PyTorch
ELM
hubert
AfriHuBERT / README.md
ajesujoba's picture
Update README.md
4cc3fcc verified
metadata
datasets:
  - mozilla-foundation/common_voice_17_0
  - TalTechNLP/VoxLingua107
  - espnet/mms_ulab_v2
  - espnet/jesus_dramas
language:
  - aab
  - aal
  - aar
  - abi
  - abn
  - abo
  - abr
  - abu
  - acd
  - ach
  - acz
  - ada
  - add
  - ade
  - adj
  - adq
  - aeb
  - aec
  - ael
  - afe
  - afo
  - afr
  - afu
  - agc
  - agh
  - ags
  - aha
  - ahg
  - ahl
  - ahp
  - ahs
  - aik
  - aiw
  - ajg
  - aka
  - akp
  - aks
  - akw
  - ala
  - ald
  - alf
  - alw
  - alz
  - amb
  - amf
  - amh
  - amo
  - anc
  - ank
  - ann
  - anu
  - anv
  - anw
  - any
  - apd
  - aqg
  - arb
  - arc
  - arq
  - arv
  - ary
  - arz
  - asa
  - ass
  - ati
  - ato
  - atu
  - aug
  - avi
  - avn
  - avu
  - awn
  - ayb
  - ayg
  - ayu
  - bab
  - bag
  - bam
  - bas
  - bau
  - baw
  - bax
  - bba
  - bbk
  - bbo
  - bbq
  - bbt
  - bbu
  - bbw
  - bby
  - bcg
  - bci
  - bcn
  - bcp
  - bcq
  - bcs
  - bcv
  - bcy
  - bda
  - bde
  - bdh
  - bdi
  - bdm
  - bdu
  - bec
  - beh
  - bej
  - bem
  - beq
  - bet
  - bev
  - bex
  - bez
  - bfa
  - bfd
  - bfj
  - bfm
  - bfo
  - bga
  - bgf
  - bgj
  - bhr
  - bhs
  - bhy
  - bib
  - bil
  - bim
  - bin
  - bip
  - biv
  - biz
  - bjg
  - bji
  - bjt
  - bka
  - bkc
  - bkg
  - bkm
  - bkv
  - bkw
  - bky
  - ble
  - blh
  - bli
  - blm
  - blo
  - bly
  - bma
  - bmb
  - bmd
  - bmf
  - bmi
  - bmm
  - bmq
  - bmv
  - bni
  - bnx
  - bob
  - bof
  - boh
  - bol
  - bom
  - boo
  - bot
  - bov
  - box
  - bqa
  - bqc
  - bqg
  - bqj
  - bqo
  - bqt
  - bqv
  - bqw
  - bqx
  - brf
  - bri
  - brt
  - bsc
  - bse
  - bsf
  - bsp
  - bsq
  - bss
  - bst
  - btg
  - btt
  - btu
  - bub
  - bud
  - bui
  - bum
  - bun
  - bus
  - buu
  - bux
  - buz
  - bvh
  - bvi
  - bvm
  - bvw
  - bwo
  - bwq
  - bwr
  - bwu
  - bww
  - bxb
  - bxg
  - bxk
  - bxq
  - bxs
  - byc
  - byj
  - byn
  - byp
  - bys
  - byv
  - bza
  - bze
  - bzv
  - bzw
  - bzy
  - bzz
  - cbj
  - cce
  - ccg
  - cch
  - ccl
  - cdr
  - cfa
  - cfd
  - cfg
  - chw
  - cib
  - cjk
  - ckl
  - cko
  - ckx
  - cky
  - cla
  - cli
  - cll
  - cme
  - cnq
  - coh
  - cou
  - cra
  - cry
  - csk
  - cuv
  - dag
  - dai
  - das
  - dav
  - dbb
  - dbd
  - dbi
  - dbm
  - dbv
  - dde
  - ddn
  - dee
  - deg
  - dez
  - dga
  - dgd
  - dgh
  - dgi
  - dhm
  - dib
  - did
  - dig
  - dik
  - dil
  - dim
  - dio
  - dip
  - dir
  - diu
  - diw
  - dje
  - djm
  - dks
  - dkx
  - dmo
  - dnj
  - dnn
  - doo
  - dop
  - dos
  - dot
  - dow
  - dox
  - doy
  - doz
  - dri
  - drs
  - dsh
  - dsq
  - dtm
  - dts
  - dua
  - dug
  - dur
  - dwr
  - dya
  - dyi
  - dyo
  - dyu
  - dza
  - dzd
  - dzg
  - ebo
  - ebr
  - ebu
  - ega
  - ego
  - eja
  - eka
  - ekr
  - elm
  - emk
  - emn
  - enb
  - eng
  - enn
  - eot
  - erh
