--- license: odc-by task_categories: - text-generation - translation pretty_name: FineTranslations size_categories: - n>1T language: - abk - abq - abs - acm - adh - adi - ady - aeb - afr - agx - aii - aim - ain - ajz - akb - aln - als - alt - amh - anp - aoz - apc - apt - arb - arg - arq - ars - ary - arz - asm - ast - atb - ava - awa - ayp - ayr - azb - azj - bak - bam - ban - bar - bas - bbc - bbk - bcl - bdq - bel - ben - bew - bho - bhp - bis - biu - bjn - bod - bos - brh - brx - bts - btx - bug - bul - bwi - bxr - cat - cbk - ccp - ceb - ces - cfm - cha - che - chr - chu - chv - cjs - ckb - ckt - cmn - cnh - cnw - cos - crh - crj - crk - crl - crs - csb - csw - csy - ctd - cym - czt - dak - dan - dar - deu - dik - diu - div - dje - dks - dln - dng - dnw - doi - dru - dsb - dtp - dty - dzo - ekk - ell - enl - enm - epo - ess - eus - eve - ewo - ext - fao - fas - ffm - fij - fil - fin - fit - fkv - fmu - fra - fro - frp - fry - fuf - fur - fuv - gag - gaz - gcf - gla - gle - glg - glk - glv - gmh - gnb - goh - gom - gos - grc - gsw - gug - guj - guz - hac - hae - hak - hat - hau - haw - hbo - heb - her - hif - hil - hin - hmr - hne - hns - hrv - hrx - hsb - hun - hwc - hye - hyw - iba - ibg - ibo - ife - ike - ikt - ilo - ina - ind - inh - isl - ita - ivv - jav - jpn - jun - kaa - kab - kac - kak - kal - kam - kan - kas - kat - kaz - kbd - kca - kdh - kdr - kea - kei - kgp - kha - khk - khm - kik - kin - kir - kiu - kjb - kjh - kmr - knc - koi - kor - kos - kpv - krj - krl - kru - ksh - ksw - ktj - ktz - kua - kum - kwn - kyu - kzj - lad - lao - lat - lbe - ldn - lew - lez - lfn - lim - lin - lis - lit - lki - lld - lmk - lnd - lrc - ltg - ltz - lud - lug - luo - lus - lvs - lwg - lzh - mag - mah - mai - mak - mal - mar - mas - mbf - mdf - mer - mfe - mfg - mfy - mhi - mhr - mhy - min - mip - mjw - mkd - mlt - mni - mnk - mns - mnw - moh - mph - mqy - mri - mrj - mrw - mtg - mui - mup - mus - mvp - mwf - mwl - mww - mya - myv - myx - mzh - nah - nan - nap - naq - nbu - nde - ndo - nds - new - nio - njn - njo - nld - nmf - nmz - nno - nob - nog - non - npi - npo - nrf - nri - nrm - nse - nus - nya - nyn - nzm - obo - oci - ojb - olo - orv - ory - oss - ota - oto - otw - pam - pan - pap - pbt - pcd - pck - pcm - pfl - plt - pmq - pmx - pnb - pnt - pol - por - pov - ppk - pps - prg - pui - pxm - quc - qul - qup - qus - quz - raw - rcf - rel - rhg - ria - rjs - rmc - rml - rmn - rmy - rnl - roh - ron - rtm - rue - run - rus - sah - san - sat - sck - scn - sda - sdc - sdh - ses - sgc - sgh - sid - sin - sju - skr - slk - slv - sma - sme - smj - smn - smo - sms - smt - sna - snd - som - sot - spa - srd - srp - ssw - sun - swe - swg - swh - syc - syl - szl - tab - tam - taq - tat - tcy - tcz - tel - tet - tgk - tha - thl - tig - tir - tkl - tkr - tlh - tly - tok - ton - tpi - tpw - trc - trp - trs - ttj - tuk - tur - tuv - twx - tyv - tzl - tzm - udm - uig - ukr - urd - uzn - uzs - vap - vie - vot - vro - war - way - wba - wbm - wes - whk - wlx - wol - wsg - wwa - xal - xho - xmm - xmv - xog - yaz - ydd - yor - yrk - yrl - yua - yue - zea - zgh - zom - zsm - zul configs: - config_name: all data_files: data/*/* - config_name: abk_Cyrl data_files: data/abk_Cyrl/* - config_name: abq_Cyrl data_files: data/abq_Cyrl/* - config_name: abs_Latn data_files: data/abs_Latn/* - config_name: acm_Arab data_files: data/acm_Arab/* - config_name: adh_Latn data_files: data/adh_Latn/* - config_name: adi_Latn data_files: data/adi_Latn/* - config_name: ady_Cyrl data_files: data/ady_Cyrl/* - config_name: aeb_Arab data_files: data/aeb_Arab/* - config_name: afr_Latn data_files: data/afr_Latn/* - config_name: agx_Cyrl data_files: data/agx_Cyrl/* - config_name: aii_Syrc data_files: data/aii_Syrc/* - config_name: aim_Latn data_files: data/aim_Latn/* - config_name: ain_Latn data_files: data/ain_Latn/* - config_name: ajz_Latn data_files: data/ajz_Latn/* - config_name: akb_Latn data_files: data/akb_Latn/* - config_name: aln_Latn data_files: data/aln_Latn/* - config_name: als_Latn data_files: data/als_Latn/* - config_name: alt_Cyrl data_files: data/alt_Cyrl/* - config_name: amh_Ethi data_files: data/amh_Ethi/* - config_name: anp_Deva data_files: data/anp_Deva/* - config_name: aoz_Latn data_files: data/aoz_Latn/* - config_name: apc_Arab data_files: data/apc_Arab/* - config_name: apt_Latn data_files: data/apt_Latn/* - config_name: arb_Arab data_files: data/arb_Arab/* - config_name: arb_Latn data_files: data/arb_Latn/* - config_name: arg_Latn data_files: data/arg_Latn/* - config_name: arq_Arab data_files: data/arq_Arab/* - config_name: ars_Arab data_files: data/ars_Arab/* - config_name: ary_Arab data_files: data/ary_Arab/* - config_name: arz_Arab data_files: data/arz_Arab/* - config_name: asm_Beng data_files: data/asm_Beng/* - config_name: asm_Latn data_files: data/asm_Latn/* - config_name: ast_Latn data_files: data/ast_Latn/* - config_name: atb_Latn data_files: data/atb_Latn/* - config_name: ava_Cyrl data_files: data/ava_Cyrl/* - config_name: awa_Deva data_files: data/awa_Deva/* - config_name: ayp_Arab data_files: data/ayp_Arab/* - config_name: ayr_Latn data_files: data/ayr_Latn/* - config_name: azb_Arab data_files: data/azb_Arab/* - config_name: azj_Latn data_files: data/azj_Latn/* - config_name: bak_Cyrl data_files: data/bak_Cyrl/* - config_name: bam_Latn data_files: data/bam_Latn/* - config_name: ban_Latn data_files: data/ban_Latn/* - config_name: bar_Latn data_files: data/bar_Latn/* - config_name: bas_Latn data_files: data/bas_Latn/* - config_name: bbc_Latn data_files: data/bbc_Latn/* - config_name: bbk_Latn data_files: data/bbk_Latn/* - config_name: bcl_Latn data_files: data/bcl_Latn/* - config_name: bdq_Latn data_files: data/bdq_Latn/* - config_name: bel_Cyrl data_files: data/bel_Cyrl/* - config_name: ben_Beng data_files: data/ben_Beng/* - config_name: ben_Latn data_files: data/ben_Latn/* - config_name: bew_Latn data_files: data/bew_Latn/* - config_name: bho_Deva data_files: data/bho_Deva/* - config_name: bhp_Latn data_files: data/bhp_Latn/* - config_name: bis_Latn data_files: data/bis_Latn/* - config_name: biu_Latn data_files: data/biu_Latn/* - config_name: bjn_Arab data_files: data/bjn_Arab/* - config_name: bjn_Latn data_files: data/bjn_Latn/* - config_name: bod_Tibt data_files: data/bod_Tibt/* - config_name: bos_Latn data_files: data/bos_Latn/* - config_name: brh_Arab data_files: data/brh_Arab/* - config_name: brx_Deva data_files: data/brx_Deva/* - config_name: bts_Latn data_files: data/bts_Latn/* - config_name: btx_Latn data_files: data/btx_Latn/* - config_name: bug_Latn data_files: data/bug_Latn/* - config_name: bul_Cyrl data_files: data/bul_Cyrl/* - config_name: bwi_Latn data_files: data/bwi_Latn/* - config_name: bxr_Cyrl data_files: data/bxr_Cyrl/* - config_name: cat_Latn data_files: data/cat_Latn/* - config_name: cbk_Latn data_files: data/cbk_Latn/* - config_name: ccp_Latn data_files: data/ccp_Latn/* - config_name: ceb_Latn data_files: data/ceb_Latn/* - config_name: ces_Latn data_files: data/ces_Latn/* - config_name: cfm_Latn data_files: data/cfm_Latn/* - config_name: cha_Latn data_files: data/cha_Latn/* - config_name: che_Cyrl data_files: data/che_Cyrl/* - config_name: chr_Latn data_files: data/chr_Latn/* - config_name: chu_Cyrl data_files: data/chu_Cyrl/* - config_name: chv_Cyrl data_files: data/chv_Cyrl/* - config_name: cjs_Cyrl data_files: data/cjs_Cyrl/* - config_name: ckb_Arab data_files: data/ckb_Arab/* - config_name: ckt_Cyrl data_files: data/ckt_Cyrl/* - config_name: cmn_Hani data_files: data/cmn_Hani/* - config_name: cnh_Latn data_files: data/cnh_Latn/* - config_name: cnw_Latn data_files: data/cnw_Latn/* - config_name: cos_Latn data_files: data/cos_Latn/* - config_name: crh_Cyrl data_files: data/crh_Cyrl/* - config_name: crh_Latn data_files: data/crh_Latn/* - config_name: crj_Cans data_files: data/crj_Cans/* - config_name: crk_Cans data_files: data/crk_Cans/* - config_name: crk_Latn data_files: data/crk_Latn/* - config_name: crl_Cans data_files: data/crl_Cans/* - config_name: crs_Latn data_files: data/crs_Latn/* - config_name: csb_Latn data_files: data/csb_Latn/* - config_name: csw_Latn data_files: data/csw_Latn/* - config_name: csy_Latn data_files: data/csy_Latn/* - config_name: ctd_Latn data_files: data/ctd_Latn/* - config_name: cym_Latn data_files: data/cym_Latn/* - config_name: czt_Latn data_files: data/czt_Latn/* - config_name: dak_Latn data_files: data/dak_Latn/* - config_name: dan_Latn data_files: data/dan_Latn/* - config_name: dar_Cyrl data_files: data/dar_Cyrl/* - config_name: deu_Latn data_files: data/deu_Latn/* - config_name: dik_Latn data_files: data/dik_Latn/* - config_name: diu_Latn data_files: data/diu_Latn/* - config_name: div_Thaa data_files: data/div_Thaa/* - config_name: dje_Latn data_files: data/dje_Latn/* - config_name: dks_Latn data_files: data/dks_Latn/* - config_name: dln_Latn data_files: data/dln_Latn/* - config_name: dng_Cyrl data_files: data/dng_Cyrl/* - config_name: dnw_Latn data_files: data/dnw_Latn/* - config_name: doi_Deva data_files: data/doi_Deva/* - config_name: dru_Latn data_files: data/dru_Latn/* - config_name: dsb_Latn data_files: data/dsb_Latn/* - config_name: dtp_Latn data_files: data/dtp_Latn/* - config_name: dty_Deva data_files: data/dty_Deva/* - config_name: dzo_Tibt data_files: data/dzo_Tibt/* - config_name: ekk_Latn data_files: data/ekk_Latn/* - config_name: ell_Grek data_files: data/ell_Grek/* - config_name: enl_Latn data_files: data/enl_Latn/* - config_name: enm_Latn data_files: data/enm_Latn/* - config_name: epo_Latn data_files: data/epo_Latn/* - config_name: ess_Latn data_files: data/ess_Latn/* - config_name: eus_Latn data_files: data/eus_Latn/* - config_name: eve_Cyrl data_files: data/eve_Cyrl/* - config_name: ewo_Latn data_files: data/ewo_Latn/* - config_name: ext_Latn data_files: data/ext_Latn/* - config_name: fao_Latn data_files: data/fao_Latn/* - config_name: fas_Arab data_files: data/fas_Arab/* - config_name: ffm_Latn data_files: data/ffm_Latn/* - config_name: fij_Latn data_files: data/fij_Latn/* - config_name: fil_Latn data_files: data/fil_Latn/* - config_name: fin_Latn data_files: data/fin_Latn/* - config_name: fit_Latn data_files: data/fit_Latn/* - config_name: fkv_Latn data_files: data/fkv_Latn/* - config_name: fmu_Deva data_files: data/fmu_Deva/* - config_name: fra_Latn data_files: data/fra_Latn/* - config_name: fro_Latn data_files: data/fro_Latn/* - config_name: frp_Latn data_files: data/frp_Latn/* - config_name: fry_Latn data_files: data/fry_Latn/* - config_name: fuf_Latn data_files: data/fuf_Latn/* - config_name: fur_Latn data_files: data/fur_Latn/* - config_name: fuv_Latn data_files: data/fuv_Latn/* - config_name: gag_Latn data_files: data/gag_Latn/* - config_name: gaz_Latn data_files: data/gaz_Latn/* - config_name: gcf_Latn data_files: data/gcf_Latn/* - config_name: gla_Latn data_files: data/gla_Latn/* - config_name: gle_Latn data_files: data/gle_Latn/* - config_name: glg_Latn data_files: data/glg_Latn/* - config_name: glk_Arab data_files: data/glk_Arab/* - config_name: glv_Latn data_files: data/glv_Latn/* - config_name: gmh_Latn data_files: data/gmh_Latn/* - config_name: gnb_Latn data_files: data/gnb_Latn/* - config_name: goh_Latn data_files: data/goh_Latn/* - config_name: gom_Deva data_files: data/gom_Deva/* - config_name: gom_Latn data_files: data/gom_Latn/* - config_name: gos_Latn data_files: data/gos_Latn/* - config_name: grc_Grek data_files: data/grc_Grek/* - config_name: gsw_Latn data_files: data/gsw_Latn/* - config_name: gug_Latn data_files: data/gug_Latn/* - config_name: guj_Gujr data_files: data/guj_Gujr/* - config_name: guj_Latn data_files: data/guj_Latn/* - config_name: guz_Latn data_files: data/guz_Latn/* - config_name: hac_Arab data_files: data/hac_Arab/* - config_name: hae_Latn data_files: data/hae_Latn/* - config_name: hak_Hani data_files: data/hak_Hani/* - config_name: hat_Latn data_files: data/hat_Latn/* - config_name: hau_Latn data_files: data/hau_Latn/* - config_name: haw_Latn data_files: data/haw_Latn/* - config_name: hbo_Hebr data_files: data/hbo_Hebr/* - config_name: heb_Hebr data_files: data/heb_Hebr/* - config_name: her_Latn data_files: data/her_Latn/* - config_name: hif_Latn data_files: data/hif_Latn/* - config_name: hil_Latn data_files: data/hil_Latn/* - config_name: hin_Deva data_files: data/hin_Deva/* - config_name: hin_Latn data_files: data/hin_Latn/* - config_name: hmr_Latn data_files: data/hmr_Latn/* - config_name: hne_Deva data_files: data/hne_Deva/* - config_name: hns_Latn data_files: data/hns_Latn/* - config_name: hrv_Latn data_files: data/hrv_Latn/* - config_name: hrx_Latn data_files: data/hrx_Latn/* - config_name: hsb_Latn data_files: data/hsb_Latn/* - config_name: hun_Latn data_files: data/hun_Latn/* - config_name: hwc_Latn data_files: data/hwc_Latn/* - config_name: hye_Armn data_files: data/hye_Armn/* - config_name: hyw_Armn data_files: data/hyw_Armn/* - config_name: iba_Latn data_files: data/iba_Latn/* - config_name: ibg_Latn data_files: data/ibg_Latn/* - config_name: ibo_Latn data_files: data/ibo_Latn/* - config_name: ife_Latn data_files: data/ife_Latn/* - config_name: ike_Cans data_files: data/ike_Cans/* - config_name: ikt_Latn data_files: data/ikt_Latn/* - config_name: ilo_Latn data_files: data/ilo_Latn/* - config_name: ina_Latn data_files: data/ina_Latn/* - config_name: ind_Latn data_files: data/ind_Latn/* - config_name: inh_Cyrl data_files: data/inh_Cyrl/* - config_name: isl_Latn data_files: data/isl_Latn/* - config_name: ita_Latn data_files: data/ita_Latn/* - config_name: ivv_Latn data_files: data/ivv_Latn/* - config_name: jav_Latn data_files: data/jav_Latn/* - config_name: jpn_Jpan data_files: data/jpn_Jpan/* - config_name: jun_Orya data_files: data/jun_Orya/* - config_name: kaa_Cyrl data_files: data/kaa_Cyrl/* - config_name: kaa_Latn data_files: data/kaa_Latn/* - config_name: kab_Latn data_files: data/kab_Latn/* - config_name: kac_Latn data_files: data/kac_Latn/* - config_name: kak_Latn data_files: data/kak_Latn/* - config_name: kal_Latn data_files: data/kal_Latn/* - config_name: kam_Latn data_files: data/kam_Latn/* - config_name: kan_Knda data_files: data/kan_Knda/* - config_name: kan_Latn data_files: data/kan_Latn/* - config_name: kas_Deva data_files: data/kas_Deva/* - config_name: kas_Latn data_files: data/kas_Latn/* - config_name: kat_Geor data_files: data/kat_Geor/* - config_name: kaz_Cyrl data_files: data/kaz_Cyrl/* - config_name: kbd_Cyrl data_files: data/kbd_Cyrl/* - config_name: kca_Cyrl data_files: data/kca_Cyrl/* - config_name: kdh_Latn data_files: data/kdh_Latn/* - config_name: kdr_Latn data_files: data/kdr_Latn/* - config_name: kea_Latn data_files: data/kea_Latn/* - config_name: kei_Latn data_files: data/kei_Latn/* - config_name: kgp_Latn data_files: data/kgp_Latn/* - config_name: kha_Latn data_files: data/kha_Latn/* - config_name: khk_Cyrl data_files: data/khk_Cyrl/* - config_name: khm_Khmr data_files: data/khm_Khmr/* - config_name: kik_Latn data_files: data/kik_Latn/* - config_name: kin_Latn data_files: data/kin_Latn/* - config_name: kir_Cyrl data_files: data/kir_Cyrl/* - config_name: kiu_Latn data_files: data/kiu_Latn/* - config_name: kjb_Latn data_files: data/kjb_Latn/* - config_name: kjh_Cyrl data_files: data/kjh_Cyrl/* - config_name: kmr_Cyrl data_files: data/kmr_Cyrl/* - config_name: kmr_Latn data_files: data/kmr_Latn/* - config_name: knc_Latn data_files: data/knc_Latn/* - config_name: koi_Cyrl data_files: data/koi_Cyrl/* - config_name: kor_Hang data_files: data/kor_Hang/* - config_name: kos_Latn data_files: data/kos_Latn/* - config_name: kpv_Cyrl data_files: data/kpv_Cyrl/* - config_name: krj_Latn data_files: data/krj_Latn/* - config_name: krl_Latn data_files: data/krl_Latn/* - config_name: kru_Deva data_files: data/kru_Deva/* - config_name: ksh_Latn data_files: data/ksh_Latn/* - config_name: ksw_Mymr data_files: data/ksw_Mymr/* - config_name: ktj_Latn data_files: data/ktj_Latn/* - config_name: ktz_Latn data_files: data/ktz_Latn/* - config_name: kua_Latn data_files: data/kua_Latn/* - config_name: kum_Cyrl data_files: data/kum_Cyrl/* - config_name: kwn_Latn data_files: data/kwn_Latn/* - config_name: kyu_Kali data_files: data/kyu_Kali/* - config_name: kzj_Latn data_files: data/kzj_Latn/* - config_name: lad_Latn data_files: data/lad_Latn/* - config_name: lao_Laoo data_files: data/lao_Laoo/* - config_name: lat_Latn data_files: data/lat_Latn/* - config_name: lbe_Cyrl data_files: data/lbe_Cyrl/* - config_name: ldn_Latn data_files: data/ldn_Latn/* - config_name: lew_Latn data_files: data/lew_Latn/* - config_name: lez_Cyrl data_files: data/lez_Cyrl/* - config_name: lfn_Cyrl data_files: data/lfn_Cyrl/* - config_name: lim_Latn data_files: data/lim_Latn/* - config_name: lin_Latn data_files: data/lin_Latn/* - config_name: lis_Lisu data_files: data/lis_Lisu/* - config_name: lit_Latn data_files: data/lit_Latn/* - config_name: lki_Arab data_files: data/lki_Arab/* - config_name: lld_Latn data_files: data/lld_Latn/* - config_name: lmk_Latn data_files: data/lmk_Latn/* - config_name: lnd_Latn data_files: data/lnd_Latn/* - config_name: lrc_Arab data_files: data/lrc_Arab/* - config_name: ltg_Latn data_files: data/ltg_Latn/* - config_name: ltz_Latn data_files: data/ltz_Latn/* - config_name: lud_Latn data_files: data/lud_Latn/* - config_name: lug_Latn data_files: data/lug_Latn/* - config_name: luo_Latn data_files: data/luo_Latn/* - config_name: lus_Latn data_files: data/lus_Latn/* - config_name: lvs_Latn data_files: data/lvs_Latn/* - config_name: lwg_Latn data_files: data/lwg_Latn/* - config_name: lzh_Hani data_files: data/lzh_Hani/* - config_name: mag_Deva data_files: data/mag_Deva/* - config_name: mah_Latn data_files: data/mah_Latn/* - config_name: mai_Deva data_files: data/mai_Deva/* - config_name: mak_Latn data_files: data/mak_Latn/* - config_name: mal_Latn data_files: data/mal_Latn/* - config_name: mal_Mlym data_files: data/mal_Mlym/* - config_name: mar_Deva data_files: data/mar_Deva/* - config_name: mar_Latn data_files: data/mar_Latn/* - config_name: mas_Latn data_files: data/mas_Latn/* - config_name: mbf_Latn data_files: data/mbf_Latn/* - config_name: mdf_Cyrl data_files: data/mdf_Cyrl/* - config_name: mer_Latn data_files: data/mer_Latn/* - config_name: mfe_Latn data_files: data/mfe_Latn/* - config_name: mfg_Latn data_files: data/mfg_Latn/* - config_name: mfy_Latn data_files: data/mfy_Latn/* - config_name: mhi_Latn data_files: data/mhi_Latn/* - config_name: mhr_Cyrl data_files: data/mhr_Cyrl/* - config_name: mhy_Latn data_files: data/mhy_Latn/* - config_name: min_Latn data_files: data/min_Latn/* - config_name: mip_Latn data_files: data/mip_Latn/* - config_name: mjw_Latn data_files: data/mjw_Latn/* - config_name: mkd_Cyrl data_files: data/mkd_Cyrl/* - config_name: mlt_Latn data_files: data/mlt_Latn/* - config_name: mni_Beng data_files: data/mni_Beng/* - config_name: mni_Latn data_files: data/mni_Latn/* - config_name: mnk_Latn data_files: data/mnk_Latn/* - config_name: mns_Cyrl data_files: data/mns_Cyrl/* - config_name: mnw_Mymr data_files: data/mnw_Mymr/* - config_name: moh_Latn data_files: data/moh_Latn/* - config_name: mph_Latn data_files: data/mph_Latn/* - config_name: mqy_Latn data_files: data/mqy_Latn/* - config_name: mri_Latn data_files: data/mri_Latn/* - config_name: mrj_Cyrl data_files: data/mrj_Cyrl/* - config_name: mrw_Latn data_files: data/mrw_Latn/* - config_name: mtg_Latn data_files: data/mtg_Latn/* - config_name: mui_Latn data_files: data/mui_Latn/* - config_name: mup_Deva data_files: data/mup_Deva/* - config_name: mus_Latn data_files: data/mus_Latn/* - config_name: mvp_Latn data_files: data/mvp_Latn/* - config_name: mwf_Latn data_files: data/mwf_Latn/* - config_name: mwl_Latn data_files: data/mwl_Latn/* - config_name: mww_Latn data_files: