File size: 15,719 Bytes
579fe1b a5fbed6 579fe1b a5fbed6 579fe1b a5fbed6 579fe1b 7fc2adf 579fe1b 7fc2adf 579fe1b 7fc2adf 579fe1b 7fc2adf 579fe1b a5fbed6 579fe1b 7fc2adf 579fe1b 7fc2adf 579fe1b 7fc2adf 579fe1b 7fc2adf 579fe1b 7fc2adf 579fe1b a5fbed6 579fe1b |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 |
---
language:
- ca
- da
- es
- fr
- gl
- is
- it
- nb
- pt
- ro
- sv
- multilingual
license: cc-by-4.0
tags:
- translation
- opus-mt-tc
model-index:
- name: opus-mt-tc-big-gmq-itc
results:
- task:
type: translation
name: Translation dan-cat
dataset:
name: flores101-devtest
type: flores_101
args: dan cat devtest
metrics:
- type: bleu
value: 33.4
name: BLEU
- type: chrf
value: 0.59224
name: chr-F
- type: bleu
value: 38.3
name: BLEU
- type: chrf
value: 0.63387
name: chr-F
- type: bleu
value: 26.4
name: BLEU
- type: chrf
value: 0.54446
name: chr-F
- type: bleu
value: 25.7
name: BLEU
- type: chrf
value: 0.55237
name: chr-F
- type: bleu
value: 36.9
name: BLEU
- type: chrf
value: 0.62233
name: chr-F
- type: bleu
value: 31.8
name: BLEU
- type: chrf
value: 0.58235
name: chr-F
- type: bleu
value: 24.3
name: BLEU
- type: chrf
value: 0.52453
name: chr-F
- type: bleu
value: 22.7
name: BLEU
- type: chrf
value: 0.4893
name: chr-F
- type: bleu
value: 26.2
name: BLEU
- type: chrf
value: 0.52704
name: chr-F
- type: bleu
value: 18.0
name: BLEU
- type: chrf
value: 0.45387
name: chr-F
- type: bleu
value: 18.6
name: BLEU
- type: chrf
value: 0.47303
name: chr-F
- type: bleu
value: 24.9
name: BLEU
- type: chrf
value: 0.51381
name: chr-F
- type: bleu
value: 21.6
name: BLEU
- type: chrf
value: 0.48224
name: chr-F
- type: bleu
value: 18.1
name: BLEU
- type: chrf
value: 0.45786
name: chr-F
- type: bleu
value: 28.9
name: BLEU
- type: chrf
value: 0.55984
name: chr-F
- type: bleu
value: 33.8
name: BLEU
- type: chrf
value: 0.60102
name: chr-F
- type: bleu
value: 23.4
name: BLEU
- type: chrf
value: 0.52145
name: chr-F
- type: bleu
value: 22.2
name: BLEU
- type: chrf
value: 0.52619
name: chr-F
- type: bleu
value: 32.2
name: BLEU
- type: chrf
value: 0.58836
name: chr-F
- type: bleu
value: 27.6
name: BLEU
- type: chrf
value: 0.54845
name: chr-F
- type: bleu
value: 21.8
name: BLEU
- type: chrf
value: 0.50661
name: chr-F
- type: bleu
value: 32.4
name: BLEU
- type: chrf
value: 0.58542
name: chr-F
- type: bleu
value: 39.3
name: BLEU
- type: chrf
value: 0.63688
name: chr-F
- type: bleu
value: 26.0
name: BLEU
- type: chrf
value: 0.53989
name: chr-F
- type: bleu
value: 25.9
name: BLEU
- type: chrf
value: 0.55232
name: chr-F
- type: bleu
value: 36.5
name: BLEU
- type: chrf
value: 0.61882
name: chr-F
- type: bleu
value: 31.0
name: BLEU
- type: chrf
value: 0.57419
name: chr-F
- type: bleu
value: 23.8
name: BLEU
- type: chrf
value: 0.52175
name: chr-F
- task:
type: translation
name: Translation dan-fra
dataset:
name: tatoeba-test-v2021-08-07
type: tatoeba_mt
args: dan-fra
metrics:
- type: bleu
value: 63.8
name: BLEU
- type: chrf
value: 0.76671
name: chr-F
- type: bleu
value: 56.2
name: BLEU
- type: chrf
value: 0.74658
name: chr-F
- type: bleu
value: 57.8
name: BLEU
- type: chrf
value: 0.74944
name: chr-F
- type: bleu
value: 54.8
name: BLEU
- type: chrf
value: 0.72328
name: chr-F
- type: bleu
value: 51.0
name: BLEU
- type: chrf
value: 0.69354
name: chr-F
- type: bleu
value: 49.2
name: BLEU
- type: chrf
value: 0.66008
name: chr-F
- type: bleu
value: 54.4
name: BLEU
- type: chrf
value: 0.70854
name: chr-F
- type: bleu
value: 55.9
name: BLEU
- type: chrf
value: 0.73672
name: chr-F
- type: bleu
value: 59.2
name: BLEU
- type: chrf
value: 0.73014
name: chr-F
- type: bleu
value: 56.6
name: BLEU
- type: chrf
value: 0.73211
name: chr-F
- type: bleu
value: 48.7
name: BLEU
- type: chrf
value: 0.68146
name: chr-F
- type: bleu
value: 55.3
name: BLEU
- type: chrf
value: 0.71373
name: chr-F
---
# opus-mt-tc-big-gmq-itc
## Table of Contents
- [Model Details](#model-details)
- [Uses](#uses)
- [Risks, Limitations and Biases](#risks-limitations-and-biases)
- [How to Get Started With the Model](#how-to-get-started-with-the-model)
- [Training](#training)
- [Evaluation](#evaluation)
- [Citation Information](#citation-information)
- [Acknowledgements](#acknowledgements)
## Model Details
Neural machine translation model for translating from North Germanic languages (gmq) to Italic languages (itc).
