This is an implementation of the hotshotco/Hotshot-XL as a Cog model. Cog packages machine learning models as standard containers.
Before running the image, you need to fetch the weights:
cog run python ./scripts/download_weights.py
You can then run the image with:
cog predict -i prompt="go-pro video of a polar bear diving in the ocean, 8k, HD, dslr, nature footage" -i seed=6226
"go-pro video of a polar bear diving in the ocean, 8k, HD, dslr, nature footage"