V2. train 12 different outfit see what happenedTraining with<ol><li>Zoe Crop Linen Brown Midi Skirt</li><li>High Waist Skirt with Pocket</li><li>Fall Spring</li><li>Grey Tank Top</li><li>Smile</li><li>White Vest</li><li>White Tight Trouser</li><li>Grey Blazer</li><li>Satin Pajamas Soft Polyester Sleep</li><li>Jeans</li><li>Blazer</li><li>Medium Shoot,</li></ol> if one of the training word use , it will change the background into a dirty seamless background use weight below 0.6 so that the background will follow the prompt

DBXLORA-Concept Outfit

Training Word zoe crop linen brown midi skirt, high waist skirt with pocket, fall spring, grey tank top, smile, white vest, white tight trouser, grey blazer, satin pajamas soft polyester sleep, jeans, blazer, medium shoot, if one of the training word use , it will change the background into a dirty seamless background use weight below 0.6 so that the background will follow the promptTraining Params<pre><code>{
 "unetLR": 0.0005,
 "clipSkip": 1,
 "loraType": "lora",
 "keepTokens": 0,
 "networkDim": 32,
 "numRepeats": 7,
 "resolution": 512,
 "lrScheduler": "cosine_with_restarts",
 "minSnrGamma": 5,
 "targetSteps": 1085,
 "enableBucket": true,
 "networkAlpha": 16,
 "optimizerArgs": "weight_decay=0.1",
 "optimizerType": "AdamW8Bit",
 "textEncoderLR": 0.00005,
 "maxTrainEpochs": 10,
 "shuffleCaption": true,
 "trainBatchSize": 2,
 "flipAugmentation": true,
 "lrSchedulerNumCycles": 3
}</code></pre>

Files with SHA256: ca5321ec52ec4e4d0f40fe2ddb573dd26457a79fca77cebc8bf8c9b7c0cbd0ed