tiedeman commited on
Commit
9d43e28
1 Parent(s): e2105eb

Initial commit

Browse files
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ *.spm filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,1685 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: transformers
3
+ language:
4
+ - da
5
+ - de
6
+ - en
7
+ - es
8
+ - fo
9
+ - fr
10
+ - is
11
+ - nb
12
+ - nn
13
+ - no
14
+ - non
15
+ - pt
16
+ - sv
17
+
18
+ tags:
19
+ - translation
20
+ - opus-mt-tc-bible
21
+
22
+ license: apache-2.0
23
+ model-index:
24
+ - name: opus-mt-tc-bible-big-deu_eng_fra_por_spa-gmq
25
+ results:
26
+ - task:
27
+ name: Translation deu-dan
28
+ type: translation
29
+ args: deu-dan
30
+ dataset:
31
+ name: flores200-devtest
32
+ type: flores200-devtest
33
+ args: deu-dan
34
+ metrics:
35
+ - name: BLEU
36
+ type: bleu
37
+ value: 35.1
38
+ - name: chr-F
39
+ type: chrf
40
+ value: 0.62152
41
+ - task:
42
+ name: Translation deu-fao
43
+ type: translation
44
+ args: deu-fao
45
+ dataset:
46
+ name: flores200-devtest
47
+ type: flores200-devtest
48
+ args: deu-fao
49
+ metrics:
50
+ - name: BLEU
51
+ type: bleu
52
+ value: 11.5
53
+ - name: chr-F
54
+ type: chrf
55
+ value: 0.33611
56
+ - task:
57
+ name: Translation deu-isl
58
+ type: translation
59
+ args: deu-isl
60
+ dataset:
61
+ name: flores200-devtest
62
+ type: flores200-devtest
63
+ args: deu-isl
64
+ metrics:
65
+ - name: BLEU
66
+ type: bleu
67
+ value: 19.1
68
+ - name: chr-F
69
+ type: chrf
70
+ value: 0.48648
71
+ - task:
72
+ name: Translation deu-nno
73
+ type: translation
74
+ args: deu-nno
75
+ dataset:
76
+ name: flores200-devtest
77
+ type: flores200-devtest
78
+ args: deu-nno
79
+ metrics:
80
+ - name: BLEU
81
+ type: bleu
82
+ value: 24.0
83
+ - name: chr-F
84
+ type: chrf
85
+ value: 0.53530
86
+ - task:
87
+ name: Translation deu-nob
88
+ type: translation
89
+ args: deu-nob
90
+ dataset:
91
+ name: flores200-devtest
92
+ type: flores200-devtest
93
+ args: deu-nob
94
+ metrics:
95
+ - name: BLEU
96
+ type: bleu
97
+ value: 25.1
98
+ - name: chr-F
99
+ type: chrf
100
+ value: 0.55748
101
+ - task:
102
+ name: Translation deu-swe
103
+ type: translation
104
+ args: deu-swe
105
+ dataset:
106
+ name: flores200-devtest
107
+ type: flores200-devtest
108
+ args: deu-swe
109
+ metrics:
110
+ - name: BLEU
111
+ type: bleu
112
+ value: 34.2
113
+ - name: chr-F
114
+ type: chrf
115
+ value: 0.62138
116
+ - task:
117
+ name: Translation eng-dan
118
+ type: translation
119
+ args: eng-dan
120
+ dataset:
121
+ name: flores200-devtest
122
+ type: flores200-devtest
123
+ args: eng-dan
124
+ metrics:
125
+ - name: BLEU
126
+ type: bleu
127
+ value: 47.0
128
+ - name: chr-F
129
+ type: chrf
130
+ value: 0.70321
131
+ - task:
132
+ name: Translation eng-fao
133
+ type: translation
134
+ args: eng-fao
135
+ dataset:
136
+ name: flores200-devtest
137
+ type: flores200-devtest
138
+ args: eng-fao
139
+ metrics:
140
+ - name: BLEU
141
+ type: bleu
142
+ value: 14.1
143
+ - name: chr-F
144
+ type: chrf
145
+ value: 0.35857
146
+ - task:
147
+ name: Translation eng-isl
148
+ type: translation
149
+ args: eng-isl
150
+ dataset:
151
+ name: flores200-devtest
152
+ type: flores200-devtest
153
+ args: eng-isl
154
+ metrics:
155
+ - name: BLEU
156
+ type: bleu
157
+ value: 24.4
158
+ - name: chr-F
159
+ type: chrf
160
+ value: 0.52585
161
+ - task:
162
+ name: Translation eng-nno
163
+ type: translation
164
+ args: eng-nno
165
+ dataset:
166
+ name: flores200-devtest
167
+ type: flores200-devtest
168
+ args: eng-nno
169
+ metrics:
170
+ - name: BLEU
171
+ type: bleu
172
+ value: 33.8
173
+ - name: chr-F
174
+ type: chrf
175
+ value: 0.61372
176
+ - task:
177
+ name: Translation eng-nob
178
+ type: translation
179
+ args: eng-nob
180
+ dataset:
181
+ name: flores200-devtest
182
+ type: flores200-devtest
183
+ args: eng-nob
184
+ metrics:
185
+ - name: BLEU
186
+ type: bleu
187
+ value: 34.4
188
+ - name: chr-F
189
+ type: chrf
190
+ value: 0.62508
191
+ - task:
192
+ name: Translation eng-swe
193
+ type: translation
194
+ args: eng-swe
195
+ dataset:
196
+ name: flores200-devtest
197
+ type: flores200-devtest
198
+ args: eng-swe
199
+ metrics:
200
+ - name: BLEU
201
+ type: bleu
202
+ value: 46.0
203
+ - name: chr-F
204
+ type: chrf
205
+ value: 0.69703
206
+ - task:
207
+ name: Translation fra-dan
208
+ type: translation
209
+ args: fra-dan
210
+ dataset:
211
+ name: flores200-devtest
212
+ type: flores200-devtest
213
+ args: fra-dan
214
+ metrics:
215
+ - name: BLEU
216
+ type: bleu
217
+ value: 34.1
218
+ - name: chr-F
219
+ type: chrf
220
+ value: 0.61025
221
+ - task:
222
+ name: Translation fra-isl
223
+ type: translation
224
+ args: fra-isl
225
+ dataset:
226
+ name: flores200-devtest
227
+ type: flores200-devtest
228
+ args: fra-isl
229
+ metrics:
230
+ - name: BLEU
231
+ type: bleu
232
+ value: 18.8
233
+ - name: chr-F
234
+ type: chrf
235
+ value: 0.48273
236
+ - task:
237
+ name: Translation fra-nno
238
+ type: translation
239
+ args: fra-nno
240
+ dataset:
241
+ name: flores200-devtest
242
+ type: flores200-devtest
243
+ args: fra-nno
244
+ metrics:
245
+ - name: BLEU
246
+ type: bleu
247
+ value: 24.3
248
+ - name: chr-F
249
+ type: chrf
250
+ value: 0.53032
251
+ - task:
252
+ name: Translation fra-nob
253
+ type: translation
254
+ args: fra-nob
255
+ dataset:
256
+ name: flores200-devtest
257
+ type: flores200-devtest
258
+ args: fra-nob
259
+ metrics:
260
+ - name: BLEU
261
+ type: bleu
262
+ value: 25.0
263
+ - name: chr-F
264
+ type: chrf
265
+ value: 0.54933
266
+ - task:
267
+ name: Translation fra-swe
268
+ type: translation
269
+ args: fra-swe
270
+ dataset:
271
+ name: flores200-devtest
272
+ type: flores200-devtest
273
+ args: fra-swe
274
+ metrics:
275
+ - name: BLEU
276
+ type: bleu
277
+ value: 32.8
278
+ - name: chr-F
279
+ type: chrf
280
+ value: 0.60612
281
+ - task:
282
+ name: Translation por-dan
283
+ type: translation
284
+ args: por-dan
285
+ dataset:
286
+ name: flores200-devtest
287
+ type: flores200-devtest
288
+ args: por-dan
289
+ metrics:
290
+ - name: BLEU
291
+ type: bleu
292
+ value: 36.2
293
+ - name: chr-F
294
+ type: chrf
295
+ value: 0.62221
296
+ - task:
297
+ name: Translation por-fao
298
+ type: translation
299
+ args: por-fao
300
+ dataset:
301
+ name: flores200-devtest
302
+ type: flores200-devtest
303
+ args: por-fao
304
+ metrics:
305
+ - name: BLEU
306
+ type: bleu
307
+ value: 11.5
308
+ - name: chr-F
309
+ type: chrf
310
+ value: 0.33159
311
+ - task:
312
+ name: Translation por-isl
313
+ type: translation
314
+ args: por-isl
315
+ dataset:
316
+ name: flores200-devtest
317
+ type: flores200-devtest
318
+ args: por-isl
319
+ metrics:
320
+ - name: BLEU
321
+ type: bleu
322
+ value: 19.6
323
+ - name: chr-F
324
+ type: chrf
325
+ value: 0.48357
326
