Representations of language in a model of visually grounded speech signal

Connections are not prepared for this record yet, or its source metadata has changed. The original record remains available while background snapshots are rebuilt.