
|
| 279 |
-
value: 37.76
|
| 280 |
-
- name: 5-shot
|
| 281 |
-
type: accuracy
|
| 282 |
-
value: 46.25
|
| 283 |
-
- task:
|
| 284 |
-
type: text-generation
|
| 285 |
-
dataset:
|
| 286 |
-
name: LaRoSeDa_binary
|
| 287 |
-
type: LaRoSeDa_binary
|
| 288 |
-
metrics:
|
| 289 |
-
- name: 0-shot
|
| 290 |
-
type: macro-f1
|
| 291 |
-
value: 96.33
|
| 292 |
-
- name: 1-shot
|
| 293 |
-
type: macro-f1
|
| 294 |
-
value: 94.62
|
| 295 |
-
- name: 3-shot
|
| 296 |
-
type: macro-f1
|
| 297 |
-
value: 95.06
|
| 298 |
-
- name: 5-shot
|
| 299 |
-
type: macro-f1
|
| 300 |
-
value: 95.76
|
| 301 |
-
- task:
|
| 302 |
-
type: text-generation
|
| 303 |
-
dataset:
|
| 304 |
-
name: LaRoSeDa_multiclass
|
| 305 |
-
type: LaRoSeDa_multiclass
|
| 306 |
-
metrics:
|
| 307 |
-
- name: 0-shot
|
| 308 |
-
type: macro-f1
|
| 309 |
-
value: 43.65
|
| 310 |
-
- name: 1-shot
|
| 311 |
-
type: macro-f1
|
| 312 |
-
value: 64.30
|
| 313 |
-
- name: 3-shot
|
| 314 |
-
type: macro-f1
|
| 315 |
-
value: 64.22
|
| 316 |
-
- name: 5-shot
|
| 317 |
-
type: macro-f1
|
| 318 |
-
value: 64.81
|
| 319 |
-
- task:
|
| 320 |
-
type: text-generation
|
| 321 |
-
dataset:
|
| 322 |
-
name: WMT_EN-RO
|
| 323 |
-
type: WMT_EN-RO
|
| 324 |
-
metrics:
|
| 325 |
-
- name: 0-shot
|
| 326 |
-
type: bleu
|
| 327 |
-
value: 13.30
|
| 328 |
-
- name: 1-shot
|
| 329 |
-
type: bleu
|
| 330 |
-
value: 28.59
|
| 331 |
-
- name: 3-shot
|
| 332 |
-
type: bleu
|
| 333 |
-
value: 29.48
|
| 334 |
-
- name: 5-shot
|
| 335 |
-
type: bleu
|
| 336 |
-
value: 29.31
|
| 337 |
-
- task:
|
| 338 |
-
type: text-generation
|
| 339 |
-
dataset:
|
| 340 |
-
name: WMT_RO-EN
|
| 341 |
-
type: WMT_RO-EN
|
| 342 |
-
metrics:
|
| 343 |
-
- name: 0-shot
|
| 344 |
-
type: bleu
|
| 345 |
-
value: 1.11
|
| 346 |
-
- name: 1-shot
|
| 347 |
-
type: bleu
|
| 348 |
-
value: 18.97
|
| 349 |
-
- name: 3-shot
|
| 350 |
-
type: bleu
|
| 351 |
-
value: 31.99
|
| 352 |
-
- name: 5-shot
|
| 353 |
-
type: bleu
|
| 354 |
-
value: 32.60
|
| 355 |
-
- task:
|
| 356 |
-
type: text-generation
|
| 357 |
-
dataset:
|
| 358 |
-
name: XQuAD_EM
|
| 359 |
-
type: XQuAD_EM
|
| 360 |
-
metrics:
|
| 361 |
-
- name: 0-shot
|
| 362 |
-
type: exact_match
|
| 363 |
-
value: 17.31
|
| 364 |
-
- name: 1-shot
|
| 365 |
-
type: exact_match
|
| 366 |
-
value: 12.44
|
| 367 |
-
- name: 3-shot
|
| 368 |
-
type: exact_match
|
| 369 |
-
value: 13.11
|
| 370 |
-
- name: 5-shot
|
| 371 |
-
type: exact_match
|
| 372 |
-
value: 20.67
|
| 373 |
-
- task:
|
| 374 |
-
type: text-generation
|
| 375 |
-
dataset:
|
| 376 |
-
name: XQuAD_F1
|
| 377 |
-
type: XQuAD_F1
|
| 378 |
-
metrics:
|
| 379 |
-
- name: 0-shot
|
| 380 |
-
type: f1
|
| 381 |
-
value: 29.90
|
| 382 |
-
- name: 1-shot
|
| 383 |
-
type: f1
|
| 384 |
-
value: 24.24
|
| 385 |
-
- name: 3-shot
|
| 386 |
-
type: f1
|
| 387 |
-
value: 25.64
|
| 388 |
-
- name: 5-shot
|
| 389 |
-
type: f1
|
| 390 |
-
value: 36.86
|
| 391 |
-
- task:
|
| 392 |
-
type: text-generation
|
| 393 |
-
dataset:
|
| 394 |
-
name: STS_Spearman
|
| 395 |
-
type: STS_Spearman
|
| 396 |
-
metrics:
|
| 397 |
-
- name: 1-shot
|
| 398 |
-
type: spearman
|
| 399 |
-
value: 76.50
|
| 400 |
-
- name: 3-shot
|
| 401 |
-
type: spearman
|
| 402 |
-
value: 73.63
|
| 403 |
-
- name: 5-shot
|
| 404 |
-
type: spearman
|
| 405 |
-
value: 77.58
|
| 406 |
-
- task:
|
| 407 |
-
type: text-generation
|
| 408 |
-
dataset:
|
| 409 |
-
name: STS_Pearson
|
| 410 |
-
type: STS_Pearson
|
| 411 |
-
metrics:
|
| 412 |
-
- name: 1-shot
|
| 413 |
-
type: pearson
|
| 414 |
-
value: 75.15
|
| 415 |
-
- name: 3-shot
|
| 416 |
-
type: pearson
|
| 417 |
-
value: 72.69
|
| 418 |
-
- name: 5-shot
|
| 419 |
-
type: pearson
|
| 420 |
-
value: 77.63
|
| 421 |
-
|
| 422 |
-
---
|
| 423 |
-
|
| 424 |
-
# Model Card for Model ID
|
| 425 |
-
|
| 426 |
-
|
| 427 |
-
|
| 428 |
-
|
| 429 |
-
|
| 430 |
-
|
| 431 |
-
|
| 432 |
-
|
| 433 |
-
|
| 434 |
-
|
| 435 |
-
|
| 436 |
-
|
| 437 |
-
|
| 438 |
-
|
| 439 |
-
|
| 440 |
-
|
| 441 |
-
<!-- - **
|
| 442 |
-
- **
|
| 443 |
-
- **
|
| 444 |
-
- **
|
| 445 |
-
- **
|
| 446 |
-
|
| 447 |
-
|
| 448 |
-
|
| 449 |
-
|
| 450 |
-
|
| 451 |
-
|
| 452 |