+ - task:
327
+ name: Translation por-nno
328
+ type: translation
329
+ args: por-nno
330
+ dataset:
331
+ name: flores200-devtest
332
+ type: flores200-devtest
333
+ args: por-nno
334
+ metrics:
335
+ - name: BLEU
336
+ type: bleu
337
+ value: 26.3
338
+ - name: chr-F
339
+ type: chrf
340
+ value: 0.54369
341
+ - task:
342
+ name: Translation por-nob
343
+ type: translation
344
+ args: por-nob
345
+ dataset:
346
+ name: flores200-devtest
347
+ type: flores200-devtest
348
+ args: por-nob
349
+ metrics:
350
+ - name: BLEU
351
+ type: bleu
352
+ value: 26.4
353
+ - name: chr-F
354
+ type: chrf
355
+ value: 0.56054
356
+ - task:
357
+ name: Translation por-swe
358
+ type: translation
359
+ args: por-swe
360
+ dataset:
361
+ name: flores200-devtest
362
+ type: flores200-devtest
363
+ args: por-swe
364
+ metrics:
365
+ - name: BLEU
366
+ type: bleu
367
+ value: 34.1
368
+ - name: chr-F
369
+ type: chrf
370
+ value: 0.61388
371
+ - task:
372
+ name: Translation spa-dan
373
+ type: translation
374
+ args: spa-dan
375
+ dataset:
376
+ name: flores200-devtest
377
+ type: flores200-devtest
378
+ args: spa-dan
379
+ metrics:
380
+ - name: BLEU
381
+ type: bleu
382
+ value: 24.7
383
+ - name: chr-F
384
+ type: chrf
385
+ value: 0.55091
386
+ - task:
387
+ name: Translation spa-isl
388
+ type: translation
389
+ args: spa-isl
390
+ dataset:
391
+ name: flores200-devtest
392
+ type: flores200-devtest
393
+ args: spa-isl
394
+ metrics:
395
+ - name: BLEU
396
+ type: bleu
397
+ value: 14.2
398
+ - name: chr-F
399
+ type: chrf
400
+ value: 0.44469
401
+ - task:
402
+ name: Translation spa-nno
403
+ type: translation
404
+ args: spa-nno
405
+ dataset:
406
+ name: flores200-devtest
407
+ type: flores200-devtest
408
+ args: spa-nno
409
+ metrics:
410
+ - name: BLEU
411
+ type: bleu
412
+ value: 18.6
413
+ - name: chr-F
414
+ type: chrf
415
+ value: 0.48898
416
+ - task:
417
+ name: Translation spa-nob
418
+ type: translation
419
+ args: spa-nob
420
+ dataset:
421
+ name: flores200-devtest
422
+ type: flores200-devtest
423
+ args: spa-nob
424
+ metrics:
425
+ - name: BLEU
426
+ type: bleu
427
+ value: 18.8
428
+ - name: chr-F
429
+ type: chrf
430
+ value: 0.50901
431
+ - task:
432
+ name: Translation spa-swe
433
+ type: translation
434
+ args: spa-swe
435
+ dataset:
436
+ name: flores200-devtest
437
+ type: flores200-devtest
438
+ args: spa-swe
439
+ metrics:
440
+ - name: BLEU
441
+ type: bleu
442
+ value: 22.7
443
+ - name: chr-F
444
+ type: chrf
445
+ value: 0.54182
446
+ - task:
447
+ name: Translation deu-dan
448
+ type: translation
449
+ args: deu-dan
450
+ dataset:
451
+ name: flores101-devtest
452
+ type: flores_101
453
+ args: deu dan devtest
454
+ metrics:
455
+ - name: BLEU
456
+ type: bleu
457
+ value: 34.8
458
+ - name: chr-F
459
+ type: chrf
460
+ value: 0.62006
461
+ - task:
462
+ name: Translation deu-isl
463
+ type: translation
464
+ args: deu-isl
465
+ dataset:
466
+ name: flores101-devtest
467
+ type: flores_101
468
+ args: deu isl devtest
469
+ metrics:
470
+ - name: BLEU
471
+ type: bleu
472
+ value: 18.8
473
+ - name: chr-F
474
+ type: chrf
475
+ value: 0.48236
476
+ - task:
477
+ name: Translation deu-swe
478
+ type: translation
479
+ args: deu-swe
480
+ dataset:
481
+ name: flores101-devtest
482
+ type: flores_101
483
+ args: deu swe devtest
484
+ metrics:
485
+ - name: BLEU
486
+ type: bleu
487
+ value: 33.7
488
+ - name: chr-F
489
+ type: chrf
490
+ value: 0.61778
491
+ - task:
492
+ name: Translation eng-swe
493
+ type: translation
494
+ args: eng-swe
495
+ dataset:
496
+ name: flores101-devtest
497
+ type: flores_101
498
+ args: eng swe devtest
499
+ metrics:
500
+ - name: BLEU
501
+ type: bleu
502
+ value: 45.5
503
+ - name: chr-F
504
+ type: chrf
505
+ value: 0.69435
506
+ - task:
507
+ name: Translation fra-dan
508
+ type: translation
509
+ args: fra-dan
510
+ dataset:
511
+ name: flores101-devtest
512
+ type: flores_101
513
+ args: fra dan devtest
514
+ metrics:
515
+ - name: BLEU
516
+ type: bleu
517
+ value: 34.0
518
+ - name: chr-F
519
+ type: chrf
520
+ value: 0.61019
521
+ - task:
522
+ name: Translation fra-isl
523
+ type: translation
524
+ args: fra-isl
525
+ dataset:
526
+ name: flores101-devtest
527
+ type: flores_101
528
+ args: fra isl devtest
529
+ metrics:
530
+ - name: BLEU
531
+ type: bleu
532
+ value: 18.1
533
+ - name: chr-F
534
+ type: chrf
535
+ value: 0.47647
536
+ - task:
537
+ name: Translation fra-swe
538
+ type: translation
539
+ args: fra-swe
540
+ dataset:
541
+ name: flores101-devtest
542
+ type: flores_101
543
+ args: fra swe devtest
544
+ metrics:
545
+ - name: BLEU
546
+ type: bleu
547
+ value: 32.2
548
+ - name: chr-F
549
+ type: chrf
550
+ value: 0.60354
551
+ - task:
552
+ name: Translation por-isl
553
+ type: translation
554
+ args: por-isl
555
+ dataset:
556
+ name: flores101-devtest
557
+ type: flores_101
558
+ args: por isl devtest
559
+ metrics:
560
+ - name: BLEU
561
+ type: bleu
562
+ value: 19.1
563
+ - name: chr-F
564
+ type: chrf
565
+ value: 0.47937
566
+ - task:
567
+ name: Translation por-swe
568
+ type: translation
569
+ args: por-swe
570
+ dataset:
571
+ name: flores101-devtest
572
+ type: flores_101
573
+ args: por swe devtest
574
+ metrics:
575
+ - name: BLEU
576
+ type: bleu
577
+ value: 33.1
578
+ - name: chr-F
579
+ type: chrf
580
+ value: 0.60857
581
+ - task:
582
+ name: Translation spa-dan
583
+ type: translation
584
+ args: spa-dan
585
+ dataset:
586
+ name: flores101-devtest
587
+ type: flores_101
588
+ args: spa dan devtest
589
+ metrics:
590
+ - name: BLEU
591
+ type: bleu
592
+ value: 24.4
593
+ - name: chr-F
594
+ type: chrf
595
+ value: 0.54890
596
+ - task:
597
+ name: Translation spa-nob
598
+ type: translation
599
+ args: spa-nob
600
+ dataset:
601
+ name: flores101-devtest
602
+ type: flores_101
603
+ args: spa nob devtest
604
+ metrics:
605
+ - name: BLEU
606
+ type: bleu
607
+ value: 18.3
608
+ - name: chr-F
609
+ type: chrf
610
+ value: 0.50610
611
+ - task:
612
+ name: Translation spa-swe
613
+ type: translation
614
+ args: spa-swe
615
+ dataset:
616
+ name: flores101-devtest
617
+ type: flores_101
618
+ args: spa swe devtest
619
+ metrics:
620
+ - name: BLEU
621
+ type: bleu
622
+ value: 22.4
623
+ - name: chr-F
624
+ type: chrf
625
+ value: 0.54011
626
+ - task:
627
+ name: Translation deu-dan
628
+ type: translation
629
+ args: deu-dan
630
+ dataset:
631
+ name: ntrex128
632
+ type: ntrex128
633
+ args: deu-dan
634
+ metrics:
635
+ - name: BLEU
636
+ type: bleu
637
+ value: 29.1
638
+ - name: chr-F
639
+ type: chrf
640
+ value: 0.56412
641
+ - task:
642
+ name: Translation deu-fao
643
+ type: translation
644
+ args: deu-fao
645
+ dataset:
646
+ name: ntrex128
647
+ type: ntrex128
648
+ args: deu-fao
649
+ metrics:
650
+ - name: BLEU
651
+ type: bleu
652
