tiedeman commited on
Commit
a6933ed
1 Parent(s): 054bab6

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 +1,3090 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ library_name: transformers
3
+ language:
4
+ - af
5
+ - ang
6
+ - bar
7
+ - bi
8
+ - bzj
9
+ - de
10
+ - djk
11
+ - drt
12
+ - en
13
+ - enm
14
+ - es
15
+ - fr
16
+ - frr
17
+ - fy
18
+ - gos
19
+ - gsw
20
+ - hrx
21
+ - hwc
22
+ - icr
23
+ - jam
24
+ - kri
25
+ - ksh
26
+ - lb
27
+ - li
28
+ - nds
29
+ - nl
30
+ - ofs
31
+ - pcm
32
+ - pdc
33
+ - pfl
34
+ - pih
35
+ - pis
36
+ - pt
37
+ - rop
38
+ - sco
39
+ - srm
40
+ - srn
41
+ - stq
42
+ - swg
43
+ - tcs
44
+ - tpi
45
+ - vls
46
+ - wae
47
+ - yi
48
+ - zea
49
+
50
+ tags:
51
+ - translation
52
+ - opus-mt-tc-bible
53
+
54
+ license: apache-2.0
55
+ model-index:
56
+ - name: opus-mt-tc-bible-big-gmw-deu_eng_fra_por_spa
57
+ results:
58
+ - task:
59
+ name: Translation afr-deu
60
+ type: translation
61
+ args: afr-deu
62
+ dataset:
63
+ name: flores200-devtest
64
+ type: flores200-devtest
65
+ args: afr-deu
66
+ metrics:
67
+ - name: BLEU
68
+ type: bleu
69
+ value: 29.5
70
+ - name: chr-F
71
+ type: chrf
72
+ value: 0.58456
73
+ - task:
74
+ name: Translation afr-eng
75
+ type: translation
76
+ args: afr-eng
77
+ dataset:
78
+ name: flores200-devtest
79
+ type: flores200-devtest
80
+ args: afr-eng
81
+ metrics:
82
+ - name: BLEU
83
+ type: bleu
84
+ value: 54.7
85
+ - name: chr-F
86
+ type: chrf
87
+ value: 0.74857
88
+ - task:
89
+ name: Translation afr-fra
90
+ type: translation
91
+ args: afr-fra
92
+ dataset:
93
+ name: flores200-devtest
94
+ type: flores200-devtest
95
+ args: afr-fra
96
+ metrics:
97
+ - name: BLEU
98
+ type: bleu
99
+ value: 37.2
100
+ - name: chr-F
101
+ type: chrf
102
+ value: 0.62537
103
+ - task:
104
+ name: Translation afr-por
105
+ type: translation
106
+ args: afr-por
107
+ dataset:
108
+ name: flores200-devtest
109
+ type: flores200-devtest
110
+ args: afr-por
111
+ metrics:
112
+ - name: BLEU
113
+ type: bleu
114
+ value: 36.1
115
+ - name: chr-F
116
+ type: chrf
117
+ value: 0.61751
118
+ - task:
119
+ name: Translation afr-spa
120
+ type: translation
121
+ args: afr-spa
122
+ dataset:
123
+ name: flores200-devtest
124
+ type: flores200-devtest
125
+ args: afr-spa
126
+ metrics:
127
+ - name: BLEU
128
+ type: bleu
129
+ value: 23.2
130
+ - name: chr-F
131
+ type: chrf
132
+ value: 0.51647
133
+ - task:
134
+ name: Translation deu-eng
135
+ type: translation
136
+ args: deu-eng
137
+ dataset:
138
+ name: flores200-devtest
139
+ type: flores200-devtest
140
+ args: deu-eng
141
+ metrics:
142
+ - name: BLEU
143
+ type: bleu
144
+ value: 41.4
145
+ - name: chr-F
146
+ type: chrf
147
+ value: 0.67103
148
+ - task:
149
+ name: Translation deu-fra
150
+ type: translation
151
+ args: deu-fra
152
+ dataset:
153
+ name: flores200-devtest
154
+ type: flores200-devtest
155
+ args: deu-fra
156
+ metrics:
157
+ - name: BLEU
158
+ type: bleu
159
+ value: 36.8
160
+ - name: chr-F
161
+ type: chrf
162
+ value: 0.62658
163
+ - task:
164
+ name: Translation deu-por
165
+ type: translation
166
+ args: deu-por
167
+ dataset:
168
+ name: flores200-devtest
169
+ type: flores200-devtest
170
+ args: deu-por
171
+ metrics:
172
+ - name: BLEU
173
+ type: bleu
174
+ value: 34.8
175
+ - name: chr-F
176
+ type: chrf
177
+ value: 0.60909
178
+ - task:
179
+ name: Translation deu-spa
180
+ type: translation
181
+ args: deu-spa
182
+ dataset:
183
+ name: flores200-devtest
184
+ type: flores200-devtest
185
+ args: deu-spa
186
+ metrics:
187
+ - name: BLEU
188
+ type: bleu
189
+ value: 24.2
190
+ - name: chr-F
191
+ type: chrf
192
+ value: 0.52584
193
+ - task:
194
+ name: Translation eng-deu
195
+ type: translation
196
+ args: eng-deu
197
+ dataset:
198
+ name: flores200-devtest
199
+ type: flores200-devtest
200
+ args: eng-deu
201
+ metrics:
202
+ - name: BLEU
203
+ type: bleu
204
+ value: 37.5
205
+ - name: chr-F
206
+ type: chrf
207
+ value: 0.64560
208
+ - task:
209
+ name: Translation eng-fra
210
+ type: translation
211
+ args: eng-fra
212
+ dataset:
213
+ name: flores200-devtest
214
+ type: flores200-devtest
215
+ args: eng-fra
216
+ metrics:
217
+ - name: BLEU
218
+ type: bleu
219
+ value: 49.1
220
+ - name: chr-F
221
+ type: chrf
222
+ value: 0.70736
223
+ - task:
224
+ name: Translation eng-por
225
+ type: translation
226
+ args: eng-por
227
+ dataset:
228
+ name: flores200-devtest
229
+ type: flores200-devtest
230
+ args: eng-por
231
+ metrics:
232
+ - name: BLEU
233
+ type: bleu
234
+ value: 49.5
235
+ - name: chr-F
236
+ type: chrf
237
+ value: 0.71065
238
+ - task:
239
+ name: Translation eng-spa
240
+ type: translation
241
+ args: eng-spa
242
+ dataset:
243
+ name: flores200-devtest
244
+ type: flores200-devtest
245
+ args: eng-spa
246
+ metrics:
247
+ - name: BLEU
248
+ type: bleu
249
+ value: 27.4
250
+ - name: chr-F
251
+ type: chrf
252
+ value: 0.55738
253
+ - task:
254
+ name: Translation lim-deu
255
+ type: translation
256
+ args: lim-deu
257
+ dataset:
258
+ name: flores200-devtest
259
+ type: flores200-devtest
260
+ args: lim-deu
261
+ metrics:
262
+ - name: BLEU
263
+ type: bleu
264
+ value: 16.1
265
+ - name: chr-F
266
+ type: chrf
267
+ value: 0.45062
268
+ - task:
269
+ name: Translation lim-eng
270
+ type: translation
271
+ args: lim-eng
272
+ dataset:
273
+ name: flores200-devtest
274
+ type: flores200-devtest
275
+ args: lim-eng
276
+ metrics:
277
+ - name: BLEU
278
+ type: bleu
279
+ value: 21.8
280
+ - name: chr-F
281
+ type: chrf
282
+ value: 0.48217
283
+ - task:
284
+ name: Translation lim-fra
285
+ type: translation
286
+ args: lim-fra
287
+ dataset:
288
+ name: flores200-devtest
289
+ type: flores200-devtest
290
+ args: lim-fra
291
+ metrics:
292
+ - name: BLEU
293
+ type: bleu
294
+ value: 18.5
295
+ - name: chr-F
296
+ type: chrf
297
+ value: 0.44347
298
+ - task:
299
+ name: Translation lim-por
300
+ type: translation
301
+ args: lim-por
302
+ dataset:
303
+ name: flores200-devtest
304
+ type: flores200-devtest
305
+ args: lim-por
306
+ metrics:
307
+ - name: BLEU
308
+ type: bleu
309
+ value: 16.8
310
+ - name: chr-F
311
+ type: chrf
312
+ value: 0.42527
313
+ - task:
314
+ name: Translation lim-spa
315
+ type: translation
316
+ args: lim-spa
317
+ dataset:
318
+ name: flores200-devtest
319
+ type: flores200-devtest
320
+ args: lim-spa
321
+ metrics:
322
+ - name: BLEU
323
+ type: bleu
324
+ value: 13.6
325
+ - name: chr-F
326
+ type: chrf
327
+ value: 0.39232
328
+ - task:
329
+ name: Translation ltz-deu
330
+ type: translation
331
+ args: ltz-deu
332
+ dataset:
333
+ name: flores200-devtest
334
+ type: flores200-devtest
335
+ args: ltz-deu
336
+ metrics:
337
+ - name: BLEU
338
+ type: bleu
339
+ value: 31.3
340
+ - name: chr-F
341
+ type: chrf
342
+ value: 0.60114
343
+ - task:
344
+ name: Translation ltz-eng
345
+ type: translation
346
+ args: ltz-eng
347
+ dataset:
348
+ name: flores200-devtest
349
+ type: flores200-devtest
350
+ args: ltz-eng
351
+ metrics:
352
+ - name: BLEU
353
+ type: bleu
354
+ value: 39.3
355
+ - name: chr-F
356
+ type: chrf
357
+ value: 0.64345
358
+ - task:
359
+ name: Translation ltz-fra
360
+ type: translation
361
+ args: ltz-fra
362
+ dataset:
363
+ name: flores200-devtest
364
+ type: flores200-devtest
365
+ args: ltz-fra
366
+ metrics:
367
+ - name: BLEU
368
+ type: bleu
369
+ value: 33.8
370
+ - name: chr-F
371
+ type: chrf
372
+ value: 0.59368
373
+ - task:
374
+ name: Translation ltz-por
375
+ type: translation
376
+ args: ltz-por
377
+ dataset:
378
+ name: flores200-devtest
379
+ type: flores200-devtest
380
+ args: ltz-por
381
+ metrics:
382
+ - name: BLEU
383
+ type: bleu
384
+ value: 24.8
385
+ - name: chr-F
386
+ type: chrf
387
+ value: 0.51545
388
+ - task:
389
+ name: Translation ltz-spa
390
+ type: translation
391
+ args: ltz-spa
392
+ dataset:
393
+ name: flores200-devtest
394
+ type: flores200-devtest
395
+ args: ltz-spa
396
+ metrics:
397
+ - name: BLEU
398
+ type: bleu
399
+ value: 17.5
400
+ - name: chr-F
401
+ type: chrf
402
+ value: 0.44821
403
+ - task:
404
+ name: Translation nld-deu
405
+ type: translation
406
+ args: nld-deu
407
+ dataset:
408
+ name: flores200-devtest
409
+ type: flores200-devtest
410
+ args: nld-deu
411
+ metrics:
412
+ - name: BLEU
413
+ type: bleu
414
+ value: 22.4
415
+ - name: chr-F
416
+ type: chrf
417
+ value: 0.53650
418
+ - task:
419
+ name: Translation nld-eng
420
+ type: translation
421
+ args: nld-eng
422
+ dataset:
423
+ name: flores200-devtest
424
+ type: flores200-devtest
425
+ args: nld-eng
426
+ metrics:
427
+ - name: BLEU
428
+ type: bleu
429
+ value: 30.6
430
+ - name: chr-F
431
+ type: chrf
432
+ value: 0.59102
433
+ - task:
434
+ name: Translation nld-fra
435
+ type: translation
436
+ args: nld-fra
437
+ dataset:
438
+ name: flores200-devtest
439
+ type: flores200-devtest
440
+ args: nld-fra
441
+ metrics:
442
+ - name: BLEU
443
+ type: bleu
444
+ value: 28.7
445
+ - name: chr-F
446
+ type: chrf
447
+ value: 0.56608
448
+ - task:
449
+ name: Translation nld-por
450
+ type: translation
451
+ args: nld-por
452
+ dataset:
453
+ name: flores200-devtest
454
+ type: flores200-devtest
455
+ args: nld-por
456
+ metrics:
457
+ - name: BLEU
458
+ type: bleu
459
+ value: 26.7
460
+ - name: chr-F
461
+ type: chrf
462
+ value: 0.54728
463
+ - task:
464
+ name: Translation nld-spa
465
+ type: translation
466
+ args: nld-spa
467
+ dataset:
468
+ name: flores200-devtest
469
+ type: flores200-devtest
470
+ args: nld-spa
471
+ metrics:
472
+ - name: BLEU
473
+ type: bleu
474
+ value: 20.1
475
+ - name: chr-F
476
+ type: chrf
477
+ value: 0.49175
478
+ - task:
479
+ name: Translation tpi-deu
480
+ type: translation
481
+ args: tpi-deu
482
+ dataset:
483
+ name: flores200-devtest
484
+ type: flores200-devtest
485
+ args: tpi-deu
486
+ metrics:
487
+ - name: BLEU
488
+ type: bleu
489
+ value: 10.9
490
+ - name: chr-F
491
+ type: chrf
492
+ value: 0.40350
493
+ - task:
494
+ name: Translation tpi-eng
495
+ type: translation
496
+ args: tpi-eng
497
+ dataset:
498
+ name: flores200-devtest
499
+ type: flores200-devtest
500
+ args: tpi-eng
501
+ metrics:
502
+ - name: BLEU
503
+ type: bleu
504
+ value: 19.6
505
+ - name: chr-F
506
+ type: chrf
507
+ value: 0.48289
508
+ - task:
509
+ name: Translation tpi-fra
510
+ type: translation
511
+ args: tpi-fra
512
+ dataset:
513
+ name: flores200-devtest
514
+ type: flores200-devtest
515
+ args: tpi-fra
516
+ metrics:
517
+ - name: BLEU
518
+ type: bleu
519
+ value: 16.1
520
+ - name: chr-F
521
+ type: chrf
522
+ value: 0.43428
523
+ - task:
524
+ name: Translation tpi-por
525
+ type: translation
526
+ args: tpi-por
527
+ dataset:
528
+ name: flores200-devtest
529
+ type: flores200-devtest
530
+ args: tpi-por
531
+ metrics:
532
+ - name: BLEU
533
+ type: bleu
534
+ value: 15.4
535
+ - name: chr-F
536
+ type: chrf
537
+ value: 0.42966
538
+ - task:
539
+ name: Translation tpi-spa
540
+ type: translation
541
+ args: tpi-spa
542
+ dataset:
543
+ name: flores200-devtest
544
+ type: flores200-devtest
545
+ args: tpi-spa
546
+ metrics:
547
+ - name: BLEU
548
+ type: bleu
549
+ value: 12.9
550
+ - name: chr-F
551
+ type: chrf
552
+ value: 0.39730
553
+ - task:
554
+ name: Translation afr-deu
555
+ type: translation
556
+ args: afr-deu
557
+ dataset:
558
+ name: flores101-devtest
559
+ type: flores_101
560
+ args: afr deu devtest
561
+ metrics:
562
+ - name: BLEU
563
+ type: bleu
564
+ value: 28.4
565
+ - name: chr-F
566
+ type: chrf
567
+ value: 0.57831
568
+ - task:
569
+ name: Translation afr-eng
570
+ type: translation
571
+ args: afr-eng
572
+ dataset:
573
+ name: flores101-devtest
574
+ type: flores_101
575
+ args: afr eng devtest
576
+ metrics:
577
+ - name: BLEU
578
+ type: bleu
579
+ value: 53.8
580
+ - name: chr-F
581
+ type: chrf
