FHDR Uncensored
by kpsss34
Open source · 248k downloads · 389 likes
3.2
(389 reviews)ImageAPI & LocalAbout
The FHDR Uncensored model is an AI-powered solution specialized in enhancing and generating high-definition images, with an optimized approach to reduce training times while preserving quality. It excels at restoring and improving blurry, underexposed, or low-resolution images, producing detailed and natural results. Its core capabilities include color correction, detail sharpening, and contrast adjustment, making it ideal for photographers, digital artists, or content creators. What sets it apart is its targeted training method, which focuses on specific layers of the network for superior output without sacrificing performance. It is particularly well-suited for projects requiring realistic and impactful images, even from low-quality sources.
Documentation
I trained this model using the Diffusers library by randomly selecting layers and blocks (not training every layer), which reduced the training time and is expected to yield better results.
Python
import torch
from diffusers import FluxPipeline
pipe = FluxPipeline.from_pretrained("kpsss34/FHDR_Uncensored", torch_dtype=torch.bfloat16)
pipe.enable_model_cpu_offload()
prompt = "a women..."
image = pipe(
prompt,
height=1024,
width=1024,
guidance_scale=4.0,
num_inference_steps=40,
max_sequence_length=512,
generator=torch.Generator("cpu").manual_seed(0)
).images[0]
image.save("outputs.png")
Capabilities & Tags
diffuserssafetensorsggufarttext-to-imageenendpoints_compatible
Links & Resources