Weight used: 0.70
Model used for samples: Realistic Vision v1.3 with epi_noiseoffset and bad_prompt_v2
CFG scale for last 2 images were 5.5 not 6
Note: Multiple subjects is inherently tougher so please don't expect perfect results.
Description
Training info (useful in picking resolutions):
max_resolution: 768x768
image_count: 19
bucket 0: resolution (448, 704), count: 1
bucket 1: resolution (512, 896), count: 1
bucket 2: resolution (576, 512), count: 1
bucket 3: resolution (576, 832), count: 2
bucket 4: resolution (576, 896), count: 1
bucket 5: resolution (640, 640), count: 1
bucket 6: resolution (704, 512), count: 1
bucket 7: resolution (704, 704), count: 1
bucket 8: resolution (704, 832), count: 1
bucket 9: resolution (768, 640), count: 3
bucket 10: resolution (832, 576), count: 4
bucket 11: resolution (896, 576), count: 2