This model is part of the [OPUS-MT project](https://github.com/Helsinki-NLP/Opus-MT), an effort to make neural machine translation models widely available and accessible for many languages in the world. All models are originally trained using the amazing framework of [Marian NMT](https://marian-nmt.github.io/), an efficient NMT implementation written in pure C++. The models have been converted to pyTorch using the transformers library by huggingface. Training data is taken from [OPUS](https://opus.nlpl.eu/) and training pipelines use the procedures of [OPUS-MT-train](https://github.com/Helsinki-NLP/Opus-MT-train).
**Model Description:**
- **Developed by:** Language Technology Research Group at the University of Helsinki
- **Model Type:** Translation (transformer-big)
- **Release**: 2022-08-09
- **License:** CC-BY-4.0
- **Language(s):**
- Source Language(s): dan isl nno nob nor swe
- Target Language(s): cat fra glg ita lat por ron spa
- Language Pair(s): dan-cat dan-fra dan-glg dan-ita dan-por dan-ron dan-spa isl-cat isl-fra isl-ita isl-por isl-ron isl-spa nob-cat nob-fra nob-glg nob-ita nob-por nob-ron nob-spa swe-cat swe-fra swe-glg swe-ita swe-por swe-ron swe-spa
- Valid Target Language Labels: >>acf<< >>aoa<< >>arg<< >>ast<< >>cat<< >>cbk<< >>ccd<< >>cks<< >>cos<< >>cri<< >>crs<< >>dlm<< >>drc<< >>egl<< >>ext<< >>fab<< >>fax<< >>fra<< >>frc<< >>frm<< >>fro<< >>frp<< >>fur<< >>gcf<< >>gcr<< >>glg<< >>hat<< >>idb<< >>ist<< >>ita<< >>itk<< >>kea<< >>kmv<< >>lad<< >>lad_Latn<< >>lat<< >>lat_Latn<< >>lij<< >>lld<< >>lmo<< >>lou<< >>mcm<< >>mfe<< >>mol<< >>mwl<< >>mxi<< >>mzs<< >>nap<< >>nrf<< >>oci<< >>osc<< >>osp<< >>osp_Latn<< >>pap<< >>pcd<< >>pln<< >>pms<< >>pob<< >>por<< >>pov<< >>pre<< >>pro<< >>qbb<< >>qhr<< >>rcf<< >>rgn<< >>roh<< >>ron<< >>ruo<< >>rup<< >>ruq<< >>scf<< >>scn<< >>sdc<< >>sdn<< >>spa<< >>spq<< >>spx<< >>src<< >>srd<< >>sro<< >>tmg<< >>tvy<< >>vec<< >>vkp<< >>wln<< >>xfa<< >>xum<<
- **Original Model**: [opusTCv20210807_transformer-big_2022-08-09.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/gmq-itc/opusTCv20210807_transformer-big_2022-08-09.zip)
- **Resources for more information:**
- [OPUS-MT-train GitHub Repo](https://github.com/Helsinki-NLP/OPUS-MT-train)
- More information about released models for this language pair: [OPUS-MT gmq-itc README](https://github.com/Helsinki-NLP/Tatoeba-Challenge/tree/master/models/gmq-itc/README.md)
- [More information about MarianNMT models in the transformers library](https://huggingface.co/docs/transformers/model_doc/marian)
- [Tatoeba Translation Challenge](https://github.com/Helsinki-NLP/Tatoeba-Challenge/
This is a multilingual translation model with multiple target languages. A sentence initial language token is required in the form of `>>id<<` (id = valid target language ID), e.g. `>>fra<<`
## Uses
This model can be used for translation and text-to-text generation.
