MeMAD H2020 Project
Published December 24, 2020 | Version v1
Other Open

MeMAD end-to-end speech translation models

Creators

  • 1. University of Helsinki

Contributors

Data curator:

  • 1. University of Helsinki

Description

A public collection of pre-trained translation, segmentation and truecasing models for the MeMAD end-to-end speech translation system, available via Github.

Files

speech-translation-models.zip

Files (2.7 GB)

Name Size Download all
md5:567c3f5f6b2a98ab68d7a09a421ec953
2.7 GB Preview Download

Additional details

Funding

European Commission
MeMAD - Methods for Managing Audiovisual Data: Combining Automatic Efficiency with Human Accuracy 780069