-
|
| 453 |
-
|
| 454 |
-
|
| 455 |
-
|
| 456 |
-
|
| 457 |
-
|
| 458 |
-
|
| 459 |
-
|
| 460 |
-
|
| 461 |
-
|
| 462 |
-
|
| 463 |
-
|
| 464 |
-
|
| 465 |
-
|
| 466 |
-
|
| 467 |
-
|
| 468 |
-
|
| 469 |
-
|
| 470 |
-
|
| 471 |
-
|
| 472 |
-
|
| 473 |
-
|
| 474 |
-
|
| 475 |
-
|
| 476 |
-
|
| 477 |
-
|
| 478 |
-
|
| 479 |
-
|
| 480 |
-
|
| 481 |
-
|
| 482 |
-
|
| 483 |
-
|
| 484 |
-
|
| 485 |
-
|
| 486 |
-
|
| 487 |
-
|
| 488 |
-
|
| 489 |
-
|
| 490 |
-
|
| 491 |
-
|
| 492 |
-
|
| 493 |
-
|
| 494 |
-
<
|
| 495 |
-
<
|
| 496 |
-
<
|
| 497 |
-
<td><strong
|
| 498 |
-
<td><strong><center>
|
| 499 |
-
<td><strong><center>
|
| 500 |
-
<td><strong><center>
|
| 501 |
-
<td><strong><center>
|
| 502 |
-
<td><strong><center>
|
| 503 |
-
</
|
| 504 |
-
<
|
| 505 |
-
|
| 506 |
-
|
| 507 |
-
<
|
| 508 |
-
|
| 509 |
-
|
| 510 |
-
<
|
| 511 |
-
|
| 512 |
-
|
| 513 |
-
<
|
| 514 |
-
|
| 515 |
-
|
| 516 |
-
<
|
| 517 |
-
|
| 518 |
-
|
| 519 |
-
</
|
| 520 |
-
</
|
| 521 |
-
|
| 522 |
-
|
| 523 |
-
|
| 524 |
-
|
| 525 |
-
|
| 526 |
-
<
|
| 527 |
-
<
|
| 528 |
-
<
|
| 529 |
-
<td
|
| 530 |
-
</
|
| 531 |
-
<
|
| 532 |
-
|
| 533 |
-
<
|
| 534 |
-
<td
|
| 535 |
-
<td colspan="2"><center><strong>Few-shot</strong></center></td>
|
| 536 |
-
<td colspan="2"><center><strong>Finetuned</strong></center></td>
|
| 537 |
-
</
|
| 538 |
-
<
|
| 539 |
-
|
| 540 |
-
<
|
| 541 |
-
<td><
|
| 542 |
-
<td><center><strong>Binary<br>(Macro F1)</strong></center></td>
|
| 543 |
-
<td><center><strong>Multiclass<br>(Macro F1)</strong></center></td>
|
| 544 |
-
<td><center><strong>
|
| 545 |
-
<td><center><strong>
|
| 546 |
-
<td><center><strong>EN-RO<br>(Bleu)</strong></center></td>
|
| 547 |
-
<td><center><strong>RO-EN<br>(Bleu)</strong></center>
|
| 548 |
-
</
|
| 549 |
-
<
|
| 550 |
-
|
| 551 |
-
|
| 552 |
-
<
|
| 553 |
-
|
| 554 |
-
|
| 555 |
-
<
|
| 556 |
-
|
| 557 |
-
|
| 558 |
-
<
|
| 559 |
-
|
| 560 |
-
|
| 561 |
-
<
|
| 562 |
-
|
| 563 |
-
|
| 564 |
-
</
|
| 565 |
-
</
|
| 566 |
-
|
| 567 |
-
|
| 568 |
-
|
| 569 |
-
|
| 570 |
-
<
|
| 571 |
-
<
|
| 572 |
-
<
|
| 573 |
-
<td
|
| 574 |
-
</
|
| 575 |
-
<
|
| 576 |
-
|
| 577 |
-
<
|
| 578 |
-
<td
|
| 579 |
-
<td colspan="2"><center><strong>Few-shot</strong></center></td>
|
| 580 |
-
<td colspan="2"><center><strong>Finetuned</strong></center></td>
|
| 581 |
-
</
|
| 582 |
-
<
|
| 583 |
-
|
| 584 |
-
<
|
| 585 |
-
<td><
|
| 586 |
-
<td><center><strong>(EM)</strong></center></td>
|
| 587 |
-
<td><center><strong>(F1)</strong></center></td>
|
| 588 |
-
<td><center><strong>(
|
| 589 |
-
<td><center><strong>(
|
| 590 |
-
<td><center><strong>(Spearman)</strong></center></td>
|
| 591 |
-
<td><center><strong>(Pearson)</strong></center></td>
|
| 592 |
-
</
|
| 593 |
-
<
|
| 594 |
-
|
| 595 |
-
|
| 596 |
-
<
|
| 597 |
-
|
| 598 |
-
|
| 599 |
-
<
|
| 600 |
-
|
| 601 |
-
|
| 602 |
-
<
|
| 603 |
-
|
| 604 |
-
|
| 605 |
-
<
|
| 606 |
-
|
| 607 |
-
|
| 608 |
-
</
|
| 609 |
-
</
|
| 610 |
-
|
| 611 |
-
|
| 612 |
-
|
| 613 |
-
|
| 614 |
-
|
| 615 |
-
|
| 616 |
-
<
|
| 617 |
-
<
|
| 618 |
-
<
|
| 619 |
-
<td><strong
|
| 620 |
-
<td><strong><center>
|
| 621 |
-
<td><strong><center>
|
| 622 |
-
</
|
| 623 |
-
<
|
| 624 |
-
|
| 625 |
-
|
| 626 |
-
<
|
| 627 |
-
|
| 628 |
-
|
| 629 |
-
<
|
| 630 |
-
|
| 631 |
-
|
| 632 |
-
<
|
| 633 |
-
|
| 634 |
-
|
| 635 |
-
<
|
| 636 |
-
|
| 637 |
-
|
| 638 |
-
</
|
| 639 |
-
</
|
| 640 |
-
|
| 641 |
-
|
| 642 |
-
|
| 643 |
-
|
| 644 |
-
|
| 645 |
-
<
|
| 646 |
-
<
|
| 647 |
-
<
|
| 648 |
-
<td><strong
|
| 649 |
-
</
|
| 650 |
-
<
|
| 651 |
-
|
| 652 |
-
|
| 653 |
-
<
|
| 654 |
-
|
| 655 |
-
|
| 656 |
-
<
|
| 657 |
-
|
| 658 |
-
|
| 659 |
-
<
|
| 660 |
-
|
| 661 |
-
|
| 662 |
-
<
|
| 663 |
-
|
| 664 |
-
|
| 665 |
-
</
|
| 666 |
-
</
|
| 667 |
-
|
| 668 |
-
|
| 669 |
-
|
| 670 |
-
|
| 671 |
-
|
| 672 |
-
|
|
| 673 |
-
|
| 674 |
-
|
| 675 |
-
|RoGemma-7b-Instruct-
|
| 676 |
-
|
| 677 |
-
|
| 678 |
-
|
| 679 |
-
|
| 680 |
-
|
| 681 |
-
|
| 682 |
-
|
| 683 |
-
|
| 684 |
-
|
| 685 |
-
|
| 686 |
-
|
| 687 |
-
|
| 688 |
-
|
| 689 |
-
}
|
| 690 |
-
|
| 691 |
-
|
| 692 |
-
|
|
|
|
|
|
|
| 693 |
[More Information Needed] -->
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: cc-by-nc-4.0
|
| 3 |
+
language:
|
| 4 |
+
- ro
|
| 5 |
+
base_model:
|
| 6 |
+
- google/gemma-7b
|
| 7 |
+
datasets:
|
| 8 |
+
- OpenLLM-Ro/ro_sft_alpaca
|
| 9 |
+
- OpenLLM-Ro/ro_sft_alpaca_gpt4
|
| 10 |
+
- OpenLLM-Ro/ro_sft_dolly
|
| 11 |
+
- OpenLLM-Ro/ro_sft_selfinstruct_gpt4
|
| 12 |
+
- OpenLLM-Ro/ro_sft_norobots
|
| 13 |
+
- OpenLLM-Ro/ro_sft_orca
|
| 14 |
+
- OpenLLM-Ro/ro_sft_camel
|
| 15 |
+
- OpenLLM-Ro/ro_sft_oasst
|
| 16 |
+
- OpenLLM-Ro/ro_sft_ultrachat
|
| 17 |
+
- OpenLLM-Ro/ro_sft_magpie_mt
|
| 18 |
+
- OpenLLM-Ro/ro_sft_magpie_reasoning
|
| 19 |
+
model-index:
|
| 20 |
+
- name: OpenLLM-Ro/RoGemma-7b-Instruct-2025-04-23
|
| 21 |
+
results:
|
| 22 |
+
- task:
|
| 23 |
+
type: text-generation
|
| 24 |
+
dataset:
|
| 25 |
+
name: RoMT-Bench
|
| 26 |
+
type: RoMT-Bench
|
| 27 |
+
metrics:
|
| 28 |
+
- name: Score
|
| 29 |
+
type: Score
|
| 30 |
+
value: 6.28
|
| 31 |
+
- task:
|
| 32 |
+
type: text-generation
|
| 33 |
+
dataset:
|
| 34 |
+
name: RoCulturaBench
|
| 35 |
+
type: RoCulturaBench
|
| 36 |
+
metrics:
|
| 37 |
+
- name: Score
|
| 38 |
+
type: Score
|
| 39 |
+
value: 3.65
|
| 40 |
+
- task:
|
| 41 |
+
type: text-generation
|
| 42 |
+
dataset:
|
| 43 |
+
name: Romanian_Academic_Benchmarks
|
| 44 |
+
type: Romanian_Academic_Benchmarks
|
| 45 |
+
metrics:
|
| 46 |
+
- name: Average accuracy
|
| 47 |
+
type: accuracy
|
| 48 |
+
value: 50.52
|
| 49 |
+
- task:
|
| 50 |
+
type: text-generation
|
| 51 |
+
dataset:
|
| 52 |
+
name: OpenLLM-Ro/ro_arc_challenge
|
| 53 |
+
type: OpenLLM-Ro/ro_arc_challenge
|
| 54 |
+
metrics:
|
| 55 |
+
- name: Average accuracy
|
| 56 |
+
type: accuracy
|
| 57 |
+
value: 47.70
|
| 58 |
+
- task:
|
| 59 |
+
type: text-generation
|
| 60 |
+
dataset:
|
| 61 |
+
name: OpenLLM-Ro/ro_mmlu
|
| 62 |
+
type: OpenLLM-Ro/ro_mmlu
|
| 63 |
+
metrics:
|
| 64 |
+
- name: Average accuracy
|
| 65 |
+
type: accuracy
|
| 66 |
+
value: 51.66
|
| 67 |
+
- task:
|
| 68 |
+
type: text-generation
|
| 69 |
+
dataset:
|
| 70 |
+
name: OpenLLM-Ro/ro_winogrande
|
| 71 |
+
type: OpenLLM-Ro/ro_winogrande
|
| 72 |
+
metrics:
|
| 73 |
+
- name: Average accuracy
|
| 74 |
+
type: accuracy
|
| 75 |
+
value: 66.32
|
| 76 |
+
- task:
|
| 77 |
+
type: text-generation
|
| 78 |
+
dataset:
|
| 79 |
+
name: OpenLLM-Ro/ro_hellaswag
|
| 80 |
+
type: OpenLLM-Ro/ro_hellaswag
|
| 81 |
+
metrics:
|
| 82 |
+
- name: Average accuracy
|
| 83 |
+
type: accuracy
|
| 84 |
+
value: 53.59
|
| 85 |
+
- task:
|
| 86 |
+
type: text-generation
|
| 87 |
+
dataset:
|
| 88 |
+
name: OpenLLM-Ro/ro_gsm8k
|
| 89 |
+
type: OpenLLM-Ro/ro_gsm8k
|
| 90 |
+
metrics:
|
| 91 |
+
- name: Average accuracy
|
| 92 |
+
type: accuracy
|
| 93 |
+
value: 36.04
|
| 94 |
+
- task:
|
| 95 |
+
type: text-generation
|
| 96 |
+
dataset:
|
| 97 |
+
name: OpenLLM-Ro/ro_truthfulqa
|
| 98 |
+
type: OpenLLM-Ro/ro_truthfulqa
|
| 99 |
+
metrics:
|
| 100 |
+
- name: Average accuracy
|
| 101 |
+
type: accuracy
|
| 102 |
+
value: 47.81
|
| 103 |
+
- task:
|
| 104 |
+
type: text-generation
|
| 105 |
+
dataset:
|
| 106 |
+
name: LaRoSeDa_binary
|
| 107 |
+
type: LaRoSeDa_binary
|
| 108 |
+
metrics:
|
| 109 |
+
- name: Average macro-f1
|
| 110 |
+
type: macro-f1
|
| 111 |
+
value: 95.44
|
| 112 |
+
- task:
|
| 113 |
+
type: text-generation
|
| 114 |
+
dataset:
|
| 115 |
+
name: LaRoSeDa_multiclass
|
| 116 |
+
type: LaRoSeDa_multiclass
|
| 117 |
+
metrics:
|
| 118 |
+
- name: Average macro-f1
|
| 119 |
+
type: macro-f1
|
| 120 |
+
value: 59.24
|
| 121 |
+
- task:
|
| 122 |
+
type: text-generation
|
| 123 |
+
dataset:
|
| 124 |
+
name: WMT_EN-RO
|
| 125 |
+
type: WMT_EN-RO
|
| 126 |
+
metrics:
|
| 127 |
+
- name: Average bleu
|
| 128 |
+
type: bleu
|
| 129 |
+
value: 25.17
|
| 130 |
+
- task:
|
| 131 |
+
type: text-generation
|
| 132 |
+
dataset:
|
| 133 |
+
name: WMT_RO-EN
|
| 134 |
+
type: WMT_RO-EN
|
| 135 |
+
metrics:
|
| 136 |
+
- name: Average bleu
|
| 137 |
+
type: bleu
|
| 138 |
+
value: 21.17
|
| 139 |
+
- task:
|
| 140 |
+
type: text-generation
|
| 141 |
+
dataset:
|
| 142 |
+
name: XQuAD
|
| 143 |
+
type: XQuAD
|
| 144 |
+
metrics:
|
| 145 |
+
