【模型介绍】模型采用200+镭射风格渐变的背景图训练而成，适合平面设计，海报，发布会，壁纸及PPT设计的背景图生成。【模型使用参数】大模型：麦橘写实/其他真实系模型VAE：vae-ft-mse-840000-ema-pruned迭代步数：18-25步（步数越高，弥散效果越明显）采样方法：Eular-a推荐分辨率：768*1024，横竖构图均可实现lora效果正向提示词：leishe, gradient, gradient background, colorful, no humans反向提示词：1girl,1boy,long hair, (worst quality, low quality:2),(normal quality:1.5),monochrome,((watermark:1.4)),artist name,character name,text,bad anatomy,sketch

Graphic Design平面设计-背景图(镭射渐变)V1.0

leishe.safetensors

v1.0

Graphic Design - Background (Laser Gradient) V1.0

[Model introduction] The model was trained using 200+ laser-style gradient background maps, suitable for generating background images for graphic design, posters, press conferences, wallpapers, and PPT designs. [Model usage parameters] Large model: realistic/other real system models VAE: vae-ft-mse-840000-ema-pruned number of iteration steps: 18-25 steps (the higher the number of steps, the more obvious the dispersion effect) Sampling method: eular-a Recommended resolution: 768*1024, both horizontal and vertical compositions can achieve Lora effects. Positive hint: leishe, gradient, gradient background, colorful, no humans reverse hint: 1girl,1boy, long hair, (worst quality, low quality: 2), (normal quality: 1.5), monochrome , (watermark: 1.4)), artist name, character name, text, bad anatomy, sketch

Files with SHA256: a8347a92eea96983ea3a5bee78cc10f9af342b60753c17bb42ce5d595ce53861