582
+ value: 0.74272
583
+ - task:
584
+ name: Translation afr-fra
585
+ type: translation
586
+ args: afr-fra
587
+ dataset:
588
+ name: flores101-devtest
589
+ type: flores_101
590
+ args: afr fra devtest
591
+ metrics:
592
+ - name: BLEU
593
+ type: bleu
594
+ value: 36.4
595
+ - name: chr-F
596
+ type: chrf
597
+ value: 0.61936
598
+ - task:
599
+ name: Translation afr-por
600
+ type: translation
601
+ args: afr-por
602
+ dataset:
603
+ name: flores101-devtest
604
+ type: flores_101
605
+ args: afr por devtest
606
+ metrics:
607
+ - name: BLEU
608
+ type: bleu
609
+ value: 35.4
610
+ - name: chr-F
611
+ type: chrf
612
+ value: 0.61309
613
+ - task:
614
+ name: Translation afr-spa
615
+ type: translation
616
+ args: afr-spa
617
+ dataset:
618
+ name: flores101-devtest
619
+ type: flores_101
620
+ args: afr spa devtest
621
+ metrics:
622
+ - name: BLEU
623
+ type: bleu
624
+ value: 22.9
625
+ - name: chr-F
626
+ type: chrf
627
+ value: 0.51393
628
+ - task:
629
+ name: Translation deu-spa
630
+ type: translation
631
+ args: deu-spa
632
+ dataset:
633
+ name: flores101-devtest
634
+ type: flores_101
635
+ args: deu spa devtest
636
+ metrics:
637
+ - name: BLEU
638
+ type: bleu
639
+ value: 23.9
640
+ - name: chr-F
641
+ type: chrf
642
+ value: 0.52438
643
+ - task:
644
+ name: Translation eng-deu
645
+ type: translation
646
+ args: eng-deu
647
+ dataset:
648
+ name: flores101-devtest
649
+ type: flores_101
650
+ args: eng deu devtest
651
+ metrics:
652
+ - name: BLEU
653
+ type: bleu
654
+ value: 37.2
655
+ - name: chr-F
656
+ type: chrf
657
+ value: 0.64236
658
+ - task:
659
+ name: Translation eng-spa
660
+ type: translation
661
+ args: eng-spa
662
+ dataset:
663
+ name: flores101-devtest
664
+ type: flores_101
665
+ args: eng spa devtest
666
+ metrics:
667
+ - name: BLEU
668
+ type: bleu
669
+ value: 27.1
670
+ - name: chr-F
671
+ type: chrf
672
+ value: 0.55524
673
+ - task:
674
+ name: Translation nld-deu
675
+ type: translation
676
+ args: nld-deu
677
+ dataset:
678
+ name: flores101-devtest
679
+ type: flores_101
680
+ args: nld deu devtest
681
+ metrics:
682
+ - name: BLEU
683
+ type: bleu
684
+ value: 22.1
685
+ - name: chr-F
686
+ type: chrf
687
+ value: 0.53435
688
+ - task:
689
+ name: Translation nld-eng
690
+ type: translation
691
+ args: nld-eng
692
+ dataset:
693
+ name: flores101-devtest
694
+ type: flores_101
695
+ args: nld eng devtest
696
+ metrics:
697
+ - name: BLEU
698
+ type: bleu
699
+ value: 30.0
700
+ - name: chr-F
701
+ type: chrf
702
+ value: 0.58686
703
+ - task:
704
+ name: Translation nld-fra
705
+ type: translation
706
+ args: nld-fra
707
+ dataset:
708
+ name: flores101-devtest
709
+ type: flores_101
710
+ args: nld fra devtest
711
+ metrics:
712
+ - name: BLEU
713
+ type: bleu
714
+ value: 28.2
715
+ - name: chr-F
716
+ type: chrf
717
+ value: 0.56292
718
+ - task:
719
+ name: Translation deu-eng
720
+ type: translation
721
+ args: deu-eng
722
+ dataset:
723
+ name: generaltest2022
724
+ type: generaltest2022
725
+ args: deu-eng
726
+ metrics:
727
+ - name: BLEU
728
+ type: bleu
729
+ value: 31.0
730
+ - name: chr-F
731
+ type: chrf
732
+ value: 0.56042
733
+ - task:
734
+ name: Translation deu-fra
735
+ type: translation
736
+ args: deu-fra
737
+ dataset:
738
+ name: generaltest2022
739
+ type: generaltest2022
740
+ args: deu-fra
741
+ metrics:
742
+ - name: BLEU
743
+ type: bleu
744
+ value: 37.6
745
+ - name: chr-F
746
+ type: chrf
747
+ value: 0.61145
748
+ - task:
749
+ name: Translation eng-deu
750
+ type: translation
751
+ args: eng-deu
752
+ dataset:
753
+ name: generaltest2022
754
+ type: generaltest2022
755
+ args: eng-deu
756
+ metrics:
757
+ - name: BLEU
758
+ type: bleu
759
+ value: 32.5
760
+ - name: chr-F
761
+ type: chrf
762
+ value: 0.60090
763
+ - task:
764
+ name: Translation deu-eng
765
+ type: translation
766
+ args: deu-eng
767
+ dataset:
768
+ name: multi30k_test_2016_flickr
769
+ type: multi30k-2016_flickr
770
+ args: deu-eng
771
+ metrics:
772
+ - name: BLEU
773
+ type: bleu
774
+ value: 40.1
775
+ - name: chr-F
776
+ type: chrf
777
+ value: 0.60974
778
+ - task:
779
+ name: Translation deu-fra
780
+ type: translation
781
+ args: deu-fra
782
+ dataset:
783
+ name: multi30k_test_2016_flickr
784
+ type: multi30k-2016_flickr
785
+ args: deu-fra
786
+ metrics:
787
+ - name: BLEU
788
+ type: bleu
789
+ value: 38.8
790
+ - name: chr-F
791
+ type: chrf
792
+ value: 0.62493
793
+ - task:
794
+ name: Translation eng-deu
795
+ type: translation
796
+ args: eng-deu
797
+ dataset:
798
+ name: multi30k_test_2016_flickr
799
+ type: multi30k-2016_flickr
800
+ args: eng-deu
801
+ metrics:
802
+ - name: BLEU
803
+ type: bleu
804
+ value: 35.3
805
+ - name: chr-F
806
+ type: chrf
807
+ value: 0.64164
808
+ - task:
809
+ name: Translation eng-fra
810
+ type: translation
811
+ args: eng-fra
812
+ dataset:
813
+ name: multi30k_test_2016_flickr
814
+ type: multi30k-2016_flickr
815
+ args: eng-fra
816
+ metrics:
817
+ - name: BLEU
818
+ type: bleu
819
+ value: 50.7
820
+ - name: chr-F
821
+ type: chrf
822
+ value: 0.71137
823
+ - task:
824
+ name: Translation deu-eng
825
+ type: translation
826
+ args: deu-eng
827
+ dataset:
828
+ name: multi30k_test_2017_flickr
829
+ type: multi30k-2017_flickr
830
+ args: deu-eng
831
+ metrics:
832
+ - name: BLEU
833
+ type: bleu
834
+ value: 40.6
835
+ - name: chr-F
836
+ type: chrf
837
+ value: 0.63118
838
+ - task:
839
+ name: Translation deu-fra
840
+ type: translation
841
+ args: deu-fra
842
+ dataset:
843
+ name: multi30k_test_2017_flickr
844
+ type: multi30k-2017_flickr
845
+ args: deu-fra
846
+ metrics:
847
+ - name: BLEU
848
+ type: bleu
849
+ value: 37.0
850
+ - name: chr-F
851
+ type: chrf
852
+ value: 0.62614
853
+ - task:
854
+ name: Translation eng-deu
855
+ type: translation
856
+ args: eng-deu
857
+ dataset:
858
+ name: multi30k_test_2017_flickr
859
+ type: multi30k-2017_flickr
860
+ args: eng-deu
861
+ metrics:
862
+ - name: BLEU
863
+ type: bleu
864
+ value: 33.4
865
+ - name: chr-F
866
+ type: chrf
867
+ value: 0.62518
868
+ - task:
869
+ name: Translation eng-fra
870
+ type: translation
871
+ args: eng-fra
872
+ dataset:
873
+ name: multi30k_test_2017_flickr
874
+ type: multi30k-2017_flickr
875
+ args: eng-fra
876
+ metrics:
877
+ - name: BLEU
878
+ type: bleu
879
+ value: 50.3
880
+ - name: chr-F
881
+ type: chrf
882
+ value: 0.71402
883
+ - task:
884
+ name: Translation deu-eng
885
+ type: translation
886
+ args: deu-eng
887
+ dataset:
888
+ name: multi30k_test_2017_mscoco
889
+ type: multi30k-2017_mscoco
890
+ args: deu-eng
891
+ metrics:
892
+ - name: BLEU
893
+ type: bleu
894
+ value: 32.1
895
+ - name: chr-F
896
+ type: chrf
897
+ value: 0.55495
898
+ - task:
899
+ name: Translation deu-fra
900
+ type: translation
901
+ args: deu-fra
902
+ dataset:
903
+ name: multi30k_test_2017_mscoco
904
+ type: multi30k-2017_mscoco
905
+ args: deu-fra
906
+ metrics:
907
+ - name: BLEU
908
+ type: bleu
909
+ value: 34.7
910
+ - name: chr-F
911
+ type: chrf
912
+ value: 0.59307
913
+ - task:
914
+ name: Translation eng-deu
915
+ type: translation
916
+ args: eng-deu
917
+ dataset:
918
+ name: multi30k_test_2017_mscoco
919
+ type: multi30k-2017_mscoco
920
+ args: eng-deu
921
+ metrics:
922
+ - name: BLEU
923
+ type: bleu
924
+ value: 29.7
925
+ - name: chr-F
926
+ type: chrf
927
+ value: 0.58028
928
+ - task:
929
+ name: Translation eng-fra
930
+ type: translation
931
+ args: eng-fra
932
+ dataset:
933
+ name: multi30k_test_2017_mscoco
934
+ type: multi30k-2017_mscoco
935
+ args: eng-fra
936
+ metrics:
937
+ - name: BLEU
938
+ type: bleu
939
+ value: 54.7
940
+ - name: chr-F
941
+ type: chrf
942
+ value: 0.73637
943
+ - task:
944
+ name: Translation deu-eng
945
+ type: translation
946
+ args: deu-eng
947
+ dataset:
948
+ name: multi30k_test_2018_flickr
949
+ type: multi30k-2018_flickr
950
+ args: deu-eng
951
+ metrics:
952
+ - name: BLEU
953
+ type: bleu
954
+ value: 36.7
955
+ - name: chr-F
956
+ type: chrf
957
+ value: 0.59367
958
+ - task:
959
+ name: Translation deu-fra
960
+ type: translation
961
+ args: deu-fra
962
+ dataset:
963
+ name: multi30k_test_2018_flickr
964
+ type: multi30k-2018_flickr
965
+ args: deu-fra
966
+ metrics:
967
+ - name: BLEU
968
+ type: bleu
969
+ value: 31.3
970
+ - name: chr-F
971
+ type: chrf
972
+ value: 0.57388
973
+ - task:
974
+ name: Translation eng-deu
975
+ type: translation
976
+ args: eng-deu
977
+ dataset:
978
+ name: multi30k_test_2018_flickr
979
+ type: multi30k-2018_flickr
980
+ args: eng-deu
981
+ metrics:
982
+ - name: BLEU
983
+ type: bleu
984
+ value: 30.8
985
+ - name: chr-F
986
+ type: chrf
987
+ value: 0.59998
988
+ - task:
989
+ name: Translation eng-fra
990
+ type: translation
991
+ args: eng-fra
992
+ dataset:
993
+ name: multi30k_test_2018_flickr
994
+ type: multi30k-2018_flickr
995
+ args: eng-fra
996
+ metrics:
997
+ - name: BLEU
998
+ type: bleu
999
+ value: 41.6
1000
+ - name: chr-F
1001
+ type: chrf
1002
+ value: 0.65354
1003
+ - task:
1004
+ name: Translation eng-fra
1005
+ type: translation
1006
+ args: eng-fra
1007
+ dataset:
1008
+ name: newsdiscusstest2015
1009
+ type: newsdiscusstest2015
1010
+ args: eng-fra
1011
+ metrics:
1012
+ - name: BLEU
1013
+ type: bleu
1014
+ value: 37.7
1015
+ - name: chr-F
1016
+ type: chrf
1017
+ value: 0.63308
1018
+ - task:
1019
+ name: Translation deu-eng
1020
+ type: translation
1021
+ args: deu-eng
1022
+ dataset:
1023
+ name: newstestALL2020
1024
+ type: newstestALL2020
1025
+ args: deu-eng
1026
+ metrics:
1027
+ - name: BLEU
1028
+ type: bleu
1029
+ value: 34.5
1030
+ - name: chr-F
1031
+ type: chrf
1032
+ value: 0.60800
1033
+ - task:
1034
+ name: Translation eng-deu
1035
+ type: translation
1036
+ args: eng-deu
1037
+ dataset:
1038
+ name: newstestALL2020
1039
+ type: newstestALL2020
1040
+ args: eng-deu
1041
+ metrics:
1042
+ - name: BLEU
1043
+ type: bleu
1044
+ value: 31.7
1045
+ - name: chr-F
1046
+ type: chrf
1047
+ value: 0.60078
1048
+ - task:
1049
+ name: Translation afr-deu
1050
+ type: translation
1051
+ args: afr-deu
1052
+ dataset:
1053
+ name: ntrex128
1054
+ type: ntrex128
1055
+ args: afr-deu
1056
+ metrics:
1057
+ - name: BLEU
1058
+ type: bleu
1059
+ value: 25.8
1060
+ - name: chr-F
1061
+ type: chrf
1062
+ value: 0.55289
1063
+ - task:
1064
+ name: Translation afr-eng
1065
+ type: translation
1066
+ args: afr-eng
1067
+ dataset:
1068
+ name: ntrex128
1069
+ type: ntrex128
1070
+ args: afr-eng
1071
+ metrics:
1072
+ - name: BLEU
1073
+ type: bleu
1074
+ value: 51.8
1075
+ - name: chr-F
1076
+ type: chrf
1077
+ value: 0.72558
1078
+ - task:
1079
+ name: Translation afr-fra
1080
+ type: translation
1081
+ args: afr-fra
1082
+ dataset:
1083
+ name: ntrex128
1084
+ type: ntrex128
1085
+ args: afr-fra
1086
+ metrics:
1087
+ - name: BLEU
1088
+ type: bleu
1089
+ value: 29.3
1090
+ - name: chr-F
1091
+ type: chrf
1092
+ value: 0.56601
1093
+ - task:
1094
+ name: Translation afr-por
1095
+ type: translation
1096
+ args: afr-por
1097
+ dataset:
1098
+ name: ntrex128
1099
+ type: ntrex128
1100
+ args: afr-por
1101
+ metrics:
1102
+ - name: BLEU
1103
+ type: bleu
1104
+ value: 28.1
1105
+ - name: chr-F
1106
+ type: chrf
1107
+ value: 0.55396
1108
+ - task:
1109
+ name: Translation afr-spa
1110
+ type: translation
1111
+ args: afr-spa
1112
+ dataset:
1113
+ name: ntrex128
1114
+ type: ntrex128
1115
+ args: afr-spa
1116
+ metrics:
1117
+ - name: BLEU
1118
+ type: bleu
1119
+ value: 33.7
1120
+ - name: chr-F
1121
+ type: chrf
1122
+ value: 0.58558
1123
+ - task:
1124
+ name: Translation deu-eng
1125
+ type: translation
1126
+ args: deu-eng
1127
+ dataset:
1128
+ name: ntrex128
1129
+ type: ntrex128
1130
+ args: deu-eng
1131
+ metrics:
1132
+ - name: BLEU
1133
+ type: bleu
1134
+ value: 33.8
1135
+ - name: chr-F
