metadata

tags:
  - text-to-image
  - image-to-image
  - stable-diffusion
  - lora
  - diffusers
  - template:sd-lora
widget:
  - text: graphis-01 cityscape
    parameters:
      negative_prompt: text, watermark, ugly, messy, bad, photo
    output:
      url: images/graphis-01-25.jpg
  - text: graphis-01 cityscape at sunset, girl, cyberpunk, synthwave
    parameters:
      negative_prompt: text, watermark, ugly, messy, bad, photo
    output:
      url: images/graphis-01-10.jpg
  - text: graphis-01, bubble tea
    parameters:
      negative_prompt: text, watermark, ugly, messy, bad, photo
    output:
      url: images/graphis-01-01.jpg
  - text: graphis-01 girl, cyberpunk, synthwave
    parameters:
      negative_prompt: text, watermark, ugly, messy, bad, photo
    output:
      url: images/graphis-01-12.jpg
  - text: graphis-01 flower
    parameters:
      negative_prompt: text, watermark, ugly, messy, bad, photo
    output:
      url: images/graphis-01-19.jpg
  - text: graphis-01 golden gate bridge
    parameters:
      negative_prompt: text, watermark, ugly, messy, bad, photo
    output:
      url: images/graphis-01-20.jpg
base_model: stabilityai/stable-diffusion-xl-base-1.0
instance_prompt: graphis-01
license: creativeml-openrail-m

Graphis-01

Prompt
graphis-01, bubble tea

Negative Prompt
text, watermark, ugly, messy, bad, photo

Prompt
graphis-01 girl, cyberpunk, synthwave

Negative Prompt
text, watermark, ugly, messy, bad, photo

Prompt
graphis-01 flower

Negative Prompt
text, watermark, ugly, messy, bad, photo

Prompt
graphis-01 golden gate bridge

Negative Prompt
text, watermark, ugly, messy, bad, photo

Model description

This fun LoRA model generates stylized graphics with a flat, textured look featuring smooth, grainy gradients. It was trained on a custom dataset of my own digital illustrations showcasing graphic shapes on various paper textures. For optimal results, use the trigger word "graphis-01" in your prompts. Experiment with different LoRA weightings to achieve various stylistic effects. Recommended weightings range from 0.6 to 1.6 (when using ComfyUI). For best results, pair this model with an artistic base model such as Zavy Chroma XL. Image-to-image generations yeiled some very interesting results as well. Have fun!

Trigger words

You should use graphis-01 to trigger the image generation.

Download model

Weights for this model are available in Safetensors format.

Download them in the Files & versions tab.