Try the Europarl corpus (http://www.statmt.org/europarl/) or the TED corpus (https://github.com/ajinkyakulkarni14/TED-Multilingual-Parall...).