tiedeman commited on
Commit
70c4613
1 Parent(s): 80435d3

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 +1,2529 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: transformers
3
+ language:
4
+ - anp
5
+ - as
6
+ - awa
7
+ - bho
8
+ - bn
9
+ - bpy
10
+ - de
11
+ - dv
12
+ - en
13
+ - es
14
+ - fr
15
+ - gbm
16
+ - gu
17
+ - hi
18
+ - hif
19
+ - hne
20
+ - hns
21
+ - kok
22
+ - ks
23
+ - lah
24
+ - mag
25
+ - mai
26
+ - mr
27
+ - ne
28
+ - or
29
+ - pa
30
+ - pi
31
+ - pt
32
+ - rhg
33
+ - rmy
34
+ - rom
35
+ - sa
36
+ - sd
37
+ - si
38
+ - skr
39
+ - syl
40
+ - ur
41
+
42
+ tags:
43
+ - translation
44
+ - opus-mt-tc-bible
45
+
46
+ license: apache-2.0
47
+ model-index:
48
+ - name: opus-mt-tc-bible-big-inc-deu_eng_fra_por_spa
49
+ results:
50
+ - task:
51
+ name: Translation asm-eng
52
+ type: translation
53
+ args: asm-eng
54
+ dataset:
55
+ name: flores200-devtest
56
+ type: flores200-devtest
57
+ args: asm-eng
58
+ metrics:
59
+ - name: BLEU
60
+ type: bleu
61
+ value: 21.9
62
+ - name: chr-F
63
+ type: chrf
64
+ value: 0.48584
65
+ - task:
66
+ name: Translation asm-por
67
+ type: translation
68
+ args: asm-por
69
+ dataset:
70
+ name: flores200-devtest
71
+ type: flores200-devtest
72
+ args: asm-por
73
+ metrics:
74
+ - name: BLEU
75
+ type: bleu
76
+ value: 10.1
77
+ - name: chr-F
78
+ type: chrf
79
+ value: 0.35028
80
+ - task:
81
+ name: Translation awa-deu
82
+ type: translation
83
+ args: awa-deu
84
+ dataset:
85
+ name: flores200-devtest
86
+ type: flores200-devtest
87
+ args: awa-deu
88
+ metrics:
89
+ - name: BLEU
90
+ type: bleu
91
+ value: 16.5
92
+ - name: chr-F
93
+ type: chrf
94
+ value: 0.47173
95
+ - task:
96
+ name: Translation awa-eng
97
+ type: translation
98
+ args: awa-eng
99
+ dataset:
100
+ name: flores200-devtest
101
+ type: flores200-devtest
102
+ args: awa-eng
103
+ metrics:
104
+ - name: BLEU
105
+ type: bleu
106
+ value: 24.5
107
+ - name: chr-F
108
+ type: chrf
109
+ value: 0.50582
110
+ - task:
111
+ name: Translation awa-fra
112
+ type: translation
113
+ args: awa-fra
114
+ dataset:
115
+ name: flores200-devtest
116
+ type: flores200-devtest
117
+ args: awa-fra
118
+ metrics:
119
+ - name: BLEU
120
+ type: bleu
121
+ value: 21.4
122
+ - name: chr-F
123
+ type: chrf
124
+ value: 0.49682
125
+ - task:
126
+ name: Translation awa-por
127
+ type: translation
128
+ args: awa-por
129
+ dataset:
130
+ name: flores200-devtest
131
+ type: flores200-devtest
132
+ args: awa-por
133
+ metrics:
134
+ - name: BLEU
135
+ type: bleu
136
+ value: 21.5
137
+ - name: chr-F
138
+ type: chrf
139
+ value: 0.49663
140
+ - task:
141
+ name: Translation awa-spa
142
+ type: translation
143
+ args: awa-spa
144
+ dataset:
145
+ name: flores200-devtest
146
+ type: flores200-devtest
147
+ args: awa-spa
148
+ metrics:
149
+ - name: BLEU
150
+ type: bleu
151
+ value: 15.1
152
+ - name: chr-F
153
+ type: chrf
154
+ value: 0.43740
155
+ - task:
156
+ name: Translation ben-deu
157
+ type: translation
158
+ args: ben-deu
159
+ dataset:
160
+ name: flores200-devtest
161
+ type: flores200-devtest
162
+ args: ben-deu
163
+ metrics:
164
+ - name: BLEU
165
+ type: bleu
166
+ value: 16.6
167
+ - name: chr-F
168
+ type: chrf
169
+ value: 0.47330
170
+ - task:
171
+ name: Translation ben-eng
172
+ type: translation
173
+ args: ben-eng
174
+ dataset:
175
+ name: flores200-devtest
176
+ type: flores200-devtest
177
+ args: ben-eng
178
+ metrics:
179
+ - name: BLEU
180
+ type: bleu
181
+ value: 30.5
182
+ - name: chr-F
183
+ type: chrf
184
+ value: 0.58077
185
+ - task:
186
+ name: Translation ben-fra
187
+ type: translation
188
+ args: ben-fra
189
+ dataset:
190
+ name: flores200-devtest
191
+ type: flores200-devtest
192
+ args: ben-fra
193
+ metrics:
194
+ - name: BLEU
195
+ type: bleu
196
+ value: 22.6
197
+ - name: chr-F
198
+ type: chrf
199
+ value: 0.50884
200
+ - task:
201
+ name: Translation ben-por
202
+ type: translation
203
+ args: ben-por
204
+ dataset:
205
+ name: flores200-devtest
206
+ type: flores200-devtest
207
+ args: ben-por
208
+ metrics:
209
+ - name: BLEU
210
+ type: bleu
211
+ value: 21.4
212
+ - name: chr-F
213
+ type: chrf
214
+ value: 0.50054
215
+ - task:
216
+ name: Translation ben-spa
217
+ type: translation
218
+ args: ben-spa
219
+ dataset:
220
+ name: flores200-devtest
221
+ type: flores200-devtest
222
+ args: ben-spa
223
+ metrics:
224
+ - name: BLEU
225
+ type: bleu
226
+ value: 15.2
227
+ - name: chr-F
228
+ type: chrf
229
+ value: 0.44159
230
+ - task:
231
+ name: Translation bho-deu
232
+ type: translation
233
+ args: bho-deu
234
+ dataset:
235
+ name: flores200-devtest
236
+ type: flores200-devtest
237
+ args: bho-deu
238
+ metrics:
239
+ - name: BLEU
240
+ type: bleu
241
+ value: 12.6
242
+ - name: chr-F
243
+ type: chrf
244
+ value: 0.42660
245
+ - task:
246
+ name: Translation bho-eng
247
+ type: translation
248
+ args: bho-eng
249
+ dataset:
250
+ name: flores200-devtest
251
+ type: flores200-devtest
252
+ args: bho-eng
253
+ metrics:
254
+ - name: BLEU
255
+ type: bleu
256
+ value: 22.7
257
+ - name: chr-F
258
+ type: chrf
259
+ value: 0.50609
260
+ - task:
261
+ name: Translation bho-fra
262
+ type: translation
263
+ args: bho-fra
264
+ dataset:
265
+ name: flores200-devtest
266
+ type: flores200-devtest
267
+ args: bho-fra
268
+ metrics:
269
+ - name: BLEU
270
+ type: bleu
271
+ value: 16.8
272
+ - name: chr-F
273
+ type: chrf
274
+ value: 0.44889
275
+ - task:
276
+ name: Translation bho-por
277
+ type: translation
278
+ args: bho-por
279
+ dataset:
280
+ name: flores200-devtest
281
+ type: flores200-devtest
282
+ args: bho-por
283
+ metrics:
284
+ - name: BLEU
285
+ type: bleu
286
+ value: 16.9
287
+ - name: chr-F
288
+ type: chrf
289
+ value: 0.44582
290
+ - task:
291
+ name: Translation bho-spa
292
+ type: translation
293
+ args: bho-spa
294
+ dataset:
295
+ name: flores200-devtest
296
+ type: flores200-devtest
297
+ args: bho-spa
298
+ metrics:
299
+ - name: BLEU
300
+ type: bleu
301
+ value: 13.1
302
+ - name: chr-F
303
+ type: chrf
304
+ value: 0.40581
305
+ - task:
306
+ name: Translation guj-deu
307
+ type: translation
308
+ args: guj-deu
309
+ dataset:
310
+ name: flores200-devtest
311
+ type: flores200-devtest
312
+ args: guj-deu
313
+ metrics:
314
+ - name: BLEU
315
+ type: bleu
316
+ value: 16.8
317
+ - name: chr-F
318
+ type: chrf
319
+ value: 0.46665
320
+ - task:
321
+ name: Translation guj-eng
322
+ type: translation
323
+ args: guj-eng
324
+ dataset:
325
+ name: flores200-devtest
326
+ type: flores200-devtest
327
+ args: guj-eng
328
+ metrics:
329
+ - name: BLEU
330
+ type: bleu
331
+ value: 34.3
332
+ - name: chr-F
333
+ type: chrf
334
+ value: 0.61383
335
+ - task:
336
+ name: Translation guj-fra
337
+ type: translation
338
+ args: guj-fra
339
+ dataset:
340
+ name: flores200-devtest
341
+ type: flores200-devtest
342
+ args: guj-fra
343
+ metrics:
344
+ - name: BLEU
345
+ type: bleu
346
+ value: 22.3
347
+ - name: chr-F
348
+ type: chrf
349
+ value: 0.50410
350
+ - task:
351
+ name: Translation guj-por
352
+ type: translation
353
+ args: guj-por
354
+ dataset:
355
+ name: flores200-devtest
356
+ type: flores200-devtest
357
+ args: guj-por
358
+ metrics:
359
+ - name: BLEU
360
+ type: bleu
361
+ value: 21.3
362
+ - name: chr-F
363
+ type: chrf
364
+ value: 0.49257
365
+ - task:
366
+ name: Translation guj-spa
367
+ type: translation
368
+ args: guj-spa
369
+ dataset:
370
+ name: flores200-devtest
371
+ type: flores200-devtest
372
+ args: guj-spa
373
+ metrics:
374
+ - name: BLEU
375
+ type: bleu
376
+ value: 15.6
377
+ - name: chr-F
378
+ type: chrf
379
+ value: 0.44565
380
+ - task:
381
+ name: Translation hin-deu
382
+ type: translation
383
+ args: hin-deu
384
+ dataset:
385
+ name: flores200-devtest
386
+ type: flores200-devtest
387
+ args: hin-deu
388
+ metrics:
389
+ - name: BLEU
390
+ type: bleu
391
+ value: 20.4
392
+ - name: chr-F
393
+ type: chrf
394
+ value: 0.50226
395
+ - task:
396
+ name: Translation hin-eng
397
+ type: translation
398
+ args: hin-eng
399
+ dataset:
400
+ name: flores200-devtest
401
+ type: flores200-devtest
402
+ args: hin-eng
403
+ metrics:
404
+ - name: BLEU
405
+ type: bleu
406
+ value: 37.3
407
+ - name: chr-F
408
+ type: chrf
409
+ value: 0.63336
410
+ - task:
411
+ name: Translation hin-fra
412
+ type: translation
413
+ args: hin-fra
414
+ dataset:
415
+ name: flores200-devtest
416
+ type: flores200-devtest
417
+ args: hin-fra
418
+ metrics:
419
+ - name: BLEU
420
+ type: bleu
421
+ value: 25.9
422
+ - name: chr-F
423
+ type: chrf
424
+ value: 0.53701
425
+ - task:
426
+ name: Translation hin-por
427
+ type: translation
428
+ args: hin-por
429
+ dataset:
430
+ name: flores200-devtest
431
+ type: flores200-devtest
432
+ args: hin-por
433
+ metrics:
434
+ - name: BLEU
435
+ type: bleu
436
+ value: 25.5
437
+ - name: chr-F
438
+ type: chrf
439
+ value: 0.53448
440
+ - task:
441
+ name: Translation hin-spa
442
+ type: translation
443
+ args: hin-spa
444
+ dataset:
445
+ name: flores200-devtest
446
+ type: flores200-devtest
447
+ args: hin-spa
448
+ metrics:
449
+ - name: BLEU
450
+ type: bleu
451
+ value: 17.2
452
+ - name: chr-F
453
+ type: chrf
454
+ value: 0.46171
455
+ - task:
456
+ name: Translation hne-deu
457
+ type: translation
458
+ args: hne-deu
459
+ dataset:
460
+ name: flores200-devtest
461
+ type: flores200-devtest
462
+ args: hne-deu
463
+ metrics:
464
+ - name: BLEU
465
+ type: bleu
466
+ value: 19.0
467
+ - name: chr-F
468
+ type: chrf
469
+ value: 0.49698
