Skip to content

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

chore: update workflow templates to v0.8.42 (#12491)
ComfyUI Wiki committed 9 months ago
1978f59
Remove workaround for old pytorch. (#12480)
comfyanonymous committed 9 months ago
88e6370
feat(api-nodes): add Bria RMBG nodes (#12465)
Alexander Piskun committed 9 months ago
2c1d06a
Node Replacement API (#12014)
Jedrzej Kosinski committed 9 months ago
596ed68
chore(api-nodes): remove "gpt-4o" model (#12467)
Alexander Piskun committed 9 months ago
ce4a1ab
Update frontend package to 1.38.14 (#12469)
Christian Byrne committed 9 months ago
df1e5e8
Add left padding to LTXAV text encoder. (#12456)
comfyanonymous committed 9 months ago
712efb4
Fix some custom nodes. (#12455)
comfyanonymous committed 9 months ago
726af73
dynamic_vram: Training fixes (#12442)
rattus committed 9 months ago
8902907
e03fe8b
Add a tip for common error. (#12414)
comfyanonymous committed 9 months ago
66c1852
Update ace15.py to allow min_p sampling (#12373)
askmyteapot committed 9 months ago
e5ae670
3fe61ce
2a4328d
[API Nodes] enable Magnific Upscalers (#12179)
Alexander Piskun committed 9 months ago
2b7cc7e
Make built in lora training work on anima. (#12402)
comfyanonymous committed 9 months ago
76a7fa9
[Trainer] training with proper offloading (#12189)
Kohaku-Blueleaf committed 9 months ago
cdcf411