Changed base model to oilPainting_oilPaintingV10.safetensor.
Reduced the number of training images of beautiful women. LoRA wouldn't listen to base model and always produced a similar style beautiful woman. Training images is down to 25, but diversity is better.
Try this as a starting prompt:
"Create a spectacular mountain landscape. Use textured impasto brush strokes." <lora:add_sharpness:1> <lora:ImpastoBrush LoRa v1.0.1:0.8>
Negative prompt: "rmadanegative402_sd15-neg, blurry, out of focus, low quality, grainy, low resolution, smudged, indistinct, unclear, vague lines, (muted colors, dull colors, low contrast, grey, overcast, dark, dim lighting, shadows, darkness, fog, haze, cloudy sky, desaturated, flat lighting, monochrome, grayscale, soft focus, washed out), cartoon, photorealistic, illustration, worst quality, letterbox, picture frame, black bars, borders, edges"
Steps: 20, Sampler: DPM++ 2M, Schedule type: Karras, CFG scale: 7, Seed: 1713772848, Size: 512x768, Model hash: 9da0c2b94d, Model: oilPainting_oilPaintingV10, Denoising strength: 0.4, Hires upscale: 2, Hires upscaler: SwinIR_4x, Lora hashes: "add_sharpness: accc639dc81f, ImpastoBrush LoRa v1.0.1: d393ab654392", Version: v1.10.1
