Latgalian (ltg) 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
10000 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

Training corpus sample, encoded with different BPE vocabulary sizes

Vocab sizeltgwiki sample
original {| border=0 align=right cellpadding=0 cellspacing=0 width=000 style="margin: 0 0 0em 0em; background: #f0f0f0; border: 0px #aaaaaa solid; border-colla
vylki byun da 000 ceņtimetru garuma, placu augstums 0,0-0,00 metri i svors 00-00 kilogrami (rekords - 00 kilogrami). augums apaug ar rupim palākim mot
benedikts zazerskis (0000–0000) — sakstagalīts, krimys gubernators nu 0000 da 0000 godam, ir dabuojis generala pakuopi. uzturiejs sakarus ar dzymtū pu
1000 ▁{| ▁border =0 ▁align = right ▁cellpadding =0 ▁cellspacing =0 ▁width =000 ▁style =" margin : ▁0 ▁0 ▁0 em ▁0 em ; ▁background : ▁# f 0 f 0 f 0; ▁border : ▁0 px ▁# aaaaaa ▁solid ; ▁border - c oll a
▁vy l ki ▁b yun ▁da ▁000 ▁c eņ t im e tr u ▁gar uma , ▁pl ac u ▁augst ums ▁0,0 - 0 , 00 ▁m e tr i ▁i ▁sv ors ▁00 - 00 ▁k il og r am i ▁( r ek or ds ▁- ▁00 ▁k il og r am i ). ▁aug ums ▁ap aug ▁ar ▁r up im ▁pal ā k im ▁m ot
▁b en ed ik ts ▁z az ers k is ▁(0000 – 0000 ) ▁— ▁s akst ag al īt s , ▁kr im ys ▁g ub er n at ors ▁nu ▁0000 ▁da ▁0000 ▁god am , ▁ir ▁dab uoj is ▁g en er al a ▁p ak uo p i . ▁ uz tur iejs ▁s ak ar us ▁ar ▁dz ym t ū ▁p u
3000 ▁{| ▁border =0 ▁align = right ▁cellpadding =0 ▁cellspacing =0 ▁width =000 ▁style =" margin : ▁0 ▁0 ▁0 em ▁0 em ; ▁background : ▁# f 0 f 0 f 0; ▁border : ▁0 px ▁# aaaaaa ▁solid ; ▁border - c oll a
▁vyl ki ▁byun ▁da ▁000 ▁c eņt im etru ▁gar uma , ▁pl ac u ▁augstums ▁0,0 - 0 ,00 ▁m etr i ▁i ▁svors ▁00-00 ▁k il ogr ami ▁( r ek or ds ▁- ▁00 ▁k il ogr ami ). ▁aug ums ▁ap aug ▁ar ▁rup im ▁palā kim ▁m ot
▁ben ed ik ts ▁z az ers k is ▁(0000 – 0000 ) ▁— ▁s akst ag al īts , ▁kr im ys ▁g ub ern at ors ▁nu ▁0000 ▁da ▁0000 ▁godam , ▁ir ▁dab uojis ▁gen er ala ▁pak uop i . ▁uz tur iejs ▁sak ar us ▁ar ▁dzym tū ▁pu
5000 ▁{| ▁border =0 ▁align = right ▁cellpadding =0 ▁cellspacing =0 ▁width =000 ▁style =" margin : ▁0 ▁0 ▁0 em ▁0 em ; ▁background : ▁# f 0 f 0 f 0; ▁border : ▁0 px ▁# aaaaaa ▁solid ; ▁border - c oll a
▁vyl ki ▁byun ▁da ▁000 ▁c eņt im etru ▁gar uma , ▁placu ▁augstums ▁0,0 -0 ,00 ▁metr i ▁i ▁svors ▁00-00 ▁kil ogr ami ▁( r ek or ds ▁- ▁00 ▁kil ogr ami ). ▁aug ums ▁ap aug ▁ar ▁rup im ▁palā kim ▁mot
▁ben ed ik ts ▁z az ers kis ▁(0000 –0000 ) ▁— ▁sakst ag al īts , ▁kr im ys ▁gub ern ators ▁nu ▁0000 ▁da ▁0000 ▁godam , ▁ir ▁dab uojis ▁gener ala ▁pakuop i . ▁uz tur iejs ▁sak ar us ▁ar ▁dzym tū ▁pu
10000 ▁{| ▁border =0 ▁align = right ▁cellpadding =0 ▁cellspacing =0 ▁width =000 ▁style =" margin : ▁0 ▁0 ▁0 em ▁0 em ; ▁background : ▁# f 0 f 0 f 0; ▁border : ▁0 px ▁# aaaaaa ▁solid ; ▁border - c oll a
▁vylki ▁byun ▁da ▁000 ▁c eņt im etru ▁garuma , ▁placu ▁augstums ▁0,0 -0 ,00 ▁metri ▁i ▁svors ▁00-00 ▁kil ogrami ▁( rek or ds ▁- ▁00 ▁kil ogrami ). ▁augums ▁ap aug ▁ar ▁rup im ▁palā kim ▁mot
▁ben ed ikts ▁z az erskis ▁(0000–0000 ) ▁— ▁sakst ag al īts , ▁krim ys ▁gubern ators ▁nu ▁0000 ▁da ▁0000 ▁godam , ▁ir ▁dab uojis ▁gener ala ▁pakuop i . ▁uz tur iejs ▁sak ar us ▁ar ▁dzymtū ▁pu