- name: Average exact_match
|
| 146 |
+
type: exact_match
|
| 147 |
+
value: 15.88
|
| 148 |
+
- task:
|
| 149 |
+
type: text-generation
|
| 150 |
+
dataset:
|
| 151 |
+
name: XQuAD
|
| 152 |
+
type: XQuAD
|
| 153 |
+
metrics:
|
| 154 |
+
- name: Average f1
|
| 155 |
+
type: f1
|
| 156 |
+
value: 29.16
|
| 157 |
+
- task:
|
| 158 |
+
type: text-generation
|
| 159 |
+
dataset:
|
| 160 |
+
name: STS
|
| 161 |
+
type: STS
|
| 162 |
+
metrics:
|
| 163 |
+
- name: Average spearman
|
| 164 |
+
type: spearman
|
| 165 |
+
value: 75.90
|
| 166 |
+
- task:
|
| 167 |
+
type: text-generation
|
| 168 |
+
dataset:
|
| 169 |
+
name: STS
|
| 170 |
+
type: STS
|
| 171 |
+
metrics:
|
| 172 |
+
- name: Average pearson
|
| 173 |
+
type: pearson
|
| 174 |
+
value: 75.16
|
| 175 |
+
- task:
|
| 176 |
+
type: text-generation
|
| 177 |
+
dataset:
|
| 178 |
+
name: RoMT-Bench
|
| 179 |
+
type: RoMT-Bench
|
| 180 |
+
metrics:
|
| 181 |
+
- name: First turn
|
| 182 |
+
type: Score
|
| 183 |
+
value: 6.97
|
| 184 |
+
- name: Second turn
|
| 185 |
+
type: Score
|
| 186 |
+
value: 5.58
|
| 187 |
+
- task:
|
| 188 |
+
type: text-generation
|
| 189 |
+
dataset:
|
| 190 |
+
name: OpenLLM-Ro/ro_arc_challenge
|
| 191 |
+
type: OpenLLM-Ro/ro_arc_challenge
|
| 192 |
+
metrics:
|
| 193 |
+
- name: 0-shot
|
| 194 |
+
type: accuracy
|
| 195 |
+
value: 46.19
|
| 196 |
+
- name: 1-shot
|
| 197 |
+
type: accuracy
|
| 198 |
+
value: 46.53
|
| 199 |
+
- name: 3-shot
|
| 200 |
+
type: accuracy
|
| 201 |
+
value: 46.02
|
| 202 |
+
- name: 5-shot
|
| 203 |
+
type: accuracy
|
| 204 |
+
value: 48.33
|
| 205 |
+
- name: 10-shot
|
| 206 |
+
type: accuracy
|
| 207 |
+
value: 49.27
|
| 208 |
+
- name: 25-shot
|
| 209 |
+
type: accuracy
|
| 210 |
+
value: 49.87
|
| 211 |
+
- task:
|
| 212 |
+
type: text-generation
|
| 213 |
+
dataset:
|
| 214 |
+
name: OpenLLM-Ro/ro_mmlu
|
| 215 |
+
type: OpenLLM-Ro/ro_mmlu
|
| 216 |
+
metrics:
|
| 217 |
+
- name: 0-shot
|
| 218 |
+
type: accuracy
|
| 219 |
+
value: 51.13
|
| 220 |
+
- name: 1-shot
|
| 221 |
+
type: accuracy
|
| 222 |
+
value: 50.94
|
| 223 |
+
- name: 3-shot
|
| 224 |
+
type: accuracy
|
| 225 |
+
value: 52.67
|
| 226 |
+
- name: 5-shot
|
| 227 |
+
type: accuracy
|
| 228 |
+
value: 51.90
|
| 229 |
+
- task:
|
| 230 |
+
type: text-generation
|
| 231 |
+
dataset:
|
| 232 |
+
name: OpenLLM-Ro/ro_winogrande
|
| 233 |
+
type: OpenLLM-Ro/ro_winogrande
|
| 234 |
+
metrics:
|
| 235 |
+
- name: 0-shot
|
| 236 |
+
type: accuracy
|
| 237 |
+
value: 67.40
|
| 238 |
+
- name: 1-shot
|
| 239 |
+
type: accuracy
|
| 240 |
+
value: 65.04
|
| 241 |
+
- name: 3-shot
|
| 242 |
+
type: accuracy
|
| 243 |
+
value: 65.67
|
| 244 |
+
- name: 5-shot
|
| 245 |
+
type: accuracy
|
| 246 |
+
value: 67.17
|
| 247 |
+
- task:
|
| 248 |
+
type: text-generation
|
| 249 |
+
dataset:
|
| 250 |
+
name: OpenLLM-Ro/ro_hellaswag
|
| 251 |
+
type: OpenLLM-Ro/ro_hellaswag
|
| 252 |
+
metrics:
|
| 253 |
+
- name: 0-shot
|
| 254 |
+
type: accuracy
|
| 255 |
+
value: 58.03
|
| 256 |
+
- name: 1-shot
|
| 257 |
+
type: accuracy
|
| 258 |
+
value: 56.63
|
| 259 |
+
- name: 3-shot
|
| 260 |
+
type: accuracy
|
| 261 |
+
value: 52.47
|
| 262 |
+
- name: 5-shot
|
| 263 |
+
type: accuracy
|
| 264 |
+
value: 48.63
|
| 265 |
+
- name: 10-shot
|
| 266 |
+
type: accuracy
|
| 267 |
+
value: 52.18
|
| 268 |
+
- task:
|
| 269 |
+
type: text-generation
|
| 270 |
+
dataset:
|
| 271 |
+
name: OpenLLM-Ro/ro_gsm8k
|
| 272 |
+
type: OpenLLM-Ro/ro_gsm8k
|
| 273 |
+
metrics:
|
| 274 |
+
- name: 1-shot
|
| 275 |
+
type: accuracy
|
| 276 |
+
value: 24.11
|
| 277 |
+
- name: 3-shot
|
| 278 |
+
type: accuracy
|
| 279 |
+
value: 37.76
|
| 280 |
+
- name: 5-shot
|
| 281 |
+
type: accuracy
|
| 282 |
+
value: 46.25
|
| 283 |
+
- task:
|
| 284 |
+
type: text-generation
|
| 285 |
+
dataset:
|
| 286 |
+
name: LaRoSeDa_binary
|
| 287 |
+
type: LaRoSeDa_binary
|
| 288 |
+
metrics:
|
| 289 |
+
- name: 0-shot
|
| 290 |
+
type: macro-f1
|
| 291 |
+
value: 96.33
|
| 292 |
+
- name: 1-shot
|
| 293 |
+
type: macro-f1
|
| 294 |
+
value: 94.62
|
| 295 |
+
- name: 3-shot
|
| 296 |
+
type: macro-f1
|
| 297 |
+
value: 95.06
|
| 298 |
+
