metadata
license: mit
library_name: diffusers
base_model:
- black-forest-labs/FLUX.1-dev
tags:
- uncensored
flux-uncensored
Summary
Flux base model merged with uncensored LoRA. This model is not for those looking for "safe" or watered-down outputs. It’s optimized for real-world use with fewer constraints.
Specs
- Model: Flux base
- LoRA: Uncensored version, merged directly
Usage
pretty straight forward plug-and-play model
from diffusers import FluxPipeline
import torch
pipe = FluxPipeline.from_pretrained("shauray/FLUX-UNCENSORED-merged", torch_dtype=torch.float16)
pipe.enable_model_cpu_offload()
prompt = "A mystic cat with a sign that says hello world!"
image = pipe(prompt, guidance_scale=3.5, num_inference_steps=28, generator=torch.manual_seed(0)).images[0]
image.save("flux-dev-loaded.png")
this README has what you'd need, it's a merge from Uncensored LoRA on CivitAI