1136
+ type: chrf
1137
+ value: 0.61722
1138
+ - task:
1139
+ name: Translation deu-fra
1140
+ type: translation
1141
+ args: deu-fra
1142
+ dataset:
1143
+ name: ntrex128
1144
+ type: ntrex128
1145
+ args: deu-fra
1146
+ metrics:
1147
+ - name: BLEU
1148
+ type: bleu
1149
+ value: 28.6
1150
+ - name: chr-F
1151
+ type: chrf
1152
+ value: 0.55908
1153
+ - task:
1154
+ name: Translation deu-por
1155
+ type: translation
1156
+ args: deu-por
1157
+ dataset:
1158
+ name: ntrex128
1159
+ type: ntrex128
1160
+ args: deu-por
1161
+ metrics:
1162
+ - name: BLEU
1163
+ type: bleu
1164
+ value: 25.7
1165
+ - name: chr-F
1166
+ type: chrf
1167
+ value: 0.54059
1168
+ - task:
1169
+ name: Translation deu-spa
1170
+ type: translation
1171
+ args: deu-spa
1172
+ dataset:
1173
+ name: ntrex128
1174
+ type: ntrex128
1175
+ args: deu-spa
1176
+ metrics:
1177
+ - name: BLEU
1178
+ type: bleu
1179
+ value: 30.8
1180
+ - name: chr-F
1181
+ type: chrf
1182
+ value: 0.56887
1183
+ - task:
1184
+ name: Translation eng-deu
1185
+ type: translation
1186
+ args: eng-deu
1187
+ dataset:
1188
+ name: ntrex128
1189
+ type: ntrex128
1190
+ args: eng-deu
1191
+ metrics:
1192
+ - name: BLEU
1193
+ type: bleu
1194
+ value: 29.8
1195
+ - name: chr-F
1196
+ type: chrf
1197
+ value: 0.58492
1198
+ - task:
1199
+ name: Translation eng-fra
1200
+ type: translation
1201
+ args: eng-fra
1202
+ dataset:
1203
+ name: ntrex128
1204
+ type: ntrex128
1205
+ args: eng-fra
1206
+ metrics:
1207
+ - name: BLEU
1208
+ type: bleu
1209
+ value: 35.2
1210
+ - name: chr-F
1211
+ type: chrf
1212
+ value: 0.61349
1213
+ - task:
1214
+ name: Translation eng-por
1215
+ type: translation
1216
+ args: eng-por
1217
+ dataset:
1218
+ name: ntrex128
1219
+ type: ntrex128
1220
+ args: eng-por
1221
+ metrics:
1222
+ - name: BLEU
1223
+ type: bleu
1224
+ value: 33.4
1225
+ - name: chr-F
1226
+ type: chrf
1227
+ value: 0.59785
1228
+ - task:
1229
+ name: Translation eng-spa
1230
+ type: translation
1231
+ args: eng-spa
1232
+ dataset:
1233
+ name: ntrex128
1234
+ type: ntrex128
1235
+ args: eng-spa
1236
+ metrics:
1237
+ - name: BLEU
1238
+ type: bleu
1239
+ value: 40.1
1240
+ - name: chr-F
1241
+ type: chrf
1242
+ value: 0.63935
1243
+ - task:
1244
+ name: Translation ltz-deu
1245
+ type: translation
1246
+ args: ltz-deu
1247
+ dataset:
1248
+ name: ntrex128
1249
+ type: ntrex128
1250
+ args: ltz-deu
1251
+ metrics:
1252
+ - name: BLEU
1253
+ type: bleu
1254
+ value: 21.9
1255
+ - name: chr-F
1256
+ type: chrf
1257
+ value: 0.51469
1258
+ - task:
1259
+ name: Translation ltz-eng
1260
+ type: translation
1261
+ args: ltz-eng
1262
+ dataset:
1263
+ name: ntrex128
1264
+ type: ntrex128
1265
+ args: ltz-eng
1266
+ metrics:
1267
+ - name: BLEU
1268
+ type: bleu
1269
+ value: 32.4
1270
+ - name: chr-F
1271
+ type: chrf
1272
+ value: 0.58627
1273
+ - task:
1274
+ name: Translation ltz-fra
1275
+ type: translation
1276
+ args: ltz-fra
1277
+ dataset:
1278
+ name: ntrex128
1279
+ type: ntrex128
1280
+ args: ltz-fra
1281
+ metrics:
1282
+ - name: BLEU
1283
+ type: bleu
1284
+ value: 22.8
1285
+ - name: chr-F
1286
+ type: chrf
1287
+ value: 0.50491
1288
+ - task:
1289
+ name: Translation ltz-por
1290
+ type: translation
1291
+ args: ltz-por
1292
+ dataset:
1293
+ name: ntrex128
1294
+ type: ntrex128
1295
+ args: ltz-por
1296
+ metrics:
1297
+ - name: BLEU
1298
+ type: bleu
1299
+ value: 18.7
1300
+ - name: chr-F
1301
+ type: chrf
1302
+ value: 0.45364
1303
+ - task:
1304
+ name: Translation ltz-spa
1305
+ type: translation
1306
+ args: ltz-spa
1307
+ dataset:
1308
+ name: ntrex128
1309
+ type: ntrex128
1310
+ args: ltz-spa
1311
+ metrics:
1312
+ - name: BLEU
1313
+ type: bleu
1314
+ value: 21.6
1315
+ - name: chr-F
1316
+ type: chrf
1317
+ value: 0.47568
1318
+ - task:
1319
+ name: Translation nld-deu
1320
+ type: translation
1321
+ args: nld-deu
1322
+ dataset:
1323
+ name: ntrex128
1324
+ type: ntrex128
1325
+ args: nld-deu
1326
+ metrics:
1327
+ - name: BLEU
1328
+ type: bleu
1329
+ value: 25.7
1330
+ - name: chr-F
1331
+ type: chrf
1332
+ value: 0.55943
1333
+ - task:
1334
+ name: Translation nld-eng
1335
+ type: translation
1336
+ args: nld-eng
1337
+ dataset:
1338
+ name: ntrex128
1339
+ type: ntrex128
1340
+ args: nld-eng
1341
+ metrics:
1342
+ - name: BLEU
1343
+ type: bleu
1344
+ value: 36.1
1345
+ - name: chr-F
1346
+ type: chrf
1347
+ value: 0.63470
1348
+ - task:
1349
+ name: Translation nld-fra
1350
+ type: translation
1351
+ args: nld-fra
1352
+ dataset:
1353
+ name: ntrex128
1354
+ type: ntrex128
1355
+ args: nld-fra
1356
+ metrics:
1357
+ - name: BLEU
1358
+ type: bleu
1359
+ value: 27.5
1360
+ - name: chr-F
1361
+ type: chrf
1362
+ value: 0.55832
1363
+ - task:
1364
+ name: Translation nld-por
1365
+ type: translation
1366
+ args: nld-por
1367
+ dataset:
1368
+ name: ntrex128
1369
+ type: ntrex128
1370
+ args: nld-por
1371
+ metrics:
1372
+ - name: BLEU
1373
+ type: bleu
1374
+ value: 27.3
1375
+ - name: chr-F
1376
+ type: chrf
1377
+ value: 0.54714
1378
+ - task:
1379
+ name: Translation nld-spa
1380
+ type: translation
1381
+ args: nld-spa
1382
+ dataset:
1383
+ name: ntrex128
1384
+ type: ntrex128
1385
+ args: nld-spa
1386
+ metrics:
1387
+ - name: BLEU
1388
+ type: bleu
1389
+ value: 32.1
1390
+ - name: chr-F
1391
+ type: chrf
1392
+ value: 0.57692
1393
+ - task:
1394
+ name: Translation afr-deu
1395
+ type: translation
1396
+ args: afr-deu
1397
+ dataset:
1398
+ name: tatoeba-test-v2021-08-07
1399
+ type: tatoeba_mt
1400
+ args: afr-deu
1401
+ metrics:
1402
+ - name: BLEU
1403
+ type: bleu
1404
+ value: 48.8
1405
+ - name: chr-F
1406
+ type: chrf
1407
+ value: 0.68492
1408
+ - task:
1409
+ name: Translation afr-eng
1410
+ type: translation
1411
+ args: afr-eng
1412
+ dataset:
1413
+ name: tatoeba-test-v2021-08-07
1414
+ type: tatoeba_mt
1415
+ args: afr-eng
1416
+ metrics:
1417
+ - name: BLEU
1418
+ type: bleu
1419
+ value: 59.6
1420
+ - name: chr-F
1421
+ type: chrf
1422
+ value: 0.72943
1423
+ - task:
1424
+ name: Translation afr-spa
1425
+ type: translation
1426
+ args: afr-spa
1427
+ dataset:
1428
+ name: tatoeba-test-v2021-08-07
1429
+ type: tatoeba_mt
1430
+ args: afr-spa
1431
+ metrics:
1432
+ - name: BLEU
1433
+ type: bleu
1434
+ value: 58.4
1435
+ - name: chr-F
1436
+ type: chrf
1437
+ value: 0.72793
1438
+ - task:
1439
+ name: Translation deu-deu
1440
+ type: translation
1441
+ args: deu-deu
1442
+ dataset:
1443
+ name: tatoeba-test-v2021-08-07
1444
+ type: tatoeba_mt
1445
+ args: deu-deu
1446
+ metrics:
1447
+ - name: BLEU
1448
+ type: bleu
1449
+ value: 34.8
1450
+ - name: chr-F
1451
+ type: chrf
1452
+ value: 0.59840
1453
+ - task:
1454
+ name: Translation deu-eng
1455
+ type: translation
1456
+ args: deu-eng
1457
+ dataset:
1458
+ name: tatoeba-test-v2021-08-07
1459
+ type: tatoeba_mt
1460
+ args: deu-eng
1461
+ metrics:
1462
+ - name: BLEU
1463
+ type: bleu
1464
+ value: 48.5
1465
+ - name: chr-F
1466
+ type: chrf
1467
+ value: 0.65957
1468
+ - task:
1469
+ name: Translation deu-fra
1470
+ type: translation
1471
+ args: deu-fra
1472
+ dataset:
1473
+ name: tatoeba-test-v2021-08-07
1474
+ type: tatoeba_mt
1475
+ args: deu-fra
1476
+ metrics:
1477
+ - name: BLEU
1478
+ type: bleu
1479
+ value: 50.2
1480
+ - name: chr-F
1481
+ type: chrf
1482
+ value: 0.68054
1483
+ - task:
1484
+ name: Translation deu-por
1485
+ type: translation
1486
+ args: deu-por
1487
+ dataset:
1488
+ name: tatoeba-test-v2021-08-07
1489
+ type: tatoeba_mt
1490
+ args: deu-por
1491
+ metrics:
1492
+ - name: BLEU
1493
+ type: bleu
1494
+ value: 42.8
1495
+ - name: chr-F
1496
+ type: chrf
1497
+ value: 0.63368
1498
+ - task:
1499
+ name: Translation deu-spa
1500
+ type: translation
1501
+ args: deu-spa
1502
+ dataset:
1503
+ name: tatoeba-test-v2021-08-07
1504
+ type: tatoeba_mt
1505
+ args: deu-spa
1506
+ metrics:
1507
+ - name: BLEU
1508
+ type: bleu
1509
+ value: 49.3
1510
+ - name: chr-F
1511
+ type: chrf
1512
+ value: 0.68198
1513
+ - task:
1514
+ name: Translation eng-deu
1515
+ type: translation
1516
+ args: eng-deu
1517
+ dataset:
1518
+ name: tatoeba-test-v2021-08-07
1519
+ type: tatoeba_mt
1520
+ args: eng-deu
1521
+ metrics:
1522
+ - name: BLEU
1523
+ type: bleu
1524
+ value: 40.5
1525
+ - name: chr-F
1526
+ type: chrf
1527
+ value: 0.62497
1528
+ - task:
1529
+ name: Translation eng-eng
1530
+ type: translation
1531
+ args: eng-eng
1532
+ dataset:
1533
+ name: tatoeba-test-v2021-08-07
1534
+ type: tatoeba_mt
1535
+ args: eng-eng
1536
+ metrics:
1537
+ - name: BLEU
1538
+ type: bleu
1539
+ value: 57.3
1540
+ - name: chr-F
1541
+ type: chrf
1542
+ value: 0.79878
1543
+ - task:
1544
+ name: Translation eng-fra
1545
+ type: translation
1546
+ args: eng-fra
1547
+ dataset:
1548
+ name: tatoeba-test-v2021-08-07
1549
+ type: tatoeba_mt
1550
+ args: eng-fra
1551
+ metrics:
1552
+ - name: BLEU
1553
+ type: bleu
1554
+ value: 50.9
1555
+ - name: chr-F
1556
+ type: chrf
1557
+ value: 0.68567
1558
+ - task:
1559
+ name: Translation eng-por
1560
+ type: translation
1561
+ args: eng-por
1562
+ dataset:
1563
+ name: tatoeba-test-v2021-08-07
1564
+ type: tatoeba_mt
1565
+ args: eng-por
1566
+ metrics:
1567
+ - name: BLEU
1568
+ type: bleu
1569
+ value: 53.4
1570
+ - name: chr-F
1571
+ type: chrf
1572
+ value: 0.72204
1573
+ - task:
1574
+ name: Translation eng-spa
1575
+ type: translation
1576
+ args: eng-spa
1577
+ dataset:
1578
+ name: tatoeba-test-v2021-08-07
1579
+ type: tatoeba_mt
1580
+ args: eng-spa
1581
+ metrics:
1582
+ - name: BLEU
1583
+ type: bleu
1584
+ value: 55.3
1585
+ - name: chr-F
1586
+ type: chrf
1587
+ value: 0.72539
1588
+ - task:
1589
+ name: Translation fry-eng
1590
+ type: translation
1591
+ args: fry-eng
1592
+ dataset:
1593
+ name: tatoeba-test-v2021-08-07
1594
+ type: tatoeba_mt
1595
+ args: fry-eng
1596
+ metrics:
1597
+ - name: BLEU
1598
+ type: bleu
1599
+ value: 37.0
1600
+ - name: chr-F
1601
+ type: chrf
1602
+ value: 0.55137
1603
+ - task:
1604
+ name: Translation gos-deu
1605
+ type: translation
1606
+ args: gos-deu
1607
+ dataset:
1608
+ name: tatoeba-test-v2021-08-07
1609
+ type: tatoeba_mt
1610
+ args: gos-deu
1611
+ metrics:
1612
+ - name: BLEU
1613
+ type: bleu
1614
+ value: 24.7
1615
+ - name: chr-F
1616
+ type: chrf
1617
+ value: 0.46120
1618
+ - task:
1619
+ name: Translation gos-eng
1620
+ type: translation
1621
+ args: gos-eng
1622
+ dataset:
1623
+ name: tatoeba-test-v2021-08-07
1624
+ type: tatoeba_mt
1625
+ args: gos-eng
1626
+ metrics:
1627
+ - name: BLEU
1628
+ type: bleu
1629
+ value: 22.3
1630
+ - name: chr-F
1631
+ type: chrf
1632
+ value: 0.38628
1633
+ - task:
1634
+ name: Translation gsw-eng
1635
+ type: translation
1636
+ args: gsw-eng
1637
+ dataset:
1638
+ name: tatoeba-test-v2021-08-07
1639
+ type: tatoeba_mt
1640
+ args: gsw-eng
1641
+ metrics:
1642
+ - name: BLEU
1643
+ type: bleu
1644
+ value: 27.5
1645
+ - name: chr-F
1646
+ type: chrf
1647
+ value: 0.43003
1648
+ - task:
1649
+ name: Translation ltz-deu
1650
+ type: translation
1651
+ args: ltz-deu
1652
+ dataset:
1653
+ name: tatoeba-test-v2021-08-07
1654
+ type: tatoeba_mt
1655
+ args: ltz-deu
1656
+ metrics:
1657
+ - name: BLEU
1658
+ type: bleu
1659
+ value: 32.0
1660
+ - name: chr-F
1661
+ type: chrf
1662
+ value: 0.48474
1663
+ - task:
1664
+ name: Translation ltz-eng
1665
+ type: translation
1666
+ args: ltz-eng
1667
+ dataset:
1668
+ name: tatoeba-test-v2021-08-07
1669
+ type: tatoeba_mt
1670
+ args: ltz-eng
1671
+ metrics:
1672
+ - name: BLEU
1673
+ type: bleu
1674
+ value: 56.4
1675
+ - name: chr-F
1676
+ type: chrf
1677
+ value: 0.65366
1678
+ - task:
1679
+ name: Translation multi-multi
1680
+ type: translation
1681
+ args: multi-multi
1682
+ dataset:
1683
+ name: tatoeba-test-v2020-07-28-v2023-09-26
1684
+ type: tatoeba_mt
1685
+ args: multi-multi
1686