+ value: 12.5
653
+ - name: chr-F
654
+ type: chrf
655
+ value: 0.35495
656
+ - task:
657
+ name: Translation deu-isl
658
+ type: translation
659
+ args: deu-isl
660
+ dataset:
661
+ name: ntrex128
662
+ type: ntrex128
663
+ args: deu-isl
664
+ metrics:
665
+ - name: BLEU
666
+ type: bleu
667
+ value: 18.8
668
+ - name: chr-F
669
+ type: chrf
670
+ value: 0.48309
671
+ - task:
672
+ name: Translation deu-nno
673
+ type: translation
674
+ args: deu-nno
675
+ dataset:
676
+ name: ntrex128
677
+ type: ntrex128
678
+ args: deu-nno
679
+ metrics:
680
+ - name: BLEU
681
+ type: bleu
682
+ value: 22.0
683
+ - name: chr-F
684
+ type: chrf
685
+ value: 0.51535
686
+ - task:
687
+ name: Translation deu-nob
688
+ type: translation
689
+ args: deu-nob
690
+ dataset:
691
+ name: ntrex128
692
+ type: ntrex128
693
+ args: deu-nob
694
+ metrics:
695
+ - name: BLEU
696
+ type: bleu
697
+ value: 27.6
698
+ - name: chr-F
699
+ type: chrf
700
+ value: 0.56152
701
+ - task:
702
+ name: Translation deu-swe
703
+ type: translation
704
+ args: deu-swe
705
+ dataset:
706
+ name: ntrex128
707
+ type: ntrex128
708
+ args: deu-swe
709
+ metrics:
710
+ - name: BLEU
711
+ type: bleu
712
+ value: 29.6
713
+ - name: chr-F
714
+ type: chrf
715
+ value: 0.58061
716
+ - task:
717
+ name: Translation eng-dan
718
+ type: translation
719
+ args: eng-dan
720
+ dataset:
721
+ name: ntrex128
722
+ type: ntrex128
723
+ args: eng-dan
724
+ metrics:
725
+ - name: BLEU
726
+ type: bleu
727
+ value: 37.6
728
+ - name: chr-F
729
+ type: chrf
730
+ value: 0.61894
731
+ - task:
732
+ name: Translation eng-fao
733
+ type: translation
734
+ args: eng-fao
735
+ dataset:
736
+ name: ntrex128
737
+ type: ntrex128
738
+ args: eng-fao
739
+ metrics:
740
+ - name: BLEU
741
+ type: bleu
742
+ value: 15.9
743
+ - name: chr-F
744
+ type: chrf
745
+ value: 0.38410
746
+ - task:
747
+ name: Translation eng-isl
748
+ type: translation
749
+ args: eng-isl
750
+ dataset:
751
+ name: ntrex128
752
+ type: ntrex128
753
+ args: eng-isl
754
+ metrics:
755
+ - name: BLEU
756
+ type: bleu
757
+ value: 23.9
758
+ - name: chr-F
759
+ type: chrf
760
+ value: 0.52027
761
+ - task:
762
+ name: Translation eng-nno
763
+ type: translation
764
+ args: eng-nno
765
+ dataset:
766
+ name: ntrex128
767
+ type: ntrex128
768
+ args: eng-nno
769
+ metrics:
770
+ - name: BLEU
771
+ type: bleu
772
+ value: 34.0
773
+ - name: chr-F
774
+ type: chrf
775
+ value: 0.60754
776
+ - task:
777
+ name: Translation eng-nob
778
+ type: translation
779
+ args: eng-nob
780
+ dataset:
781
+ name: ntrex128
782
+ type: ntrex128
783
+ args: eng-nob
784
+ metrics:
785
+ - name: BLEU
786
+ type: bleu
787
+ value: 36.9
788
+ - name: chr-F
789
+ type: chrf
790
+ value: 0.62327
791
+ - task:
792
+ name: Translation eng-swe
793
+ type: translation
794
+ args: eng-swe
795
+ dataset:
796
+ name: ntrex128
797
+ type: ntrex128
798
+ args: eng-swe
799
+ metrics:
800
+ - name: BLEU
801
+ type: bleu
802
+ value: 41.3
803
+ - name: chr-F
804
+ type: chrf
805
+ value: 0.66129
806
+ - task:
807
+ name: Translation fra-dan
808
+ type: translation
809
+ args: fra-dan
810
+ dataset:
811
+ name: ntrex128
812
+ type: ntrex128
813
+ args: fra-dan
814
+ metrics:
815
+ - name: BLEU
816
+ type: bleu
817
+ value: 27.1
818
+ - name: chr-F
819
+ type: chrf
820
+ value: 0.54102
821
+ - task:
822
+ name: Translation fra-fao
823
+ type: translation
824
+ args: fra-fao
825
+ dataset:
826
+ name: ntrex128
827
+ type: ntrex128
828
+ args: fra-fao
829
+ metrics:
830
+ - name: BLEU
831
+ type: bleu
832
+ value: 10.8
833
+ - name: chr-F
834
+ type: chrf
835
+ value: 0.32337
836
+ - task:
837
+ name: Translation fra-isl
838
+ type: translation
839
+ args: fra-isl
840
+ dataset:
841
+ name: ntrex128
842
+ type: ntrex128
843
+ args: fra-isl
844
+ metrics:
845
+ - name: BLEU
846
+ type: bleu
847
+ value: 18.4
848
+ - name: chr-F
849
+ type: chrf
850
+ value: 0.47296
851
+ - task:
852
+ name: Translation fra-nno
853
+ type: translation
854
+ args: fra-nno
855
+ dataset:
856
+ name: ntrex128
857
+ type: ntrex128
858
+ args: fra-nno
859
+ metrics:
860
+ - name: BLEU
861
+ type: bleu
862
+ value: 21.6
863
+ - name: chr-F
864
+ type: chrf
865
+ value: 0.50532
866
+ - task:
867
+ name: Translation fra-nob
868
+ type: translation
869
+ args: fra-nob
870
+ dataset:
871
+ name: ntrex128
872
+ type: ntrex128
873
+ args: fra-nob
874
+ metrics:
875
+ - name: BLEU
876
+ type: bleu
877
+ value: 25.7
878
+ - name: chr-F
879
+ type: chrf
880
+ value: 0.54026
881
+ - task:
882
+ name: Translation fra-swe
883
+ type: translation
884
+ args: fra-swe
885
+ dataset:
886
+ name: ntrex128
887
+ type: ntrex128
888
+ args: fra-swe
889
+ metrics:
890
+ - name: BLEU
891
+ type: bleu
892
+ value: 27.9
893
+ - name: chr-F
894
+ type: chrf
895
+ value: 0.56278
896
+ - task:
897
+ name: Translation por-dan
898
+ type: translation
899
+ args: por-dan
900
+ dataset:
901
+ name: ntrex128
902
+ type: ntrex128
903
+ args: por-dan
904
+ metrics:
905
+ - name: BLEU
906
+ type: bleu
907
+ value: 30.0
908
+ - name: chr-F
909
+ type: chrf
910
+ value: 0.56288
911
+ - task:
912
+ name: Translation por-fao
913
+ type: translation
914
+ args: por-fao
915
+ dataset:
916
+ name: ntrex128
917
+ type: ntrex128
918
+ args: por-fao
919
+ metrics:
920
+ - name: BLEU
921
+ type: bleu
922
+ value: 12.7
923
+ - name: chr-F
924
+ type: chrf
925
+ value: 0.35059
926
+ - task:
927
+ name: Translation por-isl
928
+ type: translation
929
+ args: por-isl
930
+ dataset:
931
+ name: ntrex128
932
+ type: ntrex128
933
+ args: por-isl
934
+ metrics:
935
+ - name: BLEU
936
+ type: bleu
937
+ value: 17.8
938
+ - name: chr-F
939
+ type: chrf
940
+ value: 0.47577
941
+ - task:
942
+ name: Translation por-nno
943
+ type: translation
944
+ args: por-nno
945
+ dataset:
946
+ name: ntrex128
947
+ type: ntrex128
948
+ args: por-nno
949
+ metrics:
950
+ - name: BLEU
951
+ type: bleu
952
+ value: 23.0
953
+ - name: chr-F
954
+ type: chrf
955
+ value: 0.52158
956
+ - task:
957
+ name: Translation por-nob
958
+ type: translation
959
+ args: por-nob
960
+ dataset:
961
+ name: ntrex128
962
+ type: ntrex128
963
+ args: por-nob
964
+ metrics:
965
+ - name: BLEU
966
+ type: bleu
967
+ value: 27.4
968
+ - name: chr-F
969
+ type: chrf
970
+ value: 0.55788
971
+ - task:
972
+ name: Translation por-swe
973
+ type: translation
974
+ args: por-swe
975
+ dataset:
976
+ name: ntrex128
977
+ type: ntrex128
978
+ args: por-swe
979
+ metrics:
980
+ - name: BLEU
981
+ type: bleu
982
+ value: 29.3
983
+ - name: chr-F
984
+ type: chrf
985
+ value: 0.57790
986
+ - task:
987
+ name: Translation spa-dan
988
+ type: translation
989
+ args: spa-dan
990
+ dataset:
991
+ name: ntrex128
992
+ type: ntrex128
993
+ args: spa-dan
994
+ metrics:
995
+ - name: BLEU
996
+ type: bleu
997
+ value: 27.5
