The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.
Fix memory leak by properly detaching model finalizer (#9979)
Guy Nivc8d2117f02bc
diff --git a/comfy/model_management.py b/comfy/model_management.pyindex d880f1970..c5b817b62 100644--- a/comfy/model_management.py+++ b/comfy/model_management.py@@ -645,7 +645,9 @@ def load_models_gpu(models, memory_required=0, force_patch_weights=False, minimuif loaded_model.model.is_clone(current_loaded_models[i].model):to_unload = [i] + to_unloadfor i in to_unload:- current_loaded_models.pop(i).model.detach(unpatch_all=False)+ model_to_unload = current_loaded_models.pop(i)+ model_to_unload.model.detach(unpatch_all=False)+ model_to_unload.model_finalizer.detach()total_memory_required = {}for loaded_model in models_to_load: