metadata

language:
  - en
license: cc-by-nc-nd-4.0
library_name: diffusers
tags:
  - art
  - people
  - diffusion
  - Cinematic
  - Photography
  - Landscape
  - Interior
  - Food
  - Car
  - Wildlife
  - Architecture
  - Neuron
  - Inferentia
thumbnail: >-
  https://storage.googleapis.com/run-diffusion-public-assets/juggernaut-xi/juggernaut-collage-256.webp
base_model: stabilityai/stable-diffusion-xl-base-1.0
pipeline_tag: text-to-image

Juggernaut XI v11 by RunDiffusion (Official) - Neuron

Amazing prompt adherence ✅
Massively improved aesthetics
Better hands, eyes, faces, and compostion
Fully trained from the ground up using the GPT4 Vision Captioning tool by LEOSAM 🛠️
Expanded and cleaner dataset with higher quality images 🖼️
Improved classifications of shots (Full Body, Midshots, Portraits, etc) 📸
Enhanced text generation capability 📝
Two different prompting techniques, Natural and Tagging style 🏷️
Enhanced by RunDiffusion Photo for refinement of details 🧐

Read more about this version here https://rundiffusion.com/juggernaut-xi

Prompting Guide 📘 Because everything has been trained from the ground up, prompting is a bit different. (Simpler, don't worry) @Kandoo has created a guide to help you seamlessly integrate this powerful model into your workflow, enabling you to leverage its advanced capabilities without feeling overwhelmed. Download it here: https://rundiffusion.com/juggernaut-xl#nav

Don't forget to follow RunDiffusion team on Twitter.

https://x.com/RunDiffusion

Usage

from diffusers import DPMSolverMultistepScheduler
from optimum.neuron import NeuronStableDiffusionXLPipeline

pipeline = NeuronStableDiffusionXLPipeline.from_pretrained("Shekswess/Juggernaut-XI-v11-Neuron", device_ids=[0, 1])
pipeline.scheduler = DPMSolverMultistepScheduler.from_config(pipeline.scheduler.config)

prompt = "A swirling beautiful exploding scene of magical wonders and surreal ideas and objects with portraits of beautiful woman with silk back to camera, flowers, light, cosmic wonder, nebula, high-resolution"
negative_prompt = "fake eyes, deformed eyes, bad eyes, cgi, 3D, digital, airbrushed, hands, hand"
image = pipeline(prompt=prompt, negative_prompt=negative_prompt).images[0].save("output.png")

Original Model

Model

Precision

BFloat16 (bf16) For Matrix Multiplication Operations.