Fantassified Icons V2

🚀 New and Shiny ｡･:*:･ﾟ’☆

This is an updated text-to-image model based on StableDiffusion, which can generate fantasy - style icons. It's a finetuned version, offering better performance compared to its previous iteration.

🚀 Quick Start

About

• This is an updated version of v1. It was created using a dataset that mostly consists of the old version's data. However, it's much better as the developer has learned a lot since the dreambooth days.
• It generates icons inspired by fantasy games with mostly plain backgrounds.
• No trigger words are required.
• There's an issue where either the local hires fix isn't working well, or potions look strange when hires is enabled. This needs further testing (possibly requiring low denoising strength).
• This model is not recommended for generating people and faces as they were not a concern during training. The focus was on items, but users can experiment on their own.
• The examples were made with this VAE at 20 steps, 512x512, CFG 7, Euler a (try DPM ++2M for a sharper look).

Examples

If you enjoy this, consider buying the developer a coffee (ノ◕ヮ◕)ノ*:・゚✧

💻 Usage Examples

Basic Usage

import torch
from diffusers import StableDiffusionPipeline, DDIMScheduler

scheduler =  DDIMScheduler.from_pretrained("proximasanfinetuning/fantassified_icons_v2", subfolder="scheduler")
pipe = StableDiffusionPipeline.from_pretrained("proximasanfinetuning/fantassified_icons_v2", scheduler=scheduler).to("cuda")
prompt = "A lemon themed high quality hamburger"
images = pipe(prompt, num_images_per_prompt=6, num_inference_steps=25).images
images[0]

📄 License

This model is licensed under a modified CreativeML OpenRAIL - M license.

• Utilizing and hosting the Fantassified Icons 1.0 model and its derivatives on platforms that earn, will earn, or plan to earn revenue or donations requires prior authorization. To request permission, please email proximasan@protonmail.com.
• You are permitted to host the model card and files on both commercial and non - commercial websites, apps, etc. as long as you properly credit the model by stating its full name and providing a link to the model card (https://huggingface.co/proximasanfinetuning/fantassified_icons_v2), without performing any actual inference or finetuning.
• The Fantassified Icons 1.0 model and its derivatives can be hosted on non - commercial websites, apps, etc. as long as no revenue or donations are received. Proper credit must be given by stating the full model name and including a link to the model card (https://huggingface.co/proximasanfinetuning/fantassified_icons_v2).
• The outputs of the model or its derivatives can be used for commercial purposes as long as the usage is limited to teams of 10 or fewer individuals.
• You can't use the model to deliberately produce nor share illegal or harmful outputs or content.
• The authors claim no rights on the outputs you generate. You are free to use them and are accountable for their use which must not go against the provisions set in the license.
• You may re - distribute the weights. If you do, please be aware you have to include the same use restrictions as the ones in the license and share a copy of the modified CreativeML OpenRAIL - M to all your users (please read the license entirely and carefully). Please read the full license here: https://huggingface.co/proximasanfinetuning/fantassified_icons_v2/blob/main/license.txt