Skip to main content

CLI for Stable Diffusion tasks.

Project description

PyPI-Server Monthly Downloads Test Coveralls License MIT Project generated with PyScaffold

CLI for Stable Diffusion tasks.

Read more: https://artificialhoney.github.io/giger

Installation

pip install giger git+https://github.com/ai-forever/Real-ESRGAN.git

Usage

Please check first of all the help function and the examples.

giger --help

Also make sure to always obtain the latest version.

giger --version

Increase verbosity to also output the diffusers logging.

giger -vv

template

Use a jinja2 template file and supply data from file. Overwrite variables and print out to console.

giger template --config hero=viking --data hero.yml "$(cat hero.txt.j2)"

Use an inline jinja2 template file and supply data. Write out to file.

giger template --config hero=viking "A {{hero}} with long hair and sword" --output viking.txt

You can also pipe from another command to the template task.

echo "A {{hero}} with long hair and sword" | giger template --config hero=viking  --output viking.txt

prompt

Generate a prompt with multiple well-known input keywords to choose.

giger prompt "A viking with long hair and sword" --time "Ancient" --type "Comic Book" --art_style "Concept art" --realism "Photorealistic" --rendering_engine "Octane render" --lightning_style "Cinematic" --camera_position "Ultra-Wide-Angle Shot" --resolution "8k"

You can also pipe from another command to the prompt task.

echo "A viking with long hair and sword" | giger prompt --time "Ancient" --type "Comic Book" --art_style "Concept art" --realism "Photorealistic" --rendering_engine "Octane render" --lightning_style "Cinematic" --camera_position "Ultra-Wide-Angle Shot" --resolution "8k"

image

The commands pull the chosen model from huggingface.co. You choose one with the --model option. Also the batch and image sizes can be configured and one can pass the prompt via pipe.

Please see the help function for more information.

txt2img

giger image "A viking with long hair and sword, Concept art, Photorealistic, Octane render, Cinematic, Ultra-Wide-Angle Shot, 8k" --output $HOME/Desktop/ --name viking

img2img

giger image "A viking with long hair and sword, Concept art, Photorealistic, Octane render, Cinematic, Ultra-Wide-Angle Shot, 8k" --output $HOME/Desktop/ --name viking --input input.png

controlnet

giger image "A viking with long hair and sword, Concept art, Photorealistic, Octane render, Cinematic, Ultra-Wide-Angle Shot, 8k" --output $HOME/Desktop/ --name viking --input input.png --controlnet_model "lllyasviel/sd-controlnet-hed"

roop

Simply change the face in an input image and render the result to disc.

giger roop --face face.jpg --input target.png --output output.png

upscale

Simply upscale an image and render the result to disc.

giger upscale --input image.png --output image@4x.png --scale 4

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

giger-0.2.1.tar.gz (183.4 kB view hashes)

Uploaded Source

Built Distribution

giger-0.2.1-py3-none-any.whl (16.4 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page