metadata
tags:
- text-to-image
- lora
- diffusers
- template:diffusion-lora
widget:
- text: >-
Watercolor of a majestic dragon soaring through clouds, with its scales
shimmering in shades of blue and green
output:
url: images/ComfyUI_00022_.png
- text: >-
surreal, watercolor, painting, traditional media, a human figure with a
clock for a face, time melting and dripping around, abstract cityscape in
the background, muted earth tones blending into soft blues and purples,
delicate brushstrokes, fluid transitions, ethereal, dreamlike
output:
url: images/ComfyUI_00001_.png
- text: >-
watercolor, painting, traditional media, jack-o'-lantern, cracked, broken,
fluid transition, in to a face,
output:
url: images/ComfyUI_00014_.png
- text: surreal walercolour painting of a duck
output:
url: images/ComfyUI_00015_.png
base_model: black-forest-labs/FLUX.1-dev
instance_prompt: watercolor, watercolour
license: mit
Dreamlike Watercolour Style
- Prompt
- Watercolor of a majestic dragon soaring through clouds, with its scales shimmering in shades of blue and green
- Prompt
- surreal, watercolor, painting, traditional media, a human figure with a clock for a face, time melting and dripping around, abstract cityscape in the background, muted earth tones blending into soft blues and purples, delicate brushstrokes, fluid transitions, ethereal, dreamlike
- Prompt
- watercolor, painting, traditional media, jack-o'-lantern, cracked, broken, fluid transition, in to a face,
- Prompt
- surreal walercolour painting of a duck
Model description
Recommended model strength: 0.8 - 1.2, with the ideal strength at 1.
Other words that where common in captioning and may help refine your images: Traditional media, surreal, liquid transition, painting (while not intentional trigger words, they do get you closer to the intended result)
Example prompt structure: "Watercolor painting of a"... followed by you usual prompt
Trigger words
You should use watercolor to trigger the image generation.
You should use watercolour to trigger the image generation.
Download model
Weights for this model are available in Safetensors format.
Download them in the Files & versions tab.