470
+ - task:
471
+ name: Translation hne-eng
472
+ type: translation
473
+ args: hne-eng
474
+ dataset:
475
+ name: flores200-devtest
476
+ type: flores200-devtest
477
+ args: hne-eng
478
+ metrics:
479
+ - name: BLEU
480
+ type: bleu
481
+ value: 38.5
482
+ - name: chr-F
483
+ type: chrf
484
+ value: 0.63936
485
+ - task:
486
+ name: Translation hne-fra
487
+ type: translation
488
+ args: hne-fra
489
+ dataset:
490
+ name: flores200-devtest
491
+ type: flores200-devtest
492
+ args: hne-fra
493
+ metrics:
494
+ - name: BLEU
495
+ type: bleu
496
+ value: 25.3
497
+ - name: chr-F
498
+ type: chrf
499
+ value: 0.52835
500
+ - task:
501
+ name: Translation hne-por
502
+ type: translation
503
+ args: hne-por
504
+ dataset:
505
+ name: flores200-devtest
506
+ type: flores200-devtest
507
+ args: hne-por
508
+ metrics:
509
+ - name: BLEU
510
+ type: bleu
511
+ value: 25.0
512
+ - name: chr-F
513
+ type: chrf
514
+ value: 0.52788
515
+ - task:
516
+ name: Translation hne-spa
517
+ type: translation
518
+ args: hne-spa
519
+ dataset:
520
+ name: flores200-devtest
521
+ type: flores200-devtest
522
+ args: hne-spa
523
+ metrics:
524
+ - name: BLEU
525
+ type: bleu
526
+ value: 16.7
527
+ - name: chr-F
528
+ type: chrf
529
+ value: 0.45443
530
+ - task:
531
+ name: Translation mag-deu
532
+ type: translation
533
+ args: mag-deu
534
+ dataset:
535
+ name: flores200-devtest
536
+ type: flores200-devtest
537
+ args: mag-deu
538
+ metrics:
539
+ - name: BLEU
540
+ type: bleu
541
+ value: 19.7
542
+ - name: chr-F
543
+ type: chrf
544
+ value: 0.50359
545
+ - task:
546
+ name: Translation mag-eng
547
+ type: translation
548
+ args: mag-eng
549
+ dataset:
550
+ name: flores200-devtest
551
+ type: flores200-devtest
552
+ args: mag-eng
553
+ metrics:
554
+ - name: BLEU
555
+ type: bleu
556
+ value: 38.0
557
+ - name: chr-F
558
+ type: chrf
559
+ value: 0.63906
560
+ - task:
561
+ name: Translation mag-fra
562
+ type: translation
563
+ args: mag-fra
564
+ dataset:
565
+ name: flores200-devtest
566
+ type: flores200-devtest
567
+ args: mag-fra
568
+ metrics:
569
+ - name: BLEU
570
+ type: bleu
571
+ value: 25.8
572
+ - name: chr-F
573
+ type: chrf
574
+ value: 0.53616
575
+ - task:
576
+ name: Translation mag-por
577
+ type: translation
578
+ args: mag-por
579
+ dataset:
580
+ name: flores200-devtest
581
+ type: flores200-devtest
582
+ args: mag-por
583
+ metrics:
584
+ - name: BLEU
585
+ type: bleu
586
+ value: 25.9
587
+ - name: chr-F
588
+ type: chrf
589
+ value: 0.53537
590
+ - task:
591
+ name: Translation mag-spa
592
+ type: translation
593
+ args: mag-spa
594
+ dataset:
595
+ name: flores200-devtest
596
+ type: flores200-devtest
597
+ args: mag-spa
598
+ metrics:
599
+ - name: BLEU
600
+ type: bleu
601
+ value: 16.9
602
+ - name: chr-F
603
+ type: chrf
604
+ value: 0.45822
605
+ - task:
606
+ name: Translation mai-deu
607
+ type: translation
608
+ args: mai-deu
609
+ dataset:
610
+ name: flores200-devtest
611
+ type: flores200-devtest
612
+ args: mai-deu
613
+ metrics:
614
+ - name: BLEU
615
+ type: bleu
616
+ value: 16.2
617
+ - name: chr-F
618
+ type: chrf
619
+ value: 0.46791
620
+ - task:
621
+ name: Translation mai-eng
622
+ type: translation
623
+ args: mai-eng
624
+ dataset:
625
+ name: flores200-devtest
626
+ type: flores200-devtest
627
+ args: mai-eng
628
+ metrics:
629
+ - name: BLEU
630
+ type: bleu
631
+ value: 30.4
632
+ - name: chr-F
633
+ type: chrf
634
+ value: 0.57461
635
+ - task:
636
+ name: Translation mai-fra
637
+ type: translation
638
+ args: mai-fra
639
+ dataset:
640
+ name: flores200-devtest
641
+ type: flores200-devtest
642
+ args: mai-fra
643
+ metrics:
644
+ - name: BLEU
645
+ type: bleu
646
+ value: 22.1
647
+ - name: chr-F
648
+ type: chrf
649
+ value: 0.50585
650
+ - task:
651
+ name: Translation mai-por
652
+ type: translation
653
+ args: mai-por
654
+ dataset:
655
+ name: flores200-devtest
656
+ type: flores200-devtest
657
+ args: mai-por
658
+ metrics:
659
+ - name: BLEU
660
+ type: bleu
661
+ value: 22.0
662
+ - name: chr-F
663
+ type: chrf
664
+ value: 0.50490
665
+ - task:
666
+ name: Translation mai-spa
667
+ type: translation
668
+ args: mai-spa
669
+ dataset:
670
+ name: flores200-devtest
671
+ type: flores200-devtest
672
+ args: mai-spa
673
+ metrics:
674
+ - name: BLEU
675
+ type: bleu
676
+ value: 15.3
677
+ - name: chr-F
678
+ type: chrf
679
+ value: 0.44366
680
+ - task:
681
+ name: Translation mar-deu
682
+ type: translation
683
+ args: mar-deu
684
+ dataset:
685
+ name: flores200-devtest
686
+ type: flores200-devtest
687
+ args: mar-deu
688
+ metrics:
689
+ - name: BLEU
690
+ type: bleu
691
+ value: 14.5
692
+ - name: chr-F
693
+ type: chrf
694
+ value: 0.44725
695
+ - task:
696
+ name: Translation mar-eng
697
+ type: translation
698
+ args: mar-eng
699
+ dataset:
700
+ name: flores200-devtest
701
+ type: flores200-devtest
702
+ args: mar-eng
703
+ metrics:
704
+ - name: BLEU
705
+ type: bleu
706
+ value: 31.4
707
+ - name: chr-F
708
+ type: chrf
709
+ value: 0.58500
710
+ - task:
711
+ name: Translation mar-fra
712
+ type: translation
713
+ args: mar-fra
714
+ dataset:
715
+ name: flores200-devtest
716
+ type: flores200-devtest
717
+ args: mar-fra
718
+ metrics:
719
+ - name: BLEU
720
+ type: bleu
721
+ value: 19.5
722
+ - name: chr-F
723
+ type: chrf
724
+ value: 0.47027
725
+ - task:
726
+ name: Translation mar-por
727
+ type: translation
728
+ args: mar-por
729
+ dataset:
730
+ name: flores200-devtest
731
+ type: flores200-devtest
732
+ args: mar-por
733
+ metrics:
734
+ - name: BLEU
735
+ type: bleu
736
+ value: 19.3
737
+ - name: chr-F
738
+ type: chrf
739
+ value: 0.47216
740
+ - task:
741
+ name: Translation mar-spa
742
+ type: translation
743
+ args: mar-spa
744
+ dataset:
745
+ name: flores200-devtest
746
+ type: flores200-devtest
747
+ args: mar-spa
748
+ metrics:
749
+ - name: BLEU
750
+ type: bleu
751
+ value: 14.2
752
+ - name: chr-F
753
+ type: chrf
754
+ value: 0.42178
755
+ - task:
756
+ name: Translation npi-deu
757
+ type: translation
758
+ args: npi-deu
759
+ dataset:
760
+ name: flores200-devtest
761
+ type: flores200-devtest
762
+ args: npi-deu
763
+ metrics:
764
+ - name: BLEU
765
+ type: bleu
766
+ value: 16.4
767
+ - name: chr-F
768
+ type: chrf
769
+ value: 0.46631
770
+ - task:
771
+ name: Translation npi-eng
772
+ type: translation
773
+ args: npi-eng
774
+ dataset:
775
+ name: flores200-devtest
776
+ type: flores200-devtest
777
+ args: npi-eng
778
+ metrics:
779
+ - name: BLEU
780
+ type: bleu
781
+ value: 32.3
782
+ - name: chr-F
783
+ type: chrf
784
+ value: 0.59776
785
+ - task:
786
+ name: Translation npi-fra
787
+ type: translation
788
+ args: npi-fra
789
+ dataset:
790
+ name: flores200-devtest
791
+ type: flores200-devtest
792
+ args: npi-fra
793
+ metrics:
794
+ - name: BLEU
795
+ type: bleu
796
+ value: 22.5
797
+ - name: chr-F
798
+ type: chrf
799
+ value: 0.50548
800
+ - task:
801
+ name: Translation npi-por
802
+ type: translation
803
+ args: npi-por
804
+ dataset:
805
+ name: flores200-devtest
806
+ type: flores200-devtest
807
+ args: npi-por
808
+ metrics:
809
+ - name: BLEU
810
+ type: bleu
811
+ value: 21.7
812
+ - name: chr-F
813
+ type: chrf
814
+ value: 0.50202
815
+ - task:
816
+ name: Translation npi-spa
817
+ type: translation
818
+ args: npi-spa
819
+ dataset:
820
+ name: flores200-devtest
821
+ type: flores200-devtest
822
+ args: npi-spa
823
+ metrics:
824
+ - name: BLEU
825
+ type: bleu
826
+ value: 15.3
827
+ - name: chr-F
828
+ type: chrf
829
+ value: 0.43804
830
+ - task:
831
+ name: Translation pan-deu
832
+ type: translation
833
+ args: pan-deu
834
+ dataset:
835
+ name: flores200-devtest
836
+ type: flores200-devtest
837
+ args: pan-deu
838
+ metrics:
839
+ - name: BLEU
840
+ type: bleu
841
+ value: 18.7
842
+ - name: chr-F
843
+ type: chrf
844
+ value: 0.48421
845
+ - task:
846
+ name: Translation pan-eng
847
+ type: translation
848
+ args: pan-eng
849
+ dataset:
850
+ name: flores200-devtest
851
+ type: flores200-devtest
852
+ args: pan-eng
853
+ metrics:
854
+ - name: BLEU
855
+ type: bleu
856
+ value: 33.8
857
+ - name: chr-F
858
+ type: chrf
859
+ value: 0.60676
860
+ - task:
861
+ name: Translation pan-fra
862
+ type: translation
863
+ args: pan-fra
864
+ dataset:
865
+ name: flores200-devtest
866
+ type: flores200-devtest
867
+ args: pan-fra
868
+ metrics:
869
+ - name: BLEU
870
+ type: bleu
871
+ value: 23.5
872
+ - name: chr-F
873
+ type: chrf
874
+ value: 0.51368
875
+ - task:
876
+ name: Translation pan-por
877
+ type: translation
878
+ args: pan-por
879
+ dataset:
880
+ name: flores200-devtest
881
+ type: flores200-devtest
882
+ args: pan-por
883
+ metrics:
884
+ - name: BLEU
885
+ type: bleu
886
+ value: 22.7
887
+ - name: chr-F
888
+ type: chrf
889
+ value: 0.50586
890
+ - task:
891
+ name: Translation pan-spa
892
+ type: translation
893
+ args: pan-spa
894
+ dataset:
895
+ name: flores200-devtest
896
+ type: flores200-devtest
897
+ args: pan-spa
898
+ metrics:
899
+ - name: BLEU
900
+ type: bleu
901
+ value: 16.5
902
+ - name: chr-F
903
+ type: chrf
904
+ value: 0.44653
905
+ - task:
906
+ name: Translation san-eng
907
+ type: translation
908
+ args: san-eng
909
+ dataset:
910
+ name: flores200-devtest
911
+ type: flores200-devtest
912
+ args: san-eng
913
+ metrics:
914
+ - name: BLEU
915
+ type: bleu
916
+ value: 11.8
917
+ - name: chr-F
918
+ type: chrf
919
+ value: 0.36887
920
+ - task:
921
+ name: Translation sin-deu
922
+ type: translation