- name: 5-shot
|
| 299 |
+
type: macro-f1
|
| 300 |
+
value: 95.76
|
| 301 |
+
- task:
|
| 302 |
+
type: text-generation
|
| 303 |
+
dataset:
|
| 304 |
+
name: LaRoSeDa_multiclass
|
| 305 |
+
type: LaRoSeDa_multiclass
|
| 306 |
+
metrics:
|
| 307 |
+
- name: 0-shot
|
| 308 |
+
type: macro-f1
|
| 309 |
+
value: 43.65
|
| 310 |
+
- name: 1-shot
|
| 311 |
+
type: macro-f1
|
| 312 |
+
value: 64.30
|
| 313 |
+
- name: 3-shot
|
| 314 |
+
type: macro-f1
|
| 315 |
+
value: 64.22
|
| 316 |
+
- name: 5-shot
|
| 317 |
+
type: macro-f1
|
| 318 |
+
value: 64.81
|
| 319 |
+
- task:
|
| 320 |
+
type: text-generation
|
| 321 |
+
dataset:
|
| 322 |
+
name: WMT_EN-RO
|
| 323 |
+
type: WMT_EN-RO
|
| 324 |
+
metrics:
|
| 325 |
+
- name: 0-shot
|
| 326 |
+
type: bleu
|
| 327 |
+
value: 13.30
|
| 328 |
+
- name: 1-shot
|
| 329 |
+
type: bleu
|
| 330 |
+
value: 28.59
|
| 331 |
+
- name: 3-shot
|
| 332 |
+
type: bleu
|
| 333 |
+
value: 29.48
|
| 334 |
+
- name: 5-shot
|
| 335 |
+
type: bleu
|
| 336 |
+
value: 29.31
|
| 337 |
+
- task:
|
| 338 |
+
type: text-generation
|
| 339 |
+
dataset:
|
| 340 |
+
name: WMT_RO-EN
|
| 341 |
+
type: WMT_RO-EN
|
| 342 |
+
metrics:
|
| 343 |
+
- name: 0-shot
|
| 344 |
+
type: bleu
|
| 345 |
+
value: 1.11
|
| 346 |
+
- name: 1-shot
|
| 347 |
+
type: bleu
|
| 348 |
+
value: 18.97
|
| 349 |
+
- name: 3-shot
|
| 350 |
+
type: bleu
|
| 351 |
+
value: 31.99
|
| 352 |
+
- name: 5-shot
|
| 353 |
+
type: bleu
|
| 354 |
+
value: 32.60
|
| 355 |
+
- task:
|
| 356 |
+
type: text-generation
|
| 357 |
+
dataset:
|
| 358 |
+
name: XQuAD_EM
|
| 359 |
+
type: XQuAD_EM
|
| 360 |
+
metrics:
|
| 361 |
+
- name: 0-shot
|
| 362 |
+
type: exact_match
|
| 363 |
+
value: 17.31
|
| 364 |
+
- name: 1-shot
|
| 365 |
+
type: exact_match
|
| 366 |
+
value: 12.44
|
| 367 |
+
- name: 3-shot
|
| 368 |
+
type: exact_match
|
| 369 |
+
value: 13.11
|
| 370 |
+
- name: 5-shot
|
| 371 |
+
type: exact_match
|
| 372 |
+
value: 20.67
|
| 373 |
+
- task:
|
| 374 |
+
type: text-generation
|
| 375 |
+
dataset:
|
| 376 |
+
name: XQuAD_F1
|
| 377 |
+
type: XQuAD_F1
|
| 378 |
+
metrics:
|
| 379 |
+
- name: 0-shot
|
| 380 |
+
type: f1
|
| 381 |
+
value: 29.90
|
| 382 |
+
- name: 1-shot
|
| 383 |
+
type: f1
|
| 384 |
+
value: 24.24
|
| 385 |
+
- name: 3-shot
|
| 386 |
+
type: f1
|
| 387 |
+
value: 25.64
|
| 388 |
+
- name: 5-shot
|
| 389 |
+
type: f1
|
| 390 |
+
value: 36.86
|
| 391 |
+
- task:
|
| 392 |
+
type: text-generation
|
| 393 |
+
dataset:
|
| 394 |
+
name: STS_Spearman
|
| 395 |
+
type: STS_Spearman
|
| 396 |
+
metrics:
|
| 397 |
+
- name: 1-shot
|
| 398 |
+
type: spearman
|
| 399 |
+
value: 76.50
|
| 400 |
+
- name: 3-shot
|
| 401 |
+
type: spearman
|
| 402 |
+
value: 73.63
|
| 403 |
+
- name: 5-shot
|
| 404 |
+
type: spearman
|
| 405 |
+
value: 77.58
|
| 406 |
+
- task:
|
| 407 |
+
type: text-generation
|
| 408 |
+
dataset:
|
| 409 |
+
name: STS_Pearson
|
| 410 |
+
type: STS_Pearson
|
| 411 |
+
metrics:
|
| 412 |
+
- name: 1-shot
|
| 413 |
+
type: pearson
|
| 414 |
+
value: 75.15
|
| 415 |
+
- name: 3-shot
|
| 416 |
+
type: pearson
|
| 417 |
+
value: 72.69
|
| 418 |
+
- name: 5-shot
|
| 419 |
+
type: pearson
|
| 420 |
+
value: 77.63
|
| 421 |
+
|
| 422 |
+
---
|
| 423 |
+
|
| 424 |
+
# Model Card for Model ID
|
| 425 |
+
|
| 426 |
+
This model points/is identical to [RoGemma-7b-Instruct-2025-04-23](https://huggingface.co/OpenLLM-Ro/RoGemma-7b-Instruct-2025-04-23).
|
| 427 |
+
|
| 428 |
+
<!-- Provide a quick summary of what the model is/does. -->
|
| 429 |
+
|
| 430 |
+
RoGemma is a family of pretrained and fine-tuned generative text models for Romanian. This is the repository for the **instruct 7B model**. Links to other models can be found at the bottom of this page.
|
| 431 |
+
|
| 432 |
+
## Model Details
|
| 433 |
+
|
| 434 |
+
### Model Description
|
| 435 |
+
|
| 436 |
+
<!-- Provide a longer summary of what this model is. -->
|
| 437 |
+
OpenLLM-Ro represents the first open-source effort to build a LLM specialized for Romanian. OpenLLM-Ro developed and publicly releases a collection of Romanian LLMs, both in the form of foundational model and instruct and chat variants.
