metadata
tags:
- text-to-image
- stable-diffusion
- lora
- diffusers
- template:sd-lora
widget:
- text: a bottle of water with the text "clean" on it, flmft photo style
output:
url: images/ComfyUI_00240_.png
- text: a fluffy white kitten, flmft photo style
output:
url: images/ComfyUI_00236_.png
- text: a man, flmft photo style
output:
url: images/ComfyUI_00230_.png
- text: a gray day with mt fuji in the distance, flmft photo style
output:
url: images/ComfyUI_00242_ (1).png
- text: a celestial being, flmft photo style
output:
url: images/ComfyUI_00244_.png
- text: a japanese woman with a bit smile, flmft photo style
output:
url: images/ComfyUI_00247_.png
- text: >-
a messy desk with an pile of old cassette tapes from different indie bands
with scrawled labels on them, still life, flmft photo style
output:
url: images/ComfyUI_00253_ (1).png
base_model: black-forest-labs/FLUX.1-dev
instance_prompt: flmft photo style
license: creativeml-openrail-m
Here is my first crack at a realism Flux model, and I guess the first realism model I've ever shared publicly.
All the training data was opensource/open license photographs. I do intend to revisit it and expand and improve on it. It can benefit from emphasizing the style by adding "vintage" or "faded film" to the prompt.
Weights for this model are available in Safetensors format.