923
+ args: sin-deu
924
+ dataset:
925
+ name: flores200-devtest
926
+ type: flores200-devtest
927
+ args: sin-deu
928
+ metrics:
929
+ - name: BLEU
930
+ type: bleu
931
+ value: 14.2
932
+ - name: chr-F
933
+ type: chrf
934
+ value: 0.44676
935
+ - task:
936
+ name: Translation sin-eng
937
+ type: translation
938
+ args: sin-eng
939
+ dataset:
940
+ name: flores200-devtest
941
+ type: flores200-devtest
942
+ args: sin-eng
943
+ metrics:
944
+ - name: BLEU
945
+ type: bleu
946
+ value: 26.8
947
+ - name: chr-F
948
+ type: chrf
949
+ value: 0.54777
950
+ - task:
951
+ name: Translation sin-fra
952
+ type: translation
953
+ args: sin-fra
954
+ dataset:
955
+ name: flores200-devtest
956
+ type: flores200-devtest
957
+ args: sin-fra
958
+ metrics:
959
+ - name: BLEU
960
+ type: bleu
961
+ value: 19.0
962
+ - name: chr-F
963
+ type: chrf
964
+ value: 0.47283
965
+ - task:
966
+ name: Translation sin-por
967
+ type: translation
968
+ args: sin-por
969
+ dataset:
970
+ name: flores200-devtest
971
+ type: flores200-devtest
972
+ args: sin-por
973
+ metrics:
974
+ - name: BLEU
975
+ type: bleu
976
+ value: 18.4
977
+ - name: chr-F
978
+ type: chrf
979
+ value: 0.46935
980
+ - task:
981
+ name: Translation sin-spa
982
+ type: translation
983
+ args: sin-spa
984
+ dataset:
985
+ name: flores200-devtest
986
+ type: flores200-devtest
987
+ args: sin-spa
988
+ metrics:
989
+ - name: BLEU
990
+ type: bleu
991
+ value: 13.7
992
+ - name: chr-F
993
+ type: chrf
994
+ value: 0.42143
995
+ - task:
996
+ name: Translation urd-deu
997
+ type: translation
998
+ args: urd-deu
999
+ dataset:
1000
+ name: flores200-devtest
1001
+ type: flores200-devtest
1002
+ args: urd-deu
1003
+ metrics:
1004
+ - name: BLEU
1005
+ type: bleu
1006
+ value: 17.1
1007
+ - name: chr-F
1008
+ type: chrf
1009
+ value: 0.46542
1010
+ - task:
1011
+ name: Translation urd-eng
1012
+ type: translation
1013
+ args: urd-eng
1014
+ dataset:
1015
+ name: flores200-devtest
1016
+ type: flores200-devtest
1017
+ args: urd-eng
1018
+ metrics:
1019
+ - name: BLEU
1020
+ type: bleu
1021
+ value: 29.3
1022
+ - name: chr-F
1023
+ type: chrf
1024
+ value: 0.56935
1025
+ - task:
1026
+ name: Translation urd-fra
1027
+ type: translation
1028
+ args: urd-fra
1029
+ dataset:
1030
+ name: flores200-devtest
1031
+ type: flores200-devtest
1032
+ args: urd-fra
1033
+ metrics:
1034
+ - name: BLEU
1035
+ type: bleu
1036
+ value: 22.3
1037
+ - name: chr-F
1038
+ type: chrf
1039
+ value: 0.50276
1040
+ - task:
1041
+ name: Translation urd-por
1042
+ type: translation
1043
+ args: urd-por
1044
+ dataset:
1045
+ name: flores200-devtest
1046
+ type: flores200-devtest
1047
+ args: urd-por
1048
+ metrics:
1049
+ - name: BLEU
1050
+ type: bleu
1051
+ value: 20.3
1052
+ - name: chr-F
1053
+ type: chrf
1054
+ value: 0.48010
1055
+ - task:
1056
+ name: Translation urd-spa
1057
+ type: translation
1058
+ args: urd-spa
1059
+ dataset:
1060
+ name: flores200-devtest
1061
+ type: flores200-devtest
1062
+ args: urd-spa
1063
+ metrics:
1064
+ - name: BLEU
1065
+ type: bleu
1066
+ value: 14.7
1067
+ - name: chr-F
1068
+ type: chrf
1069
+ value: 0.43032
1070
+ - task:
1071
+ name: Translation asm-por
1072
+ type: translation
1073
+ args: asm-por
1074
+ dataset:
1075
+ name: flores101-devtest
1076
+ type: flores_101
1077
+ args: asm por devtest
1078
+ metrics:
1079
+ - name: BLEU
1080
+ type: bleu
1081
+ value: 10.0
1082
+ - name: chr-F
1083
+ type: chrf
1084
+ value: 0.34689
1085
+ - task:
1086
+ name: Translation ben-eng
1087
+ type: translation
1088
+ args: ben-eng
1089
+ dataset:
1090
+ name: flores101-devtest
1091
+ type: flores_101
1092
+ args: ben eng devtest
1093
+ metrics:
1094
+ - name: BLEU
1095
+ type: bleu
1096
+ value: 30.4
1097
+ - name: chr-F
1098
+ type: chrf
1099
+ value: 0.57906
1100
+ - task:
1101
+ name: Translation ben-fra
1102
+ type: translation
1103
+ args: ben-fra
1104
+ dataset:
1105
+ name: flores101-devtest
1106
+ type: flores_101
1107
+ args: ben fra devtest
1108
+ metrics:
1109
+ - name: BLEU
1110
+ type: bleu
1111
+ value: 21.9
1112
+ - name: chr-F
1113
+ type: chrf
1114
+ value: 0.50109
1115
+ - task:
1116
+ name: Translation guj-spa
1117
+ type: translation
1118
+ args: guj-spa
1119
+ dataset:
1120
+ name: flores101-devtest
1121
+ type: flores_101
1122
+ args: guj spa devtest
1123
+ metrics:
1124
+ - name: BLEU
1125
+ type: bleu
1126
+ value: 15.2
1127
+ - name: chr-F
1128
+ type: chrf
1129
+ value: 0.44065
1130
+ - task:
1131
+ name: Translation mar-deu
1132
+ type: translation
1133
+ args: mar-deu
1134
+ dataset:
1135
+ name: flores101-devtest
1136
+ type: flores_101
1137
+ args: mar deu devtest
1138
+ metrics:
1139
+ - name: BLEU
1140
+ type: bleu
1141
+ value: 13.8
1142
+ - name: chr-F
1143
+ type: chrf
1144
+ value: 0.44067
1145
+ - task:
1146
+ name: Translation mar-por
1147
+ type: translation
1148
+ args: mar-por
1149
+ dataset:
1150
+ name: flores101-devtest
1151
+ type: flores_101
1152
+ args: mar por devtest
1153
+ metrics:
1154
+ - name: BLEU
1155
+ type: bleu
1156
+ value: 18.6
1157
+ - name: chr-F
1158
+ type: chrf
1159
+ value: 0.46685
1160
+ - task:
1161
+ name: Translation mar-spa
1162
+ type: translation
1163
+ args: mar-spa
1164
+ dataset:
1165
+ name: flores101-devtest
1166
+ type: flores_101
1167
+ args: mar spa devtest
1168
+ metrics:
1169
+ - name: BLEU
1170
+ type: bleu
1171
+ value: 14.0
1172
+ - name: chr-F
1173
+ type: chrf
1174
+ value: 0.41662
1175
+ - task:
1176
+ name: Translation pan-eng
1177
+ type: translation
1178
+ args: pan-eng
1179
+ dataset:
1180
+ name: flores101-devtest
1181
+ type: flores_101
1182
+ args: pan eng devtest
1183
+ metrics:
1184
+ - name: BLEU
1185
+ type: bleu
1186
+ value: 33.0
1187
+ - name: chr-F
1188
+ type: chrf
1189
+ value: 0.59922
1190
+ - task:
1191
+ name: Translation pan-por
1192
+ type: translation
1193
+ args: pan-por
1194
+ dataset:
1195
+ name: flores101-devtest
1196
+ type: flores_101
1197
+ args: pan por devtest
1198
+ metrics:
1199
+ - name: BLEU
1200
+ type: bleu
1201
+ value: 21.9
1202
+ - name: chr-F
1203
+ type: chrf
1204
+ value: 0.49373
1205
+ - task:
1206
+ name: Translation pan-spa
1207
+ type: translation
1208
+ args: pan-spa
1209
+ dataset:
1210
+ name: flores101-devtest
1211
+ type: flores_101
1212
+ args: pan spa devtest
1213
+ metrics:
1214
+ - name: BLEU
1215
+ type: bleu
1216
+ value: 15.4
1217
+ - name: chr-F
1218
+ type: chrf
1219
+ value: 0.43910
1220
+ - task:
1221
+ name: Translation ben-deu
1222
+ type: translation
1223
+ args: ben-deu
1224
+ dataset:
1225
+ name: ntrex128
1226
+ type: ntrex128
1227
+ args: ben-deu
1228
+ metrics:
1229
+ - name: BLEU
1230
+ type: bleu
1231
+ value: 14.6
1232
+ - name: chr-F
1233
+ type: chrf
1234
+ value: 0.45180
1235
+ - task:
1236
+ name: Translation ben-eng
1237
+ type: translation
1238
+ args: ben-eng
1239
+ dataset:
1240
+ name: ntrex128
1241
+ type: ntrex128
1242
+ args: ben-eng
1243
+ metrics:
1244
+ - name: BLEU
1245
+ type: bleu
1246
+ value: 29.5
1247
+ - name: chr-F
1248
+ type: chrf
1249
+ value: 0.57247
1250
+ - task:
1251
+ name: Translation ben-fra
1252
+ type: translation
1253
+ args: ben-fra
1254
+ dataset:
1255
+ name: ntrex128
1256
+ type: ntrex128
1257
+ args: ben-fra
1258
+ metrics:
1259
+ - name: BLEU
1260
+ type: bleu
1261
+ value: 18.0
1262
+ - name: chr-F
1263
+ type: chrf
1264
+ value: 0.46475
1265
+ - task:
1266
+ name: Translation ben-por
1267
+ type: translation
1268
+ args: ben-por
1269
+ dataset:
1270
+ name: ntrex128
1271
+ type: ntrex128
1272
+ args: ben-por
1273
+ metrics:
1274
+ - name: BLEU
1275
+ type: bleu
1276
+ value: 16.8
1277
+ - name: chr-F
1278
+ type: chrf
1279
+ value: 0.45486
1280
+ - task:
1281
+ name: Translation ben-spa
1282
+ type: translation
1283
+ args: ben-spa
1284
+ dataset:
1285
+ name: ntrex128
1286
+ type: ntrex128
1287
+ args: ben-spa
1288
+ metrics:
1289
+ - name: BLEU
1290
+ type: bleu
1291
+ value: 21.1
1292
+ - name: chr-F
1293
+ type: chrf
1294
+ value: 0.48738
1295
+ - task:
1296
+ name: Translation guj-deu
1297
+ type: translation
1298
+ args: guj-deu
1299
+ dataset:
1300
+ name: ntrex128
1301
+ type: ntrex128
1302
+ args: guj-deu
1303
+ metrics:
1304
+ - name: BLEU
1305
+ type: bleu
1306
+ value: 13.9
1307
+ - name: chr-F
1308
+ type: chrf
1309
+ value: 0.43539
1310
+ - task:
1311
+ name: Translation guj-eng
1312
+ type: translation
1313
+ args: guj-eng
1314
+ dataset:
1315
+ name: ntrex128
1316
+ type: ntrex128
1317
+ args: guj-eng
1318
+ metrics:
1319
+ - name: BLEU
1320
+ type: bleu
1321
+ value: 31.6
1322
+ - name: chr-F
1323
+ type: chrf
1324
+ value: 0.58894
1325
+ - task:
1326
+ name: Translation guj-fra
1327
+ type: translation
1328
+ args: guj-fra
1329
+ dataset:
1330
+ name: ntrex128
1331
+ type: ntrex128
1332
+ args: guj-fra
1333
+ metrics:
1334
+ - name: BLEU
1335
+ type: bleu
1336
+ value: 16.9
1337
+ - name: chr-F
1338
+ type: chrf
1339
+ value: 0.45075
1340
+ - task:
1341
+ name: Translation guj-por
1342
+ type: translation
1343
+ args: guj-por
1344
+ dataset:
1345
+ name: ntrex128
1346
+ type: ntrex128
1347
+ args: guj-por
1348
+ metrics:
1349
+ - name: BLEU
1350
+ type: bleu
1351
+ value: 15.2
1352
+ - name: chr-F
1353
+ type: chrf
1354
+ value: 0.43567
1355
+ - task:
1356
+ name: Translation guj-spa
1357
+ type: translation
1358
+ args: guj-spa
1359
+ dataset:
1360
+ name: ntrex128
1361
+ type: ntrex128
1362
+ args: guj-spa
1363
+ metrics:
1364
+ - name: BLEU
1365
+ type: bleu
1366
+ value: 20.2
1367
