StatsGary's picture
Update README.md
64de6b0
|
raw
history blame
3.01 kB
metadata
license: creativeml-openrail-m
tags:
  - pytorch
  - diffusers
  - stable-diffusion
  - text-to-image
  - diffusion-models-class
  - dreambooth-hackathon
  - landscape
widget:
  - text: a photo of norweigen-fjords fjords

DreamBooth model for the norweigen-fjords concept trained by StatsGary on the StatsGary/dreambooth-hackathon-images dataset.

This is a Stable Diffusion model fine-tuned on the norweigen-fjords concept with DreamBooth. It can be used by modifying the instance_prompt: a viking on the fjords

This model was created as part of the DreamBooth Hackathon 🔥. Visit the organisation page for instructions on how to take part!

Description

This is a Stable Diffusion model fine-tuned on fjords images for the landscape theme. Below are some examples of the images generated on the back of the model:

Lobster swimming in a Fjord

The below example uses a prompt similar to lobster swimming in a fjord to generate the output:

lobster.jpeg

Viking warrior in a Fjord

This represents a generated Viking warrior on or near a Fjord. The prompt used to generate is prompt=a viking warrior on a fjord: viking_on_fjord.jpeg

A yellow submarine (inspired by The Beetles)

Here, we see a yellow submarine inspired by the popular Beetles album. The prompt used to generate is prompt=a beetles like yellow submarines on a fjord*: Beetles_submarine.jpeg

A cruise ship on a fjord

This is based on the prompt=a cruise ship on a fjord:

6bd7a6b7-9716-478e-81ea-7f58b59707e8.jpeg

Taj Mahal on a Fjord

This generates landmarks near or on the fjord:

68dd6b17-bb8c-45e7-bfe6-79442f633121.jpeg

Generating your own predictions

The following Python code will allow you to get up and running quickly, just replace the prompt field for your own generation, wait for HuggingFace to compute and you should have your own Stable Diffusion object generated against a backdrop of the fjords. Idyllic!

from diffusers import StableDiffusionPipeline

pipeline = StableDiffusionPipeline.from_pretrained('StatsGary/norweigen-fjords-fjords')
image = pipeline(prompt='a viking on a fjord').images[0]
image

Supporting article(s)

I have undertaken a blog to explain this: