Launchpad/Lunchpad

feat: include examples

fix: move to cuda before generate

fix: map to cpu

fix: no map location in main process

fix: move to cuda in generate function

feat: ZeroGPU implementation

fix: image transform order

fix: more imports + loading files

fix: model import

fix: masked fill and imports

change colors

adjust README

add requirements + gradio demo

git attributes

Merge pull request #12 from amaiasalvador/master

Update README.md with correct pytorch version

Update requirements.txt with torch and torchvision versions

Merge pull request #7 from amaiasalvador/master

ingredient encoder parameters removed when args.ingrs_only = True

transformer no longer returns unnecessary attention weights. fix: allow backward when training ingredient decoder

fix: removed perminv_ingrs deprecated argument. Improved instructions.

Initial commit