This LORA was created by using over 100 images that I had created in MidJourney. These images were then mirrored to increase the training set. I then had ChatGPT analyze about a dozen images and make recommendations for the LORA settings in CivitAI. The result of that experiment is what is being shown here.

Settings:

The LORA prompt works best around 8 and the CFG around 8. However play around with it, from 5 / 5 up to 10 / 10

Here are some prompts to put before your basic descriptive prompt:

(((masterpiece))), ((ultra-detailed)), ((highly aesthetic)), ((visionary art style)), ((surreal fantasy painting)), with ((decorative swirling patterns)) and ((glowing radiant light)), rendered in ((8k resolution)). (((Digital illustration))) filled with ((intricate spirals)) and ((lace-like textures)) that flow into organic, natural shapes. The composition has a ((dreamlike atmosphere)) with striking contrasts of ((warm golden sunset tones)) against ((cool twilight blues)). Painterly brushstrokes create a ((whimsical yet refined composition)) enhanced by glowing highlights and cinematic depth. (((sharp focus))), ((vibrant colors)), ((surreal landscape)), and a ((highly detailed illustration)) with immersive visual richness.

Here are some negative prompts to use:

(pumpkin:1.3, pumpkins:1.3, apple:1.3, apples:1.3, fruit:1.2, vegetables:1.2, harvest:1.2, Halloween:1.2, orange fruit:1.3, red fruit:1.3, gourds:1.3, squash:1.3, bad anatomy:1.6, bad proportions:1.4, malformed body:1.6, deformed body:1.6, distorted body:1.6, mutated body:1.6, extra limbs:1.5, missing limbs:1.4, broken legs:1.5, duplicate body parts:1.5, cloned face:1.5, poorly drawn face:1.6, poorly drawn hands:1.6, malformed hands:1.6, extra fingers:1.6, missing fingers:1.6, mutated fingers:1.6, twisted fingers:1.5, bad hands:1.6, bad eyes:1.5, asymmetrical eyes:1.4, misaligned eyes:1.4, wonky eyes:1.4, blur:1.3, blurry:1.3, lowres:1.4, pixelated:1.3, artifacts:1.3, grainy:1.2, text:1.4, watermark:1.4, logo:1.4, signature:1.4, cropped head:1.5, cropped body:1.5, out of frame:1.3, ugly:1.5, disfigured:1.6, unnatural colors:1.3, worst quality:1.6, low quality:1.6)

Please comment.