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 memory estimation function to ltxav text encoder. (#11716)
25bc1b5 comfyanonymous committed 10 months ago
Add device selection for LTXAVTextEncoderLoader (#11700)
1c705f7 Jukka Seppänen committed 10 months ago
Fix stable release workflow not pulling latest comfy kitchen. (#11695)
c0c9720 comfyanonymous committed 10 months ago
feat(api-nodes): add WAN2.6 ReferenceToVideo (#11644)
79e9454 Alexander Piskun committed 10 months ago
Disable comfy kitchen cuda if pytorch cuda less than 13 (#11681)
edee33f comfyanonymous committed 10 months ago
Skip fp4 matrix mult on devices that don't support it. (#11677)
2c03884 comfyanonymous committed 10 months ago
Revert "Use rope functions from comfy kitchen. (#11647)" (#11648)
1618002 comfyanonymous committed 10 months ago
Initial ops changes to use comfy_kitchen: Initial nvfp4 checkpoint support. (#11635)
6da00dd comfyanonymous committed 10 months ago