+ - name: chr-F
1368
+ type: chrf
1369
+ value: 0.47525
1370
+ - task:
1371
+ name: Translation hin-deu
1372
+ type: translation
1373
+ args: hin-deu
1374
+ dataset:
1375
+ name: ntrex128
1376
+ type: ntrex128
1377
+ args: hin-deu
1378
+ metrics:
1379
+ - name: BLEU
1380
+ type: bleu
1381
+ value: 15.0
1382
+ - name: chr-F
1383
+ type: chrf
1384
+ value: 0.46336
1385
+ - task:
1386
+ name: Translation hin-eng
1387
+ type: translation
1388
+ args: hin-eng
1389
+ dataset:
1390
+ name: ntrex128
1391
+ type: ntrex128
1392
+ args: hin-eng
1393
+ metrics:
1394
+ - name: BLEU
1395
+ type: bleu
1396
+ value: 31.5
1397
+ - name: chr-F
1398
+ type: chrf
1399
+ value: 0.59842
1400
+ - task:
1401
+ name: Translation hin-fra
1402
+ type: translation
1403
+ args: hin-fra
1404
+ dataset:
1405
+ name: ntrex128
1406
+ type: ntrex128
1407
+ args: hin-fra
1408
+ metrics:
1409
+ - name: BLEU
1410
+ type: bleu
1411
+ value: 19.2
1412
+ - name: chr-F
1413
+ type: chrf
1414
+ value: 0.48208
1415
+ - task:
1416
+ name: Translation hin-por
1417
+ type: translation
1418
+ args: hin-por
1419
+ dataset:
1420
+ name: ntrex128
1421
+ type: ntrex128
1422
+ args: hin-por
1423
+ metrics:
1424
+ - name: BLEU
1425
+ type: bleu
1426
+ value: 17.6
1427
+ - name: chr-F
1428
+ type: chrf
1429
+ value: 0.46509
1430
+ - task:
1431
+ name: Translation hin-spa
1432
+ type: translation
1433
+ args: hin-spa
1434
+ dataset:
1435
+ name: ntrex128
1436
+ type: ntrex128
1437
+ args: hin-spa
1438
+ metrics:
1439
+ - name: BLEU
1440
+ type: bleu
1441
+ value: 21.8
1442
+ - name: chr-F
1443
+ type: chrf
1444
+ value: 0.49436
1445
+ - task:
1446
+ name: Translation mar-deu
1447
+ type: translation
1448
+ args: mar-deu
1449
+ dataset:
1450
+ name: ntrex128
1451
+ type: ntrex128
1452
+ args: mar-deu
1453
+ metrics:
1454
+ - name: BLEU
1455
+ type: bleu
1456
+ value: 12.8
1457
+ - name: chr-F
1458
+ type: chrf
1459
+ value: 0.43119
1460
+ - task:
1461
+ name: Translation mar-eng
1462
+ type: translation
1463
+ args: mar-eng
1464
+ dataset:
1465
+ name: ntrex128
1466
+ type: ntrex128
1467
+ args: mar-eng
1468
+ metrics:
1469
+ - name: BLEU
1470
+ type: bleu
1471
+ value: 27.3
1472
+ - name: chr-F
1473
+ type: chrf
1474
+ value: 0.55151
1475
+ - task:
1476
+ name: Translation mar-fra
1477
+ type: translation
1478
+ args: mar-fra
1479
+ dataset:
1480
+ name: ntrex128
1481
+ type: ntrex128
1482
+ args: mar-fra
1483
+ metrics:
1484
+ - name: BLEU
1485
+ type: bleu
1486
+ value: 16.2
1487
+ - name: chr-F
1488
+ type: chrf
1489
+ value: 0.43957
1490
+ - task:
1491
+ name: Translation mar-por
1492
+ type: translation
1493
+ args: mar-por
1494
+ dataset:
1495
+ name: ntrex128
1496
+ type: ntrex128
1497
+ args: mar-por
1498
+ metrics:
1499
+ - name: BLEU
1500
+ type: bleu
1501
+ value: 15.4
1502
+ - name: chr-F
1503
+ type: chrf
1504
+ value: 0.43555
1505
+ - task:
1506
+ name: Translation mar-spa
1507
+ type: translation
1508
+ args: mar-spa
1509
+ dataset:
1510
+ name: ntrex128
1511
+ type: ntrex128
1512
+ args: mar-spa
1513
+ metrics:
1514
+ - name: BLEU
1515
+ type: bleu
1516
+ value: 19.1
1517
+ - name: chr-F
1518
+ type: chrf
1519
+ value: 0.46271
1520
+ - task:
1521
+ name: Translation nep-deu
1522
+ type: translation
1523
+ args: nep-deu
1524
+ dataset:
1525
+ name: ntrex128
1526
+ type: ntrex128
1527
+ args: nep-deu
1528
+ metrics:
1529
+ - name: BLEU
1530
+ type: bleu
1531
+ value: 13.0
1532
+ - name: chr-F
1533
+ type: chrf
1534
+ value: 0.42940
1535
+ - task:
1536
+ name: Translation nep-eng
1537
+ type: translation
1538
+ args: nep-eng
1539
+ dataset:
1540
+ name: ntrex128
1541
+ type: ntrex128
1542
+ args: nep-eng
1543
+ metrics:
1544
+ - name: BLEU
1545
+ type: bleu
1546
+ value: 29.1
1547
+ - name: chr-F
1548
+ type: chrf
1549
+ value: 0.56277
1550
+ - task:
1551
+ name: Translation nep-fra
1552
+ type: translation
1553
+ args: nep-fra
1554
+ dataset:
1555
+ name: ntrex128
1556
+ type: ntrex128
1557
+ args: nep-fra
1558
+ metrics:
1559
+ - name: BLEU
1560
+ type: bleu
1561
+ value: 16.5
1562
+ - name: chr-F
1563
+ type: chrf
1564
+ value: 0.44663
1565
+ - task:
1566
+ name: Translation nep-por
1567
+ type: translation
1568
+ args: nep-por
1569
+ dataset:
1570
+ name: ntrex128
1571
+ type: ntrex128
1572
+ args: nep-por
1573
+ metrics:
1574
+ - name: BLEU
1575
+ type: bleu
1576
+ value: 15.4
1577
+ - name: chr-F
1578
+ type: chrf
1579
+ value: 0.43686
1580
+ - task:
1581
+ name: Translation nep-spa
1582
+ type: translation
1583
+ args: nep-spa
1584
+ dataset:
1585
+ name: ntrex128
1586
+ type: ntrex128
1587
+ args: nep-spa
1588
+ metrics:
1589
+ - name: BLEU
1590
+ type: bleu
1591
+ value: 19.3
1592
+ - name: chr-F
1593
+ type: chrf
1594
+ value: 0.46553
1595
+ - task:
1596
+ name: Translation pan-deu
1597
+ type: translation
1598
+ args: pan-deu
1599
+ dataset:
1600
+ name: ntrex128
1601
+ type: ntrex128
1602
+ args: pan-deu
1603
+ metrics:
1604
+ - name: BLEU
1605
+ type: bleu
1606
+ value: 14.1
1607
+ - name: chr-F
1608
+ type: chrf
1609
+ value: 0.44036
1610
+ - task:
1611
+ name: Translation pan-eng
1612
+ type: translation
1613
+ args: pan-eng
1614
+ dataset:
1615
+ name: ntrex128
1616
+ type: ntrex128
1617
+ args: pan-eng
1618
+ metrics:
1619
+ - name: BLEU
1620
+ type: bleu
1621
+ value: 31.6
1622
+ - name: chr-F
1623
+ type: chrf
1624
+ value: 0.58427
1625
+ - task:
1626
+ name: Translation pan-fra
1627
+ type: translation
1628
+ args: pan-fra
1629
+ dataset:
1630
+ name: ntrex128
1631
+ type: ntrex128
1632
+ args: pan-fra
1633
+ metrics:
1634
+ - name: BLEU
1635
+ type: bleu
1636
+ value: 17.3
1637
+ - name: chr-F
1638
+ type: chrf
1639
+ value: 0.45593
1640
+ - task:
1641
+ name: Translation pan-por
1642
+ type: translation
1643
+ args: pan-por
1644
+ dataset:
1645
+ name: ntrex128
1646
+ type: ntrex128
1647
+ args: pan-por
1648
+ metrics:
1649
+ - name: BLEU
1650
+ type: bleu
1651
+ value: 15.9
1652
+ - name: chr-F
1653
+ type: chrf
1654
+ value: 0.44264
1655
+ - task:
1656
+ name: Translation pan-spa
1657
+ type: translation
1658
+ args: pan-spa
1659
+ dataset:
1660
+ name: ntrex128
1661
+ type: ntrex128
1662
+ args: pan-spa
1663
+ metrics:
1664
+ - name: BLEU
1665
+ type: bleu
1666
+ value: 20.0
1667
+ - name: chr-F
1668
+ type: chrf
1669
+ value: 0.47199
1670
+ - task:
1671
+ name: Translation sin-deu
1672
+ type: translation
1673
+ args: sin-deu
1674
+ dataset:
1675
+ name: ntrex128
1676
+ type: ntrex128
1677
+ args: sin-deu
1678
+ metrics:
1679
+ - name: BLEU
1680
+ type: bleu
1681
+ value: 12.4
1682
+ - name: chr-F
1683
+ type: chrf
1684
+ value: 0.42280
1685
+ - task:
1686
+ name: Translation sin-eng
1687
+ type: translation
1688
+ args: sin-eng
1689
+ dataset:
1690
+ name: ntrex128
1691
+ type: ntrex128
1692
+ args: sin-eng
1693
+ metrics:
1694
+ - name: BLEU
1695
+ type: bleu
1696
+ value: 24.6
1697
+ - name: chr-F
1698
+ type: chrf
1699
+ value: 0.52576
1700
+ - task:
1701
+ name: Translation sin-fra
1702
+ type: translation
1703
+ args: sin-fra
1704
+ dataset:
1705
+ name: ntrex128
1706
+ type: ntrex128
1707
+ args: sin-fra
1708
+ metrics:
1709
+ - name: BLEU
1710
+ type: bleu
1711
+ value: 15.6
1712
+ - name: chr-F
1713
+ type: chrf
1714
+ value: 0.43594
1715
+ - task:
1716
+ name: Translation sin-por
1717
+ type: translation
1718
+ args: sin-por
1719
+ dataset:
1720
+ name: ntrex128
1721
+ type: ntrex128
1722
+ args: sin-por
1723
+ metrics:
1724
+ - name: BLEU
1725
+ type: bleu
1726
+ value: 14.4
1727
+ - name: chr-F
1728
+ type: chrf
1729
+ value: 0.42751
1730
+ - task:
1731
+ name: Translation sin-spa
1732
+ type: translation
1733
+ args: sin-spa
1734
+ dataset:
1735
+ name: ntrex128
1736
+ type: ntrex128
1737
+ args: sin-spa
1738
+ metrics:
1739
+ - name: BLEU
1740
+ type: bleu
1741
+ value: 18.3
1742
+ - name: chr-F
1743
+ type: chrf
1744
+ value: 0.45890
1745
+ - task:
1746
+ name: Translation urd-deu
1747
+ type: translation
1748
+ args: urd-deu
1749
+ dataset:
1750
+ name: ntrex128
1751
+ type: ntrex128
1752
+ args: urd-deu
1753
+ metrics:
1754
+ - name: BLEU
1755
+ type: bleu
1756
+ value: 15.6
1757
+ - name: chr-F
1758
+ type: chrf
1759
+ value: 0.45737
1760
+ - task:
1761
+ name: Translation urd-eng
1762
+ type: translation
1763
+ args: urd-eng
1764
+ dataset:
1765
+ name: ntrex128
1766
+ type: ntrex128
1767
+ args: urd-eng
1768
+ metrics:
1769
+ - name: BLEU
1770
+ type: bleu
1771
+ value: 28.6
1772
+ - name: chr-F
1773
+ type: chrf
1774
+ value: 0.56781
1775
+ - task:
1776
+ name: Translation urd-fra
1777
+ type: translation
1778
+ args: urd-fra
1779
+ dataset:
1780
+ name: ntrex128
1781
+ type: ntrex128
1782
+ args: urd-fra
1783
+ metrics:
1784
+ - name: BLEU
1785
+ type: bleu
1786
+ value: 18.9
1787
+ - name: chr-F
1788
+ type: chrf
1789
+ value: 0.47298
1790
+ - task:
1791
+ name: Translation urd-por
1792
+ type: translation
1793
+ args: urd-por
1794
+ dataset:
1795
+ name: ntrex128
1796
+ type: ntrex128
1797
+ args: urd-por
1798
+ metrics:
1799
+ - name: BLEU
1800
+ type: bleu
1801
+ value: 16.2
1802
+ - name: chr-F
1803
+ type: chrf
1804
+ value: 0.45273
1805
+ - task:
1806
+ name: Translation urd-spa
1807
+ type: translation
1808
+ args: urd-spa
1809
+ dataset:
1810
+ name: ntrex128
1811
+ type: ntrex128
1812
+ args: urd-spa
1813
+ metrics:
1814
+ - name: BLEU
1815
+ type: bleu
1816
+ value: 21.0
1817
+ - name: chr-F
1818
+ type: chrf
1819
+ value: 0.48644
1820
+ - task:
1821
+ name: Translation awa-eng
1822
+ type: translation
1823
+ args: awa-eng
1824
+ dataset:
1825
