Inupiaq (ik) subword embeddings

Vocab size vocab model 25 dim 50 dim 100 dim 200 dim 300 dim
1000 vocab model txt | bin
bokeh | umap | matrix
txt | bin
bokeh | umap | matrix
txt | bin
bokeh | umap | matrix
txt | bin
bokeh | umap | matrix
txt | bin
bokeh | umap | matrix
3000 vocab model txt | bin
bokeh | umap | matrix
txt | bin
bokeh | umap | matrix
txt | bin
bokeh | umap | matrix
txt | bin
bokeh | umap | matrix
txt | bin
bokeh | umap | matrix
5000 vocab model txt | bin
bokeh | umap | matrix
txt | bin
bokeh | umap | matrix
txt | bin
bokeh | umap | matrix
txt | bin
bokeh | umap | matrix
txt | bin
bokeh | umap | matrix

Embedding matrix plots

Training corpus sample, encoded with different BPE vocabulary sizes

Vocab sizeikwiki sample
original jeesus kristus kristo redentor, rio de janeiro, brazil. jeesus kristus on jumala poeg ning vanas testamendis ennustatud messias. kristlased käsitlevad
* emma bodfish suli david baumgartner (0000), iñupiat grammar. utqiaġvigmi: utqiaġvium minuaqtuġviata iñupiatun savagvianni * jonathan m. nielson (000
when it comes to planning a wedding, i always advise people to start with a budget. the budget is the framework that tells you how much wedding you ca
1000 ▁jeesus ▁krist us ▁krist o ▁r ed ent or , ▁ri o ▁de ▁jane ir o , ▁b r az il . ▁jeesus ▁krist us ▁on ▁j um ala ▁p o eg ▁ning ▁van as ▁testamend is ▁en n ust atud ▁m es si as . ▁kristlased ▁käsitle v ad
▁* ▁ em ma ▁b od f ish ▁suli ▁d av id ▁b a um g art n er ▁(0000), ▁iñupiat ▁g ram m ar . ▁utqiaġv ig mi : ▁utqiaġv ium ▁m in uaq tu ġv i ata ▁iñupiatun ▁sav ag v i ann i ▁* ▁j on at h an ▁m . ▁ni el son ▁( 000
▁wh en ▁it ▁com es ▁to ▁pl ann ing ▁a ▁wedding , ▁i ▁al way s ▁a d v is e ▁people ▁to ▁start ▁with ▁a ▁budget . ▁the ▁budget ▁is ▁the ▁f ram ew ork ▁that ▁t ell s ▁y ou ▁h ow ▁m u ch ▁wedding ▁y ou ▁c a
3000 ▁jeesus ▁kristus ▁krist o ▁r edent or , ▁rio ▁de ▁jane iro , ▁b razil . ▁jeesus ▁kristus ▁on ▁jumala ▁poeg ▁ning ▁vanas ▁testamendis ▁ennustatud ▁messias . ▁kristlased ▁käsitlevad
▁* ▁emma ▁bod fish ▁suli ▁david ▁baumg artner ▁(0000), ▁iñupiat ▁gram mar . ▁utqiaġv igmi : ▁utqiaġv ium ▁min uaqtu ġviata ▁iñupiatun ▁savag vianni ▁* ▁j onath an ▁m . ▁ni elson ▁( 000
▁when ▁it ▁comes ▁to ▁planning ▁a ▁wedding , ▁i ▁alway s ▁ad vise ▁people ▁to ▁start ▁with ▁a ▁budget . ▁the ▁budget ▁is ▁the ▁f ramew ork ▁that ▁tells ▁you ▁how ▁much ▁wedding ▁you ▁c a
5000 ▁jeesus ▁kristus ▁kristo ▁redentor , ▁rio ▁de ▁janeiro , ▁brazil . ▁jeesus ▁kristus ▁on ▁jumala ▁poeg ▁ning ▁vanas ▁testamendis ▁ennustatud ▁messias . ▁kristlased ▁käsitlevad
▁* ▁emma ▁bodfish ▁suli ▁david ▁baumgartner ▁(0000), ▁iñupiat ▁grammar . ▁utqiaġvigmi : ▁utqiaġvium ▁minuaqtuġviata ▁iñupiatun ▁savagvianni ▁* ▁jonathan ▁m . ▁nielson ▁( 000
▁when ▁it ▁comes ▁to ▁planning ▁a ▁wedding , ▁i ▁always ▁advise ▁people ▁to ▁start ▁with ▁a ▁budget . ▁the ▁budget ▁is ▁the ▁framework ▁that ▁tells ▁you ▁how ▁much ▁wedding ▁you ▁ca