The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. The fastest local inference engine in the world.
Fix regression in VAE code on old pytorch versions.
comfyanonymous4c5c4ddedaba
diff --git a/comfy/ldm/modules/diffusionmodules/model.py b/comfy/ldm/modules/diffusionmodules/model.pyindex 74322c272..ed1e88212 100644--- a/comfy/ldm/modules/diffusionmodules/model.py+++ b/comfy/ldm/modules/diffusionmodules/model.py@@ -91,7 +91,7 @@ class Upsample(nn.Module):def forward(self, x):scale_factor = self.scale_factor- if not isinstance(scale_factor, tuple):+ if isinstance(scale_factor, (int, float)):scale_factor = (scale_factor,) * (x.ndim - 2)if x.ndim == 5 and scale_factor[0] > 1.0:@@ -109,7 +109,7 @@ class Upsample(nn.Module):else:x = aelse:- x = interpolate_up(x, self.scale_factor)+ x = interpolate_up(x, scale_factor)if self.with_conv:x = self.conv(x)return x