data/mww_Latn/* - config_name: mya_Mymr data_files: data/mya_Mymr/* - config_name: myv_Cyrl data_files: data/myv_Cyrl/* - config_name: myx_Latn data_files: data/myx_Latn/* - config_name: mzh_Latn data_files: data/mzh_Latn/* - config_name: nah_Latn data_files: data/nah_Latn/* - config_name: nan_Latn data_files: data/nan_Latn/* - config_name: nap_Latn data_files: data/nap_Latn/* - config_name: naq_Latn data_files: data/naq_Latn/* - config_name: nbu_Latn data_files: data/nbu_Latn/* - config_name: nde_Latn data_files: data/nde_Latn/* - config_name: ndo_Latn data_files: data/ndo_Latn/* - config_name: nds_Latn data_files: data/nds_Latn/* - config_name: new_Deva data_files: data/new_Deva/* - config_name: nio_Cyrl data_files: data/nio_Cyrl/* - config_name: njn_Latn data_files: data/njn_Latn/* - config_name: njo_Latn data_files: data/njo_Latn/* - config_name: nld_Latn data_files: data/nld_Latn/* - config_name: nmf_Latn data_files: data/nmf_Latn/* - config_name: nmz_Latn data_files: data/nmz_Latn/* - config_name: nno_Latn data_files: data/nno_Latn/* - config_name: nob_Latn data_files: data/nob_Latn/* - config_name: nog_Cyrl data_files: data/nog_Cyrl/* - config_name: non_Latn data_files: data/non_Latn/* - config_name: npi_Deva data_files: data/npi_Deva/* - config_name: npi_Latn data_files: data/npi_Latn/* - config_name: npo_Latn data_files: data/npo_Latn/* - config_name: nrf_Latn data_files: data/nrf_Latn/* - config_name: nri_Latn data_files: data/nri_Latn/* - config_name: nrm_Latn data_files: data/nrm_Latn/* - config_name: nse_Latn data_files: data/nse_Latn/* - config_name: nus_Latn data_files: data/nus_Latn/* - config_name: nya_Latn data_files: data/nya_Latn/* - config_name: nyn_Latn data_files: data/nyn_Latn/* - config_name: nzm_Latn data_files: data/nzm_Latn/* - config_name: obo_Latn data_files: data/obo_Latn/* - config_name: oci_Latn data_files: data/oci_Latn/* - config_name: ojb_Latn data_files: data/ojb_Latn/* - config_name: olo_Latn data_files: data/olo_Latn/* - config_name: orv_Cyrl data_files: data/orv_Cyrl/* - config_name: ory_Latn data_files: data/ory_Latn/* - config_name: ory_Orya data_files: data/ory_Orya/* - config_name: oss_Cyrl data_files: data/oss_Cyrl/* - config_name: ota_Arab data_files: data/ota_Arab/* - config_name: oto_Latn data_files: data/oto_Latn/* - config_name: otw_Latn data_files: data/otw_Latn/* - config_name: pam_Latn data_files: data/pam_Latn/* - config_name: pan_Guru data_files: data/pan_Guru/* - config_name: pan_Latn data_files: data/pan_Latn/* - config_name: pap_Latn data_files: data/pap_Latn/* - config_name: pbt_Arab data_files: data/pbt_Arab/* - config_name: pcd_Latn data_files: data/pcd_Latn/* - config_name: pck_Latn data_files: data/pck_Latn/* - config_name: pcm_Latn data_files: data/pcm_Latn/* - config_name: pfl_Latn data_files: data/pfl_Latn/* - config_name: plt_Latn data_files: data/plt_Latn/* - config_name: pmq_Latn data_files: data/pmq_Latn/* - config_name: pmx_Latn data_files: data/pmx_Latn/* - config_name: pnb_Arab data_files: data/pnb_Arab/* - config_name: pnt_Grek data_files: data/pnt_Grek/* - config_name: pol_Latn data_files: data/pol_Latn/* - config_name: por_Latn data_files: data/por_Latn/* - config_name: pov_Latn data_files: data/pov_Latn/* - config_name: ppk_Latn data_files: data/ppk_Latn/* - config_name: pps_Latn data_files: data/pps_Latn/* - config_name: prg_Latn data_files: data/prg_Latn/* - config_name: pui_Latn data_files: data/pui_Latn/* - config_name: pxm_Latn data_files: data/pxm_Latn/* - config_name: quc_Latn data_files: data/quc_Latn/* - config_name: qul_Latn data_files: data/qul_Latn/* - config_name: qup_Latn data_files: data/qup_Latn/* - config_name: qus_Latn data_files: data/qus_Latn/* - config_name: quz_Latn data_files: data/quz_Latn/* - config_name: raw_Latn data_files: data/raw_Latn/* - config_name: rcf_Latn data_files: data/rcf_Latn/* - config_name: rel_Latn data_files: data/rel_Latn/* - config_name: rhg_Latn data_files: data/rhg_Latn/* - config_name: ria_Latn data_files: data/ria_Latn/* - config_name: rjs_Deva data_files: data/rjs_Deva/* - config_name: rmc_Latn data_files: data/rmc_Latn/* - config_name: rml_Latn data_files: data/rml_Latn/* - config_name: rmn_Latn data_files: data/rmn_Latn/* - config_name: rmy_Cyrl data_files: data/rmy_Cyrl/* - config_name: rmy_Latn data_files: data/rmy_Latn/* - config_name: rnl_Latn data_files: data/rnl_Latn/* - config_name: roh_Latn data_files: data/roh_Latn/* - config_name: ron_Cyrl data_files: data/ron_Cyrl/* - config_name: ron_Latn data_files: data/ron_Latn/* - config_name: rtm_Latn data_files: data/rtm_Latn/* - config_name: rue_Cyrl data_files: data/rue_Cyrl/* - config_name: run_Latn data_files: data/run_Latn/* - config_name: rus_Cyrl data_files: data/rus_Cyrl/* - config_name: sah_Cyrl data_files: data/sah_Cyrl/* - config_name: san_Deva data_files: data/san_Deva/* - config_name: san_Latn data_files: data/san_Latn/* - config_name: sat_Latn data_files: data/sat_Latn/* - config_name: sck_Deva data_files: data/sck_Deva/* - config_name: scn_Latn data_files: data/scn_Latn/* - config_name: sda_Latn data_files: data/sda_Latn/* - config_name: sdc_Latn data_files: data/sdc_Latn/* - config_name: sdh_Arab data_files: data/sdh_Arab/* - config_name: ses_Latn data_files: data/ses_Latn/* - config_name: sgc_Latn data_files: data/sgc_Latn/* - config_name: sgh_Cyrl data_files: data/sgh_Cyrl/* - config_name: sid_Latn data_files: data/sid_Latn/* - config_name: sin_Sinh data_files: data/sin_Sinh/* - config_name: sju_Latn data_files: data/sju_Latn/* - config_name: skr_Arab data_files: data/skr_Arab/* - config_name: slk_Latn data_files: data/slk_Latn/* - config_name: slv_Latn data_files: data/slv_Latn/* - config_name: sma_Latn data_files: data/sma_Latn/* - config_name: sme_Latn data_files: data/sme_Latn/* - config_name: smj_Latn data_files: data/smj_Latn/* - config_name: smn_Latn data_files: data/smn_Latn/* - config_name: smo_Latn data_files: data/smo_Latn/* - config_name: sms_Latn data_files: data/sms_Latn/* - config_name: smt_Latn data_files: data/smt_Latn/* - config_name: sna_Latn data_files: data/sna_Latn/* - config_name: snd_Arab data_files: data/snd_Arab/* - config_name: snd_Deva data_files: data/snd_Deva/* - config_name: snd_Latn data_files: data/snd_Latn/* - config_name: som_Latn data_files: data/som_Latn/* - config_name: sot_Latn data_files: data/sot_Latn/* - config_name: spa_Latn data_files: data/spa_Latn/* - config_name: srd_Latn data_files: data/srd_Latn/* - config_name: srp_Cyrl data_files: data/srp_Cyrl/* - config_name: srp_Latn data_files: data/srp_Latn/* - config_name: ssw_Latn