998
+ - name: chr-F
999
+ type: chrf
1000
+ value: 0.55607
1001
+ - task:
1002
+ name: Translation spa-fao
1003
+ type: translation
1004
+ args: spa-fao
1005
+ dataset:
1006
+ name: ntrex128
1007
+ type: ntrex128
1008
+ args: spa-fao
1009
+ metrics:
1010
+ - name: BLEU
1011
+ type: bleu
1012
+ value: 12.5
1013
+ - name: chr-F
1014
+ type: chrf
1015
+ value: 0.34781
1016
+ - task:
1017
+ name: Translation spa-isl
1018
+ type: translation
1019
+ args: spa-isl
1020
+ dataset:
1021
+ name: ntrex128
1022
+ type: ntrex128
1023
+ args: spa-isl
1024
+ metrics:
1025
+ - name: BLEU
1026
+ type: bleu
1027
+ value: 18.4
1028
+ - name: chr-F
1029
+ type: chrf
1030
+ value: 0.48566
1031
+ - task:
1032
+ name: Translation spa-nno
1033
+ type: translation
1034
+ args: spa-nno
1035
+ dataset:
1036
+ name: ntrex128
1037
+ type: ntrex128
1038
+ args: spa-nno
1039
+ metrics:
1040
+ - name: BLEU
1041
+ type: bleu
1042
+ value: 22.2
1043
+ - name: chr-F
1044
+ type: chrf
1045
+ value: 0.51741
1046
+ - task:
1047
+ name: Translation spa-nob
1048
+ type: translation
1049
+ args: spa-nob
1050
+ dataset:
1051
+ name: ntrex128
1052
+ type: ntrex128
1053
+ args: spa-nob
1054
+ metrics:
1055
+ - name: BLEU
1056
+ type: bleu
1057
+ value: 26.8
1058
+ - name: chr-F
1059
+ type: chrf
1060
+ value: 0.55824
1061
+ - task:
1062
+ name: Translation spa-swe
1063
+ type: translation
1064
+ args: spa-swe
1065
+ dataset:
1066
+ name: ntrex128
1067
+ type: ntrex128
1068
+ args: spa-swe
1069
+ metrics:
1070
+ - name: BLEU
1071
+ type: bleu
1072
+ value: 28.8
1073
+ - name: chr-F
1074
+ type: chrf
1075
+ value: 0.57851
1076
+ - task:
1077
+ name: Translation deu-dan
1078
+ type: translation
1079
+ args: deu-dan
1080
+ dataset:
1081
+ name: tatoeba-test-v2021-08-07
1082
+ type: tatoeba_mt
1083
+ args: deu-dan
1084
+ metrics:
1085
+ - name: BLEU
1086
+ type: bleu
1087
+ value: 57.8
1088
+ - name: chr-F
1089
+ type: chrf
1090
+ value: 0.74051
1091
+ - task:
1092
+ name: Translation deu-isl
1093
+ type: translation
1094
+ args: deu-isl
1095
+ dataset:
1096
+ name: tatoeba-test-v2021-08-07
1097
+ type: tatoeba_mt
1098
+ args: deu-isl
1099
+ metrics:
1100
+ - name: BLEU
1101
+ type: bleu
1102
+ value: 31.7
1103
+ - name: chr-F
1104
+ type: chrf
1105
+ value: 0.61256
1106
+ - task:
1107
+ name: Translation deu-nob
1108
+ type: translation
1109
+ args: deu-nob
1110
+ dataset:
1111
+ name: tatoeba-test-v2021-08-07
1112
+ type: tatoeba_mt
1113
+ args: deu-nob
1114
+ metrics:
1115
+ - name: BLEU
1116
+ type: bleu
1117
+ value: 52.9
1118
+ - name: chr-F
1119
+ type: chrf
1120
+ value: 0.71413
1121
+ - task:
1122
+ name: Translation deu-nor
1123
+ type: translation
1124
+ args: deu-nor
1125
+ dataset:
1126
+ name: tatoeba-test-v2021-08-07
1127
+ type: tatoeba_mt
1128
+ args: deu-nor
1129
+ metrics:
1130
+ - name: BLEU
1131
+ type: bleu
1132
+ value: 52.7
1133
+ - name: chr-F
1134
+ type: chrf
1135
+ value: 0.71253
1136
+ - task:
1137
+ name: Translation deu-swe
1138
+ type: translation
1139
+ args: deu-swe
1140
+ dataset:
1141
+ name: tatoeba-test-v2021-08-07
1142
+ type: tatoeba_mt
1143
+ args: deu-swe
1144
+ metrics:
1145
+ - name: BLEU
1146
+ type: bleu
1147
+ value: 58.2
1148
+ - name: chr-F
1149
+ type: chrf
1150
+ value: 0.72650
1151
+ - task:
1152
+ name: Translation eng-dan
1153
+ type: translation
1154
+ args: eng-dan
1155
+ dataset:
1156
+ name: tatoeba-test-v2021-08-07
1157
+ type: tatoeba_mt
1158
+ args: eng-dan
1159
+ metrics:
1160
+ - name: BLEU
1161
+ type: bleu
1162
+ value: 60.6
1163
+ - name: chr-F
1164
+ type: chrf
1165
+ value: 0.74708
1166
+ - task:
1167
+ name: Translation eng-fao
1168
+ type: translation
1169
+ args: eng-fao
1170
+ dataset:
1171
+ name: tatoeba-test-v2021-08-07
1172
+ type: tatoeba_mt
1173
+ args: eng-fao
1174
+ metrics:
1175
+ - name: BLEU
1176
+ type: bleu
1177
+ value: 29.0
1178
+ - name: chr-F
1179
+ type: chrf
1180
+ value: 0.48304
1181
+ - task:
1182
+ name: Translation eng-isl
1183
+ type: translation
1184
+ args: eng-isl
1185
+ dataset:
1186
+ name: tatoeba-test-v2021-08-07
1187
+ type: tatoeba_mt
1188
+ args: eng-isl
1189
+ metrics:
1190
+ - name: BLEU
1191
+ type: bleu
1192
+ value: 33.2
1193
+ - name: chr-F
1194
+ type: chrf
1195
+ value: 0.58312
1196
+ - task:
1197
+ name: Translation eng-nno
1198
+ type: translation
1199
+ args: eng-nno
1200
+ dataset:
1201
+ name: tatoeba-test-v2021-08-07
1202
+ type: tatoeba_mt
1203
+ args: eng-nno
1204
+ metrics:
1205
+ - name: BLEU
1206
+ type: bleu
1207
+ value: 42.7
1208
+ - name: chr-F
1209
+ type: chrf
1210
+ value: 0.62606
1211
+ - task:
1212
+ name: Translation eng-nob
1213
+ type: translation
1214
+ args: eng-nob
1215
+ dataset:
1216
+ name: tatoeba-test-v2021-08-07
1217
+ type: tatoeba_mt
1218
+ args: eng-nob
1219
+ metrics:
1220
+ - name: BLEU
1221
+ type: bleu
1222
+ value: 57.4
1223
+ - name: chr-F
1224
+ type: chrf
1225
+ value: 0.72340
1226
+ - task:
1227
+ name: Translation eng-nor
1228
+ type: translation
1229
+ args: eng-nor
1230
+ dataset:
1231
+ name: tatoeba-test-v2021-08-07
1232
+ type: tatoeba_mt
1233
+ args: eng-nor
1234
+ metrics:
1235
+ - name: BLEU
1236
+ type: bleu
1237
+ value: 56.2
1238
+ - name: chr-F
1239
+ type: chrf
1240
+ value: 0.71514
1241
+ - task:
1242
+ name: Translation eng-swe
1243
+ type: translation
1244
+ args: eng-swe
1245
+ dataset:
1246
+ name: tatoeba-test-v2021-08-07
1247
+ type: tatoeba_mt
1248
+ args: eng-swe
1249
+ metrics:
1250
+ - name: BLEU
1251
+ type: bleu
1252
+ value: 60.5
1253
+ - name: chr-F
1254
+ type: chrf
1255
+ value: 0.73720
1256
+ - task:
1257
+ name: Translation fra-dan
1258
+ type: translation
1259
+ args: fra-dan
1260
+ dataset:
1261
+ name: tatoeba-test-v2021-08-07
1262
+ type: tatoeba_mt
1263
+ args: fra-dan
1264
+ metrics:
1265
+ - name: BLEU
1266
+ type: bleu
1267
+ value: 64.1
1268
+ - name: chr-F
1269
+ type: chrf
1270
+ value: 0.78018
1271
+ - task:
1272
+ name: Translation fra-nob
1273
+ type: translation
1274
+ args: fra-nob
1275
+ dataset:
1276
+ name: tatoeba-test-v2021-08-07
1277
+ type: tatoeba_mt
1278
+ args: fra-nob
1279
+ metrics:
1280
+ - name: BLEU
1281
+ type: bleu
1282
+ value: 59.1
1283
+ - name: chr-F
1284
+ type: chrf
1285
+ value: 0.74252
1286
+ - task:
1287
+ name: Translation fra-nor
1288
+ type: translation
1289
+ args: fra-nor
1290
+ dataset:
1291
+ name: tatoeba-test-v2021-08-07
1292
+ type: tatoeba_mt
1293
+ args: fra-nor
1294
+ metrics:
1295
+ - name: BLEU
1296
+ type: bleu
1297
+ value: 60.3
1298
+ - name: chr-F
1299
+ type: chrf
1300
+ value: 0.74407
1301
+ - task:
1302
+ name: Translation fra-swe
1303
+ type: translation
1304
+ args: fra-swe
1305
+ dataset:
1306
+ name: tatoeba-test-v2021-08-07
1307
+ type: tatoeba_mt
1308
+ args: fra-swe
1309
+ metrics:
1310
+ - name: BLEU
1311
+ type: bleu
1312
