Daryl Lim Claude Opus 4.6 commited on
Commit
3d1e1ac
·
1 Parent(s): 09377d3

Restrict language mapping to Table 14 evaluated languages

Browse files

Filter langid_to_language from ~400 entries to 183 languages
evaluated in Section A.9 of the MADLAD-400 paper. Only languages
from the WMT, Flores-200, NTREX, and Gatones evaluation sets
are retained.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Files changed (1) hide show
  1. langmap/langid_mapping.py +4 -237
langmap/langid_mapping.py CHANGED
@@ -1,5 +1,7 @@
1
- # Mapping the BCP-47 codes used in the MADLAD-400 models to the language names
 
2
  # [MADLAD-400: A Multilingual And Document-Level Large Audited Dataset](https://arxiv.org/pdf/2309.04662)
 
3
  langid_to_language = {
4
  # Page 16
5
  "<2en>": "English",
@@ -66,12 +68,10 @@ langid_to_language = {
66
  "<2yue>": "Cantonese",
67
  "<2ne>": "Nepali",
68
  "<2kn>": "Kannada",
69
- "<2kaa>": "Kara-Kalpak",
70
  "<2gu>": "Gujarati",
71
  "<2si>": "Sinhala",
72
  "<2cy>": "Welsh",
73
  "<2eo>": "Esperanto",
74
- "<2la>": "Latin",
75
  "<2hy>": "Armenian",
76
  "<2ky>": "Kyrghyz",
77
  "<2tg>": "Tajik",
@@ -88,7 +88,6 @@ langid_to_language = {
88
  "<2lo>": "Lao",
89
  "<2ha>": "Hausa",
90
  "<2dv>": "Dhivehi",
91
- "<2fy>": "W. Frisian",
92
  "<2lb>": "Luxembourgish",
93
  "<2ckb>": "Kurdish (Sorani)",
94
  "<2mg>": "Malagasy",
@@ -96,7 +95,6 @@ langid_to_language = {
96
  "<2am>": "Amharic",
97
  "<2ug>": "Uyghur",
98
  "<2ht>": "Haitian Creole",
99
- "<2grc>": "Ancient Greek",
100
  "<2hmn>": "Hmong",
101
  "<2sd>": "Sindhi",
102
  "<2jv>": "Javanese",
@@ -113,316 +111,85 @@ langid_to_language = {
113
  "<2ny>": "Chichewa",
114
  "<2sm>": "Samoan",
115
  "<2sn>": "Shona",
116
- "<2co>": "Corsican",
117
  "<2zu>": "Zulu",
118
  "<2ig>": "Igbo",
119
  "<2yo>": "Yoruba",
120
  "<2pap>": "Papiamento",
121
  "<2st>": "Sesotho",
122
- "<2haw>": "Hawaiian",
123
  "<2as>": "Assamese",
124
  "<2oc>": "Occitan",
125
  "<2cv>": "Chuvash",
126
  "<2lus>": "Mizo",
127
- "<2tet>": "Tetum",
128
- "<2gsw>": "Swiss German",
129
- "<2sah>": "Yakut",
130
- "<2br>": "Breton",
131
- "<2rm>": "Romansh",
132
  "<2sa>": "Sanskrit",
133
  "<2bo>": "Tibetan",
134
  "<2om>": "Oromo",
135
- "<2se>": "N. Sami",
136
  "<2ce>": "Chechen",
137
- "<2cnh>": "Hakha Chin",
138
  # Page 18
139
  "<2ilo>": "Ilocano",
140
- "<2hil>": "Hiligaynon",
141
- "<2udm>": "Udmurt",
142
- "<2os>": "Ossetian",
143
  "<2lg>": "Luganda",
144
  "<2ti>": "Tigrinya",
145
  "<2vec>": "Venetian",
146
  "<2ts>": "Tsonga",
147
- "<2tyv>": "Tuvinian",
148
- "<2kbd>": "Kabardian",
149
  "<2ee>": "Ewe",
150
- "<2iba>": "Iban",
151