+ metrics:
1687
+ - name: BLEU
1688
+ type: bleu
1689
+ value: 50.3
1690
+ - name: chr-F
1691
+ type: chrf
1692
+ value: 0.68713
1693
+ - task:
1694
+ name: Translation nds-deu
1695
+ type: translation
1696
+ args: nds-deu
1697
+ dataset:
1698
+ name: tatoeba-test-v2021-08-07
1699
+ type: tatoeba_mt
1700
+ args: nds-deu
1701
+ metrics:
1702
+ - name: BLEU
1703
+ type: bleu
1704
+ value: 45.9
1705
+ - name: chr-F
1706
+ type: chrf
1707
+ value: 0.65251
1708
+ - task:
1709
+ name: Translation nds-eng
1710
+ type: translation
1711
+ args: nds-eng
1712
+ dataset:
1713
+ name: tatoeba-test-v2021-08-07
1714
+ type: tatoeba_mt
1715
+ args: nds-eng
1716
+ metrics:
1717
+ - name: BLEU
1718
+ type: bleu
1719
+ value: 44.7
1720
+ - name: chr-F
1721
+ type: chrf
1722
+ value: 0.61858
1723
+ - task:
1724
+ name: Translation nds-fra
1725
+ type: translation
1726
+ args: nds-fra
1727
+ dataset:
1728
+ name: tatoeba-test-v2021-08-07
1729
+ type: tatoeba_mt
1730
+ args: nds-fra
1731
+ metrics:
1732
+ - name: BLEU
1733
+ type: bleu
1734
+ value: 43.9
1735
+ - name: chr-F
1736
+ type: chrf
1737
+ value: 0.60412
1738
+ - task:
1739
+ name: Translation nds-por
1740
+ type: translation
1741
+ args: nds-por
1742
+ dataset:
1743
+ name: tatoeba-test-v2021-08-07
1744
+ type: tatoeba_mt
1745
+ args: nds-por
1746
+ metrics:
1747
+ - name: BLEU
1748
+ type: bleu
1749
+ value: 39.5
1750
+ - name: chr-F
1751
+ type: chrf
1752
+ value: 0.58778
1753
+ - task:
1754
+ name: Translation nds-spa
1755
+ type: translation
1756
+ args: nds-spa
1757
+ dataset:
1758
+ name: tatoeba-test-v2021-08-07
1759
+ type: tatoeba_mt
1760
+ args: nds-spa
1761
+ metrics:
1762
+ - name: BLEU
1763
+ type: bleu
1764
+ value: 43.9
1765
+ - name: chr-F
1766
+ type: chrf
1767
+ value: 0.63404
1768
+ - task:
1769
+ name: Translation nld-deu
1770
+ type: translation
1771
+ args: nld-deu
1772
+ dataset:
1773
+ name: tatoeba-test-v2021-08-07
1774
+ type: tatoeba_mt
1775
+ args: nld-deu
1776
+ metrics:
1777
+ - name: BLEU
1778
+ type: bleu
1779
+ value: 55.7
1780
+ - name: chr-F
1781
+ type: chrf
1782
+ value: 0.72998
1783
+ - task:
1784
+ name: Translation nld-eng
1785
+ type: translation
1786
+ args: nld-eng
1787
+ dataset:
1788
+ name: tatoeba-test-v2021-08-07
1789
+ type: tatoeba_mt
1790
+ args: nld-eng
1791
+ metrics:
1792
+ - name: BLEU
1793
+ type: bleu
1794
+ value: 60.9
1795
+ - name: chr-F
1796
+ type: chrf
1797
+ value: 0.74362
1798
+ - task:
1799
+ name: Translation nld-fra
1800
+ type: translation
1801
+ args: nld-fra
1802
+ dataset:
1803
+ name: tatoeba-test-v2021-08-07
1804
+ type: tatoeba_mt
1805
+ args: nld-fra
1806
+ metrics:
1807
+ - name: BLEU
1808
+ type: bleu
1809
+ value: 48.0
1810
+ - name: chr-F
1811
+ type: chrf
1812
+ value: 0.68461
1813
+ - task:
1814
+ name: Translation nld-por
1815
+ type: translation
1816
+ args: nld-por
1817
+ dataset:
1818
+ name: tatoeba-test-v2021-08-07
1819
+ type: tatoeba_mt
1820
+ args: nld-por
1821
+ metrics:
1822
+ - name: BLEU
1823
+ type: bleu
1824
+ value: 49.3
1825
+ - name: chr-F
1826
+ type: chrf
1827
+ value: 0.68798
1828
+ - task:
1829
+ name: Translation nld-spa
1830
+ type: translation
1831
+ args: nld-spa
1832
+ dataset:
1833
+ name: tatoeba-test-v2021-08-07
1834
+ type: tatoeba_mt
1835
+ args: nld-spa
1836
+ metrics:
1837
+ - name: BLEU
1838
+ type: bleu
1839
+ value: 51.6
1840
+ - name: chr-F
1841
+ type: chrf
1842
+ value: 0.69971
1843
+ - task:
1844
+ name: Translation yid-eng
1845
+ type: translation
1846
+ args: yid-eng
1847
+ dataset:
1848
+ name: tatoeba-test-v2021-08-07
1849
+ type: tatoeba_mt
1850
+ args: yid-eng
1851
+ metrics:
1852
+ - name: BLEU
1853
+ type: bleu
1854
+ value: 31.5
1855
+ - name: chr-F
1856
+ type: chrf
1857
+ value: 0.49807
1858
+ - task:
1859
+ name: Translation yid-fra
1860
+ type: translation
1861
+ args: yid-fra
1862
+ dataset:
1863
+ name: tatoeba-test-v2021-08-07
1864
+ type: tatoeba_mt
1865
+ args: yid-fra
1866
+ metrics:
1867
+ - name: BLEU
1868
+ type: bleu
1869
+ value: 31.9
1870
+ - name: chr-F
1871
+ type: chrf
1872
+ value: 0.54147
1873
+ - task:
1874
+ name: Translation eng-fra
1875
+ type: translation
1876
+ args: eng-fra
1877
+ dataset:
1878
+ name: tico19-test
1879
+ type: tico19-test
1880
+ args: eng-fra
1881
+ metrics:
1882
+ - name: BLEU
1883
+ type: bleu
1884
+ value: 39.5
1885
+ - name: chr-F
1886
+ type: chrf
1887
+ value: 0.62559
1888
+ - task:
1889
+ name: Translation eng-por
1890
+ type: translation
1891
+ args: eng-por
1892
+ dataset:
1893
+ name: tico19-test
1894
+ type: tico19-test
1895
+ args: eng-por
1896
+ metrics:
1897
+ - name: BLEU
1898
+ type: bleu
1899
+ value: 49.8
1900
+ - name: chr-F
1901
+ type: chrf
1902
+ value: 0.72765
1903
+ - task:
1904
+ name: Translation eng-spa
1905
+ type: translation
1906
+ args: eng-spa
1907
+ dataset:
1908
+ name: tico19-test
1909
+ type: tico19-test
1910
+ args: eng-spa
1911
+ metrics:
1912
+ - name: BLEU
1913
+ type: bleu
1914
+ value: 51.6
1915
+ - name: chr-F
1916
+ type: chrf
1917
+ value: 0.72905
1918
+ - task:
1919
+ name: Translation deu-eng
1920
+ type: translation
1921
+ args: deu-eng
1922
+ dataset:
1923
+ name: newstest2008
1924
+ type: wmt-2008-news
1925
+ args: deu-eng
1926
+ metrics:
1927
+ - name: BLEU
1928
+ type: bleu
1929
+ value: 27.0
1930
+ - name: chr-F
1931
+ type: chrf
1932
+ value: 0.54434
1933
+ - task:
1934
+ name: Translation deu-fra
1935
+ type: translation
1936
+ args: deu-fra
1937
+ dataset:
1938
+ name: newstest2008
1939
+ type: wmt-2008-news
1940
+ args: deu-fra
1941
+ metrics:
1942
+ - name: BLEU
1943
+ type: bleu
1944
+ value: 26.2
1945
+ - name: chr-F
1946
+ type: chrf
1947
+ value: 0.55076
1948
+ - task:
1949
+ name: Translation deu-spa
1950
+ type: translation
1951
+ args: deu-spa
1952
+ dataset:
1953
+ name: newstest2008
1954
+ type: wmt-2008-news
1955
+ args: deu-spa
1956
+ metrics:
1957
+ - name: BLEU
1958
+ type: bleu
1959
+ value: 25.6
1960
+ - name: chr-F
1961
+ type: chrf
1962
+ value: 0.54056
1963
+ - task:
1964
+ name: Translation eng-deu
1965
+ type: translation
1966
+ args: eng-deu
1967
+ dataset:
1968
+ name: newstest2008
1969
+ type: wmt-2008-news
1970
+ args: eng-deu
1971
+ metrics:
1972
+ - name: BLEU
1973
+ type: bleu
1974
+ value: 23.0
1975
+ - name: chr-F
1976
+ type: chrf
1977
+ value: 0.52906
1978
+ - task:
1979
+ name: Translation eng-fra
1980
+ type: translation
1981
+ args: eng-fra
1982
+ dataset:
1983
+ name: newstest2008
1984
+ type: wmt-2008-news
1985
+ args: eng-fra
1986
+ metrics:
1987
+ - name: BLEU
1988
+ type: bleu
1989
+ value: 26.8
1990
+ - name: chr-F
1991
+ type: chrf
1992
+ value: 0.55247
1993
+ - task:
1994
+ name: Translation eng-spa
1995
+ type: translation
1996
+ args: eng-spa
1997
+ dataset:
1998
+ name: newstest2008
1999
+ type: wmt-2008-news
2000
+ args: eng-spa
2001
+ metrics:
2002
+ - name: BLEU
2003
+ type: bleu
2004
+ value: 29.6
2005
+ - name: chr-F
2006
+ type: chrf
2007
+ value: 0.56423
2008
+ - task:
2009
+ name: Translation deu-eng
2010
+ type: translation
2011
+ args: deu-eng
2012
+ dataset:
2013
+ name: newstest2009
2014
+ type: wmt-2009-news
2015
+ args: deu-eng
2016
+ metrics:
2017
+ - name: BLEU
2018
+ type: bleu
2019
+ value: 26.7
2020
+ - name: chr-F
2021
+ type: chrf
2022
+ value: 0.53972
2023
+ - task:
2024
+ name: Translation deu-fra
2025
+ type: translation
2026
+ args: deu-fra
2027
+ dataset:
2028
+ name: newstest2009
2029
+ type: wmt-2009-news
2030
+ args: deu-fra
2031
+ metrics:
2032
+ - name: BLEU
2033
+ type: bleu
2034
+ value: 25.6
2035
+ - name: chr-F
2036
+ type: chrf
2037
+ value: 0.53975
2038
+ - task:
2039
+ name: Translation deu-spa
2040
+ type: translation
2041
+ args: deu-spa
2042
+ dataset:
2043
+ name: newstest2009
2044
+ type: wmt-2009-news
2045
+ args: deu-spa
2046
+ metrics:
2047
+ - name: BLEU
2048
+ type: bleu
2049
+ value: 25.6
2050
+ - name: chr-F
2051
+ type: chrf
2052
+ value: 0.53677
2053
+ - task:
2054
+ name: Translation eng-deu
2055
+ type: translation
2056
+ args: eng-deu
2057
+ dataset:
2058
+ name: newstest2009
2059
+ type: wmt-2009-news
2060
+ args: eng-deu
2061
+ metrics:
2062
+ - name: BLEU
2063
+ type: bleu
2064
+ value: 22.1
2065
+ - name: chr-F
2066
+ type: chrf
2067
+ value: 0.53097
2068
+ - task:
2069
+ name: Translation eng-fra
2070
+ type: translation
2071
+ args: eng-fra
2072
+ dataset:
2073
+ name: newstest2009
2074
+ type: wmt-2009-news
2075
+ args: eng-fra
2076
+ metrics:
2077
+ - name: BLEU
2078
+ type: bleu
2079
+ value: 29.1
2080
+ - name: chr-F
2081
+ type: chrf
2082
+ value: 0.57542
2083
+ - task:
2084
+ name: Translation eng-spa
2085
+ type: translation
2086
+ args: eng-spa
2087
+ dataset:
2088
+ name: newstest2009
2089
+ type: wmt-2009-news
2090
+ args: eng-spa
2091
+ metrics:
2092
+ - name: BLEU
2093
+ type: bleu
2094
+ value: 29.8
2095
+ - name: chr-F
2096
+ type: chrf
2097
+ value: 0.57733
2098
+ - task:
2099
+ name: Translation deu-eng
2100
+ type: translation
2101
+ args: deu-eng
2102
+ dataset:
2103
+ name: newstest2010
2104
+ type: wmt-2010-news
2105
+ args: deu-eng
2106
+ metrics:
2107
+ - name: BLEU
2108
+ type: bleu
2109
+ value: 30.2
2110
+ - name: chr-F
2111
+ type: chrf
2112
+ value: 0.58278
2113
+ - task:
2114
+ name: Translation deu-fra
2115
+ type: translation
2116
+ args: deu-fra
2117
+ dataset:
2118
+ name: newstest2010
2119
+ type: wmt-2010-news
2120
+ args: deu-fra
2121
+ metrics:
2122
+ - name: BLEU
2123
+ type: bleu
2124
+ value: 29.0
2125
+ - name: chr-F
2126
+ type: chrf
2127
+ value: 0.57876
2128
+ - task:
2129
+ name: Translation deu-spa
2130
+ type: translation
2131
+ args: deu-spa
2132
+ dataset:
2133
+ name: newstest2010
2134
+ type: wmt-2010-news
2135
+ args: deu-spa
2136
+ metrics:
2137
+ - name: BLEU
2138
+ type: bleu
2139
+ value: 32.6
2140
+ - name: chr-F
2141
+ type: chrf
2142
+ value: 0.59402
2143
+ - task:
2144
+ name: Translation eng-deu
2145
+ type: translation
2146
+ args: eng-deu
2147
+ dataset:
2148
+ name: newstest2010
2149
+ type: wmt-2010-news
2150
+ args: eng-deu
2151
+ metrics:
2152
+ - name: BLEU
2153
+ type: bleu
2154
+ value: 25.3
2155
+ - name: chr-F
2156
+ type: chrf
2157
+ value: 0.54587
2158
+ - task:
2159
+ name: Translation eng-fra
2160
+ type: translation
2161
+ args: eng-fra
2162
+ dataset:
2163
+ name: newstest2010
2164
+ type: wmt-2010-news
2165
+ args: eng-fra
2166
+ metrics:
2167
+ - name: BLEU
2168
+ type: bleu
2169
+ value: 32.0
2170
+ - name: chr-F
2171
+ type: chrf
2172
+ value: 0.59460
2173
+ - task:
2174
+ name: Translation eng-spa
2175
+ type: translation
2176
+ args: eng-spa
2177
+ dataset:
2178
+ name: newstest2010
2179
+ type: wmt-2010-news
2180
+ args: eng-spa
2181
+ metrics:
2182
+ - name: BLEU
2183
+ type: bleu
2184
+ value: 36.3
2185
+ - name: chr-F
2186
+ type: chrf
2187
+ value: 0.61861
2188
+ - task:
2189
+ name: Translation deu-eng
2190
+ type: translation
2191
+ args: deu-eng
2192
+ dataset:
2193
+ name: newstest2011
2194
+ type: wmt-2011-news
2195
+ args: deu-eng
2196
+ metrics:
2197
+ - name: BLEU
2198
+ type: bleu
2199
+ value: 26.8
2200
+ - name: chr-F
2201
+ type: chrf
2202
+ value: 0.55074
2203
+ - task:
2204
+ name: Translation deu-fra
2205
+ type: translation
2206
+ args: deu-fra
2207
+ dataset:
2208
+ name: newstest2011
2209
+ type: wmt-2011-news
2210
+ args: deu-fra
2211
+ metrics:
2212
+ - name: BLEU
2213
+ type: bleu
2214
+ value: 27.4
2215
+ - name: chr-F
2216
+ type: chrf
2217
+ value: 0.55879
2218
+ - task:
2219
+ name: Translation deu-spa
2220
+ type: translation
2221
+ args: deu-spa
2222
+ dataset:
2223
+ name: newstest2011
2224
+ type: wmt-2011-news
2225
+ args: deu-spa
2226
+ metrics:
2227
+ - name: BLEU
2228
+ type: bleu
2229
+ value: 30.2
2230
+ - name: chr-F
2231
+ type: chrf
2232
+ value: 0.56593
2233
+ - task:
2234
+ name: Translation eng-deu
2235
