This is Mashup , a woman made from random generated images , her age is 25-30 , 155-165cm tall , small to medium sized chest etc (no nudes used, only different angles of faces and body) . I've taken the most beautiful woman portraits/full body shots (60 images) that i've made with z-image, sdxl , krea2 , etc and put them together to make this lora .

Settings : each image seen 65x , even if i have 30 images (best lora i have made) i make sure the images are seen 60+ times , at epoch 9 it should be ok but to capture everything i put 15 epochs . The more images the more steps , but over 100 images for a character is bad .

{
  "lr": 0.0001,
  "steps": 4000,
  "engine": "ai-toolkit",
  "epochs": 15,
  "batchSize": 1,
  "ecosystem": "krea2",
  "keepTokens": 0,
  "networkDim": 32,
  "resolution": 1024,
  "lrScheduler": "constant",
  "minSnrGamma": null,
  "noiseOffset": null,
  "networkAlpha": 32,
  "optimizerType": "adamw8bit",
  "shuffleTokens": false,
  "textEncoderLr": null,
  "flipAugmentation": false,
  "trainTextEncoder": false
}