  "<2av>": "Avar",
152
- "<2kha>": "Khasi",
153
  "<2to>": "Tonga (Tonga Islands)",
154
  "<2tn>": "Tswana",
155
  "<2nso>": "Sepedi",
156
  "<2fj>": "Fijian",
157
  "<2zza>": "Zaza",
158
  "<2ak>": "Twi",
159
- "<2ada>": "Adangme",
160
- "<2otq>": "Querétaro Otomi",
161
  "<2dz>": "Dzongkha",
162
- "<2bua>": "Buryat",
163
- "<2cfm>": "Falam Chin",
164
  "<2ln>": "Lingala",
165
- "<2chm>": "Meadow Mari",
166
  "<2gn>": "Guarani",
167
- "<2krc>": "Karachay-Balkar",
168
- "<2wa>": "Walloon",
169
- "<2hif>": "Fiji Hindi",
170
  "<2yua>": "Yucateco",
171
- "<2srn>": "Sranan Tongo",
172
  "<2war>": "Waray (Philippines)",
173
- "<2rom>": "Romani",
174
- "<2bik>": "Central Bikol",
175
- "<2pam>": "Pampanga",
176
  "<2sg>": "Sango",
177
- "<2lu>": "Luba-Katanga",
178
  "<2ady>": "Adyghe",
179
  "<2kbp>": "Kabiyè",
180
- "<2syr>": "Syriac",
181
  "<2ltg>": "Latgalian",
182
- "<2myv>": "Erzya",
183
  "<2iso>": "Isoko",
184
  "<2kac>": "Kachin",
185
  "<2bho>": "Bhojpuri",
186
  "<2ay>": "Aymara",
187
- "<2kum>": "Kumyk",
188
  "<2qu>": "Quechua",
189
- "<2za>": "Zhuang",
190
  "<2pag>": "Pangasinan",
191
- "<2ngu>": "Guerrero Nahuatl",
192
  "<2ve>": "Venda",
193
- "<2pck>": "Paite Chin",
194
- "<2zap>": "Zapotec",
195
- "<2tyz>": "Tày",
196
- "<2hui>": "Huli",
197
  "<2bbc>": "Batak Toba",
198
- "<2tzo>": "Tzotzil",
199
  "<2tiv>": "Tiv",
200
- "<2ksd>": "Kuanua",
201
  "<2gom>": "Goan Konkani",
202
  "<2min>": "Minangkabau",
203
- "<2ang>": "Old English",
204
- "<2nhe>": "E. Huasteca Nahuatl",
205
- "<2bgp>": "E. Baluchi",
206
- "<2nzi>": "Nzima",
207
- "<2nnb>": "Nande",
208
- "<2nv>": "Navajo",
209
- # '<2zxx>': 'Noise',
210
  "<2bci>": "Baoulé",
211
- "<2kv>": "Komi",
212
- "<2new>": "Newari",
213
- "<2mps>": "Dadibi",
214
- "<2alt>": "S. Altai",
215
- "<2meu>": "Motu",
216
  "<2bew>": "Betawi",
217
  "<2fon>": "Fon",
218
- "<2iu>": "Inuktitut",
219
- "<2abt>": "Ambulas",
220
  # Page 19
221
- "<2mgh>": "Makhuwa-Meetto",
222
- "<2mnw>": "Mon",
223
- "<2tvl>": "Tuvalu",
224
- "<2dov>": "Dombe",
225
- "<2tlh>": "Klingon",
226
- "<2ho>": "Hiri Motu",
227
- "<2kw>": "Cornish",
228
- "<2mrj>": "Hill Mari",
229
  "<2meo>": "Kedah Malay",
230
  "<2crh>": "Crimean Tatar",
231
- "<2mbt>": "Matigsalug Manobo",
232
- "<2emp>": "N. Emberá",
233
  "<2ace>": "Achinese",
234
- "<2ium>": "Iu Mien",
235
- "<2mam>": "Mam",
236
- "<2gym>": "Ngäbere",
237
  "<2mai>": "Maithili",
238
- "<2crs>": "Seselwa Creole French",
239
- "<2pon>": "Pohnpeian",
240
- "<2ubu>": "Umbu-Ungu",
241
- "<2fip>": "Fipa",
242
- "<2quc>": "K’iche’",
243
- "<2gv>": "Manx",
244
- "<2kj>": "Kuanyama",
245
- "<2btx>": "Batak Karo",
246