+ value: 62.1
1313
+ - name: chr-F
1314
+ type: chrf
1315
+ value: 0.75644
1316
+ - task:
1317
+ name: Translation multi-multi
1318
+ type: translation
1319
+ args: multi-multi
1320
+ dataset:
1321
+ name: tatoeba-test-v2020-07-28-v2023-09-26
1322
+ type: tatoeba_mt
1323
+ args: multi-multi
1324
+ metrics:
1325
+ - name: BLEU
1326
+ type: bleu
1327
+ value: 56.4
1328
+ - name: chr-F
1329
+ type: chrf
1330
+ value: 0.72858
1331
+ - task:
1332
+ name: Translation por-dan
1333
+ type: translation
1334
+ args: por-dan
1335
+ dataset:
1336
+ name: tatoeba-test-v2021-08-07
1337
+ type: tatoeba_mt
1338
+ args: por-dan
1339
+ metrics:
1340
+ - name: BLEU
1341
+ type: bleu
1342
+ value: 65.6
1343
+ - name: chr-F
1344
+ type: chrf
1345
+ value: 0.79528
1346
+ - task:
1347
+ name: Translation por-nor
1348
+ type: translation
1349
+ args: por-nor
1350
+ dataset:
1351
+ name: tatoeba-test-v2021-08-07
1352
+ type: tatoeba_mt
1353
+ args: por-nor
1354
+ metrics:
1355
+ - name: BLEU
1356
+ type: bleu
1357
+ value: 58.0
1358
+ - name: chr-F
1359
+ type: chrf
1360
+ value: 0.73559
1361
+ - task:
1362
+ name: Translation por-swe
1363
+ type: translation
1364
+ args: por-swe
1365
+ dataset:
1366
+ name: tatoeba-test-v2021-08-07
1367
+ type: tatoeba_mt
1368
+ args: por-swe
1369
+ metrics:
1370
+ - name: BLEU
1371
+ type: bleu
1372
+ value: 60.2
1373
+ - name: chr-F
1374
+ type: chrf
1375
+ value: 0.75566
1376
+ - task:
1377
+ name: Translation spa-dan
1378
+ type: translation
1379
+ args: spa-dan
1380
+ dataset:
1381
+ name: tatoeba-test-v2021-08-07
1382
+ type: tatoeba_mt
1383
+ args: spa-dan
1384
+ metrics:
1385
+ - name: BLEU
1386
+ type: bleu
1387
+ value: 57.7
1388
+ - name: chr-F
1389
+ type: chrf
1390
+ value: 0.73310
1391
+ - task:
1392
+ name: Translation spa-nob
1393
+ type: translation
1394
+ args: spa-nob
1395
+ dataset:
1396
+ name: tatoeba-test-v2021-08-07
1397
+ type: tatoeba_mt
1398
+ args: spa-nob
1399
+ metrics:
1400
+ - name: BLEU
1401
+ type: bleu
1402
+ value: 60.9
1403
+ - name: chr-F
1404
+ type: chrf
1405
+ value: 0.76501
1406
+ - task:
1407
+ name: Translation spa-nor
1408
+ type: translation
1409
+ args: spa-nor
1410
+ dataset:
1411
+ name: tatoeba-test-v2021-08-07
1412
+ type: tatoeba_mt
1413
+ args: spa-nor
1414
+ metrics:
1415
+ - name: BLEU
1416
+ type: bleu
1417
+ value: 60.1
1418
+ - name: chr-F
1419
+ type: chrf
1420
+ value: 0.75815
1421
+ - task:
1422
+ name: Translation spa-swe
1423
+ type: translation
1424
+ args: spa-swe
1425
+ dataset:
1426
+ name: tatoeba-test-v2021-08-07
1427
+ type: tatoeba_mt
1428
+ args: spa-swe
1429
+ metrics:
1430
+ - name: BLEU
1431
+ type: bleu
1432
+ value: 60.7
1433
+ - name: chr-F
1434
+ type: chrf
1435
+ value: 0.74222
1436
+ - task:
1437
+ name: Translation eng-isl
1438
+ type: translation
1439
+ args: eng-isl
1440
+ dataset:
1441
+ name: newstest2021
1442
+ type: wmt-2021-news
1443
+ args: eng-isl
1444
+ metrics:
1445
+ - name: BLEU
1446
+ type: bleu
1447
+ value: 21.9
1448
+ - name: chr-F
1449
+ type: chrf
1450
+ value: 0.51196
1451
+ ---
1452
+ # opus-mt-tc-bible-big-deu_eng_fra_por_spa-gmq
1453
+
1454
+ ## Table of Contents
1455
+ - [Model Details](#model-details)
1456
+ - [Uses](#uses)
1457
+ - [Risks, Limitations and Biases](#risks-limitations-and-biases)
1458
+ - [How to Get Started With the Model](#how-to-get-started-with-the-model)
1459
+ - [Training](#training)
1460
+ - [Evaluation](#evaluation)
1461
+ - [Citation Information](#citation-information)
1462
+ - [Acknowledgements](#acknowledgements)
1463
+
1464
+ ## Model Details
1465
+
1466
+ Neural machine translation model for translating from unknown (deu+eng+fra+por+spa) to North Germanic languages (gmq).
1467
+
1468
+ 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).
1469
+ **Model Description:**
1470
+ - **Developed by:** Language Technology Research Group at the University of Helsinki
1471
+ - **Model Type:** Translation (transformer-big)
1472
+ - **Release**: 2024-05-30
1473
+ - **License:** Apache-2.0
1474
+ - **Language(s):**
1475
+ - Source Language(s): deu eng fra por spa
1476
+ - Target Language(s): dan fao isl nno nob non nor swe
1477
+ - Valid Target Language Labels: >>dan<< >>fao<< >>isl<< >>jut<< >>nno<< >>nob<< >>non<< >>nor<< >>nrn<< >>ovd<< >>rmg<< >>swe<< >>xxx<<
1478
+ - **Original Model**: [opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-30.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/deu+eng+fra+por+spa-gmq/opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-30.zip)
1479
+ - **Resources for more information:**
1480
+ - [OPUS-MT dashboard](https://opus.nlpl.eu/dashboard/index.php?pkg=opusmt&test=all&scoreslang=all&chart=standard&model=Tatoeba-MT-models/deu%2Beng%2Bfra%2Bpor%2Bspa-gmq/opusTCv20230926max50%2Bbt%2Bjhubc_transformer-big_2024-05-30)
1481
+ - [OPUS-MT-train GitHub Repo](https://github.com/Helsinki-NLP/OPUS-MT-train)
1482
+ - [More information about MarianNMT models in the transformers library](https://huggingface.co/docs/transformers/model_doc/marian)
1483
+ - [Tatoeba Translation Challenge](https://github.com/Helsinki-NLP/Tatoeba-Challenge/)
1484
+ - [HPLT bilingual data v1 (as part of the Tatoeba Translation Challenge dataset)](https://hplt-project.org/datasets/v1)
1485
+ - [A massively parallel Bible corpus](https://aclanthology.org/L14-1215/)
1486
+
1487
+ 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. `>>dan<<`
1488
+
1489
+ ## Uses
1490
+
1491
+ This model can be used for translation and text-to-text generation.
1492
+
1493
+ ## Risks, Limitations and Biases
1494
+
1495
+ **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.**
1496
+
1497
+ 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)).
1498
+
1499
+ ## How to Get Started With the Model
1500
+
1501
+ A short example code:
1502
+
1503
+ ```python
1504
+ from transformers import MarianMTModel, MarianTokenizer
1505
+
1506
+ src_text = [
1507
+ ">>dan<< Replace this with text in an accepted source language.",
1508
+ ">>swe<< This is the second sentence."
1509
+ ]
1510
+
1511
+ model_name = "pytorch-models/opus-mt-tc-bible-big-deu_eng_fra_por_spa-gmq"
1512
+ tokenizer = MarianTokenizer.from_pretrained(model_name)
1513
+ model = MarianMTModel.from_pretrained(model_name)
1514
+ translated = model.generate(**tokenizer(src_text, return_tensors="pt", padding=True))
1515
+
1516
+ for t in translated:
1517
+ print( tokenizer.decode(t, skip_special_tokens=True) )
1518
+ ```
1519
+
1520
+ You can also use OPUS-MT models with the transformers pipelines, for example:
1521
+
1522
+ ```python
1523
+ from transformers import pipeline
1524
+ pipe = pipeline("translation", model="Helsinki-NLP/opus-mt-tc-bible-big-deu_eng_fra_por_spa-gmq")
1525
+ print(pipe(">>dan<< Replace this with text in an accepted source language."))
