This model is designed for use with the embedding, it has been trained using OneTrainer with pivotal tuning.This LoRA (+ embedding) was trained on a dataset of images generated with DALL-E3 and later refined with Flux.
The major feature is the overwhelming bokeh which, as the name suggests, I love!
I'm curious to see what you create with it :)
Notes
ComfyUI workflow for the samples generation should be embedded in the images, in case you want to check more settings.
IMHO it works best between 0.4 and 0.8.
As said above, it's intended for use along with this embedding, they've been trained concurrently with OneTrainer (pivotal tuning)!