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.

Set default hunyuan refiner shift to 4.0 (#9833)

comfyanonymous7757d5a657cb
diff --git a/comfy/supported_models.py b/comfy/supported_models.py
index ba1b8c313..472ea0ae9 100644
--- a/comfy/supported_models.py
+++ b/comfy/supported_models.py
@@ -1329,7 +1329,7 @@ class HunyuanImage21Refiner(HunyuanVideo):
}
sampling_settings = {
- "shift": 1.0,
+ "shift": 4.0,
}
latent_format = latent_formats.HunyuanImage21Refiner