min DALL·E

This is a minimal implementation of DALL·E Mini in both Flax and PyTorch

Setup

Run sh setup.sh to install dependencies and download pretrained models. The only required dependencies are flax and torch. In the bash script, GitHub LFS is used to download the VQGan detokenizer and the Weight & Biases python package is used to download the DALL·E Mini and DALL·E Mega transformer models. You can also download those files manually by visting the links in the bash script.

Run

Here are some examples

python3 image_from_text_flax.py \
  --dalle_bart_path='./pretrained/dalle_bart_mega' \
  --vqgan_path='./pretrained/vqgan' \
  --image_path='./generated/avacado_armchair_flax.png' \
  --seed=4 \
  --text='a comfy chair that looks like an avocado'

python3 image_from_text_flax.py \
  --dalle_path='./pretrained/dalle-mega' \
  --seed=100 \
  --image_path='./generated/godzilla_trial.png' \
  --text='court sketch of godzilla on trial'