This model is created from a large dataset of various colour "Bridal style lingerie"Each image in the dataset was carefully picked for quality and look and added for training.I hate to make this early access so i APPRECIATE ANY TIPS :) and i did 5 runs of this before i was happy with the outcome and i will leave it up to you to verify the integrity of the produced images.So i added a single trigger word covering multiple colours and styles.There are many variations in the "br1dal white lingerie" triggerword and at 0.8 to 1 strength will produce some very nice setsThe "br1dal black lingerie" is similar with less variation in the set.The "br1dal red lingerie" is similar with less variation in the set.The "br1dal yellow lingerie" is similar with less variation in the set but will take colour mixing very well particulary "br1dal yellow lingerie with pink lace trimming" etcExample:<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/87fcf8d3-c5d2-4c68-84bc-b62a65317b9d/width=525/87fcf8d3-c5d2-4c68-84bc-b62a65317b9d.jpeg" />The "br1dal pink lingerie" is similar with less variation in the set but will take colour mixing very well particulary "br1dal yellow lingerie with pink lace trimming" etcThe "br1dal purple lingerie" is similar with less variation in the set.The "br1dal green lingerie" is similar with less variation in the set.<hr />Use the Euler sampler with simple scheduler.Steps 30 generally does well with better detail the higher you ventureCFG:1I use 2.2 for Flux GuidanceMax Shift: 1.5Base Shift: 1.0<hr />For Upscaling this works very well for me.<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/9e1b06ca-63f9-4335-ae66-eb3b0f5b0be6/width=525/9e1b06ca-63f9-4335-ae66-eb3b0f5b0be6.jpeg" /><hr /><hr />That covers everything, looking forward to seeing your generations and don't forget to post some of your usage here.

Files with SHA256: 29ab1b62ed5ece1572cf46fda0594e7a28f8c261abe3820c58f71de367929180