The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.
Make lowvram more aggressive on low memory machines.
comfyanonymous1aa9cf329249
diff --git a/comfy/model_management.py b/comfy/model_management.pyindex 07c137270..de4bd4426 100644--- a/comfy/model_management.py+++ b/comfy/model_management.py@@ -318,7 +318,7 @@ class LoadedModel:return self.model is other.modeldef minimum_inference_memory():- return (1024 * 1024 * 1024)+ return (1024 * 1024 * 1024) * 1.2def unload_model_clones(model, unload_weights_only=True, force_unload=True):to_unload = []