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.

Closer memory usage factors for PID (#14123)

comfyanonymousd8d860a5883d
diff --git a/comfy/supported_models.py b/comfy/supported_models.py
index 4723caff5..776091d42 100644
--- a/comfy/supported_models.py
+++ b/comfy/supported_models.py
@@ -1214,7 +1214,7 @@ class PixelDiTT2I(supported_models_base.BASE):
}
latent_format = latent_formats.PixelDiTPixel
- memory_usage_factor = 0.18
+ memory_usage_factor = 0.04
supported_inference_dtypes = [torch.bfloat16, torch.float32]
vae_key_prefix = ["vae."]
@@ -1263,7 +1263,7 @@ class PiD(PixelDiTT2I):
"shift": 1.5, # close approximation of the original distill 4 steps [0.999, 0.866, 0.634, 0.342, 0]
}
- memory_usage_factor = 0.07
+ memory_usage_factor = 0.04
def get_model(self, state_dict, prefix="", device=None):
return model_base.PiD(self, device=device)