- "<2ape>": "Bukiyip",
247
- "<2chk>": "Chuukese",
248
- "<2rcf>": "Réunion Creole French",
249
  "<2shn>": "Shan",
250
- "<2tzh>": "Tzeltal",
251
- "<2mdf>": "Moksha",
252
- "<2ppk>": "Uma",
253
  "<2ss>": "Swati",
254
- "<2gag>": "Gagauz",
255
- "<2cab>": "Garifuna",
256
  "<2kri>": "Krio",
257
- "<2seh>": "Sena",
258
- "<2ibb>": "Ibibio",
259
- "<2tbz>": "Ditammari",
260
- "<2bru>": "E. Bru",
261
- "<2enq>": "Enga",
262
- "<2ach>": "Acoli",
263
- "<2cuk>": "San Blas Kuna",
264
  "<2kmb>": "Kimbundu",
265
  "<2wo>": "Wolof",
266
- "<2kek>": "Kekchí",
267
- "<2qub>": "Huallaga Huánuco Quechua",
268
- "<2tab>": "Tabassaran",
269
- "<2bts>": "Batak Simalungun",
270
- "<2kos>": "Kosraean",
271
- "<2rwo>": "Rawa",
272
- "<2cak>": "Kaqchikel",
273
- "<2tuc>": "Mutu",
274
- "<2bum>": "Bulu",
275
- "<2cjk>": "Chokwe",
276
- "<2gil>": "Gilbertese",
277
- "<2stq>": "Saterfriesisch",
278
- "<2tsg>": "Tausug",
279
- "<2quh>": "S. Bolivian Quechua",
280
- "<2mak>": "Makasar",
281
- "<2arn>": "Mapudungun",
282
  "<2ban>": "Balinese",
283
- "<2jiv>": "Shuar",
284
- "<2sja>": "Epena",
285
- "<2yap>": "Yapese",
286
- "<2tcy>": "Tulu",
287
- "<2toj>": "Tojolabal",
288
- "<2twu>": "Termanu",
289
- "<2xal>": "Kalmyk",
290
- "<2amu>": "Guerrero Amuzgo",
291
- "<2rmc>": "Carpathian Romani",
292
- "<2hus>": "Huastec",
293
- "<2nia>": "Nias",
294
- "<2kjh>": "Khakas",
295
  "<2bm>": "Bambara",
296
- "<2guh>": "Guahibo",
297
- "<2mas>": "Masai",
298
- "<2acf>": "St Lucian Creole French",
299
- "<2dtp>": "Kadazan Dusun",
300
- "<2ksw>": "S’gaw Karen",
301
- "<2bzj>": "Belize Kriol English",
302
  # Page 20
303
  "<2din>": "Dinka",
304
- "<2zne>": "Zande",
305
  "<2mad>": "Madurese",
306
- "<2msi>": "Sabah Malay",
307
  "<2mag>": "Magahi",
308
- "<2mkn>": "Kupang Malay",
309
  "<2kg>": "Kongo",
310
- "<2lhu>": "Lahu",
311
- "<2ch>": "Chamorro",
312
- "<2qvi>": "Imbabura H. Quichua",
313
- "<2mh>": "Marshallese",
314
- "<2djk>": "E. Maroon Creole",
315
- "<2sus>": "Susu",
316
- "<2mfe>": "Morisien",
317
- "<2srm>": "Saramaccan",
318
  "<2dyu>": "Dyula",
319
- "<2ctu>": "Chol",
320
- "<2gui>": "E. Bolivian Guaraní",
321
- "<2pau>": "Palauan",
322
- "<2inb>": "Inga",
323
- "<2bi>": "Bislama",
324
  "<2mni>": "Meiteilon (Manipuri)",
325
- "<2guc>": "Wayuu",
326
- "<2jam>": "Jamaican Creole English",
327
- "<2wal>": "Wolaytta",
328
- "<2jac>": "Popti’",
329
- "<2bas>": "Basa (Cameroon)",
330
- "<2gor>": "Gorontalo",
331
  "<2skr>": "Saraiki",
332
- "<2nyu>": "Nyungwe",
333
- "<2noa>": "Woun Meu",
334
- "<2sda>": "Toraja-Sa’dan",
335
- "<2gub>": "Guajajára",
336
- "<2nog>": "Nogai",
337
- "<2cni>": "Asháninka",
338
- "<2teo>": "Teso",