data_files: data/ssw_Latn/* - config_name: sun_Latn data_files: data/sun_Latn/* - config_name: swe_Latn data_files: data/swe_Latn/* - config_name: swg_Latn data_files: data/swg_Latn/* - config_name: swh_Latn data_files: data/swh_Latn/* - config_name: syc_Syrc data_files: data/syc_Syrc/* - config_name: syl_Latn data_files: data/syl_Latn/* - config_name: szl_Latn data_files: data/szl_Latn/* - config_name: tab_Cyrl data_files: data/tab_Cyrl/* - config_name: tam_Latn data_files: data/tam_Latn/* - config_name: tam_Taml data_files: data/tam_Taml/* - config_name: taq_Tfng data_files: data/taq_Tfng/* - config_name: tat_Cyrl data_files: data/tat_Cyrl/* - config_name: tat_Latn data_files: data/tat_Latn/* - config_name: tcy_Knda data_files: data/tcy_Knda/* - config_name: tcz_Latn data_files: data/tcz_Latn/* - config_name: tel_Latn data_files: data/tel_Latn/* - config_name: tel_Telu data_files: data/tel_Telu/* - config_name: tet_Latn data_files: data/tet_Latn/* - config_name: tgk_Cyrl data_files: data/tgk_Cyrl/* - config_name: tha_Thai data_files: data/tha_Thai/* - config_name: thl_Deva data_files: data/thl_Deva/* - config_name: tig_Ethi data_files: data/tig_Ethi/* - config_name: tir_Ethi data_files: data/tir_Ethi/* - config_name: tkl_Latn data_files: data/tkl_Latn/* - config_name: tkr_Cyrl data_files: data/tkr_Cyrl/* - config_name: tlh_Latn data_files: data/tlh_Latn/* - config_name: tly_Latn data_files: data/tly_Latn/* - config_name: tok_Latn data_files: data/tok_Latn/* - config_name: ton_Latn data_files: data/ton_Latn/* - config_name: tpi_Latn data_files: data/tpi_Latn/* - config_name: tpw_Latn data_files: data/tpw_Latn/* - config_name: trc_Latn data_files: data/trc_Latn/* - config_name: trp_Latn data_files: data/trp_Latn/* - config_name: trs_Latn data_files: data/trs_Latn/* - config_name: ttj_Latn data_files: data/ttj_Latn/* - config_name: tuk_Arab data_files: data/tuk_Arab/* - config_name: tuk_Cyrl data_files: data/tuk_Cyrl/* - config_name: tuk_Latn data_files: data/tuk_Latn/* - config_name: tur_Latn data_files: data/tur_Latn/* - config_name: tuv_Latn data_files: data/tuv_Latn/* - config_name: twx_Latn data_files: data/twx_Latn/* - config_name: tyv_Cyrl data_files: data/tyv_Cyrl/* - config_name: tzl_Latn data_files: data/tzl_Latn/* - config_name: tzm_Tfng data_files: data/tzm_Tfng/* - config_name: udm_Cyrl data_files: data/udm_Cyrl/* - config_name: uig_Arab data_files: data/uig_Arab/* - config_name: uig_Cyrl data_files: data/uig_Cyrl/* - config_name: uig_Latn data_files: data/uig_Latn/* - config_name: ukr_Cyrl data_files: data/ukr_Cyrl/* - config_name: urd_Arab data_files: data/urd_Arab/* - config_name: urd_Latn data_files: data/urd_Latn/* - config_name: uzn_Cyrl data_files: data/uzn_Cyrl/* - config_name: uzn_Latn data_files: data/uzn_Latn/* - config_name: uzs_Arab data_files: data/uzs_Arab/* - config_name: vap_Latn data_files: data/vap_Latn/* - config_name: vie_Latn data_files: data/vie_Latn/* - config_name: vot_Latn data_files: data/vot_Latn/* - config_name: vro_Latn data_files: data/vro_Latn/* - config_name: war_Latn data_files: data/war_Latn/* - config_name: way_Latn data_files: data/way_Latn/* - config_name: wba_Latn data_files: data/wba_Latn/* - config_name: wbm_Latn data_files: data/wbm_Latn/* - config_name: wes_Latn data_files: data/wes_Latn/* - config_name: whk_Latn data_files: data/whk_Latn/* - config_name: wlx_Latn data_files: data/wlx_Latn/* - config_name: wol_Latn data_files: data/wol_Latn/* - config_name: wsg_Telu data_files: data/wsg_Telu/* - config_name: wwa_Latn data_files: data/wwa_Latn/* - config_name: xal_Cyrl data_files: data/xal_Cyrl/* - config_name: xho_Latn data_files: data/xho_Latn/* - config_name: xmm_Latn data_files: data/xmm_Latn/* - config_name: xmv_Latn data_files: data/xmv_Latn/* - config_name: xog_Latn data_files: data/xog_Latn/* - config_name: yaz_Latn data_files: data/yaz_Latn/* - config_name: ydd_Hebr data_files: data/ydd_Hebr/* - config_name: yor_Latn data_files: data/yor_Latn/* - config_name: yrk_Cyrl data_files: data/yrk_Cyrl/* - config_name: yrl_Latn data_files: data/yrl_Latn/* - config_name: yua_Latn data_files: data/yua_Latn/* - config_name: yue_Hani data_files: data/yue_Hani/* - config_name: zea_Latn data_files: data/zea_Latn/* - config_name: zgh_Tfng data_files: data/zgh_Tfng/* - config_name: zom_Latn data_files: data/zom_Latn/* - config_name: zsm_Arab data_files: data/zsm_Arab/* - config_name: zsm_Latn data_files: data/zsm_Latn/* - config_name: zul_Latn data_files: data/zul_Latn/* --- # 💬 FineTranslations
FineTranslations
> The world's knowledge in 1+1T tokens of parallel text # Table of Contents - [💬 FineTranslations](#-finetranslations) * [What is it?](#what-is-it) * [What is it for?](#what-is-it-for) * [Languages and available subsets](#languages-and-available-subsets) * [How to download and use 💬 FineTranslations](#how-to-download-and-use-finetranslations) + [Using 🏭 `datatrove`](#using-datatrove) + [Using `huggingface_hub`](#using-huggingface_hub) + [Using `datasets`](#using-datasets) * [Dataset processing steps](#dataset-processing-steps) + [1. Sourcing the data](#1-sourcing-the-data) + [2. Running translation at scale](#2-running-translation-at-scale) + [3. Post-processing](#3-post-processing) + [4. Edu-filtering](#4-edu-filtering) - [Dataset card for 💬 FineTranslations](#dataset-card-for-finetranslations) * [Dataset Description](#dataset-description) + [Dataset Summary](#dataset-summary) * [Dataset Structure](#dataset-structure) + [Data Instances](#data-instances) + [Data Fields](#data-fields) + [Data Splits](#data-splits) * [Dataset Creation](#dataset-creation) + [Curation Rationale](#curation-rationale) + [Source Data](#source-data) + [Data processing steps](#data-processing-steps-1) + [Annotations](#annotations) + [Personal and Sensitive Information and opt-out](#personal-and-sensitive-information-and-opt-out) * [Considerations for Using the Data](#considerations-for-using-the-data) + [Social Impact of Dataset](#social-impact-of-dataset) + [Discussion of Biases](#discussion-of-biases) + [Other Known Limitations](#other-known-limitations) * [Additional Information](#additional-information) + [Licensing Information](#licensing-information) * [Citation Information](#citation-information) ## What is it? This dataset contains over 1 trillion tokens of parallel text in English and 500+ languages. It was obtained by translating data from **🥂 [FineWeb2](https://huggingface.co/datasets/HuggingFaceFW/fineweb-2)** into English using [Gemma3 27B](https://huggingface.co/google/gemma-3-27b-it). We relied on [datatrove's inference