+ type: translation
2236
+ args: eng-deu
2237
+ dataset:
2238
+ name: newstest2011
2239
+ type: wmt-2011-news
2240
+ args: eng-deu
2241
+ metrics:
2242
+ - name: BLEU
2243
+ type: bleu
2244
+ value: 22.7
2245
+ - name: chr-F
2246
+ type: chrf
2247
+ value: 0.52619
2248
+ - task:
2249
+ name: Translation eng-fra
2250
+ type: translation
2251
+ args: eng-fra
2252
+ dataset:
2253
+ name: newstest2011
2254
+ type: wmt-2011-news
2255
+ args: eng-fra
2256
+ metrics:
2257
+ - name: BLEU
2258
+ type: bleu
2259
+ value: 34.1
2260
+ - name: chr-F
2261
+ type: chrf
2262
+ value: 0.60960
2263
+ - task:
2264
+ name: Translation eng-spa
2265
+ type: translation
2266
+ args: eng-spa
2267
+ dataset:
2268
+ name: newstest2011
2269
+ type: wmt-2011-news
2270
+ args: eng-spa
2271
+ metrics:
2272
+ - name: BLEU
2273
+ type: bleu
2274
+ value: 38.5
2275
+ - name: chr-F
2276
+ type: chrf
2277
+ value: 0.62056
2278
+ - task:
2279
+ name: Translation deu-eng
2280
+ type: translation
2281
+ args: deu-eng
2282
+ dataset:
2283
+ name: newstest2012
2284
+ type: wmt-2012-news
2285
+ args: deu-eng
2286
+ metrics:
2287
+ - name: BLEU
2288
+ type: bleu
2289
+ value: 28.4
2290
+ - name: chr-F
2291
+ type: chrf
2292
+ value: 0.56290
2293
+ - task:
2294
+ name: Translation deu-fra
2295
+ type: translation
2296
+ args: deu-fra
2297
+ dataset:
2298
+ name: newstest2012
2299
+ type: wmt-2012-news
2300
+ args: deu-fra
2301
+ metrics:
2302
+ - name: BLEU
2303
+ type: bleu
2304
+ value: 27.3
2305
+ - name: chr-F
2306
+ type: chrf
2307
+ value: 0.55931
2308
+ - task:
2309
+ name: Translation deu-spa
2310
+ type: translation
2311
+ args: deu-spa
2312
+ dataset:
2313
+ name: newstest2012
2314
+ type: wmt-2012-news
2315
+ args: deu-spa
2316
+ metrics:
2317
+ - name: BLEU
2318
+ type: bleu
2319
+ value: 31.5
2320
+ - name: chr-F
2321
+ type: chrf
2322
+ value: 0.57369
2323
+ - task:
2324
+ name: Translation eng-deu
2325
+ type: translation
2326
+ args: eng-deu
2327
+ dataset:
2328
+ name: newstest2012
2329
+ type: wmt-2012-news
2330
+ args: eng-deu
2331
+ metrics:
2332
+ - name: BLEU
2333
+ type: bleu
2334
+ value: 23.3
2335
+ - name: chr-F
2336
+ type: chrf
2337
+ value: 0.52668
2338
+ - task:
2339
+ name: Translation eng-fra
2340
+ type: translation
2341
+ args: eng-fra
2342
+ dataset:
2343
+ name: newstest2012
2344
+ type: wmt-2012-news
2345
+ args: eng-fra
2346
+ metrics:
2347
+ - name: BLEU
2348
+ type: bleu
2349
+ value: 31.6
2350
+ - name: chr-F
2351
+ type: chrf
2352
+ value: 0.59076
2353
+ - task:
2354
+ name: Translation eng-spa
2355
+ type: translation
2356
+ args: eng-spa
2357
+ dataset:
2358
+ name: newstest2012
2359
+ type: wmt-2012-news
2360
+ args: eng-spa
2361
+ metrics:
2362
+ - name: BLEU
2363
+ type: bleu
2364
+ value: 38.8
2365
+ - name: chr-F
2366
+ type: chrf
2367
+ value: 0.62361
2368
+ - task:
2369
+ name: Translation deu-eng
2370
+ type: translation
2371
+ args: deu-eng
2372
+ dataset:
2373
+ name: newstest2013
2374
+ type: wmt-2013-news
2375
+ args: deu-eng
2376
+ metrics:
2377
+ - name: BLEU
2378
+ type: bleu
2379
+ value: 31.8
2380
+ - name: chr-F
2381
+ type: chrf
2382
+ value: 0.58065
2383
+ - task:
2384
+ name: Translation deu-fra
2385
+ type: translation
2386
+ args: deu-fra
2387
+ dataset:
2388
+ name: newstest2013
2389
+ type: wmt-2013-news
2390
+ args: deu-fra
2391
+ metrics:
2392
+ - name: BLEU
2393
+ type: bleu
2394
+ value: 30.0
2395
+ - name: chr-F
2396
+ type: chrf
2397
+ value: 0.56431
2398
+ - task:
2399
+ name: Translation deu-spa
2400
+ type: translation
2401
+ args: deu-spa
2402
+ dataset:
2403
+ name: newstest2013
2404
+ type: wmt-2013-news
2405
+ args: deu-spa
2406
+ metrics:
2407
+ - name: BLEU
2408
+ type: bleu
2409
+ value: 31.5
2410
+ - name: chr-F
2411
+ type: chrf
2412
+ value: 0.56965
2413
+ - task:
2414
+ name: Translation eng-deu
2415
+ type: translation
2416
+ args: eng-deu
2417
+ dataset:
2418
+ name: newstest2013
2419
+ type: wmt-2013-news
2420
+ args: eng-deu
2421
+ metrics:
2422
+ - name: BLEU
2423
+ type: bleu
2424
+ value: 26.9
2425
+ - name: chr-F
2426
+ type: chrf
2427
+ value: 0.55423
2428
+ - task:
2429
+ name: Translation eng-fra
2430
+ type: translation
2431
+ args: eng-fra
2432
+ dataset:
2433
+ name: newstest2013
2434
+ type: wmt-2013-news
2435
+ args: eng-fra
2436
+ metrics:
2437
+ - name: BLEU
2438
+ type: bleu
2439
+ value: 33.1
2440
+ - name: chr-F
2441
+ type: chrf
2442
+ value: 0.58760
2443
+ - task:
2444
+ name: Translation eng-spa
2445
+ type: translation
2446
+ args: eng-spa
2447
+ dataset:
2448
+ name: newstest2013
2449
+ type: wmt-2013-news
2450
+ args: eng-spa
2451
+ metrics:
2452
+ - name: BLEU
2453
+ type: bleu
2454
+ value: 35.1
2455
+ - name: chr-F
2456
+ type: chrf
2457
+ value: 0.59825
2458
+ - task:
2459
+ name: Translation deu-eng
2460
+ type: translation
2461
+ args: deu-eng
2462
+ dataset:
2463
+ name: newstest2014
2464
+ type: wmt-2014-news
2465
+ args: deu-eng
2466
+ metrics:
2467
+ - name: BLEU
2468
+ type: bleu
2469
+ value: 32.9
2470
+ - name: chr-F
2471
+ type: chrf
2472
+ value: 0.59617
2473
+ - task:
2474
+ name: Translation eng-deu
2475
+ type: translation
2476
+ args: eng-deu
2477
+ dataset:
2478
+ name: newstest2014
2479
+ type: wmt-2014-news
2480
+ args: eng-deu
2481
+ metrics:
2482
+ - name: BLEU
2483
+ type: bleu
2484
+ value: 28.0
2485
+ - name: chr-F
2486
+ type: chrf
2487
+ value: 0.58847
2488
+ - task:
2489
+ name: Translation eng-fra
2490
+ type: translation
2491
+ args: eng-fra
2492
+ dataset:
2493
+ name: newstest2014
2494
+ type: wmt-2014-news
2495
+ args: eng-fra
2496
+ metrics:
2497
+ - name: BLEU
2498
+ type: bleu
2499
+ value: 39.9
2500
+ - name: chr-F
2501
+ type: chrf
2502
+ value: 0.65294
2503
+ - task:
2504
+ name: Translation deu-eng
2505
+ type: translation
2506
+ args: deu-eng
2507
+ dataset:
2508
+ name: newstest2015
2509
+ type: wmt-2015-news
2510
+ args: deu-eng
2511
+ metrics:
2512
+ - name: BLEU
2513
+ type: bleu
2514
+ value: 33.8
2515
+ - name: chr-F
2516
+ type: chrf
2517
+ value: 0.59741
2518
+ - task:
2519
+ name: Translation eng-deu
2520
+ type: translation
2521
+ args: eng-deu
2522
+ dataset:
2523
+ name: newstest2015
2524
+ type: wmt-2015-news
2525
+ args: eng-deu
2526
+ metrics:
2527
+ - name: BLEU
2528
+ type: bleu
2529
+ value: 31.0
2530
+ - name: chr-F
2531
+ type: chrf
2532
+ value: 0.59474
2533
+ - task:
2534
+ name: Translation deu-eng
2535
+ type: translation
2536
+ args: deu-eng
2537
+ dataset:
2538
+ name: newstest2016
2539
+ type: wmt-2016-news
2540
+ args: deu-eng
2541
+ metrics:
2542
+ - name: BLEU
2543
+ type: bleu
2544
+ value: 40.6
2545
+ - name: chr-F
2546
+ type: chrf
2547
+ value: 0.64981
2548
+ - task:
2549
+ name: Translation eng-deu
2550
+ type: translation
2551
+ args: eng-deu
2552
+ dataset:
2553
+ name: newstest2016
2554
+ type: wmt-2016-news
2555
+ args: eng-deu
2556
+ metrics:
2557
+ - name: BLEU
2558
+ type: bleu
2559
+ value: 37.1
2560
+ - name: chr-F
2561
+ type: chrf
2562
+ value: 0.63839
2563
+ - task:
2564
+ name: Translation deu-eng
2565
+ type: translation
2566
+ args: deu-eng
2567
+ dataset:
2568
+ name: newstest2017
2569
+ type: wmt-2017-news
2570
+ args: deu-eng
2571
+ metrics:
2572
+ - name: BLEU
2573
+ type: bleu
2574
+ value: 35.5
2575
+ - name: chr-F
2576
+ type: chrf
2577
+ value: 0.60957
2578
+ - task:
2579
+ name: Translation eng-deu
2580
+ type: translation
2581
+ args: eng-deu
2582
+ dataset:
2583
+ name: newstest2017
2584
+ type: wmt-2017-news
2585
+ args: eng-deu
2586
+ metrics:
2587
+ - name: BLEU
2588
+ type: bleu
2589
+ value: 30.0
2590
+ - name: chr-F
2591
+ type: chrf
2592
+ value: 0.58967
2593
+ - task:
2594
+ name: Translation deu-eng
2595
+ type: translation
2596
+ args: deu-eng
2597
+ dataset:
2598
+ name: newstest2018
2599
+ type: wmt-2018-news
2600
+ args: deu-eng
2601
+ metrics:
2602
+ - name: BLEU
2603
+ type: bleu
2604
+ value: 43.4
2605
+ - name: chr-F
2606
+ type: chrf
2607
+ value: 0.66739
2608
+ - task:
2609
+ name: Translation eng-deu
2610
+ type: translation
2611
+ args: eng-deu
2612
+ dataset:
2613
+ name: newstest2018
2614
+ type: wmt-2018-news
2615
+ args: eng-deu
2616
+ metrics:
2617
+ - name: BLEU
2618
+ type: bleu
2619
+ value: 44.9
2620
+ - name: chr-F
2621
+ type: chrf
2622
+ value: 0.68858
2623
+ - task:
2624
+ name: Translation deu-eng
2625
+ type: translation
2626
+ args: deu-eng
2627
+ dataset:
2628
+ name: newstest2019
2629
+ type: wmt-2019-news
2630
+ args: deu-eng
2631
+ metrics:
2632
+ - name: BLEU
2633
+ type: bleu
2634
+ value: 39.6
2635
+ - name: chr-F
2636
+ type: chrf
2637
+ value: 0.63671
2638
+ - task:
2639
+ name: Translation deu-fra
2640
+ type: translation
2641
+ args: deu-fra
2642
+ dataset:
2643
+ name: newstest2019
2644
+ type: wmt-2019-news
2645
+ args: deu-fra
2646
+ metrics:
2647
+ - name: BLEU
2648
+ type: bleu
2649
+ value: 36.1
2650
+ - name: chr-F
2651
+ type: chrf
2652
+ value: 0.63043
2653
+ - task:
2654
+ name: Translation eng-deu
2655
+ type: translation
2656
+ args: eng-deu
2657
+ dataset:
2658
+ name: newstest2019
2659
+ type: wmt-2019-news
2660
+ args: eng-deu
2661
+ metrics:
2662
+ - name: BLEU
2663
+ type: bleu
2664
+ value: 41.4
2665
+ - name: chr-F
2666
+ type: chrf
2667
+ value: 0.65934
2668
+ - task:
2669
+ name: Translation deu-eng
2670
+ type: translation
2671
+ args: deu-eng
2672
+ dataset:
2673
+ name: newstest2020
2674
+ type: wmt-2020-news
2675
+ args: deu-eng
2676
+ metrics:
2677
+ - name: BLEU
2678
+ type: bleu
2679
+ value: 34.5
2680
+ - name: chr-F
2681
+ type: chrf
2682
+ value: 0.60800
2683
+ - task:
2684
+ name: Translation deu-fra
2685
+ type: translation
2686
+ args: deu-fra
2687
+ dataset:
2688
+ name: newstest2020
2689
+ type: wmt-2020-news
2690
+ args: deu-fra
2691
+ metrics:
2692
+ - name: BLEU
2693
+ type: bleu
2694
+ value: 33.1
2695
+ - name: chr-F
2696
+ type: chrf
2697
+ value: 0.60544
2698
+ - task:
2699
+ name: Translation eng-deu
2700
+ type: translation
2701
+ args: eng-deu
2702
+ dataset:
2703
+ name: newstest2020
2704
+ type: wmt-2020-news
2705
+ args: eng-deu
2706
+ metrics:
2707
+ - name: BLEU
2708
+ type: bleu
2709
+ value: 31.7
2710
+ - name: chr-F
2711
+ type: chrf
2712
+ value: 0.60078
2713
+ - task:
2714
+ name: Translation deu-eng
2715
+ type: translation
2716
+ args: deu-eng
2717
+ dataset:
2718
+ name: newstest2021
2719
+ type: wmt-2021-news
2720
+ args: deu-eng
2721
+ metrics:
2722
+ - name: BLEU
2723
+ type: bleu
2724
+ value: 31.9
2725
+ - name: chr-F
2726
+ type: chrf
2727
+ value: 0.60048
2728
+ - task:
2729
+ name: Translation deu-fra
2730
+ type: translation
2731
+ args: deu-fra
2732
+ dataset:
2733
+ name: newstest2021
2734
+ type: wmt-2021-news
2735
+ args: deu-fra
2736
+ metrics:
2737
+ - name: BLEU
2738
+ type: bleu
2739
+ value: 31.8
2740
+ - name: chr-F
2741
+ type: chrf
2742
+ value: 0.59590
2743
+ - task:
2744
+ name: Translation eng-deu
2745
+ type: translation
2746
+ args: eng-deu
2747
+ dataset:
2748
+ name: newstest2021
2749
+ type: wmt-2021-news
2750
+ args: eng-deu
2751
+ metrics:
2752
+ - name: BLEU
2753
+ type: bleu
2754
+ value: 25.6
2755
+ - name: chr-F
2756
+ type: chrf
2757
+ value: 0.56133
2758
+ ---
2759
+ # opus-mt-tc-bible-big-gmw-deu_eng_fra_por_spa
2760
+
2761
+ ## Table of Contents
2762
+ - [Model Details](#model-details)
2763
+ - [Uses](#uses)
2764
+ - [Risks, Limitations and Biases](#risks-limitations-and-biases)
2765
+ - [How to Get Started With the Model](#how-to-get-started-with-the-model)
2766
+ - [Training](#training)
2767
+ - [Evaluation](#evaluation)
2768
+ - [Citation Information](#citation-information)
2769
+ - [Acknowledgements](#acknowledgements)
2770
+
2771
+ ## Model Details
2772
+
2773
+ Neural machine translation model for translating from West Germanic languages (gmw) to unknown (deu+eng+fra+por+spa).