1526
+ ```
1527
+
1528
+ ## Training
1529
+
1530
+ - **Data**: opusTCv20230926max50+bt+jhubc ([source](https://github.com/Helsinki-NLP/Tatoeba-Challenge))
1531
+ - **Pre-processing**: SentencePiece (spm32k,spm32k)
1532
+ - **Model Type:** transformer-big
1533
+ - **Original MarianNMT Model**: [opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-30.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/deu+eng+fra+por+spa-gmq/opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-30.zip)
1534
+ - **Training Scripts**: [GitHub Repo](https://github.com/Helsinki-NLP/OPUS-MT-train)
1535
+
1536
+ ## Evaluation
1537
+
1538
+ * [Model scores at the OPUS-MT dashboard](https://opus.nlpl.eu/dashboard/index.php?pkg=opusmt&test=all&scoreslang=all&chart=standard&model=Tatoeba-MT-models/deu%2Beng%2Bfra%2Bpor%2Bspa-gmq/opusTCv20230926max50%2Bbt%2Bjhubc_transformer-big_2024-05-30)
1539
+ * test set translations: [opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-29.test.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/deu+eng+fra+por+spa-gmq/opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-29.test.txt)
1540
+ * test set scores: [opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-29.eval.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/deu+eng+fra+por+spa-gmq/opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-29.eval.txt)
1541
+ * benchmark results: [benchmark_results.txt](benchmark_results.txt)
1542
+ * benchmark output: [benchmark_translations.zip](benchmark_translations.zip)
1543
+
1544
+ | langpair | testset | chr-F | BLEU | #sent | #words |
1545
+ |----------|---------|-------|-------|-------|--------|
1546
+ | deu-dan | tatoeba-test-v2021-08-07 | 0.74051 | 57.8 | 9998 | 74644 |
1547
+ | deu-isl | tatoeba-test-v2021-08-07 | 0.61256 | 31.7 | 969 | 5951 |
1548
+ | deu-nob | tatoeba-test-v2021-08-07 | 0.71413 | 52.9 | 3525 | 31978 |
1549
+ | deu-nor | tatoeba-test-v2021-08-07 | 0.71253 | 52.7 | 3651 | 32928 |
1550
+ | deu-swe | tatoeba-test-v2021-08-07 | 0.72650 | 58.2 | 3410 | 22701 |
1551
+ | eng-dan | tatoeba-test-v2021-08-07 | 0.74708 | 60.6 | 10795 | 79385 |
1552
+ | eng-fao | tatoeba-test-v2021-08-07 | 0.48304 | 29.0 | 294 | 1933 |
1553
+ | eng-isl | tatoeba-test-v2021-08-07 | 0.58312 | 33.2 | 2503 | 19023 |
1554
+ | eng-nno | tatoeba-test-v2021-08-07 | 0.62606 | 42.7 | 460 | 3428 |
1555
+ | eng-nob | tatoeba-test-v2021-08-07 | 0.72340 | 57.4 | 4539 | 36119 |
1556
+ | eng-nor | tatoeba-test-v2021-08-07 | 0.71514 | 56.2 | 5000 | 39552 |
1557
+ | eng-swe | tatoeba-test-v2021-08-07 | 0.73720 | 60.5 | 10362 | 68067 |
1558
+ | fra-dan | tatoeba-test-v2021-08-07 | 0.78018 | 64.1 | 1731 | 11312 |
1559
+ | fra-nob | tatoeba-test-v2021-08-07 | 0.74252 | 59.1 | 323 | 2175 |
1560
+ | fra-nor | tatoeba-test-v2021-08-07 | 0.74407 | 60.3 | 477 | 3097 |
1561
+ | fra-swe | tatoeba-test-v2021-08-07 | 0.75644 | 62.1 | 1407 | 9170 |
1562
+ | por-dan | tatoeba-test-v2021-08-07 | 0.79528 | 65.6 | 873 | 5258 |
1563
+ | por-nor | tatoeba-test-v2021-08-07 | 0.73559 | 58.0 | 481 | 4030 |
1564
+ | por-swe | tatoeba-test-v2021-08-07 | 0.75566 | 60.2 | 320 | 1938 |
1565
+ | spa-dan | tatoeba-test-v2021-08-07 | 0.73310 | 57.7 | 5000 | 35937 |
1566
+ | spa-isl | tatoeba-test-v2021-08-07 | 0.52169 | 18.7 | 238 | 1220 |
1567
+ | spa-nob | tatoeba-test-v2021-08-07 | 0.76501 | 60.9 | 885 | 6762 |
1568
+ | spa-nor | tatoeba-test-v2021-08-07 | 0.75815 | 60.1 | 960 | 7217 |
1569
+ | spa-swe | tatoeba-test-v2021-08-07 | 0.74222 | 60.7 | 1351 | 8357 |
1570
+ | deu-dan | flores101-devtest | 0.62006 | 34.8 | 1012 | 24638 |
1571
+ | deu-isl | flores101-devtest | 0.48236 | 18.8 | 1012 | 22834 |
1572
+ | deu-swe | flores101-devtest | 0.61778 | 33.7 | 1012 | 23121 |
1573
+ | eng-swe | flores101-devtest | 0.69435 | 45.5 | 1012 | 23121 |
1574
+ | fra-dan | flores101-devtest | 0.61019 | 34.0 | 1012 | 24638 |
1575
+ | fra-isl | flores101-devtest | 0.47647 | 18.1 | 1012 | 22834 |
1576
+ | fra-swe | flores101-devtest | 0.60354 | 32.2 | 1012 | 23121 |
1577
+ | por-isl | flores101-devtest | 0.47937 | 19.1 | 1012 | 22834 |
1578
+ | por-swe | flores101-devtest | 0.60857 | 33.1 | 1012 | 23121 |
1579
+ | spa-dan | flores101-devtest | 0.54890 | 24.4 | 1012 | 24638 |
1580
+ | spa-nob | flores101-devtest | 0.50610 | 18.3 | 1012 | 23873 |
1581
+ | spa-swe | flores101-devtest | 0.54011 | 22.4 | 1012 | 23121 |
1582
+ | deu-dan | flores200-devtest | 0.62152 | 35.1 | 1012 | 24638 |
1583
+ | deu-isl | flores200-devtest | 0.48648 | 19.1 | 1012 | 22834 |
1584
+ | deu-nno | flores200-devtest | 0.53530 | 24.0 | 1012 | 24316 |
1585
+ | deu-nob | flores200-devtest | 0.55748 | 25.1 | 1012 | 23873 |
1586
+ | deu-swe | flores200-devtest | 0.62138 | 34.2 | 1012 | 23121 |
1587
+ | eng-dan | flores200-devtest | 0.70321 | 47.0 | 1012 | 24638 |
1588
+ | eng-isl | flores200-devtest | 0.52585 | 24.4 | 1012 | 22834 |
1589
+ | eng-nno | flores200-devtest | 0.61372 | 33.8 | 1012 | 24316 |
1590
+ | eng-nob | flores200-devtest | 0.62508 | 34.4 | 1012 | 23873 |
1591
+ | eng-swe | flores200-devtest | 0.69703 | 46.0 | 1012 | 23121 |
1592
+ | fra-dan | flores200-devtest | 0.61025 | 34.1 | 1012 | 24638 |
1593
+ | fra-isl | flores200-devtest | 0.48273 | 18.8 | 1012 | 22834 |
1594
+ | fra-nno | flores200-devtest | 0.53032 | 24.3 | 1012 | 24316 |
1595
+ | fra-nob | flores200-devtest | 0.54933 | 25.0 | 1012 | 23873 |
1596
+ | fra-swe | flores200-devtest | 0.60612 | 32.8 | 1012 | 23121 |
1597
+ | por-dan | flores200-devtest | 0.62221 | 36.2 | 1012 | 24638 |
1598
+ | por-isl | flores200-devtest | 0.48357 | 19.6 | 1012 | 22834 |
1599
+ | por-nno | flores200-devtest | 0.54369 | 26.3 | 1012 | 24316 |
1600
+ | por-nob | flores200-devtest | 0.56054 | 26.4 | 1012 | 23873 |
1601
+ | por-swe | flores200-devtest | 0.61388 | 34.1 | 1012 | 23121 |
1602
+ | spa-dan | flores200-devtest | 0.55091 | 24.7 | 1012 | 24638 |
1603
+ | spa-isl | flores200-devtest | 0.44469 | 14.2 | 1012 | 22834 |
1604
+ | spa-nno | flores200-devtest | 0.48898 | 18.6 | 1012 | 24316 |
1605
+ | spa-nob | flores200-devtest | 0.50901 | 18.8 | 1012 | 23873 |
1606
+ | spa-swe | flores200-devtest | 0.54182 | 22.7 | 1012 | 23121 |
1607
+ | eng-isl | newstest2021 | 0.51196 | 21.9 | 1000 | 25233 |
1608
+ | deu-dan | ntrex128 | 0.56412 | 29.1 | 1997 | 47643 |
1609
+ | deu-isl | ntrex128 | 0.48309 | 18.8 | 1997 | 46643 |
1610
+ | deu-nno | ntrex128 | 0.51535 | 22.0 | 1997 | 46512 |
1611
+ | deu-nob | ntrex128 | 0.56152 | 27.6 | 1997 | 45501 |
1612
+ | deu-swe | ntrex128 | 0.58061 | 29.6 | 1997 | 44889 |
1613
+ | eng-dan | ntrex128 | 0.61894 | 37.6 | 1997 | 47643 |
1614
+ | eng-isl | ntrex128 | 0.52027 | 23.9 | 1997 | 46643 |
1615
+ | eng-nno | ntrex128 | 0.60754 | 34.0 | 1997 | 46512 |
1616
+ | eng-nob | ntrex128 | 0.62327 | 36.9 | 1997 | 45501 |
1617
+ | eng-swe | ntrex128 | 0.66129 | 41.3 | 1997 | 44889 |
1618
+ | fra-dan | ntrex128 | 0.54102 | 27.1 | 1997 | 47643 |
1619
+ | fra-isl | ntrex128 | 0.47296 | 18.4 | 1997 | 46643 |
1620
+ | fra-nno | ntrex128 | 0.50532 | 21.6 | 1997 | 46512 |
1621
+ | fra-nob | ntrex128 | 0.54026 | 25.7 | 1997 | 45501 |
1622
+ | fra-swe | ntrex128 | 0.56278 | 27.9 | 1997 | 44889 |
1623
+ | por-dan | ntrex128 | 0.56288 | 30.0 | 1997 | 47643 |
1624
+ | por-isl | ntrex128 | 0.47577 | 17.8 | 1997 | 46643 |
1625
+ | por-nno | ntrex128 | 0.52158 | 23.0 | 1997 | 46512 |
1626
+ | por-nob | ntrex128 | 0.55788 | 27.4 | 1997 | 45501 |
1627
+ | por-swe | ntrex128 | 0.57790 | 29.3 | 1997 | 44889 |
1628
+ | spa-dan | ntrex128 | 0.55607 | 27.5 | 1997 | 47643 |
1629
+ | spa-isl | ntrex128 | 0.48566 | 18.4 | 1997 | 46643 |
1630
+ | spa-nno | ntrex128 | 0.51741 | 22.2 | 1997 | 46512 |
1631
+ | spa-nob | ntrex128 | 0.55824 | 26.8 | 1997 | 45501 |
1632
+ | spa-swe | ntrex128 | 0.57851 | 28.8 | 1997 | 44889 |
1633
+
1634
+ ## Citation Information
1635
+
1636
+ * Publications: [Democratizing neural machine translation with OPUS-MT](https://doi.org/10.1007/s10579-023-09704-w) and [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.)
