Cree (cr) 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 sizecrwiki sample
original ote e ici takok cipicic otinikatew anihe mackikiatikw. awik e osami otakikomitc nama aric minikwaniwon, tepirak micta nipira mawotcitakaniwon ekotca e
despite their ancient pedigree, the julii caesares were not especially politically influential, although they had enjoyed some revival of their politi
enko anaha piresiw tekaci ka wapisitc e itikititc kirika ka micta mitcetitc ite e ici micta nameskarik acitc e kickapiskarik ekota meka e ici pinawetc
1000 ▁ote ▁e ▁ici ▁ta kok ▁cip icic ▁otin ikatew ▁anihe ▁mackikiatikw . ▁awik ▁e ▁osam i ▁otakikom itc ▁nama ▁aric ▁minik waniwon , ▁tepirak ▁micta ▁nipira ▁mawotc itak aniwon ▁ekotca ▁e
▁des p ite ▁the ir ▁an c ie n t ▁pe d ig ree , ▁the ▁jul i i ▁caesar es ▁we re ▁not ▁es p ec ial ly ▁p ol itic al ly ▁in f l u ent ial , ▁al t h ou g h ▁the y ▁h ad ▁en j o y ed ▁some ▁re v iv al ▁of ▁the ir ▁p ol it i
▁enko ▁anaha ▁pir esiw ▁tekaci ▁ka ▁wapisitc ▁e ▁itikititc ▁kirika ▁ka ▁micta ▁mitcet itc ▁ite ▁e ▁ici ▁micta ▁names ka rik ▁acitc ▁e ▁kic ka pis ka rik ▁ekota ▁meka ▁e ▁ici ▁p ina wetc
3000 ▁ote ▁e ▁ici ▁takok ▁cipicic ▁otinikatew ▁anihe ▁mackikiatikw . ▁awik ▁e ▁osami ▁otakikomitc ▁nama ▁aric ▁minikwaniwon , ▁tepirak ▁micta ▁nipira ▁mawotc itakaniwon ▁ekotca ▁e
▁des p ite ▁their ▁an c ie nt ▁pe d ig ree , ▁the ▁jul i i ▁caesar es ▁were ▁not ▁es pec ial ly ▁politic ally ▁influ ential , ▁al th ough ▁they ▁had ▁en jo y ed ▁some ▁re v ival ▁of ▁their ▁pol it i
▁enko ▁anaha ▁piresiw ▁tekaci ▁ka ▁wapisitc ▁e ▁itikititc ▁kirika ▁ka ▁micta ▁mitcetitc ▁ite ▁e ▁ici ▁micta ▁names karik ▁acitc ▁e ▁kicka pis karik ▁ekota ▁meka ▁e ▁ici ▁pinawetc
5000 ▁ote ▁e ▁ici ▁takok ▁cipicic ▁otinikatew ▁anihe ▁mackikiatikw . ▁awik ▁e ▁osami ▁otakikomitc ▁nama ▁aric ▁minikwaniwon , ▁tepirak ▁micta ▁nipira ▁mawotc itakaniwon ▁ekotca ▁e
▁des p ite ▁their ▁an cient ▁pe d ig ree , ▁the ▁julii ▁caesar es ▁were ▁not ▁espec ial ly ▁politic ally ▁influential , ▁alth ough ▁they ▁had ▁enjoy ed ▁some ▁re v ival ▁of ▁their ▁pol it i
▁enko ▁anaha ▁piresiw ▁tekaci ▁ka ▁wapisitc ▁e ▁itikititc ▁kirika ▁ka ▁micta ▁mitcetitc ▁ite ▁e ▁ici ▁micta ▁names karik ▁acitc ▁e ▁kickapiskarik ▁ekota ▁meka ▁e ▁ici ▁pinawetc