+ name: tatoeba-test-v2021-08-07
1826
+ type: tatoeba_mt
1827
+ args: awa-eng
1828
+ metrics:
1829
+ - name: BLEU
1830
+ type: bleu
1831
+ value: 40.8
1832
+ - name: chr-F
1833
+ type: chrf
1834
+ value: 0.60390
1835
+ - task:
1836
+ name: Translation ben-eng
1837
+ type: translation
1838
+ args: ben-eng
1839
+ dataset:
1840
+ name: tatoeba-test-v2021-08-07
1841
+ type: tatoeba_mt
1842
+ args: ben-eng
1843
+ metrics:
1844
+ - name: BLEU
1845
+ type: bleu
1846
+ value: 49.4
1847
+ - name: chr-F
1848
+ type: chrf
1849
+ value: 0.64078
1850
+ - task:
1851
+ name: Translation hin-eng
1852
+ type: translation
1853
+ args: hin-eng
1854
+ dataset:
1855
+ name: tatoeba-test-v2021-08-07
1856
+ type: tatoeba_mt
1857
+ args: hin-eng
1858
+ metrics:
1859
+ - name: BLEU
1860
+ type: bleu
1861
+ value: 49.1
1862
+ - name: chr-F
1863
+ type: chrf
1864
+ value: 0.64929
1865
+ - task:
1866
+ name: Translation mar-eng
1867
+ type: translation
1868
+ args: mar-eng
1869
+ dataset:
1870
+ name: tatoeba-test-v2021-08-07
1871
+ type: tatoeba_mt
1872
+ args: mar-eng
1873
+ metrics:
1874
+ - name: BLEU
1875
+ type: bleu
1876
+ value: 48.0
1877
+ - name: chr-F
1878
+ type: chrf
1879
+ value: 0.64074
1880
+ - task:
1881
+ name: Translation multi-multi
1882
+ type: translation
1883
+ args: multi-multi
1884
+ dataset:
1885
+ name: tatoeba-test-v2020-07-28-v2023-09-26
1886
+ type: tatoeba_mt
1887
+ args: multi-multi
1888
+ metrics:
1889
+ - name: BLEU
1890
+ type: bleu
1891
+ value: 44.6
1892
+ - name: chr-F
1893
+ type: chrf
1894
+ value: 0.61360
1895
+ - task:
1896
+ name: Translation urd-eng
1897
+ type: translation
1898
+ args: urd-eng
1899
+ dataset:
1900
+ name: tatoeba-test-v2021-08-07
1901
+ type: tatoeba_mt
1902
+ args: urd-eng
1903
+ metrics:
1904
+ - name: BLEU
1905
+ type: bleu
1906
+ value: 35.0
1907
+ - name: chr-F
1908
+ type: chrf
1909
+ value: 0.52963
1910
+ - task:
1911
+ name: Translation ben-eng
1912
+ type: translation
1913
+ args: ben-eng
1914
+ dataset:
1915
+ name: tico19-test
1916
+ type: tico19-test
1917
+ args: ben-eng
1918
+ metrics:
1919
+ - name: BLEU
1920
+ type: bleu
1921
+ value: 38.2
1922
+ - name: chr-F
1923
+ type: chrf
1924
+ value: 0.64568
1925
+ - task:
1926
+ name: Translation ben-fra
1927
+ type: translation
1928
+ args: ben-fra
1929
+ dataset:
1930
+ name: tico19-test
1931
+ type: tico19-test
1932
+ args: ben-fra
1933
+ metrics:
1934
+ - name: BLEU
1935
+ type: bleu
1936
+ value: 22.0
1937
+ - name: chr-F
1938
+ type: chrf
1939
+ value: 0.49799
1940
+ - task:
1941
+ name: Translation ben-por
1942
+ type: translation
1943
+ args: ben-por
1944
+ dataset:
1945
+ name: tico19-test
1946
+ type: tico19-test
1947
+ args: ben-por
1948
+ metrics:
1949
+ - name: BLEU
1950
+ type: bleu
1951
+ value: 27.2
1952
+ - name: chr-F
1953
+ type: chrf
1954
+ value: 0.55115
1955
+ - task:
1956
+ name: Translation ben-spa
1957
+ type: translation
1958
+ args: ben-spa
1959
+ dataset:
1960
+ name: tico19-test
1961
+ type: tico19-test
1962
+ args: ben-spa
1963
+ metrics:
1964
+ - name: BLEU
1965
+ type: bleu
1966
+ value: 29.9
1967
+ - name: chr-F
1968
+ type: chrf
1969
+ value: 0.56847
1970
+ - task:
1971
+ name: Translation hin-eng
1972
+ type: translation
1973
+ args: hin-eng
1974
+ dataset:
1975
+ name: tico19-test
1976
+ type: tico19-test
1977
+ args: hin-eng
1978
+ metrics:
1979
+ - name: BLEU
1980
+ type: bleu
1981
+ value: 46.6
1982
+ - name: chr-F
1983
+ type: chrf
1984
+ value: 0.70694
1985
+ - task:
1986
+ name: Translation hin-fra
1987
+ type: translation
1988
+ args: hin-fra
1989
+ dataset:
1990
+ name: tico19-test
1991
+ type: tico19-test
1992
+ args: hin-fra
1993
+ metrics:
1994
+ - name: BLEU
1995
+ type: bleu
1996
+ value: 26.7
1997
+ - name: chr-F
1998
+ type: chrf
1999
+ value: 0.53932
2000
+ - task:
2001
+ name: Translation hin-por
2002
+ type: translation
2003
+ args: hin-por
2004
+ dataset:
2005
+ name: tico19-test
2006
+ type: tico19-test
2007
+ args: hin-por
2008
+ metrics:
2009
+ - name: BLEU
2010
+ type: bleu
2011
+ value: 33.4
2012
+ - name: chr-F
2013
+ type: chrf
2014
+ value: 0.60581
2015
+ - task:
2016
+ name: Translation hin-spa
2017
+ type: translation
2018
+ args: hin-spa
2019
+ dataset:
2020
+ name: tico19-test
2021
+ type: tico19-test
2022
+ args: hin-spa
2023
+ metrics:
2024
+ - name: BLEU
2025
+ type: bleu
2026
+ value: 35.7
2027
+ - name: chr-F
2028
+ type: chrf
2029
+ value: 0.61585
2030
+ - task:
2031
+ name: Translation mar-eng
2032
+ type: translation
2033
+ args: mar-eng
2034
+ dataset:
2035
+ name: tico19-test
2036
+ type: tico19-test
2037
+ args: mar-eng
2038
+ metrics:
2039
+ - name: BLEU
2040
+ type: bleu
2041
+ value: 31.8
2042
+ - name: chr-F
2043
+ type: chrf
2044
+ value: 0.59329
2045
+ - task:
2046
+ name: Translation mar-fra
2047
+ type: translation
2048
+ args: mar-fra
2049
+ dataset:
2050
+ name: tico19-test
2051
+ type: tico19-test
2052
+ args: mar-fra
2053
+ metrics:
2054
+ - name: BLEU
2055
+ type: bleu
2056
+ value: 19.3
2057
+ - name: chr-F
2058
+ type: chrf
2059
+ value: 0.46574
2060
+ - task:
2061
+ name: Translation mar-por
2062
+ type: translation
2063
+ args: mar-por
2064
+ dataset:
2065
+ name: tico19-test
2066
+ type: tico19-test
2067
+ args: mar-por
2068
+ metrics:
2069
+ - name: BLEU
2070
+ type: bleu
2071
+ value: 23.6
2072
+ - name: chr-F
2073
+ type: chrf
2074
+ value: 0.51463
2075
+ - task:
2076
+ name: Translation mar-spa
2077
+ type: translation
2078
+ args: mar-spa
2079
+ dataset:
2080
+ name: tico19-test
2081
+ type: tico19-test
2082
+ args: mar-spa
2083
+ metrics:
2084
+ - name: BLEU
2085
+ type: bleu
2086
+ value: 25.7
2087
+ - name: chr-F
2088
+ type: chrf
2089
+ value: 0.52551
2090
+ - task:
2091
+ name: Translation nep-eng
2092
+ type: translation
2093
+ args: nep-eng
2094
+ dataset:
2095
+ name: tico19-test
2096
+ type: tico19-test
2097
+ args: nep-eng
2098
+ metrics:
2099
+ - name: BLEU
2100
+ type: bleu
2101
+ value: 40.7
2102
+ - name: chr-F
2103
+ type: chrf
2104
+ value: 0.66283
2105
+ - task:
2106
+ name: Translation nep-fra
2107
+ type: translation
2108
+ args: nep-fra
2109
+ dataset:
2110
+ name: tico19-test
2111
+ type: tico19-test
2112
+ args: nep-fra
2113
+ metrics:
2114
+ - name: BLEU
2115
+ type: bleu
2116
+ value: 22.8
2117
+ - name: chr-F
2118
+ type: chrf
2119
+ value: 0.50397
2120
+ - task:
2121
+ name: Translation nep-por
2122
+ type: translation
2123
+ args: nep-por
2124
+ dataset:
2125
+ name: tico19-test
2126
+ type: tico19-test
2127
+ args: nep-por
2128
+ metrics:
2129
+ - name: BLEU
2130
+ type: bleu
2131
+ value: 28.1
2132
+ - name: chr-F
2133
+ type: chrf
2134
+ value: 0.55951
2135
+ - task:
2136
+ name: Translation nep-spa
2137
+ type: translation
2138
+ args: nep-spa
2139
+ dataset:
2140
+ name: tico19-test
2141
+ type: tico19-test
2142
+ args: nep-spa
2143
+ metrics:
2144
+ - name: BLEU
2145
+ type: bleu
2146
+ value: 30.3
2147
+ - name: chr-F
2148
+ type: chrf
2149
+ value: 0.57272
2150
+ - task:
2151
+ name: Translation urd-eng
2152
+ type: translation
2153
+ args: urd-eng
2154
+ dataset:
2155
+ name: tico19-test
2156
+ type: tico19-test
2157
+ args: urd-eng
2158
+ metrics:
2159
+ - name: BLEU
2160
+ type: bleu
2161
+ value: 30.5
2162
+ - name: chr-F
2163
+ type: chrf
2164
+ value: 0.57473
2165
+ - task:
2166
+ name: Translation urd-fra
2167
+ type: translation
2168
+ args: urd-fra
2169
+ dataset:
2170
+ name: tico19-test
2171
+ type: tico19-test
2172
+ args: urd-fra
2173
+ metrics:
2174
+ - name: BLEU
2175
+ type: bleu
2176
+ value: 19.6
2177
+ - name: chr-F
2178
+ type: chrf
2179
+ value: 0.46725
2180
+ - task:
2181
+ name: Translation urd-por
2182
+ type: translation
2183
+ args: urd-por
2184
+ dataset:
2185
+ name: tico19-test
2186
+ type: tico19-test
2187
+ args: urd-por
2188
+ metrics:
2189
+ - name: BLEU
2190
+ type: bleu
2191
+ value: 23.5
2192
+ - name: chr-F
2193
+ type: chrf
2194
+ value: 0.50913
2195
+ - task:
2196
+ name: Translation urd-spa
2197
+ type: translation
2198
+ args: urd-spa
2199
+ dataset:
2200
+ name: tico19-test
2201
+ type: tico19-test
2202
+ args: urd-spa
2203
+ metrics:
2204
+ - name: BLEU
2205
+ type: bleu
2206
+ value: 25.8
2207
+ - name: chr-F
2208
+ type: chrf
2209
+ value: 0.52387
2210
+ - task:
2211
+ name: Translation hin-eng
2212
+ type: translation
2213
+ args: hin-eng
2214
+ dataset:
2215
+ name: newstest2014
2216
+ type: wmt-2014-news
2217
+ args: hin-eng
2218
+ metrics:
2219
+ - name: BLEU
2220
+ type: bleu
2221
+ value: 30.3
2222
+ - name: chr-F
2223
+ type: chrf
2224
+ value: 0.59329
2225
+ - task:
2226
+ name: Translation guj-eng
2227
+ type: translation
2228
+ args: guj-eng
2229
+ dataset:
2230
+ name: newstest2019
2231
+ type: wmt-2019-news
2232
+ args: guj-eng
2233
+ metrics:
2234
+ - name: BLEU
2235
+ type: bleu
2236
+ value: 26.9
2237
+ - name: chr-F
2238
+ type: chrf
2239
+ value: 0.53383
2240
+ ---
2241
+ # opus-mt-tc-bible-big-inc-deu_eng_fra_por_spa
2242
+
2243
+ ## Table of Contents
2244
+ - [Model Details](#model-details)
2245
+ - [Uses](#uses)
2246
+ - [Risks, Limitations and Biases](#risks-limitations-and-biases)
2247
+ - [How to Get Started With the Model](#how-to-get-started-with-the-model)
2248
+ - [Training](#training)
2249
+ - [Evaluation](#evaluation)
2250
+ - [Citation Information](#citation-information)
2251
+ - [Acknowledgements](#acknowledgements)
2252
+
2253
+ ## Model Details
2254
+
2255
+ Neural machine translation model for translating from Indic languages (inc) to unknown (deu+eng+fra+por+spa).
2256
+
2257
+ 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).
2258
+ **Model Description:**
2259
+ - **Developed by:** Language Technology Research Group at the University of Helsinki
2260
+ - **Model Type:** Translation (transformer-big)
2261
+ - **Release**: 2024-05-30
2262
+ - **License:** Apache-2.0
2263
+ - **Language(s):**
2264
+ - Source Language(s): anp asm awa ben bho bpy div dty gbm guj hif hin hne hns kas kok lah mag mai mar nep npi ori pan pli rhg rmy rom san sin skr snd syl urd
2265
+ - Target Language(s): deu eng fra por spa
2266
+ - Valid Target Language Labels: >>deu<< >>eng<< >>fra<< >>por<< >>spa<< >>xxx<<
2267
+ - **Original Model**: [opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-30.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/inc-deu+eng+fra+por+spa/opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-30.zip)
2268
+ - **Resources for more information:**
2269
+ - [OPUS-MT dashboard](https://opus.nlpl.eu/dashboard/index.php?pkg=opusmt&test=all&scoreslang=all&chart=standard&model=Tatoeba-MT-models/inc-deu%2Beng%2Bfra%2Bpor%2Bspa/opusTCv20230926max50%2Bbt%2Bjhubc_transformer-big_2024-05-30)
2270
+ - [OPUS-MT-train GitHub Repo](https://github.com/Helsinki-NLP/OPUS-MT-train)
2271
+ - [More information about MarianNMT models in the transformers library](https://huggingface.co/docs/transformers/model_doc/marian)
2272
+ - [Tatoeba Translation Challenge](https://github.com/Helsinki-NLP/Tatoeba-Challenge/)
2273
+ - [HPLT bilingual data v1 (as part of the Tatoeba Translation Challenge dataset)](https://hplt-project.org/datasets/v1)
2274
+ - [A massively parallel Bible corpus](https://aclanthology.org/L14-1215/)
2275
+
2276
+ 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. `>>deu<<`
2277
+
2278
+ ## Uses
2279
+
2280
+ This model can be used for translation and text-to-text generation.
2281
+
2282
+ ## Risks, Limitations and Biases
2283
+
2284
+ **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.**
2285
+
2286
+ 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)).
2287
+
2288
+ ## How to Get Started With the Model
2289
+
2290
+ A short example code:
2291
+
2292
+ ```python
2293
+ from transformers import MarianMTModel, MarianTokenizer
2294
+
2295
+ src_text = [
2296
+ ">>deu<< Replace this with text in an accepted source language.",
2297
+ ">>spa<< This is the second sentence."
2298
+ ]
2299
+
2300
+ model_name = "pytorch-models/opus-mt-tc-bible-big-inc-deu_eng_fra_por_spa"
2301
+ tokenizer = MarianTokenizer.from_pretrained(model_name)
2302
+ model = MarianMTModel.from_pretrained(model_name)
2303
+ translated = model.generate(**tokenizer(src_text, return_tensors="pt", padding=True))
2304
+
2305
+ for t in translated:
2306
+ print( tokenizer.decode(t, skip_special_tokens=True) )
2307
+ ```
2308
+
2309
+ You can also use OPUS-MT models with the transformers pipelines, for example:
2310
+
2311
+ ```python
2312
+ from transformers import pipeline
2313
+ pipe = pipeline("translation", model="Helsinki-NLP/opus-mt-tc-bible-big-inc-deu_eng_fra_por_spa")
2314
+ print(pipe(">>deu<< Replace this with text in an accepted source language."))
2315
+ ```
2316
+
2317
+ ## Training
2318
+
2319
+ - **Data**: opusTCv20230926max50+bt+jhubc ([source](https://github.com/Helsinki-NLP/Tatoeba-Challenge))
2320
+ - **Pre-processing**: SentencePiece (spm32k,spm32k)
2321
+ - **Model Type:** transformer-big
2322
+ - **Original MarianNMT Model**: [opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-30.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/inc-deu+eng+fra+por+spa/opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-30.zip)
2323
+ - **Training Scripts**: [GitHub Repo](https://github.com/Helsinki-NLP/OPUS-MT-train)
2324
+
2325
+ ## Evaluation
2326
+
2327
+ * [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/inc-deu%2Beng%2Bfra%2Bpor%2Bspa/opusTCv20230926max50%2Bbt%2Bjhubc_transformer-big_2024-05-30)
2328
+ * test set translations: [opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-29.test.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/inc-deu+eng+fra+por+spa/opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-29.test.txt)
2329
+ * test set scores: [opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-29.eval.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/inc-deu+eng+fra+por+spa/opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-29.eval.txt)
2330
+ * benchmark results: [benchmark_results.txt](benchmark_results.txt)
2331
+ * benchmark output: [benchmark_translations.zip](benchmark_translations.zip)
2332
+
2333
+ | langpair | testset | chr-F | BLEU | #sent | #words |
2334
+ |----------|---------|-------|-------|-------|--------|
2335
+ | awa-eng | tatoeba-test-v2021-08-07 | 0.60390 | 40.8 | 279 | 1335 |
2336
+ | ben-eng | tatoeba-test-v2021-08-07 | 0.64078 | 49.4 | 2500 | 13978 |
2337
+ | hin-eng | tatoeba-test-v2021-08-07 | 0.64929 | 49.1 | 5000 | 33943 |
2338
+ | mar-eng | tatoeba-test-v2021-08-07 | 0.64074 | 48.0 | 10396 | 67527 |
2339
+ | urd-eng | tatoeba-test-v2021-08-07 | 0.52963 | 35.0 | 1663 | 12029 |
2340
+ | ben-eng | flores101-devtest | 0.57906 | 30.4 | 1012 | 24721 |
2341
+ | ben-fra | flores101-devtest | 0.50109 | 21.9 | 1012 | 28343 |
2342
+ | guj-spa | flores101-devtest | 0.44065 | 15.2 | 1012 | 29199 |
2343
+ | mar-deu | flores101-devtest | 0.44067 | 13.8 | 1012 | 25094 |
2344
+ | mar-por | flores101-devtest | 0.46685 | 18.6 | 1012 | 26519 |
2345
+ | mar-spa | flores101-devtest | 0.41662 | 14.0 | 1012 | 29199 |
2346
+ | pan-eng | flores101-devtest | 0.59922 | 33.0 | 1012 | 24721 |
2347
+ | pan-por | flores101-devtest | 0.49373 | 21.9 | 1012 | 26519 |
2348
+ | pan-spa | flores101-devtest | 0.43910 | 15.4 | 1012 | 29199 |
2349
+ | asm-eng | flores200-devtest | 0.48584 | 21.9 | 1012 | 24721 |
2350
+ | awa-deu | flores200-devtest | 0.47173 | 16.5 | 1012 | 25094 |
2351
+ | awa-eng | flores200-devtest | 0.50582 | 24.5 | 1012 | 24721 |
2352
+ | awa-fra | flores200-devtest | 0.49682 | 21.4 | 1012 | 28343 |
2353
+ | awa-por | flores200-devtest | 0.49663 | 21.5 | 1012 | 26519 |
2354
+ | awa-spa | flores200-devtest | 0.43740 | 15.1 | 1012 | 29199 |
2355
+ | ben-deu | flores200-devtest | 0.47330 | 16.6 | 1012 | 25094 |
2356
+ | ben-eng | flores200-devtest | 0.58077 | 30.5 | 1012 | 24721 |
2357
+ | ben-fra | flores200-devtest | 0.50884 | 22.6 | 1012 | 28343 |
2358
+ | ben-por | flores200-devtest | 0.50054 | 21.4 | 1012 | 26519 |
2359
+ | ben-spa | flores200-devtest | 0.44159 | 15.2 | 1012 | 29199 |
2360
+ | bho-deu | flores200-devtest | 0.42660 | 12.6 | 1012 | 25094 |
2361
+ | bho-eng | flores200-devtest | 0.50609 | 22.7 | 1012 | 24721 |
2362
+ | bho-fra | flores200-devtest | 0.44889 | 16.8 | 1012 | 28343 |
2363
+ | bho-por | flores200-devtest | 0.44582 | 16.9 | 1012 | 26519 |
2364
+ | bho-spa | flores200-devtest | 0.40581 | 13.1 | 1012 | 29199 |
2365
+ | guj-deu | flores200-devtest | 0.46665 | 16.8 | 1012 | 25094 |
2366
+ | guj-eng | flores200-devtest | 0.61383 | 34.3 | 1012 | 24721 |
2367
+ | guj-fra | flores200-devtest | 0.50410 | 22.3 | 1012 | 28343 |
2368
+ | guj-por | flores200-devtest | 0.49257 | 21.3 | 1012 | 26519 |
2369
+ | guj-spa | flores200-devtest | 0.44565 | 15.6 | 1012 | 29199 |
2370
+ | hin-deu | flores200-devtest | 0.50226 | 20.4 | 1012 | 25094 |
2371
+ | hin-eng | flores200-devtest | 0.63336 | 37.3 | 1012 | 24721 |
2372
+ | hin-fra | flores200-devtest | 0.53701 | 25.9 | 1012 | 28343 |
2373
+ | hin-por | flores200-devtest | 0.53448 | 25.5 | 1012 | 26519 |
2374
+ | hin-spa | flores200-devtest | 0.46171 | 17.2 | 1012 | 29199 |
2375
+ | hne-deu | flores200-devtest | 0.49698 | 19.0 | 1012 | 25094 |
2376
+ | hne-eng | flores200-devtest | 0.63936 | 38.5 | 1012 | 24721 |
2377
+ | hne-fra | flores200-devtest | 0.52835 | 25.3 | 1012 | 28343 |
2378
+ | hne-por | flores200-devtest | 0.52788 | 25.0 | 1012 | 26519 |
2379
+ | hne-spa | flores200-devtest | 0.45443 | 16.7 | 1012 | 29199 |
2380
+ | mag-deu | flores200-devtest | 0.50359 | 19.7 | 1012 | 25094 |
2381
