Go to file
2022-07-01 18:50:11 -04:00
examples update examples 2022-07-01 14:28:51 -04:00
min_dalle update readme, random seed if none specified 2022-07-01 18:50:11 -04:00
replicate moved flax model and conversion code to separate repository 2022-07-01 14:06:50 -04:00
.gitattributes add gitattributes file 2022-06-29 12:45:41 -04:00
.gitignore added to pypi 2022-07-01 18:16:55 -04:00
image_from_text.py added to pypi 2022-07-01 18:16:55 -04:00
LICENSE license and cleanup 2022-06-27 14:34:10 -04:00
min_dalle.ipynb added to pypi 2022-07-01 18:16:55 -04:00
README.md update readme, random seed if none specified 2022-07-01 18:50:11 -04:00
requirements.txt added to pypi 2022-07-01 18:16:55 -04:00
setup.py added to pypi 2022-07-01 18:16:55 -04:00

min(DALL·E)

Open In Colab   Replicate

This is a minimal implementation of Boris Dayma's DALL·E Mini in PyTorch. It has been stripped to the bare essentials necessary for doing inference. The only third party dependencies are numpy and torch.

It currently takes 7.4 seconds to generate an image with DALL·E Mega with PyTorch on a standard GPU runtime in Colab

The flax model, and the code for coverting it to torch, have been moved here.

Install

$ pip install min-dalle

Usage

Use the python script image_from_text.py to generate images from the command line.

$ python image_from_text.py --text='artificial intelligence' --seed=7

Artificial Intelligence

$ python image_from_text.py --text='court sketch of godzilla on trial' --mega

Godzilla Trial

To load a model once and generate multiple times, initialize MinDalleTorch, then call generate_image with some text and a seed.

from min_dalle import MinDalleTorch

model = MinDalleTorch(
    is_mega=True, 
    is_reusable=True,
    models_root='./pretrained'
)
image = model.generate_image("a comfy chair that looks like an avocado")
display(image)
image = model.generate_image("trail cam footage of gollum eating watermelon", seed=1)
display(image)

Gollum Trailcam

Model parameters will be downloaded as needed to the directory specified. The models can also be manually downloaded here.