|
| 438 |
+
|
| 439 |
+
|
| 440 |
+
- **Developed by:** OpenLLM-Ro
|
| 441 |
+
<!-- - **Funded by [optional]:** [More Information Needed] -->
|
| 442 |
+
<!-- - **Shared by [optional]:** [More Information Needed] -->
|
| 443 |
+
<!-- - **Model type:** [More Information Needed] -->
|
| 444 |
+
- **Language(s):** Romanian
|
| 445 |
+
- **License:** cc-by-nc-4.0
|
| 446 |
+
- **Finetuned from model:** [gemma-7b](https://huggingface.co/google/gemma-7b)
|
| 447 |
+
- **Trained using:** [RoAlpaca](https://huggingface.co/datasets/OpenLLM-Ro/ro_sft_alpaca), [RoAlpacaGPT4](https://huggingface.co/datasets/OpenLLM-Ro/ro_sft_alpaca_gpt4), [RoDolly](https://huggingface.co/datasets/OpenLLM-Ro/ro_sft_dolly), [RoSelfInstruct](https://huggingface.co/datasets/OpenLLM-Ro/ro_sft_selfinstruct_gpt4), [RoNoRobots](https://huggingface.co/datasets/OpenLLM-Ro/ro_sft_norobots), [RoOrca](https://huggingface.co/datasets/OpenLLM-Ro/ro_sft_orca), [RoCamel](https://huggingface.co/datasets/OpenLLM-Ro/ro_sft_camel), [RoOpenAssistant](https://huggingface.co/datasets/OpenLLM-Ro/ro_sft_oasst), [RoUltraChat](https://huggingface.co/datasets/OpenLLM-Ro/ro_sft_ultrachat), [RoMagpiePro](https://huggingface.co/datasets/OpenLLM-Ro/ro_sft_magpie_mt), [RoMagpieReasoning](https://huggingface.co/datasets/OpenLLM-Ro/ro_sft_magpie_reasoning)
|
| 448 |
+
|
| 449 |
+
|
| 450 |
+
### Model Sources
|
| 451 |
+
|
| 452 |
+
<!-- Provide the basic links for the model. -->
|
| 453 |
+
|
| 454 |
+
- **Repository:** https://github.com/OpenLLM-Ro/LLaMA-Factory
|
| 455 |
+
- **Paper:** https://arxiv.org/abs/2406.18266
|
| 456 |
+
|
| 457 |
+
## Intended Use
|
| 458 |
+
|
| 459 |
+
### Intended Use Cases
|
| 460 |
+
|
| 461 |
+
RoGemma is intented for research use in Romanian. Base models can be adapted for a variety of natural language tasks while instruction and chat tuned models are intended for assistant-like chat.
|
| 462 |
+
|
| 463 |
+
### Out-of-Scope Use
|
| 464 |
+
|
| 465 |
+
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
|
| 466 |
+
|
| 467 |
+
Use in any manner that violates the license, any applicable laws or regluations, use in languages other than Romanian.
|
| 468 |
+
|
| 469 |
+
|
| 470 |
+
|
| 471 |
+
## How to Get Started with the Model
|
| 472 |
+
|
| 473 |
+
Use the code below to get started with the model.
|
| 474 |
+
|
| 475 |
+
```python
|
| 476 |
+
from transformers import AutoTokenizer, AutoModelForCausalLM
|
| 477 |
+
|
| 478 |
+
tokenizer = AutoTokenizer.from_pretrained("OpenLLM-Ro/RoGemma-7b-Instruct")
|
| 479 |
+
model = AutoModelForCausalLM.from_pretrained("OpenLLM-Ro/RoGemma-7b-Instruct")
|
| 480 |
+
|
| 481 |
+
instruction = "Ce jocuri de societate pot juca cu prietenii mei?"
|
| 482 |
+
chat = [
|
| 483 |
+
{"role": "user", "content": instruction},
|
| 484 |
+
]
|
| 485 |
+
prompt = tokenizer.apply_chat_template(chat, tokenize=False, system_message="")
|
| 486 |
+
|
| 487 |
+
inputs = tokenizer.encode(prompt, add_special_tokens=False, return_tensors="pt")
|
| 488 |
+
outputs = model.generate(input_ids=inputs, max_new_tokens=128)
|
| 489 |
+
print(tokenizer.decode(outputs[0]))
|
| 490 |
+
```
|
| 491 |
+
|
| 492 |
+
## Academic Benchmarks
|
| 493 |
+
|
| 494 |
+
<table>
|
| 495 |
+
<tbody>
|
| 496 |
+
<tr>
|
| 497 |
+
<td><strong>Model</strong></td>
|
| 498 |
+
<td><strong><center>Average</center></strong></td>
|
| 499 |
+
<td><strong><center>ARC</center></strong></td>
|
| 500 |
+
<td><strong><center>MMLU</center></strong></td>
|
| 501 |
+
<td><strong><center>Winogrande</center></strong></td>
|
| 502 |
+
<td><strong><center>Hellaswag</center></strong></td>
|
| 503 |
+
<td><strong><center>GSM8k</center></strong></td>
|
| 504 |
+
<td><strong><center>TruthfulQA</center></strong></td>
|
| 505 |
+
</tr>
|
| 506 |
+
<tr>
|
| 507 |
+
<td>gemma-1.1-7b-it</td><td><center>41.44</center></td><td><center>40.32</center></td><td><center>47.22</center></td><td><center>55.01</center></td><td><center>47.03</center></td><td><center>9.50</center></td><td><center>49.58</center></td>
|
| 508 |
+
</tr>
|
| 509 |
+
<tr>
|
| 510 |
+
<td>RoGemma-7b-Instruct-2024-06-28</td><td><center><strong>53.41</strong></center></td><td><center><strong>52.44</strong></center></td><td><center>54.44</center></td><td><center><strong>69.36</strong></center></td><td><center><strong>61.96</strong></center></td><td><center>31.06</center></td><td><center><strong>51.23</strong></center></td>
|
| 511 |
+
</tr>
|
| 512 |
+
<tr>
|
| 513 |
+
<td>RoGemma-7b-Instruct-2024-10-09</td><td><center>50.48</center></td><td><center>52.01</center></td><td><center>52.37</center></td><td><center>66.97</center></td><td><center>56.34</center></td><td><center>25.98</center></td><td><center>49.18</center></td>
|
| 514 |
+
</tr>
|
| 515 |
+
<tr>
|
| 516 |
+
<td><em>RoGemma-7b-Instruct-2025-04-23</em></td><td><center><em>50.52</em></center></td><td><center><em>47.70</em></center></td><td><center><em>51.66</em></center></td><td><center><em>66.32</em></center></td><td><center><em>53.59</em></center></td><td><center><em><strong>36.04</strong></em></center></td><td><center><em>47.81</em></center></td>
|
| 517 |
+
</tr>
|
| 518 |
+
<tr>
|
| 519 |
+
<td>RoGemma-7b-Instruct-DPO-2024-10-09</td><td><center>48.27</center></td><td><center>46.66</center></td><td><center><strong>54.45</strong></center></td><td><center>63.73</center></td><td><center>49.33</center></td><td><center>34.98</center></td><td><center>40.45</center></td>
|
| 520 |
+
</tr>
|
| 521 |
+
</tbody>
|
| 522 |
+
</table>
|
| 523 |
+
|
| 524 |
+
## Downstream tasks
|
| 525 |
+
|
| 526 |
+
<table>
|
| 527 |
+
<tbody>
|
| 528 |
+
<tr>
|
| 529 |
+
<td></td>
|
| 530 |
+
<td colspan="4"><center><strong>LaRoSeDa</strong></center></td>
|
| 531 |
+
<td colspan="4"><center><strong>WMT</strong></center></td>
|
| 532 |
+
</tr>
|
| 533 |
+
<tr>
|
| 534 |
+
<td></td>
|
| 535 |
+
<td colspan="2"><center><strong>Few-shot</strong></center></td>
|
| 536 |
+
<td colspan="2"><center><strong>Finetuned</strong></center></td>
|
| 537 |
+
<td colspan="2"><center><strong>Few-shot</strong></center></td>
|
| 538 |
+
<td colspan="2"><center><strong>Finetuned</strong></center></td>
|
| 539 |
+
</tr>
|
| 540 |
+
<tr>
|
| 541 |
+
<td><strong>Model</strong></td>
|
| 542 |
+
<td><center><strong>Binary<br>(Macro F1)</strong></center></td>
|
| 543 |
+
<td><center><strong>Multiclass<br>(Macro F1)</strong></center></td>
|
| 544 |
+
<td><center><strong>Binary<br>(Macro F1)</strong></center></td>
|
| 545 |
+
<td><center><strong>Multiclass<br>(Macro F1)</strong></center></td>
|
| 546 |
+
<td><center><strong>EN-RO<br>(Bleu)</strong></center></td>
|
| 547 |
+
<td><center><strong>RO-EN<br>(Bleu)</strong></center></td>
|
| 548 |
+
<td><center><strong>EN-RO<br>(Bleu)</strong></center></td>
|
| 549 |
+
<td><center><strong>RO-EN<br>(Bleu)</strong></center>
|
| 550 |
+
</tr>
|
| 551 |
+
<tr>
|
| 552 |
+
<td>gemma-1.1-7b-it</td><td><center>87.54</center></td><td><center>51.48</center></td><td><center>83.87</center></td><td><center>85.61</center></td><td><center>17.96</center></td><td><center><strong>27.74</strong></center></td><td><center>25.48</center></td><td><center>36.11</center></td>
|
| 553 |
+
</tr>
|
| 554 |
+
<tr>
|
| 555 |
+
<td>RoGemma-7b-Instruct-2024-06-28</td><td><center><strong>97.86</strong></center></td><td><center><strong>65.70</strong></center></td><td><center>98.43</center></td><td><center><strong>87.17</strong></center></td><td><center><strong>27.91</strong></center></td><td><center>23.08</center></td><td><center><strong>27.99</strong></center></td><td><center><strong>39.51</strong></center></td>
|
| 556 |
+
</tr>
|
| 557 |
+
<tr>
|
| 558 |
+
<td>RoGemma-7b-Instruct-2024-10-09</td><td><center>86.96</center></td><td><center>56.72</center></td><td><center><strong>98.80</strong></center></td><td><center>85.81</center></td><td><center>24.45</center></td><td><center>14.20</center></td><td><center>25.96</center></td><td><center>39.07</center></td>
|
| 559 |
+
</tr>
|
| 560 |
+
<tr>
|
| 561 |
+
<td><em>RoGemma-7b-Instruct-2025-04-23</em></td><td><center><em>95.44</em></center></td><td><center><em>59.24</em></center></td><td><center><em>-</em></center></td><td><center><em>-</em></center></td><td><center><em>25.17</em></center></td><td><center><em>21.17</em></center></td><td><center><em>-</em></center></td><td><center><em>-</em></center></td>
|
| 562 |
+
</tr>
|
| 563 |
+
<tr>
|
| 564 |
+
<td>RoGemma-7b-Instruct-DPO-2024-10-09</td><td><center>96.45</center></td><td><center>63.23</center></td><td><center>-</center></td><td><center>-</center></td><td><center>20.73</center></td><td><center>7.87</center></td><td><center>-</center></td><td><center>-</center></td>
|
| 565 |
+
</tr>
|
| 566 |
+
</tbody>
|
| 567 |
+
</table>
|
| 568 |
+
|
| 569 |
+
|
| 570 |
+
<table>
|
| 571 |
+
<tbody>
|
| 572 |
+
<tr>
|
| 573 |
+
<td></td>
|
| 574 |
+
<td colspan="4"><center><strong>XQuAD</strong></center></td>
|
| 575 |
+
<td colspan="4"><center><strong>STS</strong></center></td>
|
| 576 |
+
</tr>
|
| 577 |
+
<tr>
|
| 578 |
+
<td></td>
|
| 579 |
+
<td colspan="2"><center><strong>Few-shot</strong></center></td>
|
| 580 |
+
<td colspan="2"><center><strong>Finetuned</strong></center></td>
|
| 581 |
+
<td colspan="2"><center><strong>Few-shot</strong></center></td>
|
| 582 |
+
<td colspan="2"><center><strong>Finetuned</strong></center></td>
|
| 583 |
+
</tr>
|
| 584 |
+
<tr>
|
| 585 |
+
<td><strong>Model</strong></td>
|
| 586 |
+
<td><center><strong>(EM)</strong></center></td>
|
| 587 |
+
<td><center><strong>(F1)</strong></center></td>
|
| 588 |
+
<td><center><strong>(EM)</strong></center></td>
|
| 589 |
+
<td><center><strong>(F1)</strong></center></td>
|
| 590 |
+
<td><center><strong>(Spearman)</strong></center></td>
|
| 591 |
+
<td><center><strong>(Pearson)</strong></center></td>
|
| 592 |
+
<td><center><strong>(Spearman)</strong></center></td>
|
| 593 |
+
<td><center><strong>(Pearson)</strong></center></td>
|
| 594 |
+
</tr>
|
| 595 |
+
<tr>
|
| 596 |
+
<td>gemma-1.1-7b-it</td><td><center><strong>42.10</strong></center></td><td><center><strong>62.30</strong></center></td><td><center><strong>60.34</strong></center></td><td><center><strong>77.40</strong></center></td><td><center>49.10</center></td><td><center>50.23</center></td><td><center>83.43</center></td><td><center>83.64</center></td>
|
| 597 |
+
</tr>
|
| 598 |
+
<tr>
|
| 599 |
+
<td>RoGemma-7b-Instruct-2024-06-28</td><td><center>17.75</center></td><td><center>28.11</center></td><td><center>52.02</center></td><td><center>68.43</center></td><td><center>73.96</center></td><td><center><strong>75.16</strong></center></td><td><center>86.45</center></td><td><center>86.31</center></td>
|
| 600 |
+
</tr>
|
| 601 |
+
<tr>
|
| 602 |
+
