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 .<pre><code>{
 "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
}</code></pre>

Mashup

i-ll put 3 epochs , no trigger words

Files with SHA256: 59494bf5e12de7c82a25f918d5c8efd66e6d9d2ca276e7c6aa97eae580ce7b91