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
Relax minimum ratio of weights loaded in memory on nvidia.
57f330c comfyanonymous committed 1 year ago
Add that Hunyuan Video and Pixart are supported to readme.
601ff9e comfyanonymous committed 1 year ago
Support preview images embedded in safetensors metadata (#6119)
3cacd3f catboxanon committed 1 year ago
Fix lowvram bug when using a model multiple times in a row.
3ad3248 comfyanonymous committed 1 year ago
Pick attention implementation based on device in llama code.
a4f59bc comfyanonymous committed 1 year ago