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
MPDynamic: force load flux img_in weight (Fixes flux1 canny+depth lora crash) (#12446)
c037004 rattus committed 9 months ago
Fix lora Extraction in offload conditions (+ dynamic_vram mode) (#12479)
ecd2a19 rattus committed 9 months ago
feat(api-nodes-Tencent): add ModelTo3DUV, 3DTextureEdit, 3DParts nodes (#12428)
e2c71ce Alexander Piskun committed 9 months ago
Remove unsafe pickle loading code that was used on pytorch older than 2.4 (#12473)
e1ede29 comfyanonymous committed 9 months ago
Add working Qwen 2512 ControlNet (Fun ControlNet) support (#12359)
dc9822b krigeta committed 9 months ago
Support generating attention masks for left padded text encoders. (#12454)
831351a comfyanonymous committed 9 months ago
Use torch RMSNorm for flux models and refactor hunyuan video code. (#12432)
e1add56 comfyanonymous committed 9 months ago
Update command to install AMD stable linux pytorch. (#12437)
e03fe8b comfyanonymous committed 9 months ago
fix(api-nodes): add separate retry budget for 429 rate limit responses (#12421)
4a93a62 Alexander Piskun committed 9 months ago
dynamic_vram: Fix windows Aimdo crash + Fix LLM performance (#12408)
d297a74 rattus committed 9 months ago
Dispatch desktop auto-bump when a ComfyUI release is published (#12398)
4993411 Benjamin Lu committed 9 months ago
fix(api-nodes): retry on connection errors during polling instead of aborting (#12393)
2c7cef4 Alexander Piskun committed 9 months ago