This package contains an international English acoustic model for Sphinx. It was trained from 100+ hours of recorded speech from the voxforge project, downsampled to 16kHz.