Cheyenne (chy) 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 |
Training corpus sample, encoded with different BPE vocabulary sizes
| Vocab size | chywiki sample |
|---|---|
| original | file:dad bareback.jpg|é-pónoma'e'omo'he. file:collectie tropenmuseum batakse man te paard tmnr 00000000.jpg|tsé-pónoma'eéxoo'êstse, indonesia, 0000 originalpancakes.jpg|e'kôsé'haseo'o, heóveamêške, pano'êhasëö'o file:maple sap buckets - beaver meadow audubon center.jpg|hoóhtsetsé-mahpe. file:acer file:making pancakes.jpg| file:korea-jeongseon-making korean pancakes (jeon) at a market-00.jpg|é-e'omâhonóne. originalpancakes.jpg|e'kôsé'haseo'o, he |
| 1000 | ▁file : d ad ▁b are ba ck . jpg | é - pó noma ' e ' omo ' he . ▁file : col lect ie ▁t rop en m useum ▁b ata k se ▁man ▁t e ▁pa ard ▁t m n r ▁0000 0000. jpg | tsé - pó noma ' e é x oo ' êstse , ▁in d one s ia , ▁0000 ▁ or igin al pan ca kes . jpg | e ' kôsé ' ha seo ' o , ▁he óve am êš ke , ▁p ano ' ê has ëö ' o ▁file : ma ple ▁sa p ▁b uck et s ▁- ▁bea ver ▁m ea do w ▁a u du bon ▁cent er . jpg | hoó htse tsé - mahpe . ▁file : a cer ▁file : ma king ▁p an ca kes . jpg | ▁file : k or ea - j e ong se on - ma king ▁k ore an ▁p an ca kes ▁( j e on ) ▁at ▁a ▁m ark et - 00. jpg | é - e ' om âhonóne . ▁ or igin al pan ca kes . jpg | e ' kôsé ' ha seo ' o , ▁he |
| 3000 | ▁file : d ad ▁b are back . jpg | é - pónoma ' e ' omo ' he . ▁file : collectie ▁tropen museum ▁bata k se ▁man ▁t e ▁pa ard ▁tmnr ▁00000000. jpg | tsé - pónoma ' e éx oo ' êstse , ▁in d onesia , ▁0000 ▁origin alpancakes . jpg | e ' kôsé ' haseo ' o , ▁heóveamêške , ▁pano ' êhasëö ' o ▁file : maple ▁sap ▁buckets ▁- ▁beaver ▁meadow ▁audubon ▁center . jpg | hoóhtsetsé - mahpe . ▁file : acer ▁file : making ▁pancakes . jpg | ▁file : korea - je ong seon - making ▁k ore an ▁pancakes ▁( je on ) ▁at ▁a ▁m ark et - 00. jpg | é - e ' om âhonóne . ▁origin alpancakes . jpg | e ' kôsé ' haseo ' o , ▁he |
| 5000 | ▁file : d ad ▁b areback . jpg | é - pónoma ' e ' omo ' he . ▁file : collectie ▁tropenmuseum ▁batak se ▁man ▁t e ▁pa ard ▁tmnr ▁00000000. jpg | tsé - pónoma ' e éx oo ' êstse , ▁in donesia , ▁0000 ▁origin alpancakes . jpg | e ' kôsé ' haseo ' o , ▁heóveamêške , ▁pano ' êhasëö ' o ▁file : maple ▁sap ▁buckets ▁- ▁beaver ▁meadow ▁audubon ▁center . jpg | hoóhtsetsé - mahpe . ▁file : acer ▁file : making ▁pancakes . jpg | ▁file : korea - jeong seon - making ▁kore an ▁pancakes ▁( jeon ) ▁at ▁a ▁m arket - 00. jpg | é - e ' om âhonóne . ▁origin alpancakes . jpg | e ' kôsé ' haseo ' o , ▁he |