  - eto
  - ets
  - etu
  - ewe
  - ewo
  - eyo
  - eza
  - fak
  - fal
  - fan
  - fap
  - fat
  - ffm
  - fie
  - fir
  - fli
  - fll
  - flr
  - fod
  - fon
  - fra
  - fub
  - fuc
  - fue
  - fuf
  - fuh
  - fuq
  - fuv
  - fvr
  - fwe
  - gaa
  - gab
  - gax
  - gaz
  - gbg
  - gbh
  - gbo
  - gbp
  - gbr
  - gby
  - gde
  - gdf
  - gdl
  - gdu
  - gea
  - gec
  - ged
  - geg
  - gej
  - gek
  - gel
  - ggb
  - ggu
  - ghl
  - gid
  - gis
  - giz
  - gjn
  - gkn
  - gkp
  - glj
  - glw
  - gmv
  - gna
  - gnd
  - gng
  - gnk
  - goa
  - gof
  - gog
  - gol
  - gou
  - gow
  - gpa
  - gqa
  - grd
  - grh
  - grj
  - gru
  - grv
  - gry
  - gud
  - guk
  - gur
  - guw
  - gux
  - guz
  - gwd
  - gwn
  - gwr
  - gxx
  - gya
  - gyz
  - hae
  - hag
  - haq
  - hau
  - hav
  - hbb
  - hbn
  - hdy
  - hed
  - her
  - hia
  - hig
  - hio
  - hmb
  - hoe
  - hoo
  - hts
  - huc
  - hwo
  - ibb
  - ibo
  - ica
  - ich
  - ida
  - idu
  - ife
  - igb
  - ige
  - igl
  - ijc
  - ijj
  - ijn
  - ijs
  - ikk
  - iko
  - ikw
  - ikx
  - ior
  - iqw
  - iri
  - irk
  - ish
  - isi
  - isn
  - isu
  - its
  - iyo
  - iyx
  - izr
  - izz
  - jab
  - jad
  - jbm
  - jbu
  - jen
  - jer
  - jgk
  - jib
  - jid
  - jit
  - jku
  - jle
  - jmb
  - jmc
  - jmr
  - jms
  - jni
  - jnj
  - jow
  - jub
  - juk
  - jum
  - juo
  - jwi
  - kab
  - kad
  - kai
  - kaj
  - kam
  - kao
  - kbj
  - kbl
  - kbo
  - kbp
  - kbr
  - kby
  - kbz
  - kcc
  - kce
  - kcg
  - kch
  - kci
  - kck
  - kcq
  - kcr
  - kcv
  - kcx
  - kde
  - kdh
  - kdj
  - kdl
  - kdm
  - kdp
  - kdu
  - kdx
  - kdz
  - kef
  - kel
  - ken
  - keo
  - ker
  - keu
  - kez
  - kfo
  - kfz
  - kga
  - kgo
  - khj
  - khq
  - khy
  - kia
  - kib
  - kid
  - kie
  - kik
  - kil
  - kin
  - kkj
  - kks
  - kku
  - klk
  - klo
  - klu
  - kma
  - kmb
  - kmi
  - kmp
  - kmq
  - kmw
  - kmy
  - kna
  - knc
  - knf
  - kng
  - kni
  - knk
  - kno
  - knp
  - knu
  - kny
  - knz
  - koe
  - kot
  - kow
  - kpa
  - kph
  - kpk
  - kpl
  - kpo
  - kqm
  - kqn
  - kqo
  - kqp
  - kqs
  - kqy
  - kri
  - krn
  - krw
  - krx
  - ksb
  - ksf
  - ksm
  - ksp
  - kss
  - ktb
  - ktc
  - ktf
  - ktu
  - kty
  - ktz
  - kua
  - kub
  - kuj
  - kul
  - kun
  - kus
  - kvf
  - kvi
  - kvj
  - kvm
  - kwb
  - kwc
  - kwg
  - kwl
  - kwn
  - kwu
  - kwv
  - kxb
  - kxc
  - kxh
  - kxj
  - kxx
  - kye
  - kyf
  - kyq
  - kza
  - kzc
  - kzr
  - lag
  - lai
  - laj
  - lam
  - lan
  - lar
  - las
  - lch
  - ldb
  - ldg
  - ldi
  - ldj
  - ldl
  - ldm
  - ldp
  - ldq
  - lea
  - led
  - lef
  - lel
  - lem
  - les
  - lgg
  - lgm
  - lgq
  - lia
  - lig
  - lin
  - lip
  - liq
  - lir
  - liu
  - liz
  - lkr
  - lky
  - lla
  - llc
  - lln
  - lma
  - lmd
  - lme
  - lmp
  - lmx
  - lns
  - lnu
  - lob
  - log
  - loh
  - lok
  - lol
  - lom
  - lop
  - loq
  - lor
  - lot
  - loz
  - lri