runner](https://github.com/huggingface/datatrove?tab=readme-ov-file#synthetic-data-generation) to deploy a **synthetic data pipeline at scale**. Its **checkpointing** and **VLLM lifecycle management** features allowed us to use leftover compute from the HF cluster without fear of preemption. The **async implementation** ensures strong GPU utilization at all times. The **💬 FineTranslations** dataset is [fully reproducible](https://github.com/huggingface/finetranslations) and available under the permissive **ODC-By 1.0 license**. This is the _base_ version. For the Edu version, see [here](https://huggingface.co/datasets/HuggingFaceFW/finetranslations). ## What is it for? The main motivation behind the creation of this dataset was improving **translation capabilities**. While models are generally strong at translating from other languages into English (X->English), the opposite is often not true, particularly for lower resource languages. Our approach was to take data that was originally in non-English languages (from **🥂 [FineWeb2](https://huggingface.co/datasets/HuggingFaceFW/fineweb-2)**, our large multilingual pre-training dataset), chunk it, and translate it. This dataset can then be used to improve English->X translations by finetuning existing models (we leave this for future work). Additionally, the resulting English data contains relevant cultural information for different countries and languages, and our experiments show that the 1T tokens we obtained perform on a similar level as our **🍷 [FineWeb](https://huggingface.co/datasets/HuggingFaceFW/fineweb)** dataset. This data can therefore also be used for English only model training (potentially as an extension of FineWeb). For the English only performance of the dataset, see the comparison below. The ablation setup is the same as in [FinePDFs](https://huggingface.co/datasets/HuggingFaceFW/finepdfs).
finetranslations-comparisons
## Languages and available subsets Each language is identified by its [ISO 639-3 code](https://iso639-3.sil.org/code_tables/639/data), and the data is grouped by language-script pairs, since some languages have content in multiple scripts. The full list of subsets is available [here](https://huggingface.co/datasets/HuggingFaceFW/finetranslations/blob/main/subsets.csv). To access data from all the languages, use the `all` subset. ## How to download and use 💬 FineTranslations See the tables above for the `subset` of the language you want to download. We currently do not provide smaller `sample` versions, but by setting `limit` or using `streaming=True` you can easily fetch a sample of the data. ### Using 🏭 [`datatrove`](https://github.com/huggingface/datatrove/) ```python from datatrove.pipeline.readers import ParquetReader # limit determines how many documents will be streamed (remove for all) # this will fetch the Portuguese filtered data data_reader = ParquetReader("hf://datasets/HuggingFaceFW/finetranslations/data/por_Latn", limit=1000) for document in data_reader(): # do something with document print(document) ############################### # OR for a processing pipeline: ############################### from datatrove.executor import LocalPipelineExecutor from datatrove.pipeline.readers import ParquetReader from datatrove.pipeline.filters import LambdaFilter from datatrove.pipeline.writers import JsonlWriter pipeline_exec = LocalPipelineExecutor( pipeline=[ ParquetReader("hf://datasets/HuggingFaceFW/finetranslations/data/por_Latn", limit=1000), LambdaFilter(lambda doc: "hugging" in doc.text), JsonlWriter("some-output-path") ], tasks=10 ) pipeline_exec.run() ``` ### Using `huggingface_hub` ```python from huggingface_hub import snapshot_download folder = snapshot_download( "HuggingFaceFW/finetranslations", repo_type="dataset", local_dir="./finetranslations/", # download the Czech filtered data allow_patterns=["data/ces_Latn/*"]) ``` ### Using `datasets` ```python from datasets import load_dataset # get data from all languages fw = load_dataset("HuggingFaceFW/finetranslations", name="all", split="train", streaming=True) ``` ## Dataset processing steps We used the 🏭 `datatrove` library to process the data. You can find the entire **working code** that created the dataset [here](https://github.com/huggingface/finetranslations). ### 1. Sourcing the data The starting point for this dataset was our previously released **🥂 [FineWeb2](https://huggingface.co/datasets/HuggingFaceFW/fineweb-2)** dataset, a large scale pre-training dataset covering over a thousand languages. As many of these language subsets consisted in large part of religious content (mostly bibles) or Wikipedia pages, we only included the languages whose subset had a `bible_wiki_ratio` (ratio of documents with this type of content) under `0.5` (around 500 languages). We processed up to 50B tokens per language. For languages that originally had more than 50B tokens, we employed quality classifiers from [FineWeb2-HQ](https://huggingface.co/datasets/epfml/FineWeb2-HQ) and kept the top 50B tokens. When a classifier wasn't available, we randomly sampled 50B tokens worth of documents. ### 2. Running translation at scale We compared a variety of models from the Qwen, Llama, Gemma, Mistral and Aya models on translation benchmarks on a large number of languages. Qwen and Gemma models showed the strongest performance across the board, but Qwen models would sometimes output Chinese even when translating from European languages into English. As such, and for simplicity, we employed [Gemma3 27B](https://huggingface.co/google/gemma-3-27b-it) to translate all languages. The main issues we observed from our early experiments were: - a large amount of toxic/adult/gambling related content, specially originating from our lower resource languages; - lack of adherence to the original formatting. In particular, new lines would often be removed or added arbitrarily; - repetition loops (that would run until the model context was full), particularly in very large documents We relied on the following measures to address them: - have the model initially classify the type of content before translating, flagging adult/spam like content early (faster processing too) - strict formatting rules in the prompt - chunk documents into at most 512 token chunks. We then rely on a sliding window approach to translate the next chunk while keeping the previous one (already translated) in the prompt for context The prompt used is as follows:
Click to view the full translation SYSTEM prompt
**EARLY EXIT (runs BEFORE anything else)**
1) First, classify <ORIGINAL>.
