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
fix(number-convert): preserve int precision for large numbers (#13147)
6580a6b Dante committed 7 months ago
fix(api-nodes): made Reve node price badges more precise (#13154)
a55835f Alexander Piskun committed 7 months ago
Fix Train LoRA crash when training_dtype is "none" with bfloat16 LoRA weights (#13145)
b53b10e Krishna Chaitanya committed 8 months ago
feat(assets): register output files as assets after prompt execution (#12812)
7d5534d Luke Mino-Altherr committed 8 months ago
fix: set default values of Color Adjustment node to zero (#13084)
2d5fd3f Kelly Yang committed 8 months ago
feat: LTX2: Support reference audio (ID-LoRA) (#13111)
e87858e Jukka Seppänen committed 8 months ago
Add warning for users who disable dynamic vram. (#13113)
6265a23 comfyanonymous committed 8 months ago
Fix fp16 intermediates giving different results. (#13100)
11c15d8 comfyanonymous committed 8 months ago
Move inline comfy.context_windows imports to top-level in model_base.py (#13083)
87cda1f Jedrzej Kosinski committed 8 months ago
Make EmptyImage node follow intermediate device/dtype. (#13079)
45d5c83 comfyanonymous committed 8 months ago
Add slice_cond and per-model context window cond resizing (#12645)
589228e drozbay committed 8 months ago
[API Nodes] mark seedream-3-0-t2i and seedance-1-0-lite models as deprecated (#13060)
e4455fd Alexander Piskun committed 8 months ago
Revert "fix: run text encoders on MPS GPU instead of CPU for Apple Silicon (#…" (#13070)
8458ae2 comfyanonymous committed 8 months ago