The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.
Bump ltxav mem estimation a bit. (#11842)
comfyanonymous1dcbd9efaf16
diff --git a/comfy/supported_models.py b/comfy/supported_models.pyindex d44c0bc37..1bf54f13f 100644--- a/comfy/supported_models.py+++ b/comfy/supported_models.py@@ -845,7 +845,7 @@ class LTXAV(LTXV):def __init__(self, unet_config):super().__init__(unet_config)- self.memory_usage_factor = 0.061 # TODO+ self.memory_usage_factor = 0.077 # TODOdef get_model(self, state_dict, prefix="", device=None):out = model_base.LTXAV(self, device=device)