mathiaslawson/web-icons

Smaple Images From Model

NB: Remeber to Add (<'outline-icon'>) at the end of your prompts [please remove the quotes and retain only the angled brackets😅]. Adding this to the end of each prompt allows you to access the facilities of the pretrained model

a icon of cat

a icon of bulb ""

a icon of bird

a icon of star

a icon of bulb a icon of lion

Web Icons

This repository contains the Web Icons model, a machine learning model for generating website icon images. The model is built using the Diffusers library and is licensed under a modified CreativeML OpenRAIL-M license. This model was fine tuned from https://huggingface.co/proximasanfinetuning/fantassified_icons_v2 with Textual Inversion

License

The Web Icons model is licensed under a modified CreativeML OpenRAIL-M license.

Usage

Here's an example of how to use the Web Icons model with the Diffusers library:

from diffusers import StableDiffusionPipeline

model_id = "mathiaslawson/web-icons"
pipe = StableDiffusionPipeline.from_pretrained(model_id)

prompt = "a icon of lion<outline-icon> "
image = pipe(prompt)["sample"][0]
image.save("dragon_icon.png")

The main changes are:

1. Added a note that the Web Icons model is pretrained on the Fantassified Icons model.
2. Updated the Acknowledgments section to credit both the Web Icons model author (Mathias Lawson) and the original Fantassified Icons model author (Proximasan).  https://huggingface.co/proximasanfinetuning/fantassified_icons_v2

Credit to goes to for base model for pretraining : https://huggingface.co/proximasanfinetuning/fantassified_icons_v2

Although its not perfect yet, the pretrained model has been able to adequately produce web-like looking icons from 3D looking icons model from  https://huggingface.co/proximasanfinetuning/fantassified_icons_v2, already showing steady capacity to produce desired results.
Contributions to the model are welcomed🙂. This is not the end, i would be improving the model till it becomes perfect for web icons.