<td>RoGemma-7b-Instruct-2024-10-09</td><td><center>26.03</center></td><td><center>41.58</center></td><td><center>46.72</center></td><td><center>60.79</center></td><td><center>73.23</center></td><td><center>71.58</center></td><td><center><strong>88.42</strong></center></td><td><center><strong>88.45</strong></center></td>
|
| 603 |
+
</tr>
|
| 604 |
+
<tr>
|
| 605 |
+
<td><em>RoGemma-7b-Instruct-2025-04-23</em></td><td><center><em>15.88</em></center></td><td><center><em>29.16</em></center></td><td><center><em>-</em></center></td><td><center><em>-</em></center></td><td><center><em><strong>75.90</strong></em></center></td><td><center><em><strong>75.16</strong></em></center></td><td><center><em>-</em></center></td><td><center><em>-</em></center></td>
|
| 606 |
+
</tr>
|
| 607 |
+
<tr>
|
| 608 |
+
<td>RoGemma-7b-Instruct-DPO-2024-10-09</td><td><center>19.14</center></td><td><center>38.10</center></td><td><center>-</center></td><td><center>-</center></td><td><center>69.38</center></td><td><center>69.34</center></td><td><center>-</center></td><td><center>-</center></td>
|
| 609 |
+
</tr>
|
| 610 |
+
</tbody>
|
| 611 |
+
</table>
|
| 612 |
+
|
| 613 |
+
|
| 614 |
+
## MT-Bench
|
| 615 |
+
|
| 616 |
+
<table>
|
| 617 |
+
<tbody>
|
| 618 |
+
<tr>
|
| 619 |
+
<td><strong>Model</strong></td>
|
| 620 |
+
<td><strong><center>Average</center></strong></td>
|
| 621 |
+
<td><strong><center>1st turn</center></strong></td>
|
| 622 |
+
<td><strong><center>2nd turn</center></strong></td>
|
| 623 |
+
<td><strong><center>Answers in Ro</center></strong></td>
|
| 624 |
+
</tr>
|
| 625 |
+
<tr>
|
| 626 |
+
<td>gemma-1.1-7b-it</td><td><center>4.83</center></td><td><center>5.11</center></td><td><center>4.55</center></td><td><center><strong>160/160</strong></center></td>
|
| 627 |
+
</tr>
|
| 628 |
+
<tr>
|
| 629 |
+
<td>RoGemma-7b-Instruct-2024-06-28</td><td><center>5.26</center></td><td><center>5.92</center></td><td><center>4.60</center></td><td><center><strong>160/160</strong></center></td>
|
| 630 |
+
</tr>
|
| 631 |
+
<tr>
|
| 632 |
+
<td>RoGemma-7b-Instruct-2024-10-09</td><td><center>5.24</center></td><td><center>5.55</center></td><td><center>4.94</center></td><td><center><strong>160/160</strong></center></td>
|
| 633 |
+
</tr>
|
| 634 |
+
<tr>
|
| 635 |
+
<td><em>RoGemma-7b-Instruct-2025-04-23</em></td><td><center><em><strong>6.28</strong></em></center></td><td><center><em><strong>6.97</strong></em></center></td><td><center><em><strong>5.58</strong></em></center></td><td><center><em><strong>160/160</strong></em></center></td>
|
| 636 |
+
</tr>
|
| 637 |
+
<tr>
|
| 638 |
+
<td>RoGemma-7b-Instruct-DPO-2024-10-09</td><td><center>5.47</center></td><td><center>5.92</center></td><td><center>5.03</center></td><td><center><strong>160/160</strong></center></td>
|
| 639 |
+
</tr>
|
| 640 |
+
</tbody>
|
| 641 |
+
</table>
|
| 642 |
+
|
| 643 |
+
## RoCulturaBench
|
| 644 |
+
|
| 645 |
+
<table>
|
| 646 |
+
<tbody>
|
| 647 |
+
<tr>
|
| 648 |
+
<td><strong>Model</strong></td>
|
| 649 |
+
<td><strong><center>Average</center></strong></td>
|
| 650 |
+
<td><strong><center>Answers in Ro</center></strong></td>
|
| 651 |
+
</tr>
|
| 652 |
+
<tr>
|
| 653 |
+
<td>gemma-1.1-7b-it</td><td><center>3.38</center></td><td><center><strong>100/100</strong></center></td>
|
| 654 |
+
</tr>
|
| 655 |
+
<tr>
|
| 656 |
+
<td>RoGemma-7b-Instruct-2024-06-28</td><td><center>3.26</center></td><td><center><strong>100/100</strong></center></td>
|
| 657 |
+
</tr>
|
| 658 |
+
<tr>
|
| 659 |
+
<td>RoGemma-7b-Instruct-2024-10-09</td><td><center>3.51</center></td><td><center><strong>100/100</strong></center></td>
|
| 660 |
+
</tr>
|
| 661 |
+
<tr>
|
| 662 |
+
<td><em>RoGemma-7b-Instruct-2025-04-23</em></td><td><center><em>3.65</em></center></td><td><center><em><strong>100/100</strong></em></center></td>
|
| 663 |
+
</tr>
|
| 664 |
+
<tr>
|
| 665 |
+
<td>RoGemma-7b-Instruct-DPO-2024-10-09</td><td><center><strong>3.94</strong></center></td><td><center><strong>100/100</strong></center></td>
|
| 666 |
+
</tr>
|
| 667 |
+
</tbody>
|
| 668 |
+
</table>
|
| 669 |
+
|
| 670 |
+
## RoGemma Model Family
|
| 671 |
+
|
| 672 |
+
| Model | Link |
|
| 673 |
+
|--------------------|:--------:|
|
| 674 |
+
|RoGemma-7b-Instruct-2024-06-28| [link](https://huggingface.co/OpenLLM-Ro/RoGemma-7b-Instruct-2024-06-28) |
|
| 675 |
+
|RoGemma-7b-Instruct-2024-10-09| [link](https://huggingface.co/OpenLLM-Ro/RoGemma-7b-Instruct-2024-10-09) |
|
| 676 |
+
|*RoGemma-7b-Instruct-2025-04-23*| [link](https://huggingface.co/OpenLLM-Ro/RoGemma-7b-Instruct-2025-04-23) |
|
| 677 |
+
|RoGemma-7b-Instruct-DPO-2024-10-09| [link](https://huggingface.co/OpenLLM-Ro/RoGemma-7b-Instruct-DPO-2024-10-09) |
|
| 678 |
+
|
| 679 |
+
|
| 680 |
+
## Citation
|
| 681 |
+
|
| 682 |
+
```
|
| 683 |
+
@misc{masala2024vorbecstiromanecsterecipetrain,
|
| 684 |
+
title={"Vorbe\c{s}ti Rom\^ane\c{s}te?" A Recipe to Train Powerful Romanian LLMs with English Instructions},
|
| 685 |
+
author={Mihai Masala and Denis C. Ilie-Ablachim and Alexandru Dima and Dragos Corlatescu and Miruna Zavelca and Ovio Olaru and Simina Terian-Dan and Andrei Terian-Dan and Marius Leordeanu and Horia Velicu and Marius Popescu and Mihai Dascalu and Traian Rebedea},
|
| 686 |
+
year={2024},
|
| 687 |
+
eprint={2406.18266},
|
| 688 |
+
archivePrefix={arXiv},
|
| 689 |
+
primaryClass={cs.CL},
|
| 690 |
+
url={https://arxiv.org/abs/2406.18266},
|
| 691 |
+
}
|
| 692 |
+
```
|
| 693 |
+
<!-- **APA:**
|
| 694 |
+
|
| 695 |
[More Information Needed] -->
|