MCIF-Short Data Share

data-share-mcif-short

Add Voxtral outputs for wmt

MCIF-Long Data Share

data-share-mcif

Removed extra pairs Europarl and fixed json schemma

mandi audio

Zipped mandi audio files

CoVoST 2 Data Share

Data release for CoVoST 2. Due to the size of the data, zip files had to be partitioned. The zip files should be concatenated prior to unzipping, i.e.,

cat covost_en.zip.part-* > covost_en.zip
unzip covost_en.zip

ACL 60/60 Unsegmented Data Share

These audio files must be manually downloaded and placed under

H2T_DATADIR/acl6060/audio/en/.

The original files are from Salesky et al. 2023.

CommonAccent Audio Share

Audio for commonAccent benchmark.