+ | mag-eng | flores200-devtest | 0.63906 | 38.0 | 1012 | 24721 |
2382
+ | mag-fra | flores200-devtest | 0.53616 | 25.8 | 1012 | 28343 |
2383
+ | mag-por | flores200-devtest | 0.53537 | 25.9 | 1012 | 26519 |
2384
+ | mag-spa | flores200-devtest | 0.45822 | 16.9 | 1012 | 29199 |
2385
+ | mai-deu | flores200-devtest | 0.46791 | 16.2 | 1012 | 25094 |
2386
+ | mai-eng | flores200-devtest | 0.57461 | 30.4 | 1012 | 24721 |
2387
+ | mai-fra | flores200-devtest | 0.50585 | 22.1 | 1012 | 28343 |
2388
+ | mai-por | flores200-devtest | 0.50490 | 22.0 | 1012 | 26519 |
2389
+ | mai-spa | flores200-devtest | 0.44366 | 15.3 | 1012 | 29199 |
2390
+ | mar-deu | flores200-devtest | 0.44725 | 14.5 | 1012 | 25094 |
2391
+ | mar-eng | flores200-devtest | 0.58500 | 31.4 | 1012 | 24721 |
2392
+ | mar-fra | flores200-devtest | 0.47027 | 19.5 | 1012 | 28343 |
2393
+ | mar-por | flores200-devtest | 0.47216 | 19.3 | 1012 | 26519 |
2394
+ | mar-spa | flores200-devtest | 0.42178 | 14.2 | 1012 | 29199 |
2395
+ | npi-deu | flores200-devtest | 0.46631 | 16.4 | 1012 | 25094 |
2396
+ | npi-eng | flores200-devtest | 0.59776 | 32.3 | 1012 | 24721 |
2397
+ | npi-fra | flores200-devtest | 0.50548 | 22.5 | 1012 | 28343 |
2398
+ | npi-por | flores200-devtest | 0.50202 | 21.7 | 1012 | 26519 |
2399
+ | npi-spa | flores200-devtest | 0.43804 | 15.3 | 1012 | 29199 |
2400
+ | pan-deu | flores200-devtest | 0.48421 | 18.7 | 1012 | 25094 |
2401
+ | pan-eng | flores200-devtest | 0.60676 | 33.8 | 1012 | 24721 |
2402
+ | pan-fra | flores200-devtest | 0.51368 | 23.5 | 1012 | 28343 |
2403
+ | pan-por | flores200-devtest | 0.50586 | 22.7 | 1012 | 26519 |
2404
+ | pan-spa | flores200-devtest | 0.44653 | 16.5 | 1012 | 29199 |
2405
+ | sin-deu | flores200-devtest | 0.44676 | 14.2 | 1012 | 25094 |
2406
+ | sin-eng | flores200-devtest | 0.54777 | 26.8 | 1012 | 24721 |
2407
+ | sin-fra | flores200-devtest | 0.47283 | 19.0 | 1012 | 28343 |
2408
+ | sin-por | flores200-devtest | 0.46935 | 18.4 | 1012 | 26519 |
2409
+ | sin-spa | flores200-devtest | 0.42143 | 13.7 | 1012 | 29199 |
2410
+ | urd-deu | flores200-devtest | 0.46542 | 17.1 | 1012 | 25094 |
2411
+ | urd-eng | flores200-devtest | 0.56935 | 29.3 | 1012 | 24721 |
2412
+ | urd-fra | flores200-devtest | 0.50276 | 22.3 | 1012 | 28343 |
2413
+ | urd-por | flores200-devtest | 0.48010 | 20.3 | 1012 | 26519 |
2414
+ | urd-spa | flores200-devtest | 0.43032 | 14.7 | 1012 | 29199 |
2415
+ | hin-eng | newstest2014 | 0.59329 | 30.3 | 2507 | 55571 |
2416
+ | guj-eng | newstest2019 | 0.53383 | 26.9 | 1016 | 17757 |
2417
+ | ben-deu | ntrex128 | 0.45180 | 14.6 | 1997 | 48761 |
2418
+ | ben-eng | ntrex128 | 0.57247 | 29.5 | 1997 | 47673 |
2419
+ | ben-fra | ntrex128 | 0.46475 | 18.0 | 1997 | 53481 |
2420
+ | ben-por | ntrex128 | 0.45486 | 16.8 | 1997 | 51631 |
2421
+ | ben-spa | ntrex128 | 0.48738 | 21.1 | 1997 | 54107 |
2422
+ | guj-deu | ntrex128 | 0.43539 | 13.9 | 1997 | 48761 |
2423
+ | guj-eng | ntrex128 | 0.58894 | 31.6 | 1997 | 47673 |
2424
+ | guj-fra | ntrex128 | 0.45075 | 16.9 | 1997 | 53481 |
2425
+ | guj-por | ntrex128 | 0.43567 | 15.2 | 1997 | 51631 |
2426
+ | guj-spa | ntrex128 | 0.47525 | 20.2 | 1997 | 54107 |
2427
+ | hin-deu | ntrex128 | 0.46336 | 15.0 | 1997 | 48761 |
2428
+ | hin-eng | ntrex128 | 0.59842 | 31.5 | 1997 | 47673 |
2429
+ | hin-fra | ntrex128 | 0.48208 | 19.2 | 1997 | 53481 |
2430
+ | hin-por | ntrex128 | 0.46509 | 17.6 | 1997 | 51631 |
2431
+ | hin-spa | ntrex128 | 0.49436 | 21.8 | 1997 | 54107 |
2432
+ | mar-deu | ntrex128 | 0.43119 | 12.8 | 1997 | 48761 |
2433
+ | mar-eng | ntrex128 | 0.55151 | 27.3 | 1997 | 47673 |
2434
+ | mar-fra | ntrex128 | 0.43957 | 16.2 | 1997 | 53481 |
2435
+ | mar-por | ntrex128 | 0.43555 | 15.4 | 1997 | 51631 |
2436
+ | mar-spa | ntrex128 | 0.46271 | 19.1 | 1997 | 54107 |
2437
+ | nep-deu | ntrex128 | 0.42940 | 13.0 | 1997 | 48761 |
2438
+ | nep-eng | ntrex128 | 0.56277 | 29.1 | 1997 | 47673 |
2439
+ | nep-fra | ntrex128 | 0.44663 | 16.5 | 1997 | 53481 |
2440
+ | nep-por | ntrex128 | 0.43686 | 15.4 | 1997 | 51631 |
2441
+ | nep-spa | ntrex128 | 0.46553 | 19.3 | 1997 | 54107 |
2442
+ | pan-deu | ntrex128 | 0.44036 | 14.1 | 1997 | 48761 |
2443
+ | pan-eng | ntrex128 | 0.58427 | 31.6 | 1997 | 47673 |
2444
+ | pan-fra | ntrex128 | 0.45593 | 17.3 | 1997 | 53481 |
2445
+ | pan-por | ntrex128 | 0.44264 | 15.9 | 1997 | 51631 |
2446
+ | pan-spa | ntrex128 | 0.47199 | 20.0 | 1997 | 54107 |
2447
+ | sin-deu | ntrex128 | 0.42280 | 12.4 | 1997 | 48761 |
2448
+ | sin-eng | ntrex128 | 0.52576 | 24.6 | 1997 | 47673 |
2449
+ | sin-fra | ntrex128 | 0.43594 | 15.6 | 1997 | 53481 |
2450
+ | sin-por | ntrex128 | 0.42751 | 14.4 | 1997 | 51631 |
2451
+ | sin-spa | ntrex128 | 0.45890 | 18.3 | 1997 | 54107 |
2452
+ | urd-deu | ntrex128 | 0.45737 | 15.6 | 1997 | 48761 |
2453
+ | urd-eng | ntrex128 | 0.56781 | 28.6 | 1997 | 47673 |
2454
+ | urd-fra | ntrex128 | 0.47298 | 18.9 | 1997 | 53481 |
2455
+ | urd-por | ntrex128 | 0.45273 | 16.2 | 1997 | 51631 |
2456
+ | urd-spa | ntrex128 | 0.48644 | 21.0 | 1997 | 54107 |
2457
+ | ben-eng | tico19-test | 0.64568 | 38.2 | 2100 | 56824 |
2458
+ | ben-fra | tico19-test | 0.49799 | 22.0 | 2100 | 64661 |
2459
+ | ben-por | tico19-test | 0.55115 | 27.2 | 2100 | 62729 |
2460
+ | ben-spa | tico19-test | 0.56847 | 29.9 | 2100 | 66563 |
2461
+ | hin-eng | tico19-test | 0.70694 | 46.6 | 2100 | 56323 |
2462
+ | hin-fra | tico19-test | 0.53932 | 26.7 | 2100 | 64661 |
2463
+ | hin-por | tico19-test | 0.60581 | 33.4 | 2100 | 62729 |
2464
+ | hin-spa | tico19-test | 0.61585 | 35.7 | 2100 | 66563 |
2465
+ | mar-eng | tico19-test | 0.59329 | 31.8 | 2100 | 56315 |
2466
+ | mar-fra | tico19-test | 0.46574 | 19.3 | 2100 | 64661 |
2467
+ | mar-por | tico19-test | 0.51463 | 23.6 | 2100 | 62729 |
2468
+ | mar-spa | tico19-test | 0.52551 | 25.7 | 2100 | 66563 |
2469
+ | nep-eng | tico19-test | 0.66283 | 40.7 | 2100 | 56824 |
2470
+ | nep-fra | tico19-test | 0.50397 | 22.8 | 2100 | 64661 |
2471
+ | nep-por | tico19-test | 0.55951 | 28.1 | 2100 | 62729 |
2472
+ | nep-spa | tico19-test | 0.57272 | 30.3 | 2100 | 66563 |
2473
+ | urd-eng | tico19-test | 0.57473 | 30.5 | 2100 | 56315 |
2474
+ | urd-fra | tico19-test | 0.46725 | 19.6 | 2100 | 64661 |
2475
+ | urd-por | tico19-test | 0.50913 | 23.5 | 2100 | 62729 |
2476
+ | urd-spa | tico19-test | 0.52387 | 25.8 | 2100 | 66563 |
2477
+
2478
+ ## Citation Information
2479
+
2480
+ * 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.)
2481
+
2482
+ ```bibtex
2483
+ @article{tiedemann2023democratizing,
2484
+ title={Democratizing neural machine translation with {OPUS-MT}},
2485
+ 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},
2486
+ journal={Language Resources and Evaluation},
2487
+ number={58},
2488
+ pages={713--755},
2489
+ year={2023},
2490
+ publisher={Springer Nature},
2491
+ issn={1574-0218},
2492
+ doi={10.1007/s10579-023-09704-w}
2493
+ }
2494
+
2495
+ @inproceedings{tiedemann-thottingal-2020-opus,
2496
+ title = "{OPUS}-{MT} {--} Building open translation services for the World",
2497
+ author = {Tiedemann, J{\"o}rg and Thottingal, Santhosh},
2498
+ booktitle = "Proceedings of the 22nd Annual Conference of the European Association for Machine Translation",
2499
+ month = nov,
2500
+ year = "2020",
2501
+ address = "Lisboa, Portugal",
2502
+ publisher = "European Association for Machine Translation",
2503
+ url = "https://aclanthology.org/2020.eamt-1.61",
2504
+ pages = "479--480",
2505
+ }
2506
+
2507
+ @inproceedings{tiedemann-2020-tatoeba,
2508
+ title = "The Tatoeba Translation Challenge {--} Realistic Data Sets for Low Resource and Multilingual {MT}",
2509
+ author = {Tiedemann, J{\"o}rg},
2510
+ booktitle = "Proceedings of the Fifth Conference on Machine Translation",
2511
+ month = nov,
2512
+ year = "2020",
2513
+ address = "Online",
2514
+ publisher = "Association for Computational Linguistics",
2515
+ url = "https://aclanthology.org/2020.wmt-1.139",
2516
+ pages = "1174--1182",
2517
+ }
2518
+ ```
2519
+
2520
+ ## Acknowledgements
2521
+
2522
+ 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/).
