The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.
Commits on master
Add a node to save animated PNG files. These work in ffpmeg unlike webp.
eff24ea comfyanonymous committed 2 years ago
Add to Readme that Stable Video Diffusion is supported.
c782cf3 comfyanonymous committed 2 years ago
Fix SaveAnimatedWEBP not working when metadata is disabled.
022033a comfyanonymous committed 2 years ago
Add sampling_settings so models can specify specific sampling settings.
32447f0 comfyanonymous committed 2 years ago
Merge branch 'hide-if-collapsed' of https://github.com/pythongosssss/ComfyUI
1ca4802 comfyanonymous committed 3 years ago
Allow controlling downscale and upscale methods in PatchModelAddDownscale.
c3ae99a comfyanonymous committed 3 years ago
Allow model config to preprocess the vae state dict on load.
6a491eb comfyanonymous committed 3 years ago
Merge branch 'fix-collapsed-clip' of https://github.com/pythongosssss/ComfyUI
d66b631 comfyanonymous committed 3 years ago
Make it easy for models to process the unet state dict on load.
ce67dcb comfyanonymous committed 3 years ago