Instruction-tuned (chat) versions of Nvidia's Minitron base models created through supervised fine-tuning (SFT)