Skip to content

The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.

Lower the default shift of lumina to reduce artifacts.

comfyanonymous8ac2dddeed27
diff --git a/comfy/supported_models.py b/comfy/supported_models.py
index 7aa152480..cdd2ba574 100644
--- a/comfy/supported_models.py
+++ b/comfy/supported_models.py
@@ -873,7 +873,7 @@ class Lumina2(supported_models_base.BASE):
sampling_settings = {
"multiplier": 1.0,
- "shift": 6.0,
+ "shift": 3.0,
}
memory_usage_factor = 1.2