  - lrm
  - lro
  - lse
  - lsm
  - lua
  - lub
  - lue
  - lug
  - luj
  - lul
  - lun
  - luo
  - lwg
  - lwo
  - lyn
  - maf
  - mas
  - maw
  - mbu
  - mbv
  - mck
  - mcn
  - mcp
  - mcs
  - mct
  - mcu
  - mcw
  - mda
  - mdd
  - mde
  - mdj
  - mdk
  - mdm
  - mdt
  - mdu
  - mdy
  - mea
  - men
  - mer
  - mev
  - mey
  - mfc
  - mfd
  - mfe
  - mfg
  - mfh
  - mfi
  - mfk
  - mfl
  - mfm
  - mfn
  - mfo
  - mfq
  - mfv
  - mfz
  - mgb
  - mgc
  - mgd
  - mgg
  - mgh
  - mgo
  - mgw
  - mhi
  - mhk
  - mho
  - mhw
  - mij
  - mjs
  - mkf
  - mkk
  - mkl
  - mku
  - mkw
  - mlg
  - mlk
  - mlq
  - mls
  - mlw
  - mma
  - mmz
  - mnf
  - mnk
  - moi
  - mor
  - mos
  - moz
  - mpe
  - mpg
  - mql
  - mqu
  - mrt
  - msc
  - mse
  - msh
  - msw
  - mtb
  - mtk
  - mua
  - mug
  - muh
  - muo
  - mur
  - muy
  - muz
  - mvz
  - mwm
  - mwn
  - mxh
  - mxl
  - mxu
  - mxx
  - myb
  - mye
  - myk
  - mym
  - myx
  - mzb
  - mzj
  - mzk
  - mzm
  - mzw
  - naj
  - naq
  - nar
  - nat
  - naw
  - nba
  - nbb
  - nbh
  - nbl
  - nbm
  - nbp
  - nbr
  - nbv
  - ncr
  - ncu
  - ndb
  - ndc
  - ndd
  - nde
  - ndi
  - ndm
  - ndo
  - ndp
  - ndr
  - ndu
  - ndv
  - ndy
  - neb
  - nfd
  - nfr
  - nfu
  - nga
  - ngc
  - ngi
  - ngl
  - ngn
  - ngs
  - ngw
  - nhb
  - nhr
  - nhu
  - nih
  - nim
  - nin
  - niq
  - nix
  - niy
  - nja
  - njj
  - njx
  - nka
  - nko
  - nku
  - nkw
  - nlj
  - nlo
  - nlu
  - nmc
  - nmi
  - nmn
  - nmz
  - nnb
  - nnc
  - nnh
  - nnj
  - nnu
  - nnw
  - nnz
  - nqg
  - nso
  - nti
  - ntm
  - nto
  - ntr
  - nuj
  - nup
  - nus
  - nwb
  - nwm
  - nxd
  - nya
  - nyb
  - nyd
  - nyf
  - nyg
  - nyi
  - nyk
  - nym
  - nyn
  - nyo
  - nyu
  - nyy
  - nza
  - nzb
  - nzi
  - nzr
  - nzy
  - odu
  - ogb
  - ogc
  - ogo
  - oki
  - oku
  - old
  - olu
  - opa
  - orc
  - org
  - orx
  - oso
  - ost
  - otr
  - oub
  - ozm
  - pai
  - pbi
  - pbl
  - pbn
  - pbo
  - pbp
  - pcm
  - pcn
  - pcw
  - pem
  - pfe
  - pga
  - pil
  - pip
  - piy
  - pkb
  - pko
  - plt
  - pmm
  - png
  - pnq
  - pny
  - pof
  - por
  - pov
  - pqa
  - pug
  - puu
  - pwb
  - pym
  - rag
  - rcf
  - rel
  - res
  - rif
  - rim
  - rin
  - rkm
  - rnd
  - rsw
  - ruf
  - rui
  - ruk
  - run
  - ruy
  - ruz
  - rwk
  - sad
  - saf
  - sag
  - saq
  - sav
  - say
  - sba
  - sbd
  - sbk
  - sbs
  - sbz
  - scv
  - sde
  - sef
  - seg
  - seh
  - sen
  - sep
  - ses
  - sfw
  - sgc
  - sgi
  - sgw
  - shc
  - she
  - shg
  - shi
  - shj
  - shk
  - shr
  - shu
  - shw
  - shy
  - sid
  - sie
  - sif
  - sig
  - sil
  - skg
  - skq
  - skt
  - sld
  - slx
  - sna
  - snf
  - sng
  - snk
  - snm
  - snq
  - snw
  - soe
  - som
  - son
  - sop
  - sor
  - sot
  - soy
  - spp
  - spy
  - srr
  - ssn
  - ssw
  - ssy
  - stj
  - stv
  - suj
  - suk
  - sur
  - sus
  - swb
  - swc
  - swh
  - swj
  - swk
  - sxb
  - sxw
