Special thanks:<ul><li><a target="_blank" rel="ugc" href="https://civarchive.com/models/459073?modelVersionId=512448">[SDXL&amp;Pony] dressing socks / adjusting legwear / sock pull / 靴下履き / 靴下直し</a></li><li><a target="_blank" rel="ugc" href="https://civarchive.com/models/317902?modelVersionId=593760">T-ponynai3 (v5.5)</a></li><li><a target="_blank" rel="ugc" href="https://civarchive.com/models/288584?modelVersionId=324619">AutismMix SDXL (Pony)</a></li></ul><hr />V2.0:Reduced "networkAlpha" from 16 to 12, changed to Prodigy, and added training steps.17 training pictures with natural language tags.Epoch #23 (17×6×23=2,346 steps) seems to perform better than Epoch #25 (2,550 steps).<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/294f4a8a-0013-491e-8623-a83b5812d36f/width=525/294f4a8a-0013-491e-8623-a83b5812d36f.jpeg" /><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/a6fab0ea-9d5a-4462-8d4b-35ee28d522c5/width=525/a6fab0ea-9d5a-4462-8d4b-35ee28d522c5.jpeg" /><pre><code>{
 "engine": "kohya",
 "unetLR": 1,
 "clipSkip": 1,
 "loraType": "lora",
 "keepTokens": 0,
 "networkDim": 2,
 "numRepeats": 6,
 "resolution": 512,
 "lrScheduler": "cosine",
 "minSnrGamma": 5,
 "noiseOffset": 0.1,
 "targetSteps": 850,
 "enableBucket": true,
 "networkAlpha": 12,
 "optimizerType": "Prodigy",
 "textEncoderLR": 0,
 "maxTrainEpochs": 25,
 "shuffleCaption": false,
 "trainBatchSize": 3,
 "flipAugmentation": false,
 "lrSchedulerNumCycles": 3
}</code></pre>Adjusting legwear (tried flux version).You need to add some specific types of legwear to get good results (e.g., white socks).Can be used with realistic lora.

Files with SHA256: e48e05bc8d07563621dcf05b986158052e13c6261bb8b2710289d542f9065bdb