transform your concept, idea into a dark creepy vibewith lora<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/c6c3a969-d108-4756-bcfb-c96b5da0bdc7/width=525/c6c3a969-d108-4756-bcfb-c96b5da0bdc7.jpeg" />without lora<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/a06948f6-9d6d-45f2-9753-73efe01026d1/width=525/a06948f6-9d6d-45f2-9753-73efe01026d1.jpeg" />s4_q2_m9-enhanced-style i use this as the main tag for trainingthe model already knows what enhanced and style are. i attach it to the made up trigger to give it weight and also a direction to take the training (enhanced style) so even if it already knows the concept--- we are saying learn more : )the result:a lora that has two flavorswith the full training tag s4_q2_m9-enhanced-styleor a lighter effect with just the s4_q2_m9i first tested this with challenge lora for tough concepts and found when you are going for a certain look / style if the training data is strong enough, this method results in:a lora with a greater freedom to apply the effect without taking over the compositionit will be a softer touch, slightly different approach, that allows you to use additional loras without washing the image out.

Files with SHA256: 786658b34c7949dfb82f0e9cb1e5eb9107fca94fd090562634a378e4791e668d