This background LoRA for <a target="_blank" rel="ugc" href="https://civarchive.com/models/353543?modelVersionId=492740">(AAA)Autism </a><code>v3</code> is intended to add detailed backgrounds with an aesthetic to my preference.Useful tags: <code>detailed background</code>, <code>simple background</code>, <code>gradient background</code>, <code>{COLOR} background</code>Recommended tags: <code>score_9, score_8, score_7</code>This is my first time using the Civitai Trainer. The detail is nowhere close to my satisfaction yet but it seems to have captured the aesthetic of the dataset.⚠️ It does not work as intended with Pony v6 XL. Will try retraining with clip skip 2 to hopefully improve compatibility.

Detailed Background

This is my first test. The training data was quite small, only 2 versions of 15 images, and only anime and furry was used so I don't know how well it will work for other uses.Training hyperparameters:<pre><code>{
 "unetLR": 0.0005,
 "clipSkip": 1,
 "loraType": "lora",
 "keepTokens": 0,
 "networkDim": 32,
 "numRepeats": 2,
 "resolution": 1024,
 "lrScheduler": "cosine_with_restarts",
 "minSnrGamma": 5,
 "noiseOffset": 0.1,
 "targetSteps": 600,
 "enableBucket": true,
 "networkAlpha": 16,
 "optimizerType": "AdamW8Bit",
 "textEncoderLR": 0.00005,
 "maxTrainEpochs": 40,
 "shuffleCaption": true,
 "trainBatchSize": 4,
 "flipAugmentation": false,
 "lrSchedulerNumCycles": 3
}</code></pre>

Files with SHA256: 1161da000d78a23d27b63676d8599ecb60ad2637c11b104424072cf0f20b08ee