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
implement multi image prompting for gpt-image-1 and fix transparency in outputs (#7763)
2c1d686 thot experiment committed 1 year ago
fp16 accumulation can now be enabled on the stable package. (#7750)
7eaff81 comfyanonymous committed 1 year ago
[API nodes] Add backbone for supporting api nodes in ComfyUI (#7745)
0738e4e Robin Huang committed 1 year ago
Replace aom-av1 with svt-av1 for saving webm videos, use preset 6 + yuv420p10le pixel format (#7736)
92cdc69 Alex Butler committed 1 year ago
Add option for using fp8_e8m0fnu for model weights. (#7733)
2d6805c comfyanonymous committed 1 year ago
Unified Weight Adapter system for better maintainability and future feature of Lora system (#7540)
1f3fba2 Kohaku-Blueleaf committed 1 year ago
Use empty t5 of size 128 for hidream, seems to give closer results.
fd27494 comfyanonymous committed 1 year ago
Hidream: Allow loading hidream text encoders in CLIPLoader and DualCLIPLoader (#7676)
f43e1d7 power88 committed 1 year ago