Hub: Training USE(DAN) on Plain Urdu(language) Text Corpus from scratch

Created on 4 Feb 2019  路  2Comments  路  Source: tensorflow/hub

Hi !
Hub Team!
Is there any way to train USE(DAN) or USE(Transformer) on Urdu (language) unsupervised plain text corpus.

I have read issues #155, #46 but there is no clear instructions about how to train USE on other than English language.

@vbardiovskyg
Thanks in Advance

text-embedding support
Source
picture MFarhatUllah

Most helpful comment

Hi, Hub Team,
I also have the same question. How can we train universal-sentence-encoder for other language?
Thanks in Advance

picture arghyachaks on 4 Mar 2019
👍3

All 2 comments

Hi, Hub Team,
I also have the same question. How can we train universal-sentence-encoder for other language?
Thanks in Advance

arghyachaks picture arghyachaks on 4 Mar 2019
👍3

As explained in https://github.com/tensorflow/hub/issues/155#issuecomment-422299981 there is no way to retrain the TF Hub modules with U.S.E. for a new language.

arnoegw picture arnoegw on 6 Mar 2019
Was this page helpful?
0 / 5 - 0 ratings

Related issues

r-wheeler picture r-wheeler  路  4Comments
bzburr picture bzburr  路  4Comments
martiansideofthemoon picture martiansideofthemoon  路  3Comments
marcoaleixo picture marcoaleixo  路  4Comments
dav-ell picture dav-ell  路  4Comments