I took a text-to-video workflow from this Kosinkadink github https://github.com/Kosinkadink/ComfyUI-AnimateDiff-Evolved and modified it. I keep telling it I want one person not two and it keeps producing 2. what am I doing wrong? and what prompts can I enter to enhance better face resemblance? (yes i know i misspelled "multiple" in my positive prompt but i corrected it and still got 2 people). ALSO do you know why this workflow keeps crapping out? i've heard numerous youtube videos lamenting on the fact that animatediff really only works nest in 16 frame batches for video, but i set it to 16 and its still crapping out.