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

Add updater test to stable release workflow. (#7887)
comfyanonymous committed 1 year ago
39c27a3
Test updater in the windows release workflow. (#7886)
comfyanonymous committed 1 year ago
b1c7291
Better vace memory estimation. (#7875)
comfyanonymous committed 1 year ago
dbc726f
Bump minimum pyav version to 14.2.0 (#7874)
comfyanonymous committed 1 year ago
7ee9645
Per device stream counters for async offload. (#7873)
comfyanonymous committed 1 year ago
0a66d4b
support more example folders (#7836)
Terry Jia committed 1 year ago
5c5457a
Add release process section to README (#7855)
Chenlei Hu committed 1 year ago
45503f6
005a91c
68f0d35
Support HiDream E1 model. (#7857)
comfyanonymous committed 1 year ago
83d0471
CFG++ for gradient estimation sampler (#7809)
chaObserv committed 1 year ago
c15909b
772b4c5
Fix stream priority to support older pytorch. (#7856)
comfyanonymous committed 1 year ago
5a50c3c
Save v pred zsnr metadata (#7840)
Pam committed 1 year ago
30159a7
ltxv: add strength parameter to conditioning. (#7849)
Andrew Kvochko committed 1 year ago
cb9ac3d
Add `/api/v2/userdata` endpoint (#7817)
Benjamin Lu committed 1 year ago
8115a78
Use stream for casting if enabled. (#7833)
comfyanonymous committed 1 year ago
c8cd7ad
542b4b3
Make loras work with --async-offload (#7824)
comfyanonymous committed 1 year ago
ac10a0d
Update portable package workflow to cu128 (#7812)
comfyanonymous committed 1 year ago
b685b8a
78992c4
Support SimpleTuner lycoris lora format for HiDream.
comfyanonymous committed 1 year ago
f935d42
ComfyUI version 0.3.30
comfyanonymous committed 1 year ago
a97f2f8
Switch LTXVPreprocess to libx264 (#7776)
comfyanonymous committed 1 year ago
5acb705
fix multiple image return from api nodes (#7772)
thot experiment committed 1 year ago
5c80da3
e2eed9e