1637
+
1638
+ ```bibtex
1639
+ @article{tiedemann2023democratizing,
1640
+ title={Democratizing neural machine translation with {OPUS-MT}},
1641
+ author={Tiedemann, J{\"o}rg and Aulamo, Mikko and Bakshandaeva, Daria and Boggia, Michele and Gr{\"o}nroos, Stig-Arne and Nieminen, Tommi and Raganato, Alessandro and Scherrer, Yves and Vazquez, Raul and Virpioja, Sami},
1642
+ journal={Language Resources and Evaluation},
1643
+ number={58},
1644
+ pages={713--755},
1645
+ year={2023},
1646
+ publisher={Springer Nature},
1647
+ issn={1574-0218},
1648
+ doi={10.1007/s10579-023-09704-w}
1649
+ }
1650
+
1651
+ @inproceedings{tiedemann-thottingal-2020-opus,
1652
+ title = "{OPUS}-{MT} {--} Building open translation services for the World",
1653
+ author = {Tiedemann, J{\"o}rg and Thottingal, Santhosh},
1654
+ booktitle = "Proceedings of the 22nd Annual Conference of the European Association for Machine Translation",
1655
+ month = nov,
1656
+ year = "2020",
1657
+ address = "Lisboa, Portugal",
1658
+ publisher = "European Association for Machine Translation",
1659
+ url = "https://aclanthology.org/2020.eamt-1.61",
1660
+ pages = "479--480",
1661
+ }
1662
+
1663
+ @inproceedings{tiedemann-2020-tatoeba,
1664
+ title = "The Tatoeba Translation Challenge {--} Realistic Data Sets for Low Resource and Multilingual {MT}",
1665
+ author = {Tiedemann, J{\"o}rg},
1666
+ booktitle = "Proceedings of the Fifth Conference on Machine Translation",
1667
+ month = nov,
1668
+ year = "2020",
1669
+ address = "Online",
1670
+ publisher = "Association for Computational Linguistics",
1671
+ url = "https://aclanthology.org/2020.wmt-1.139",
1672
+ pages = "1174--1182",
1673
+ }
1674
+ ```
1675
+
1676
+ ## Acknowledgements
1677
+
1678
+ The work is supported by the [HPLT project](https://hplt-project.org/), funded by the European Union’s Horizon Europe research and innovation programme under grant agreement No 101070350. We are also grateful for the generous computational resources and IT infrastructure provided by [CSC -- IT Center for Science](https://www.csc.fi/), Finland, and the [EuroHPC supercomputer LUMI](https://www.lumi-supercomputer.eu/).
1679
+
1680
+ ## Model conversion info
1681
+
1682
+ * transformers version: 4.45.1
1683
+ * OPUS-MT git hash: 0882077
1684
+ * port time: Tue Oct 8 09:37:58 EEST 2024
1685
+ * port machine: LM0-400-22516.local
benchmark_results.txt ADDED
@@ -0,0 +1,116 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ multi-multi tatoeba-test-v2020-07-28-v2023-09-26 0.72858 56.4 10000 73676
2
+ deu-dan flores101-devtest 0.62006 34.8 1012 24638
3
+ deu-isl flores101-devtest 0.48236 18.8 1012 22834
4
+ deu-swe flores101-devtest 0.61778 33.7 1012 23121
5
+ eng-swe flores101-devtest 0.69435 45.5 1012 23121
6
+ fra-dan flores101-devtest 0.61019 34.0 1012 24638
7
+ fra-isl flores101-devtest 0.47647 18.1 1012 22834
8
+ fra-swe flores101-devtest 0.60354 32.2 1012 23121
9
+ por-isl flores101-devtest 0.47937 19.1 1012 22834
10
+ por-swe flores101-devtest 0.60857 33.1 1012 23121
11
+ spa-dan flores101-devtest 0.54890 24.4 1012 24638
12
+ spa-nob flores101-devtest 0.50610 18.3 1012 23873
13
+ spa-swe flores101-devtest 0.54011 22.4 1012 23121
14
+ deu-dan flores200-devtest 0.62152 35.1 1012 24638
15
+ deu-fao flores200-devtest 0.33611 11.5 1012 24539
16
+ deu-isl flores200-devtest 0.48648 19.1 1012 22834
17
+ deu-nno flores200-devtest 0.53530 24.0 1012 24316
18
+ deu-nob flores200-devtest 0.55748 25.1 1012 23873
19
+ deu-swe flores200-devtest 0.62138 34.2 1012 23121
20
+ eng-dan flores200-devtest 0.70321 47.0 1012 24638
21
+ eng-fao flores200-devtest 0.35857 14.1 1012 24539
22
+ eng-isl flores200-devtest 0.52585 24.4 1012 22834
23
+ eng-nno flores200-devtest 0.61372 33.8 1012 24316
24
+ eng-nob flores200-devtest 0.62508 34.4 1012 23873
25
+ eng-swe flores200-devtest 0.69703 46.0 1012 23121
26
+ fra-dan flores200-devtest 0.61025 34.1 1012 24638
27
+ fra-fao flores200-devtest 0.29532 9.2 1012 24539
28
+ fra-isl flores200-devtest 0.48273 18.8 1012 22834
29
+ fra-nno flores200-devtest 0.53032 24.3 1012 24316
30
+ fra-nob flores200-devtest 0.54933 25.0 1012 23873
31
+ fra-swe flores200-devtest 0.60612 32.8 1012 23121
32
+ por-dan flores200-devtest 0.62221 36.2 1012 24638
33
+ por-fao flores200-devtest 0.33159 11.5 1012 24539
34
+ por-isl flores200-devtest 0.48357 19.6 1012 22834
35
+ por-nno flores200-devtest 0.54369 26.3 1012 24316
36
+ por-nob flores200-devtest 0.56054 26.4 1012 23873
37
+ por-swe flores200-devtest 0.61388 34.1 1012 23121
38
+ spa-dan flores200-devtest 0.55091 24.7 1012 24638
39
+ spa-fao flores200-devtest 0.29284 8.2 1012 24539
40
+ spa-isl flores200-devtest 0.44469 14.2 1012 22834
41
+ spa-nno flores200-devtest 0.48898 18.6 1012 24316
42
+ spa-nob flores200-devtest 0.50901 18.8 1012 23873
43
+ spa-swe flores200-devtest 0.54182 22.7 1012 23121
44
+ eng-isl newstest2021 0.51196 21.9 1000 25233
45
+ deu-dan ntrex128 0.56412 29.1 1997 47643
46
+ deu-fao ntrex128 0.35495 12.5 1997 48359
47
+ deu-isl ntrex128 0.48309 18.8 1997 46643
48
+ deu-nno ntrex128 0.51535 22.0 1997 46512
49
+ deu-nob ntrex128 0.56152 27.6 1997 45501
50
+ deu-swe ntrex128 0.58061 29.6 1997 44889
51
+ eng-dan ntrex128 0.61894 37.6 1997 47643
52
+ eng-fao ntrex128 0.38410 15.9 1997 48359
53
+ eng-isl ntrex128 0.52027 23.9 1997 46643
54
+ eng-nno ntrex128 0.60754 34.0 1997 46512
55
+ eng-nob ntrex128 0.62327 36.9 1997 45501
56
+ eng-swe ntrex128 0.66129 41.3 1997 44889
57
+ fra-dan ntrex128 0.54102 27.1 1997 47643
58
+ fra-fao ntrex128 0.32337 10.8 1997 48359
59
+ fra-isl ntrex128 0.47296 18.4 1997 46643
60
+ fra-nno ntrex128 0.50532 21.6 1997 46512
61
+ fra-nob ntrex128 0.54026 25.7 1997 45501
62
+ fra-swe ntrex128 0.56278 27.9 1997 44889
63
+ por-dan ntrex128 0.56288 30.0 1997 47643
64
+ por-fao ntrex128 0.35059 12.7 1997 48359
65
+ por-isl ntrex128 0.47577 17.8 1997 46643
66
+ por-nno ntrex128 0.52158 23.0 1997 46512