2) If it contains ANY of:
   - Pornographic/explicit sexual content (incl. escorting)
   - Online gambling/casino/betting
   - Trading/crypto/forex promotional content
   - Lists of unrelated keywords or phrases lacking complete sentences and grammatical connectors (SEO spam)
THEN immediately output exactly:
<TRANSLATION>CONTENT FLAG</TRANSLATION>
and STOP — ignore all other instructions.
3) If not, proceed with translation rules.

You are a professional translator. Follow **all** instructions exactly.

**Crucial Formatting Rules (READ CAREFULLY — HARD REQUIREMENTS):**
1. **Preserve formatting EXACTLY.**
* Do **not** add, remove, or modify any line breaks.
* You must output the **exact same number of lines** as the input.
* Each line in your output must correspond exactly to one line in the input.
* You must **never** insert additional blank lines that do not exist in the original text.
* Do **not** insert blank lines for readability.
2. **Translate every token.**
Do not skip, summarize, or ignore any word, punctuation mark, or spacing.
3. **No literal translation.**
Make the English natural and fluent, but **do not** change formatting.
4. **No hallucinations.**
Do not add explanations, commentary, or any content that isn't in the original.
5. **Output format**
Enclose the translated text **only** in:
<TRANSLATION>
</TRANSLATION>
Nothing before or after.
6. **If you cannot follow these formatting rules exactly, output:**
`ERROR: formatting rule violated`

**Additional Strict Requirements:**
- Do **NOT** insert extra whitespace.
- Do **NOT** auto-format paragraphs.
- Do **NOT** add blank lines.
- Do NOT reinterpret or restructure the text. Do NOT treat long lines as paragraphs. You must preserve every line exactly as written, even if the line is extremely long, contains many sentences, or appears to represent multiple paragraphs.
- Do NOT split any lines into multiple lines. Even if a line contains many sentences, you must keep it as a single line exactly as in the original.
- Treat every visible line break as unchangeable.
- When in doubt, copy the structure line by line.

Remember the early exit rule before you consider translating.
Click to view the full translation USER prompt
**{display_language} ({subset_language}) Text to Translate (preserve all line breaks EXACTLY):**
<ORIGINAL>{combined_chunk}</ORIGINAL>

Now translate to English (eng_Latn).
The pipeline ran on the Hugging Face cluster over a period of 3 months, making use of spare compute cycles. # 3. Post-processing We removed content that the model had flagged, removed the <TRANSLATION></TRANSLATION> markdown tags and ensured line breaks on the chunk boundaries were consistent. # 4. Edu-filtering We attempted to use the quality classifier from [FineWeb-Edu](https://huggingface.co/datasets/HuggingFaceFW/fineweb-edu) to boost English performance. However, this filtering did not lead to a performance improvement, potentially due to distribution differences between Gemma generated text and natural English from web pages. Therefore, we trained a new classifier that provided a modest performance boost when filtering for the top 10% of content. This content is available in the [finetranslations-edu dataset](https://huggingface.co/datasets/HuggingFaceFW/finetranslations-edu). # Dataset card for 💬 FineTranslations ## Dataset Description - **Homepage and Repository:** [https://huggingface.co/datasets/HuggingFaceFW/finetranslations](https://huggingface.co/datasets/HuggingFaceFW/finetranslations) - **Point of Contact:** https://huggingface.co/datasets/HuggingFaceFW/finetranslations/discussion - **License:** Open Data Commons Attribution License (ODC-By) v1.0 ### Dataset Summary This dataset contains over 1 trillion tokens of parallel text in English and 500+ languages. It was obtained by translating data from **🥂 [FineWeb2](https://huggingface.co/datasets/HuggingFaceFW/fineweb-2)** into English using [Gemma3 27B](https://huggingface.co/google/gemma-3-27b-it). ## Dataset Structure ### Data Instances The following is an example sample from the dataset. It is part of the French (`fra_Latn`) data, originally from the `CC-MAIN-2014-52` CommonCrawl snapshot. ```json { "translated_text": "A and I completed a small DIY project that is easy but yields surprising and charming results.\nWe used wooden letters that A patiently painted with wood paint.\nSubsequently, we added buttons with hot glue to embellish everything.\nEach person now has their own colorful and fun letter!", "translated_chunks": [ "A and I completed a small DIY project that is easy but yields surprising and charming results.\nWe used wooden letters that A patiently painted with wood paint.\nSubsequently, we added buttons with hot glue to embellish everything.\nEach person now has their own colorful and fun letter!" ], "og_chunks": [ "A et moi avons réalisé un petit projet brico facile mais qui donne des résultats surprenants et charmants.\nNous avons utilisé des lettres de bois que A a patiemment peint avec de la peinture à bois.\nPar la suite, nous y avons ajouté, à la colle chaude, des boutons pour garnir le tout.\nChacun a maintenant sa propre lettre colorée et amusante!" ], "og_full_text": "A et moi avons réalisé un petit projet brico facile mais qui donne des résultats surprenants et charmants.\nNous avons utilisé des lettres de bois que A a patiemment peint avec de la peinture à bois.\nPar la suite, nous y avons ajouté, à la colle chaude, des boutons pour garnir le tout.\nChacun a maintenant sa propre lettre colorée et amusante!", "og_language": "fra_Latn", "og_language_score": 0.9992175698280334, "og_token_count": 83, "og_quality_score": 0.03385915607213974, "early_stop": false, "id": "", "url": "http://mcommemaman.blogspot.com/2008/12/bricolage-personnalis.html", "warc_path": "s3://commoncrawl/crawl-data/CC-MAIN-2014-52/segments/1418802778085.5/warc/CC-MAIN-20141217075258-00136-ip-10-231-17-201.ec2.internal.warc.gz", "minhash_cluster_size": 42, "translated_token_count": 58, "edu_score_raw": 0.640625, "edu_score": 1 } ``` ### Data Fields - `translated_text` (string): the English translated text content (all chunks concatenated) - `translated_chunks` (list of strings): the English translation split into chunks - `og_chunks` (list of strings): the original text in the source language split into chunks (there is a 1-1 match between `translated_chunks` and `og_chunks`) - `og_full_text` (string): the original full text in the source language - `og_language` (string): language-script code for the original text (e.g., `fra_Latn`) - `og_language_score` (float): language prediction score for the original text as reported by the [GlotLID classifier](https://github.com/huggingface/datatrove/blob/main/src/datatrove/pipeline/filters/language_filter.py#L52) - `og_token_count` (int): token count of the original text - `og_quality_score` (float): quality score of the original text from the EPFL classifiers (if available, -1 otherwise) - `early_stop` (bool): whether translation stopped early due to formatting issues. In this case, the last chunks from the original text were dropped. - `id` (string): original unique identifier for this sample from CommonCrawl - `url` (string): url to the original page where the text was present - `warc_path` (string): s3 path for the individual CommonCrawl warc file containing this sample - `minhash_cluster_size` (int): number of samples in the FineWeb2 minhash cluster of this sample. See the deduplication section of FineWeb2 for more info. - `translated_token_count` (int): token count of the English (translated) text - `edu_score_raw` (float): raw educational score from the educational classifier - `edu_score` (int): binned educational score (0-4 scale) Notes: - `og_quality_score` is from a classifier applied on the original language, while `edu_score` was computed on the translated English text - in case of `early_stop`, some chunks might have been dropped. In this case, `og_full_text` might not match `translated_text`. Rely on the chunk variables if needed. ### Data Splits See "**Languages and available subsets**" above. ## Dataset Creation ### Curation Rationale The main motivation behind the creation of this dataset was improving **translation capabilities**. While models are generally strong at translating from other languages into English (X->English), the opposite is often not true, particularly for lower