I had the idea to create a LoRA model using a dataset of images of female DJs in different poses and places. This resulted in a very flexible model that can generate a wide variety of images.
Prompt example:
Positive: “a photo of a female dj, a woman in headphones is playing a cdj, turning knobs in a mixer”
Negative: “(worst quality, low quality, normal quality, lowres, low details, oversaturated, undersaturated, overexposed, underexposed, grayscale, bw, bad photo, bad photography, bad-hands-5, bad art:1.4), (watermark, signature, text font, username, error, logo, words, letters, digits, autograph, trademark, name:1.2), (blur, blurry, grainy), morbid, ugly, asymmetrical, mutated malformed, mutilated, poorly lit, bad shadow, draft, cropped, out of frame, cut off, censored, jpeg artifacts, out of focus, glitch, duplicate, (airbrushed, cartoon, anime, semi-realistic, cgi, render, blender, digital art, manga, amateur:1.3), (3D ,3D Game, 3D Game Scene, 3D Character:1.1), (bad hands, bad fingers, bad anatomy, bad body, bad face, bad teeth, bad arms, bad legs, deformities:1.3), oversaturated:1.3”
I haven’t tried different resolutions, but please feel free to post your version here.
Note: This LoRA model is still under development, so some strange text may appear on the DJ equipment.
Setup: For best results, I highly recommend using ComfyUI’s SeargeSDXL workflows.
These images were generated using the Searge-SDXL-Reborn-v4_0-hq workflow, but it is also possible to generate amazing images using A1111.