goodguy1963 / good-sdxl-models-plus-loras

BROKEN - DO NOT USE!

  • Public
  • 187 runs
  • Weights
Run with an API
Playground API Examples README Versions

Examples

View more

Run time and cost

This model costs approximately $0.0079 to run on Replicate, or 126 runs per $1, but this varies depending on your inputs. It is also open source and you can run it on your own computer with Docker.

This model runs on Nvidia L40S GPU hardware. Predictions typically complete within 9 seconds. The predict time for this model varies significantly based on the inputs.

Readme

SDXL Model Collection for Replicate

This repository contains a collection of high-quality SDXL models and LoRAs for image generation, optimized for deployment on Replicate.

Models Included

  • RealVisXL V5.0 by SG_161222
  • Real Dream (SDXL 5) by sinatra
  • AlbedoBase XL (3.1-Large)
  • Realism Engine SDXL by razzz
  • CyberRealistic XL by Cyberdelia

LoRAs Included

  • Detail Tweaker XL (1.0) by w4r10ck
  • Extremely Detailed (no trigger) by ntc
  • DetailedEyes XL (3.0) by bdsqlsz

Model Sources

All models and LoRAs are hosted on Hugging Face at: https://huggingface.co/hofixD/good-sdxl-models-plus-loras

Setup and Usage

The models and LoRAs will be automatically downloaded when building the container or can be manually downloaded using:

python script/download-weights

System Requirements

  • CUDA-capable GPU with at least 12GB VRAM
  • Python 3.11
  • See requirements.txt for Python package dependencies

Features

  • Text-to-Image Generation
  • Image-to-Image Processing
  • Inpainting
  • Multiple Sampling Methods
  • LoRA Support
  • Resolution Control
  • Seed Control for Reproducibility

Parameters

  • prompt: Input prompt for image generation
  • negative_prompt: Things to avoid in the generated image
  • image: Input image for img2img or inpainting
  • mask: Mask image for inpainting
  • width: Output image width (128-2048)
  • height: Output image height (128-2048)
  • num_inference_steps: Number of denoising steps (1-500)
  • guidance_scale: Scale for classifier-free guidance (1-20)
  • prompt_strength: Strength of prompt when using init image (0-1)
  • model_name: Choose from available SDXL models
  • lora_selection: Available LoRA options
  • lora_weight: LoRA effect strength (0.5-3.0)
  • sampler: Choice of different sampling methods
  • scheduler_type: Type of scheduler to use
  • seed: Random seed for reproducibility

Credits

Thanks to all model creators for their amazing work: - SG_161222 (RealVisXL) - sinatra (Real Dream) - AlbedoBase Team - razzz (Realism Engine) - Cyberdelia (CyberRealistic) - w4r10ck (Detail Tweaker XL) - ntc/sliders.ntcai.xyz (Extremely Detailed) - bdsqlsz (DetailedEyes)