339
- "<2tdx>": "Tandroy-Mahafaly Malagasy",
340
- "<2sxn>": "Sangir",
341
- "<2rki>": "Rakhine",
342
- "<2nr>": "South Ndebele",
343
- "<2frp>": "Arpitan",
344
- "<2alz>": "Alur",
345
- "<2taj>": "E. Tamang",
346
- "<2lrc>": "N. Luri",
347
- "<2cce>": "Chopi",
348
  "<2rn>": "Rundi",
349
- "<2jvn>": "Caribbean Javanese",
350
- "<2hvn>": "Sabu",
351
- "<2nij>": "Ngaju",
352
- "<2dwr>": "Dawro",
353
- "<2izz>": "Izii",
354
- "<2msm>": "Agusan Manobo",
355
- "<2bus>": "Bokobaru",
356
- "<2ktu>": "Kituba (DRC)",
357
  "<2chr>": "Cherokee",
358
- "<2maz>": "Central Mazahua",
359
- "<2tzj>": "Tz’utujil",
360
- "<2suz>": "Sunwar",
361
- "<2knj>": "W. Kanjobal",
362
- "<2bim>": "Bimoba",
363
- "<2gvl>": "Gulay",
364
- "<2bqc>": "Boko (Benin)",
365
- "<2tca>": "Ticuna",
366
- "<2pis>": "Pijin",
367
- "<2prk>": "Parauk",
368
- "<2laj>": "Lango (Uganda)",
369
- "<2mel>": "Central Melanau",
370
- "<2qxr>": "Cañar H. Quichua",
371
- "<2niq>": "Nandi",
372
- "<2ahk>": "Akha",
373
- "<2shp>": "Shipibo-Conibo",
374
  "<2hne>": "Chhattisgarhi",
375
- "<2spp>": "Supyyire Senoufo",
376
- "<2koi>": "Komi-Permyak",
377
- "<2krj>": "Kinaray-A",
378
- "<2quf>": "Lambayeque Quechua",
379
- "<2luz>": "S. Luri",
380
- "<2agr>": "Aguaruna",
381
- "<2tsc>": "Tswa",
382
- "<2mqy>": "Manggarai",
383
- "<2gof>": "Gofa",
384
  # Page 21
385
- "<2gbm>": "Garhwali",
386
- "<2miq>": "Mískito",
387
- "<2dje>": "Zarma",
388
  "<2awa>": "Awadhi",
389
- "<2bjj>": "Kanauji",
390
- "<2qvz>": "N. Pastaza Quichua",
391
- "<2sjp>": "Surjapuri",
392
- "<2tll>": "Tetela",
393
- "<2raj>": "Rajasthani",
394
- "<2kjg>": "Khmu",
395
- "<2bgz>": "Banggai",
396
  "<2quy>": "Ayacucho Quechua",
397
- "<2cbk>": "Chavacano",
398
- "<2akb>": "Batak Angkola",
399
- "<2oj>": "Ojibwa",
400
- "<2ify>": "Keley-I Kallahan",
401
  "<2mey>": "Hassaniyya",
402
  "<2ks>": "Kashimiri",
403
- "<2cac>": "Chuj",
404
- "<2brx>": "Bodo (India)",
405
- "<2qup>": "S. Pastaza Quechua",
406
- "<2syl>": "Sylheti",
407
- "<2jax>": "Jambi Malay",
408
  "<2ff>": "Fulfulde",
409
  "<2ber>": "Tamazight (Tfng)",
410
- "<2tks>": "Takestani",
411
- "<2trp>": "Kok Borok",
412
- "<2mrw>": "Maranao",
413
- "<2adh>": "Adhola",
414
- "<2smt>": "Simte",
415
- "<2srr>": "Serer",
416
- "<2ffm>": "Maasina Fulfulde",
417
- "<2qvc>": "Cajamarca Quechua",
418
- "<2mtr>": "Mewari",
419
- "<2ann>": "Obolo",
420
- "<2kaa-Latn>": "Kara-Kalpak (Latn)",
421
- "<2aa>": "Afar",
422
- "<2noe>": "Nimadi",
423
- "<2nut>": "Nung (Viet Nam)",
424
- "<2gyn>": "Guyanese Creole English",
425
- "<2kwi>": "Awa-Cuaiquer",
426
- "<2xmm>": "Manado Malay",
427
- "<2msb>": "Masbatenyo",
428
  }
 
