Fasttext: "Continuing" training of pre-trained word vector models.

Created on 25 Jul 2017  路  4Comments  路  Source: facebookresearch/fastText

I want to add vocabulary to and fine-tune existing embeddings of a pre-trained word vector model provided by FB (wiki.id.bin/vec, specifically) using my own text corpus (Tweets and public Facebook posts in Indonesian, specifically). Is there a way to do this?

alignmenupdate
Source
picture rhezab
👍13

Most helpful comment

Any updates on this? Seems like a critical feature for many tasks.

picture arshak on 16 Sep 2017
👍8

All 4 comments

Any updates on this? Seems like a critical feature for many tasks.

arshak picture arshak on 16 Sep 2017
👍8

any update for this question?

jkneng picture jkneng on 26 Feb 2018
👍6

Any insight on a possible timeline for this?

prokopyev picture prokopyev on 21 Mar 2018

This still un-merged PR is related: #423

(I haven't yet tested it)

mino98 picture mino98 on 21 Mar 2018
Was this page helpful?
0 / 5 - 0 ratings

Related issues

flybirp picture flybirp  路  4Comments
hughbzhang picture hughbzhang  路  3Comments
nomadlx picture nomadlx  路  3Comments
shriiitk picture shriiitk  路  3Comments
yasonk picture yasonk  路  3Comments