resource languages. Our approach was to take data that was originally in non-English languages (from **🥂 [FineWeb2](https://huggingface.co/datasets/HuggingFaceFW/fineweb-2)**, our large multilingual pre-training dataset), chunk it, and translate it. This dataset can then be used to improve English->X translations by finetuning existing models. Additionally, the resulting English data contains relevant cultural information for different countries and languages, and our experiments show that the 1T tokens we obtained perform on a similar level as our **🍷 [FineWeb](https://huggingface.co/datasets/HuggingFaceFW/fineweb)** dataset. This data can therefore also be used for English only model training (potentially as an extension of FineWeb). ### Source Data The source data for **💬 FineTranslations** is **🥂 [FineWeb2](https://huggingface.co/datasets/HuggingFaceFW/fineweb-2)**, a large scale pre-training dataset covering over a thousand languages sourced from CommonCrawl webpages crawled over the 2013-2024 time period. As many of these language subsets consisted in large part of religious content (mostly bibles) or Wikipedia pages, we only included the languages whose subset had a `bible_wiki_ratio` (ratio of documents with this type of content) under `0.5` (around 500 languages). We processed up to 50B tokens per language. For languages that originally had more than 50B tokens, we employed quality classifiers from [FineWeb2-HQ](https://huggingface.co/datasets/epfml/FineWeb2-HQ) and kept the top 50B tokens. When a classifier wasn't available, we randomly sampled 50B tokens worth of documents. ### Data processing steps See "**Dataset processing steps**" above. ### Annotations We augment the original samples with translation-related annotations including `translated_text`, `translated_chunks`, `og_chunks`, `og_full_text`, `og_language`, `og_language_score`, `og_token_count`, `og_quality_score`, `translated_token_count`, `early_stop`, `edu_score_raw`, and `edu_score`. The original language annotations (`og_language`, `og_language_score`) are inherited from FineWeb2 and were automatically generated by the [language filter](https://github.com/huggingface/datatrove/blob/main/src/datatrove/pipeline/filters/language_filter.py). The `minhash_cluster_size` is also inherited from FineWeb2 and was computed during the deduplication process. Translation-specific annotations track the translation process, quality, and educational scores. ### Personal and Sensitive Information and opt-out The source data (FineWeb2) anonymizes email addresses and public IP addresses. For emails, a regex pattern is applied and any occurrence of an email address is replaced with either `email@example.com` or `firstname.lastname@example.org`. For IP addresses, a regex pattern is employed and then further filtered to only anonymize IP addresses [allocated for public networks](https://www.iana.org/assignments/iana-ipv4-special-registry/iana-ipv4-special-registry.xhtml). Matched IP addresses are then replaced with one of the following randomly generated IP addresses, which at the time of dataset creation were not responding to ping requests: `22.214.171.124`, `126.96.36.199`, `188.8.131.52`, `184.108.40.206`, `220.127.116.11`, and `18.104.22.168`. The source dataset decided against applying regex patterns for phone numbers due to the high false positive rate. Despite these efforts, given that **💬 FineTranslations** is sourced from web content at large, it is very likely that some personally identifiable information (PII) will be present. If you find your own PII in **💬 FineTranslations** and would like it removed, please fill out our [PII removal/opt out form](https://forms.gle/VyNT3ZAUPZjPuWp39). CommonCrawl respects robots.txt at crawl time, but if you are a webmaster and find your website in **💬 FineTranslations** and would like to have it removed, you may also use the [PII removal/opt out form](https://forms.gle/VyNT3ZAUPZjPuWp39). ## Considerations for Using the Data ### Social Impact of Dataset With the release of this dataset we aim to improve translation capabilities, particularly for lower resource languages where English->X translation is often weak. By providing over 1 trillion tokens of parallel text data across 500+ languages, we enable researchers and practitioners to: - **Improve translation models**: Finetune existing models on this parallel data to improve English->X translation capabilities - **Train multilingual models**: Use the parallel data for training or improving multilingual models - **Enhance English models**: Leverage the translated English content, which contains cultural information from diverse languages and performs similarly to FineWeb for English-only training The dataset is fully reproducible with all code available, making the translation pipeline transparent and allowing the community to build upon our work. ### Discussion of Biases The dataset inherits biases from both the source data (FineWeb2) and the translation process: **Source data biases**: As FineWeb2 was sourced from the web, any harmful biases typically present in web content may be reproduced in this dataset. Efforts were made in the source dataset to minimize NSFW and toxic content through URL-level filtering, but some toxic or harmful content may still be present. **Translation model biases**: The translations were generated using Gemma3 27B, which may introduce its own biases: - The model may translate certain concepts or cultural references in ways that don't fully capture the original meaning - Translation quality may vary across languages, with lower resource languages potentially receiving lower quality translations - The model's training data biases may be reflected in the translations **Content filtering**: We employed early exit mechanisms to flag adult/spam content before translation, but some content that passed these filters may still be considered inappropriate. The formatting preservation requirements may also have led to some translations that don't read as naturally as human translations. ### Other Known Limitations **Translation quality**: While we compared multiple models and selected Gemma3 27B for its strong performance, translation quality is not uniform across all languages. Lower resource languages may have lower translation quality, and some translations may contain errors or awkward phrasing. **Formatting preservation**: We prompt the translation model to strictly maintain the original formatting, such as line breaks and document structure. However, in practice, the model does not always fully respect these instructions—so while our approach aims for high formatting fidelity, there may be cases where formatting inconsistencies remain or the structure is not perfectly preserved. **Model limitations**: The translation model has context limitations (we chunked documents into 512 token chunks), which means very long documents are translated in pieces. While we use a sliding window approach to maintain context, some coherence may be lost across chunk boundaries. **Language coverage**: We only included languages from FineWeb2 with a `bible_wiki_ratio` under 0.5, which excluded some languages that were predominantly religious or Wikipedia content. This means the dataset may not be representative of all possible language content. **Educational filtering**: The educational classifier was trained specifically for this dataset, but its performance may vary across different types of content. The top 10% educational content is available in a separate dataset (finetranslations-edu). We encourage users to review the translation quality for their languages of interest and consider additional filtering or post-processing if needed. ## Additional Information ### Licensing Information The dataset is released under the **Open Data Commons Attribution License (ODC-By) v1.0** [license](https://opendatacommons.org/licenses/by/1-0/). The use of this dataset is also subject to [CommonCrawl's Terms of Use](https://commoncrawl.org/terms-of-use). ## Citation Information ``` @misc{penedo2026finetranslations, title={FineTranslations}, author={Guilherme Penedo and Hynek Kydl{\'\i}{\v{c}}ek and Amir Hossein Kargaran and Leandro von Werra}, year={2026}, publisher = {Hugging Face}, journal = {Hugging Face repository}, howpublished = {\url{https://huggingface.co/datasets/HuggingFaceFW/finetranslations}} } ```