1
+ # Mapping the BCP-47 codes used in the MADLAD-400 models to the language names.
2
+ # Restricted to languages evaluated in Table 14, Section A.9 of:
3
  # [MADLAD-400: A Multilingual And Document-Level Large Audited Dataset](https://arxiv.org/pdf/2309.04662)
4
+ # Evaluation sets: WMT, Flores-200, NTREX, Gatones
5
  langid_to_language = {
6
  # Page 16
7
  "<2en>": "English",
 
68
  "<2yue>": "Cantonese",
69
  "<2ne>": "Nepali",
70
  "<2kn>": "Kannada",
 
71
  "<2gu>": "Gujarati",
72
  "<2si>": "Sinhala",
73
  "<2cy>": "Welsh",
74
  "<2eo>": "Esperanto",
 
75
  "<2hy>": "Armenian",
76
  "<2ky>": "Kyrghyz",
77
  "<2tg>": "Tajik",
 
88
  "<2lo>": "Lao",
89
  "<2ha>": "Hausa",
90
  "<2dv>": "Dhivehi",
 
91
  "<2lb>": "Luxembourgish",
92
  "<2ckb>": "Kurdish (Sorani)",
93
  "<2mg>": "Malagasy",
 
95
  "<2am>": "Amharic",
96
  "<2ug>": "Uyghur",
97
  "<2ht>": "Haitian Creole",
 
98
  "<2hmn>": "Hmong",
99
  "<2sd>": "Sindhi",
100
  "<2jv>": "Javanese",
 
111
  "<2ny>": "Chichewa",
112
  "<2sm>": "Samoan",
113
  "<2sn>": "Shona",
 
114
  "<2zu>": "Zulu",
115
  "<2ig>": "Igbo",
116
  "<2yo>": "Yoruba",
117
  "<2pap>": "Papiamento",
118
  "<2st>": "Sesotho",
 
119
  "<2as>": "Assamese",
120
  "<2oc>": "Occitan",
121
  "<2cv>": "Chuvash",
122
  "<2lus>": "Mizo",
 
 
 
 
 
123
  "<2sa>": "Sanskrit",
124
  "<2bo>": "Tibetan",
125
  "<2om>": "Oromo",
 
126
  "<2ce>": "Chechen",
 
127
  # Page 18
128
  "<2ilo>": "Ilocano",
 
 
 
