The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.
Empty cache after model unloading for normal vram and lower.
comfyanonymous0ae81c03bb03
diff --git a/comfy/model_management.py b/comfy/model_management.pyindex 09dcaa295..e148408b8 100644--- a/comfy/model_management.py+++ b/comfy/model_management.py@@ -238,6 +238,8 @@ def unload_model():current_loaded_model.model_patches_to(current_loaded_model.offload_device)current_loaded_model.unpatch_model()current_loaded_model = None+ if vram_state != VRAMState.HIGH_VRAM:+ soft_empty_cache()if vram_state != VRAMState.HIGH_VRAM:if len(current_gpu_controlnets) > 0: