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
[API Nodes] add support for seedance-1-0-pro-fast model (#10947)
fd271de Alexander Piskun committed 11 months ago
Added "system_prompt" input to Gemini nodes (#11177)
c3c6313 Alexander Piskun committed 11 months ago
chore: replace imports of deprecated V1 classes (#11127)
85c4b4a Alexander Piskun committed 11 months ago
chore(comfy_api): replace absolute imports with relative (#11145)
ec7f651 Alexander Piskun committed 11 months ago
Fix qwen scaled fp8 not working with kandinsky. Make basic t2i wf work. (#11162)
329480d comfyanonymous committed 11 months ago
Speed up lora compute and lower memory usage by doing it in fp16. (#11161)
50ca97e comfyanonymous committed 11 months ago
convert nodes_freelunch.py to the V3 schema (#10904)
117bf3f Alexander Piskun committed 11 months ago
docs: add ComfyUI-Manager documentation and update to v4.0.3b4 (#11133)
bed1267 Dr.Lt.Data committed 11 months ago
Fix regression when text encoder loaded directly on GPU. (#11129)
6fd463a comfyanonymous committed 11 months ago
Make old scaled fp8 format use the new mixed quant ops system. (#11000)
43071e3 comfyanonymous committed 11 months ago
Remove line made unnecessary (and wrong) after transformer_options was added to NextDiT's _forward definition (#11118)
0ec05b1 Jedrzej Kosinski committed 11 months ago
mp: use look-ahead actuals for stream offload VRAM calculation (#11096)
6be85c7 rattus committed 11 months ago
Fix case where text encoders where running on the CPU instead of GPU. (#11095)
ea17add comfyanonymous committed 11 months ago