  - syk
  - sym
  - tak
  - tan
  - taq
  - taz
  - tbt
  - tbz
  - tcc
  - tcd
  - tdk
  - tdv
  - tdx
  - ted
  - tek
  - tem
  - teo
  - teq
  - tex
  - tfi
  - tgd
  - tgw
  - thk
  - thy
  - thz
  - tic
  - tig
  - tii
  - tik
  - tir
  - tiv
  - tja
  - tke
  - tkg
  - tlj
  - tll
  - tma
  - tmc
  - tmh
  - tng
  - tnr
  - tod
  - tog
  - toh
  - toi
  - toq
  - tpm
  - tsb
  - tsc
  - tsn
  - tso
  - tsp
  - tsv
  - ttb
  - ttq
  - ttr
  - tug
  - tui
  - tul
  - tum
  - tuq
  - tuv
  - tuy
  - tuz
  - tvd
  - tvs
  - tvu
  - twi
  - twx
  - tye
  - tzm
  - uba
  - udl
  - udu
  - uiv
  - ukp
  - ukw
  - ula
  - uly
  - umb
  - une
  - urh
  - uss
  - uta
  - uth
  - utr
  - uya
  - vag
  - vai
  - vaj
  - ven
  - ver
  - vif
  - vig
  - vkn
  - vmk
  - vmw
  - vor
  - vum
  - vun
  - vut
  - wal
  - wan
  - wbf
  - wbi
  - wci
  - weh
  - wem
  - wes
  - wib
  - wja
  - wji
  - wlc
  - wle
  - wlx
  - wmw
  - wni
  - wob
  - wof
  - wol
  - wom
  - won
  - wss
  - wti
  - wud
  - wwa
  - xdo
  - xed
  - xho
  - xkb
  - xkg
  - xkt
  - xkv
  - xmc
  - xmg
  - xmv
  - xog
  - xom
  - xon
  - xpe
  - xrb
  - xsm
  - xsn
  - xsq
  - xtc
  - xuu
  - xwe
  - xwg
  - xwl
  - yaf
  - yal
  - yam
  - yao
  - yay
  - yaz
  - yba
  - ybb
  - ybj
  - ybl
  - yer
  - 'yes'
  - yey
  - ymk
  - ymm
  - ynq
  - yns
  - yom
  - yor
  - zag
  - zay
  - zaz
  - zbu
  - zdj
  - zem
  - zhi
  - zim
  - zin
  - zlu
  - zmb
  - zms
  - zne
  - zns
  - zul
  - zuy
  - zwa
base_model:
  - utter-project/mHuBERT-147
license: cc

AfriHuBERT: A self-supervised speech representation model for African languages

Model description

This is a compact multilingual self-supervised speech encoder based on mHuBERT-147. We performed continued pretraining through multilingual adaptive finetuning (MAFT) on over 10,000 hours of African language data aggregated from various sources. This model can be considered the fourth iteration of mHuBERT-147, specifically trained on African languages. According to the paper, this is the AfriHuBERT-n model. You can click here for the AfriHuBERT-o model.

Pretraining data

  • Dataset: AfriHuBERT was trained on data from 11 major sources, including BibleTTS, Kallaama, MMS Ulab v2, NaijaVoices, and NCHLT. All sources and their licenses are shown in the table below. Please refer to the paper for more information. Logo

Language Coverage

AfriHuBERT covers 1,230 languages in total including 1,226 indigenous African languages

BibTeX entry and citation info.

@misc{alabi2024afrihubertselfsupervisedspeechrepresentation,
      title={AfriHuBERT: A self-supervised speech representation model for African languages}, 
      author={Jesujoba O. Alabi and Xuechen Liu and Dietrich Klakow and Junichi Yamagishi},
      year={2024},
      eprint={2409.20201},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2409.20201}, 
}