HuggingFace-based Stable Diffusion Generator

• Generates a single stable diffusion image with an input prompt.
• Requires a HuggingFace account and a token.

Sample generated with v1.3 with minor edits on Photoshop and background extended with DALL-E:

Requirements

1. python 3.9+
2. HuggingFace account

How to run

1. Install the requirements

pip install -r requirements.txt

2. Accept the agreement in the Stable Diffusion v1.4 model repo to gain access to the model.

3. Create and retrieve a READ token from HuggingFace Settings

4. Login with your token in the command below

huggingface-cli login

5. You can now start generating with:

python generate.py <prompt>

ex. - python generate.py "astronaut floating in space while spaceship drifts away"

Prompt Suffixes

You can use these prompt suffixes to generate decent-looking outputs.

Places

Add your description of a place at the start of the prompt.

1. "[description of place], tone mapped, shiny, intricate, cinematic lighting, highly detailed, digital painting, artstation, concept art, smooth, sharp focus, illustration, art by terry moore and greg rutkowski and alphonse mucha" (credit: @jquesnelle)

Humans

Add your description of a human at the start of the prompt.

1. "[description of human] Clean Cel shaded vector art. Shutterstock. behance hd by lois van baarle, artgerm, Helen huang, by makoto shinkai and ilya kuvshinov, rossdraws, illustration" (credit: The_Bravinator on Reddit)