This version reduces the image training to just 768x768 and 512x768, adds some good gens from v1.0 and cleaned up the tags and increased training steps from 2k to 4k for more accurate facial features.
This version reduces the image training to just 768x768 and 512x768, adds some good gens from v1.0 and cleaned up the tags and increased training steps from 2k to 4k for more accurate facial features.