metadata

license: creativeml-openrail-m
tags:
  - pytorch
  - diffusers
  - stable-diffusion
  - text-to-image
  - diffusion-models-class
  - dreambooth-hackathon
  - wildcard
datasets:
  - TrpFrog/trpfrog-icons
widget:
  - text: a photo of trpfrog
    output:
      url: output3.png
  - text: a photo of trpfrog in spring
    output:
      url: spring.png
  - text: a photo of trpfrog in summer
    output:
      url: summer.png
  - text: a photo of trpfrog in autumn
    output:
      url: autumn.png
  - text: a photo of trpfrog in winter
    output:
      url: winter.png
  - text: a photo of trpfrog in fire
    output:
      url: output.png
  - text: a photo of trpfrog growing shiitake mushrooms
    output:
      url: output2.png

TrpFrog Diffusion

Prompt
a photo of trpfrog growing shiitake mushrooms

This is a Stable Diffusion model fine-tuned on the trpfrog concept with DreamBooth. It can be used by modifying the instance_prompt: a photo of trpfrog

Description

TrpFrog (@Trpfrog) is a famous Japanese Twitter comedian.

Below is a screenshot of the site he is creating.

This is a fine-tuned model of Stable Diffusion with the powerful help of Dreambooth. The model is made specifically for TrpFrog and faithfully reproduces its appearance.

Usage

First install the necessary packages.

pip install diffusers transformers accelerate scipy safetensors

The image can then be generated with the following script.

import torch
from diffusers import StableDiffusionPipeline

keyword = "trpfrog"
prompt = f"a photo of {keyword}"

model_id = "Prgckwb/trpfrog-diffusion"

pipe = StableDiffusionPipeline.from_pretrained(
    model_id, torch_dtype=torch.float16
).to("cuda")

image = pipe(prompt).images[0]
image.save("trpfrog.jpg")