2774
+
2775
+ 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).
2776
+ **Model Description:**
2777
+ - **Developed by:** Language Technology Research Group at the University of Helsinki
2778
+ - **Model Type:** Translation (transformer-big)
2779
+ - **Release**: 2024-05-30
2780
+ - **License:** Apache-2.0
2781
+ - **Language(s):**
2782
+ - Source Language(s): afr ang bar bis bzj deu djk drt eng enm frr fry gos gsw hrx hwc icr jam kri ksh lim ltz nds nld ofs pcm pdc pfl pih pis rop sco srm srn stq swg tcs tpi vls wae yid zea
2783
+ - Target Language(s): deu eng fra por spa
2784
+ - Valid Target Language Labels: >>deu<< >>eng<< >>fra<< >>por<< >>spa<< >>xxx<<
2785
+ - **Original Model**: [opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-30.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/gmw-deu+eng+fra+por+spa/opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-30.zip)
2786
+ - **Resources for more information:**
2787
+ - [OPUS-MT dashboard](https://opus.nlpl.eu/dashboard/index.php?pkg=opusmt&test=all&scoreslang=all&chart=standard&model=Tatoeba-MT-models/gmw-deu%2Beng%2Bfra%2Bpor%2Bspa/opusTCv20230926max50%2Bbt%2Bjhubc_transformer-big_2024-05-30)
2788
+ - [OPUS-MT-train GitHub Repo](https://github.com/Helsinki-NLP/OPUS-MT-train)
2789
+ - [More information about MarianNMT models in the transformers library](https://huggingface.co/docs/transformers/model_doc/marian)
2790
+ - [Tatoeba Translation Challenge](https://github.com/Helsinki-NLP/Tatoeba-Challenge/)
2791
+ - [HPLT bilingual data v1 (as part of the Tatoeba Translation Challenge dataset)](https://hplt-project.org/datasets/v1)
2792
+ - [A massively parallel Bible corpus](https://aclanthology.org/L14-1215/)
2793
+
2794
+ 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<<`
2795
+
2796
+ ## Uses
2797
+
2798
+ This model can be used for translation and text-to-text generation.
2799
+
2800
+ ## Risks, Limitations and Biases
2801
+
2802
+ **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.**
2803
+
2804
+ 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)).
2805
+
2806
+ ## How to Get Started With the Model
2807
+
2808
+ A short example code:
2809
+
2810
+ ```python
2811
+ from transformers import MarianMTModel, MarianTokenizer
2812
+
2813
+ src_text = [
2814
+ ">>deu<< Replace this with text in an accepted source language.",
2815
+ ">>spa<< This is the second sentence."
2816
+ ]
2817
+
2818
+ model_name = "pytorch-models/opus-mt-tc-bible-big-gmw-deu_eng_fra_por_spa"
2819
+ tokenizer = MarianTokenizer.from_pretrained(model_name)
2820
+ model = MarianMTModel.from_pretrained(model_name)
2821
+ translated = model.generate(**tokenizer(src_text, return_tensors="pt", padding=True))
2822
+
2823
+ for t in translated:
2824
+ print( tokenizer.decode(t, skip_special_tokens=True) )
2825
+ ```
2826
+
2827
+ You can also use OPUS-MT models with the transformers pipelines, for example:
2828
+
2829
+ ```python
2830
+ from transformers import pipeline
2831
+ pipe = pipeline("translation", model="Helsinki-NLP/opus-mt-tc-bible-big-gmw-deu_eng_fra_por_spa")
2832
+ print(pipe(">>deu<< Replace this with text in an accepted source language."))
2833
+ ```
2834
+
2835
+ ## Training
2836
+
2837
+ - **Data**: opusTCv20230926max50+bt+jhubc ([source](https://github.com/Helsinki-NLP/Tatoeba-Challenge))
2838
+ - **Pre-processing**: SentencePiece (spm32k,spm32k)
2839
+ - **Model Type:** transformer-big
2840
+ - **Original MarianNMT Model**: [opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-30.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/gmw-deu+eng+fra+por+spa/opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-30.zip)
2841
+ - **Training Scripts**: [GitHub Repo](https://github.com/Helsinki-NLP/OPUS-MT-train)
2842
+
2843
+ ## Evaluation
2844
+
2845
+ * [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/gmw-deu%2Beng%2Bfra%2Bpor%2Bspa/opusTCv20230926max50%2Bbt%2Bjhubc_transformer-big_2024-05-30)
2846
+ * test set translations: [opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-29.test.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/gmw-deu+eng+fra+por+spa/opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-29.test.txt)
2847
+ * test set scores: [opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-29.eval.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/gmw-deu+eng+fra+por+spa/opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-29.eval.txt)
2848
+ * benchmark results: [benchmark_results.txt](benchmark_results.txt)
2849
+ * benchmark output: [benchmark_translations.zip](benchmark_translations.zip)
2850
+
2851
+ | langpair | testset | chr-F | BLEU | #sent | #words |
2852
+ |----------|---------|-------|-------|-------|--------|
2853
+ | afr-deu | tatoeba-test-v2021-08-07 | 0.68492 | 48.8 | 1583 | 9105 |
2854
+ | afr-eng | tatoeba-test-v2021-08-07 | 0.72943 | 59.6 | 1374 | 9622 |
2855
+ | afr-spa | tatoeba-test-v2021-08-07 | 0.72793 | 58.4 | 448 | 2783 |
2856
+ | deu-deu | tatoeba-test-v2021-08-07 | 0.59840 | 34.8 | 2500 | 20806 |
2857
+ | deu-eng | tatoeba-test-v2021-08-07 | 0.65957 | 48.5 | 17565 | 149462 |
2858
+ | deu-fra | tatoeba-test-v2021-08-07 | 0.68054 | 50.2 | 12418 | 102721 |
2859
+ | deu-por | tatoeba-test-v2021-08-07 | 0.63368 | 42.8 | 10000 | 81482 |
2860
+ | deu-spa | tatoeba-test-v2021-08-07 | 0.68198 | 49.3 | 10521 | 82570 |
2861
+ | eng-deu | tatoeba-test-v2021-08-07 | 0.62497 | 40.5 | 17565 | 151568 |
2862
+ | eng-eng | tatoeba-test-v2021-08-07 | 0.79878 | 57.3 | 12062 | 115106 |
2863
+ | eng-fra | tatoeba-test-v2021-08-07 | 0.68567 | 50.9 | 12681 | 106378 |
2864
+ | eng-por | tatoeba-test-v2021-08-07 | 0.72204 | 53.4 | 13222 | 105265 |
2865
+ | eng-spa | tatoeba-test-v2021-08-07 | 0.72539 | 55.3 | 16583 | 134710 |
2866
+ | fry-eng | tatoeba-test-v2021-08-07 | 0.55137 | 37.0 | 220 | 1573 |
2867
+ | gos-deu | tatoeba-test-v2021-08-07 | 0.46120 | 24.7 | 207 | 1168 |
2868
+ | gos-eng | tatoeba-test-v2021-08-07 | 0.38628 | 22.3 | 1154 | 5635 |
2869
+ | gsw-eng | tatoeba-test-v2021-08-07 | 0.43003 | 27.5 | 205 | 990 |
2870
+ | ltz-deu | tatoeba-test-v2021-08-07 | 0.48474 | 32.0 | 347 | 2208 |
2871
+ | ltz-eng | tatoeba-test-v2021-08-07 | 0.65366 | 56.4 | 293 | 1840 |
2872
+ | nds-deu | tatoeba-test-v2021-08-07 | 0.65251 | 45.9 | 9999 | 74564 |
2873
+ | nds-eng | tatoeba-test-v2021-08-07 | 0.61858 | 44.7 | 2500 | 17589 |
2874
+ | nds-fra | tatoeba-test-v2021-08-07 | 0.60412 | 43.9 | 857 | 5676 |
2875
+ | nds-por | tatoeba-test-v2021-08-07 | 0.58778 | 39.5 | 207 | 1256 |
2876
+ | nds-spa | tatoeba-test-v2021-08-07 | 0.63404 | 43.9 | 923 | 5540 |
2877
+ | nld-deu | tatoeba-test-v2021-08-07 | 0.72998 | 55.7 | 10218 | 74131 |
2878
+ | nld-eng | tatoeba-test-v2021-08-07 | 0.74362 | 60.9 | 12696 | 89978 |
2879
+ | nld-fra | tatoeba-test-v2021-08-07 | 0.68461 | 48.0 | 11548 | 82974 |
2880
+ | nld-por | tatoeba-test-v2021-08-07 | 0.68798 | 49.3 | 2500 | 17326 |
2881
+ | nld-spa | tatoeba-test-v2021-08-07 | 0.69971 | 51.6 | 10113 | 74981 |
2882
+ | yid-eng | tatoeba-test-v2021-08-07 | 0.49807 | 31.5 | 2483 | 15452 |
2883
+ | yid-fra | tatoeba-test-v2021-08-07 | 0.54147 | 31.9 | 384 | 2455 |
2884
+ | afr-deu | flores101-devtest | 0.57831 | 28.4 | 1012 | 25094 |
2885
+ | afr-eng | flores101-devtest | 0.74272 | 53.8 | 1012 | 24721 |
2886
+ | afr-fra | flores101-devtest | 0.61936 | 36.4 | 1012 | 28343 |
2887
+ | afr-por | flores101-devtest | 0.61309 | 35.4 | 1012 | 26519 |
2888
+ | afr-spa | flores101-devtest | 0.51393 | 22.9 | 1012 | 29199 |
2889
+ | deu-spa | flores101-devtest | 0.52438 | 23.9 | 1012 | 29199 |
2890
+ | eng-deu | flores101-devtest | 0.64236 | 37.2 | 1012 | 25094 |
2891
+ | eng-spa | flores101-devtest | 0.55524 | 27.1 | 1012 | 29199 |
2892
+ | nld-deu | flores101-devtest | 0.53435 | 22.1 | 1012 | 25094 |
2893
+ | nld-eng | flores101-devtest | 0.58686 | 30.0 | 1012 | 24721 |
2894
+ | nld-fra | flores101-devtest | 0.56292 | 28.2 | 1012 | 28343 |
2895
+ | afr-deu | flores200-devtest | 0.58456 | 29.5 | 1012 | 25094 |
2896
+ | afr-eng | flores200-devtest | 0.74857 | 54.7 | 1012 | 24721 |
2897
+ | afr-fra | flores200-devtest | 0.62537 | 37.2 | 1012 | 28343 |
2898
+ | afr-por | flores200-devtest | 0.61751 | 36.1 | 1012 | 26519 |
2899
+ | afr-spa | flores200-devtest | 0.51647 | 23.2 | 1012 | 29199 |
2900
+ | deu-eng | flores200-devtest | 0.67103 | 41.4 | 1012 | 24721 |
2901
+ | deu-fra | flores200-devtest | 0.62658 | 36.8 | 1012 | 28343 |
2902
+ | deu-por | flores200-devtest | 0.60909 | 34.8 | 1012 | 26519 |
2903
+ | deu-spa | flores200-devtest | 0.52584 | 24.2 | 1012 | 29199 |
2904
+ | eng-deu | flores200-devtest | 0.64560 | 37.5 | 1012 | 25094 |
2905
+ | eng-fra | flores200-devtest | 0.70736 | 49.1 | 1012 | 28343 |
2906
+ | eng-por | flores200-devtest | 0.71065 | 49.5 | 1012 | 26519 |
2907
+ | eng-spa | flores200-devtest | 0.55738 | 27.4 | 1012 | 29199 |
2908
+ | lim-deu | flores200-devtest | 0.45062 | 16.1 | 1012 | 25094 |
2909
+ | lim-eng | flores200-devtest | 0.48217 | 21.8 | 1012 | 24721 |
2910
+ | lim-fra | flores200-devtest | 0.44347 | 18.5 | 1012 | 28343 |
2911
+ | lim-por | flores200-devtest | 0.42527 | 16.8 | 1012 | 26519 |
2912
+ | ltz-deu | flores200-devtest | 0.60114 | 31.3 | 1012 | 25094 |
2913
+ | ltz-eng | flores200-devtest | 0.64345 | 39.3 | 1012 | 24721 |
2914
+ | ltz-fra | flores200-devtest | 0.59368 | 33.8 | 1012 | 28343 |
2915
+ | ltz-por | flores200-devtest | 0.51545 | 24.8 | 1012 | 26519 |
2916
+ | ltz-spa | flores200-devtest | 0.44821 | 17.5 | 1012 | 29199 |
2917
+ | nld-deu | flores200-devtest | 0.53650 | 22.4 | 1012 | 25094 |
2918
+ | nld-eng | flores200-devtest | 0.59102 | 30.6 | 1012 | 24721 |
2919
+ | nld-fra | flores200-devtest | 0.56608 | 28.7 | 1012 | 28343 |
2920