## Risks, Limitations and Biases
**CONTENT WARNING: Readers should be aware that the model is trained on various public data sets that may contain content that is disturbing, offensive, and can propagate historical and current stereotypes.**
Significant research has explored bias and fairness issues with language models (see, e.g., [Sheng et al. (2021)](https://aclanthology.org/2021.acl-long.330.pdf) and [Bender et al. (2021)](https://dl.acm.org/doi/pdf/10.1145/3442188.3445922)).
## How to Get Started With the Model
A short example code:
```python
from transformers import MarianMTModel, MarianTokenizer
src_text = [
">>spa<< Jag �r inte religi�s.",
">>por<< Livet er for kort til � l�re seg tysk."
]
model_name = "pytorch-models/opus-mt-tc-big-gmq-itc"
tokenizer = MarianTokenizer.from_pretrained(model_name)
model = MarianMTModel.from_pretrained(model_name)
translated = model.generate(**tokenizer(src_text, return_tensors="pt", padding=True))
for t in translated:
print( tokenizer.decode(t, skip_special_tokens=True) )
# expected output:
# No soy religioso.
# A vida � muito curta para aprender alem�o.
```
You can also use OPUS-MT models with the transformers pipelines, for example:
```python
from transformers import pipeline
pipe = pipeline("translation", model="Helsinki-NLP/opus-mt-tc-big-gmq-itc")
print(pipe(">>spa<< Jag �r inte religi�s."))
# expected output: No soy religioso.
```
## Training
- **Data**: opusTCv20210807 ([source](https://github.com/Helsinki-NLP/Tatoeba-Challenge))
- **Pre-processing**: SentencePiece (spm32k,spm32k)
- **Model Type:** transformer-big
- **Original MarianNMT Model**: [opusTCv20210807_transformer-big_2022-08-09.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/gmq-itc/opusTCv20210807_transformer-big_2022-08-09.zip)
- **Training Scripts**: [GitHub Repo](https://github.com/Helsinki-NLP/OPUS-MT-train)
## Evaluation
* test set translations: [opusTCv20210807_transformer-big_2022-08-09.test.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/gmq-itc/opusTCv20210807_transformer-big_2022-08-09.test.txt)
* test set scores: [opusTCv20210807_transformer-big_2022-08-09.eval.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/gmq-itc/opusTCv20210807_transformer-big_2022-08-09.eval.txt)
* benchmark results: [benchmark_results.txt](benchmark_results.txt)
* benchmark output: [benchmark_translations.zip](benchmark_translations.zip)
| langpair | testset | chr-F | BLEU | #sent | #words |
|----------|---------|-------|-------|-------|--------|
| dan-fra | tatoeba-test-v2021-08-07 | 0.76671 | 63.8 | 1731 | 11882 |
| dan-ita | tatoeba-test-v2021-08-07 | 0.74658 | 56.2 | 284 | 2226 |
| dan-por | tatoeba-test-v2021-08-07 | 0.74944 | 57.8 | 873 | 5360 |
| dan-spa | tatoeba-test-v2021-08-07 | 0.72328 | 54.8 | 5000 | 35528 |
| isl-ita | tatoeba-test-v2021-08-07 | 0.69354 | 51.0 | 236 | 1450 |
| isl-spa | tatoeba-test-v2021-08-07 | 0.66008 | 49.2 | 238 | 1229 |
| nob-fra | tatoeba-test-v2021-08-07 | 0.70854 | 54.4 | 323 | 2269 |
| nob-spa | tatoeba-test-v2021-08-07 | 0.73672 | 55.9 | 885 | 6866 |
| swe-fra | tatoeba-test-v2021-08-07 | 0.73014 | 59.2 | 1407 | 9580 |
| swe-ita | tatoeba-test-v2021-08-07 | 0.73211 | 56.6 | 715 | 4711 |
| swe-por | tatoeba-test-v2021-08-07 | 0.68146 | 48.7 | 320 | 2032 |
| swe-spa | tatoeba-test-v2021-08-07 | 0.71373 | 55.3 | 1351 | 8235 |
| dan-cat | flores101-devtest | 0.59224 | 33.4 | 1012 | 27304 |
| dan-fra | flores101-devtest | 0.63387 | 38.3 | 1012 | 28343 |
| dan-glg | flores101-devtest | 0.54446 | 26.4 | 1012 | 26582 |
| dan-ita | flores101-devtest | 0.55237 | 25.7 | 1012 | 27306 |
| dan-por | flores101-devtest | 0.62233 | 36.9 | 1012 | 26519 |
