Version 0.5.1

v0.5.1

improve: redesigned owl-vit loss

Version 0.5.0

v0.5.0

feature: gradient checkpointing

Version 0.4.0

fix!: face diffusion working and refactoring

random diffused and remove pseudo linear sampler

BREAKING CHANGE

Version 0.3.0

refactor!: rename drawer pixel to raw

BREAKING CHANGE

Version 0.2.0

refactor!: rename k-diffusion to monster_diffusion

BREAKING CHANGE

Version 0.1.1

v0.1.1

build: less stringent python requirement

Version 0.1.0

v0.1.0

build: replace github dependencies with pypi