+ | nld-por | flores200-devtest | 0.54728 | 26.7 | 1012 | 26519 |
2921
+ | nld-spa | flores200-devtest | 0.49175 | 20.1 | 1012 | 29199 |
2922
+ | tpi-deu | flores200-devtest | 0.40350 | 10.9 | 1012 | 25094 |
2923
+ | tpi-eng | flores200-devtest | 0.48289 | 19.6 | 1012 | 24721 |
2924
+ | tpi-fra | flores200-devtest | 0.43428 | 16.1 | 1012 | 28343 |
2925
+ | tpi-por | flores200-devtest | 0.42966 | 15.4 | 1012 | 26519 |
2926
+ | deu-eng | generaltest2022 | 0.56042 | 31.0 | 1984 | 37634 |
2927
+ | deu-fra | generaltest2022 | 0.61145 | 37.6 | 1984 | 38276 |
2928
+ | eng-deu | generaltest2022 | 0.60090 | 32.5 | 2037 | 38914 |
2929
+ | deu-eng | multi30k_test_2016_flickr | 0.60974 | 40.1 | 1000 | 12955 |
2930
+ | deu-fra | multi30k_test_2016_flickr | 0.62493 | 38.8 | 1000 | 13505 |
2931
+ | eng-deu | multi30k_test_2016_flickr | 0.64164 | 35.3 | 1000 | 12106 |
2932
+ | eng-fra | multi30k_test_2016_flickr | 0.71137 | 50.7 | 1000 | 13505 |
2933
+ | deu-eng | multi30k_test_2017_flickr | 0.63118 | 40.6 | 1000 | 11374 |
2934
+ | deu-fra | multi30k_test_2017_flickr | 0.62614 | 37.0 | 1000 | 12118 |
2935
+ | eng-deu | multi30k_test_2017_flickr | 0.62518 | 33.4 | 1000 | 10755 |
2936
+ | eng-fra | multi30k_test_2017_flickr | 0.71402 | 50.3 | 1000 | 12118 |
2937
+ | deu-eng | multi30k_test_2017_mscoco | 0.55495 | 32.1 | 461 | 5231 |
2938
+ | deu-fra | multi30k_test_2017_mscoco | 0.59307 | 34.7 | 461 | 5484 |
2939
+ | eng-deu | multi30k_test_2017_mscoco | 0.58028 | 29.7 | 461 | 5158 |
2940
+ | eng-fra | multi30k_test_2017_mscoco | 0.73637 | 54.7 | 461 | 5484 |
2941
+ | deu-eng | multi30k_test_2018_flickr | 0.59367 | 36.7 | 1071 | 14689 |
2942
+ | deu-fra | multi30k_test_2018_flickr | 0.57388 | 31.3 | 1071 | 15867 |
2943
+ | eng-deu | multi30k_test_2018_flickr | 0.59998 | 30.8 | 1071 | 13703 |
2944
+ | eng-fra | multi30k_test_2018_flickr | 0.65354 | 41.6 | 1071 | 15867 |
2945
+ | eng-fra | newsdiscusstest2015 | 0.63308 | 37.7 | 1500 | 27975 |
2946
+ | deu-eng | newssyscomb2009 | 0.55170 | 28.3 | 502 | 11818 |
2947
+ | deu-fra | newssyscomb2009 | 0.56021 | 27.4 | 502 | 12331 |
2948
+ | deu-spa | newssyscomb2009 | 0.55546 | 28.1 | 502 | 12503 |
2949
+ | eng-deu | newssyscomb2009 | 0.53919 | 23.0 | 502 | 11271 |
2950
+ | eng-fra | newssyscomb2009 | 0.58384 | 29.5 | 502 | 12331 |
2951
+ | eng-spa | newssyscomb2009 | 0.58266 | 31.0 | 502 | 12503 |
2952
+ | deu-eng | newstest2008 | 0.54434 | 27.0 | 2051 | 49380 |
2953
+ | deu-fra | newstest2008 | 0.55076 | 26.2 | 2051 | 52685 |
2954
+ | deu-spa | newstest2008 | 0.54056 | 25.6 | 2051 | 52586 |
2955
+ | eng-deu | newstest2008 | 0.52906 | 23.0 | 2051 | 47447 |
2956
+ | eng-fra | newstest2008 | 0.55247 | 26.8 | 2051 | 52685 |
2957
+ | eng-spa | newstest2008 | 0.56423 | 29.6 | 2051 | 52586 |
2958
+ | deu-eng | newstest2009 | 0.53972 | 26.7 | 2525 | 65399 |
2959
+ | deu-fra | newstest2009 | 0.53975 | 25.6 | 2525 | 69263 |
2960
+ | deu-spa | newstest2009 | 0.53677 | 25.6 | 2525 | 68111 |
2961
+ | eng-deu | newstest2009 | 0.53097 | 22.1 | 2525 | 62816 |
2962
+ | eng-fra | newstest2009 | 0.57542 | 29.1 | 2525 | 69263 |
2963
+ | eng-spa | newstest2009 | 0.57733 | 29.8 | 2525 | 68111 |
2964
+ | deu-eng | newstest2010 | 0.58278 | 30.2 | 2489 | 61711 |
2965
+ | deu-fra | newstest2010 | 0.57876 | 29.0 | 2489 | 66022 |
2966
+ | deu-spa | newstest2010 | 0.59402 | 32.6 | 2489 | 65480 |
2967
+ | eng-deu | newstest2010 | 0.54587 | 25.3 | 2489 | 61503 |
2968
+ | eng-fra | newstest2010 | 0.59460 | 32.0 | 2489 | 66022 |
2969
+ | eng-spa | newstest2010 | 0.61861 | 36.3 | 2489 | 65480 |
2970
+ | deu-eng | newstest2011 | 0.55074 | 26.8 | 3003 | 74681 |
2971
+ | deu-fra | newstest2011 | 0.55879 | 27.4 | 3003 | 80626 |
2972
+ | deu-spa | newstest2011 | 0.56593 | 30.2 | 3003 | 79476 |
2973
+ | eng-deu | newstest2011 | 0.52619 | 22.7 | 3003 | 72981 |
2974
+ | eng-fra | newstest2011 | 0.60960 | 34.1 | 3003 | 80626 |
2975
+ | eng-spa | newstest2011 | 0.62056 | 38.5 | 3003 | 79476 |
2976
+ | deu-eng | newstest2012 | 0.56290 | 28.4 | 3003 | 72812 |
2977
+ | deu-fra | newstest2012 | 0.55931 | 27.3 | 3003 | 78011 |
2978
+ | deu-spa | newstest2012 | 0.57369 | 31.5 | 3003 | 79006 |
2979
+ | eng-deu | newstest2012 | 0.52668 | 23.3 | 3003 | 72886 |
2980
+ | eng-fra | newstest2012 | 0.59076 | 31.6 | 3003 | 78011 |
2981
+ | eng-spa | newstest2012 | 0.62361 | 38.8 | 3003 | 79006 |
2982
+ | deu-eng | newstest2013 | 0.58065 | 31.8 | 3000 | 64505 |
2983
+ | deu-fra | newstest2013 | 0.56431 | 30.0 | 3000 | 70037 |
2984
+ | deu-spa | newstest2013 | 0.56965 | 31.5 | 3000 | 70528 |
2985
+ | eng-deu | newstest2013 | 0.55423 | 26.9 | 3000 | 63737 |
2986
+ | eng-fra | newstest2013 | 0.58760 | 33.1 | 3000 | 70037 |
2987
+ | eng-spa | newstest2013 | 0.59825 | 35.1 | 3000 | 70528 |
2988
+ | deu-eng | newstest2014 | 0.59617 | 32.9 | 3003 | 67337 |
2989
+ | eng-deu | newstest2014 | 0.58847 | 28.0 | 3003 | 62688 |
2990
+ | eng-fra | newstest2014 | 0.65294 | 39.9 | 3003 | 77306 |
2991
+ | deu-eng | newstest2015 | 0.59741 | 33.8 | 2169 | 46443 |
2992
+ | eng-deu | newstest2015 | 0.59474 | 31.0 | 2169 | 44260 |
2993
+ | deu-eng | newstest2016 | 0.64981 | 40.6 | 2999 | 64119 |
2994
+ | eng-deu | newstest2016 | 0.63839 | 37.1 | 2999 | 62669 |
2995
+ | deu-eng | newstest2017 | 0.60957 | 35.5 | 3004 | 64399 |
2996
+ | eng-deu | newstest2017 | 0.58967 | 30.0 | 3004 | 61287 |
2997
+ | deu-eng | newstest2018 | 0.66739 | 43.4 | 2998 | 67012 |
2998
+ | eng-deu | newstest2018 | 0.68858 | 44.9 | 2998 | 64276 |
2999
+ | deu-eng | newstest2019 | 0.63671 | 39.6 | 2000 | 39227 |
3000
+ | deu-fra | newstest2019 | 0.63043 | 36.1 | 1701 | 42509 |
3001
+ | eng-deu | newstest2019 | 0.65934 | 41.4 | 1997 | 48746 |
3002
+ | deu-eng | newstest2020 | 0.60800 | 34.5 | 785 | 38220 |
3003
+ | deu-fra | newstest2020 | 0.60544 | 33.1 | 1619 | 36890 |
3004
+ | eng-deu | newstest2020 | 0.60078 | 31.7 | 1418 | 52383 |
3005
+ | deu-eng | newstest2021 | 0.60048 | 31.9 | 1000 | 20180 |
3006
+ | deu-fra | newstest2021 | 0.59590 | 31.8 | 1000 | 23757 |
3007
+ | eng-deu | newstest2021 | 0.56133 | 25.6 | 1002 | 27970 |
3008
+ | deu-eng | newstestALL2020 | 0.60800 | 34.5 | 785 | 38220 |
3009
+ | eng-deu | newstestALL2020 | 0.60078 | 31.7 | 1418 | 52383 |
3010
+ | deu-eng | newstestB2020 | 0.60795 | 34.4 | 785 | 37696 |
3011
+ | eng-deu | newstestB2020 | 0.59256 | 31.5 | 1418 | 53092 |
3012
+ | afr-deu | ntrex128 | 0.55289 | 25.8 | 1997 | 48761 |
3013
+ | afr-eng | ntrex128 | 0.72558 | 51.8 | 1997 | 47673 |
3014
+ | afr-fra | ntrex128 | 0.56601 | 29.3 | 1997 | 53481 |
3015
+ | afr-por | ntrex128 | 0.55396 | 28.1 | 1997 | 51631 |
3016
+ | afr-spa | ntrex128 | 0.58558 | 33.7 | 1997 | 54107 |
3017
+ | deu-eng | ntrex128 | 0.61722 | 33.8 | 1997 | 47673 |
3018
+ | deu-fra | ntrex128 | 0.55908 | 28.6 | 1997 | 53481 |
3019
+ | deu-por | ntrex128 | 0.54059 | 25.7 | 1997 | 51631 |
3020
+ | deu-spa | ntrex128 | 0.56887 | 30.8 | 1997 | 54107 |
3021
+ | eng-deu | ntrex128 | 0.58492 | 29.8 | 1997 | 48761 |
3022
+ | eng-fra | ntrex128 | 0.61349 | 35.2 | 1997 | 53481 |
3023
+ | eng-por | ntrex128 | 0.59785 | 33.4 | 1997 | 51631 |
3024
+ | eng-spa | ntrex128 | 0.63935 | 40.1 | 1997 | 54107 |
3025
+ | ltz-deu | ntrex128 | 0.51469 | 21.9 | 1997 | 48761 |
3026
+ | ltz-eng | ntrex128 | 0.58627 | 32.4 | 1997 | 47673 |
3027
+ | ltz-fra | ntrex128 | 0.50491 | 22.8 | 1997 | 53481 |
3028
+ | ltz-por | ntrex128 | 0.45364 | 18.7 | 1997 | 51631 |
3029
+ | ltz-spa | ntrex128 | 0.47568 | 21.6 | 1997 | 54107 |
3030
+ | nld-deu | ntrex128 | 0.55943 | 25.7 | 1997 | 48761 |
3031
+ | nld-eng | ntrex128 | 0.63470 | 36.1 | 1997 | 47673 |
3032
+ | nld-fra | ntrex128 | 0.55832 | 27.5 | 1997 | 53481 |
3033
+ | nld-por | ntrex128 | 0.54714 | 27.3 | 1997 | 51631 |
3034
+ | nld-spa | ntrex128 | 0.57692 | 32.1 | 1997 | 54107 |
3035
+ | eng-fra | tico19-test | 0.62559 | 39.5 | 2100 | 64661 |
3036
+ | eng-por | tico19-test | 0.72765 | 49.8 | 2100 | 62729 |
3037
+ | eng-spa | tico19-test | 0.72905 | 51.6 | 2100 | 66563 |
3038
+
3039
+ ## Citation Information
3040
+
3041
+ * 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.)
3042
+
3043
+ ```bibtex
3044
+ @article{tiedemann2023democratizing,
3045
+ title={Democratizing neural machine translation with {OPUS-MT}},
3046
+ 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},
3047
+ journal={Language Resources and Evaluation},
3048
+ number={58},
3049
+ pages={713--755},
3050
+ year={2023},
3051
+ publisher={Springer Nature},
3052
+ issn={1574-0218},
3053
+ doi={10.1007/s10579-023-09704-w}
3054
+ }
3055
+
3056
+ @inproceedings{tiedemann-thottingal-2020-opus,
3057
+ title = "{OPUS}-{MT} {--} Building open translation services for the World",
3058
+ author = {Tiedemann, J{\"o}rg and Thottingal, Santhosh},
3059
+ booktitle = "Proceedings of the 22nd Annual Conference of the European Association for Machine Translation",
3060
+ month = nov,
3061
+ year = "2020",
3062
+ address = "Lisboa, Portugal",
3063
+ publisher = "European Association for Machine Translation",
3064
+ url = "https://aclanthology.org/2020.eamt-1.61",
3065
+ pages = "479--480",
3066
+ }
3067
+
3068
+ @inproceedings{tiedemann-2020-tatoeba,
3069
+ title = "The Tatoeba Translation Challenge {--} Realistic Data Sets for Low Resource and Multilingual {MT}",
3070
+ author = {Tiedemann, J{\"o}rg},
3071
+ booktitle = "Proceedings of the Fifth Conference on Machine Translation",
3072
+ month = nov,
3073
+ year = "2020",
3074
+ address = "Online",
3075
+ publisher = "Association for Computational Linguistics",
3076
+ url = "https://aclanthology.org/2020.wmt-1.139",
3077
+ pages = "1174--1182",
3078
+ }
3079
+ ```
3080
+
3081
+ ## Acknowledgements
3082
+
3083
+ 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/).
