Crossmodal Supervised Learning Toolkit using High-Performance Extreme Learning Machines over the audio-visual-textual data