129
  "<2lg>": "Luganda",
130
  "<2ti>": "Tigrinya",
131
  "<2vec>": "Venetian",
132
  "<2ts>": "Tsonga",
 
 
133
  "<2ee>": "Ewe",
 
134
  "<2av>": "Avar",
 
135
  "<2to>": "Tonga (Tonga Islands)",
136
  "<2tn>": "Tswana",
137
  "<2nso>": "Sepedi",
138
  "<2fj>": "Fijian",
139
  "<2zza>": "Zaza",
140
  "<2ak>": "Twi",
 
 
141
  "<2dz>": "Dzongkha",
 
 
142
  "<2ln>": "Lingala",
 
143
  "<2gn>": "Guarani",
 
 
 
144
  "<2yua>": "Yucateco",
 
145
  "<2war>": "Waray (Philippines)",
 
 
 
146
  "<2sg>": "Sango",
 
147
  "<2ady>": "Adyghe",
148
  "<2kbp>": "Kabiyè",
 
149
  "<2ltg>": "Latgalian",
 
150
  "<2iso>": "Isoko",
151
  "<2kac>": "Kachin",
152
  "<2bho>": "Bhojpuri",
153
  "<2ay>": "Aymara",
 
154
  "<2qu>": "Quechua",
 
155
  "<2pag>": "Pangasinan",
 
156
  "<2ve>": "Venda",
 
 
 
 
157
  "<2bbc>": "Batak Toba",
 
158
  "<2tiv>": "Tiv",
 
159
  "<2gom>": "Goan Konkani",
160
  "<2min>": "Minangkabau",
 
 
 
 
 
 
 
161
  "<2bci>": "Baoulé",
 
 
 
 
 
162
  "<2bew>": "Betawi",
163
  "<2fon>": "Fon",
 
 
164
  # Page 19
 
 
 
 
 
 
 
 
165
  "<2meo>": "Kedah Malay",
166
  "<2crh>": "Crimean Tatar",
 
 
167
  "<2ace>": "Achinese",
 
 
 
168
  "<2mai>": "Maithili",
169
+ "<2quc>": "K\u2019iche\u2019",
 
 
 
 
 
 
 
 
 
 
170
  "<2shn>": "Shan",
 
 
 
171
  "<2ss>": "Swati",
 
 
172
  "<2kri>": "Krio",
 
 
 
 
 
 
 
173
  "<2kmb>": "Kimbundu",
174
  "<2wo>": "Wolof",
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
175
  "<2ban>": "Balinese",
 
 
 
 
 
 
 
 
 
 
 
 
176
  "<2bm>": "Bambara",
 
 
 
 
 
 
177
  # Page 20
178
  "<2din>": "Dinka",
 
179
  "<2mad>": "Madurese",
 
180
  "<2mag>": "Magahi",
 
181
  "<2kg>": "Kongo",
 
 
 
 
 
 
 
 
182
  "<2dyu>": "Dyula",
 
 
 
 
 
183
  "<2mni>": "Meiteilon (Manipuri)",
 
 
 
 
 
 
184
  "<2skr>": "Saraiki",
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
185
  "<2rn>": "Rundi",
 
 
 
 
 
 
 
 
186
  "<2chr>": "Cherokee",
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
187
  "<2hne>": "Chhattisgarhi",
 
 
 
 
 
 
 
 
 
188
  # Page 21
 
 
 
189
  "<2awa>": "Awadhi",
 
 
 
 
 
 
 
190
  "<2quy>": "Ayacucho Quechua",
 
 
 
 
191
  "<2mey>": "Hassaniyya",
192
  "<2ks>": "Kashimiri",
 
 
 
 
 
193
  "<2ff>": "Fulfulde",
194
  "<2ber>": "Tamazight (Tfng)",
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
195
  }