Modelpublic

lyraaaa/flux-collage-v1

sourceHugging Faceccupdated 2y agoView on Hugging Face

1likes63downloads

Model Card

collage art style lora for flux-dev

no explicit trigger word, but you should include "collage", "multimedia collage", etc in your prompt. check the demo prompts for examples. works best at ~1.2 strength

training data captions include:

—colorful collage, mixed images, watches, faces, text, vibrant, surreal
—floral collage and train tracks, vintage piano, mixed media, surreal art, portrait elements
—eclectic collage, vintage imagery, vibrant colors, pop art, surreal elements

also makes a very interesting cluttered-realism model when used on more normal prompts, eg collage, dimly lit room, couch, TV, posters, clutter

since i do not own the data for this model, i can't really claim ownership of the model itself either. have fun!

run this with diffusers:

import torch
from diffusers import FluxPipeline
import time
import random

# initialize pipeline and lora
pipe = FluxPipeline.from_pretrained("black-forest-labs/FLUX.1-dev", torch_dtype=torch.bfloat16).to("cuda")

lora_weight = 1.2
pipe.load_lora_weights(
    'bleepybloops/flux-collage-v1',
    adapter_name='collage_v1'
)
pipe.set_adapters('collage_v1', adapter_weights=[lora_weight])

# set params and generate
seed = -1
seed = seed if seed != -1 else random.randint(0, 2**32)
print(seed)

prompt = "multimedia collage, cut and paste, detective"
out = pipe(
    prompt=prompt,
    guidance_scale=3.5,
    height=720,
    width=1024,
    num_inference_steps=25,
    generator=torch.Generator("cuda").manual_seed(seed),
).images[0]

# save and display output
filename=f"{time.time()}.png"
out.save(filename)

from IPython.display import Image, display
display(Image(filename=filename))

trained with https://github.com/ostris/ai-toolkit/blob/main/notebooks/FLUX1devLoRATraining.ipynb