MT baseline results
For some language pairs among those for which benchmarks are available, results of baseline SMT systems are provided below.
The baseline systems were built upon the open-source MT toolkit Moses in a pretty standard configuration; see the WIT3 paper for details.
Performance in terms of BLEU% and TER were computed by means of the MultEval software and refer to tst2012 evaluation set. Automatic translations are linked to entries, just click for getting them.