(Translated by https://www.hiragana.jp/)
GitHub - mwydmuch/datasets4vw: Multiclass and multilabel datasets in Vowpal Wabbit format

mwydmuch / datasets4vw Public

Notifications You must be signed in to change notification settings
Fork 1
Star 12

Multiclass and multilabel datasets in Vowpal Wabbit format

12 stars 1 fork Branches Tags Activity

Notifications

Name		Name	Last commit message	Last commit date
Latest commit History 27 Commits
examples		examples
pd_sparse_multiclass		pd_sparse_multiclass
pd_sparse_multilabel		pd_sparse_multilabel
tools		tools
xml_repo		xml_repo
.gitignore		.gitignore
README.md		README.md
get_all.sh		get_all.sh
get_all_from_pd_sparse.sh		get_all_from_pd_sparse.sh
get_all_from_xml_repo.sh		get_all_from_xml_repo.sh

Repository files navigation

datasets4vw

Scripts to download and convert different datasets to Vowpal Wabbit format. Repository contains scripts for the following datasets:

Usage: bash ./<repo>/get_<dataset-name>.sh

`xml_repo` (multilabel):

Amazon-3M (amazon-3M)
Amazon-670K (amazon)
AmazonCat-13K (amazonCat)
AmazonCat-14K (amazonCat-14K)
Bibtex (bibtex)
Delicious (delicious)
Delicious-200K (deliciousLarge)
EURLex-4K (eurlex)
Mediamill (mediamill)
RCV1-2K (rcv1x)
Wiki10-31K (wiki10)
WikiLSHTC-325K (wikiLSHTC)
Wikipedia-500K (WikipediaLarge-500K)

`pd_sparse(_multiclass)`:

aloi.bin
Dmoz
imageNet
LSHTC1
sector

`pd_sparse_multilabel`:

Eur-Lex
rcv1_regions
bibtex
LSHTCwiki

Mac users need to install gnu version of sed:

brew install gnu-sed

About

Multiclass and multilabel datasets in Vowpal Wabbit format

vowpal-wabbit datasets multi-label multi-class extreme-large-datasets

Report repository

Releases

No releases published

Packages

No packages published

Languages