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 basic style model "multiply" strength.
comfyanonymous committed 1 year ago
3fc6ebc
26fb2c6
Fast previews for ltxv.
comfyanonymous committed 1 year ago
bf2650a
Update web content to release v1.4.13 (#5807)
Chenlei Hu committed 1 year ago
53646e0
Remove internal model download endpoint (#5432)
Chenlei Hu committed 1 year ago
20879c7
Remove print.
comfyanonymous committed 1 year ago
b666539
Missing parentheses.
comfyanonymous committed 1 year ago
95d8713
LTXV model merging node.
comfyanonymous committed 1 year ago
0d4e29f
Alternative fix for #5767
comfyanonymous committed 1 year ago
497db62
fix multi add makedirs error (#5786)
lky committed 1 year ago
24dc581
Support official SD3.5 Controlnets.
comfyanonymous committed 1 year ago
4c82741
Support for the official mochi lora format.
comfyanonymous committed 1 year ago
15c39ea
Flux inpaint model does not work in fp16.
comfyanonymous committed 1 year ago
b7143b7
61196d8
Skip layer guidance now works on hydit model.
comfyanonymous committed 1 year ago
b4526d3
Update README.md (#5707)
40476 committed 1 year ago
3d80271
set LTX min length to 1 for t2i (#5750)
spacepxl committed 1 year ago
7126ecf
Skip layer guidance node now works on LTX-Video.
comfyanonymous committed 1 year ago
ab885b3
Some improvements to the lowvram unloading.
comfyanonymous committed 1 year ago
839ed33
Fix some tiled VAE decoding issues with LTX-Video.
comfyanonymous committed 1 year ago
6e8cdcd
LTXV lowvram fixes.
comfyanonymous committed 1 year ago
e5c3f4b
Some fixes to the lowvram system.
comfyanonymous committed 1 year ago
bc6be6c
Remove prints.
comfyanonymous committed 1 year ago
94323a2
Remove print.
comfyanonymous committed 1 year ago
5818f6c
Add LTX-Video support to the Readme.
comfyanonymous committed 1 year ago
0b734de
Support Lightricks LTX-Video model.
comfyanonymous committed 1 year ago
5e16f1d
Fix mask issue in some attention functions.
comfyanonymous committed 1 year ago
2fd9c13
Support new flux model variants.
comfyanonymous committed 1 year ago
8f0009a