# The Dota 2 characters all have around the same duration in voice-lines, so if it is possible could more models be trained to get a wider variety of characters to use please?