hellosindh
/

sindhi-bert-base

masked-language-modeling

Model card Files Files and versions

sindhi-bert-base / tokenizer.json

Commit History

Revert: clean tokenizer matching SP model exactly

ee669ca
verified

hellosindh commited on 27 days ago

Fix: correct special token order and remove duplicate </s>

b3274ee
verified

hellosindh commited on 27 days ago

Revert: restore original vocab order

f67628e
verified

hellosindh commited on 27 days ago

Fix: swap mask token to correct index 32000 in Unigram vocab

00b38ab
verified

hellosindh commited on 27 days ago