Sranan Tongo (srn) 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

View Sranan Tongo subword embeddings in TensorFlow Projector

2-D UMAP plots

Embedding matrix plots

Training corpus sample, encoded with different BPE vocabulary sizes

Vocab sizesrnwiki sample
original disi periodi efru yari e ben wan yari nanga pasa peyna nanga ledi, ma oktu nanga gu tengi. disi e ben no referi na nanga pasar man e taki dati disi ot
disi wortu e ben wan gramatika isyu. a e ben wan ondrodeli fu a sranantongo tongo. efru yu pasa no abra disi ondrowerpi yu e ben nodi tu klik opo 'ken
0 ben wan numro. a orginali fu den arabi arabiskrifi. dy e skrifi numro leki normali letri nanga a e ben wan siri opukini masukiri um no a tuki tu gi.
1000 ▁disi ▁periodi ▁efru ▁yari ▁e ▁ben ▁wan ▁yari ▁nanga ▁pasa ▁peyna ▁nanga ▁ledi , ▁ma ▁oktu ▁nanga ▁gu ▁tengi . ▁disi ▁e ▁ben ▁no ▁referi ▁na ▁nanga ▁pasar ▁man ▁e ▁taki ▁dati ▁disi ▁o t
▁disi ▁wortu ▁e ▁ben ▁wan ▁gramatika ▁isyu . ▁a ▁e ▁ben ▁wan ▁ondrodeli ▁fu ▁a ▁sranantongo ▁tongo . ▁efru ▁yu ▁pasa ▁no ▁abra ▁disi ▁ondrowerpi ▁yu ▁e ▁ben ▁nodi ▁tu ▁klik ▁opo ▁' k en
▁0 ▁ben ▁wan ▁numro . ▁a ▁orginali ▁fu ▁den ▁arabi ▁arabiskrifi . ▁dy ▁e ▁skrifi ▁numro ▁leki ▁normali ▁letri ▁nanga ▁a ▁e ▁ben ▁wan ▁siri ▁opukini ▁masukiri ▁um ▁no ▁a ▁tuki ▁tu ▁gi .
3000 ▁disi ▁periodi ▁efru ▁yari ▁e ▁ben ▁wan ▁yari ▁nanga ▁pasa ▁peyna ▁nanga ▁ledi , ▁ma ▁oktu ▁nanga ▁gu ▁tengi . ▁disi ▁e ▁ben ▁no ▁referi ▁na ▁nanga ▁pasar ▁man ▁e ▁taki ▁dati ▁disi ▁o t
▁disi ▁wortu ▁e ▁ben ▁wan ▁gramatika ▁isyu . ▁a ▁e ▁ben ▁wan ▁ondrodeli ▁fu ▁a ▁sranantongo ▁tongo . ▁efru ▁yu ▁pasa ▁no ▁abra ▁disi ▁ondrowerpi ▁yu ▁e ▁ben ▁nodi ▁tu ▁klik ▁opo ▁' ken
▁0 ▁ben ▁wan ▁numro . ▁a ▁orginali ▁fu ▁den ▁arabi ▁arabiskrifi . ▁dy ▁e ▁skrifi ▁numro ▁leki ▁normali ▁letri ▁nanga ▁a ▁e ▁ben ▁wan ▁siri ▁opukini ▁masukiri ▁um ▁no ▁a ▁tuki ▁tu ▁gi .
5000 ▁disi ▁periodi ▁efru ▁yari ▁e ▁ben ▁wan ▁yari ▁nanga ▁pasa ▁peyna ▁nanga ▁ledi , ▁ma ▁oktu ▁nanga ▁gu ▁tengi . ▁disi ▁e ▁ben ▁no ▁referi ▁na ▁nanga ▁pasar ▁man ▁e ▁taki ▁dati ▁disi ▁o t
▁disi ▁wortu ▁e ▁ben ▁wan ▁gramatika ▁isyu . ▁a ▁e ▁ben ▁wan ▁ondrodeli ▁fu ▁a ▁sranantongo ▁tongo . ▁efru ▁yu ▁pasa ▁no ▁abra ▁disi ▁ondrowerpi ▁yu ▁e ▁ben ▁nodi ▁tu ▁klik ▁opo ▁' ken
▁0 ▁ben ▁wan ▁numro . ▁a ▁orginali ▁fu ▁den ▁arabi ▁arabiskrifi . ▁dy ▁e ▁skrifi ▁numro ▁leki ▁normali ▁letri ▁nanga ▁a ▁e ▁ben ▁wan ▁siri ▁opukini ▁masukiri ▁um ▁no ▁a ▁tuki ▁tu ▁gi .