| dan-ron | flores101-devtest | 0.58235 | 31.8 | 1012 | 26799 |
| dan-spa | flores101-devtest | 0.52453 | 24.3 | 1012 | 29199 |
| isl-cat | flores101-devtest | 0.48930 | 22.7 | 1012 | 27304 |
| isl-fra | flores101-devtest | 0.52704 | 26.2 | 1012 | 28343 |
| isl-glg | flores101-devtest | 0.45387 | 18.0 | 1012 | 26582 |
| isl-ita | flores101-devtest | 0.47303 | 18.6 | 1012 | 27306 |
| isl-por | flores101-devtest | 0.51381 | 24.9 | 1012 | 26519 |
| isl-ron | flores101-devtest | 0.48224 | 21.6 | 1012 | 26799 |
| isl-spa | flores101-devtest | 0.45786 | 18.1 | 1012 | 29199 |
| nob-cat | flores101-devtest | 0.55984 | 28.9 | 1012 | 27304 |
| nob-fra | flores101-devtest | 0.60102 | 33.8 | 1012 | 28343 |
| nob-glg | flores101-devtest | 0.52145 | 23.4 | 1012 | 26582 |
| nob-ita | flores101-devtest | 0.52619 | 22.2 | 1012 | 27306 |
| nob-por | flores101-devtest | 0.58836 | 32.2 | 1012 | 26519 |
| nob-ron | flores101-devtest | 0.54845 | 27.6 | 1012 | 26799 |
| nob-spa | flores101-devtest | 0.50661 | 21.8 | 1012 | 29199 |
| swe-cat | flores101-devtest | 0.58542 | 32.4 | 1012 | 27304 |
| swe-fra | flores101-devtest | 0.63688 | 39.3 | 1012 | 28343 |
| swe-glg | flores101-devtest | 0.53989 | 26.0 | 1012 | 26582 |
| swe-ita | flores101-devtest | 0.55232 | 25.9 | 1012 | 27306 |
| swe-por | flores101-devtest | 0.61882 | 36.5 | 1012 | 26519 |
| swe-ron | flores101-devtest | 0.57419 | 31.0 | 1012 | 26799 |
| swe-spa | flores101-devtest | 0.52175 | 23.8 | 1012 | 29199 |
## Citation Information
* Publications: [OPUS-MT � Building open translation services for the World](https://aclanthology.org/2020.eamt-1.61/) and [The Tatoeba Translation Challenge � Realistic Data Sets for Low Resource and Multilingual MT](https://aclanthology.org/2020.wmt-1.139/) (Please, cite if you use this model.)
```
@inproceedings{tiedemann-thottingal-2020-opus,
title = "{OPUS}-{MT} {--} Building open translation services for the World",
author = {Tiedemann, J{\"o}rg and Thottingal, Santhosh},
booktitle = "Proceedings of the 22nd Annual Conference of the European Association for Machine Translation",
month = nov,
year = "2020",
address = "Lisboa, Portugal",
publisher = "European Association for Machine Translation",
url = "https://aclanthology.org/2020.eamt-1.61",
pages = "479--480",
}
@inproceedings{tiedemann-2020-tatoeba,
title = "The Tatoeba Translation Challenge {--} Realistic Data Sets for Low Resource and Multilingual {MT}",
author = {Tiedemann, J{\"o}rg},
booktitle = "Proceedings of the Fifth Conference on Machine Translation",
month = nov,
year = "2020",
address = "Online",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2020.wmt-1.139",
pages = "1174--1182",
}
```
## Acknowledgements
The work is supported by the [European Language Grid](https://www.european-language-grid.eu/) as [pilot project 2866](https://live.european-language-grid.eu/catalogue/#/resource/projects/2866), by the [FoTran project](https://www.helsinki.fi/en/researchgroups/natural-language-understanding-with-cross-lingual-grounding), funded by the European Research Council (ERC) under the European Union�s Horizon 2020 research and innovation programme (grant agreement No 771113), and the [MeMAD project](https://memad.eu/), funded by the European Union�s Horizon 2020 Research and Innovation Programme under grant agreement No 780069. We are also grateful for the generous computational resources and IT infrastructure provided by [CSC -- IT Center for Science](https://www.csc.fi/), Finland.
## Model conversion info
* transformers version: 4.16.2
* OPUS-MT git hash: 8b9f0b0
* port time: Sat Aug 13 00:00:00 EEST 2022
* port machine: LM0-400-22516.local
|