A CLI tool/python module for generating images from text using guided diffusion and CLIP from OpenAI.
Full Changelog: https://github.com/afiaka87/clip-guided-diffusion/compare/v0.2.3...v0.2.4
Additional image logging to W&B. Code cleanup.
Full Changelog: https://github.com/afiaka87/clip-guided-diffusion/compare/v0.2.1...v0.2.2
Added support for some of the "quick/fast" clip guided diffusion techniques. Should help when using 100 timesteps or fewer. Saturation loss can also be used to help considerably with the 64x64 checkpoint, as well as with ddim sampling.