idlebg's picture
Update README.md
988b20d
metadata
license: other
base_model: diffusers/stable-diffusion-xl-base-0.9
instance_prompt: a 3d island
tags:
  - stable-diffusion-xl
  - stable-diffusion-xl-diffusers
  - text-to-image
  - diffusers
  - lora
inference: true
library_name: diffusers
badges:
  - alt: Name
    url: >-
      https://img.shields.io/badge/Name-FFusion%20XL%20LoRA%20%F0%9F%8F%9D%EF%B8%8F%20%20Island%20Generator-89CFF0
    src: >-
      https://img.shields.io/badge/Name-FFusion%20XL%20LoRA%20%F0%9F%8F%9D%EF%B8%8F%20%20Island%20Generator-89CFF0
  - alt: LoRA Type
    url: https://img.shields.io/badge/LoRA%20Type-LyCORIS%2FLoKr%2C%20Prodigy-blue
    src: https://img.shields.io/badge/LoRA%20Type-LyCORIS%2FLoKr%2C%20Prodigy-blue
  - alt: Refiner Compatible
    url: >-
      https://img.shields.io/badge/%F0%9F%94%A5%20Refiner%20Compatible-Yes-success
    src: >-
      https://img.shields.io/badge/%F0%9F%94%A5%20Refiner%20Compatible-Yes-success
  - alt: CLIP Tested
    url: >-
      https://img.shields.io/badge/%F0%9F%92%BB%20CLIP--ViT%2FG%20and%20CLIP--ViT%2FL%20tested-Yes-success
    src: >-
      https://img.shields.io/badge/%F0%9F%92%BB%20CLIP--ViT%2FG%20and%20CLIP--ViT%2FL%20tested-Yes-success
  - alt: Trained Resolution
    url: >-
      https://img.shields.io/badge/Trained%20Resolution-1024%20x%201024%20pixels-yellow
    src: >-
      https://img.shields.io/badge/Trained%20Resolution-1024%20x%201024%20pixels-yellow
  - alt: Training Data
    url: https://img.shields.io/badge/Training%20Data-3%20x%203000%20images-orange
    src: https://img.shields.io/badge/Training%20Data-3%20x%203000%20images-orange
  - alt: Tested Resolution
    url: >-
      https://img.shields.io/badge/Tested%20Resolution-Up%20to%201600%20pixels-brightgreen
    src: >-
      https://img.shields.io/badge/Tested%20Resolution-Up%20to%201600%20pixels-brightgreen
  - alt: Tested on
    url: >-
      https://img.shields.io/badge/Tested%20on-SDXL%200.9%20%26%20FFXL%200.001-blue
    src: >-
      https://img.shields.io/badge/Tested%20on-SDXL%200.9%20%26%20FFXL%200.001-blue
  - alt: Hugging Face Model
    url: https://img.shields.io/badge/Hugging%20Face-FFusion--BaSE-blue
    src: https://img.shields.io/badge/Hugging%20Face-FFusion--BaSE-blue
  - alt: GitHub
    url: https://img.shields.io/badge/GitHub-1e--2-green
    src: https://img.shields.io/badge/GitHub-1e--2-green
  - alt: Facebook
    url: https://img.shields.io/badge/Facebook-FFusionAI-blue
    src: https://img.shields.io/badge/Facebook-FFusionAI-blue
  - alt: Civitai
    url: https://img.shields.io/badge/Civitai-FFusionAI-blue
    src: https://img.shields.io/badge/Civitai-FFusionAI-blue
language:
  - en

FFusion XL LoRA 🏝️Island Generator

The FFusion XL LoRA Island Generator is a model designed to generate game assets like islands and objects in low polygonal landscapes, pixelated, 3D, and isometric styles, making it ideal for retro-inspired or stylized game environments.

Specifications

  • Model Name: FFusion XL LoRA Island Generator
  • LoRA Type: LyCORIS/LoKr, Prodigy
  • Trained Resolution: 1024 x 1024 pixels
  • Tested Resolution: Up to 1600 pixels
  • Training Data: The model was trained on 9,000 images, consisting of 3 different resumes each on 3,000 images, providing a diverse dataset for learning.

image.png

Refiner Example

idle-FF_00866_.png ComfyUI_00258_.png

Usage Recommendations

The FFusion XL LoRA Island Generator can be utilized to quickly create game assets for a variety of game projects. It is best suited for applications where a retro or pixelated style is desired, and where low polygonal landscapes and 3D elements are prominent. Designers and developers can leverage the model to streamline the asset creation process, saving valuable time and resources.

Limitations

  • The model's performance may vary when generating assets beyond the tested resolution of 1600 pixels. Using it for higher resolutions may lead to a decrease in output quality or increased generation time.
  • The model's ability to produce highly detailed or realistic assets might be limited due to its pixelated and low polygonal nature.
  • It is recommended to thoroughly evaluate the generated assets to ensure they align with the desired artistic direction and quality standards of the specific game project.

Ethical Considerations

As with any AI model, ethical considerations should be taken into account during usage. It is essential to avoid generating assets that might infringe upon copyrighted material or violate intellectual property rights. Additionally, the model's outputs should be monitored to ensure it does not produce inappropriate, offensive, or harmful content.

Citations

If you use the FFusion XL LoRA Island Generator in your project or research, please provide appropriate citations to acknowledge the model's contribution.

Disclaimer

The FFusion XL LoRA Island Generator is a powerful tool for generating game assets; however, it may not be perfect and might have limitations. Users are encouraged to test and validate the generated assets thoroughly before integrating them into their game projects. The developers of this model hold no responsibility for any consequences that may arise from its usage.

These are LoRA adaption weights for

stable-diffusion-xl-base-0.9

&

FFusionXL-09-SDXL

The weights were trained using experimental

kohya-ss/sd-scripts build

build

img_1 img_2 img_3 img_4 img_5 img_6 img_7

Hugging Face Model GitHub Facebook Civitai

These are LoRA adaption weights for

stable-diffusion-xl-base-0.9

&

FFusionXL-09-SDXL

The weights were trained using experimental

kohya-ss/sd-scripts build

build

Attribution: "SDXL 0.9 is licensed under the SDXL Research License, Copyright (c) Stability AI Ltd. All Rights Reserved."

License

SDXL 0.9 Research License