2523
+
2524
+ ## Model conversion info
2525
+
2526
+ * transformers version: 4.45.1
2527
+ * OPUS-MT git hash: 0882077
2528
+ * port time: Tue Oct 8 11:39:25 EEST 2024
2529
+ * port machine: LM0-400-22516.local
benchmark_results.txt ADDED
@@ -0,0 +1,178 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ multi-multi tatoeba-test-v2020-07-28-v2023-09-26 0.61360 44.6 10000 64736
2
+ asm-fra flores101-devtest 0.33042 9.5 1012 28343
3
+ asm-por flores101-devtest 0.34689 10.0 1012 26519
4
+ asm-spa flores101-devtest 0.32971 7.9 1012 29199
5
+ ben-eng flores101-devtest 0.57906 30.4 1012 24721
6
+ ben-fra flores101-devtest 0.50109 21.9 1012 28343
7
+ guj-spa flores101-devtest 0.44065 15.2 1012 29199
8
+ mar-deu flores101-devtest 0.44067 13.8 1012 25094
9
+ mar-por flores101-devtest 0.46685 18.6 1012 26519
10
+ mar-spa flores101-devtest 0.41662 14.0 1012 29199
11
+ pan-eng flores101-devtest 0.59922 33.0 1012 24721
12
+ pan-por flores101-devtest 0.49373 21.9 1012 26519
13
+ pan-spa flores101-devtest 0.43910 15.4 1012 29199
14
+ asm-deu flores200-devtest 0.30387 4.6 1012 25094
15
+ asm-eng flores200-devtest 0.48584 21.9 1012 24721
16
+ asm-fra flores200-devtest 0.32215 8.9 1012 28343
17
+ asm-por flores200-devtest 0.35028 10.1 1012 26519
18
+ asm-spa flores200-devtest 0.31983 7.6 1012 29199
19
+ awa-deu flores200-devtest 0.47173 16.5 1012 25094
20
+ awa-eng flores200-devtest 0.50582 24.5 1012 24721
21
+ awa-fra flores200-devtest 0.49682 21.4 1012 28343
22
+ awa-por flores200-devtest 0.49663 21.5 1012 26519
23
+ awa-spa flores200-devtest 0.43740 15.1 1012 29199
24
+ ben-deu flores200-devtest 0.47330 16.6 1012 25094
25
+ ben-eng flores200-devtest 0.58077 30.5 1012 24721
26
+ ben-fra flores200-devtest 0.50884 22.6 1012 28343
27
+ ben-por flores200-devtest 0.50054 21.4 1012 26519
28
+ ben-spa flores200-devtest 0.44159 15.2 1012 29199
29
+ bho-deu flores200-devtest 0.42660 12.6 1012 25094
30
+ bho-eng flores200-devtest 0.50609 22.7 1012 24721
31
+ bho-fra flores200-devtest 0.44889 16.8 1012 28343
32
+ bho-por flores200-devtest 0.44582 16.9 1012 26519
33
+ bho-spa flores200-devtest 0.40581 13.1 1012 29199
34
+ guj-deu flores200-devtest 0.46665 16.8 1012 25094
35
+ guj-eng flores200-devtest 0.61383 34.3 1012 24721
36
+ guj-fra flores200-devtest 0.50410 22.3 1012 28343
37
+ guj-por flores200-devtest 0.49257 21.3 1012 26519
38
+ guj-spa flores200-devtest 0.44565 15.6 1012 29199
39
+ hin-deu flores200-devtest 0.50226 20.4 1012 25094
40
+ hin-eng flores200-devtest 0.63336 37.3 1012 24721
41
+ hin-fra flores200-devtest 0.53701 25.9 1012 28343
42
+ hin-por flores200-devtest 0.53448 25.5 1012 26519
43
+ hin-spa flores200-devtest 0.46171 17.2 1012 29199
44
+ hne-deu flores200-devtest 0.49698 19.0 1012 25094
45
+ hne-eng flores200-devtest 0.63936 38.5 1012 24721
46
+ hne-fra flores200-devtest 0.52835 25.3 1012 28343
47
+ hne-por flores200-devtest 0.52788 25.0 1012 26519
48
+ hne-spa flores200-devtest 0.45443 16.7 1012 29199
49
+ kas_Arab-deu flores200-devtest 0.30399 2.8 1012 25094
50
+ kas_Arab-eng flores200-devtest 0.34470 5.6 1012 24721
51
+ kas_Arab-fra flores200-devtest 0.31008 4.2 1012 28343
52
+ kas_Arab-por flores200-devtest 0.29714 3.8 1012 26519
53
+ kas_Arab-spa flores200-devtest 0.29529 3.7 1012 29199
54
+ kas_Deva-deu flores200-devtest 0.21035 1.7 1012 25094
55
+ kas_Deva-eng flores200-devtest 0.26718 4.7 1012 24721
56
+ kas_Deva-fra flores200-devtest 0.21000 2.5 1012 28343
57
+ kas_Deva-por flores200-devtest 0.21490 2.6 1012 26519
58
+ kas_Deva-spa flores200-devtest 0.21290 2.6 1012 29199
59
+ mag-deu flores200-devtest 0.50359 19.7 1012 25094
60
+ mag-eng flores200-devtest 0.63906 38.0 1012 24721
61
+ mag-fra flores200-devtest 0.53616 25.8 1012 28343
62
+ mag-por flores200-devtest 0.53537 25.9 1012 26519
63
+ mag-spa flores200-devtest 0.45822 16.9 1012 29199
64
+ mai-deu flores200-devtest 0.46791 16.2 1012 25094
65
+ mai-eng flores200-devtest 0.57461 30.4 1012 24721
66
+ mai-fra flores200-devtest 0.50585 22.1 1012 28343
67
+ mai-por flores200-devtest 0.50490 22.0 1012 26519
68
+ mai-spa flores200-devtest 0.44366 15.3 1012 29199
69
+ mar-deu flores200-devtest 0.44725 14.5 1012 25094
70
+ mar-eng flores200-devtest 0.58500 31.4 1012 24721
71
+ mar-fra flores200-devtest 0.47027 19.5 1012 28343
72
+ mar-por flores200-devtest 0.47216 19.3 1012 26519
73
+ mar-spa flores200-devtest 0.42178 14.2 1012 29199
74
+ npi-deu flores200-devtest 0.46631 16.4 1012 25094
75
+ npi-eng flores200-devtest 0.59776 32.3 1012 24721
76
+ npi-fra flores200-devtest 0.50548 22.5 1012 28343
77
+ npi-por flores200-devtest 0.50202 21.7 1012 26519
78
+ npi-spa flores200-devtest 0.43804 15.3 1012 29199
79
+ pan-deu flores200-devtest 0.48421 18.7 1012 25094
80
+ pan-eng flores200-devtest 0.60676 33.8 1012 24721
81
+ pan-fra flores200-devtest 0.51368 23.5 1012 28343
82
+ pan-por flores200-devtest 0.50586 22.7 1012 26519
83
+ pan-spa flores200-devtest 0.44653 16.5 1012 29199
84
+ san-deu flores200-devtest 0.30491 4.5 1012 25094
85
+ san-eng flores200-devtest 0.36887 11.8 1012 24721
86
+ san-fra flores200-devtest 0.31881 7.8 1012 28343
87
+ san-por flores200-devtest 0.33004 7.9 1012 26519
88
+ san-spa flores200-devtest 0.30853 6.6 1012 29199
89
+ sin-deu flores200-devtest 0.44676 14.2 1012 25094
90
+ sin-eng flores200-devtest 0.54777 26.8 1012 24721
91
+ sin-fra flores200-devtest 0.47283 19.0 1012 28343
92
+ sin-por flores200-devtest 0.46935 18.4 1012 26519
93
+ sin-spa flores200-devtest 0.42143 13.7 1012 29199
94
+ urd-deu flores200-devtest 0.46542 17.1 1012 25094
95
+ urd-eng flores200-devtest 0.56935 29.3 1012 24721
96
+ urd-fra flores200-devtest 0.50276 22.3 1012 28343
97
+ urd-por flores200-devtest 0.48010 20.3 1012 26519
98
+ urd-spa flores200-devtest 0.43032 14.7 1012 29199
99
+ hin-eng newstest2014 0.59329 30.3 2507 55571
100
+ guj-eng newstest2019 0.53383 26.9 1016 17757
101
+ ben-deu ntrex128 0.45180 14.6 1997 48761
102
+ ben-eng ntrex128 0.57247 29.5 1997 47673
103
+ ben-fra ntrex128 0.46475 18.0 1997 53481
104
+ ben-por ntrex128 0.45486 16.8 1997 51631
105
+ ben-spa ntrex128 0.48738 21.1 1997 54107
106
+ div-deu ntrex128 0.20387 0.9 1997 48761
107
+ div-eng ntrex128 0.20170 1.4 1997 47673
108
+ div-fra ntrex128 0.19718 1.2 1997 53481
109
+ div-por ntrex128 0.19274 1.3 1997 51631
110
+ div-spa ntrex128 0.20807 1.5 1997 54107
111
+ guj-deu ntrex128 0.43539 13.9 1997 48761
112
+ guj-eng ntrex128 0.58894 31.6 1997 47673
113
+ guj-fra ntrex128 0.45075 16.9 1997 53481
114
+ guj-por ntrex128 0.43567 15.2 1997 51631
115
+ guj-spa ntrex128 0.47525 20.2 1997 54107
116
+ hin-deu ntrex128 0.46336 15.0 1997 48761
117
+ hin-eng ntrex128 0.59842 31.5 1997 47673
118
+ hin-fra ntrex128 0.48208 19.2 1997 53481
119
+ hin-por ntrex128 0.46509 17.6 1997 51631
120
+ hin-spa ntrex128 0.49436 21.8 1997 54107
121
+ mar-deu ntrex128 0.43119 12.8 1997 48761
122
+ mar-eng ntrex128 0.55151 27.3 1997 47673
123
+ mar-fra ntrex128 0.43957 16.2 1997 53481
124
+ mar-por ntrex128 0.43555 15.4 1997 51631
125
+ mar-spa ntrex128 0.46271 19.1 1997 54107
126
+ nep-deu ntrex128 0.42940 13.0 1997 48761
127
+ nep-eng ntrex128 0.56277 29.1 1997 47673
128
+ nep-fra ntrex128 0.44663 16.5 1997 53481
129
+ nep-por ntrex128 0.43686 15.4 1997 51631
130
+ nep-spa ntrex128 0.46553 19.3 1997 54107
131
+ pan-deu ntrex128 0.44036 14.1 1997 48761
132
+ pan-eng ntrex128 0.58427 31.6 1997 47673
133
+ pan-fra ntrex128 0.45593 17.3 1997 53481
134
+ pan-por ntrex128 0.44264 15.9 1997 51631
135
+ pan-spa ntrex128 0.47199 20.0 1997 54107
136
+ sin-deu ntrex128 0.42280 12.4 1997 48761
137
+ sin-eng ntrex128 0.52576 24.6 1997 47673
138
+ sin-fra ntrex128 0.43594 15.6 1997 53481
139
+ sin-por ntrex128 0.42751 14.4 1997 51631
140
+ sin-spa ntrex128 0.45890 18.3 1997 54107
141
+ snd_Arab-eng ntrex128 0.15440 0.6 1997 47673
142
+ snd_Arab-fra ntrex128 0.15783 0.5 1997 53481
143
+ snd_Arab-por ntrex128 0.10058 0.2 1997 51631
144
+ snd_Arab-spa ntrex128 0.13679 0.4 1997 54107
145
+ urd-deu ntrex128 0.45737 15.6 1997 48761
146
+ urd-eng ntrex128 0.56781 28.6 1997 47673
147
+ urd-fra ntrex128 0.47298 18.9 1997 53481
148
+ urd-por ntrex128 0.45273 16.2 1997 51631
149
+ urd-spa ntrex128 0.48644 21.0 1997 54107
150
+ mar-eng tatoeba-test-v2020-07-28 0.63997 47.9 10000 64831
151
+ rom-eng tatoeba-test-v2020-07-28 0.25545 6.7 671 4457
152
+ awa-eng tatoeba-test-v2021-03-30 0.61380 41.6 280 1339
153
+ awa-eng tatoeba-test-v2021-08-07 0.60390 40.8 279 1335
154
+ ben-eng tatoeba-test-v2021-08-07 0.64078 49.4 2500 13978
155
+ hin-eng tatoeba-test-v2021-08-07 0.64929 49.1 5000 33943
156
+ mar-eng tatoeba-test-v2021-08-07 0.64074 48.0 10396 67527
157
+ rom-eng tatoeba-test-v2021-08-07 0.25329 6.6 706 4690
158
+ urd-eng tatoeba-test-v2021-08-07 0.52963 35.0 1663 12029
159
+ ben-eng tico19-test 0.64568 38.2 2100 56824
160
+ ben-fra tico19-test 0.49799 22.0 2100 64661
161
+ ben-por tico19-test 0.55115 27.2 2100 62729
162
+ ben-spa tico19-test 0.56847 29.9 2100 66563
163
+ hin-eng tico19-test 0.70694 46.6 2100 56323
164
+ hin-fra tico19-test 0.53932 26.7 2100 64661
165
+ hin-por tico19-test 0.60581 33.4 2100 62729
166
+ hin-spa tico19-test 0.61585 35.7 2100 66563
167
+ mar-eng tico19-test 0.59329 31.8 2100 56315
168
+ mar-fra tico19-test 0.46574 19.3 2100 64661
169
+ mar-por tico19-test 0.51463 23.6 2100 62729
170
+ mar-spa tico19-test 0.52551 25.7 2100 66563
171
+ nep-eng tico19-test 0.66283 40.7 2100 56824
172
+ nep-fra tico19-test 0.50397 22.8 2100 64661
173
+ nep-por tico19-test 0.55951 28.1 2100 62729
174
+ nep-spa tico19-test 0.57272 30.3 2100 66563
175
+ urd-eng tico19-test 0.57473 30.5 2100 56315
176
+ urd-fra tico19-test 0.46725 19.6 2100 64661
177
+ urd-por tico19-test 0.50913 23.5 2100 62729
178
+ urd-spa tico19-test 0.52387 25.8 2100 66563
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-inc-deu_eng_fra_por_spa",
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": 61704,
17
+ "decoder_vocab_size": 61705,
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": 482,
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": 61704,
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": 61705
41
+ }
generation_config.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bad_words_ids": [
4
+ [
5
+ 61704
6
+ ]
7
+ ],
8
+ "bos_token_id": 0,
9
+ "decoder_start_token_id": 61704,
10
+ "eos_token_id": 482,
11
+ "forced_eos_token_id": 482,
12
+ "max_length": 512,
13
+ "num_beams": 4,
14
+ "pad_token_id": 61704,
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 sha256:920636616c2f7c2a09c3e398bec8317ed2562d68abf81c5769fd588bfaed0e39
3
+ size 958449620
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0087b4185dd22d27ac0845007f9d0962b34f1cc55594ccfc4a7d0d2b878964b3
3
+ size 958500869
source.spm ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8359db949348ce17f52895c7a2fb8a8dc2e6c75715d40af75300ada75a5e72ae
3
+ size 967996
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 sha256:5f671c62bb6a04015a2889680bb26ff85c9fd98433c2157c24448403ab856aef
3
+ size 802610
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"source_lang": "inc", "target_lang": "deu+eng+fra+por+spa", "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/inc-deu+eng+fra+por+spa", "tokenizer_class": "MarianTokenizer"}
vocab.json ADDED
The diff for this file is too large to render. See raw diff