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
Avoid torch compile graphbreak for older pytorch versions (#9344)
f0d5d01 Xiangxi Guo (Ryan) committed 1 year ago
Make custom node testing checkbox optional in issue templates (#9342)
644b23a filtered committed 1 year ago
Added context window support to core sampling code (#9238)
e4f7ea1 Jedrzej Kosinski committed 1 year ago
Update release workflow to python3.13 pytorch cu129 (#9315)
5ca8e2f comfyanonymous committed 1 year ago
Make torchaudio exception catching less specific (#9309)
898d88e Jedrzej Kosinski committed 1 year ago
Update test release package workflow with python 3.13 cu129. (#9306)
e1d4f36 comfyanonymous committed 1 year ago
fix(Kling Image API Node): do not pass "image_type" when no image (#9271)
f4231a8 Alexander Piskun committed 1 year ago
Bump pytorch cuda and rocm versions in readme instructions. (#9273)
0552de7 comfyanonymous committed 1 year ago
Not sure if AMD actually support fp16 acc but it doesn't crash. (#9258)
5828607 comfyanonymous committed 1 year ago
Users report gfx1201 is buggy on flux with pytorch attention. (#9244)
735bb4b comfyanonymous committed 1 year ago