67
+ por-nob ntrex128 0.55788 27.4 1997 45501
68
+ por-swe ntrex128 0.57790 29.3 1997 44889
69
+ spa-dan ntrex128 0.55607 27.5 1997 47643
70
+ spa-fao ntrex128 0.34781 12.5 1997 48359
71
+ spa-isl ntrex128 0.48566 18.4 1997 46643
72
+ spa-nno ntrex128 0.51741 22.2 1997 46512
73
+ spa-nob ntrex128 0.55824 26.8 1997 45501
74
+ spa-swe ntrex128 0.57851 28.8 1997 44889
75
+ deu-dan tatoeba-test-v2020-07-28 0.73909 57.7 10000 74655
76
+ deu-isl tatoeba-test-v2020-07-28 0.61690 32.8 968 5939
77
+ eng-dan tatoeba-test-v2020-07-28 0.74054 57.5 10000 73213
78
+ eng-swe tatoeba-test-v2020-07-28 0.73589 60.3 10000 65581
79
+ fra-dan tatoeba-test-v2020-07-28 0.78077 64.6 1742 11361
80
+ fra-nob tatoeba-test-v2020-07-28 0.73652 57.9 322 2167
81
+ fra-nor tatoeba-test-v2020-07-28 0.73456 58.3 476 3089
82
+ fra-swe tatoeba-test-v2020-07-28 0.75618 61.9 1409 9175
83
+ por-swe tatoeba-test-v2020-07-28 0.75207 60.3 299 1790
84
+ deu-dan tatoeba-test-v2021-03-30 0.73909 57.7 10000 74655
85
+ eng-dan tatoeba-test-v2021-03-30 0.74146 57.8 10437 76467
86
+ eng-isl tatoeba-test-v2021-03-30 0.58295 33.8 4994 37970
87
+ fra-nob tatoeba-test-v2021-03-30 0.73921 58.4 326 2191
88
+ fra-nor tatoeba-test-v2021-03-30 0.73715 58.8 481 3120
89
+ por-swe tatoeba-test-v2021-03-30 0.74938 60.1 319 1910
90
+ spa-isl tatoeba-test-v2021-03-30 0.52932 19.1 239 1224
91
+ spa-nob tatoeba-test-v2021-03-30 0.76219 60.7 894 6838
92
+ spa-nor tatoeba-test-v2021-03-30 0.75593 60.1 970 7300
93
+ deu-dan tatoeba-test-v2021-08-07 0.74051 57.8 9998 74644
94
+ deu-isl tatoeba-test-v2021-08-07 0.61256 31.7 969 5951
95
+ deu-nob tatoeba-test-v2021-08-07 0.71413 52.9 3525 31978
96
+ deu-nor tatoeba-test-v2021-08-07 0.71253 52.7 3651 32928
97
+ deu-swe tatoeba-test-v2021-08-07 0.72650 58.2 3410 22701
98
+ eng-dan tatoeba-test-v2021-08-07 0.74708 60.6 10795 79385
99
+ eng-fao tatoeba-test-v2021-08-07 0.48304 29.0 294 1933
100
+ eng-isl tatoeba-test-v2021-08-07 0.58312 33.2 2503 19023
101
+ eng-nno tatoeba-test-v2021-08-07 0.62606 42.7 460 3428
102
+ eng-nob tatoeba-test-v2021-08-07 0.72340 57.4 4539 36119
103
+ eng-nor tatoeba-test-v2021-08-07 0.71514 56.2 5000 39552
104
+ eng-swe tatoeba-test-v2021-08-07 0.73720 60.5 10362 68067
105
+ fra-dan tatoeba-test-v2021-08-07 0.78018 64.1 1731 11312
106
+ fra-nob tatoeba-test-v2021-08-07 0.74252 59.1 323 2175
107
+ fra-nor tatoeba-test-v2021-08-07 0.74407 60.3 477 3097
108
+ fra-swe tatoeba-test-v2021-08-07 0.75644 62.1 1407 9170
109
+ por-dan tatoeba-test-v2021-08-07 0.79528 65.6 873 5258
110
+ por-nor tatoeba-test-v2021-08-07 0.73559 58.0 481 4030
111
+ por-swe tatoeba-test-v2021-08-07 0.75566 60.2 320 1938
112
+ spa-dan tatoeba-test-v2021-08-07 0.73310 57.7 5000 35937
113
+ spa-isl tatoeba-test-v2021-08-07 0.52169 18.7 238 1220
114
+ spa-nob tatoeba-test-v2021-08-07 0.76501 60.9 885 6762
115
+ spa-nor tatoeba-test-v2021-08-07 0.75815 60.1 960 7217
116
+ spa-swe tatoeba-test-v2021-08-07 0.74222 60.7 1351 8357
benchmark_translations.zip ADDED
File without changes
config.json ADDED
@@ -0,0 +1,41 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "pytorch-models/opus-mt-tc-bible-big-deu_eng_fra_por_spa-gmq",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "relu",
5
+ "architectures": [
6
+ "MarianMTModel"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "bos_token_id": 0,
10
+ "classifier_dropout": 0.0,
11
+ "d_model": 1024,
12
+ "decoder_attention_heads": 16,
13
+ "decoder_ffn_dim": 4096,
14
+ "decoder_layerdrop": 0.0,
15
+ "decoder_layers": 6,
16
+ "decoder_start_token_id": 56247,
17
+ "decoder_vocab_size": 56248,
18
+ "dropout": 0.1,
19
+ "encoder_attention_heads": 16,
20
+ "encoder_ffn_dim": 4096,
21
+ "encoder_layerdrop": 0.0,
22
+ "encoder_layers": 6,
23
+ "eos_token_id": 571,
24
+ "forced_eos_token_id": null,
25
+ "init_std": 0.02,
26
+ "is_encoder_decoder": true,
27
+ "max_length": null,
28
+ "max_position_embeddings": 1024,
29
+ "model_type": "marian",
30
+ "normalize_embedding": false,
31
+ "num_beams": null,
32
+ "num_hidden_layers": 6,
33
+ "pad_token_id": 56247,
34
+ "scale_embedding": true,
35
+ "share_encoder_decoder_embeddings": true,
36
+ "static_position_embeddings": true,
37
+ "torch_dtype": "float32",
38
+ "transformers_version": "4.45.1",
39
+ "use_cache": true,
40
+ "vocab_size": 56248
41
+ }
generation_config.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bad_words_ids": [
4
+ [
5
+ 56247
6
+ ]
7
+ ],
8
+ "bos_token_id": 0,
9
+ "decoder_start_token_id": 56247,
10
+ "eos_token_id": 571,
11
+ "forced_eos_token_id": 571,
12
+ "max_length": 512,
13
+ "num_beams": 4,
14
+ "pad_token_id": 56247,
15
+ "transformers_version": "4.45.1"
16
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5f96bbcbe992b58f14ba667fe70f05f07af8e541f03ce44a17b43114196720c2
3
+ size 936075920
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8cebf84ce953b4841cf54ad2acae8310579885bf9dd1d67c8dc69ac788ea7623
3
+ size 936127173
source.spm ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b2a773a7f228ff70b18acd0d0dac8e53f5e67bc5c569fb6c652fc30d67d1f43a
3
+ size 809881
special_tokens_map.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"eos_token": "</s>", "unk_token": "<unk>", "pad_token": "<pad>"}
target.spm ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1e768e4c72bcf289a269b2c57c222518b508cc6589c33dedc8f94735e3f7b67d
3
+ size 803013
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"source_lang": "deu+eng+fra+por+spa", "target_lang": "gmq", "unk_token": "<unk>", "eos_token": "</s>", "pad_token": "<pad>", "model_max_length": 512, "sp_model_kwargs": {}, "separate_vocabs": false, "special_tokens_map_file": null, "name_or_path": "marian-models/opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-30/deu+eng+fra+por+spa-gmq", "tokenizer_class": "MarianTokenizer"}
vocab.json ADDED
The diff for this file is too large to render. See raw diff