3084
+
3085
+ ## Model conversion info
3086
+
3087
+ * transformers version: 4.45.1
3088
+ * OPUS-MT git hash: 0882077
3089
+ * port time: Tue Oct 8 11:18:52 EEST 2024
3090
+ * port machine: LM0-400-22516.local
benchmark_results.txt ADDED
@@ -0,0 +1,216 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ multi-multi tatoeba-test-v2020-07-28-v2023-09-26 0.68713 50.3 10000 85789
2
+ afr-deu flores101-devtest 0.57831 28.4 1012 25094
3
+ afr-eng flores101-devtest 0.74272 53.8 1012 24721
4
+ afr-fra flores101-devtest 0.61936 36.4 1012 28343
5
+ afr-por flores101-devtest 0.61309 35.4 1012 26519
6
+ afr-spa flores101-devtest 0.51393 22.9 1012 29199
7
+ deu-spa flores101-devtest 0.52438 23.9 1012 29199
8
+ eng-deu flores101-devtest 0.64236 37.2 1012 25094
9
+ eng-spa flores101-devtest 0.55524 27.1 1012 29199
10
+ nld-deu flores101-devtest 0.53435 22.1 1012 25094
11
+ nld-eng flores101-devtest 0.58686 30.0 1012 24721
12
+ nld-fra flores101-devtest 0.56292 28.2 1012 28343
13
+ afr-deu flores200-devtest 0.58456 29.5 1012 25094
14
+ afr-eng flores200-devtest 0.74857 54.7 1012 24721
15
+ afr-fra flores200-devtest 0.62537 37.2 1012 28343
16
+ afr-por flores200-devtest 0.61751 36.1 1012 26519
17
+ afr-spa flores200-devtest 0.51647 23.2 1012 29199
18
+ deu-eng flores200-devtest 0.67103 41.4 1012 24721
19
+ deu-fra flores200-devtest 0.62658 36.8 1012 28343
20
+ deu-por flores200-devtest 0.60909 34.8 1012 26519
21
+ deu-spa flores200-devtest 0.52584 24.2 1012 29199
22
+ eng-deu flores200-devtest 0.64560 37.5 1012 25094
23
+ eng-fra flores200-devtest 0.70736 49.1 1012 28343
24
+ eng-por flores200-devtest 0.71065 49.5 1012 26519
25
+ eng-spa flores200-devtest 0.55738 27.4 1012 29199
26
+ lim-deu flores200-devtest 0.45062 16.1 1012 25094
27
+ lim-eng flores200-devtest 0.48217 21.8 1012 24721
28
+ lim-fra flores200-devtest 0.44347 18.5 1012 28343
29
+ lim-por flores200-devtest 0.42527 16.8 1012 26519
30
+ lim-spa flores200-devtest 0.39232 13.6 1012 29199
31
+ ltz-deu flores200-devtest 0.60114 31.3 1012 25094
32
+ ltz-eng flores200-devtest 0.64345 39.3 1012 24721
33
+ ltz-fra flores200-devtest 0.59368 33.8 1012 28343
34
+ ltz-por flores200-devtest 0.51545 24.8 1012 26519
35
+ ltz-spa flores200-devtest 0.44821 17.5 1012 29199
36
+ nld-deu flores200-devtest 0.53650 22.4 1012 25094
37
+ nld-eng flores200-devtest 0.59102 30.6 1012 24721
38
+ nld-fra flores200-devtest 0.56608 28.7 1012 28343
39
+ nld-por flores200-devtest 0.54728 26.7 1012 26519
40
+ nld-spa flores200-devtest 0.49175 20.1 1012 29199
41
+ tpi-deu flores200-devtest 0.40350 10.9 1012 25094
42
+ tpi-eng flores200-devtest 0.48289 19.6 1012 24721
43
+ tpi-fra flores200-devtest 0.43428 16.1 1012 28343
44
+ tpi-por flores200-devtest 0.42966 15.4 1012 26519
45
+ tpi-spa flores200-devtest 0.39730 12.9 1012 29199
46
+ deu-eng generaltest2022 0.56042 31.0 1984 37634
47
+ deu-fra generaltest2022 0.61145 37.6 1984 38276
48
+ eng-deu generaltest2022 0.60090 32.5 2037 38914
49
+ deu-eng multi30k_task2_test_2016 0.21526 4.0 5000 67382
50
+ eng-deu multi30k_task2_test_2016 0.26922 2.6 5000 51501
51
+ deu-eng multi30k_test_2016_flickr 0.60974 40.1 1000 12955
52
+ deu-fra multi30k_test_2016_flickr 0.62493 38.8 1000 13505
53
+ eng-deu multi30k_test_2016_flickr 0.64164 35.3 1000 12106
54
+ eng-fra multi30k_test_2016_flickr 0.71137 50.7 1000 13505
55
+ deu-eng multi30k_test_2017_flickr 0.63118 40.6 1000 11374
56
+ deu-fra multi30k_test_2017_flickr 0.62614 37.0 1000 12118
57
+ eng-deu multi30k_test_2017_flickr 0.62518 33.4 1000 10755
58
+ eng-fra multi30k_test_2017_flickr 0.71402 50.3 1000 12118
59
+ deu-eng multi30k_test_2017_mscoco 0.55495 32.1 461 5231
60
+ deu-fra multi30k_test_2017_mscoco 0.59307 34.7 461 5484
61
+ eng-deu multi30k_test_2017_mscoco 0.58028 29.7 461 5158
62
+ eng-fra multi30k_test_2017_mscoco 0.73637 54.7 461 5484
63
+ deu-eng multi30k_test_2018_flickr 0.59367 36.7 1071 14689
64
+ deu-fra multi30k_test_2018_flickr 0.57388 31.3 1071 15867
65
+ eng-deu multi30k_test_2018_flickr 0.59998 30.8 1071 13703
66
+ eng-fra multi30k_test_2018_flickr 0.65354 41.6 1071 15867
67
+ eng-fra newsdiscusstest2015 0.63308 37.7 1500 27975
68
+ deu-eng newssyscomb2009 0.55170 28.3 502 11818
69
+ deu-fra newssyscomb2009 0.56021 27.4 502 12331
70
+ deu-spa newssyscomb2009 0.55546 28.1 502 12503
71
+ eng-deu newssyscomb2009 0.53919 23.0 502 11271
72
+ eng-fra newssyscomb2009 0.58384 29.5 502 12331
73
+ eng-spa newssyscomb2009 0.58266 31.0 502 12503
74
+ deu-eng newstest2008 0.54434 27.0 2051 49380
75
+ deu-fra newstest2008 0.55076 26.2 2051 52685
76
+ deu-spa newstest2008 0.54056 25.6 2051 52586
77
+ eng-deu newstest2008 0.52906 23.0 2051 47447
78
+ eng-fra newstest2008 0.55247 26.8 2051 52685
79
+ eng-spa newstest2008 0.56423 29.6 2051 52586
80
+ deu-eng newstest2009 0.53972 26.7 2525 65399
81
+ deu-fra newstest2009 0.53975 25.6 2525 69263
82
+ deu-spa newstest2009 0.53677 25.6 2525 68111
83
+ eng-deu newstest2009 0.53097 22.1 2525 62816
84
+ eng-fra newstest2009 0.57542 29.1 2525 69263
85
+ eng-spa newstest2009 0.57733 29.8 2525 68111
86
+ deu-eng newstest2010 0.58278 30.2 2489 61711
87
+ deu-fra newstest2010 0.57876 29.0 2489 66022
88
+ deu-spa newstest2010 0.59402 32.6 2489 65480
89
+ eng-deu newstest2010 0.54587 25.3 2489 61503
90
+ eng-fra newstest2010 0.59460 32.0 2489 66022
91
+ eng-spa newstest2010 0.61861 36.3 2489 65480
92
+ deu-eng newstest2011 0.55074 26.8 3003 74681
93
+ deu-fra newstest2011 0.55879 27.4 3003 80626
94
+ deu-spa newstest2011 0.56593 30.2 3003 79476
95
+ eng-deu newstest2011 0.52619 22.7 3003 72981
96
+ eng-fra newstest2011 0.60960 34.1 3003 80626
97
+ eng-spa newstest2011 0.62056 38.5 3003 79476
98
+ deu-eng newstest2012 0.56290 28.4 3003 72812
99
+ deu-fra newstest2012 0.55931 27.3 3003 78011
100
+ deu-spa newstest2012 0.57369 31.5 3003 79006
101
+ eng-deu newstest2012 0.52668 23.3 3003 72886
102
+ eng-fra newstest2012 0.59076 31.6 3003 78011
103
+ eng-spa newstest2012 0.62361 38.8 3003 79006
104
+ deu-eng newstest2013 0.58065 31.8 3000 64505
105
+ deu-fra newstest2013 0.56431 30.0 3000 70037
106
+ deu-spa newstest2013 0.56965 31.5 3000 70528
107
+ eng-deu newstest2013 0.55423 26.9 3000 63737
108
+ eng-fra newstest2013 0.58760 33.1 3000 70037
109
+ eng-spa newstest2013 0.59825 35.1 3000 70528
110
+ deu-eng newstest2014 0.59617 32.9 3003 67337
111
+ eng-deu newstest2014 0.58847 28.0 3003 62688
112
+ eng-fra newstest2014 0.65294 39.9 3003 77306
113
+ deu-eng newstest2015 0.59741 33.8 2169 46443
114
+ eng-deu newstest2015 0.59474 31.0 2169 44260
115
+ deu-eng newstest2016 0.64981 40.6 2999 64119
116
+ eng-deu newstest2016 0.63839 37.1 2999 62669
117
+ deu-eng newstest2017 0.60957 35.5 3004 64399
118
+ eng-deu newstest2017 0.58967 30.0 3004 61287
119
+ deu-eng newstest2018 0.66739 43.4 2998 67012
120
+ eng-deu newstest2018 0.68858 44.9 2998 64276
121
+ deu-eng newstest2019 0.63671 39.6 2000 39227
122
+ deu-fra newstest2019 0.63043 36.1 1701 42509
123
+ eng-deu newstest2019 0.65934 41.4 1997 48746
124
+ deu-eng newstest2020 0.60800 34.5 785 38220
125
+ deu-fra newstest2020 0.60544 33.1 1619 36890
126
+ eng-deu newstest2020 0.60078 31.7 1418 52383
127
+ deu-eng newstest2021 0.60048 31.9 1000 20180
128
+ deu-fra newstest2021 0.59590 31.8 1000 23757
129
+ eng-deu newstest2021 0.56133 25.6 1002 27970
130
+ deu-eng newstestALL2020 0.60800 34.5 785 38220
131
+ eng-deu newstestALL2020 0.60078 31.7 1418 52383
132
+ deu-eng newstestB2020 0.60795 34.4 785 37696
133
+ eng-deu newstestB2020 0.59256 31.5 1418 53092
134
+ afr-deu ntrex128 0.55289 25.8 1997 48761
135
+ afr-eng ntrex128 0.72558 51.8 1997 47673
136
+ afr-fra ntrex128 0.56601 29.3 1997 53481
137
+ afr-por ntrex128 0.55396 28.1 1997 51631
138
+ afr-spa ntrex128 0.58558 33.7 1997 54107
139
+ deu-eng ntrex128 0.61722 33.8 1997 47673
140
+ deu-fra ntrex128 0.55908 28.6 1997 53481
141
+ deu-por ntrex128 0.54059 25.7 1997 51631
142
+ deu-spa ntrex128 0.56887 30.8 1997 54107
143
+ eng-deu ntrex128 0.58492 29.8 1997 48761
144
+ eng-fra ntrex128 0.61349 35.2 1997 53481
145
+ eng-por ntrex128 0.59785 33.4 1997 51631
146
+ eng-spa ntrex128 0.63935 40.1 1997 54107
147
+ ltz-deu ntrex128 0.51469 21.9 1997 48761
148
+ ltz-eng ntrex128 0.58627 32.4 1997 47673
149
+ ltz-fra ntrex128 0.50491 22.8 1997 53481
150
+ ltz-por ntrex128 0.45364 18.7 1997 51631
151
+ ltz-spa ntrex128 0.47568 21.6 1997 54107
152
+ nld-deu ntrex128 0.55943 25.7 1997 48761
153
+ nld-eng ntrex128 0.63470 36.1 1997 47673
154
+ nld-fra ntrex128 0.55832 27.5 1997 53481
155
+ nld-por ntrex128 0.54714 27.3 1997 51631
156
+ nld-spa ntrex128 0.57692 32.1 1997 54107
157
+ deu-spa tatoeba-test-v2020-07-28 0.67688 48.5 10000 77529
158
+ eng-deu tatoeba-test-v2020-07-28 0.63096 42.2 10000 83347
159
+ eng-spa tatoeba-test-v2020-07-28 0.70673 52.8 10000 77311
160
+ fry-eng tatoeba-test-v2020-07-28 0.55097 37.3 205 1500
161
+ nds-por tatoeba-test-v2020-07-28 0.58940 39.9 206 1250
162
+ nld-fra tatoeba-test-v2020-07-28 0.68878 51.5 10000 69981
163
+ yid-deu tatoeba-test-v2020-07-28 0.25820 5.1 556 3332
164
+ yid-eng tatoeba-test-v2020-07-28 0.51371 33.4 1168 7741
165
+ afr-spa tatoeba-test-v2021-03-30 0.73380 58.8 450 2795
166
+ deu-eng tatoeba-test-v2021-03-30 0.66385 49.2 12664 105121
167
+ eng-deu tatoeba-test-v2021-03-30 0.62495 40.9 12664 107460
168
+ eng-por tatoeba-test-v2021-03-30 0.71810 52.7 11574 87572
169
+ eng-spa tatoeba-test-v2021-03-30 0.71178 53.7 11940 93423
170
+ gos-eng tatoeba-test-v2021-03-30 0.37022 20.8 1193 5819
171
+ gsw-eng tatoeba-test-v2021-03-30 0.44669 29.3 210 1021
172
+ nds-deu tatoeba-test-v2021-03-30 0.64880 46.2 10000 74571
173
+ nld-eng tatoeba-test-v2021-03-30 0.73889 60.4 11660 81885
174
+ nld-fra tatoeba-test-v2021-03-30 0.68227 47.7 11214 80600
175
+ nld-spa tatoeba-test-v2021-03-30 0.69697 51.2 10083 74753
176
+ yid-deu tatoeba-test-v2021-03-30 0.25287 4.8 830 5045
177
+ yid-eng tatoeba-test-v2021-03-30 0.50799 33.4 1888 11810
178
+ yid-fra tatoeba-test-v2021-03-30 0.55313 31.6 390 2419
179
+ afr-deu tatoeba-test-v2021-08-07 0.68492 48.8 1583 9105
180
+ afr-eng tatoeba-test-v2021-08-07 0.72943 59.6 1374 9622
181
+ afr-spa tatoeba-test-v2021-08-07 0.72793 58.4 448 2783
182
+ deu-deu tatoeba-test-v2021-08-07 0.59840 34.8 2500 20806
183
+ deu-eng tatoeba-test-v2021-08-07 0.65957 48.5 17565 149462
184
+ deu-fra tatoeba-test-v2021-08-07 0.68054 50.2 12418 102721
185
+ deu-por tatoeba-test-v2021-08-07 0.63368 42.8 10000 81482
186
+ deu-spa tatoeba-test-v2021-08-07 0.68198 49.3 10521 82570
187
+ eng-deu tatoeba-test-v2021-08-07 0.62497 40.5 17565 151568
188
+ eng-eng tatoeba-test-v2021-08-07 0.79878 57.3 12062 115106
189
+ eng-fra tatoeba-test-v2021-08-07 0.68567 50.9 12681 106378
190
+ eng-por tatoeba-test-v2021-08-07 0.72204 53.4 13222 105265
191
+ eng-spa tatoeba-test-v2021-08-07 0.72539 55.3 16583 134710
192
+ frr-deu tatoeba-test-v2021-08-07 0.23098 4.3 278 1880
193
+ fry-eng tatoeba-test-v2021-08-07 0.55137 37.0 220 1573
194
+ gos-deu tatoeba-test-v2021-08-07 0.46120 24.7 207 1168
195
+ gos-eng tatoeba-test-v2021-08-07 0.38628 22.3 1154 5635
196
+ gsw-eng tatoeba-test-v2021-08-07 0.43003 27.5 205 990
197
+ ltz-deu tatoeba-test-v2021-08-07 0.48474 32.0 347 2208
198
+ ltz-eng tatoeba-test-v2021-08-07 0.65366 56.4 293 1840
199
+ nds-deu tatoeba-test-v2021-08-07 0.65251 45.9 9999 74564
200
+ nds-eng tatoeba-test-v2021-08-07 0.61858 44.7 2500 17589
201
+ nds-fra tatoeba-test-v2021-08-07 0.60412 43.9 857 5676
202
+ nds-por tatoeba-test-v2021-08-07 0.58778 39.5 207 1256
203
+ nds-spa tatoeba-test-v2021-08-07 0.63404 43.9 923 5540
204
+ nld-deu tatoeba-test-v2021-08-07 0.72998 55.7 10218 74131
205
+ nld-eng tatoeba-test-v2021-08-07 0.74362 60.9 12696 89978
206
+ nld-fra tatoeba-test-v2021-08-07 0.68461 48.0 11548 82974
207
+ nld-por tatoeba-test-v2021-08-07 0.68798 49.3 2500 17326
208
+ nld-spa tatoeba-test-v2021-08-07 0.69971 51.6 10113 74981
209
+ swg-deu tatoeba-test-v2021-08-07 0.39866 13.7 1523 15632
210
+ yid-deu tatoeba-test-v2021-08-07 0.24486 4.7 853 5173
211
+ yid-eng tatoeba-test-v2021-08-07 0.49807 31.5 2483 15452
212
+ yid-fra tatoeba-test-v2021-08-07 0.54147 31.9 384 2455
213
+ yid-spa tatoeba-test-v2021-08-07 0.34871 12.9 407 2478
214
+ eng-fra tico19-test 0.62559 39.5 2100 64661
215
+ eng-por tico19-test 0.72765 49.8 2100 62729
216
+ eng-spa tico19-test 0.72905 51.6 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-gmw-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": 47962,
17
+ "decoder_vocab_size": 47963,
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": 444,
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": 47962,
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": 47963
41
+ }
generation_config.json ADDED
@@ -0,0 +1,16 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bad_words_ids": [
4
+ [
5
+ 47962
6
+ ]
7
+ ],
8
+ "bos_token_id": 0,
9
+ "decoder_start_token_id": 47962,
10
+ "eos_token_id": 444,
11
+ "forced_eos_token_id": 444,
12
+ "max_length": 512,
13
+ "num_beams": 4,
14
+ "pad_token_id": 47962,
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 sha256:2a4d90f33e28e67a522a8d95858799b058617bdca55dc8640c0132f7163a660b
3
+ size 902107420
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:22b8883f41511d323b17db124ec52e5635fc35699d1d23b3baccb411a436b1e5
3
+ size 902158661
source.spm ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:eec711a7a9398a6edbc52f8d0249552bf965f85aee845a17a237783ea95f7ba7
3
+ size 794713
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 sha256:6f0c6c5e91a1196fd729077284ba6d816a278a58f012a834f1beac1f550aeb91
3
+ size 806555
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"source_lang": "gmw", "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/gmw-deu+eng+fra+por+spa", "tokenizer_class": "MarianTokenizer"}
vocab.json ADDED
The diff for this file is too large to render. See raw diff